Ruthenium Polypyridyl Complex Bound to a Unimolecular Chair-Form G-Quadruplex

The DNA G-quadruplex is known for forming a range of topologies and for the observed lability of the assembly, consistent with its transient formation in live cells. The stabilization of a particular topology by a small molecule is of great importance for therapeutic applications. Here, we show that the ruthenium complex Λ-[Ru(phen)2(qdppz)]2+ displays enantiospecific G-quadruplex binding. It crystallized in 1:1 stoichiometry with a modified human telomeric G-quadruplex sequence, GGGTTAGGGTTAGGGTTTGGG (htel21T18), in an antiparallel chair topology, the first structurally characterized example of ligand binding to this topology. The lambda complex is bound in an intercalation cavity created by a terminal G-quartet and the central narrow lateral loop formed by T10–T11–A12. The two remaining wide lateral loops are linked through a third K+ ion at the other end of the G-quartet stack, which also coordinates three thymine residues. In a comparative ligand-binding study, we showed, using a Klenow fragment assay, that this complex is the strongest observed inhibitor of replication, both using the native human telomeric sequence and the modified sequence used in this work.

Oligonucleotides were purchased from Eurogentec Ltd. as double HPLC-purified solids and were used without further purification. Unless otherwise stated, all other materials and chemicals were sourced from Sigma Aldrich or Honeywell research chemicals. Sephadex C-25 anion exchange stationary phase and Dowex 1X2 Chloride form anion exchange resin were purchased from GE Healthcare. All solvents, unless stated in the experimental, were obtained at HPLC grade and used without further purification. Where further purification was needed, protocol from "Purification of Laboratory Chemicals, 4 th edition, Armarego et. al." was followed. Deuterated solvents for NMR analysis were purchased either through Sigma-Aldrich or Cambridge Isotope Laboratories.

Instrumentation
Unless otherwise stated, all 1 H NMR spectra were collected on a Bruker Nanobay 400 MHz instrument, with the majority of 13 C NMR spectra collected on a Bruker DPX 400.1 MHz machine operating at 100.1 MHz. Both machines were calibrated against a tetramethylsilane (TMS) internal standard and have two channels running TOPSPIN 2.4 and ICON NMR 4.2. All J-coupling constants were reported following normalisation against the used Larmor frequency, where a few couplings were omitted subject to spectral resolution.
High resolution ESI mass spectra were recorded on a Thermo Scientific LTQ Orbitrap XL running in positive ion mode. Fragmented Ions were detected on an Orbitrap Ion trap photodiode array detector and were determined via peak matching against the internally calibrated lock mass for Diisooctyl phthalate (m/z = 413.26623 Da). Data analysis was performed on the Xcalibur Qual Browser software package and all accurate masses are reported within 3 ppm.

Synthesis
Phendione (50 mg, 0.238 mmol) and 1,2-diaminoanthraquinone (56.7 mg, 0.238 mmol) were both suspended together in an ethanolic solution (7 mL) containing a trace amount of ρ-toluenesulfonic acid within a CEM microwave tube (10 mL). The violet coloured solution was degassed/evacuated with Ar for 15 minutes before being fully sealed and installed into the synthetic microwave. The sample was irradiated with 150 W at 140 o C for 20 minutes, yielding a deep red/violet solution which was ensured to be cool and then filtered in vacuo to collect the black precipitate. The powder was suspended in hot chloroform (100 mL) in the presence of powdered charcoal and filtered through a glass frit, yielding a brown solution. The filtrate was reduced to approximately 5 mL in volume (mixture of purple and yellow coloured precipitation is noted) and diethyl ether (50 mL) was added to complete the precipitation. The powder was collected via suction filtration and washed with diethyl ether (3 x 10 mL), and the target product as yellow-ochre powder (52 mg, 0.129 mmol, 54 %). Ru(phen)2Cl2 (81 mg, 0.15 mmol) and qdppz (62 mg, 0.15 mmol) were both suspended together in an aqueous ethanol solution (7 mL, 1:1) within a CEM microwave tube (10 mL). The violet coloured solution was degassed/evacuated with Ar for 15 minutes before being fully sealed and installed into the microwave. The sample was irradiated at 140W at 60 o C for 40 minutes, yielding a deep red/brown solution which was cooled and then filtered via suction filtration. Subsequent precipitation of the target compound from the filtrate was achieved by metathesis via dropwise addition of a saturated solution of aqueous potassium hexafluorophosphate (KPF6). Isolation of the PF6salt by in vacuo filtration yielded a dark orange/brown solid, which, after washing with cold water (2 x 2 mL) was allowed to dry in air. The complex was then further purified via flash chromatography on silica using 80:15:5 CH3CN/H20/KNO3 as the mobile phase (eluting as a deep orange/red band). The collected aliquots were combined and desalted via metathesis to the hexafluorophosphate salt and finally treated with Amberlite resin (IRA-400, Clform, 2.4 g) after being dissolved in a minimum amount of 2:1 H2O/ACN, to yield the complex as a dark red microcrystalline solid (68. 2 mg, 0.072 mmol, 62 %).

Enantiomeric Resolution
Semi-preparative chiral resolution of the racemic complexes, [Ru(phen)2(dppz)] 2+ and [Ru(phen)2(qdppz)] 2+ was achieved using a Hitachi Primeaide HPLC arrangement equipped with a CF6 LARIHC cyclofructan based chiral column (internal dimensions; 10 x 250 mm) supplied by AZYP separations LLC (Arlington, Texas, US). Successful baseline separations were achieved using a MeOH:ACN:TEA:AA mobile phase; with the anthraquinone enantiomers separated using a ratio of 60:40:2:0.8, whereas the phenanthroline parent complex separated at a ratio of 60:40:4:1.6. Analysis was performed at a flow rate of 5 mL min -1 where each preparative injection was 200 μL in volume with an analyte concentration of 3 mg mL -1 . Eluent fractions were combined in centrifuge tubes and to each, 2 mL of a 200 mM aqueous solution of potassium hexafluorophosphate was added. The solutions were then reduced under pressure using a DNA concentrator at 40 o C for 24 hrs to remove the organic eluent. Following metathesis, the precipitate was collected via suction filtration and washed with fractions of HPLC water (5 x 5 mL). The water-soluble chloride salts were then prepared by dissolving the enantiomers in 60:40 H2O:ACN and stirring the solutions overnight in the presence of excess Amberlite IRA-400 anion exchange resin (Clform). Complex purity was verified by electrospray mass spectrometry (ES-MS); optical purity was first confirmed by analytical chiral HPLC to quantify the quality of the enantiomeric excess of each work up (in each case EE was >99 % by peak area). Circular dichroism of the aqueous solutions ( Figure S2) confirmed the success of the separations.

Crystallisation Parameters
Crystals suitable for X-ray diffraction experiments were obtained by the vapour diffusion of diethyl ether into a saturated solution of Λ-[Ru(phen)2(qdppz)]·Cl2 in acetonitrile. Dark Red/Brown Rods grew after approximately 2 weeks at 277 K.

Data Collection and Structure Solution
Data collection was performed on a Rigaku Synergy-S diffractometer with images collected on a HyPix-6000 pixel detector. Cu Kα radiation (λ = 1.5418 Å) was used on a crystal fragment with approximate dimensions 50x50x200 μm that was cooled to 100 K using a stream of nitrogen gas. Data collection, reduction, and processing was performed using the CrysAlisPro program package. SHELXT was utilised in the Olex2 software suite to solve the structures and they were refined against F2 using a full-matrix least-squares minimisation in SHELXL, All non-hydrogen atoms were refined anisotropically, and hydrogen atoms were first placed in calculated positions and refined using a riding model. Data collection and final refinement statistics are shown in Table S1. An Ortep model is shown in Figure S3, with selected bond lengths and angles given in Table S2, and unit cell contents are shown in Figure S4. Experimental data and refined structures were uploaded to the IUCr checkCIF server and no warnings above level C were reported. The structure is deposited in the CCDC CSD under the refcode 2090875.

Crystallisation Parameters
Crystals containing the oligonucleotide d(GGGTTAGGGTTAGGGTTTGGG) and the ruthenium complex [Ru(phen)2(qdppz)] 2+ were grown from sitting drops via vapour diffusion of water at 277 K. Crystals suitable for X-ray diffraction experiments were obtained from a solution formed by the addition of following two constitutions; 1µL of a pre-annealed mixture of the single stranded oligonucleotide at 200 μM with the complex Λ-[Ru(phen)2(qdppz)]·Cl2 at 200 µM in a 10 mM KCl buffer; and 1 µL of a solution containing 80 mM sodium chloride, 20 mM barium chloride dihydrate, 12 mM spermine tetrahydrochloride, and 40 % v/v 2-methyl-2, 4-pentanediol (MPD); all buffered to pH 7.0 using 40 mM sodium cacodylate trihydrate. The sitting drop was equilibrated against 100 µL of 40 % v/v MPD in H2O forming dark orange/red hexagonal rods within 4 days of preparation.

Data Collection and Structure Solution
The data were collected at Diamond Light Source Ltd., on beamline I03 using radiation with a wavelength of 0.5604 Å from a flash cooled crystal at 100 K. 360 o of data were collected with an oscillation of 0.1 o per frame, generating 3600 images. The resulting data were processed using XDS 1 and XSCALE 2 through the xia2 3 pipeline, finding 11038 unique reflections to a resolution of 1.44 Å and an overall I/σ of 10.9. A separate dataset collected on the same crystal using radiation with a wavelength of 1.771 Å was used to build an initial substructure using the anomalous scattering of ruthenium/barium by single wavelength anomalous dispersion (SAD) methodology using the Phaser-EP pipeline in the PHENIX software package. 4,5 This substructure was used to calculate an initial phasing solution for the higher resolution dataset. The crystallographic model was built using WinCoot 6 and refined using Phenix.refine 7 to give a final Rwork of 0.1648 and an Rfree of 0.1810 reserving 10 % of the total reflections for the Rfree set. Table S3 highlights the data collection(s) and refinement statistics; the structure is deposited in the Protein Data Bank with PDB accession ID: 7OTB. Full DNA conformational analysis was conducted using a combination of programs, namely the DNATCO v3.2 and w3DNA servers. 8,9 Figures were produced using the PyMOL software suite.

Klenow fragment (exo -) preparation
The Klenow Fragment (KF) encoding gene was amplified from E. coli genomic DNA (JM109) via polymerase chain reaction (PCR). PCR was proceeded using PrimeSTAR DNA polymerase (Takara Bio) and primers (5'-GGGACCATATGGTGATTTCTTATGACAACTACG-3' and 5'-GGGAGAATTCTTAGT-GCGCCTGATCCCAG-3') sourced from Eurofin Genomics (Japan). Following digestion using Ndel and EcoRI, the cloned DNA fragments were cloned into pMal-p5x vector (New England Bio Labs). KF exo -(D355A, E357A) was prepared by mutating the constructed plasmid using a QuikChange mutagenesis kit (Stratagene) followed by using the mutated vector to transform E. coli EG2523 (New England Bio Labs). The cells were then cultured in Luria-Bertani medium containing ampicillin and worked up to an A600 of roughly 0.5 followed by addition of isopropyl β-D-1-thiogalactopyranoside (IPTG). The cultured cells were harvested and lysed via sonication and the soluble fraction was loaded onto an amylose resin packed column (New England Bio Labs). Following treatment with Factor Xa protease to eliminate the MBP-tag, the KF exo-was purified over a Hitrap Heparin column followed by purification through a Hiload Superdex 200 (GE Healthcare). The purified KF exo-was dialysed against a dialysis buffer containing 50 mM Tris-HCl (pH 7.2), 1 mM EDTA, 1 mM DTT, 100 mM NaCl and 50 % glycerol.
Concentration was determined spectrophotometrically using an extinction coefficient of 58,790 M -1 cm -1 at 280 nm. The mutant enzyme was stored at -30 o C until use.

Klenow fragment replication assay
Template strand (5'-TTAGGGTTAGGGTTAGGGTTAGGGTTTTTTTTTTTTTTTCCTATAGTGAGTCGT-ATTACCC-3' or 5'-GGGTTAGGGTTAGGGTTTGGGTTTTTTTTTTTTTTTCCTATAGTGAGTCGTATTACCC-3') and primer strand (5'-FAM-GGGTAATACGACTCACTATAGG-3') oligonucleotides were annealed in the following buffer and in the presence, if quoted, of the necessary ligand: 10 mM Tris-HCl (pH 7.5), 8 mM MgCl2, 1 µM FAM-labelled primer DNA, 1 µM template DNA, 10 µM ligand, 250 µM dNTPs and 10 mM KCl (or LiCl where indicated). After annealing (95 o C to 4 o C at -1 o C min -1 ), the mixture was incubated at 37 o C and a 100 µM solution of KF exo-was added to the reaction mixture (final concentration of 1 µM) to initiate the enzymatic reaction. At given time intervals during the reaction, aliquots of the mixture were quenched using a stopping solution containing 10 mM EDTA and 80 wt% formamide. Reaction products were separated by denaturing polyacrylamide gel electrophoresis (PAGE) using gel containing 8 M urea in a TBE buffer at 70 o C for 1 hour at 200 V. Alongside product lanes, a molecular weight marker DNA ladder (10-bp) and a bromophenol blue running aid were run in adjacent and terminal lanes. Gel images were captured using a Fujifilm Fluoreimager FLA-5100 utilising a laser excitation wavelength of 473 nm. Images were collected before and after staining with SYBR Gold (ThermoFisher Scientific) to highlight unlabelled products. Band intensities were analysed using ImageJ2 software package (NIH) by quantifying peak areas after baselining the necessary lanes. The reaction yield of full-length product (P) was quantified by calculating the ratio of intensity of the fulllength product bands to the aggregate intensity of all bands. Reaction rate analysis was performed using Dynafit software package (Biokin) after evaluating a global fit of the reaction. This was achieved by applying a kinetic model to the following two-step sequential model;

→ →
Where P0 is the starting state of the reaction, Ps represents the state immediately after unwinding of the reaction stall (motif), Pf represents the state after the replication of the full-length product is completed; and ks and kf are the rate constants that define the rate of reaction between states.

Circular Dichroism
Circular dichroism spectra were collected at 37 o C on a JASCO J-1500 CD Spectrophotometer running Spectromanager in quartz cells with path length of 1 cm. CD samples were measured at 10 µM DNA, 10 mM Tris-HCl (pH 7.5), and 10 µM of respective ligand. Samples were annealed from 95 o C to 20 o C at a rate of -1.0 o C min -1 . To determine the stability of the G-quadruplex, the temperature 10 μM DNA with or without 10 μM ligand was added to the buffer consisting of 10 mM Tris-HCl (pH 7.5) and 10 mM KCl. The melting analyses were performed using a JASCO J-1500 equipped with a temperature control system. Before measurement, samples were annealed by cooling from 90 to 4 °C at −1.0 °C min -1 . The annealed samples were transferred into cuvettes and then the CD signal at 288 nm was collected with increasing the temperature from 4 to 95 °C at 0.5 °C min -1 . To determine thermodynamic parameters, the CD melting curves were normalized and analysed by curve fitting based on the twostate model in Kaleida Graph (Synergy Software) as per previous literature 10 . 6

Supplementary Figures and Tables
2.1. Figure S1 -1 H NMR Spectra  Table S1 -Data collection and processing parameters/refinement results for the small molecule dataset   Atom colours: carbon -light/dark pink, nitrogen -blue, ruthenium -teal, oxygen -red, and chlorinegreen. Table S3 -Collection and processing statistics of 7OTB  Figure S5 -Asymmetric unit of 7OTB 2.10 Figure S6 -Ligand-adenine close contact