Streamline High-Throughput Oligonucleotide Analysis With Ion-Pairing-Free Methods

Application Note Biopharma Authors Lee Bertram and Brian Rivera Agilent Technologies, Inc. Abstract This application note presents a practical approach for medium- to high-throughput molecular confirmation of synthetic oligonucleotides using the Agilent InfinityLab Pro iQ Plus LC/MS system. Traditional ion-pair reversed-phase LC/MS methods often require dedicated instrumentation and toxic, costly reagents. In contrast, the method described here uses an ammonium bicarbonate-based method, eliminating the need for ion-pairing agents while maintaining sufficient chromatographic retention and MS sensitivity. Agilent OpenLab CDS MS spectral deconvolution simplifies data analysis workflows, enabling automated processing with minimal optimization. This approach provides robust LC/MS performance for various oligonucleotide types, including antisense oligonucleotides and siRNAs. Analysis of Oligonucleotides Using Ion-Pairing Alternatives on the Agilent Pro iQ Plus 2 Introduction Oligonucleotides are an emerging therapeutic modality that target and modulate gene expression through the silencing or degradation of mRNA. Once an mRNA target is identified, the antisense sequence must be optimized with strategically incorporated chemical modifications to improve pharmacokinetics1 , affinity, and minimize off-targeting or mismatching.2 Ion-pair reversed-phase LC/MS is commonly used to confirm the molecular weights of target oligo sequences to ensure proper synthesis. However, using alkylamines as ion-pairing agents often requires dedicated instrumentation. Furthermore, toxic and cost-prohibitive perfluorinated alcohols such as HFIP (hexafluoroisopropanol) are required for optimal chromatographic separation and MS sensitivity. In this application, an alternative reversed-phase approach was used for molecular confirmation of oligonucleotides. The method used ammonium bicarbonate instead of ion pairing, while still providing sufficient chromatographic retention and MS sensitivity. Twenty replicate injections of three different antisense oligonucleotides and five replicates of a single siRNA were performed to ensure the applicability and reproducibility of the method. Experimental Instrument configuration This experiment was conducted using the following instrument configuration: – Agilent InfinityLab Pro iQ Plus LC/MS system (G6170A) – Agilent Infinity II 1290 bio binary pump (G7120A) – Agilent Infinity II 1290 bio multisampler (G7167B) – Agilent Infinity II 1290 bio column compartment (G7116B) – Agilent Infinity II 1260 diode array detector HS (G7117C) Although this analysis used an Infinity II LC configuration, comparable results can be achieved on the Infinity III LC system with no changes to method parameters. Sample preparation All samples were resuspended in deionized (DI) water to a concentration of 50 µM and stored at –80 °C. Samples were transferred to polypropylene vials and stored in the temperature-controlled autosampler for up to two days prior to analysis. Oligonucleotide sample sequences are shown in Table 1. Oligonucleotide Name Length Sequence ASO-1 18 dU/MOErC//MOErA//MOErC/dUdUdU/MOErC// MOErA/dU/MOErA//MOErA/dU/MOErG/CdU/ MOErG/G ASO-2 20 dU/MOErC/dUdU/MOErG/TT/MOErA//MOErC// MOErA//MOErT//MOErG//MOErA//MOErA// MOErA/dU/MOErC//MOErC//MOErC/C Fomivirsen 21 G*C*G*T*T*T*G*C*T*C*T*T*C*T*T*C*T*T*G*C*G Givosiran 22 S mC*mA*mGmAmAmAfGmAfGmUfGmUfCmUfCmAm UmCmUmUmA/L96/ 23 AS mU*mG*mGfUmCfUmUfUfCmUfCfAmCfAmGfAmGfU mAmGfA*fA*mU Code Description /MOErA/ Methoxyethoxy A /MOErC/ Methoxyethoxy C /MOErT/ Methoxyethoxy T /MOErG/ Methoxyethoxy G dU Deoxyuridine fA 2-fluoroadenosine fC 2-fluorocytidine fG 2-fluoroguanadine fU 2-fluorouridine * Phosphorothioate bond A 2'-deoxyribose adenine C 2'-deoxyribose cytosine G 2'-deoxyribose guanine T 2'-deoxyribose thymine mA 2'-O-methyl A mC 2'-O-methyl C mG 2'-O-methyl G mU 2'-O-methyl U rA Ribose adenine rC Ribose cytosine rG Ribose guanine rU Ribose uracil Table 1. Oligonucleotide sample sequences. 3 LC/MS analysis Source parameters for the Pro iQ Plus system are provided in Table 2, while high-performance liquid chromatography (HPLC) parameters are provided in Table 3. Mass Spectrometry Parameters Parameter Value MS Agilent Pro iQ Plus Source Agilent Jet Stream Electrospray Ionization (AJS-ESI) Source Drying Gas Flow 13.0 L/min Gas Temp 300 °C Nebulizer Pressure 35 psi Capillary Voltage 3,000 V Sheath Gas Temp 250 °C Sheath Gas Flow 11 mL/min Nozzle Voltage 1,500 V Mode Positive Scan m/z 700–2,800 Scan Time 1,250 ms Fragmentor 180 V Gain Factor 5 Table 2. Source parameters for the Agilent Pro iQ Plus system. Parameter Value Column Agilent AdvanceBio oligonucleotide column, 2.1 × 50 mm, 2.7 µm Sampler Temperature 8 °C UV Detection 260/4 nm (Ref 360/20 nm) Peak width > 0.1 min (2.5 Hz) Mobile Phase A 20 mM ammonium bicarbonate in DI water Mobile Phase B Methanol Flow Rate 0.7 mL/min Injection Volume 2 µL Multiwash 20:80 water:methanol; flush port; 5 seconds 90:10 water:methanol; flush port; 3 seconds Column Temperature 75 °C Post Time 1.0 min Gradient Program Time (min) %B 0 5 0.1 5 3.0 40 3.1 80 3.5 80 3.6 5 Table 3. HPLC parameters used. Results and discussion Established in 1997, an alkylamine ion pair with perfluorinated alcohol as the acidic modifier is the preferred mobile phase for LC/MS analysis of oligonucleotides.3 This is due to its chromatographic performance and electrospray efficiency, especially when compared to mobile phases using acetate as the counter ion. Extensive work has further demonstrated a wide experimental design space when using this powerful ion-pairing system.4 However, there are consequences when using alkylamine and HFIP for LC/MS methods. First, optimization of buffer components is necessary since oligo modifications and sequence affect electrospray desorption and chromatography.5 Second, alkylamine may contaminate ionization sources and LC systems, leading to background peaks if polarity is switched back to positive mode. Consequently, this requires extensive cleaning/passivation of LC components and ion source surfaces. Even then, some labs may dedicate systems to negative mode due to the adsorption of alkylamines onto the LC/MS system. Finally, alkylamine-containing mobile phases have a short shelf life if not kept sealed under argon gas, and thus, mobile phases must be made fresh, sometimes daily, to ensure consistent method performance. For labs performing molecular weight confirmation workflows where many oligos are analyzed, optimizing mobile phase conditions may not be feasible. Additionally, LC/MS downtime due to instrument maintenance and daily preparation of buffers may lead to sample backlogs, which for many labs can be a considerable challenge. A more practical approach would be to use a non-ion-pairing, reversed-phase methodology. Recent work provides a more practical and cost-effective alternative to ion pairs, using an ammonium bicarbonate (NH4 HCO3 ) buffer and methanol as the strong solvent.6 This is an advantageous method for molecular weight confirmation, as ESI sensitivity and chromatographic performance are sufficient, even with minimal optimization of the mobile phase and gradient. Further, this referenced work postulates that carbon dioxide outgassing facilitates droplet formation, while ammonia evaporation contributes to proton adduction, thus allowing for positive mode analysis. This, therefore, eliminates the need to dedicate a system for negative mode analysis. 4 Because oligonucleotides are polar and negatively charged, there may be concerns that these analytes would not be retained or separate chromatographically without ion pairing. However, ammonium bicarbonate-based methods retain a wide variety of oligonucleotides and perform particularly well with modified oligos, such as antisense oligos (ASOs; Figure 1). The method demonstrated here provides a proof-of-concept for medium- to high-throughput analysis. Therefore, the gradient is 5% to 40% Mobile Phase B over three minutes. This relatively "ballistic" gradient provides excellent retention and peak shape, and the 0.7 mL/min flow rate does not negatively impact the spectral quality nor sensitivity. Like most large molecules with multiple ionizable functional groups, oligos appear as multiply charged ions when analyzed by ESI-MS. Importantly, depending on the alkylamine ion pair and source parameters, charge state distribution of the oligo may vary.7 This can be problematic with heated ionization sources, which tend to yield higher charge states. Consequently, deconvolution of the spectrum may lead to artifacts and potential misidentifications. On the other hand, ammonium bicarbonate (ABC) tends to yield oligos with lower charge states (Figure 2). The higher m/z values will have less spectral overlap with matrix interferences, impurities, or other components, thus improving confidence in the deconvoluted data. Figure 1. UV chromatograms for three antisense oligos (ASOs). Fomivirsen is a 21mer ASO with no 2′-modifications. Thus, it elutes earlier than the two shorter, fully thioated, 2′-modified ASOs. 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 Retention time (min) 0 10 20 30 40 50 60 70 80 90 100 110 Relative response (%) 5 Agilent OpenLab CDS uses a unique spectral deconvolution that is optimized for unit mass detectors. Maximum entropy calculates the most probable mass from a spectrum by increasing dominant features while minimizing noise. In contrast, OpenLab CDS spectral deconvolution simply identifies ions in the spectrum and matches them to target masses based on user inputs. These values are then fit to a linear regression or centroided to calculate a molecular weight average. This simple approach allows for easy molecular weight confirmation using generic processing methods, which can be fully automated with minimal optimization. Triplicate injections of each sample yield nearly identical results, with relative standard deviations (%RSDs) < 0.1%. The automated deconvolution settings used in the sample sequence for acquisition are shown in Table 4, and a summary of deconvolution results is shown in Table 5. Parameter Value Automatic Deconvolution Settings Run Automatic Deconvolution Enabled RT Window 1–2.7 min TIC Peak Type All peaks TIC Top (n) Peaks 3 Basic Settings Use m/z Range Disabled Low/High Molecular Weight 3,000–10,000 Maximum Charge 10 Minimum Peaks in Set 3 Advanced Settings MW Agreement (0.01%) 10 Absolute Noise Threshold 100 Relative Abundance Threshold 10 MW Algorithm Curve fit MW Algorithm Threshold (%) 40 Envelope Threshold (%) 50 Table 4. MS spectral deconvolution settings. Figure 2. Full-scan spectra for samples. Charge states of 4 to 6 are predominant for each ASO. 800 1,000 1,200 1,400 1,600 1,800 2,000 2,200 2,400 2,600 2,800 m/z 0 1 2 3 Abundance 703.1 834.9 914.7 1,026.7 1,045.1 1,113.8 1,197.6 1,219.1 1,286.3 1,440.1 1,462.7 1,598.6 1,733.4 1,828.2 1,954.6 2,088.5 2,179.2 2,337.3 2,437.3 2,583.2 2,682.4 Max: 383,860 800 1,000 1,200 1,400 1,600 1,800 2,000 2,200 2,400 2,600 2,800 m/z 0 1 Abundance 709.5 822.4 907.5 1,033.7 1,059.0 1,128.9 1,240.4 1,270.3 1,287.7 1,416.01,501.2 1,587.7 1,589.3 1,716.5 1,874.4 1,993.4 2,117.3 2,250.8 2,343.1 2,457.0 Max: 222,126 800 1,000 1,200 1,400 1,600 1,800 2,000 2,200 2,400 2600 2,800 m/z 0 2 4 Abundance 721.2 835.0 956.31,041.3 1,089.7 1,114.5 1,115.3 1,124.31,212.6 1,306.2 1,336.9 1,338.0 1404.4 1,489.2 1,584.5 1,670.6 1,671.6 1,672.8 1,691.4 1,781.6 1,873.9 1,969.6 2,063.8 2,226.0 ×10 Max: 63,956 4 ×105 ×105 6 Figure 3 shows the deconvoluted spectrum of ASO-1, an 18mer phosphorothioate with 2′-MOE modifications. Component A is a full-length product, showing good agreement with the calculated mass. Component B is likely depurination (loss of guanine). Component C is likely a sodium adduct. As shown, the InfinityLab Pro iQ Plus, which combines unit mass accuracy with simple deconvolution, can monitor oligonucleotides and identify unexpected impurities or components. Further characterization can be performed by high-resolution MS. Oligonucleotide Name Calculated Mass (Da) Average Measured Mass (Da) Delta Mass (Da) ASO-1 6,348.3 6,347.9 –0.4 ASO-2 7,309.2 7,308.8 –0.4 Fomivirsen 6,682.4 6,681.8 –0.6 Table 5. Deconvoluted results of twenty replicates of each ASO. A duplex, tri-antennary GalNAc-conjugated siRNA (Givosiran) was also analyzed to demonstrate broad applicability of the method. Slight modifications to the LC/MS parameters were required with the duplex conjugate. Specifically, the gradient program was adjusted slightly (5% to 45% in 3 minutes) to ensure baseline separation of the sense and antisense strands. Additionally, the fragmentor voltage was decreased to 120 V to minimize in-source fragmentation of the fragile oligosaccharide conjugate. Figure 4 shows an overlay of UV chromatograms for five injections of Givosiran. Although no ion pair is used, the ABC mobile phase still provides baseline separation of sense and antisense strands (Figure 5). This separation is required because any spectral overlap may result in misidentifications in deconvolution. Spectra are shown in Figure 5, with similar charge state distributions to the previously mentioned ASOs. Analysis of the deconvoluted spectrum for each peak show the antisense strand eluting earlier than the sense strand. Interestingly, the sense strand may indicate GalNAc losses (Figure 6). A summary of Givosiran results is shown in Table 6. Oligonucleotide Name Calculated Mass (Da) Average Measured Mass (Da) Delta Mass (Da) Givosiran, Antisense 7,563.8 7,563.1 0.7 Givosiran, Sense 8,736.5 8,735.6 0.9 Table 6. Deconvoluted results for each injection of Givosiran. Figure 3. Deconvoluted spectrum for ASO-1. 5,800 5,900 6,000 6,100 6,200 6,300 6,400 6,500 6,600 6,700 6,800 6,900 7,000 Mass (Daltons) 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Abundance A:6,347.88 B:6,196.24 C:6,369.33 ×105 7 Figure 4. Overlay of UV chromatograms for five injections of Givosiran. 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00 1.05 1.10 1.15 1.20 1.25 1.30 1.35 1.40 1.45 1.50 1.55 1.60 1.65 1.70 Retention time (min) 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Response (mAU) Integration Integration ×101 Figure 5. Spectra for antisense (A) and sense (B) RNA strands. Charge state distribution is similar to that of antisense oligos analyzed using a similar method. The 5– charge state is the most abundant for antisense, and the 6– charge state is the most abundant for the GalNAc-conjugated sense strand. 800 1,000 1,200 1,400 1,600 1,800 2,000 2,200 2,400 2,600 2,800 m/z 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 Abundance 761.7 843.5 946.3 953.41,039.9 1,081.1 1,087.5 1,151.3 1,254.1 1,261.7 1,265.2 1,283.0 1,422.5 1,508.7 1,513.7 1,517.8 1,519.6 1,521.9 1,613.6 1,718.91,802.11,889.8 1,891.4 1,968.9 2,122.9 2,212.8 2,313.12,397.0 2,536.02,621.62,703.32,780.3 Max: 213,395 800 1,000 1,200 1,400 1,600 1,800 2,000 2,200 2,400 2,600 2,800 m/z 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 Abundance 711.9 776.7 797.5 854.2 874.5 877.0 949.1 951.4 952.7 971.6 974.0 1,067.4 1,081.1 1,092.9 1,095.6 1,098.2 1,176.0 1,219.9 1,222.9 1,248.9 1,252.2 1,255.8 1,352.6 1,406.4 1,422.9 1,430.1 1,626.41,707.7 1,810.6 2,043.02,130.1 2,248.6 2,343.8 2,440.8 2,605.8 2,776.4 Max: 1,101,977 ×105 ×106 A B 8 Figure 6. Deconvoluted spectrum for the Givosiran sense strand. The approximately 200 Da shift observed could be loss of GalNAc. 7,800 8,000 8,200 8,400 8,600 8,800 9,000 9,200 9,400 Mass (Daltons) 0 0.5 1.0 1.5 2.0 2.5 Abundance A:8,735.42 B:8,532.26 C:8,758.10 E:8,328.63 D:8,554.26 ×106 Conclusion The method described in this application note provides a practical approach for medium- to high-throughput molecular confirmation of synthetic oligonucleotides. The Agilent Pro iQ Plus demonstrates excellent resolution and sensitivity for a unit mass detector, even at high m/z ranges that exceed the capabilities of many single quadrupole detectors. The method uses a novel ammonium bicarbonate mobile phase, providing sufficient LC/MS performance for labs analyzing many types of oligos, including antisense oligonucleotides and siRNAs. LC and MS results described in this application note are reproducible. Data analysis workflows are simplified using Agilent OpenLab CDS MS spectral deconvolution, which is automated by unattended processing methods that require minimal optimization. www.agilent.com DE-006128 This information is subject to change without notice. © Agilent Technologies, Inc. 2025 Printed in the USA, May 6, 2025 5994-8337EN References 1. Herkt, M.; Thum, T. Pharmacokinetics and Proceedings in Clinical Application of Nucleic Acid Therapeutics. Mol. Ther. 2021, 29(2), 521–539. DOI: 10.1016/j. ymthe.2020.11.008. 2. Watts, J. K.; Corey, D. R. Silencing Disease Genes in the Laboratory and the Clinic. J. Pathol. 2012, 226(2), 365–379. DOI: 10.1002/path.2993. 3. Apffel, A.; Chakel, J. A.; Fischer, S.; Lichtenwalter, K.; Hancock, W. S. Analysis of Oligonucleotides by HPLC− Electrospray Ionization Mass Spectrometry. Anal. Chem. 1997, 69(7), 1320–1325. DOI: 10.1021/ac960916h. 4. Guimaraes, G. J.; Bartlett, M. G. The Critical Role of Mobile Phase pH in the Performance of Oligonucleotide Ion-Pair Liquid Chromatography-Mass Spectrometry Methods. Future Sci. OA 2021, 7(10), FSO753. DOI: 10.2144/fsoa2021-0084. 5. Basiri, B.; Murph, M. M.; Bartlett, M. G. Assessing the Interplay Between the Physicochemical Parameters of Ion-Pairing Reagents and the Analyte Sequence on the Electrospray Desorption Process for Oligonucleotides. J. Am. Soc. Mass Spectrom. 2017, 28(8), 1647–1656. DOI: 10.1007/s13361-017-1671-6. 6. Hayashi, Y.; Sun, Y. Overcoming Challenges in Oligonucleotide Therapeutics Analysis: A Novel Nonion Pair Approach. J. Am. Soc. Mass Spectrom. 2024, 35(9), 2034–2037. DOI: 10.1021/jasms.4c00270. 7. Chen, B.; Mason, S. F.; Bartlett, M. G. The Effect of Organic Modifiers on Electrospray Ionization Charge-State Distribution and Desorption Efficiency for Oligonucleotides. J. Am. Soc. Mass Spectrom. 2013, 24(2), 257–264. DOI: 10.1007/s13361-012-0509-5.