Challenges and Solutions in Oligonucleotide Analysis, Part III: LC–MS Methods

The recent expansion of pharmaceutical portfolios to include more therapeutic oligonucleotides (ONs) has led to a dramatic increase in research of chromatographic methods for analysis and purification of these molecules. This has been accompanied by a new set of method development and troubleshooting challenges, particularly for those whose background has been focused on small molecule separations, which have not previously been addressed in any detail in the LC Troubleshooting column. Thus, I am thrilled to have Martin Gilar join me for a series of “LC Troubleshooting” installments focused on challenges encountered in LC analyses of ONs. Gilar is one of the world’s experts on this topic, and this series of articles should be a rich resource for LC practitioners working in this area.
Dwight Stoll

Mass spectrometry is an important tool for both qualitative and quantitative characterization of ONs. LC–MS is often utilized to quantify ON impurities that are structurally similar to the parent ON, and therefore difficult to resolve chromatographically.^1,2 Early reports on analysis of ONs using electrospray ionization (ESI) MS noted the unique challenges encountered in analysis of these polyanionic molecules. Among those are low ionization efficiency,³ dilution of the MS signal across several multiply charged species,⁴ cationic adducts in MS,⁵ and suppression of MS signal for ON species by mobile phase components.⁶

Cation-exchange LC (CEX-LC) is certainly not directly compatible with MS detection due to the high concentrations of inorganic salts used in the mobile phase,⁷ but even ion-pair reversed-phase liquid chromatography (IP-RPLC) and hydrophilic interaction chromatography (HILIC) with semivolatile mobile phases such as triethylammonium acetate (TEAA) or ammonium acetate (NH₄Ac), respectively, can suppress ONs signals to a degree.⁸ Nevertheless, IP-RPLC and HILIC mobile phases with concentrations of alkylammonium acetate or NH₄Ac higher than 10 mM have been used while retaining acceptable detection sensitivity (~ 5–20 pmol of ON injected on 2.1-mm-i.d. column provides useful signal). Postcolumn addition of a sheath solution,⁴ and postcolumn cation-exchange suppressors⁹ have also been used to enhance LC–MS sensitivity.

A major breakthrough in LC–MS of ONs was achieved by Apffel and colleagues in 1997.¹⁰ They introduced IP-RPLC mobile phases composed of triethylamine buffered with hexafluoroisopropanol (HFIP). HFIP is a weakly acidic, volatile alcohol that is highly compatible with ESI MS ionization of ONs. Alkylamine-HFIP mobile phases provide 2 to 3 orders of magnitude greater LC–MS sensitivity compared to alkylamine-acetate or NH₄Ac mobile phases. The alkylamine-HFIP mobile phases enable quantitation of about 250 femtomole ON injected on 2.1-mm-i.d. column.¹¹ The major drawback associated with using IP-RPLC conditions together with MS detection is that residual alkylamine in the LC system may interfere with the subsequent LC–MS analyses in the positive ESI MS mode. This is one of the drivers motivating the search for alternatives to IP-RPLC for LC–MS analysis of ONs. Recently, several groups have been evaluating HILIC as an “ion-pair-free” LC–MS method for analysis of ONs.^12-15

How to Interpret an Oligonucleotide ESI-MS Spectrum

Figure 1 shows an example of an MS spectrum for a 30 nt long ON acquired with TEA-HFIP mobile phase. The MS signal is distributed among several charge states labeled in Figure 1a. The distribution of charge states depends on the ON length, ESI desolvation parameters, ion-pairing system choice, and other factors. Panel B shows the detail of -3 charge state. The MS signal shows the presence of alkali cation adducts. The level of adducts in Figure 1b is acceptable but may increase over time (lifetime of the mobile phase). The main sources of the adducts are solvents, IP reagents, and HFIP. Short 10–20 nt ONs typically show one to two dominant charge states. The number of charge states increases with the ON length and so does the level of adduction. That is, longer ONs tend to be observed with stronger adduct signals compared to shorter ONs. MS analysis of ONs longer than about 50 nt is rather challenging.

Manufacturers of mass spectrometers suitable for ON analysis by MS also produce software built to “deconvolve” mass spectra for the purpose of calculating the ON mass from the observed mass-to-charge ratios of several charge states (for example, -3, -4, -5, and so on) and adduct ions (for example, (ON + K⁺). For certain applications, the mass spectral resolution of ON isotope peaks (for example, ONs with 1 or more ¹³C atoms) is desired. Quadrupole MS instruments (typically exhibiting a resolving power of 2000) do not afford resolution of isotope peaks for ONs longer than about 20 nt; this is illustrated for the -4 m/z charge state shown in Figure 1c. In such a case, the ON mass is reported as the average mass. Time-of-flight (TOF) instruments with resolving powers on the order of 20,000 can resolve the isotope peaks (Figure 1d); if the isotope peaks can be resolved, then the monoisotopic mass of the ON can be reported (1941.66 Da, -4 m/z peak is labeled with asterisk). However, for molecules of this size, the monoisotopic peak is present in relatively low abundance. If the monoisotopic peak is missed and incorrectly assigned, this can result in a deconvolution mass error of one Dalton. Obtaining the monoisotopic mass measurement for ONs on the order of 100 nt will require an MS instrument with a resolving power higher than 100,000, and the monoisotopic peak will be present at very low abundance (data not shown). Therefore, ONs masses are typically reported as an average mass.

IP-RPLC Compatibility With ESI-MS

Mobile phases with 5–25 mM concentration of alkylamine-acetate buffers are useful for pharmaceutical LC-MS applications, where ON sample concentrations are relatively high (1–100 mM).^8,16 For clinical analyses of ONs in biological fluids, the alkylamine-HFIP mobile phase systems are more suitable due to the improvement in detection sensitivity that HFIP provides (analyte concentrations in the range of 1–100 nM). HFIP behaves as a weak acid with pK_a of approximately 9.3. Typically, buffers are prepared such that the molar ratio of alkylamine–HFIP is about 1:10 in water. For example, 5/50 mM or 10/100 mM of amine–HFIP concentrations have pH values in the range of 8.5 to 9.0.⁸ Further increases in alkylamine concentration will push the pH above 9.5, with a concomitant reduction in the fraction of the alkylamine that is protonated and positively charged (most alkyamines have pK_as around 10.5). Decreases in the concentration of positively charged alkylamine in the mobile phase will lead to sharp declines in ON retention and separation.¹⁷

The high cost and environmental concerns related to HFIP have motivated a search for alternatives,^18,19 such as other perfluorinated alcohols or volatile acids. However, HFIP remains the predominant buffering constituent used in contemporary IP-RPLC-MS oligonucleotide analysis. The problem of metal cation adduction (for example, K⁺, Na⁺, and so on) in ON MS spectra was mentioned above. The inclusion of alkylamines in the mobile phase helps to mitigate this problem. Protonated alkylamines compete with the alkali ions for interaction with the negatively charged ON backbone, but the amines are removed from the ON during the ionization and desolvation process. As a result, the level of sodium and potassium adduction is typically acceptable when using IP-RPLC–MS conditions (Figure 2a). Adduction becomes more pronounced for long ONs, such as 100-nt gRNA. Sodium and potassium are the most common adducts; trace levels of alkali metals originate from solvents, IP reagents and HFIP. As the mobile phase “ages,” the level of adduction increases, presumably due to Na⁺ and K⁺ ions leaching from the glass solvent bottles (compare Figure 2d to 2c). Diminished MS signal has also been linked to IP reagent oxidation.^20,21 Figure 2 illustrates that high-quality HFIP is preferred for LC–MS; HFIP obtained from different commercial sources contain highly variable levels of potassium.

HILIC–MS Analysis of ONs

Some LC–MS practitioners are hesitant to use IP reagents in the mobile phase, because the alkylamines can linger in the LC–MS instrument, even after switching to IP-free mobile phase, and can cause signal suppression mass spectral interference in subsequent analyses that use positive ESI-MS mode. Dedicating an LC–MS system for analyses using ion-pairing conditions is a common practice in some industrial laboratories. HILIC–MS has been explored as an alternative to IP-RPLC for analysis of ONs. HILIC has somewhat different separation selectivity compared to IP-RPLC, further justifying its use in some cases.^12,15,22,23 Figure 3 shows that HILIC coupled with a simple QDa MS detector has sufficient sensitivity to analyze 20 pmol of ON injected on column. Higher-end MS instruments can achieve substantially lower detection limits. Figure 3 illustrates an important drawback of HILIC–MS analyses that is not widely appreciated—alkali metal ion adducts are more prominent in the spectra of ONs larger than 25 nt in the absence of IPRP reagents in the mobile phase (see above for discussion of competition effects). When analyzing samples containing ONs at high concentrations, the adducts are not a problem, but when more sensitivity is needed the relatively more abundant adduct signal may compromise MS data clarity because the ON mass is spread out across several mass spectral peaks.

RNA Mapping Applications

LC–MS analysis of long nucleic acids (for example, guide RNAs of about 100 nt, messenger RNAs of about 1000 nt) is generally difficult. Alkali ion adduction, signals spread across multiple charge states, and sample heterogeneity (for example, mRNA polyA tail heterogeneity) all contribute to low MS sensitivity. These problems could be mitigated by first digesting the RNA (or DNA) with nucleases into shorter ONs prior to LC–MS analysis.²⁴ This approach, known as RNA mapping, is conceptually similar to peptide mapping (digestion of proteins with specific enzymes such as trypsin).

RNA mapping for an 80 nt RNA is illustrated in Figure 4, where the RNA is digested with RNase T1 enzyme. RNase T1 selectively cleaves the RNA strand at the 3’ side of guanidine nucleotides, leaving a phosphate group at 3’ end of the resulting fragment. The 80 nt ON sequence shown in Figure 4 was chosen to illustrate a few limitations of the RNA mapping approach. First, the abundance of G or GG motifs in the sequence results in many short ON products and Gp mononucleotide(s). Such short oligonucleotides are not as informative for confirmation of the 80 nt RNA sequence as are 6–12 nt long ONs. In addition, isobaric ONs (peaks 9a and 9b) can also be produced by the digest. In this case, ONs 9a and 9b have the same base composition, and thus the same mass, but different sequences of the same bases. MS/MS sequencing may be used to provide further information on the specific sequences that are present. Alternatively, parallel digests with enzymes having complementary selectivity can be used to obtain full sequence coverage from simple LC–MS experiments.²⁵

RNA mapping is often performed for sequence verification of mRNAs.²⁶ The PolyA tail microheterogeneity can be also assessed by LC–MS. Figure 5 shows the separation of RNA fragments produced from the digestion of Fluc-beta mRNA (1922 nt) with RNase T1. The enzyme cleaves the polyA tail from the 3’ end of the mRNA; the polyA tail is the last eluting peak observed in Figure 5a. The MS spectra across the polyA tail peak were deconvoluted with MaxEnt1 software. The resulting MS spectrum shown in Figure 5b illustrates that the polyA tail exists as multiple species differing by a single adenosine nucleotide. This finding was confirmed by other methods.²⁷ The polyA tail ranged from 122–137 nt ONs, whereas the dominant form is about 126–128 nt.

Summary

LC–MS analysis is an essential tool for contemporary ON analysis. The MS can provide evidence of ON modifications even when physical resolution of structurally similar ONs cannot be achieved by chromatographic methods alone. LC–MS analysis involving both IP-RPLC and HILIC modes is common for short (15–30 nt) ONs commonly used as therapeutic compounds. LC–MS analysis of longer nucleic acids (for example, guide RNAs of about 100 nt, messenger RNAs of about 1000–5000 nt) remains difficult. RNA mapping methods have been developed to address this challenge by converting large RNA molecules into short oligonucleotides.

While MS detection is very powerful in this context, it cannot distinguish between isobaric ONs, such as phosphorothioate diastereomers or isobaric ONs species (for example, deaminations observed in different sequence positions). Thus, the combination of LC and MS is very powerful. In this installment we have highlighted some challenges encountered with these methods, such as the formation of alkali metal ion adducts, and recommended practical solutions to these problems. It is important to note that nonspecific adsorption (NSA) adsorption of ONs on LC–MS hardware has to be mitigated¹¹, otherwise the lower detection limit can be compromised. We will discuss this issue in some detail in a future installment in this series.

BEH is a trademark of Waters Technologies Corporation.

References

1. Roussis, S. G.; Cedillo, I.; Rentel, C. Semi-Quantitative Determination of Co-Eluting Impurities in Oligonucleotide Drugs Using Ion-Pair Reversed-Phase Liquid Chromatography Mass Spectrometry. J Chromatogr A 2019, 1584, 106–114. DOI: 10.1016/j.chroma.2018.11.034

2. Gaus, H. J.; Owens, S. R.; Winniman, M. et al. On-Line HPLC Electrospray Mass Spectrometry of Phosphorothioate Oligonucleotide Metabolites. Anal Chem 1997, 69 (3), 313–319. DOI: 10.1021/ac960557q

3. Liu, C.; Wu, Q.; Harms, A. C.; Smith R. D. On-Line Microdialysis Sample Cleanup for Electrospray Ionization Mass Spectrometry of Nucleic Acid Samples. Anal Chem 1996, 68 (18), 3295–3299. DOI: 10.1021/ac960286j

4. Huber, C. G.; Krajete, A. Analysis of Nucleic Acids by Capillary Ion-Pair Reversed-Phase HPLC Coupled to Negative-Ion Electrospray Ionization Mass Spectrometry. Anal Chem 1999, 71 (17), 3730–3739. DOI: 10.1021/ac990378j

5. Greig, M.; Griffey, R. H. Utility of Organic Bases for Improved Electrospray Mass Spectrometry of Oligonucleotides. Rapid Commun Mass Spectrom. 1995, 9 (1), 97–102. DOI: 10.1002/rcm.1290090121

6. Bleicher, K.; Bayer, E. Various Factors Influencing the Signal Intensity of Oligonucleotides in Electrospray Mass Spectrometry. Biol Mass Spectrom 1994, 23 (6), 320–322. DOI: 10.1002/bms.1200230604

7. Thayer, J. R.; Puri, N.; Burnett, C. et al. Identification of RNA Linkage Isomers by Anion Exchange Purification with Electrospray Ionization Mass Spectrometry of Automatically Desalted Phosphodiesterase-II Digests. Anal Biochem 2010, 399 (1), 110–117. DOI: 10.1016/j.ab.2009.11.009

8. Donegan, M.; Nguyen, J. M.; Gilar, M. Effect of Ion-Pairing Reagent Hydrophobicity on Liquid Chromatography and Mass Spectrometry Analysis of Oligonucleotides. J Chromatogr A 2022, 1666, 462860. DOI: 10.1016/j.chroma.2022.462860

9. Huber, C. G.; Buchmeiser, M. R. On-Line Cation Exchange for Suppression of Adduct Formation in Negative-Ion Electrospray Mass Spectrometry of Nucleic Acids. Anal Chem 1998, 70 (24), 5288–5295. DOI: 10.1021/ac980791b

10. Apffel, A.; Chakel, J. A.; Fischer, S. et al. Analysis of Oligonucleotides by HPLC−Electrospray Ionization Mass Spectrometry. Anal Chem 1997, 69 (7), 1320–1325. DOI: 10.1021/ac960916h

11. Basiri, B.; Sutton, J. M.; Hooshfar, S. et al. Direct Identification of Microribonucleic Acid miR-451 from Plasma Using Liquid Chromatography Mass Spectrometry. J Chromatogr A 2019, 1584, 97–105. DOI: 10.1016/j.chroma.2018.11.029

12. Goyon, A.; Nguyen, D.; Boulanouar, S. et al. Characterization of Impurities in Therapeutic RNAs at the Single Nucleotide Level. Anal Chem 2022, 94 (48), 16960–16966. DOI: 10.1021/acs.analchem.2c04681

13. Lobue, P. A.; Jora, M.; Addepalli, B. et al. Oligonucleotide Analysis by Hydrophilic Interaction Liquid Chromatography-Mass Spectrometry in the Absence of Ion-Pair Reagents. J Chromatogr A 2019, 1595, 39–48. DOI: 10.1016/j.chroma.2019.02.016

14. Abernathy, S.; Rayhan, A.; Limbach, P. A. Stationary Phase Effects in Hydrophilic Interaction Liquid Chromatographic Separation of Oligonucleotides. Analyst 2025, 150 (1), 185–196. DOI: 10.1039/d4an01155d

15. Gilar, M.; Koshel, B. M.; Birdsall, R. E. Ion-Pair Reversed-Phase and Hydrophilic Interaction Chromatography Methods for Analysis of Phosphorothioate Oligonucleotides. J Chromatogr A 2023, 1712, 464475. DOI: 10.1016/j.chroma.2023.464475

16. Roussis, S. G.; Pearce, M.; Rentel, C. Small Alkyl Amines as Ion-Pair Reagents for the Separation of Positional Isomers of Impurities in Phosphate Diester Oligonucleotides. J Chromatogr A 2019, 1594, 105–111. DOI: 10.1016/j.chroma.2019.02.026

17. Li, N.; El Zahar, N. M.; Saad, J. G. et al. Alkylamine Ion-Pairing Reagents and the Chromatographic Separation of Oligonucleotides. J Chromatogr A 2018, 1580, 110–119. DOI: 10.1016/j.chroma.2018.10.040

18. Basiri, B.; van Hattum, H.; van Dongen, W. D. et al. The Role of Fluorinated Alcohols as Mobile Phase Modifiers for LC-MS Analysis of Oligonucleotides. J Am Soc Mass Spectrom 2017, 28 (1), 190–199. DOI: 10.1007/s13361-016-1500-3

19. Liu, R.; Ruan, Y.; Liu, Z. et al. The Role of Fluoroalcohols as Counter Anions for Ion‐Pairing Reversed‐phase Liquid Chromatography/High‐Resolution Electrospray Ionization Mass Spectrometry Analysis of Oligonucleotides. Rapid Commun Mass Spectrom 2019, 33 (7), 697–709. DOI: 10.1002/rcm.8386

20. Birdsall, R. E.; Gilar, M.; Shion, H. et al. Reduction of Metal Adducts in Oligonucleotide Mass Spectra in Ion‐pair Reversed‐phase Chromatography/Mass Spectrometry Analysis. Rapid Commun Mass Spectrom 2016, 30 (14), 1667–1679. DOI: 10.1002/rcm.7596

21. Guimaraes, G. J.; Saad, J. G.; Annavarapu, V. et al. Mobile Phase Aging and Its Impact on Electrospray Ionization of Oligonucleotides. J. Am. Soc. Mass Spectrom. 2023, 34 (12), 2691–2699. DOI: 10.1021/jasms.3c00264

22. Gilar, M.; Stoll, D., R. Challenges and Solutions in Oligonucleotide Analysis, Part II: A Detailed Look at Ion-Pairing Reversed-Phase Separations. LCGC Int 2026, 3 (2), 8-15. DOI: 10.56530/lcgc.int.mh1387e9

23. Sorensen, M. J.; Paulines, M. J.; Maloney, T. D. Evaluating Orthogonality between Ion-Pair Reversed Phase, Anion Exchange, and Hydrophilic Interaction Liquid Chromatography for the Separation of Synthetic Oligonucleotides. J Chromatogr A 2023, 1705, 464184. DOI: 10.1016/j.chroma.2023.464184

24. Goyon, A.; Blevins, M. S.; Napolitano, J. G. et al. Characterization of Antisense Oligonucleotide and Guide Ribonucleic Acid Diastereomers by Hydrophilic Interaction Liquid Chromatography Coupled to Mass Spectrometry. J Chromatogr A 2023, 1708, 464327. DOI: 10.1016/j.chroma.2023.464327

25. Menneteau, T.; Addepalli, B.; Johnston, T. et al. Advancing Gene Therapy: Enzyme Selection For Effective RNA Oligonucleotide Mapping. LCGC Int 2025, 26–33. DOI: 10.56530/lcgc.int.nv8977m6

26. Jiang, T.; Yu, N.; Kim, J. Oligonucleotide Sequence Mapping of Large Therapeutic mRNAs via Parallel Ribonuclease Digestions and LC-MS/MS. Anal Chem 2019, 91 (13), 8500–8506. DOI: 10.1021/acs.analchem.9b01664

27. Gilar, M.; Doneanu, C.; Gaye, M. M. Liquid Chromatography Methods for Analysis of mRNA Poly(A) Tail Length and Heterogeneity. Anal Chem 2023, 95 (38), 14308–14316. DOI: 10.1021/acs.analchem.3c02552

Download the PDF Here