Acrylamide-dT: a polymerisable nucleoside for DNA incorporation

The synthesis of a novel modified nucleoside phosphoramidite, Acrylamide-dT-CE phosphoramidite, obtained in three steps from commercially available starting materials, is reported. It was readily incorporated into thrombin binding aptamer (TBA) sequences using automated solid-phase synthesis under ultra-mild conditions, with the modification shown not to adversely affect duplex stability, G-quadruplex structure, or thrombin binding. The reaction and integration of the modified strands with acrylamide polymers was evidenced by gel electrophoresis. The Acrylamide-dT functional handle promises to be an ideal synthon for preparing DNA–polymer hybrids for use in various macromolecular materials applications.


Introduction
Nucleobase modications are commonplace for enhancing the properties of oligonucleotides for therapeutic and diagnostic applications. Such modications are typically employed for improving duplex stability, 1 increasing binding affinity and selectivity towards certain targets 2,3 or introducing detection elements such as uorescent 4 or redox-active 5 groups. The C5 site of pyrimidines and the N7/C8 sites of purines are usually selected for modication as they are more synthetically accessible than other sites and do not interfere with base pairing interactions (Fig. 1a). 6 And largely too for synthetic reasons, the most common nucleobase for functionalisation is thymine (or uracil), as fewer protective steps are required. Functional reactive handles sited at the C5 position of thymine/uracil are typically amino, 7 alkyne, 8 carboxylate, 9,10 halogenated, 11 or thiol 12 groups. Once inserted into DNA, these handles may be used to conjugate the oligonucleotide (i.e. using post-DNA synthesis) to various motifs via amide coupling, 13 palladium catalysed coupling, 14 or alkyne-azide click chemistry. 15 Here, we introduce an acrylamide-containing group at the C5 position of thymine to create the novel nucleoside derivative Acrylamide-dT in the form of its phosphoramidite (Fig. 1b). This new handle is ideal for reacting with nucleophiles such as thiols via the Michael addition reaction, or for forming polymers via free radical polymerisation chemistry. Currently for such purposes, oligonucleotides can be modied with the commercially available phosphoramidite Acrydite™ (Fig. 1b). 16 For example, Jie and coworkers used this acrylamide reagent to integrate uorescent DNA into hydrogels, whereas others have incorporated it into aptameric systems for molecular imprinting applications. 9,10,17,18 However, the placement of Acrydite™ is limited to the 5 0 end of the oligonucleotide. We reasoned that a synthon such as Acrylamide-dT would allow placement of acrylamide moieties not only internally, but also at multiple positions within a strand.
Multiple incorporations of polymerisable groups into DNA have previously been performed using the commercially available Carboxy-dT-CE monomer (Fig. 1b), with the ester moiety reported as being converted to an amide upon oligonucleotide deprotection with ammonia. 9 The work by Turner et al. on DNAincorporated molecularly imprinted nanoparticles highlights that multiple anchorage points of the DNA strand to the polymer network is advantageous for target binding. 9,10 A structural comparison of this Carboxy-dT-CE monomer to Acrylamide-dT suggests heightened reactivity for the latter through its primary alkene which is less sterically hindered. Herein we report the ready incorporation of Acrylamide-dT into DNA sequences at both internal and end positions and also show that these strands can undergo polymerisation with acrylamide/ bisacrylamide monomers in gel electrophoresis experiments.

Synthesis of Acrylamide-dT and its phosphoramidite
In approaching the synthesis of the target Acrylamide-dT and its corresponding phosphoramidite, our rst strategy involved replacing the iodine at C5 of commercially available 5-iodo-2 0deoxyuridine with an acrylamide group via the Heck reaction with N,N 0 -methylenebisacrylamide (Scheme 1). A microwave assisted procedure adapted from Fujimoto 19 was followed to couple the two entities together to form Acrylamide-dT in good yield (70%). Excess N,N 0 -methylenebisacrylamide was used to minimise coupling to both sides of the molecule. The microwave reaction was monitored by thin layer chromatography, which took approximately 10 minutes to complete. The product was isolated by precipitation from cold chloroform, and washed with cold chloroform to remove excess N,N 0 -methylenebisacrylamide. The next step was to perform tritylation and phosphitylation to allow incorporation of the target molecule into DNA sequences using automated synthesis. Unfortunately, tritylation of the 5 0 hydroxyl position of Acrylamide-dT with a dimethoxytrityl group proved difficult, resulting in extremely low yields of compound 3 (<5%) even under prolonged reaction conditions (3 days). An alternative route involving a change in order was therefore devised in which tritylation was performed rst on 5-iodo-2 0 -deoxyuridine to give compound 2 (ref. 20) prior to the palladium catalysed addition of N,N 0 -methylenebisacrylamide. This route proved successful with identical microwave conditions applied to produce compound 3 in good yield. The enhanced hydrophobicity of compound 3 led to its isolation by column chromatography using DCM/methanol rather than precipitation from chloroform. The remaining 3 0 alcohol of compound 3 was then phosphitylated to obtain the target Acrylamide-dT-CE phosphoramidite in 72% yield (Scheme 1).

Incorporation of Acrylamide-dT into DNA
Preliminary 1 H, 13 C, and 2D NMR, and mass spectrometry experiments on Acrylamide-dT showed that the phosphoramidite would not be able to withstand standard oligonucleotide deprotection conditions used in DNA synthesis as ammonia was found to add to the terminal alkene to form a primary amine (see ESI †). However no such problems were found using the ultramild deprotection conditions of methanolic potassium carbonate. Therefore these conditions were chosen for automated oligonucleotide synthesis. Given the ongoing interest in functionalising aptameric DNA sequences with polymerisable groups, 9,10,17,18 and our recent work Scheme 1 Synthesis of Acrylamide-dT and its derivatives. on such systems, 21 it was decided to incorporate Acrylamide-dT into the thrombin binding aptamer (TBA). Different positions in the TBA sequence were chosen, including internal T sites and endfunctionalisation, as shown in Table 1. Those T sites in the central TGT loop known not to interact with the thrombin protein were selected for replacement with Acrylamide-dT. 22 DNA synthesis proceeded very smoothly with up to three incorporations achieved successfully, and high average stepwise yields of >90% for each coupling. Each strand was puried by reversed-phase HPLC (RP-HPLC) and characterised by analytical RP-HPLC and ESMS (see ESI †). The strands were found to be stable, with no appreciable degradation aer storing samples under standard conditions (ultrapure water at À20 C) over several months.
Structure and binding properties of Acrylamide-dT modied DNA Circular dichroism (CD) spectroscopy was used to probe the effect of introducing acrylamide groups on the structure of the modied aptamer strands (Fig. 2). In the presence of potassium, TBA has a characteristic CD maximum at ca. 295 nm and a CD minimum at ca. 265 nm, which is indicative of a G-quadruplex conformation. 23 The data showed no signicant difference in peak positions of the CD signals of the modied strands compared to the unmodied TBA strand. The strong peaks at ca. 295 nm highlight this effect and indicate that all modied strands can similarly adopt a G-quadruplex conformation.
Thermal melting (T m ) experiments were performed to assess whether the inclusion of acrylamide groups had any effect on the stability of the duplexes formed with the 15-mer complementary strand ( Table 2). The average T m values of the modied strands were similar if not slightly higher than the unmodied version. This suggests that the acrylamide groups enhance the stability of the duplex to some extent, possibly since the acrylamide moiety has sites for additional hydrogen bonding interactions.
Gel electromobility shi assays were then used to demonstrate affinity of the Acrylamide-dT modied oligonucleotides to the thrombin protein (see ESI †). The images revealed near complete binding to thrombin for all modied stands at a ratio of 1 : 1 DNA : thrombin. Furthermore, the intensities of the bands were similar to those observed with unmodied TBA, indicating similar thrombin binding affinities, with the dissociation constant (K d ) ratio for TBA and Acryl-endT estimated from relative band intensities as: Reaction of Acrylamide-dT modied DNA with acrylamide gels An important aspect of the work was to check whether DNA functionalisation would not adversely affect the ability of the acrylamide groups to polymerise. A polyacrylamide gel This journal is © The Royal Society of Chemistry 2019 RSC Adv., 2019, 9, 31511-31516 | 31513 Paper experiment was therefore designed to test this in the presence of other acrylamide monomers. A mixture of acrylamide, bisacrylamide, ammonium persulphate, TEMED, and DNA (unmodied or Acrylamide-dT modied) was loaded into wells of a polyacrylamide gel and allowed to set. It was anticipated that if the acrylamide group on the strands could react with the gel, the strands would become covalently bound and immobilised. Conversely, unmodied TBA would move down the gel in the normal way as it does not possess the acrylamide functionality. Once run, the gels were stained and visualised to determine the position of the strands (Fig. 3). As expected, each Acrylamide-dT containing DNA strand remained in the wells at the top of the gel. The absence of any band intensity lower down each gel suggests a near-quantitative reaction under these conditions. In contrast, samples of unmodied TBA moved down the gel along with the current, with there being no indication of any reaction with acrylamide monomers, for example via amine-containing nucleobases. This result clearly shows the ability of the incorporated acrylamide groups to react with other acrylamide monomers to anchor the DNA within the polymer network. As expected, running gels of the three modied strands in the absence of acrylamide, bisacrylamide and radical initiators resulted in no retardation of the bands (see ESI †).

Conclusion
In conclusion, a new polymerisable nucleosidic monomer, Acrylamide-dT, and its associated phosphoramidite has been synthesised, with the latter made in three steps from commercially available starting materials. Under ultramild DNA synthesis conditions, it can be readily incorporated intact into aptamer sequences, both internally and externally (i.e. at strand ends) and in multiple positions, thus expanding the repertoire for DNA modication chemistry. The modications were found to have no adverse effect on the structures or thrombin binding properties of the DNA strands, as evidenced by CD, T m and gel experiments. Furthermore, the acrylamide groups within these strands were shown to react with other acrylamide monomers under gel electrophoresis conditions to form acrylamide polymers. We expect Acrylamide-dT to be a useful synthon in the future for constructing new DNA-polymer hybrids for various applications such as stimuli-responsive hydrogels or molecular imprinting. Although this methodology has currently been tested only on thymine/uracil sites, given the ease of synthesis, it should be possible to similarly attach acrylamide groups to other nucleobases to broaden the applicability further.

Materials and methods
Reagents and solvents were purchased from commercial suppliers and used without further purication, unless otherwise stated. 5 0 -O-(4,4 0 -dimethoxytrityl)-5-iodo-2 0 -deoxyuridine 2 was synthesised following an established procedure. 20 Column chromatography was carried out using open columns packed with Merck grade 60 silica gel topped with 0.5 cm of sand. TLC analysis was performed on Merck silica gel 60 silica sheets. 1 H, 13 C, and 31 P NMR spectra were obtained on Bruker AVIII300 or AVIII400 spectrometers. Chemical shis (d) are given in ppm and are relative to the residual solvent peak. Electrospray mass (ESI-MS) spectra were measured by either Waters micromass LCT electrospray time-of-ight (ES-TOF), Waters Xevo G2-XS, or Synapt G2S mass spectrometers. Milli-Q water puried with a Millipore Elix-Gradient A10 system (resistivity > 18 mU cm, TOC # 5 ppb, Millipore, France) was used for DNA sample preparation.
Synthesis of Acrylamide-dT 5-Iodo-2 0 -deoxyuridine (1.00 g, 2.82 mmol) and palladium acetate (60 mg, 0.28 mmol) were suspended in DMF (3 ml) in a 10 ml microwave tube equipped with a small magnetic stirring bar. Tributylamine (0.67 ml, 2.82 mmol) and N,N 0 -methylenebisacrylamide (1.09 g, 7.06 mmol), were then added. The suspension was stirred and degassed with argon for 10 minutes. The tube was then tightly sealed, and the mixture irradiated in a microwave for 10 minutes at 100 C. Aer the irradiation period, the reaction vessel was cooled to room temperature before opening. The reaction was then ltered through Celite®. The product was isolated by precipitation from cold chloroform, washing with cold chloroform, to yield an off white solid (748 mg, 70%). 1 13  The tube was tightly sealed and the mixture irradiated in a microwave for 10 minutes at 100 C. Aer the irradiation period, the reaction vessel was cooled to room temperature before opening. The reaction was then ltered through Celite®, rinsing with DCM (50 ml), and the ltrate washed with water (50 ml). The organic phase was dried over magnesium sulphate and evaporated under vacuum. The crude product was puried by ash chromatography with an eluent of 9 : 1 DCM : 7 N ammonia in methanol. The appropriate fractions were combined and evaporated to give an off white solid (700 mg, 67%). 1

Synthesis of unmodied and Acrylamide-dT modied DNA oligonucleotides
All oligonucleotides were synthesised using solid phase synthesis on an Applied Biosystems ABI 394 DNA/RNA synthesiser using commercially supplied DNA synthesis grade solvents and reagents. Standard phosphoramidites of Pac-dA, iPr-Pac-dG, Ac-dC, dT from Link Technologies, and Acrylamide-dT-CE phosphoramidite were used for ultramild synthesis. The phosphoramidites were dissolved in anhydrous acetonitrile to 0.1 M prior to synthesis. Strands were synthesised at a 1 mmol scale on Syn-Base™ CPG 1000/110 solid supports from Link Technologies. Phosphoramidites were activated with 5-ethylthio-1H-tetrazole (0.25 M) in acetonitrile prior to coupling and coupling times of 25 seconds were used. Then, phenoxyacetic anhydride and methylimidazole were added to cap unreacted material, and iodine (0.02 M) in THF/pyridine/water (7 : 2 : 1) was added to oxidise the phosphotriester formed. Upon sequence completion, the resins were placed in freshly prepared 1 ml solutions of potassium carbonate (0.05 M) in methanol and le overnight to cleave strands from the resin and remove protecting groups. The solutions were neutralised with acetic acid (6 ml) and the solvent was removed on a Thermo Scientic speed vac. The dried powders were redissolved in 1 ml Milli-Q water and desalted with a NAP-10 column from GE Healthcare to remove residual resin and potassium carbonate. The solutions were then concentrated to 1 ml and stored in the freezer for purication (see ESI for further details †).
Circular dichroism spectroscopy 500 ml aqueous solutions of each DNA sample (5 mM) in KCl (10 mM) and Tris$HCl buffer (10 mM, pH 7.5) were prepared. Prior to CD analysis, the samples were heated to 95 C for 5 minutes and cooled slowly to room temperature. CD spectra of samples were recorded on a Jasco J-810 spectropolarimeter, scanning from 350-200 nm at a rate of 100 nm min À1 . Three accumulations were performed for each sample and the data produced an average of the three scans. A baseline correction was manually performed on each sample by subtracting the blank and offsetting results at 350 nm.
Thermal melting experiments 100 ml aqueous solutions of each DNA sample (10 mM) and complementary DNA (10 mM) in NaCl (100 mM) and sodium phosphate buffer (10 mM, pH 7) were prepared. 10 ml of each sample was withdrawn, added to 10 mL of SsoAdvanced™ universal SYBR® green supermix from Bio-rad, and the thermal melting of the resulting solutions was performed on a M550 double beam scanning UV/visible spectrophotometer. Samples were heated from 15 C / 90 C and cooled from 90 C / 20 C at a rate of 1 C min À1 , and values were obtained from the maxima of the negative rst derivative of the melting curve.

Polyacrylamide gel experiments
Experiments were performed on 12% native polyacrylamide gels with 1Â TBE buffer and 10 mM potassium chloride, using 1Â TBE buffer with 10 mM potassium chloride as a running buffer. Gels were run on Bio-rad Mini-PROTEAN® gel kits with a Biorad PowerPac (highest voltage: 5000 V/500 mA/400 W). Aer electrophoresis, gels were stained with Diamond™ nucleic acid dye and visualised under UV with an AlphaImager HP gel imager from Alpha Innotech. Gel electromobility shi assays. 50 ml aqueous solutions of DNA (1 mM) and thrombin (1 mM) in KCl (10 mM), Tris$HCl buffer (10 mM, pH 7.5), and glycerol (3%) were prepared. 10 ml of each sample was loaded into wells and gels were run at 100 V for 1 hour.
DNA-acrylamide gel copolymerisation experiments. 100 ml aqueous samples containing 1 mM DNA, 12% acrylamide, 1Â TBE buffer, and 10 mM potassium chloride were prepared. 10% APS in water (1 ml) and TEMED (1 ml) were added and 40 ml of each sample was then immediately loaded into two adjacent wells of a gel (20 ml per well). A three well gap was le between unmodied and Acrylamide-dT modied DNA as the solutions were found to react into neighbouring wells. The solutioncontaining wells were then le to set (typically 5 minutes). Once set, 10 ml of 12% polyacrylamide gel solution was added to the same wells and le to set. The gels were then rinsed with HPLC water and run at 100 V for 1 hour.

Conflicts of interest
There are no conicts to declare.