Synthesis of Site‐Specifically Phosphate‐Caged siRNAs and Evaluation of Their RNAi Activity and Stability

A complete set of new photolabile nucleoside phosphoramidites were synthesized, then site‐specifically incorporated into sense or antisense strands of siRNA for phosphate caging. Single caging modification was made along siRNA strands and their photomodulation of gene silencing were examined by using the firefly luciferase reporter gene. Several key phosphate positions were then identified. Furthermore, multiple caging modifications at these key positions led to significantly enhanced photomodulation of gene silencing activity, suggesting a synergistic effect. The caging group on both the terminally phosphate‐caged siRNA and the single‐stranded caged RNA has comparatively high stability, whereas hydrolysis of the caged group from the internally caged siRNA was observed, irrespective of the presence of Mg²⁺. Molecular dynamic simulations demonstrated that enhanced hydrolysis of the caging group on internally phosphate‐caged siRNAs was due to easy fragmentation of the caging group upon formation of the pentavalent intermediate of the phosphotriester with attack by water. The caging group in the terminally phosphate‐caged siRNA or single‐stranded caged RNA prefers to form π–π stacks with nearby nucleobases. In addition to providing explanations for previous observations, this study sheds further light on the design of caged oligonucleotides and indicates the direction of future development of nucleic acid drugs with phosphate modifications.     

Intrinsic Acid–Base Properties of a Hexa‐2′‐deoxynucleoside Pentaphosphate,d(ApGpGpCpCpT): Neighboring Effects and Isomeric Equilibria

The intrinsic acid‐base properties of the hexa‐2′‐deoxynucleoside pentaphosphate, d(ApGpGpCpCpT) [=(A1?G2?G3?C4?C5?T6)=(HNPP)^5?] have been determined by ¹H NMR shift experiments. The pK_a values of the individual sites of the adenosine (A), guanosine (G), cytidine (C), and thymidine (T) residues were measured in water under single‐strand conditions (i.e., 10 % D₂O, 47 °C, I=0.1 M , NaClO₄). These results quantify the release of H⁺ from the two (N7)H⁺ (G?G), the two (N3)H⁺ (C?C), and the (N1)H⁺ (A) units, as well as from the two (N1)H (G?G) and the (N3)H (T) sites. Based on measurements with 2′‐deoxynucleosides at 25 °C and 47 °C, they were transferred to pK_a values valid in water at 25 °C and I=0.1 M . Intramolecular stacks between the nucleobases A1 and G2 as well as most likely also between G2 and G3 are formed. For HNPP three pK_a clusters occur, that is those encompassing the pK_a values of 2.44, 2.97, and 3.71 of G2(N7)H⁺, G3(N7)H⁺, and A1(N1)H⁺, respectively, with overlapping buffer regions. The tautomer populations were estimated, giving for the release of a single proton from five‐fold protonated H₅(HNPP)^±, the tautomers (G2)N7, (G3)N7, and (A1)N1 with formation degrees of about 74, 22, and 4 %, respectively. Tautomer distributions reveal pathways for proton‐donating as well as for proton‐accepting reactions both being expected to be fast and to occur practically at no “cost”. The eight pK_a values for H₅(HNPP)^± are compared with data for nucleosides and nucleotides, revealing that the nucleoside residues are in part affected very differently by their neighbors. In addition, the intrinsic acidity constants for the RNA derivative r(A1?G2?G3? C4?C5?U6), where U=uridine, were calculated. Finally, the effect of metal ions on the pK_a values of nucleobase sites is briefly discussed because in this way deprotonation reactions can easily be shifted to the physiological pH range.     

Selective Enzymatic Demethylation of N2,N2‐Dimethylguanosine in RNA and Its Application in High‐Throughput tRNA Sequencing

The abundant Watson–Crick face methylations in biological RNAs such as N¹‐methyladenosine (m¹A), N¹‐methylguanosine (m¹G), N³‐methylcytosine (m³C), and N²,N²‐dimethylguanosine (m²₂G) cause significant obstacles for high‐throughput RNA sequencing by impairing cDNA synthesis. One strategy to overcome this obstacle is to remove the methyl group on these modified bases prior to cDNA synthesis using enzymes. The wild‐type E. coli AlkB and its D135S mutant can remove most of m¹A, m¹G, m³C modifications in transfer RNA (tRNA), but they work poorly on m²₂G. Here we report the design and evaluation of a series of AlkB mutants against m²₂G‐containing model RNA substrates that we synthesize using an improved synthetic method. We show that the AlkB D135S/L118V mutant efficiently and selectively converts m²₂G modification to N²‐methylguanosine (m²G). We also show that this new enzyme improves the efficiency of tRNA sequencing.     

Chemical Proteomic Discovery of Isotype-Selective Covalent Inhibitors of the RNA Methyltransferase NSUN2

5-Methylcytosine (m⁵C) is an RNA modification prevalent on tRNAs, where it can protect tRNAs from endonucleolytic cleavage to maintain protein synthesis. The NSUN family (NSUN1-7 in humans) of RNA methyltransferases are capable of installing the methyl group onto the C⁵ position of cytosines in RNA. NSUNs are implicated in a wide range of (patho)physiological processes, but selective and cell-active inhibitors of these enzymes are lacking. Here, we use cysteine-directed activity-based protein profiling (ABPP) to discover azetidine acrylamides that act as stereoselective covalent inhibitors of human NSUN2. Despite targeting a conserved catalytic cysteine in the NSUN family, the NSUN2 inhibitors show negligible cross-reactivity with other human NSUNs and exhibit good proteome-wide selectivity. We verify that the azetidine acrylamides inhibit the catalytic activity of recombinant NSUN2, but not NSUN6, and demonstrate that these compounds stereoselectively disrupt NSUN2-tRNA interactions in cancer cells, leading to a global reduction in tRNA m⁵C content. Our findings thus highlight the potential to create isotype-selective and cell-active inhibitors of NSUN2 with covalent chemistry targeting a conserved catalytic cysteine.     

RNA Methylation m6A: A New Code and Drug Target?

The m⁶A‐RNA modification is a dynamic and reversible process, which has emerged as a new RNA code for the regulation of gene expression. The functional network of methyltransferases (writers), demethylases (erasers), and binding proteins (readers) modulate the level of m⁶A modification. Dysfunction of RNA methylation has been associated with various fundamental biological processes and human diseases. Herein, we briefly introduce an understanding‐enabled manipulation on m⁶A‐RNA modification with an emphasis on the use of small‐molecule intervention.      

7‐Iodo‐5‐aza‐7‐deazaguanine ribonucleoside: crystal structure,physical properties,base‐pair stability and functionalization

The positional change of nitrogen‐7 of the RNA constituent guanosine to the bridgehead position‐5 leads to the base‐modified nucleoside 5‐aza‐7‐deazaguanosine. Contrary to guanosine, this molecule cannot form Hoogsteen base pairs and the Watson–Crick proton donor site N3—H becomes a proton‐acceptor site. This causes changes in nucleobase recognition in nucleic acids and has been used to construct stable `all‐purine' DNA and DNA with silver‐mediated base pairs. The present work reports the single‐crystal X‐ray structure of 7‐iodo‐5‐aza‐7‐deazaguanosine, C<sub>10</sub>H<sub>12</sub>IN<sub>5</sub>O<sub>5</sub> ( 1 ). The iodinated nucleoside shows an anti conformation at the glycosylic bond and an N conformation (O4′‐endo) for the ribose moiety, with an antiperiplanar orientation of the 5′‐hydroxy group. Crystal packing is controlled by interactions between nucleobase and sugar moieties. The 7‐iodo substituent forms a contact to oxygen‐2′ of the ribose moiety. Self‐pairing of the nucleobases does not take place. A Hirshfeld surface analysis of 1 highlights the contacts of the nucleobase and sugar moiety (O—H…O and N—H…O). The concept of pK‐value differences to evaluate base‐pair stability was applied to purine–purine base pairing and stable base pairs were predicted for the construction of `all‐purine' RNA. Furthermore, the 7‐iodo substituent of 1 was functionalized with benzofuran to detect motional constraints by fluorescence spectroscopy.     

Chemo‐Enzymatic Synthesis of Position‐Specifically Modified RNA for Biophysical Studies including Light Control and NMR Spectroscopy

The investigation of non‐coding RNAs requires RNAs containing modifications at every possible position within the oligonucleotide. Here, we present the chemo‐enzymatic RNA synthesis containing photoactivatable or ¹³C,¹⁵N‐labelled nucleosides. All four ribonucleotides containing ortho‐nitrophenylethyl (NPE) photocages, photoswitchable azobenzene C‐nucleotides and ¹³C,¹⁵N‐labelled nucleotides were incorporated position‐specifically in high yields. We applied this approach for the synthesis of light‐inducible 2′dG‐sensing riboswitch variants and detected ligand‐induced structural reorganization upon irradiation by NMR spectroscopy. This chemo‐enzymatic method opens the possibility to incorporate a wide range of modifications at any desired position of RNAs of any lengths beyond the limits of solid‐phase synthesis.     

Oligonucleotide Analogues with Integrated Bases and Backbone. Part 32

The G[s ]G dinucleoside 6 and the G[s ]G* dinucleoside 8 were prepared by alkylation of the guanosine thiols derived from 2 and 5 , respectively, by the C(8)‐chloromethylated guanosine 4 that was obtained from alcohol 3 . Dinucleosides 6 and 8 were deacylated to 7 and 9 , and fully deprotected to 10 and 11 , respectively. The G[n ]G dinucleoside 16 was obtained by reductive amination of aldehyde 13 with an iminophosphorane derived from azide 14 and deprotection of the resulting dimer 15 . In the solid state of 6 , and in a solution of 6 and 8 in CDCl₃, H? N(1/I) and H? N(1/II) are engaged in intramolecular H‐bonds to the C?O of the isobutyryl protecting groups, and HN of the isobutyryl group of unit I forms an interresidue, intramolecular H‐bond to N(7/II), leading to a syn orientation of the nucleobase at unit I, to a tg orientation of the sulfanyl moiety, and to an orthogonal orientation of the nucleobases, preventing any base pairing. The silylated and isopropylidenated dinucleosides 7 and 9 are present in DMSO solution as solvated monoplexes. Broad ¹H‐NMR signals of the nucleosides 7 and 16 in CHCl₃ solution evidence equilibrating G‐quadruplexes. The quadruplex formation of 7 and 16 was established by ¹H‐NMR spectroscopy (only of 16 ), vapour pressure osmometry, mass spectrometry, and CD spectroscopy. The C(6(I))‐hydroxymethylated analogue 9 in CDCl₃ and the fully deprotected dinucleosides 10 and 11 in H₂O form only weakly π? π stacked associates, but no G‐quadruplexes, as evidenced by CD spectroscopy.     

Novel Fluoride‐Labile Nucleobase‐Protecting Groups for the Synthesis of 3′(2′)‐O‐Aminoacylated RNA Sequences

With the aim to develop a general approach to a total synthesis of aminoacylated t‐RNAs and analogues, we describe the synthesis of stabilized, aminoacylated RNA fragments, which, upon ligation, could lead to aminoacylated t‐RNA structures. Novel RNA phosphoramidites with fluoride‐labile 2′‐O‐[(triisopropylsilyl)oxy]methyl (=tom) sugar‐protecting and N‐{{2‐[(triisopropylsilyl)oxy]benzyl}oxy}carbonyl (=tboc) base‐protecting groups were prepared (Schemes 4 and 5), as well as a solid support containing an immobilized N⁶‐tboc‐protected adenosine with an orthogonal (photolabile) 2′‐O‐[(S)‐1‐(2‐nitrophenyl)ethoxy]methyl (=(S)‐npeom) group (Scheme 6). From these building blocks, a hexameric oligoribonucleotide was prepared by automated synthesis under standard conditions (Scheme 7). After the detachment from the solid support, the resulting fully protected sequence 34 was aminoacylated with L ‐phenylalanine derivatives carrying photolabile N‐protecting groups (→ 42 and 43 ; Scheme 9). Upon removal of the fluoride‐labile sugar‐ and nucleobase‐protecting groups, the still stabilized, partially with the photolabile group protected precursors 44 and 45 , respectively, of an aminoacylated RNA sequence were obtained (Scheme 9 and Fig. 3). Photolysis of 45 under mild conditions resulted in the efficient formation of the 3′(2′)‐O‐aminoacylated RNA sequence 46 (Fig. 4). Additionally, we carried out model investigations concerning the stability of ester bonds of aminoacylated ribonucleotide derivatives under acidic conditions (Table) and established conditions for the purification and handling of 3′(2′)‐O‐aminoacylated RNA sequences and their stabilized precursors.     

A Vastly Increased Chemical Variety of RNA Modifications Containing a Thioacetal Structure

Recently discovered new chemical entities in RNA modifications have involved surprising functional groups that enlarge the chemical space of RNA. Using LC‐MS, we found over 100 signals of RNA constituents that contained a ribose moiety in tRNAs from E. coli. Feeding experiments with variegated stable isotope labeled compounds identified 37 compounds that are new structures of RNA modifications. One structure was elucidated by deuterium exchange and high‐resolution mass spectrometry. The structure of msms²i⁶A (2‐methylthiomethylenethio‐N6‐isopentenyl‐adenosine) was confirmed by methione‐D3 feeding experiments and by synthesis of the nucleobase. The msms²i⁶A contains a thioacetal, shown in vitro to be biosynthetically derived from ms²i⁶A by the radical‐SAM enzyme MiaB. This enzyme performs thiomethylation, forming ms²i⁶A from i⁶A in a first turnover. The new thioacetal is formed by a second turnover. Along with the pool of 36 new modifications, this work describes a new layer of RNA modification chemistry.     

The 2‐Cyano‐2,2‐dimethylethanimine‐N‐oxymethyl Group for the 2′‐Hydroxyl Protection of Ribonucleosides in the Solid‐Phase Synthesis of RNA Sequences

The reaction of 2‐cyano‐2‐methyl propanal with 2′‐O‐aminooxymethylribonucleosides leads to stable and yet reversible 2′‐O‐(2‐cyano‐2,2‐dimethylethanimine‐N‐oxymethyl)ribonucleosides. Following N‐protection of the nucleobases, 5′‐dimethoxytritylation and 3′‐phosphitylation, the resulting 2′‐protected ribonucleoside phosphoramidite monomers are employed in the solid‐phase synthesis of three chimeric RNA sequences, each differing in their ratios of purine/pyrimidine. When the activation of phosphoramidite monomers is performed in the presence of 5‐benzylthio‐1H‐tetrazole, coupling efficiencies averaging 99 % are obtained within 180 s. Upon completion of the RNA‐chain assemblies, removal of the nucleobase and phosphate protecting groups and release of the sequences from the solid support are carried out under standard basic conditions, whereas the cleavage of 2′‐O‐(2‐cyano‐2,2‐dimethylethanimine‐N‐oxymethyl) protective groups is effected (without releasing RNA alkylating side‐products) by treatment with tetra‐n‐butylammonium fluoride (0.5 m) in dry DMSO over a period of 24–48 h at 55 °C. Characterization of the fully deprotected RNA sequences by polyacrylamide gel electrophoresis (PAGE), enzymatic hydrolysis, and matrix‐assisted laser desorption/ionization (MALDI) mass spectrometry confirmed the identity and quality of these sequences. Thus, the use of 2′‐O‐aminooxymethylribonucleosides in the design of new 2′‐hydroxyl protecting groups is a powerful approach to the development of a straightforward, efficient, and cost‐effective method for the chemical synthesis of high‐quality RNA sequences in the framework of RNA interference applications.     

Photocaging strategy for functionalisation of oligonucleotides and its applications for oligonucleotide labelling and cyclisation

We have used a photocaging strategy to develop novel phosphoramidites and expand the repertoire of protecting groups for modification of oligonucleotides by solid-phase synthesis. We synthesised five photolabile phosphoramidites and four new photolabile controlled pore glasses (CPGs). By using these photolabile phosphoramidites and CPGs, modified oligodeoxynucleotides (ODNs) with phosphate, amine, acid, thiol and carbonyl moieties at 5' and/or 3' ends were readily synthesised. To the best of our knowledge, this is the first report of introducing a carbonyl at the 5' end and thiol groups at both ends of ODNs with photolabile modifiers. Terminal labelling was also easily realised in solution or by on-column solid-phase synthesis. By using the photolabile amine modifier and the photolabile acid CPG, cyclisation of an oligodeoxynucleotide was achieved with good yields. This study provides an alternative way to introduce functional groups into oligonucleotides and expand the scope of oligonucleotide bio-orthogonal labelling.     

Structural Basis for Recognition of Guanosine by a Synthetic Tricyclic Cytosine Analogue: Guanidinium G‐Clamp

An oligonucleotide analogue containing a novel heterocyclic analogue, the guanidinium G‐clamp, was designed to allow formation of five H‐bonds to guanosine. The guanidinium group was introduced postsynthetically by treatment of the deprotected oligonucleotide containing a free amino group with a solution of 1H‐pyrazole‐1‐carboxamidine and purified by a combination of size‐exclusion chromatography and reversed‐phase HPLC. A single incorporation of this modification into an oligodeoxynucleotide sequence was found to increase duplex stability by 13° and 16° per modification to RNA and DNA, respectively. Crystals of a self‐complementary decamer sequence containing this modification were grown and diffracted to 1‐Å resolution. The structure was solved by molecular replacement and revealed that the modification forms additional H‐bonds to O(6) and N(7) of guanosine through the amino and imino N‐atoms, respectively. The origins of enhanced duplex stability are also attributed to increased stacking interactions mediated by the phenoxazine moiety of the G‐clamp and formation of H‐bond networks between the positively charged guanidinium group, H₂O molecules, and negatively charged O‐atoms from phosphates on the adjacent strand.     

Nucleobase‐Modified PNA Suppresses Translation by Forming a Triple Helix with a Hairpin Structure in mRNA In Vitro and in Cells

Compounds that bind specifically to double‐stranded regions of RNA have potential as regulators of structure‐based RNA function; however, sequence‐selective recognition of double‐stranded RNA is challenging. The modification of peptide nucleic acid (PNA) with unnatural nucleobases enables the formation of PNA–RNA triplexes. Herein, we demonstrate that a 9‐mer PNA forms a sequence‐specific PNA–RNA triplex with a dissociation constant of less than 1 nm at physiological pH. The triplex formed within the 5′ untranslated region of an mRNA reduces the protein expression levels both in vitro and in cells. A single triplet mismatch destabilizes the complex, and in this case, no translation suppression is observed. The triplex‐forming PNAs are unique and potent compounds that hold promise as inhibitors of cellular functions that are controlled by double‐stranded RNAs, such as RNA interference, RNA editing, and RNA localization mediated by protein–RNA interactions.