Oligonucleotides should be synthesized using standard protocols appropriate to the phosphoramidites and automated DNA/RNA synthesizer, leaving the product oligonucleotide uncleaved from the solid support in the original plastic synthesis column, with the 5ʹ-terminal dimethoxytrityl group removed to yield the free 5ʹ-hydroxyl (section 1). All oligonucleotides used in this demonstration were prepared using 1,000 Å controlled pore glass (CPG) resin as the solid support and conducted at the 0.2 or 1 µmole scale. Representative examples of synthesizer columns, resins, reagents, and phosphoramidites are provided in the Table of Materials. For larger-scale reactions, volumes and times used in subsequent steps may need to be adjusted.
The triphosphorylation reaction is conducted on-column in a custom-built reaction chamber (Figure 2, section 3) using standard, commercially available components listed in the Table of Materials and follows the scheme illustrated in Figure 1 (section 4)28. It is essential that conditions be kept strictly anhydrous during triphosphorylation, and that all solvents and reagents be prepared over molecular sieves in advance and allowed to fully dry before use (section 2). Triphosphorylation typically takes 2 h to occur, and afterward, the washed and dried column can be treated according to standard oligonucleotide deprotection and purification procedures (section 5).
After deprotection, oligonucleotide triphosphates are purified by denaturing polyacrylamide gel electrophoresis (PAGE), showing a single major product band by UV back-shadowing that can be excised and eluted from the gel. The 5′-triphosphate product is readily separated from reaction side products for short oligonucleotides, as shown for DNA trinucleotide 5ʹ-triphosphates, pppAAA and pppCCC, and L-RNA trinucleotide 5ʹ-triphosphate pppGAA in Figure 3A,B. Both the 5′-hydroxyl and 5′-triphosphate products for AAA and CCC DNA trimers were excised and identified by mass spectrometry and correspondingly labeled in Figure 3A. Additional bands, as visible for the AAA DNA trimer, generally do not contain enough material to recover and identify. The presence of these bands, however, correlates with additional product masses in the unpurified reaction products (Figure 3C), typically representing 5′-diphosphate, monophosphate, and H-phosphonate side products, as discussed below.
After PAGE purification, larger oligonucleotides can be eluted using the crush and soak method42 and subsequent ethanol precipitation. However, oligonucleotides less than 15 nt cannot be ethanol precipitated efficiently and, thus, require a modified procedure for gel elution (step 5.11.3). The disposable size exclusion column listed in the Table of Materials is rated only for use with oligonucleotides longer than 10 nt. However, we have found that oligonucleotides as short as trimers can be effectively desalted using the manufacturer's recommended protocol. Nevertheless, it is recommended when desalting short oligonucleotides (as in steps 5.6 and 5.11.3) that the column eluate be collected in fractions, and product fractions be identified by absorbance at 260 nm using a UV-Vis spectrophotometer. A size exclusion column optimized for shorter oligonucleotides is provided in the Table of Materials as an alternative choice. The final yield from 1 µmole scale oligonucleotide synthesis after purification is 50-300 nmol.
Triphosphorylation can be confirmed by mass spectrometry, where the triphosphorylated product has a mass +239.94 Da greater than the 5′-hydroxyl oligonucleotide, although the presence of materials corresponding to the 5′-di- and monophosphate (+159.96 and +79.98 Da, respectively) are often observed. A 5′-H-phosphonate side product with a mass +63.98 Da from the 5′-OH mass may also be observed, and high levels of this product indicate conditions during triphosphorylation were not sufficiently anhydrous. Prior to purification, deprotected oligonucleotides will typically show all these products (Figure 3C), while purified material will show a peak corresponding to the 5′-triphosphate product along with 5′-di- and monophosphates (Figure 3D,E).
Mass spectrometry alone will typically not give a rigorous measure of 5′-triphosphate purity due to differential rates of ionization and fragmentation of the triphosphate during ionization. To measure final product purity, reverse-phase liquid chromatography and tandem ESI-MS (RP-LC/ESI-MS) are recommended, particularly for longer oligonucleotides. Analysis of D-RNA 5ʹ-triphosphates pppACGAGG and pppGAGACCGCAACUUA by RP-LC/ESI-MS (Figure 4A,B, respectively) show typical final product purity, containing 20% 5ʹ-diphosphate as these two species are difficult to separate when present on longer oligonucleotides.
Synthetic 5′-triphosphate oligonucleotides typically function as well or better than materials prepared enzymatically in biochemical studies. In section 6, as an example, 5′-triphosphate 14 nt RNA substrates prepared either synthetically or by in vitro transcription were compared in an RNA-catalyzed self-replication reaction14,15,43,44,45. Ribozyme E catalyzes the joining of substrates A and B to yield a new copy of E in an autocatalytic reaction capable of exponential growth (Figure 5A). E and 32P-labeled A components were prepared by in vitro transcription, and triphosphorylated substrate B was prepared either synthetically, as described above, or by in vitro transcription14. Self-replication reaction progress was monitored by taking periodic samples that were analyzed by denaturing PAGE and quantified via a fluorescent/phosphorescent gel scanner. The resulting data, fit to a logistic growth function, revealed that either transcribed or synthetic B substrate supports exponential growth, but synthetic B gives a slightly greater amount of product (Figure 5B). This result may reflect compositional heterogeneity at the 5′-end of RNA prepared by in vitro transcription23,24.
Chemical triphosphorylation also enables the synthesis of oligonucleotide triphosphates that cannot be prepared biologically, either in vitro or in cells. In section 7, nonbiological oligonucleotide triphosphates composed of L-RNA, the enantiomer of natural D-RNA, prepared as in sections 1-5, were used as substrates for the D-RNA "cross-chiral" polymerase ribozyme 27.3t (Figure 6A), which catalyzes the template-directed polymerization of a longer L-RNA product from short L-RNA oligonucleotide 5′-triphosphates in a sequence-general manner. As an example, the ribozyme can synthesize an L-RNA version of the hammerhead self-cleavage motif (Figure 6B)18. Purified L-RNA trinucleotide triphosphates were combined with a fluorescein-labeled L-RNA primer and L-RNA template (Figure 6C) and reacted with the cross-chiral ligase. Samples over the course of the reaction were analyzed by PAGE and imaged using a fluorescent/phosphorescent gel scanner to demonstrate synthesis of an L-RNA version of the hammerhead ribozyme encoded by the template (Figure 6D).

Figure 3: Purification of trinucleotide 5ʹ-triphosphates. (A) PAGE analysis (visualized by UV-back-shadowing) of triphosphorylation of DNA trinucleotides tri-deoxyadenosine (AAA, blue) and tri-deoxycytidine (CCC, red), intentionally overloaded to visualize minor side-products. Both the 5ʹ-triphosphate product (ppp) and 5ʹ-hydroxyl (OH) starting material were excised and identified by MALDI-MS. (B) Preparative PAGE of triphosphorylation of L-RNA trinucleotide GAA, with major product band excised and identified as the 5ʹ-triphosphate (ppp) by ESI-MS. (C) MALDI-MS of crude reaction products after deprotection and (D) purified products from (A). 5ʹ-triphosphate (ppp; pppAAA expected 1,119 Da, observed 1,118 Da; pppCCC expected 1,047 Da, observed 1,046); 5ʹ-diphosphate (pp), 5ʹ-monophosphate (p), 5ʹ-hydroxyl (OH), and 5ʹ-H-phosphonate (Hp) are labeled. (E) Deconvoluted mass spectrum from direct injection ESI-MS of isolated 5ʹ-triphosphate product from (B), with identified peaks labeled (expected 1,181.6 Da, observed 1,181.0 Da). 5ʹ-diphosphate (pp) products are also observed, as are sodium ion peaks for both the tri- and di-phosphate products (+22 Da). Common contaminant peaks are labeled with an asterisk. For ease of comparison, mass spectra were normalized to the highest intensity measured in each spectrum and are reported as a percentage relative to that value. Abbreviations: PAGE = polyacrylamide gel electrophoresis; MALDI-MS = matrix-assisted laser desorption/ionization; ESI-MS = electrospray ionization mass spectrometry. Please click here to view a larger version of this figure.

Figure 4: Analytical RP-LC of 6 nt and 14 nt D-RNA oligonucleotide triphosphates. (A) 5ʹ-pppACGAGG-3ʹ and (B) 5ʹ-pppGAGACCGCAACUUA-3ʹ. Tandem ESI-MS identified the major peak of both (~70%) as the 5ʹ-triphosphate (ppp), with lesser amounts of the 5ʹ-diphosphate (pp). Abbreviations: RP-LC = reverse-phase liquid chromatography; nt = nucleotides; ESI-MS = electrospray ionization mass spectrometry. Please click here to view a larger version of this figure.

Figure 5: Comparison of oligonucleotide 5ʹ-triphosphate substrates prepared by chemical synthesis or in vitro transcription. (A) The self-replicating ribozyme E ligates RNA A and 5′-triphosphorylated RNA B. (B) Comparison of self-replication reactions using 10 µM A and 10 µM B, either synthetic (open circles) or in vitro transcribed (filled circles). (B) Data were fit to the logistic growth equation: [E] = a / (1 + be-ct), where a is the final yield, b is the degree of sigmoidicity, and c is the exponential growth rate. Growth rates for the two reactions were identical, at 1.14 h-1, while the final extent was 10% higher for reactions with synthetic B. Please click here to view a larger version of this figure.

Figure 6: Cross-chiral L-RNA polymerization with a ribozyme. (A) The D-RNA 27.3t polymerase ribozyme, which catalyzes template-dependent ligation of L-RNA. (B) The L-RNA product synthesized by 27.3t forms part of a hammerhead endonuclease motif. (C) L-RNA polymerization catalyzed by 27.3t using a biotinylated L-RNA template (brown), an end-labeled L-RNA primer (magenta), and four L-RNA trinucleotide triphosphates (cyan), prepared synthetically. (D) PAGE analysis of extension products of (B) at 4 h and 24 h, showing each trinucleotide incorporation up to full-length product (black dot). Unreacted L-RNA primer is included as a reference marker. Abbreviations: PAGE = polyacrylamide gel electrophoresis; M = reference marker. Please click here to view a larger version of this figure.