NMR analyses on N-hydroxymethylated nucleobases – implications for formaldehyde toxicity and nucleic acid demethylases

NMR studies reveal that formaldehyde, a toxic pollutant and metabolite, reacts with nucleotides to form N-hydroxymethylated adducts of varying stabilities.


NMR Experiments
NMR experiments were carried out using either a Bruker AVII 500 spectrometer equipped with a TXI probe, a Bruker AVIII 600 spectrometer equipped with a Prodigy N2 broadband cryoprobe, or a Bruker AVIII 700 spectrometer equipped with an inverse TCI 1 H/ 13 C/ 15 N cryoprobe. All spectrometers were operated using TOPSPIN 3 software. 1 H chemical shifts are reported in ppm relative to the solvent resonance (δH 4.7 ppm), while signal intensities were calibrated relative to 3-(trimethylsilyl)-2,2,3,3-tetradeuteropropionic acid (TSP, δH 0 ppm), which was added to each sample. The deuterium signal of D2O was used as an internal lock signal.
Samples containing nucleotides and HCHO were prepared by mixing stocks of nucleotides in D2O (pD adjusted) with HCHO in D2O. TSP was added before transferring to either 3 mm or 5 mm NMR tubes for analysis by 1 H NMR. The total lapse time between mixing and data acquisition was 5-7 minutes. Concentrations of products were quantified relative to the concentration of added TSP (5.  Figure S4 (2-25-fold dilutions) were carried out by pre-incubating stock solutions of nucleotides (10 mM) and HCHO (8 equivalents) for one week, before splitting the stock solutions and diluting with the appropriate quantities of D2O.
EXSY analyses were carried out using a 1-dimensional gradient-selected NOESY pulse sequence employing selective refocusing with a Gaussian pulse. [3][4][5] Experiments were run accumulating 16 transients with mixing times (τm) of 10-1200 ms, with the 1 H-resonance of hydrated HCHO (δH 4.89 ppm at 37 °C) being selectively irradiated. Adduct formation rates were calculated as follows: the normalised intensities of the EXSY correlations (i.e. the intensities of the 1 H-resonances corresponding to the adduct N-hydroxymethyl protons normalised to the intensity of the irradiated HCHO resonance) were plotted as a function of τm, and the initial intensity build-up rates were determined (see Figure S11). Assuming a bimolecular mechanism for adduct formation, the build-up rates represent k1[nucleotideeq], where k1 is the rate constant for adduct formation, and [nucleotideeq] is the concentration of unreacted nucleotide at equilibrium (see below). The initial adduct formation rate for 3hmUMP and 3hmTMP were calculated by dividing their accumulation rate by [ FTO catalysis was monitored using both 1 H NMR and the gradient-selected 1 dimensional heteronuclear single quantum correlation ( 1 H-13 C-HSQC) method. Samples were prepared containing FTO (prepared as reported 1 , with a final concentration of 20 µM), the nucleoside selectively 13 C-labelled of the methyl group (400 µM), 2-oxoglutarate (2OG, 5 mM), sodium ascorbate (1 mM), ferrous iron (20 µM) in ammonium formate buffer in D2O pH* 7.5. The samples were then transferred to 3 mM MATCH NMR tubes (Hilgenberg) and monitored by NMR. 1 H analyses employed NOESY water presaturation, while the 1 dimensional heteronuclear single quantum correlation (1D-1 H-13 C-HSQC) method was derived from the standard 2D 1 H-13 C-HSQC pulse sequence to remove both the variable t1 period and 13 C decoupling during data acquisition. The 1/2JCH delays were optimised for 145 Hz.

MS-Based FTO Activity Assay
A reaction mixture containing RNA oligonucleotide (AUUGUGG-m6A-CUGCAGC, 1 µM), 2OG (10 µM), ascorbate (100 µM), ferrous iron (10 µM) and FTO (100 nM) in 50 mM Tris buffer in H2O at pH 7.5 was prepared in a 2 mL 96-well plate (Greiner) and reaction progression was monitored by MS using an Agilent RapidFire RF360 high throughput system paired with an Agilent quad time of flight (Q-TOF) mass spectrometer. Aliquots from the mixture were periodically subjected to MS analysis (at 1 minute intervals over the first 14 minutes after mixing, then at 23 minutes and 33 minute after mixing). The sample was passed through an Agilent C8 RapidFire cartridge, which isolated the oligonucleotides; the oligonucleotides were then eluted from the cartridge with 600mM octyl ammonium acetate (OAA) (20 %) and acetonitrile (80 %), and injected into the spectrometer. The RapidFire RF360 system was operated using RapidFire RF360 integrated software, and the mass spectrometer was operated using Agilent MassHunter Workstation Data Acquisition software. Signal intensities were quantified as the total ion count, and analysed using RapidFire integrator software (Agilent). Detailed RapidFire-MS procedures will be published elsewhere.

General Methods
All chemicals, including dried solvents, were from Sigma-Aldrich and used without further purification. Solvents used for work-up and chromatography were from Aldrich at HPLC grade. Silica gel 60 F254 analytical thin layer chromatography (TLC) plates were from Merck. Prepacked SNAP columns were used for chromatography on a Biotage SP1 Purification system. Proton and Carbon NMR spectra were acquired using AVIIIHD 500 or Bruker AVIIIHD 400 or AVIIIHD 600 with N2 cryoprobe. Shifts are reported in δ ppm. Abbreviations s, d, t, q, and m denoting singlet, doublet, triplet, quartet and multiplet respectively used in 1 H NMR. Coupling constants, J, are registered in Hz to a resolution of 0.5 Hz. High Resolution (HR) mass spectrometry data (m/z) were obtained from a Bruker MicroTOF instrument using an ESI source and Time of Flight (TOF) analyzer. Values are reported as ratio of mass to charge in Daltons. Melting points were obtained using a Leica VMTG heated-stage microscope or Stuart SMP-40 automatic melting point apparatus. All reactions were carried out in an oven-dried round bottom flask. A magnetic stirrer was used to ensure homogenous mixing during the reaction. Synthesis of oligonucleotides were carried out as reported. 6

3', 5'-O-Bis(t-butyldimethylsilyl)thymidine
This compound was synthesised using a reported procedure. 7 To a stirred solution of 2'deoxythymidine (2 g, 8.3 mmol) and imidazole (2.25 g, 33.026 mmol) in dry DMF in a 50 mL round bottom flask was added TBDMSCl (2.74 g, 18.16 mmol) portionwise. The reaction was stirred for four hours. The resulting mixture was concentrated under reduced pressure, then diluted with EtOAc. The EtOAc layer was washed with water (3 x 10 mL), and finally with brine. The organic layer was dried over MgSO4 and concentrated in vacuo. Mixture was purified (19:1 to 4:1 cyclohexane / EtOAc) by column chromatography which resulted in an off-white solid

(3-( 13 C)-Methyl)thymidine
This compound was prepared according to a modified version of a reported procedure. 8

3',5'-O-Bis(t-butylsilyl)-2'-O-(t-butyldimethylsilyl)adenosine
This compound was prepared according to a modified version of a reported procedure. 9 To a stirred suspension of adenosine (2.12 g, 8 mmol) in 40 mL anhydrous DMF at 0 °C, di-tbutylsilyl ditrifluoromethanesulfonate (3.87 mL, 8.8 mmol) was added drop wise under an N2 atmosphere. After consumption of starting material (30 min, as assessed by TLC), the reaction was quenched immediately with imidazole (2.7 g, 40 mmol) at 0 °C. After 5 minutes, the reaction was warmed to room temperature. Then, t-butyldimethylsilyl chloride (1.45 g, 9.6 mmol) was added portion wise and the reaction was refluxed at 60 °C for 12 hr. The suspension was cooled down to room temperature, water was added and the precipitate was collected by suction filtration. The filtrate was discarded, and the white precipitate was washed with cold MeOH. The MeOH layer was evaporated under reduced pressure and the product was crystallised from CH2Cl2 to give a white solid

(6-( 13 C)-Methyl)adenosine
This compound was prepared according to a modified version of a reported procedure. 8