The synthetic route of dU^Az phosphoramidite is outlined in Scheme 1. dU^Az nucleoside 1 was synthesized from the corresponding 2'-deoxy-5-iodouridine (2) through a palladium-catalyzed cross-coupling reaction [12] with 4-ethynylazobenzene 3[13]. Tritylation at the primary hydroxyl group of 1 with DMTrCl and phosphitylation at the secondary hydroxyl group yielded phosphoramidite 5. The amidite 5 was incorporated into the oligodeoxynucleotide using conventional solid-phase phosphoramidite synthesis and purified by reverse-phase HPLC (29% yield). The ON sequences used in this study are shown in Table 1.

We initially investigated the efficiency of the dU^Az cis-trans photoisomerization property in ON by UV spectra and HPLC analysis. UV spectra of trans/cis ON 7, showed that photoisomerization of trans-dU^Az to cis-dU^Az decreased absorbance at 365 nm and increased absorbance at 310 nm and 450 nm (Figure 2a). The λ_max of cis-form (340 nm) was blue-shifted compared to that of the trans-form (365 nm), as was the case with previous reports [6,7,14]. The trans-form dU^Az was photoisomerized to the cis-form by a 10-second irradiation of 365 nm monochromic light with 60% conversion, as determined by the HPLC peak areas (Figure 2b). In addition, subsequent 10-second irradiation of 450 nm yielded the trans form isomer with 80%. The HPLC analysis showed no side products from the reactions.

Even when the photoirradiation was repeated three times, the efficiency of the dU^Az cis-trans photoisomerization was not attenuated (Figure 2c). It can therefore be concluded that dU^Az has a rapid and highly efficient cis-trans photoisomerization property and the potential to work as a photo-switch for various biomolecules.

We investigated the differences in the thermal stability of 12-bp duplexes containing dU^Az in the trans- and cis-forms by monitoring the melting temperature (T_m) following the way of azobenzene- modified nucleoside containing ONs (Table 2) [15,16]. DNA duplex 7/8 showed a modest T_m difference (ΔT_m) between the trans- and cis-forms, namely, the T_m value of the cis-form was 2 °C higher than that of the trans-form. On the other hand, the ON 7/RNA 12 duplex showed a larger T_m difference. The T_m value of the cis-form was 5 °C higher than that of the trans-form. It is noteworthy that the cis-ON 7/RNA 12 duplex showed a T_m value comparable to that of natural DNA 6/RNA 12 duplex. According to past studies, the cis-form photochromic moieties generically destabilize the duplex because of its interference with the vicinity bases stacking interaction [4,5,17,18,19].In this study, ON containing dU^Az showed a higher hybridization ability when dU^Az is cis-form rather than trans-form, unlike ONs containing the exiting photochromic nucleoside. Brown et al. have reported that hydrophobic buta-1,3-diynyl anthracene in ON leads to significant destabilization of the duplex, probably because the aromatic moiety is exposed to the aqueous environment [9]. The azobenzene moiety of trans-dU^Az also would extend to the outside of the major groove, a highly polar aqueous phase. This may have an impact on the groove hydration and the formation of interstrand cation bridges, and lead to destabilization of the duplex containing trans-dU^Az.

Finally, we investigated the mismatch discrimination ability of ON containing dU^Az. The T_m values of mismatched DNA duplexes containing dU^Az were found to be 14 or 15 °C lower than that of ON7/DNA8 in both trans- and cis-form (Table 3). Toward complementary ssRNA, ON containing dU^Az could also discriminate mismatched bases comparable to ON7 (Table S1 in Supplementary Material). These results indicate that the mismatch discrimination ability of ON containing trans-/cis-dU^Az is not spoiled by the C5-substituted-azobenzene moiety of dU^Az.

We achieved synthesis of the photoisomeric nucleoside, dU^Az, for which the hybridization can be controlled by using different wavelengths of light. The ΔT_m value between the trans- and cis-form is more remarkable in the DNA/RNA duplex than the DNA duplex. Although dU^Az photoisomerization induced modest T_m differences, the modification of ONs with multiple dU^Az units or the introduction of substituents to the azobenzene moiety [20] could enhance the ΔT_m value between the trans- and cis-forms. Our strategy indicated the possibility of photo-switches based on dU^Az-modified ONs for the development of unique molecular machines and the control of various biological phenomena.

Reagents and solvents were purchased from commercial suppliers and were used without purification unless otherwise specified. All experiments involving air and/or moisture-sensitive compounds were carried out under N₂ or Ar atmosphere. All reactions were monitored with analytical TLC (Merck Kieselgel 60 F254). Column chromatography was carried out with a Fuji Silysia FL-100D. Physical data were measured as follows: NMR spectra were recorded on a JEOL JNM-ECS-500 spectrometer in CDCl₃ or DMSO-d₆ as the solvent with tetramethylsilane as an internal standard. IR spectra were recorded on a JASCO FT/IR-4200 spectrometer. Optical rotations were recorded on a JASCO P-2200 instrument. FAB mass spectra were measured on a JEOL JMS-700 mass spectrometer.

Under an argon atmosphere, 4-ethynylazobenzene (3 [13], 1.06 g, 5.12 mmol), Pd(PPh₃)₄ (592 mg, 0.512 mmol), and CuI (113 mg,0.512 mmol) was dissolved in dry DMF (50 mL). Then, Et₃N (3.6 mL) and 2'-deoxy-5-iodouridine (2, 1.81 g, 5.12 mmol) were added. The reaction mixture was stirred at 60 °C for 4 h. The resultant mixture was filtered over Celite. The filtrate was concentrated in vacuo. The residue was purified by silica gel column chromatography and eluted with CHCl₃/MeOH (20:1), to give compound 1 (1.80 g, 81%) as a light-orange powder: M.p. 208–210 °C; IR (KBr): ν 3439 (NH, OH), 1617 (C=O), 1289 (N=N) cm⁻¹; −3.7 (c 1.00, DMSO); ¹H-NMR (500 MHz, DMSO-d₆): δ 11.7 (1H, brs, NH), 8.47 (1H, s, H-6), 7.94–7.90 (4H , m), 7.69–7.57 (5H, m), 6.14 (1H, t, J = 6.5 Hz, H-1'), 5.27 (1H, d, J = 4.0 Hz, H-3'), 5.20 (1H, t, J = 5.0 Hz, C-H4'), 4.30–4.26 (1H, m, OH), 3.82 (1H, m, OH), 3.71–3.58 (2H, m, H-5'), 2.21–2.17 (2H, m, H_-2'); ¹³C-NMR (125 MHz, DMSO-d₆): δ 161.3, 151.9, 151.0, 149.4, 132.2, 131.8, 129.5, 125.4, 122.9, 122.6, 97.8, 91.5, 87.6, 85.6, 84.9, 69.8, 60.8, 40.2; FAB-LRMS m/z = 433 (MH⁺); FAB-HRMS calcd for C₂₃H₂₁N₄O₅ 433.1506, found 433.1524.

To a solution of compound 1 (141 mg, 0.324 mmol) in dry pyridine (3 mL) was added DMTrCl (131 mg, 0.389 mmol) at room temperature, and the reaction mixture was stirred for 4 h. The reaction was quenched by the addition of MeOH with 10 min stirring. The solvent was removed in vacuo, and the residue was partitioned between CHCl₃ and H₂O. The separated organic layer was washed with H₂O, followed by brine. The organic layer was dried (Na₂SO₄) and concentrated in vacuo. The residue was purified by silica gel column chromatography and eluted with CHCl₃/MeOH (20:1 with 0.5% Et₃N) to give Compound 4 (239 mg, 88%) as an orange foam: IR (KBr): ν 3437, 3410(NH, OH), 1701 (C=O), 1272 (N=N) cm⁻¹; 36.2 (c 1.00, CHCl₃); ¹H-NMR (500 MHz, CDCl₃): δ 8.51 (1H, brs, NH), 8.29 (1H, s, H-6), 7.90 (2H, d, J = 7.5 Hz), 7.70 (2H, d, J = 8.5 Hz), 7.52–7.45 (5H, m), 7.37–7.28 (6H, m), 7.16 (1H, dd, J = 6.5 and 1.0 Hz), 7.10 (2H, d, J = 8.0 Hz), 6.82–6.79 (4H, m) 6.38 (1H, dd, J = 7.5, 6.5 Hz, H-1'), 4.60–4.59 (1H, m, H-3'), 4.14–4.13 (1H, m, H-4'), 3.70 (3H, s, OMe), 3.69 (3H, s, OMe), 3.50 (1H, dd, J = 8.0 and 3.0 Hz, H-5'), 3.34 (1H, dd, J = 8.0 and 3.0 Hz, H-5'), 2.57–2.53 (1H, m, H-2'), 2.40–2.34 (1H, m, H-2'), 2.09 (1H, brs, OH); ¹³C-NMR (125 MHz, CDCl₃): δ 158.6, 152.6, 151.7, 148.8, 144.3, 135.4, 132.4, 131.3, 129.9, 129.1, 128.1, 127.9, 127.1, 125.1, 122.9, 122.5, 113.4, 100.4, 93.6, 87.2, 86.7, 85.9, 82.2, 72.4, 63.3, 55.2, 41.7; FAB-LRMS m/z = 757 (MNa⁺); FAB-HRMS calcd for C₄₄H₃₈N₄O₇Na 757.2633, found 757.2633.

To a solution of compound 4 (188 mg, 0.26 mmol) in dry MeCN (5 mL) was added N, N-diisopropylamine (0.13 mL,0.76 mmol) and 2-cyanoethyl-N, N'-diisopropylchlorophosphoramidite (0.09 mL, 0.40mmol) at room temperature, and the reaction mixture was stirred for 1.5 h. The resultant mixture was partitioned between AcOEt and H₂O. The separated organic layer was washed with saturated aqueous NaHCO₃, followed by brine. The organic layer was dried (Na₂SO₄) and concentrated in vacuo. The residue was purified by silica gel column chromatography and eluted with CHCl₃/MeOH (20:1 with 0.5% Et₃N), to give a 17:3 diastereomeric mixture of 5 (324 mg, 82%) as an orange foam: IR (KBr): ν 3610 (NH), 1699 (C=O), 1272 (N=N) cm⁻¹; 32.5 (c 1.00, CHCl₃); ¹H-NMR (500 MHz, CDCl₃): δ 9.08 (1H, brs, NH), 8.35 (0.85H, s, H-6), 8.30 (0.15H, s, H-6), 7.89 (2H, d,J = 7.5 Hz), 7.67 (2H, d, J = 8.5 Hz), 7.55–7.04 (14H, m), , 6.67–6.75 (4H, m), , 6.35 (1H, dd, J = 7.5, 6.0 Hz, H-1'), 4.68–4.61 (1H, m, H-3'), 4.26 (1H, m, H-4'), 3.70 (3H, s, OMe), 3.69 (3H, s, OMe), 3.67–3.53 (5H, m, CH₂CH₂CN, H-5'), 3.31 (1H, dd, J = 8.5, 2.5 Hz, H-5'), 2.65–2.56 (1H, m, H-2'), 2.47–2.36 (3H, m, H-2', ((CH₃)₂CH)₂N), 1.18 (12H, d, J = 6.5 Hz, ((CH₃)₂CH)₂N); ¹³C-NMR (125 MHz, CDCl₃): δ 161.2, 158.5(9), 158.5(6), 152.6, 151.5, 149.1, 144.35, 142.5, 135.4, 132.3, 132.0, 131.1, 130.0 (d, J (C, P) = 6.0 Hz), 129.1, 128.7, 128.0, 127.9 ,127.0, 125.1, 122.8, 122.4, 120.5, 117.3, 113.3, 100.3, 93.4, 86.3 (d, J (C, P) = 3.5 Hz), 85.9, 82.4, 77.3, 77.0, 76.8, 73.4, 73.2, 63.0, 58.2, 58.1, 55.1, 43.2 (d, J (C, P) = 13.0 Hz), 40.8 (d, J (C, P) = 5.0 Hz), 25.6, 24.5(9), 24.5(3), 24.4(8), 20.2 (d, J (C, P ) = 7.0 Hz); ³¹P-NMR (200 MHz, CDCl₃): δ 149.09, 148.66; FAB-LRMS m/z = 957 (MNa⁺); FAB-HRMS calcd for C₅₃H₅₅N₆O₈PNa 957.3711, found 957.3711.

Solid-phase oligonucleotide synthesis was performed on an nS-8 Oligonucleotides Synthesizer (GeneDesign, Inc., Osaka, Japan) using commercially available reagents and phosphoramidites with 5-(bis-3, 5-trifluoromethylphenyl)-1H-tetrazole (0.25 M concentration in acetonitrile) as the activator. dU^Az phosphoramidite was chemically synthesized as described above. All of the reagents were assembled, and the oligonucleotides were synthesized according to the standard synthesis cycle (trityl on mode). Cleavage from the solid support and deprotection were accomplished with concentrated ammonium hydroxide solution at 55 °C for 12 h. The crude oligonucleotides were purified with Sep-Pak Plus C18 cartridges (Waters) followed by RP-HPLC on a XBridge^TM OST C18 Column, 2.5 μm, 10 × 50 mm (Waters) using MeCN in 0.1 M triethylammonium acetate buffer (pH 7.0). The purified oligonucleotides were quantified by UV absorbance at 260 nm and confirmed by MALDI-TOF mass spectrometry (Table 4).

Melting temperatures (T_m) were determined by measuring the change in absorbance at 260 nm as a function of temperature using a Shimadzu UV-Vis Spectrophotometer UV-1650PC equipped with a T_m analysis accessory TMSPC-8. Equimolecular amounts of the target DNA/RNA and oligonucleotides were dissolved in 10 mM sodium phosphate buffer (pH 7.0) containing 100 mM NaCl to give a final strand concentration of 4.0 µM. The melting samples were denatured at 100 °C and annealed slowly to room temperature. Absorbance was recorded in the forward and reverse directions at temperatures of 5 to 90 °C at a rate of 0.5 °C/min.

The trans-to-cis isomerization was performed with a UV-LED lamp (ZUV-C30H; OMRON) and a ZUV-L10H lens unit (760 mW/cm²). The cis-to-trans isomerization was performed with a Xenon lamp (MAX-303; Asahi Spectra Co., Ltd., Tokyo, Japan) and XHQA420 optical filter. Absorbance spectra of trans-cis ON 7 were measured by a Shimadzu UV-Vis Spectrophotometer UV-1650PC. Conditions: ON 7 (4.0 µM), NaCl (100mM) in sodium phosphate buffer (10 mM, pH 7.0).