Marc M. Lemaître, Andrew S. Murphy and Robert L. Somers
Glen Research Corporation, 22825 Davis Drive, Sterling, VA 20164, USA
Preparation of oligoribonucleotides with a phosphorothioate backbone requires a sulfurizing reagent that is efficient and stable in solution. One of the most efficient sulfurizing reagents, 3H-1,2-benzodithiol-3-one-1,1-dioxide (Beaucage Reagent), is stable in solution but lacks long-term stability on the DNA synthesizer. This report evaluates Beaucage Reagent and 3-((N,N-dimethyl-aminomethylidene)amino)-3H-1,2,4-dithiazole-5-thione (DDTT) (Sulfurizing Reagent II) as sulfurizing reagents when using TOM-protected and TBDMS-protected RNA phosphoramidites, and standard DNA phosphoramidites for oligonucleotide synthesis. The product oligonucleotides were evaluated by ion exchange HPLC, ESI-Mass, and 31P NMR. It was confirmed that Ion Exchange HPLC can resolve oligonucleotides containing a single phosphodiester linkage from those containing all phosphorothioate linkages.
The replacement of one non-bridging oxygen atom in the phosphodiester linkage of DNA or RNA by sulfur creates a phosphorothioate (PS) linkage. This is one of the oldest and most studied backbone modifications of nucleic acids (Figure 1). Eckstein synthesized the first dinucleoside containing a phosphorothioate linkage1. Interest in phosphorothioate oligonucleotides (PSO) as biological tools began with the discovery by De Clercq et al. that PS RNA was more resistant to RNases that natural RNA2.
Elemental sulfur was one of the first sulfurization reagents used by Burgers and Eckstein to sulfurize phosphite triesters3. Since then, a number of sulfurizing reagents have been used for chemically modifying the phosphate backbone of oligonucleotides. Some are shown in Figure 2, Page 2, Dibenzyl tetrasulfide4 (2a), Beaucage Reagent5 (2b), 3-Ethoxy-1,2,4-dithiazolidin-5-one (EDITH)6 (2c), 1,2,4-Dithiazolidine-3,5-dione (DtsNH)6 (2d), 3-Amino-1,2,4-dithiazole-5-thione7 (2e).
After this initial work, thousands of papers describe activities and properties of such modified oligonucleotides in cell culture as well as in vivo. Indeed, PSO is considered to be the first-generation antisense oligonucleotide (ASO)8-10.
Though not perfect, the phosphorothioate modification offers several advantages:
And the use of chimeric oligos containing several different modifications along with a normal phosphodiester segment reduces side effects in total PS modification11,12.
In addition to their use as DNA analogues in ASO, PS modifications have some advantages for RNA as well13. Recently, Overhoff and Sczakiel14 showed that full PSO could promote the delivery of siRNAs in cell culture. This siRNA uptake is sequence-independent and the optimal length seems to vary between 30 and 70 nucleotides, depending on the cell line. Even though this method is not yet as efficient as the cationic lipids, it opens the way to possible new methods.
Another recent paper15 describes a method for the inactivation of miRNA that may help to elucidate their functions. It uses oligonucleotides called antagomirs (23-mers, complementary to a target miRNA) consisting of 2´-OMe bases, a cholesteryl group at the 3´ terminus, phosphorothioates at positions 1 and 2 at the 5´ end, and at the last four at the 3´ end. These molecules promote the cleavage of complementary miRNAs and thus should allow analysis of their function. Presumably the role of the PS linkages is the stabilization against degradation in the mouse experiments, which is consistent with its function in the antisense field in such in vivo situations.
Finally, two recent papers16,17 show that modifications including PS do not systematically abolish siRNA activity, opening the road for some potentially less expensive stabilization of such molecules. Incorporation of 2’-OMe or 2’-F (at various positions in the sense as well as in the antisense strand) in combination with PS linkages should confer increased resistance to degradation by nucleases, as well as prolonged serum retention. And it is also possible that such an easy modification of siRNA may increase specificity by reducing or eliminating sense strand recruitment in the RISC complex and thus reducing a source of off-target effect.
The most common method for making PSO today is to use 3H-1,2-Benzodithiol-3-one-1,1-dioxide5,18 (Beaucage Reagent) (2b) as the sulfurizing reagent during oligonucleotide synthesis in place of the iodine-based oxidizer. Despite its ease of use and long history, this product has some disadvantages such as its limited stability in solution while on a synthesizer. Also, it may not display optimal kinetics for the sulfurization of RNA in solid phase phosphoramidite synthesis. This work compares Beaucage Reagent with an alternative sulfurizing reagent, 3-((N,N-dimethylaminomethylidene)amino)-3H-1,2,4-dithiazole-5-thione (DDTT) (Sulfurizing Reagent II) (Figure 3) to evaluate kinetics for sulfurizing RNA, performance for sulfurizing DNA, and stability in storage and while on the synthesizer.
Synthesis columns. phosphoramidites, and synthesis reagents were from Glen Research, Sterling, VA. TRIS was from JT Baker. HPLC grade acetonitrile was from B&J. Ammonium hydroxide, butanol, DMSO, sodium perchlorate, and triethylammonium trihydrofluoride were from Sigma-Aldrich. Alcoholic methylamine and aqueous methylamine were from Fluka. Reagent alcohol and sterile water were from Fisher Scientific. HPLC water and sodium acetate were from EMD.
RNA oligonucleotides and PSO RNA were prepared using an ABI 394 DNA synthesizer using standard protocols. The sulfurizing reagent was loaded onto port 10 of the synthesizer. Sulfurization times and coupling times were modified as indicated. RNA synthesis was carried out on a T-CPG or a T-Q-support. After synthesis, columns were dried under a stream of argon and the contents transferred to a 4mL glass vial. The oligos were cleaved and deprotected using 1mL of 40% aqueous methylamine: 33% ethanolic methylamine (1:1) at room temperature overnight. The oligos were then filtered into a 2mL centrifuge tube and dried under vacuum in a Speedvac. The oligos were dissolved in 100µL dry DMSO and 125µL triethylammonium trihydrofluoride was added to the vial. The tube was heated to 65°C for at least one hour and then cooled briefly in a freezer. 25µL of 3M sodium acetate was added and mixed thoroughly before 1mL of 1-butanol was added and mixed. The tube was cooled on dry ice for 30 minutes and centrifuged at 13,000rpm for 10 minutes. The pellet was gently washed twice with reagent alcohol, dried under vacuum, and dissolved in 1mL of sterile water. The concentration of each oligo was calculated using Beer's law and the extinction coefficient at 260 nm.
DNA was prepared using an ABI 394 DNA synthesizer using standard protocols and Ac-dC phosphoramidite. Oligonucleotides that required base deprotection were deprotected and cleaved as described19.
The oligos were analyzed on Ion Exchange (IEX) HPLC using a concentration range of 5-20 A260/mL. IEX was performed on a Beckman System Gold HPLC using a Dionex DNAPac PA200 polymeric ion exchange column (4.6x250mm). Buffer A - 25mM TRIS, 10mM Sodium Perchlorate, pH 7.4, 20% acetonitrile, Buffer B - 25mM TRIS, 600mM Sodium Perchlorate, pH 7.4, 20% acetonitrile. Simple RNA sequences were analyzed at room temperature and longer sequences were analyzed at 60°C using a column heater. All analyses were run with a gradient elution; 0-60% B in 30 minutes at 1mL/min, monitored at 260nm.
Samples for ESI-Mass were converted from the sodium salt to the ammonium salt as follows. Aliquots of each oligo were evaporated to dryness in a Speedvac. Samples were dissolved in 100µL of water. 25µL of 3M ammonium acetate was added to each vial. Samples were vortexed before adding 1 mL of butanol and the samples were then vortexed again. The samples were chilled on dry ice for at least 30 minutes. The samples were centrifuged for 10 minutes, washed with ethanol and dried under vacuum.
Seven sequences were used to evaluate the performance of the sulfurizing reagents. TOM protected and TBDMS protected RNA phosphoramidites were evaluated in this study since we offer both types of phosphoramidites.
Initial studies used the sequence, 5’-UUU UUU UUT T-3’, to compare the performance of 0.05M Beaucage Reagent, 0.03M DDTT, and 0.05M DDTT. This sequence offered the advantage in that there was no need for the overnight base deprotection step. These sulfurizing reagents were evaluated using increasing times of 30, 60, 120, and 240 secs for sulfurization with TOM RNA phosphoramidites. Results are shown in Figure 4. The activator solution was 0.25M 5-(ethylthio)-1H-tetrazole (ETT) with a coupling time of 4 minutes. The coupling efficiency was calculated based on the purity by ion exchange and length of oligonucleotide.
Setting the minimum coupling efficiency at ≥98%, then a sulfurizing time between one minute and four minutes was suitable for similar sequences using 0.05M Sulfurizing Reagent II. A four minute sulfurizing time was also sufficient while using the 0.03M Sulfurizing Reagent II for this oligo. When using the 0.05M Beaucage Reagent, only a four minute sulfurization time was suitable under the same conditions. Figure 5 shows the IEX chromatograms of oligos sulfurized with 0.05M DDTT versus Beaucage Reagent with sulfurization time of 60 seconds for each.
The performance of the 0.05M Sulfurizing Reagent II and Beaucage Reagent was also evaluated on TBDMS-RNA phosphoramidites. Results are in Figure 6 (Page 4). The 30 second sulfurizing time was omitted for TBDMS-RNA phosphoramidites based on the decreasing coupling efficiency results for 60 and 120 seconds. Coupling time was also 4 minutes using ETT as the activator. Using the TBDMS phosphoramidites, the 0.05M Sulfurizing Reagent II performed well for all time periods from 1 to 4 minutes. 0.05M Beaucage Reagent performed well only when using a sulfurizing time of 4 minutes.
Additional sequences were prepared that represented the n-1 oligonucleotide and a mixed oxidation sequence with a single internal PO linkage. A sulfurization time of 60 seconds was used for these oligonucleotides. These sequences were compared to the full-length oligo on ion exchange HPLC. See Table 1, Page 4 for the n-1 and mixed oxidation sequences.
Figure 7, Page 4 shows the ion exchange HPLC of a) a mixture of the full-length oligo and the mixed oxidation oligo, b) a mixture of the full-length oligo and the n-1 oligo, c) a mixture of the n-1 oligo and the mixed oxidation oligo. Under these conditions, ion-exchange HPLC can separate phosphorothioates that contain one or more PO linkages. Additionally, ion-exchange HPLC can also separate failure sequences, such as n-1, from the full-length PSO.
Samples were selected for analysis by ESI-Mass and 31P NMR. See Tables 2 and 3, respectively, on Page 5. The mass was confirmed for each sequence indicating that no modifications of the bases occurred during synthesis or the subsequent work up of the sample. The 31P NMR can distinguish between the PS linkage and PO linkage as seen in Figure 8, Page 5. None of the sulfurizing reagents showed any extraneous oxidation of the phosphite linkages. The 31P NMR of the mixed oxidation sequence shows the presence of a phosphodiester linkage and the phosphorothioate.