(W1130-04-25) Software-Optimized Oligonucleotide Purification with Preparative HPLC/MS

Purpose: In recent years, synthetic oligonucleotides (ONs) such as aptamers, guide and small interfering RNA, or antisense ONs have moved into the focus of life science and diagnostics research. These molecules typically exhibit chain lengths of fewer than 100 nucleotides and thus can be analyzed using ion pair-reversed phase (IP-RP) HPLC. For higher sensitivity in mass spectrometry (MS) based analyses, triethylamine (TEA) and hexafluoroisopropanol (HFIP) are considered the standard ion-pair reagents. This work shows the purification of two ONs on a preparative scale using dibutyl amine (DBA) and tris(hydroxymethyl)aminomethane (TRIS) for ion-pairing and pH adjustment. DBA has already been shown to yield higher resolution than TEA with short-chained ONs [1], while TRIS serves to substitute expensive HFIP. Although TRIS is not MS-compatible, the target ONs were detected and collected by the MS signal. This was achieved by active splitting of the full flow and delivering the split flow to the MS with a compatible make-up solvent. To optimize the separation, focused gradients were automatically calculated by Automated Purification Software, based on the target peak retention time in the analytical scouting runs. Using these software algorithms, the separation was shortened from 30 to 13 min. Fractions of the collected ONs reached a purity of 99%.

Methods: Both samples were analyzed using an HPLC/MS separation method (HA/HFIP method, to determine the purity and the target masses for purification. To enable an optimized scale-up by the Automated Purification Software, the analytical and preparative methods need to use the same mobile phase. Therefore, another method was applied (DBA/ TRIS method), which does not require the costly HFIP reagent at the high flow rate used for purification. This second method is not compatible with mass spectrometry, as the TRIS buffer is not volatile. Nevertheless, mass-based fraction collection was used as only a minor part of the flow is transported to the MSD using an active splitter and volatile make-up solvent.

Results: Both ON samples were separated using the analytical path of the 1290 Infinity II Autoscale Preparative LC/MSD. The optimized gradient from 50 to 71 %B separated the full-length product (FLP) from the incomplete ON fragments. Figure 1 displays UV and MSD signals of the long ON separation. The mass spectra of the FLP were extracted and deconvoluted by the software to provide the molecular mass of the ONs and trigger ions for fraction collection in the preparative runs. To purify the two ONs, the method was changed to DBA/TRIS. Automated Purification Software created a linear gradient for analytical scouting and–based on target peak selection in this analytical run–calculated a focused gradient for purification. With this optimized, shortened gradient, fractions were collected selectively using target masses. The purification of the long ON is shown in figure 2. Purification of the short ON produced similar results. Collected fractions were re-analyzed and showed purities of typically > 99%, see figure 3.

Conclusion: Two ON samples of 30 to 50 nucleotides were successfully purified using preparative HPLC with fraction collection triggered by MSD signals. Both samples were analyzed using an MS-compatible method (HA/HFIP), which yielded the mass spectra required to trigger fraction collection selectively. For preparative purification, another method using DBA/TRIS was used, avoiding costly HFIP reagent. Although TRIS is not MS-compatible, a suitable make-up solvent in the split flow to the MSD enabled mass-based fraction collection with high selectivity and confidence. Automated Purification Software turned a generic analytical gradient into an optimized focused gradient to shorten the preparative runs and get the pure compounds faster. Fraction slices of both ONs were successfully collected, allowing re-analysis and selective pooling of fractions that met purity requirements. All runs were conducted using AdvanceBio Oligonucleotide columns, which provide fast and easy scale-up from analytical to preparative applications

References: (1) Catani, M. et al. Oligonucleotides: Current Trends and Innovative Applications in the Synthesis, Characterization, and Purification. Biotechnology Journal 2020, 15, 8.
(2) Purification of Single-Stranded RNA Oligonucleotides Using High-Performance Liquid Chromatography. Agilent Technologies application note, publication number 5994-3514EN, 2021.
(3) Evaluation of Different Ion-Pairing Reagents for LC/UV and LC/MS Analysis of Oligonucleotides. Agilent Technologies application note, publication number 5994-2957EN, 2021.

Acknowledgements: This work was supported by the Swedish Knowledge Foundation through the KKS SYNERGY project "Improved Methods for Process and Quality Controls using Digital Tools" (grant number 20210021). The authors gratefully acknowledge Olle Stålberg, PhD, of Qiagen DNA Synthesis for supplying the oligonucleotide samples

Figure 1. Separation of the long ON, HA/HFIP method. A: UV chromatogram, B: MSD full scan, C: deconvoluted spectrum of FLP, D: multiply charged ions of FLP. Trigger ions in bold font.

Figure 2. Mass-based purification run of the long ON

Figure 3. Fraction reanalysis of the short (A) and long ON (B). Fractions are numbered and sorted by increasing retention.