Synthesis of Oligoribonucleotides Containing a 2′-Amino-5′-S-phosphorothiolate Linkage

Oligoribonucleotides containing a photocaged 2′-amino-5′-S-phophorothiolate linkage have potential applications as therapeutic agents and biological probes to investigate the RNA structure and function. We envisioned that oligoribonucleotides containing a 2′-amino-5′-S-phosphorothiolate linkage could provide an approach to identify the general base within catalytic RNAs by chemogenetic suppression. To enable preliminary tests of this idea, we developed synthetic approaches to a dinucleotide, trinucleotide, and oligoribonucleotide containing a photocaged 2′-amino-5′-S-phosphorothiolate linkage. We incorporated the photocaged 2′-amino-5′-S-phosphorothiolate linkage into an oligoribonucleotide substrate for the hepatitis delta virus (HDV) ribozyme and investigated the pH dependence of its cleavage following UV irradiation both in the presence and absence of the ribozyme. The substrate exhibited a pH-rate profile characteristic of the modified linkage but reacted slower when bound to the ribozyme. Cleavage inhibition by the HDV ribozyme could reflect a non-productive ground-state interaction with the modified substrate’s nucleophilic 2′-NH2 or a poor fit of the modified transition state at the ribozyme’s active site.


■ INTRODUCTION
The synthesis of modified nucleosides, nucleotides, and oligonucleotides has been extensively investigated and motivated, in part, by creation of potential therapeutic agents (antisense, antiviral, and anticancer agents) 1−6 and biological probes for the investigation of the relationship between the RNA structure and function. 7 Chemical synthesis provides access to both naturally occurring and designed modified nucleotides and oligonucleotides, endowing biochemists and chemical biologists with tools to probe RNA chemistry and biology deeply and comprehensively. 8,9 For example, the replacement of RNA's 2′-OH with a 2′-NH 2 ( Figure 1A) maintains the hydrogen bonding capacity of 2′-OH but alters nucleophilicity, pK a , and metal-ion coordination properties. 10 These defined changes in chemical properties form the basis of biochemical strategies to define the functional roles of RNA's 2′-hydroxyl groups at specific locations. Owing to the weak nucleophilicity of the amino group toward the adjacent phosphodiester bond, 2′-amino substitution renders the ribose phosphate backbone inert to cleavage via internal transphosphorylation. 11 Analogously, substitution of the 5′-bridging oxygen atom of the phosphodiester linkage with a sulfur atom ( Figure 1B) alters hydrogen bonding, metal-ion coordination properties, and leaving group ability. However, in contrast to the 2′-amino group, a 5′-sulfur renders the phosphodiester backbone much more susceptible to transphosphorylation, owing to the greater leaving ability of sulfur relative to oxygen. This hyperactivation of the leaving group underpins a chemogenetic strategy to identify groups that activate the 5′oxygen leaving group within the active site of a biological catalyst. 12−14 These 2′-NH 2 and 5′-S-RNA modifications have been used independently in studies of RNA, including as mechanistic probes for ribozyme-catalyzed reactions. 12−32 Many reports have described the synthesis and incorporation of 2′-aminomodified nucleosides or nucleotides into RNA, including as a photocaged precursor ( Figure 1A,A*). 15−31 Nevertheless, 2′-NH 2 substitution of the nucleophilic 2′-OH at the cleavage site of an endonucleolytic ribozyme has limited use as a mechanism probe on its own because the modification essentially abolishes cleavage. 27,32 In contrast, the inherent instability of RNA containing a 5′-S-phosphorothiolate linkage makes working with this modification more challenging ( Figure 1B). Protection of 2′-hydroxyl with a photolabile group such as an o-nitrobenzyl group, which could be removed by UV irradiation, has facilitated the use of this modification (Figure 1B*). 12−14 We previously developed a strategy to identify the general acid in an enzymatic reaction using sulfur substitution of the leaving group. 12,13 The better leaving ability of the sulfur obviates the need for general acid catalysis. As a consequence, mutations to the general acid that adversely affect catalysis in the context of the natural oxygen leaving group become suppressed in the context of the sulfur leaving group, whereas mutations elsewhere remain deleterious.
We have been interested in an analogous strategy to identify a potential general base in catalysis. However, there appear to be no simple chemical modifications of the 2′-hydroxyl group that would suppress the need for a general base. A nucleotide analogue whose nucleophilic hydroxyl group ionizes fully within the pH range of the ribozyme reaction could provide a suitable probe, but this is not obviously accessed within the nucleotide framework. An amino group represents another possibility, but as noted above, oligonucleotides bearing 2′amino groups do not undergo backbone cleavage via attack of the nitrogen at the adjacent phosphorus center.
Alternatively, Eckstein and co-workers have shown that cleavage of a dinucleotide with a 2′-amino group can occur readily when the adjacent phosphorus bears a 5′-sulfur leaving group (k ∼ 10 −4 s −1 with half-life time ∼2 h). 33 Moreover, the cleavage reaction occurs independently of pH at pH values >7, indicating no susceptibility of the linkage to base catalysis. Accordingly, mutations that disable the ability of a ribozyme to deprotonate the nucleophile would be expected to affect cleavage of a substrate containing a 2′-amino group nucleophile and a 5′-S leaving group less adversely than a substrate containing only the 5′-S leaving group (Figure 2). Testing this approach in a ribozyme reaction requires installation of 2′-NH 2 , 5′-S modifications beyond dinucleotides and into oligoribonucleotides. Here, we report the synthesis of oligoribonucleotides containing a photocaged 2′-amino-5′-Sphosphorothiolate linkage (Figure 1C*) and determine its cleavage rate versus pH in the presence and absence of the hepatitis delta virus (HDV) ribozyme.    35 We adapted these two strategies to enable the synthesis of RNAs containing 2′-amino-5′-Sphosphorothiolate linkages. We prepared the 2′-photocaged 2′-amino-5′-S-dinucleotide (5′-C 2′-NHX -ps-G-3′), the trinucleotide derivative (5′-C 2′-NHX -ps-GG-3′), and 2′-photocaged 2′amino-3′-phosphoramidites.
Unfortunately, the trinucleotide 13 synthesized either by solid-phase synthesis or by solution methods still failed to afford the full-length RNA due to the failure of the second ligation step of our two-step ligation approach. We then investigated a possible solid-phase synthetic approach.
The corresponding RNAs containing a photocaged 2′amino-5′-O-phosphonate linkage (26a and 26b) were prepared by the solid-phase synthesis with the first coupling to phosphoramidite 25 40 and then coupling to 4a (Scheme 9).
Characterization and pH-Dependent Cleavage of a Ribozyme Substrate Containing a 2′-Amino-5′-S-phosphorothiolate Linkage. All photocaged RNA oligonucleotides 24a, 24b, 26a, and 26b were analyzed by MALDI-TOF mass spectrometry, confirming their molecular weights. HPLC confirmed that under neutral conditions, 26a and 26b were photodeprotected to the corresponding 5′-C 2′-NH2 -GGGUCGGC-3′ (∼25% conversion) and 5′-UUC 2′-NH 2 -GGGUCGGC-3′ (∼30% conversion) after UV irradiation (365 nm, 15−30 min). The UV deprotection rates were k (26a) = 0.27 min −1 and k (26b) = 0.065 min −1 , respectively. The photodeprotection of the shorter oligonucleotide (26a, 9 mer) occurred about 4 times faster than the longer oligonucleotide (26b, 11 mer). After 3′-radiolabeling, the RNA oligonucleotides 24b and 26b were treated with Ag + solution. As expected, 24b cleaves in the presence of Ag + ion, confirming the We then studied the pH-dependent cleavage reaction of 5′radiolabeled 24b in the presence and absence of the antigenomic HDV ribozyme 12,41 (1 μM) and 10 mM MgCl 2 ( Figure 5). As expected, the cleavage rate of 24b increases in a log-linear fashion at pH values below the pK a of the 2′-amino group and becomes independent of pH at pH values above the pK a (6.2). 42 This pH rate profile resembles that for the cleavage of the corresponding U 2′-NH 2 -ps-U dinucleotide. 33 We found that in the presence of the HDV ribozyme, 24b underwent cleavage 3−9-fold slower than in the absence of the HDV ribozyme throughout the tested pH range. This result indicates that ribozyme binding to the substrate inhibits cleavage of the 2′-amino-5′-S-phosphorothiolate linkage. The inhibition may reflect a non-productive ground-state interaction involving 2′-OH in the natural reaction. 43−46 The possible non-productive ground-state interactions in the enzyme substrate complex most likely involve hydrogen bonding or metal coordination to the nucleophilic amino group. These interactions could diminish nucleophilicity through interaction with the amino group's lone pair of electrons or disfavor acquisition of the in-line conformation required for reaction. 46 Alternatively, the HDV ribozyme may not be able to accommodate the transition state for 2′-Ntransphosphorylation of the 2′-amino-5′-S-phosphorothiolate linkage. We have shown previously using model systems that amine nucleophiles react at phosphodiesters bearing sulfurleaving groups via expanded transition states, with less bonding to both the nucleophile and the leaving group, relative to analogous reactions of phosphodiesters bearing oxygen leaving groups. 47 Possibly, the expanded transition state does not fit well at the HDV-active site, resulting in slower cleavage relative to the corresponding reaction in the absence of ribozyme.
Synthesis by the Coupling of 5 to 8c. Under argon, N,Obis(trimethylsilyl)trifluoroacetamide (0.135 mL, 0.51 mmol) was added to a solution of 5 (18 mg, 0.017 mmol) and 8c (19 mg, 0.026 mmol) in anhydrous THF (5.0 mL). The mixture was stirred under reflux for 1.5 h. The solvent was removed, and the residue was treated with 80% acetic acid (5.0 mL) at rt for 10 min. After evaporation of the mixture under vacuum, the remaining residue was treated with a mixture of NH 4 OH/EtOH (3:1, v/v) (8.0 mL) at 55°C in a sealed vial in an oven for 4 h. After cooling the mixture, the solvent was removed, and the residue was treated with 0.6 mL of NMP/Et 3 N/ Et 3 N−3HF solution (by mixing 275 μL of NMP and 140 μL of triethylamine with 180 μL Et 3 N−3HF) at 65°C in a water bath for 1 h. The reaction mixture was diluted with water (1 mL) and washed with chloroform (3 × 0.2 mL). The aqueous phase was evaporated, and the residue was purified by silica gel chromatography, eluting with acetonitrile/water/triethylamine (90:10:1 v/v/v) to afford the desired dinucleotide 11 as a colorless film (12 mg, 75% yield). The purity of 11 is estimated to be ∼90% by reverse-phase HPLC using a C18 column (HPLC conditions: Thermo Scientific Acclaim C18, 5 μm 120 Å 4. Trinucleotide 5′-C 2′-NHX -ps-GG-3′ (13). Synthesis via Solid-Phase Synthesis. The synthesis was started by using an Expedite 8909 synthesizer via a modified 1 μmol RNA protocol (trityl on). After standard detritylation, the 1 μmol i-Pr-Pac-G-RNA-CPG column was double coupling to 5′-tritylthioguanosine phosphoramidite (12) 34 (68 mg, 0.075 mmol) in dry acetonitrile (0.75 mL) and followed by standard capping and oxidation. The CPG column was then removed from the synthesizer and treated with the solution of AgNO 3 (26 mg) in water (3 mL) at rt for 1 h. The CPG column was washed with water (10 mL) and further treated with the solution of DTT (23 mg) in water (3 mL) at rt for 30 min. The CPG column was subsequently rinsed with water (5 mL), acetonitrile (5 mL), and CH 2 Cl 2 (5 mL). The CPG column was then treated with the solution of 2,2′dithiobis(5-nitropyridine) (46.5 mg, 0.150 mmol) in dry DMF (3 mL) at rt overnight. The CPG column was rinsed with acetonitrile (5 mL) and CH 2 Cl 2 (5 mL) and dried under vacuum for 30 min. Under argon, N,O-bis(trimethylsilyl)trifluoroacetamide (80 μL, 0.30 mmol) was added to a dried 10 mL flask containing the dried CPG and the 2′-photocaged amino 3′-H-phosphonate (5) (10 mg, 10 μmol) in anhydrous THF (3 mL). The mixture was stirred under argon at reflux for 1 h. The solvent was removed, and the solid supports were treated with 3% trichloroacetic acid (2 mL) in CH 2 Cl 2 at rt for 5 min. After rinsing with CH 2 Cl 2 (5 mL), the supports were treated with a mixture of concentrated ammonium hydroxide/ethanol (3:1, v/v) (2 mL) at rt overnight and then at 55°C for 1 h in a sealed tube in an oven. After cooling down in ice, the supernatant solution was removed, and the support was rinsed with an ethanol/acetonitrile/ water (3:1:1) mixture. The solutions were combined and evaporated to dryness. The residue was desilylated with a mixture of NMP/Et 3 N/  (15). 5′-O-Dimethoxytrityl-N 2 -phenoxyacetylguanosine (14) (0.500 g, 0.695 mmol) was co-evaporated with toluene (2 × 10 mL), dried under vacuum, and then dissolved into DMF (10 mL). To the resulting solution, imidazole (1.62 g, 23.9 mmol) was added, followed by TBSCl (746 mg, 5.00 mmol). The mixture was stirred under argon at rt overnight. The solvent was removed, and the residue was dissolved into dichloromethane (30 mL). The dichloromethane solution was washed with saturated aqueous NaHCO 3 and brine. The solvent was removed, and the residue was purified by silica gel chromatography, eluting with 2% methanol in chloroform to afford 15 as a white foam: 0.519 g (79% yield). 1  2′,3′-O-Di-(tert-butyldimethylsilyl)-N 2 -phenoxyacetylguanosine (16). Compound 15 (0.438 g, 0.462 mmol) was treated with 3% trichloroacetic acid in dichloromethane (10 mL) at rt for 5 min. The mixture was diluted with dichloromethane (20 mL). The solution was washed with saturated aqueous NaHCO 3 and brine. The solvent was removed, and the residue was purified by silica gel chromatography, eluting with 5% methanol in chloroform to afford 16 as a white foam: 0.280 g, (94% yield). 1 13  Trinucleotide 5′-C 2′-NHX -ps-GG-3′ (13). Synthesis via Solution Method. Under argon to a solution of guanosine derivative 16 (71 mg, 0.11 mmol) and 5′-disulfide phosphoramidite 17 34 (91 mg, 0.11 mmol) in CH 3 CN (1.0 mL), the standard activator solution (0.45 tetrazole in acetonitrile, 1.0 mL) was added. After stirring the mixture at rt for 30 min, it was oxidized with 10% tert-butyl hydroperoxide (1.0 mL) at rt for 10 min. The solvent was removed, and the residue was isolated by silica gel chromatography, eluting with 5% methanol in chloroform to afford the protected dinucleotide derivative 18: 54 mg (37% yield). 31 ) were added, and the mixture was stirred at rt for 24 h. The solvent was removed, and the residue was dissolved into CH 2 Cl 2 . The solution was subsequently washed with saturated NaHCO 3 , water, and brine and dried over anhydrous MgSO 4 . The solvent was removed, and the residue was treated with the solution of 2,2′-dithiobis(5-nitropyridine) (11 mg, 36 μmol) in dry DMF (2 mL) at rt overnight. The solvent was removed, and the residue was purified by silica gel chromatography, eluting with 5% MeOH in CH 2 Cl 2 containing 5% Et 3 N to afford a mixture of 5′disulfide derivatives (confirmed by MS). Under argon, to the mixture of the above-prepared 5′-disulfide derivatives in THF (10 mL), N,Obis(trimethylsilyl)trifluoroacetamide (319 μL, 1.20 mmol) and the 2′photocaged amino 3′-H-phosphonate (5) (13 mg, 12 μmol) were added. The mixture was stirred under argon at reflux for 1 h. The solvent was removed, and the residue was treated with 80% AcOH (3 mL) in CH 2 Cl 2 at rt for 30 min. The solvent was removed, and the residue was dried under vacuum and then treated with saturated ammonia in methanol in a 4°C refrigerator overnight. The solvent was removed, and the residue was desilylated with a mixture of NMP (450 μL), Et 3 N (225 μL), and Et 3 N−3HF (300 μL) at 65°C in a water bath for 25 min. The solvent was removed at rt under vacuum. The residue was dissolved into water (1.5 mL) and washed with chloroform (4 × 1.0 mL). The aqueous phase was desalted via a C18 Sep-Pak column. The product was then purified by reverse-phase HPLC column to afford the desired trinucleotide 13 as a colorless film  (20). Compound 19 35 (723 mg, 1.48 mmol) was dried by co-evaporation with dry pyridine (2 × 5 mL) under vacuum. Under argon, to the solution of dried 19 in dry pyridine (10 mL), TsCl (423 mg, 2.22 mmol) was added, and the mixture was stirred at rt for 40 h. The reaction was quenched by the addition of methanol (1.0 mL). After 10 min, the solvent was removed, and the residue was isolated by silica gel chromatography, eluting with 1.5−3% methanol in dichloromethane to afford 20 as a white foam: 526 mg (55% yield). 1  5′-Acetylthio-5′-deoxy-N 2 -isobutyryl-2′-O-(o-nitrobenzyl)guanosine (21a) and 3′-O, 5′-S-Diacetyl-5′-deoxy-N 2 -isobutyryl-2′-O-(o-nitrobenzyl)-5′-thioguanosine (21b). Under argon, to the solution of 20 (676 mg, 1.05 mmol) in DMF (10 mL), potassium thioacetate (240 mg, 2.10 mmol) was added. The mixture was stirred at 60°C in an oil bath for 17 h. The solvent was removed, and the residue was isolated by silica gel chromatography, eluting with 1−3% methanol in dichloromethane to afford 21a: 365 mg (63% yield, lower spot on TLC) and 21b: 127 mg (21% yield, higher spot on TLC) as light yellow foams.