Development of an Artificial Nucleic Acid Skeleton Allowing for Unnatural-Type Triplex DNA Formation with Duplex DNA Having a TA Inversion Site

Abstract

Triplex DNA formation has generated much interest as a genomic targeting tool that directly targets duplex DNA. However, fundamental limitations in the base pairs of target duplex DNA sequences that can form stable triplex DNA have limited the application. Recently, we have reported on the recognition of CG and ^5mCG base pairs by artificial nucleic acid derivatives with a 2′-deoxynebularine skeleton. Therefore, we attempted to explore the basic skeleton that is important for the development of new artificial nucleic acids allowing for the recognition of TA base pairs. In this study, we focused on a benzimidazole skeleton and introduced a hydroxyl group to enable one-point hydrogen bonding. We have synthesized artificial nucleoside analogues with hydroxyl group on the benzimidazole and incorporated their amidite derivatives into triplex forming oligonucleotides (TFOs). The gel shift assay was performed to evaluate the triplex DNA formation ability of synthesized TFOs, and TFOs containing hydroxybenzimidazole were successfully recognized TA base pairs for all four different sequences. Moreover, compared to the results for the TFOs containing benzimidazole, which suggested hydrogen bonding formation at the hydroxyl group. Therefore, hydroxybenzimidazole would be an important artificial nucleic acid skeleton for TA base pair recognition.

Introduction

Duplex DNA is one of the important biological components that store genetic information. Therefore, molecules that act directly on duplex DNA may be applied to various applications targeting the genome in the future.^1–5) Among them, triplex DNA formation using triplex-forming oligonucleotides (TFOs) is expected to be deployed in nucleic acid therapeutics and genome editing tools.^6–12) As a property of the antiparallel type that forms stable triplex DNA under physiological conditions, it is formed by an interaction with the homopurine strand of the target duplex DNA from the major groove side by TFOs, which comprise purine-rich nucleobases. The guanine nucleobase (G) and adenine nucleobase (A) or thymine nucleobase (T) in TFOs form two reverse Hoogsteen hydrogen bonds with the guanine base (G) and adenine base (A) in duplex DNA, respectively, thereby stabilizing the triplex complex of G/GC, A/AT, and T/AT base triplets^12,13) (Fig. 1A). Since there is only one hydrogen-bonding site on the major groove side of the pyrimidine nucleobase in the base pair of the inverted CG and TA base pairs, TFOs composed of natural nucleosides cannot form stable triplex DNA¹³⁾ (Fig. 1B). Although a number of researchers have developed artificial nucleoside analogues that recognize CG and TA base pairs in duplex DNA, novel nucleoside derivatives have yet to be identified.^14–18)

Fig. 1. Chemical Structures of (A) Natural Base Triplets, (B) an Unrecognizable TA Site, (C) an Artificial Base Triplet; the Guanidino-dN and ^5mCG Base Pair, (D) a Novel Base Skeleton for the Recognition of a TA Base Pair, and (E) Nucleic Acid Derivatives with a Benzimidazole Skeleton in the Nucleobase Part Used in the Present Study

Our group also developed artificial nucleoside analogues that recognize CG or TA base pairs and reported a number of useful nucleoside derivatives.^19–22) However, the stability of the triplex complex was insufficient, particularly with regards to the stability of TA base pair recognition, which was affected by the neighboring nucleobases of artificial nucleoside analogues.²³⁾ We recently reported the recognition of CG and ^5mCG base pairs by artificial nucleic acid derivatives with a 2′-deoxynebularine (dN) skeleton.²⁴⁾ Since nebularine is a purine ring, it was expected to increase stability by promoting stacking interactions with neighboring nucleobases. Furthermore, TFOs containing guanidino-dN successfully avoided steric repulsion with the methyl group at the 5-position of the ^5mC nucleobase to ensure the recognition of the ^5mCG base pair (Fig. 1C). The avoidance of the methyl group may be applied to the design of new molecules that recognize the T nucleobase with a methyl group in the same position. Therefore, we attempted to investigate the basic skeleton that is important for the development of novel nucleoside analogues allowing for the recognition of TA base pairs (Fig. 1D). We focused on adopting a simple benzimidazole skeleton in consideration of further derivatization in the future. Moreover, a hydroxyl group was introduced at the appropriate position of the benzimidazole ring to allow for the formation of single hydrogen bonding (Fig. 1D). Therefore, we designed the nucleoside analogues of 6-hydroxy-1H-benzimidazole derivative (1) and chose to use benzimidazole derivative (2) as a control compound (Fig. 1E).

We herein described the synthesis of new nucleoside analogues with benzimidazole skeleton derivatives and an evaluation of triplex DNA formation by TFOs including these analogues. The ability to form triplex DNA was evaluated by a gel shift assay, and the recognition ability of base pairs in duplex DNA was assessed. We found that TFOs containing hydroxybenzimidazole recognized TA base pairs without sequence dependency.

Results and Discussion

Synthesis of TFOs Having Benzimidazole Derivatives

The main synthetic routes were depicted in Chart 1. The synthesis of 6-hydroxy-1H-benzimidazole used 4-amino-3-nitrophenol as a starting material. The phenolic hydroxyl group was protected with a benzoyl group to give 3. A condensation reaction at a low temperature provided a highly selective β-isomer 4. After removing all protecting groups to give 1, stereochemistry and regioselectivity were confirmed by measuring two dimensional-correlation spectroscopy (2D-COSY) and nuclear Overhauser effect spectroscopy (NOESY) NMR. The benzoyl group was removed at the same time as the deprotection of the toluoyl group; therefore, the benzoyl group was re-protected and the diol compound 5 was successfully obtained in good yield. As a control compound, benzimidazole derivative, which does not have a phenolic hydroxyl group, was condensed with the sugar moiety to obtain the compound 6, which was then deprotected to obtain the diol compound 2. The synthesized diol compounds of benzimidazole derivatives were converted to the corresponding phosphoramidite compounds (7 and 8) under general reaction conditions. These amidite compounds were applied to the automated DNA synthesizer in order to synthesize TFOs with their benzimidazole derivatives. These artificial nucleic acids were introduced into four different sequences to test the effects of neighboring nucleobases. After the purification of TFOs by reverse-phase HPLC, the DMTr group at the 5′-end of TFOs was deprotected by a treatment with 5% acetic acid. These synthesized TFOs (TFO1–4(Z)) were identified using matrix assisted laser desorption/ionization-time of flight (MALDI-TOF) MS (Table 1 and Supplementary Figs. S1 and S2).

Chart 1. Synthesis Scheme of Benzimidazole Derivatives and Sequences of TFOs

Table 1. MALDI-TOF MS Results of Synthesized TFOs in the Present Study

Z =	1			2
TFOs	Calcd.a)	Found	Yieldb)	Calcd.a)	Found	Yieldb)
3′-AZA-5′	5731.00	5730.30	20	5715.01	5714.98	25
3′-AZG-5′	5746.99	5747.29	20	5731.00	5729.53	22
3′-GZA-5′	5746.99	5746.16	18	5731.00	5730.73	22
3′-GZG-5′	5762.98	5763.93	16	5746.99	5748.73	20

a) Negative mode measurements, [M−H]⁻. b) Isolated yield (%).

Evaluation of the Triplex-Forming Abilities

The triplex-forming abilities of synthesized TFOs were analyzed using a non-denatured polyacrylamide gel-shift assay in the presence of 20 mM MgCl₂ (Figs. 2, 3). A low-mobility band resulting from triplex DNA formation was observed in contrast to high-mobility unreacted FAM-labeled duplex DNA. For example, a TFO with hydroxybenzimidazole derivative 1 in the sequence of 3′-AZA-5′ showed a triplex band around the addition of 100 nM TFOs for duplex DNA containing a TA base pair (Fig. 2). On the other hand, triplex bands were observed at the addition of 500 nM TFOs for duplex DNA containing AT or CG base pairs. Triplex DNA formation was not clearly observed for GC base pairs. Furthermore, relatively stable triplex DNA formation was observed even when TFO sequences with different neighboring nucleobases of the artificial nucleoside analogues were used. Regarding TFO having benzimidazole derivative 2 without the hydroxyl group synthesized as a control compound, triplex DNA formation was observed in the 3′-G and 5′-G sequences (Fig. 3). The fluorescence intensities of these bands in the gel-shift assay were quantified to assess equilibrium association constants (K_s values), which are summarized in Table 2. TFO with hydroxybenzimidazole derivative 1 showed relatively high K_s values for the TA base pair in all four different sequences. Some sequences also showed affinity for AT base pairs. Overall, this artificial nucleoside skeleton showed higher values for the TA and AT base pairs than for the CG and GC base pairs. The control compound, benzimidazole derivative 2, showed a value of 10.9 for the TA base pair in the 3′-GZG-5′ sequence, which might be due to the shape-fitting of the aromatic ring moiety to the TA base pair; however, the value for the others was either small or not calculable under these assay conditions. Therefore, we evaluated it using a duplex DNA containing dUA base pair (Supplementary Fig. S3). The results showed similar K_s values for TFOs having two derivatives. Interestingly, their values were similar to or slightly lower than the K_s values of TFOs having benzimidazole for TA base pair. These findings suggest that the benzimidazole skeleton is able to exist in the major groove side of the T or dU nucleobase by the stacking interaction with the neighboring nucleobases in TFOs. The differences in the K_s values by TFOs for the four different sequences may be due to the effects of the magnitude of the stacking interaction. Thymidine (T), a natural nucleoside, forms a base triplet structure with the AT base pair in duplex DNA through two hydrogen bonds (Fig. 1A). However, K_s values differed depending on the neighboring nucleobases: 41.8 × 10⁶ M⁻¹ for 3′-AZA5′, 20.8 × 10⁶ M⁻¹ for 3′-AZG-5′, 31.2 × 10⁶ M⁻¹ for 3′-GZA-5′, and 10.9 × 10⁶ M⁻¹ for 3′-GZG-5′ sequences, indicating stable and selective affinity for AT base pairs. The mechanisms underlying this interaction warrant further study, which will be performed when synthesizing 6-hydroxy-1H-benzimidazole derivatives. Although the selectivity needs further improvement, we have identified a new basic skeleton that recognizes a TA base pair in duplex DNA in any sequence.

Fig. 2. Gel Results of Triplex DNA Formation

Conditions: FAM-labeled duplex DNA (24 bp; 100 nM) was incubated with increasing concentrations of TFO1–4(1) (18-mer; 0–1000 nM) in buffer containing 20 mM Tris–HCl and 20 mM MgCl₂ at 37 °C for 16 h and pH 7.5 Electrophoresis was performed with a 10% non-denaturing poly-acrylamide gel.

Fig. 3. Gel Results of Triplex DNA Formation

Conditions: FAM-labeled duplex DNA (24 bp; 100 nM) was incubated with increasing concentrations of TFO1–4(2) (18-mer; 0–1000 nM) in buffer containing 20 mM Tris–HCl and 20 mM MgCl₂ at 37 °C for 16 h and pH 7.5. Electrophoresis was performed with a 10% non-denaturing poly-acrylamide gel.

Table 2. Equilibrium Association Constants (K_s) of Each TFO^a)

Sequences		3′-GGAAGG NZN′ GAGGAGGGA　5′-GAGGGAAGG NXN′ GAGGAGGGAAGC　3′-CTCCCTTCC MYM′ CTCCTCCCTTCG-FAM
TFOs	Z =	1				2				Tb)
	XY =	TA	AT	CG	GC	TA	AT	CG	GC	TA	AT	CG	GC
TFO1 (3′-AZA)		2.9	0.8	0.5	—c)	—c)	—c)	—c)	—c)	—c)	41.8	1.4	—c)
TFO2 (3′-AZG)		16.5	14.1	6.9	4.1	1.0	—c)	—c)	—c)	2.3	20.8	4.2	2.5
TFO3 (3′-GZA)		13.3	15.3	6.4	5.9	6.0	0.2	—c)	—c)	4.1	31.2	11.2	5.0
TFO4 (3′-GZG)		30.0	5.9	7.6	12.7	10.9	1.4	1.7	0.8	4.7	10.9	6.2	5.2

a) Conditions: FAM-labeled duplex DNA (24 bp; 100 nM) was incubated with increasing concentrations of TFOs (18-mer; 0–1000 nM) in buffer containing 20 mM Tris–HCl and 20 mM MgCl₂ at 37 °C for 16 h and pH 7.5. Electrophoresis was performed with a 10% non-denaturing polyacrylamide gel. K_s (10⁶ M⁻¹) = [Triplex]/([TFO][Duplex]). The mean of each K_s value was calculated from three independent experiments, errors are less than 10%. b) These data were from reference 22. c) The value is less than 0.1.

Conclusion

We newly developed the basic skeleton of the recognition molecule that interacts with TA base pairs without sequence dependency. Although some sequences also interact with AT base pairs, we expect antigene oligonucleotides to be developed using this compound if the target sequence is carefully designed. Furthermore, the synthesis and evaluation of the hydroxybenzimidazole derivative is expected to lead to the development of novel artificial nucleoside analogues that recognize TA base pairs with high selectivity and stability on going now.

Experimental

Evaluation of Triplex-Forming Abilities of TFOs

FAM-labeled duplex DNA (100 nM) was incubated with increasing concentrations of TFO1–4(Z) (0–1000 nM) in buffer containing 20 mM Tris–HCl and 20 mM MgCl₂ at 37 °C for 16 h and pH 7.5. Electrophoresis was performed at 4 °C for 5 h using a 10% non-denaturing polyacrylamide gel. Faster mobility bands were observed as duplex DNA, while slower mobility bands were detected as triplex DNA. The gel was visualized using the Luminoimage analyzer LAS-4000 (FUJIFILM Tokyo, Japan) (Figs. 2, 3 and Supplementary Fig. S3), and the fluorescence intensity of each band was quantified for the calculation of equilibrium association constants. K_s (10⁶ M⁻¹) = [Triplex]/([TFO][Duplex]). All K_s values were similarly calculated using a narrow range of TFO concentrations corresponding to each TFO. The mean of each K_s value was calculated from three independent experiments.

Acknowledgments

The present study was supported by a Grant-in-Aid for Scientific Research (B) (Grant Number JP19H03351 and JP23H02610) from the Japan Society for the Promotion of Science (JSPS), AMED under Grant Number JP21am0401026, and the JST SPRING Program (Grant Number JPMJSP2136, Japan for R.N.).

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

References

• 1) McGorman B., Fantoni N. Z., O’Carroll S., Ziemele A., El-Sagheer A. H., Brown T., Kellett A., Nucleic Acids Res., 50, 5467–5481 (2022).
• 2) Dalla Pozza M., Abdullrahman A., Cardin C. J., Gasser G., Hall J. P., Chem. Sci., 13, 10193–10215 (2022).
• 3) Bekkouche I., Shishonin A. Y., Vetcher A. A., Polymers, 15, 858 (2023).
• 4) Hu Y., Cecconello A., Idili A., Ricci F., Willner I., Angew. Chem. Int. Ed., 56, 15210–15233 (2017).
• 5) Ni D., Huang K., Wang H., Zhou W., Guo M., Baimanov D., Xue Y., Chen Y., Liu Y., Nano Today, 48, 101699 (2023).
• 6) Lohani N., Rajeswari M. R., Anticancer. Agents Med. Chem., 20, 1943–1955 (2020).
• 7) Li C., Zhou Z., Ren C., Deng Y., Peng F., Wang Q., Zhang H., Jiang Y., Front. Pharmacol., 13, 1007723 (2022).
• 8) Hennessy J., McGorman B., Molphy Z., Farrell N. P., Singleton D., Brown T., Kellett A., Angew. Chem. Int. Ed., 61, e202110455 (2022).
• 9) Wang L., Notomi R., Sasaki S., Taniguchi Y., RSC Med. Chem., 14, 1482–1491 (2023).
• 10) Wang L., Okamiura H., Sasaki S., Taniguchi Y., Chem. Comm., 56, 9731–9734 (2020).
• 11) Taniguchi Y., Sasaki S., Org. Biomol. Chem., 10, 8336–8341 (2012).
• 12) Reza F., Glazer P. M., Methods Mol. Biol., 1114, 115–142 (2014).
• 13) Fox K. R., “Handbook of Chemical Biology of Nucleic Acids”, Springer, Singapore, 2023, pp. 2613–2643.
• 14) Nishizawa S., Tu G., Ogata D., Miyauchi K., Ohkubo A., Bioorg. Med. Chem., 71, 116934 (2022).
• 15) Ohkubo A., Yamada K., Ito Y., Yoshimura K., Miyauchi K., Kanamori T., Masaki Y., Seio K., Yuasa H., Sekine M., Nucleic Acids Res., 43, 5675–5686 (2015).
• 16) Hari Y., Kashima S., Matsuda Y., Sakata A., Takamine R., Ijitsu S., Obika S., Heterocycles, 90, 432–441 (2015).
• 17) Guianvarc’h D., Fourrey J.-L., Maurisse R., Sun J.-S., Benhida R., Bioorg. Med. Chem., 11, 2751–2759 (2003).
• 18) Wang Y., Rusling D. A., Powers V. E., Lack O., Osborne S. D., Fox K. R., Brown T., Biochemistry, 44, 5884–5892 (2005).
• 19) Notomi R., Wang L., Sasaki S., Taniguchi Y., RSC Adv., 11, 21390–21396 (2021).
• 20) Notomi R., Wang L., Osuki T., Okamura H., Sasaki S., Taniguchi Y., Bioorg. Med. Chem., 28, 115782 (2020).
• 21) Wang L., Taniguchi Y., Okamura H., Sasaki S., Nucleic Acids Res., 46, 8679–8688 (2018).
• 22) Okamura H., Taniguchi Y., Sasaki S., Angew. Chem. Int. Ed., 55, 12445–12449 (2016).
• 23) Taniguchi Y., Magata Y., Osuki T., Notomi R., Wang L., Okamura H., Sasaki S., Org. Biomol. Chem., 18, 2845–2851 (2020).
• 24) Notomi R., Sasaki S., Taniguchi Y., Nucleic Acids Res., 50, 12071–12081 (2022).