Communication Cite This: Bioconjugate Chem. XXXX, XXX, XXX−XXX
pubs.acs.org/bc
Post-Synthetic Modification of Oligonucleotides via Orthogonal Amidation and Copper Catalyzed Cycloaddition Reactions Daniel Zewge,*,† Gabor Butora,† Edward C. Sherer,† David M. Tellers,*,‡ Daniel R. Sidler,† Joseph Gouker,† Greg Copeland,† Vasant Jadhav,§ Zhen Li,† Joseph Armstrong,† and Ian W. Davies† †
Department of Process Chemistry, Merck & Co. Inc., Rahway, New Jersey 07065, United States Department of Medicinal Chemistry and §Department of RNA Biology, Merck & Co. Inc., West Point, Pennsylvania 19486, United States
‡
S Supporting Information *
ABSTRACT: An efficient multicomponent orthogonal protocol was developed for post-synthetic oligonucleotide modification using commercially available 2′-O-methyl ester and 2′-O-propargyl nucleoside scaffolds. Amidation of methyl esters with primary amines was achieved in the presence of 2′propargyl groups which were utilized for subsequent copper catalyzed cycloaddition with GalNAc-azide. The methodology was applied to generate siRNA composed of multiple amide and triazole conjugates. Computational methods were used to illustrate the impact of substitution at the 2′-position. This a powerful post-oligomerization technique for rapidly introducing diversity to oligonucleotide design.
T
methodology to a mild and robust amide formation reaction. Importantly, this reaction can be coupled with the azide cycloaddition reaction to further increase the diversity along the oligonucleotide backbone.
he therapeutic potential of siRNA has generally been regarded to be limited by the inability to safely and efficiently deliver the oligonucleotide double strand (guide and passenger strand) to the cytosol.1−3 Multiple approaches have been employed to overcome this limitation including encapsulation of siRNA in lipid nanoparticles, incorporation of siRNA with polymers, and direct conjugation of delivery facilitating groups to the siRNA.4−12 Enabled delivery vehicles, such as lipid nanoparticles and polymers, serve to protect the siRNA from the harsh endolysosomal environment. The conjugate approach allows for attachment of groups which target the oligonucleotide to specific cells and modulate the overall charge/hydrophobicity of the duplex for improved cellular uptake. Despite their differences, each approach fundamentally relies on extensive and selective modification of the native oligonucleotide backbone to control properties such as potency, immunogenicity, and chemical and enzymatic stability. The majority of these modifications are incorporated at the 2′ position of the nucleotide and are typically installed in the nucleoside prior to solid phase oligonucleotide synthesis (Figure 1). This approach limits rapid access to large, chemically diverse pools of modified oligonucleotides as each monomer building block must be prepared in a multistep manner prior to solid phase oligonucleotide synthesis. Postoligomerization modification, in contrast, provides a way of incorporating such diversity in a high throughput manner (Figure 1).13 To demonstrate this approach, we recently reported the development and application of a post-resin synthesis copper catalyzed azide cyclization on alkyne-modified oligonucleotides.14 Herein we describe the extension of this © XXXX American Chemical Society
■
RESULTS AND DISCUSSION Driven by our interest in optimizing delivery of siRNA to hepatocytes via multivalent N-acetylgalactosamine (GalNAc) and endosomal escape strategies, we sought to develop a methodology which would allow us to systematically examine both aspects in parallel without the need to develop the synthesis of the individual phosphoroamidite starting materials. Orthogonal methodologies have been extensively explored for conjugating small and large molecules on the 3′ and 5′ terminal positions of oligonucleotides.13 In contrast, to the best of our knowledge, orthogonal chemistry to modify the internal positions of siRNA strands has remained elusive.15 To accomplish this, we sought to identify reactions with mild and efficient protocols in this complex chemical environment. Given our success with the copper catalyzed azide cycloaddition reaction,14 we opted to identify an orthogonal reaction which was compatible with alkynes, high yielding and employed readily available starting materials. A selection of approaches are highlighted in Figure 2. Of the reactions examined, amidation proved to be the most robust in terms of meeting these criteria. While additional Received: April 29, 2018 Revised: May 31, 2018
A
DOI: 10.1021/acs.bioconjchem.8b00298 Bioconjugate Chem. XXXX, XXX, XXX−XXX
Communication
Bioconjugate Chemistry
Figure 1. Synthetic scheme highlighting two different approaches for diversity installation on the 2′ position of the nucleotide.
Figure 2. Selection of orthogonal reactions examined. The amidation reaction proved to be highly compatible with the cycloaddition reaction.
optimization of the other reactions was possible, we opted to focus on further examination of the amidation reaction. In particular, we envisioned developing and employing an orthogonal modification protocol for installing lipidic and carbohydrate substrates at multiple positions of the passenger strand (sense strand, Table 1) whose corresponding guide strand (antisense strand) is complementary to the apolipopro-
tein B mRNA. Optimized amidation conditions involved mild heating of the resin-bound oligonucleotide with pentyl amine (10 equiv per site) resulting in clean conversion to the desired carboxamide-containing nucleotide as determined by HPLC. The oligonucleotide-containing resin is then treated with 40% aqueous methylamine to remove the oligonucleotide from solid support and cleave remaining amide and cyanoethyl protecting groups.16 Reaction mixtures were lyophilized to remove volatiles, redissolved, and subjected to the previously disclosed cycloaddition reaction.14 A representative example of this process is shown in Figure 3. We sought to exemplify this protocol by preparing five positional modifications along the 2′ position of the nucleotides with pentyl amine and GalNac azide. We were pleased to find complete incorporation at the selected positions with both reactions when the amidation was performed first (Figure 3),
Table 1. ApoB Passenger Parent Sequence Modified Positions (5′ to 3′) ApoB sequence 5′-3′ Strand Strand Strand Strand
A B C D
amidation positions
triazole positions
4,8,12,16,20 12,14,16,18,20 13,14,15,16,17 8,10,12,14,16
2,6,10,14,18 2,4,6,8,10 5,6,7,8,9 7,9,11,13,15 B
DOI: 10.1021/acs.bioconjchem.8b00298 Bioconjugate Chem. XXXX, XXX, XXX−XXX
Communication
Bioconjugate Chemistry
Figure 3. Amidation and cycloaddition illustration to form Strand C
Modeling to Inform Design. The methodology described in this work provides a useful way for “presenting” different moieties off the three-dimensional oligonucleotide template. When design strategy is considered, leveraging computational approaches provides a convenient method for prioritizing designs and ensuring different spatial presentations are covered, i.e., a means for converting substitution patterns in Table 1 to three-dimensional figures. To illustrate this strategy, models of helices A−D were constructed in Schrödinger’s Maestro and are presented in Figure 5. Structures A and D highlight how dispersion of the substituents can be controlled by selection of where the substitution occurs. Structure A has the pentylamide and GalNAc residues spread uniformly throughout the helix whereas Structure D has pentylamide and GalNAc more closely bundled near the center of the helix. Structure B and C demonstrate the structural ramifications of sequential versus alternating substitutions. Structure B has sequential GalNAc and pentylamide substitutions resulting in GalNAc/pentylamide distribution in a 360° arc on the nucleotide. In constrast, structure C has alternating substitutions of GalNAc/pentylamide which result in these moieties being clustered on the same face of the oligonucleotide. Finally, Duplex C has the GalNAc ligand clustered relatively tightly compared to the other duplexes. The improved distribution through the other duplexes may be responsible for the improved binding compared to Duplex C.19,20 Additional experimentation across a multitude of positions should be performed in order to validate this hypothesis. Summary and Conclusions. We have developed an orthogonal oligonucleotide modification methodology which enables amide bond and triazole formation. This approach was successfully used for installing up to five pentylamines and five azido N-acetylgalactosamine ligands along the 2′ positions of the oligonucleotides. Computational methods were employed to aid in visualization of each structure. This platform can be used to access highly modified oligonucleotides with readily available starting materials and serves as a useful, differentiated extension of existing oligonucleotide synthesis methodologies.
followed by the copper catalyzed azide cycloaddition reaction (CuBr.SMe2/GalNAc azide in DMSO:H2O).17 Reversing the order of the reactions resulted in hydrolysis of methyl ester scaffolds generating 2′-carboxylc acids.18 The post-cycloaddition material was purified on a C-18 cartridge with on column deprotection of 5′-O-DMT using 1% TFA and aqueous acetonitrile (35%) wash to afford the final multiconjugated products in ∼20% overall yield (based on starting CPG) with residual copper 98% conversion compared to pentamethylamine adduct), and identity (see Supporting Information). CPGcleaved strands were then directly used for subsequent multiclick reactions. Procedure for Deacylation of GalNAc Azide. Fully protected GalNAc azide (1 g, 2.2 mmol) was treated with
excess methylamine (4 mL) and aged over 1 h at RT. Reaction mixture was lyophilized and reconstituted in DMSO/H2O (20 mL, ∼110 mM solution) and directly used for cycloaddition modifications. Procedure for Multclick of Passenger Strand with GalNAc Azide. To the lyophilized crude amidation product obtained above was added 900 μL of 20% aqueous ACN. To this mixture was added 50 μL of GalNAc azide (∼2 equiv per site) followed by 50 μL of CuBr·SMe2 (8 mM in DMSO, ∼0.2 equiv per site). Reaction mixture was aged in a glovebox at 50 °C over 8 h with stirring. The crude reaction mixture was checked for purity and identity via UPLC/MS (Supporting Information) and then subjected to spin dialysis to remove residual low molecular weight impurities (
