Oligonucleotide Purity Analysis With Side-Product Identification and Quantitation
Oligonucleotides play a crucial role in modern drug development, therefore their accurate characterization is essential. However, given the sheer complexity and diversity of the types of oligonucleotides that are in clinical development, a range of diverse analytical approaches are necessary.
This app note highlights a solution using MS technology that can provide an in-depth analysis of oligonucleotides, ensuring identity confirmation, purity determination and synthesis side product identification.
Download this app note to discover:
How to enhance oligonucleotide analysis accuracy The benefits of using UHPLC for efficient separation of oligonucleotide impurities Automatic, meaningful reports to reduce analysis turnaround time
These drug candidates include multiple classes such as antisense oligonucleotides, small/short interfering RNA, micro-RNA, immunostimulatory oligonucleotides, aptamers, and splice-switching oligonucleotides (1). Accurate analytical characterization of oligonucleotides as APIs is necessary to confirm their identity, to determine their purity, and to identify and quantify synthesis side products. Determining the molecular weight and confirming the nucleotide sequence of an oligonucleotide are fundamental criteria for establishing the molecule’s identity. Oligonucleotide synthesis is a complex process that requires more than 100 sequential chemical reactions to make a single 25-base sequence, and the key to understanding and optimizing this chemistry is the identification of side products. Furthermore, quality of each synthesized oligonucleotide must be evaluated prior to use to ensure that the correct sequence was made, and that purity meets regulatory standards. This technical note describes the use of the Bruker timsTOF Pro 2 mass spectrometer equipped with the VIP-HESI ion source for the in-depth analysis of oligonucleotides. UHPLC was used to efficiently separate impurities from the full-length product (FLP) and to provide unambiguous confirmation of FLP sequences and synthesis side-products. Their in-depth characterization was carried out by the highly automated OligoQuest™ workflow within Bruker´s BioPharma Compass® software. Keywords: Oligonucleotides, Sequence confirmation, MS/MS, Side product quantitation, Side product identification, Quality control, High resolution accurate mass spectrometryMaterials and Methods The 24-mer RNA FLP (dubbed “mod3”) with 2´ O-methylation of the ribose at each position was synthesized by Axolabs GmbH (product number X119083K2; sequence: 5OH c a c g c g u g c u u u u g c a c g g c g u g c 3OH). Seven isomers of mod3 were also analyzed (see Figure 1). The sample was diluted with Eluent A to 0.1 µg/µL prior to analysis and 0.4 µg was injected onto the UHPLC system for subsequent MS analysis. A Bruker Elute UHPLC - equipped with an Azura UVD 2.1S UV detector (KNAUER) recording the 260 nm UV chromatogram - was connected to the Bruker timsTOF Pro 2 via the VIP-HESI ion source to separate side-products, concentrate target compounds and remove salts. For a pure sample such as the 24-mer it is beneficial to use a low-speed auto MS/MS cycle with only one precursor being selected. This ensures that the charge states with the highest intensity will be selected for MS/MS. Gradient Time [min] Flow [mL/min] % A % B 0.0 0.25 99 1 1.0 0.25 99 1 3.0 0.25 96 4 16.0 0.25 90 10 16.2 0.25 5 95 16.8 0.25 5 95 17.0 0.25 99 1 23.5 0.25 99 1 Chromatography UHPLC column Waters XBridge Oligonucleotide BEH C18 , 130Å 2.5 µm, 2.1 x 50 mm, 70°C column oven temperature Eluent A (aqueous phase) in deionized water 0.24% (v/v) Triethylamine (TEA), 1.00% (v/v) Hexafluoro-2-propanol (HFIP), 1.00% (v/v) Methanol (MeOH) Eluent B (organic phase) in Acetonitrile 10% (v/v) Isopropanol (IPA) MS Parameters – autoMSMS Acquisition AutoMSMS spectra were acquired on a Bruker timsTOF Pro 2 in negative ion mode. MS acquisition parameter in the autoMSMS analysis (timsControl Version 4.1.12) are listed in tabular form below Deflection 1 delta-70 V Funnel 1 RF 350 Vpp isCID 0 Funnel 2 RF 400 Vpp Multipole RF 600 Vpp Quadrupole ion energy 4 eV Quadrupole low mass 500 m/z Collision energy 10 eV Pre-pulse storage 10 µs MS/MS Parameters – auto MS/MS Acquisition AutoMSMS spectra were acquired on a Bruker timsTOF Pro 2 in negative ion mode. MS acquisition parameters in the autoMSMS analysis (timsCONTROL Version 4.1.12) are listed in tabular form below Total cycle time 1 s MS spectra rate 2 Hz MS/MS spectra rate (fixed) 2 Hz No. precursors 1 Normalized threshold 31 counts/1000 scans Average scans 5 Scan range, isol. width, collision energy 500 -3000 m/z, isol. width 3 m/z, 15.5-97.6 eV MS/MS Parameters – targeted MS/MS Acquisition Targeted MS/MS spectra were acquired in case of the mod3-c1 side product to increase the sequence coverage for that oligo and to locate the loss of c. Here, the mod3-c1 oligo eluted at 12.17 min. MS Settings Scan mode MRM MS/MS Settings m/z 1912, width (m/z) 5, isCID 0, CE (eV) 61.9, x Acq. 1.0, Rt range 11.9-12.3 min VIP-HESI Source Parameters Nebulizer Dry gas Dry temp Sheath gas tempSheath gas flow 4 bar 8 L/min 220°C 450°C 4 L/minData processing The LC-UV-MS(/MS) data were processed in BioPharma Compass® 2023b using the OligoQuest autoMSMS tutorial workflow method based on the user defined sequences of 24-mer RNA variants, which include residue-specific modifications. Here, 2´ O-methylated nucleotides were used, abbreviated by a, c, g and u in the sequences. OligoQuest enables the automated rapid verification of molecular mass, sequence and the assessment of purity by quantifying chromatographic peaks using the UV and MS signal intensities. In addition, it can identify sequence variants and synthetic impurities based on the input of the target sequence and further workflow parameters, which allows screening for failure sequences, addition of nucleotides or nucleotide exchange variants. Impurities with incomplete MS/MS coverage were targeted in a second round of analysis with the OligoQuest targetedMSMS workflow method in which only selected m/z and Rt ranges were used for MS/MS spectra acquisition. MS precursor ion spectra were not measured in this case. Results and Discussion The Full-Length Product The Result table (Figure 1) provides a quick confirmation that the base peak in the dataset matches the expected molecular weight of the oligo with good mass accuracy and the sample purity. The Expected table shows the analysis result for the currently selected mod3 sample at greater detail listing all identified molecular species, including sequence variants defined in the method´s matching parameters. Here we observe the confirmation of the mod3 FLP, 5 side product candidates: mod3>> [a,g]16, mod3-c1, mod3+g1 and mod3>>[u,c] and mod3>>[a,c]2. The variants with an added g residue (row 8) or the exchange of u-to-c (row 14) were solely identified based on molecular weight without significant support from fragment ion data (MSMS Score = 0). Therefore, the location of these variants in the sequence could not be determined. In the following study the focus is on the mod3 sample and sequence. The UV-chromatogram (Figure 2) shows mod3 as base peak (93.4% area) at 9.28 min and side products in earlier chromatographic peaks with relative peak areas from 0.7-2.5%. These values reflect the peak areas within the UV chromatogram. The overall purity provided in the Result table (Figure 1) was calculated at 93.4% as well. Purity also includes unrelated MS peaks within the target molecule’s chromatographic peak, such as coeluting side-products or unidentified by-products. Figure 1 Result summaries of all 8 RNA 24mers (top) from one batch OligoQuest autoMSMS analysis; detailed results for the selected mod3 sample in the Expected table (bottom); deselected entries are excluded from result reports.They cannot be detected by LC-UV analysis alone but in combination with MS. Adduct peaks from, e.g., sodium or triethylamine are not acknowledged as side products but contribute to the FLP abundance. For each sample multiple quality attributes are determined, which have been previously defined in the workflow method. For this sample, all reporting attributes are shown in green, confirming that all narrow definitions of acceptance criteria were matched (Figure 3). The deconvoluted mass spectrum obtained from the mod3 peak in the chromatogram can be overlayed with the theoretical isotope pattern which was calculated based on the elemental composition of the sequence to provide information about the quality of the match (Figure 4). Partial c/u conversions (shift of +1 Da) for example can be detected sensitively by this comparison (see Figure 8). Accurate mass and isotope pattern add credibility to proposed side products for which MS/MS data were not obtained or insufficient. Figure 3 Multi Attributes view, displaying the Narrow and Wide acceptance criteria for 5 quality attributes and the values determined as Sample Result. Figure 2 UV 260 nm Chromatogram and the Chromatogram Peaks table used for quantitative assessment of side products.Figure 4 Overlay between experimental (blue) and calculated (red) isotope pattern of the mod3 FLP allows to assess, e.g., the absence of C/U conversion. The Sequence Map (Figure 5) supplies a clear overview about the quality of the match between MS/MS data and the selected sequence in the Expected table. The 5´-fragment assignment matches are shown in red bricks, 3´-fragments are shown in blue bricks (ppm errors inside). The sequence with index numbers is counted from 5´-end above and from 3´-end below. The green bottom line uses more stringent sequence validation criteria, serving to confirm only residues that are bracketed by 3’ and 5’ fragment ions: the number inside reports the redundancy in the validation for each individual residue. The intensity of the fragment ions is color coded in 3 intensity ranges to guide the manual validation of weak fragment ions. Figure 5 Sequence Map shows the sequence coverage (green bar) of the MS/MS fragment ions for the mod3 24mer.To verify the individual matches in the Sequence Map, fragment ions and charge states can be selected in BioPharma Compass and the respective profile spectrum of the fragment is shown together with its theoretical isotope pattern (Figure 6). Side Products Characterization As shown in Figure 1 for mod3, 5 side product candidates were returned from the OligoQuest autoMSMS analysis: mod3-c1, mod3+g1, mod3>> [a,g]16, mod3>>[u,c] and mod3>>[a,c]2. They were obtained by screening the mod3 sequence 5OH c a c g c g u g c u u u u g c a c g g c g u g c 3OH for residue losses (-c1), additions (+g) or exchanges ([a,g]16, [a,c]2 and [u,c]). The residue-specific assignment was omitted here if the MS/MS score did not suggest a clear location of the sequence variation. Figure 6 The z9 (-2) fragment in the MS/MS spectrum of mod3. The theoretical pattern is overlayed automatically in red; another -1 fragment at m/z 1487.17 is also present. Figure 7 Sequence matches for the a/g (top) and a/c (bottom) exchanges at positions 2 and 16 in mod3 side products. a2g a2c a16g a16cFigure 8 A Analysis of the -c1 variant by autoMSMS; the intact mass spectrum shows the presence of an added species at -1 Da. (B, C) The targetedMSMS analysis finds that variant as undergoing an additional u9c conversion with a high quality MS/MS spectrum, a conclusive sequence match and the highest MSMS score. D Summary of the targetedMSMS analysis of mod3-c1. D In Figure 7 the sequence matches for the exchange variants a/c and a/g are shown. The variants a16g and a2c are the best matches and were, therefore, marked “Confirmed” in the Expected table (Figure 1). The elucidation of the -c1 side product benefits from the high isotopic fidelity of the data: the MS spectrum shows the presence of an additional sequence (Mr 7649.346), which is 1 Da lighter than the -c1 variant, and the match of the autoMSMS data to the sequence is not good enough to confirm the loss of c1 and exclude any loss of c at another position (Figure 8A). A targeted analysis of the most abundant charge state -4 at 12-13 min yielded a much better coverage of the sequence (Figure 8B) which allowed to assign the loss of c to residue 1. However, automatic screening for further sequence variants of mod3-c1 yielded the u9c-variant (mod3-c1,u9c) as best match (Figure 8C and D). This helps explain the isotope pattern of the mod3-c1. In fact, one can quantify the composition of that isotope pattern´s constituents based on the isotope pattern: the chromatographic peak at 12.17 min is composed of 22% mod3-c1 and 78% mod3-c1,u9c; both species are coeluting with this UHPLC separation and are only 1 Da apart. A B C-c1 (autoMSMS)-c1 (targetedMSMS)-c1, u9c (targetedMSMS)Bruker Daltonics is continually improving its products and reserves the right to change specifications without notice. © Bruker Daltonics 09-2023, LCMS-215, 1904245 For Research Use Only. Not for use in clinical diagnostic procedures. firstname.lastname@example.org – www.bruker.com Bruker Scientific L