Synthesis and Evaluation of Oligonucleotides of High Purity

Abstract

We have developed and evaluated methods for the production of highly pure oligonucleotides.

Presently the solid phase synthesis in an automated DNA synthesiser applying the phosphoramidite chemistry can be regarded as a standard. During the synthesis several undesirable by-products arise:

- incomplete coupling (1%) leads to 5′-truncated sequences. These sequences are acetylated at their 5′-hydroxyl group to prevent further elongation in subsequent coupling steps, but this “capping step” is incomplete, the capping-yield is 90%, leading to accumulation of sequences of the length n-1 with internal deletions.

- the glycosidic bond to N-protected purines, especially adenine, is susceptible to acid leading to depurination and subsequently to strand scission during alkaline deprotection of the oligonucleotide. This gives rise to 3′- and to 5′-truncated sequences. The 3′-truncated sequences will not be removed by standard Rp HPLC as they are tritylated.

- the reactions involved in synthesis and deprotection may cause base modifications (full length product with damaged bases).

- insufficient deprotection procedures may result in incomplete removal of protecting groups, especially from the bases (full length products with altered bases).

We have set up two different schemes (Fig. 1 and Fig. 2) for synthesis and purification, which should provide highly pure oligonucleotides with the potential of adapting to large scale production:

- accumulation of n-1 sequences (failure of capping) will be avoided by a double capping procedure using phosphite in the first capping step and an acetic anhydride capping reagent in the second capping step, as described in the literature¹.

- 3′-truncated sequences are removed by different methqds in the two schemes. In scheme I (Fig. 1) the 3′-truncated sequences can be washed off, as the 3′-full length product still is anchored to the solid support after deprotection. In scheme II (Fig. 2) the 3′truncated sequences are digested by snake venom phosphodiesterase. The 3′-full length product is protected against digestion by a 3′ - 3′-inverted end. An oligo with a correct 3′-end is, in both schemes, eventually obtained by cleaving with RNase between the ribo unit and the requested DNA-sequence.

- 5′-truncated sequences are removed by Rp HPLC using the DMTr group of the last coupling step (trityl-on synthesis) as a hydrophobic tag.

Very labile protecting groups will be used to avoid problems with deprotection.