Solvent Selection in the Detritylation Reaction for the Solid-Phase Synthesis of Oligonucleotides

Abstract

This study examined the choice of solvent in the detritylation reaction for solid-phase oligonucleotide synthesis. In addition to being chemically inert under detritylation conditions, such a solvent should not bind with acid used in this reaction. These considerations explained the only two choices of solvent currently used in the detritylation reaction: dichloromethane and toluene. Other solvents, such as ethyl acetate, acetonitrile, THF, and nitromethane are shown by NMR spectroscopy to bind with trichloroacetic acid. Therefore, these are undesirable solvents for the detritylation reaction, confirmed by solid-phase synthesis experiments.

Since its inception in the early 1980s, solid-phase synthesis (SPS) of oligonucleotides[1] by the phosphoramidite chemistry[2] has allowed for rapid access to natural and modified oligonucleotides on milligram to kilogram scales. Phosphoramidites are the key building blocks in SPS of oligonucleotides usually protected with a 4,4′-dimethoxytitryl (DMTr) group at the 5′-OH. First demonstrated by Khorana and co-workers,[3] the DMTr group is ideal in oligonucleotide synthesis due to its ease of removal, while stable enough for isolation, purification, and storage. In SPS of oligonucleotides, effective release of the DMTr group during the detritylation reaction is crucial to the success of syntheses. This treatment typically involves the use of a solution of acid, usually di- or trichloroacetic acid in dichloromethane or toluene and can lead to effective detritylation in as short as 20 s on a 1 μmol scale.[4] While dichloromethane is an effective solvent in this reaction on small scales, it is not recommended for large-scale, industrial production of oligonucleotides, due to its hazardous nature. Toluene, on the other hand, often leads to swelling of the polymer matrix, such as polystyrene,[5] [6] leading to increased back pressure in flowthrough columns. As such, efforts have been made to explore potential alternative solvents that are not only effective for the detritylation reaction but also industry-friendly in terms of their environmental and health and safety profiles, and suitability for large-scale manufacturing.[7]

In our effort to explore alternative solvents for this purpose, it became obvious to us that such an alternative is, in fact, not straightforward. A few factors must be taken into consideration for the choice of suitable solvents in the detritylation reaction during SPS of oligonucleotides. We wish to summarize our findings in this communication.

For the detritylation reaction, ideal solvent should be relatively easy to dry and keep anhydrous, chemically stable under the detritylation conditions, and does not bind to acid. Therefore, dichloromethane, toluene, acetonitrile, ethyl acetate, THF, and nitromethane were evaluated. Protic solvents such as alcohol were not considered due to their tendency to react with trityl cations.

The initial evaluation of solvents used in the detritylation reaction focused on the potential reaction of solvents with trityl cations, considering the Lewis basicity of solvents. As DMTr cations are more difficult to work with, tritylium hexafluorophosphate (trityl PF₆) was used as a model system due to its commercial access and relative ease of handling. Using enthalpies of complex formation with boron trifluoride in dichloromethane, the Lewis basicity of nonprotogenic solvents were determined by Maria and Gal (Table [1]).[8]

This experiment turned out to be rather demanding as the solvents must be dried thoroughly and solutions handled in an inert atmosphere. Trace amount of water present can react with trityl cation, which will complicate the interpretation of data.

Trityl PF₆ dissolves well in dichloromethane, acetonitrile, and THF, but has limited solubility in toluene and ethyl acetate. When dissolved, trityl PF₆ gives a yellow to light orange color in dichloromethane, acetonitrile, ethyl acetate, as well as in a mixture of dichloromethane–toluene (1:1, v/v), but not in THF where the yellow color disappeared upon standing for a few minutes (Table [1]). Interestingly, a polymeric gel was formed in concentrated solution of trityl PF₆ in THF. In fact, the polymerization of THF initiated by tritylium salts is well-known in the literature.[9] [10] This observation may rule out THF as a solvent for the detritylation reaction. The other solvents did not appear to react with trityl PF₆ at appreciable rates.

Table 1 Properties of Solvent Compared in this Study: Enthalpies of Complex Formation with BF₃ in Dichloromethane, Water Content, and Physical Appearance of Trityl PF₆ in these Solvents

	DCM	Toluene	MeCN	THF	EtOAc
–ΔH° (BF3, 298.15 K) (kJ·mol–1)a	10.0 ± 3.0	n/d	60.39 ± 0.46	90.40 ± 0.28	75.55 ± 0.31
Water content (ppm)b	2.3 ± 0.1	5.0 ± 2.0c	8.0 ± 2.0	6.5 ± 0.5	13.6 ± 0.8
Appearance of trityl PF6 solutionsd

^a Enthalpies of complex formation with BF₃ in dichloromethane.

^b Water contents were determined by Karl Fischer titration.

^c Water content in toluene.

^d All five samples were prepared by adding 40 mg of trityl PF₆ to dried solvents (50 mL) in a nitrogen atmosphere. n/a: not determined in ref. 8.

Another consideration in choosing solvents for the detritylation reaction is their potential binding to acid, which will hinder the detritylation reaction. While acid binding with acetonitrile has been studied for the detritylation reaction, detailed NMR studies have not been conducted.[11] Therefore, a more comprehensive comparison of the solvent–acid binding by NMR analysis was undertaken to shed more light on the type of interaction.

In these experiments, all solvents were carefully dried, with water content (Table [1] and Table [2]) determined by Karl Fischer titration. ¹H and ¹³C NMR spectra of solvent, ¹⁵N NMR spectra in the case of acetonitrile, and ¹H and ¹³C NMR spectra of trichloroacetic acid (TCA) were recorded in CD₂Cl₂. Upon mixing trichloroacetic acid with the solvent, the corresponding NMR spectra were recorded again. It was found that solvents such as acetonitrile, ethyl acetate, acetone (data not shown), THF, and nitromethane, but not dichloromethane and toluene, showed significant changes in chemical shifts when these solvents are present together with trichloroacetic acid. For comparison, the NMR observations of acetonitrile and dichloromethane are described below, while observations for the other solvents can be found in the Supporting Information.

As shown in Figure [1a–c], the carboxyl proton of TCA shifted from 11.32 (TCA in CD₂Cl₂) to 10.63 ppm (a mixture of TCA and acetonitrile in CD₂Cl₂), while the CH₃ proton in acetonitrile remained almost unchanged between acetonitrile in CD₂Cl₂ and a mixture of acetonitrile and TCA in CD₂Cl₂. Similarly, a rather large change in the ¹³C NMR chemical shift of carbonyl was seen for TCA (166.8 to 163.0 ppm, Figure [1e,f]) between that of TCA in CD₂Cl₂ or a mixture of TCA and acetonitrile in CD₂Cl₂. A significant change in the ¹⁵N shift was also seen for acetonitrile alone in CD₂Cl₂ (245.2 ppm) to 238.0 ppm when acetonitrile and TCA are both present in CD₂Cl₂. The changes outlined here suggest an interaction between TCA and acetonitrile (Figure [1h]), leading to changes in chemical shifts of nitrogen and carbonyl carbon in acetonitrile and TCA, respectively. Binding of acetonitrile with a few acids was also reported in the literature.[12] [13] [14] On the other hand, virtually no difference in chemical shift was seen for dichloromethane, with or without trichloroacetic acid, in both ¹H and ¹³C NMR spectra recorded in C₆D₆ (Figure [1a]′–f′).

Taken together, the NMR study showed significant changes in the chemical shift of the carboxy proton and carbonyl carbon of trichloroacetic acid when both acid and solvents such as acetonitrile, ethyl acetate, acetone, THF, and nitromethane are present, indicating acid binding by these solvents (Table [2]). On the other hand, very little change was seen for these chemical shifts of trichloroacetic acid in the presence of dichloromethane (or toluene). This data are of significant value in choosing the appropriate solvents for the detritylation reaction, as acid binding by solvent will decrease the detritylation rate, leading to oligonucleotide synthesis failure. The proposed structures of trichloroacetic acid binding with solvents are shown in Figure [2].

To demonstrate the influence of solvents used in the detritylation reactions on the overall success of solid-phase oligonucleotide synthesis, T₁₈ was assembled on a 1 μmol scale using the standard cycle conditions with the corresponding solvent pre-washed before introducing detritylation solution (Table [3]) on an ABI 3400 synthesizer. THF and nitromethane were not included in this comparison, as THF was shown to react with trityl cations, and nitromethane is associated with explosive hazard.[15]

After syntheses were complete, the products were cleaved off the solid support, deprotected under standard ammonolysis conditions, and analyzed by anion exchange HPLC off a DNAPac 200 column. As shown in Figures [3a,b], T₁₈ was successfully synthesized with very little truncated sequences when toluene or dichloromethane was used as solvent for the detritylation reaction. The synthesis, however, failed when acetonitrile or ethyl acetate was used (Figure [3c,d]).

It should also be noted that ABI 3400 utilizes a conductivity detector to monitor trityl cations that was intended for early detection and diagnosis of instrument-related problems. The average stepwise yield (ASWY) in trityl assay report was found to be highly unreliable when some solvents were used in the detritylation reaction. For example, while very little coupling took place when ethyl acetate was used as solvent (as judged by HPLC and the color of detritylation flowthrough), the reported ASWY was virtually quantitative. On the other hand, while the overall synthesis was successful when toluene was used in the detritylation reaction, no trityl cations were detected (Table [4]). In addition, the relative success in the detritylation reaction is also reflected in the intensity of trityl cations seen in the flowthrough (Table [4]).

This work demonstrated the important considerations in the choice of solvents for the detritylation reaction in solid-phase synthesis of oligonucleotides by the phosphoramidite chemistry. Taken together, solvents suitable for this reaction should be chemically unreactive under the detritylation conditions, should not bind to the acid, and should be relatively easy to dry and keep dry. With these considerations in mind, it became obvious that, among common solvents, only dichloromethane and toluene are suitable for this reaction. Other solvents, such as ethyl acetate, acetonitrile, nitromethane, acetone, and THF bind to acid, leading to partial or virtually no detritylation. In addition, THF reacts with trityl cations and polymerizes over time.

While there is very limited room for the improvement of detritylation conditions in terms of choice of solvent, more detailed expanded investigations of appropriate trityl cation scavengers[16] and alternative acids are warranted to identify conditions that not only allow for efficient detritylation but are also industry-friendly when syntheses are carried out on large scales. These aspects of consideration are currently underway in this lab.

Furthermore, in the phosphoramidite chemistry based SPS of oligonucleotides, the detritylation step is preceded by oxidation, which requires thorough washing with acetonitrile to remove the excess of oxidation reagents. Given the observation that acetonitrile binds to acid compromising the detritylation efficiency, thorough washing by dichloromethane or toluene is recommended to displace acetonitrile prior to the delivery of acid solution for detritylation step during large-scale production.

Conflict of Interest

The authors declare no conflict of interest.

• References and Notes

• 1 Matteucci MD, Caruthers MH. J. Am. Chem. Soc. 1981; 103: 3185
  CrossrefSearch in Google Scholar
• 2 Beaucage SL, Caruthers MH. Tetrahedron Lett. 1981; 22: 859
  CrossrefSearch in Google Scholar
• 3 Schaller H, Weimann G, Lerch B, Khorana HG. J. Am. Chem. Soc. 1963; 85: 3821
  CrossrefSearch in Google Scholar
• 4 Tram K, Sanghvi YS, Yan H. Nucleosides, Nucleotides Nucleic Acids 2011; 30: 12
  CrossrefPubMedSearch in Google Scholar
• 5 Moore MN, Andrade MA, Scozzari AN, Krotz AH. Org. Process Res. Dev. 2004; 8: 271
  CrossrefSearch in Google Scholar
• 6 Kim SW, Cho YI, Jung KE. Bull. Korean Chem. Soc. 2021; 42: 1296
  CrossrefSearch in Google Scholar
• 7 Andrews BI, Antia FD, Brueggemeier SB, Diorazio LJ, Koenig SG, Kopach ME, Lee H, Olbrich M, Watson AL. J. Org. Chem. 2021; 86: 49
  CrossrefPubMedSearch in Google Scholar
• 8 Maria P.-C, Gal J.-F. J. Phys. Chem. 1985; 89: 1296
  CrossrefSearch in Google Scholar
• 9 Dreyfuss MP, Westfahl JC, Dreyfuss P. Macromolecules 1968; 1: 437
  CrossrefSearch in Google Scholar
• 10 Smirnov YN, Volkov VP, Oleinik EF, Komarov BA, Rozenberg BA, Yenikolopyan NS. Polymer Sci. U.S.S.R. 1974; 16: 846
  CrossrefSearch in Google Scholar
• 11 Paul CH, Royappa AT. Nucleic Acids Res. 1996; 24: 3048
  CrossrefPubMedSearch in Google Scholar
• 12 Appelman EH, Dunkelberg O, Kol M. J. Fluorine Chem. 1992; 56: 199
  CrossrefSearch in Google Scholar
• 13 Salnikov GE, Genaev AM, Vasiliev VG, Shubin VG. Org. Biomol. Chem. 2012; 10: 2282
  CrossrefPubMedSearch in Google Scholar
• 14 Rozenberg M, Loewenschuss A, Nielsen CJ. J. Mol. Struct. 2020; 1199: 126948
  CrossrefSearch in Google Scholar
• 15 Makovky A, Lenji L. Chem. Rev. 1958; 58: 627
  CrossrefSearch in Google Scholar
• 16 Zhou X, Kiesman WF, Yan W, Jiang H, Antia FD, Yang J, Fillon YA, Xiao L, Shi X. J. Org. Chem. 2022; 87: 2087
  CrossrefPubMedSearch in Google Scholar

Corresponding Authors

Publication History

Received: 16 June 2023

Accepted after revision: 02 August 2023

Article published online:
11 October 2023

© 2023. Thieme. All rights reserved

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany

• References and Notes

• 1 Matteucci MD, Caruthers MH. J. Am. Chem. Soc. 1981; 103: 3185
  CrossrefSearch in Google Scholar
• 2 Beaucage SL, Caruthers MH. Tetrahedron Lett. 1981; 22: 859
  CrossrefSearch in Google Scholar
• 3 Schaller H, Weimann G, Lerch B, Khorana HG. J. Am. Chem. Soc. 1963; 85: 3821
  CrossrefSearch in Google Scholar
• 4 Tram K, Sanghvi YS, Yan H. Nucleosides, Nucleotides Nucleic Acids 2011; 30: 12
  CrossrefPubMedSearch in Google Scholar
• 5 Moore MN, Andrade MA, Scozzari AN, Krotz AH. Org. Process Res. Dev. 2004; 8: 271
  CrossrefSearch in Google Scholar
• 6 Kim SW, Cho YI, Jung KE. Bull. Korean Chem. Soc. 2021; 42: 1296
  CrossrefSearch in Google Scholar
• 7 Andrews BI, Antia FD, Brueggemeier SB, Diorazio LJ, Koenig SG, Kopach ME, Lee H, Olbrich M, Watson AL. J. Org. Chem. 2021; 86: 49
  CrossrefPubMedSearch in Google Scholar
• 8 Maria P.-C, Gal J.-F. J. Phys. Chem. 1985; 89: 1296
  CrossrefSearch in Google Scholar
• 9 Dreyfuss MP, Westfahl JC, Dreyfuss P. Macromolecules 1968; 1: 437
  CrossrefSearch in Google Scholar
• 10 Smirnov YN, Volkov VP, Oleinik EF, Komarov BA, Rozenberg BA, Yenikolopyan NS. Polymer Sci. U.S.S.R. 1974; 16: 846
  CrossrefSearch in Google Scholar
• 11 Paul CH, Royappa AT. Nucleic Acids Res. 1996; 24: 3048
  CrossrefPubMedSearch in Google Scholar
• 12 Appelman EH, Dunkelberg O, Kol M. J. Fluorine Chem. 1992; 56: 199
  CrossrefSearch in Google Scholar
• 13 Salnikov GE, Genaev AM, Vasiliev VG, Shubin VG. Org. Biomol. Chem. 2012; 10: 2282
  CrossrefPubMedSearch in Google Scholar
• 14 Rozenberg M, Loewenschuss A, Nielsen CJ. J. Mol. Struct. 2020; 1199: 126948
  CrossrefSearch in Google Scholar
• 15 Makovky A, Lenji L. Chem. Rev. 1958; 58: 627
  CrossrefSearch in Google Scholar
• 16 Zhou X, Kiesman WF, Yan W, Jiang H, Antia FD, Yang J, Fillon YA, Xiao L, Shi X. J. Org. Chem. 2022; 87: 2087
  CrossrefPubMedSearch in Google Scholar

• Supplementary Material