DNA with zwitterionic and negatively charged phosphate modifications: Formation of DNA triplexes, duplexes and cell uptake studies

Two phosphate modifications were introduced into the DNA backbone using the Staudinger reaction between the 3’,5’-dinucleoside β-cyanoethyl phosphite triester formed during DNA synthesis and sulfonyl azides, 4-(azidosulfonyl)-N,N,N-trimethylbutan-1-aminium iodide (N+ azide) or p-toluenesulfonyl (tosyl or Ts) azide, to provide either a zwitterionic phosphoramidate with N+ modification or a negatively charged phosphoramidate for Ts modification in the DNA sequence. The incorporation of these N+ and Ts modifications led to the formation of thermally stable parallel DNA triplexes, regardless of the number of modifications incorporated into the oligodeoxynucleotides (ONs). For both N+ and Ts-modified ONs, the antiparallel duplexes formed with complementary RNA were more stable than those formed with complementary DNA (except for ONs with modification in the middle of the sequence). Additionally, the incorporation of N+ modifications led to the formation of duplexes with a thermal stability that was less dependent on the ionic strength than native DNA duplexes. The thermodynamic analysis of the melting curves revealed that it is the reduction in unfavourable entropy, despite the decrease in favourable enthalpy, which is responsible for the stabilisation of duplexes with N+ modification. N+ONs also demonstrated greater resistance to nuclease digestion by snake venom phosphodiesterase I than the corresponding Ts-ONs. Cell uptake studies showed that Ts-ONs can enter the nucleus of mouse fibroblast NIH3T3 cells without any transfection reagent, whereas, N+ONs remain concentrated in vesicles within the cytoplasm. These results indicate that both N+ and Ts-modified ONs are promising for various in vivo applications.


Synthesis and purification of chemically modified ONs
4-(Azidosulfonyl)-N,N,N-trimethylbutan-1-aminium iodide [1] and tosyl azide (p-toluenesulfonyl azide, TsN3) [2] were synthesised as described and used for the synthesis of the modified ONs. One should note that TsN3 should be handled with caution as it has the shock sensitivity of tetryl (N-methyl-N-2,4,6-tetranitroaniline) and the explosiveness of TNT [3] . Unmodified oligonucleotides (ONs) were purchased from Integrated DNA Technologies (IDT, Singapore). The N+ and Ts-modified ONs were synthesised using a Mermaid-4 automated DNA synthesiser (BioAutomation Corp.) using 5-ethylthio-1H-tetrazole (ETT) as an activator. Oxidation and deprotection times were 60 s (repeat 3 times) and coupling time 90 s for 5 μmol synthesis scale. When the modifications were incorporated, the automatic synthesis was paused before the capping step (after coupling step and wash), the column was taken out from the DNA synthesiser and connected to a micro tube pump. For N+ modification, 4-(azidosulfonyl)-N,N,N-trimethylbutan-1-aminium iodide in DMF (saturated and degassed, 2 mL) was pumped through the column for 30 min at 37 °C. For the Ts-modification, TsN3 in MeCN (0.5 M, 2 mL) was pumped through the column for 30 min at 37 °C. The column was placed back in the synthesiser to continue the ON synthesis. The resulting ONs were cleaved from the solid support, and deprotected with concentrated aqueous ammonia (≈ 28%) at 55 °C for 12 h. The ONs were purified by HPLC using IE-column (TSKgel Super Q-5PW). Buffer A (20 mM Tris-HCl, 1 mM Na2-EDTA, pH 9.0), buffer B (20 mM Tris-HCl, 1 mM Na2-EDTA, 1 M NaCl, pH 9.0). Gradients: 3.7 min 100% A, convex curve gradient to 30% B in 7.3 min, linear gradient to 50% B in 18.5 min, concave gradient to 100% B in 7.4 min, keep 100% B for 7.4 min and then 100% A in 7.4 min. Collected individual UV-absorbing fractions (λ = 260 nm) were desalted using a NAP-25 column. The composition of each fraction was confirmed by ESIMS (Table 1 and Figures S2-S5), and the fractions containing the desired modified ONs were identified and freeze-dried.
The purity of the chemically modified ONs was confirmed using 20% denaturing gel ( Figure S1). Gels were prepared in TBE buffer (pH 8.0) in 7 M urea with 0.5 mm thickness, 17.5 × 14.5 cm 2 (19:1 acrylamide/bisacrylamide ratio). Samples were incubated at high temperature in 7 M urea (7.5 µL) to disrupt high order assemblies. TBE buffer (pH 8.0) was used as a running buffer. Gel electrophoresis was performed at 15 W per gel at 4 °C to avoid overheat. After the electrophoresis, gels were stained with 5% Stains-All® solution in water/formamide 1: 1 for 5-10 min and then destained in H2O until complete fading of the dye from the gel background.

ESIMS analysis of the modified ONs
ESIMS spectra were recorded using a Thermo Scientific Q-Exactive Focus Hybrid Quadrupole-Orbitrap Mass Spectrometer. The modified ONs were prepared into 4 μM strand concentration with water/methanol 4:1 (HPLC grade from Fisher Scientific). Samples (5 µL) were injected via a Dionex Ultimate 3000 HPLC system running at 0.1 mL/min CH3OH.  Table 1.  Table 1.

Determination of the melting temperature of duplexes and triplexes using UV-vis spectrometry
Parallel and antiparallel duplexes were formed by mixing two strands (each at a concentration of 1.0 µM) in the corresponding buffer solution as listed in Table 1. Each solution was heated to 80 °C for 5 min and cooled to room temperature. Triplexes were formed by first mixing the two strands of the Watson-Crick duplex, each at a concentration of 1.0 µM in the corresponding buffer solution. The solution was heated to 80 °C for 5 min and cooled to room temperature, and the third strand (TFO) was added and then kept at 15 °C for at least 30 min. Melting temperature measurements were performed on a Cary 100Bio UV-vis spectrometer using quartz cuvettes with 10 mm pathlengths and a 2 × 6 multicell block with a Peltier temperature controller. The melting temperature (Tm, °C) for parallel and antiparallel duplexes was determined as the maximum of the first derivative plots of the melting curves obtained by measuring the absorbance at 260 nm against increasing temperature (0.5 °C per min). The Tm values for parallel ONs/DNA duplexes is presented in Table S1.
For parallel triplex at pH 5.0 and pH 6.0, a melting curve with two transition states ( Figure S6C, blue line) was obtained due to the triplex and duplex melting curves were overlaid. A melting triplex profile was obtained by subtracting the melting curve of duplex D1 ( Figure S6C, black line) from the triplex melting curve. The Tm value for a parallel triplex was determined as a cross point of the obtained triplex melting curve ( Figure S6C, red line) with its median of the upper and lower baselines ( Figure S6C).
All melting temperatures are within the uncertainty of ± 0.5 °C as determined by repetitive experiments.  Table S1. Tm [°C] Data for parallel duplex melting, taken from UV melting curves (λ = 260 nm).

Determination of melting temperature using CD measurements
To confirm the Tm values of antiparallel duplexes formed by ONs/RNA obtained from UV-vis denaturation experiment, CD denaturation experiments were conducted using the same buffer and DNA concentrations as for the UV-vis measurements. The CD spectra were recorded using a Chirascan CD spectrophotometer (150 W Xe arc) from Applied Photophysics with a Quantum Northwest TC125 temperature controller. The CD spectra (average of at least 3 scans) were recorded between 220 and 350 nm with 1 nm intervals, 120 nm/min scan rate and 10 mm path length followed by subtraction of a background spectrum (buffer only). CD denaturation and renaturation experiments were performed by recording spectra every 2.5 °C with equilibration for 2.5 min at each temperature from 10 to 75 °C ( Figure S7A). The comparison of Tm values obtained from UV-vis and CD melting experiments is shown in Table   S2.

Determination of thermodynamic parameters of antiparallel duplexes formed at different salt concentrations
In order to analyse thermodynamic parameters of N + and Ts-modified ONs toward complementary DNA and RNA at different salt concentrations, melting profiles obtained from UV melting experiment were converted into a fraction folded (Ɵ) vs temperature representation ( Figure S9A). The signal change at 260 nm was extracted, converted to fraction folded Ɵ and plotted against temperature to give the Tm value ( Figure S7B). The conversion of UV signal at 260 nm (CT) to Ɵ T at a given temperature was followed by adequate upper and lower baselines chosen: Ɵ T = (L0T -CT) / (L0T -L1T) (Equation 1) L0T and L1T correspond to the baseline values of the unfolded and folded species, respectively. Ɵ is a number between 0 and 1: Ɵ = 0 for T>>Tm, Ɵ = 1 for T<<Tm, and Ɵ = 0.5 for T = Tm.
By definition, the free Gibbs enthalpy may be written as: △G 0 = -RT ln(Ka) = △H 0 -T × △S 0 (Equation 2) Where R = 8.3145 J/(K·mol), T is the temperature in Kelvin, △H 0 is the standard enthalpy of the reaction, and △S 0 is the standard entropy, assuming that △cp = 0 [4] Equation 2 can be deduced as: ln(Ka) = -△H 0 / R × (1/T) + △S 0 / R (Equation 3) Therefore, the following step required a van't Hoff plot of the natural logarithm of the affinity constant (ln(Ka)) as a function of the reciprocal of the temperature (1/T in K −1 ) [5] . For bimolecular equilibrium A + B  C: S14

Ka = [C] / [A] [B] = (Cc × Ɵ) / ([CA× (1 -Ɵ) ] × [ CB × (1 -Ɵ) ]) (Equation 4) When A and B is present at the same initial strand concentration C0 Ka = Ɵ / (C0 × (1 -Ɵ) 2 ) (Equation 5)
where C0 is the initial strand concentration and Ɵ is ƟT at each temperature. It should be noted that the analysis should be restricted between the temperature range for which 0.03 < Ɵ < 0.97 as it is relatively difficult to evaluate the affinity constant when almost all or almost none of the molecules are associated [6] .
Following the calculations described above, Figure S9A was converted into Figure S9B. The van't Hoff relation (ln(Ka) vs. 1/T) should give a straight line, with a slope of −△H 0 /R and the y-axis intercept of △S 0 /R. The same procedure was used for determination of thermodynamic parameters for all complexes studied in Table 3.

Evaluation of triplex formation using size-exclusion HPLC
To confirm triplex formation, we performed size-exclusion (SE) HPLC evaluation of several samples at pH 50 and 6.0 at rt (25 °C). Preformed complex samples (1 µM, 100 µL) were analysed on an Ultimate 3000 HPLC system, equipped with an autosampler, a diode array detector detecting absorbance at 260 nm and a Thermo Acclaim SEC-300 column (4.6 × 300 mm; 5-μm hydrophilic polymethacrylate resin spherical particles, 300 Å pore size). 10 mM Na-cacodylate buffer (pH 5.0 and pH 6.0, respectively) supplemented with 100 mM NaCl and 10 mM MgCl2 was used as a mobile phase. As shown in Figure S15B, samples of D1/ON1, D1/m-Ts-ON9 and D1/3Ts-ON12 at pH 6.0 showed no triplex formation at rt (25 °C

Enzymatic digestion of ONs by phosphodiesterase Ⅰ
The nuclease stability of the modified ONs was evaluated using snake venom phosphodiesterase (phosphodiesterase I, Sigma, made into 0.8 units /mL stock) and compared with unmodified sequence ON1. N+ONs, Ts-ONs and ON1 at 7.5 µM were incubated with phosphodiesterase Ⅰ (0.16 units /mL, 12 µL) in 60 µL of 5 mM Tris·HCl buffer (pH 8.0, 10 mM MgCl2) at 37 °C. 10 µL aliquots were collected at 0, 10, 30, 60 and 120 min, heated at 90 °C for 5 min to deactivate the enzyme and analysed by SE-HPLC to evaluate the amount of intact ONs remaining ( Figure S17). The percentage of intact ONs in each sample was calculated and plotted against the digestion time (