2022 Volume 70 Issue 10 Pages 726-730
Several psoralen-conjugated oligonucleotides (Ps-Oligos) have been developed as photo-crosslinkable oligonucleotides targeting DNA or RNA. To avoid potential off-target effects, it is important to investigate the selective photo-crosslinking reactivity of Ps-Oligos to DNA or RNA. However, the selectivity of these Ps-Oligos has not been reported in detail thus far. In this study, we evaluated the photo-crosslinking properties of two Ps-Oligos, 5′-Ps-Oligo and a novel Ps-Oligo containing 2′-O-{[(4,5′,8-trimethylpsoralen)-4′-ylmethoxy]ethylaminocarbonyl}adenosine (APs2-Oligo). Notably, 5′-Ps-Oligo preferentially crosslinked with DNA, whereas APs2-Oligo preferentially crosslinked with RNA. These results demonstrate the interesting crosslinking properties of Ps-Oligos, which will provide useful information for the molecular design of novel Ps-Oligos in future studies.
Psoralen derivatives are used as dermal photosensitizing molecules for the treatment of various skin pigmentation disorders.1–3) These compounds are also known as DNA intercalators and can form covalent bonds via the [2 + 2] cycloaddition reaction with pyrimidine bases upon UV irradiation.4,5) Therefore, psoralens have been applied as one of DNA intercrosslinking6,7) agents in various research fields.8–10) Several research groups have developed psoralen-conjugated oligonucleotides (Ps-Oligos) targeting DNA or RNA. These Ps-Oligos can be used to form crosslinkages in a sequence-specific manner via duplex formation. Some of Ps-Oligos have been used as antisense oligonucleotides and suppressed tumor growth by crosslinking with target mRNA and interfering with mRNA translation.11–14) In contrast, we have developed Ps-Oligos, which can distinguish a point mutation (from a purine base to a pyrimidine base) in DNA in a sequence-specific manner based on the photo-crosslinking strategy.15,16) These Ps-Oligos can be used for the detection of oncogenes, such as ras, egfr, myc, and erb-B, which are well-known causes of tumorigenesis.17,18) The selective detection method for these oncogenes may provide valuable insight into the development mechanisms of these diseases. Thus, Ps-Oligos are expected to be a genome-targeting technology.
As described above, Ps-Oligos are widely used as photo-crosslinkable oligonucleotides that target DNA and RNA. Therefore, to avoid potential off-target effects, it is essential to investigate the selective photo-crosslinking reactivity of Ps-Oligos to DNA and RNA. However, the selectivity of these Ps-Oligos has not been reported in detail thus far, and information on the relationship between the Ps-Oligo structure and its target selectivity is limited. This study aims to assess the photo-crosslinking property of 5′-Ps-Oligo (Fig. 1A), which possesses a psoralen moiety at its 5′-end, with cDNA and RNA. The psoralen moiety of 5′-Ps-Oligo is inserted into the 3′ side of its complementary nucleobase. Subsequently, a novel Ps-Oligo containing 2′-O-{[(4,5′,8-trimethylpsoralen)-4′-ylmethoxy]ethylaminocarbonyl}adenosine (APs2-Oligo, Fig. 1B) is synthesized and its photo-crosslinking property to cDNA and RNA is evaluated to further investigate the photo-crosslinking property of Ps-Oligos. It is proposed that by introducing psoralen at the 2′-O-position of adenosine, the psoralen moiety of Aps2-Oligo is inserted into the 5′ side of its complementary nucleobase. Thus, APs2-Oligo may crosslink with the same target thymidine (or uridine) as 5′-Ps-Oligo, which form base pairs with adenosine, in the target gene. Notably, APs2-Oligo showed significant photo-crosslinking efficiency with RNA compared with DNA, suggesting that the photo-crosslinking selectivity of Ps-Oligo can be controlled by structural modification of Ps-Oligo.
Chemical structure of psoralen-conjugated oligonucleotides (5′-Ps-Oligo (A) and Aps2-Oligo (B)), sequences of oligonucleotides, and the thermal stability of the duplex (Tm value) of the psoralen-conjugated oligonucleotides to its complementary oligonucleotides.
The preparation of APs2-Oligo was started with the synthesis of N6-Bz-5′-O-DMTr-2′-O-{[(4,5′,8-trimethylpsoralen)-4′-ylmethoxy]ethylaminocarbonyl}-adenosine-3′-(2-cyanoethyl-N,N′-diisopropylphosphoramidite) (Chart 1, 1). We decided to introduce psoralen via carbamate linkage, since it is synthetically easier to prepare than the previously reported 2′-O-psoralenylmethoxyethyladenosine.12) First, the alcohols at the 5′ and 3′ positions were simultaneously protected by disiloxane 2 according to a procedure reported in literature to give compound 3.19) Protected compound 3 was treated with carbonyldiimidazole to generate reactive intermediate 4 for the introduction of the psoralen units. Subsequently, intermediate 4 was mixed with 4′-(2-aminoethoxy)methyl-4,5′,8-trimethylpsoralen 5, which was prepared separately according to a reported procedure.20) This reaction gave compound 6 in 31% yield with a psoralen moiety at the 2′ position via an ethoxyaminocarbonyl linker. Next, the disiloxane group was removed in the presence of triethylamine trihydrofluoride (TEA‧3HF) to give diol 7 in 85% yield. The primary alcohol of 7 was capped with a 4,4′-dimethoxytrityl group to obtain compound 8 in 74% yield, and the final treatment of the secondary alcohol of 8 with 2-Cyanoethyl N,N,N,N-tetraisopropylphosphordiamidite produced compound 1 in 29% yield. Using the amidite compound 1, the synthesis of psoralen-introduced oligonucleotide (APs2-Oligo) was performed at Ajinomoto Genedesign (Osaka, Japan) as described in the supplementary materials.
The photo-crosslinking behaviors of 5′-Ps-Oligo and APs2-Oligo were evaluated and the results were compared (Fig. 1). The fidelity in target sequence recognition by oligonucleotides is an important factor that influences photo-crosslinking efficiency, which is attributed to the difference in the thermal stability of the target hybrids. To properly evaluate the photo-crosslinking selectivity of 5′-Ps-Oligo toward DNA or RNA, two sequences were prepared for target DNA and RNA, as shown in Fig. 1. These sequences showed slightly higher Tm value (Supplementary Fig. S5) during duplex formation with APs2-Oligo (Tm = 69 °C for DNA and RNA) (Fig. 1B) than 5′-Ps-Oligo (Tm = 68 °C for DNA and RNA) (Fig. 1A). These differences might be probably attributed to the additional base-pairs of adenosine of APs2-Oligo.
Subsequently, the photo-crosslinking efficiencies of 5′-Ps-Oligo and APs2-Oligo were evaluated using denaturing polyacrylamide gel electrophoresis (denaturing PAGE). UV irradiation was conducted for a maximum of 120 s, and samples were collected at different time steps (0, 1, 5, 10, 20, 30, 60, and 120 s). Consequently, 5′-Ps-Oligo was selectively crosslinked with DNA (Fig. 2A) but not efficiently with RNA (Fig. 2B). The crosslinking efficiency of 5′-Ps-Oligo with DNA was 40% higher than that with RNA. On the other hand, APs2-Oligo crosslinked with both DNA (Fig. 3A) and RNA (Fig. 3B), and interestingly, the crosslinking efficiency of APs2-Oligo with RNA was higher than that with DNA (Fig. 3C).
Analysis of the photo-crosslinking reaction with (A) DNA and (B) RNA was performed using denaturing PAGE (20% polyacrylamide/7 M urea/25% formamide/TBE). [5′-Ps-Oligo] = [DNA] = [RNA] = 1 µM in 10 mM phosphate buffer containing 0.1 M NaCl (pH 7.0). (C) Quantification of photocross-linking efficiency of 5′-Ps-Oligo with DNA or RNA.
Analysis of the photo-crosslinking reaction with (A) DNA and (B) RNA was performed using denaturing PAGE (20% polyacrylamide/7 M urea/25% formamide/TBE). [APs2-Oligo] = [DNA] = [RNA] = 1 µM in 10 mM phosphate buffer containing 0.1 M NaCl (pH 7.0). (C) Quantification of photocross-linking efficiency of APs2-Oligo with DNA or RNA.
To determine whether the RNA selectivity of APs2-Oligo in a crosslinking reaction is influenced by differences in the nucleobase structure (DNA: thymine, RNA: uracil), a photo-crosslinking experiment was conducted using DNA in which the target thymine was replaced with uracil (DNA-U). Interestingly, the crosslinking efficiencies of APs2-Oligo with DNA and DNA-U were nearly the same (Fig. 4). These results suggest that the preferential crosslinking of APs2-Oligo to RNA is not caused by structural differences in the target nucleobase. Based on these results, we propose that certain structural differences in the duplex (DNA duplex generally forms B-form double helix, while DNA–RNA duplex adopts the A-form for both strands)21–23) may affect the target preference of APs2-Oligo during photo-crosslinking. We conducted structural simulations of duplex models for both 5′-Ps-Oligo (Supplementary Fig. S6) and APs2-Oligo (Supplementary Fig. S7). The psoralen moiety of 5′-Ps-Oligo stacks with target thymine in DNA models, while the psoralen of 5′-Ps-Oligo is away from the target uracil in RNA models. This is considered as the reason why the 5′-Ps-Oligo crosslinked selectively to DNA. On the other hand, from the structural simulations, the psoralen moiety of APs2-Oligo stacks with both thymine (in DNA) and uracil (in RNA). These results can explain why APs2-Oligo could crosslinked with both DNA and RNA. Although we do not have enough and clear evidence to explain the RNA crosslinking preference of APs2-Oligo, APs2-Oligo probably forms a better crosslinking intermediate with the target RNA upon duplex formation, and there is no doubt that APs2-Oligo is more suitable than 5′-Ps-Oligo as a candidate of nucleic acid medicines targeting RNA.
(A) Sequences of DNA-U and APs2-Oligo. (B) Analysis of photo-crosslinking reaction with DNA-U was performed using denaturing PAGE (20% polyacrylamide/7 M urea/25% formamide/TBE). [APs2-Oligo] = [DNA-U] = 1 µM in 10 mM phosphate buffer containing 0.1 M NaCl (pH 7.0). (C) Quantification of photocross-linking efficiency of APs2-Oligo with DNA-U.
A novel oligonucleotide containing 2′-O-{[(4,5′,8-trimethyl- psoralen)-4′-ylmethoxy]ethylaminocarbonyl}adenosine (APs2-Oligo) are developed and the photo-crosslinking behavior of the APs2-Oligo toward target DNA and RNA was evaluated. APs2-Oligo preferentially crosslinked with RNA, whereas previously developed 5′-Ps-Oligo was crosslinked selectively to DNA. Our results suggest that the RNA-selective crosslinking property of APs2-Oligo is not caused by structural differences in the target nucleobase. From the structural simulation analysis, it is proposed that APs2-Oligo forms a more suitable crosslinking intermediate with the target RNA upon duplex formation. This study provides a valuable knowledge for the molecular design of novel Ps-Oligos targeting DNA (5′-Ps-Oligo) or RNA (APs2-Oligo) in the future. Further investigations are ongoing by our research group.
The UV melting profiles of the 5′-Ps-Oligo and APs2-Oligo duplexes with their target oligonucleotides were measured using a UV spectrophotometer equipped with a programmed thermal controller at a ramp rate of 1.0 °C/min. The sample solutions were prepared in 10 mM phosphate buffer (pH 7.0) containing 0.1 M NaCl, and the oligonucleotide concentration was set at 1.0 µM. Before the measurements, all the samples were heated at 95 °C for 5 min and slowly cooled from 95 to 25 °C at 1.0 °C/min for annealing.
Photo-crosslinking ReactionEquimolar mixtures of TAMRA labeled 5′-Ps-Oligo or APs2-Oligo and their complementary oligonucleotides were prepared in a similar manner to the samples prepared for UV-melting profile measurements. After incubation, the reaction mixtures were UV-irradiated for 0–120 s using a UV-LED irradiator (OMRON ZUV-C30H, 365 nm), followed by analysis using denaturing PAGE (20% polyacrylamide/7 M urea/25% formamide). The crosslinking efficiencies were quantified by measuring TAMRA fluorescence.
This study was supported by the Grant-in-Aid for Transformative Research Areas (A) “Material Symbiosis” (Grant Number: 20H05874 awarded to A.Y. and T.Y.) from MEXT, Japan. This study was also supported by JSPS KAKENHI (Grant Numbers 22H00593 and 22K14839 to A.Y. and Y.M., respectively), Japan.
The authors declare no conflict of interest.
This article contains supplementary materials.
