Oligonucleotide synthesis is the chemical synthesis of relatively short fragments of nucleic acids with defined chemical structure (sequence). The technique is extremely useful in current laboratory practice because it provides a rapid and inexpensive access to custom-made oligonucleotides of the desired sequence. Whereas enzymes synthesize DNA and RNA only in a 5' to 3' direction, chemical oligonucleotide synthesis does not have this limitation, although it is most often carried out in the opposite, 3' to 5' direction. Currently, the process is implemented as solid-phase synthesis using phosphoramidite method and phosphoramidite building blocks derived from protected 2'-deoxynucleosides (dA, dC, dG, and T), ribonucleosides (A, C, G, and U), or chemically modified nucleosides, e.g. LNA or BNA.
To obtain the desired oligonucleotide, the building blocks are sequentially coupled to the growing oligonucleotide chain in the order required by the sequence of the product (see Synthetic cycle below). The process has been fully automated since the late 1970s. Upon the completion of the chain assembly, the product is released from the solid phase to solution, deprotected, and collected. The occurrence of side reactions sets practical limits for the length of synthetic oligonucleotides (up to about 200 nucleotide residues) because the number of errors accumulates with the length of the oligonucleotide being synthesized. Products are often isolated by high-performance liquid chromatography (HPLC) to obtain the desired oligonucleotides in high purity. Typically, synthetic oligonucleotides are single-stranded DNA or RNA molecules around 15–25 bases in length.
Oligonucleotides find a variety of applications in molecular biology and medicine. They are most commonly used as antisense oligonucleotides, small interfering RNA, primers for DNA sequencing and amplification, probes for detecting complementary DNA or RNA via molecular hybridization, tools for the targeted introduction of mutations and restriction sites, and for the synthesis of artificial genes. An emerging application of Oligonucleotide synthesis is the re-creation of viruses from sequence alone — either harmless, such as Phi X 174, or dangerous such as the 1917 influenza virus or SARS-CoV-2.
History The evolution of oligonucleotide synthesis saw four major methods of the formation of internucleosidic linkages and has been reviewed in the literature in great detail.
Early work and contemporary H-phosphonate synthesis
In the early 1950s, Alexander Todd's group pioneered H-phosphonate and phosphate triester methods of oligonucleotide synthesis. The reaction of compounds 1 and 2 to form H-phosphonate diester 3 is an H-phosphonate coupling in solution while that of compounds 4 and 5 to give 6 is a phosphotriester coupling (see phosphotriester synthesis below).
Thirty years later, this work inspired, independently, two research groups to adopt the H-phosphonate chemistry to the solid-phase synthesis using nucleoside H-phosphonate monoesters 7 as building blocks and pivaloyl chloride, 2,4,6-triisopropylbenzenesulfonyl chloride (TPS-Cl), and other compounds as activators. The practical implementation of H-phosphonate method resulted in a very short and simple synthetic cycle consisting of only two steps, detritylation and coupling (Scheme 2). Oxidation of internucleosidic H-phosphonate diester linkages in 8 to phosphodiester linkages in 9 with a solution of iodine in aqueous pyridine is carried out at the end of the chain assembly rather than as a step in the synthetic cycle. If desired, the oxidation may be carried out under anhydrous conditions. Alternatively, 8 can be converted to phosphorothioate 10 or phosphoroselenoate 11 (X = Se), or oxidized by CCl4 in the presence of primary or secondary amines to phosphoramidate analogs 12. The method is very convenient in that various types of phosphate modifications (phosphate/phosphorothioate/phosphoramidate) may be introduced to the same oligonucleotide for modulation of its properties.
Most often, H-phosphonate building blocks are protected at the 5'-hydroxy group and at the amino group of nucleic bases A, C, and G in the same manner as phosphoramidite building blocks (see below). However, protection at the amino group is not mandatory.
Phosphodiester synthesis
In the 1950s, Har Gobind Khorana and co-workers developed a phosphodiester method where 3'-O-acetylnucleoside-5'-O-phosphate 2 (Scheme 3) was activated with N,N-dicyclohexylcarbodiimide (DCC) or 4-toluenesulfonyl chloride (Ts-Cl). The activated species were reacted with a 5'-O-protected nucleoside 1 to give a protected dinucleoside monophosphate 3. Upon the removal of 3'-O-acetyl group using base-catalyzed hydrolysis, further chain elongation was carried out. Following this methodology, sets of tri- and tetradeoxyribonucleotides were synthesized and were enzymatically converted to longer oligonucleotides, which allowed elucidation of the genetic code. The major limitation of the phosphodiester method consisted in the formation of pyrophosphate oligomers and oligonucleotides branched at the internucleosidic phosphate. The method seems to be a step back from the more selective chemistry described earlier; however, at that time, most phosphate-protecting groups available now had not yet been introduced. The lack of the convenient protection strategy necessitated taking a retreat to a slower and less selective chemistry to achieve the ultimate goal of the study.
Phosphotriester synthesis
In the 1960s, groups led by R. Letsinger and C. Reese developed a phosphotriester approach. The defining difference from the phosphodiester approach was the protection of the phosphate moiety in the building block 1 (Scheme 4) and in the product 3 with 2-cyanoethyl group. This precluded the formation of oligonucleotides branched at the internucleosidic phosphate. The higher selectivity of the method allowed the use of more efficient coupling agents and catalysts, which dramatically reduced the length of the synthesis. The method, initially developed for the solution-phase synthesis, was also implemented on low-cross-linked "popcorn" polystyrene, and later on controlled pore glass (CPG, see "Solid support material" below), which initiated a massive research effort in solid-phase synthesis of oligonucleotides and eventually led to the automation of the oligonucleotide chain assembly.
Phosphite triester synthesis In the 1970s, substantially more reactive P(III) derivatives of nucleosides, 3'-O-chlorophosphites, were successfully used for the formation of internucleosidic linkages. This led to the discovery of the phosphite triester methodology. The group led by M. Caruthers took the advantage of less aggressive and more selective 1H-tetrazolidophosphites and implemented the method on solid phase. Very shortly after, the workers from the same group further improved the method by using more stable nucleoside phosphoramidites as building blocks. The use of 2-cyanoethyl phosphite-protecting group in place of a less user-friendly methyl group led to the nucleoside phosphoramidites currently used in oligonucleotide synthesis (see Phosphoramidite building blocks below). Many later improvements to the manufacturing of building blocks, oligonucleotide synthesizers, and synthetic protocols made the phosphoramidite chemistry a very reliable and expedient method of choice for the preparation of synthetic oligonucleotides.
Synthesis by the phosphoramidite method
Building blocks
Nucleoside phosphoramidites
As mentioned above, the naturally occurring nucleotides (nucleoside-3'- or 5'-phosphates) and their phosphodiester analogs are insufficiently reactive to afford an expeditious synthetic preparation of oligonucleotides in high yields. The selectivity and the rate of the formation of internucleosidic linkages is dramatically improved by using 3'-O-(N,N-diisopropyl phosphoramidite) derivatives of nucleosides (nucleoside phosphoramidites) that serve as building blocks in phosphite triester methodology. To prevent undesired side reactions, all other functional groups present in nucleosides have to be rendered unreactive (protected) by attaching protecting groups. Upon the completion of the oligonucleotide chain assembly, all the protecting groups are removed to yield the desired oligonucleotides. Below, the protecting groups currently used in commercially available and most common nucleoside phosphoramidite building blocks are briefly reviewed: The 5'-hydroxyl group is protected by an acid-labile DMT (4,4'-dimethoxytrityl) group. Thymine and uracil, nucleic bases of thymidine and uridine, respectively, do not have exocyclic amino groups and hence do not require any protection. Although the nucleic base of guanosine and 2'-deoxyguanosine does have an exocyclic amino group, its basicity is low to an extent that it does not react with phosphoramidites under the conditions of the coupling reaction. However, a phosphoramidite derived from the N2-unprotected 5'-O-DMT-2'-deoxyguanosine is poorly soluble in acetonitrile, the solvent commonly used in oligonucleotide synthesis. In contrast, the N2-protected versions of the same compound dissolve in acetonitrile well and hence are widely used. Nucleic bases adenine and cytosine bear the exocyclic amino groups reactive with the activated phosphoramidites under the conditions of the coupling reaction. By the use of additional steps in the synthetic cycle or alternative coupling agents and solvent systems, the oligonucleotide chain assembly may be carried out using dA and dC phosphoramidites with unprotected amino groups. However, these approaches currently remain in the research stage. In routine oligonucleotide synthesis, exocyclic amino groups in nucleosides are kept permanently protected over the entire length of the oligonucleotide chain assembly. The protection of the exocyclic amino groups has to be orthogonal to that of the 5'-hydroxy group because the latter is removed at the end of each synthetic cycle. The simplest to implement, and hence the most widely used, strategy is to install a base-labile protection group on the exocyclic amino groups. Most often, two protection schemes are used. In the first, the standard and more robust scheme (Figure), Bz (benzoyl) protection is used for A, dA, C, and dC, while G and dG are protected with isobutyryl group. More recently, Ac (acetyl) group is used to protect C and dC as shown in Figure. In the second, mild protection scheme, A and dA are protected with isobutyryl or phenoxyacetyl groups (PAC). C and dC bear acetyl protection, and G and dG are protected with 4-isopropylphenoxyacetyl (iPr-PAC) or dimethylformamidino (dmf) groups. Mild protecting groups are removed more readily than the standard protecting groups. However, the phosphoramidites bearing these groups are less stable when stored in solution. The phosphite group is protected by a base-labile 2-cyanoethyl protecting group. Once a phosphoramidite has been coupled to the solid support-bound oligonucleotide and the phosphite moieties have been converted to the P(V) species, the presence of the phosphate protection is not mandatory for the successful conducting of further coupling reactions.
In RNA synthesis, the 2'-hydroxy group is protected with TBDMS (t-butyldimethylsilyl) group. or with TOM (tri-iso-propylsilyloxymethyl) group, both being removable by treatment with fluoride ion. The phosphite moiety also bears a diisopropylamino (iPr2N) group reactive under acidic conditions. Upon activation, the diisopropylamino group leaves to be substituted by the 5'-hydroxy group of the support-bound oligonucleotide (see "Step 2: Coupling" below).
Non-nucleoside phosphoramidites