Analysis of Oligonucleotides Using a Single Quadrupole Mass Spectrometer with Hydrophilic Interaction Chromatography (HILIC)

Significance of the topic

Oligonucleotide therapeutics are an expanding class of modalities that require rigorous analytical workflows to confirm identity, purity and impurity profiles. Reliable methods that avoid toxic or instrument-contaminating additives (e.g., ion-pair reagents) and are compatible with mass spectrometry are especially valuable for QC, method development and routine monitoring of synthetic processes. This work demonstrates that hydrophilic interaction chromatography (HILIC) coupled with a single-quadrupole LC-MS and dedicated oligonucleotide processing software can provide comprehensive characterization of synthetic oligonucleotides and common truncated impurities without ion-pair reagents.

Objectives and study overview

The study aimed to evaluate HILIC as an alternative separation mode to reverse-phase ion-pair chromatography for synthetic oligonucleotide analysis, and to demonstrate impurity detection/quantification using a single-quadrupole mass spectrometer (LCMS-2050) together with LabSolutions Insight Biologics software. The focus was to:

• separate a full-length product (FLP) and nucleotide-deficient variants (n-1 and n-3) using HILIC;
• establish linearity and detection capability for impurities added at low relative concentrations (2–50% of FLP);
• compare mobile-phase chemistries (ammonium acetate vs ammonium bicarbonate) regarding MS sensitivity and retention stability.

Sample

A simulated mixture was prepared containing:

• a 20-mer full-length product (FLP) modeled on the mipomersen sequence;
• a 19-mer n-1(5’) truncated variant;
• a 17-mer n-3(5’) truncated variant.

Component masses were consistent with sequence lengths and common base and backbone modifications (methylated bases, phosphorothioate linkers, methoxy/deoxy ribose variants).

Methodology

HILIC separations were performed using a HILICpak VN-50 2D column on a Nexera XS inert system. Two mobile-phase regimes were tested: ammonium acetate (100 mmol/L CH3COONH4 in water/acetonitrile mixtures) and ammonium bicarbonate (100 mmol/L NH4HCO3, pH 8). Gradients used high organic starting conditions with 0.2 mL/min flow, 60 °C column temperature and 15 µL injection volume. UV detection used a photodiode array (190–800 nm) with major monitoring at 260 nm.

Used Instrumentation

• Liquid chromatograph: Nexera XS inert (Shimadzu).
• Column: HILICpak VN-50 2D (Shodex).
• Mass spectrometer: LCMS-2050 single-quadrupole with ESI (DUIS), negative-ion mode; scan m/z 650–2000.
• Data processing: LabSolutions Insight Biologics for oligonucleotide sequence input, isotope/multiple-charge consolidation and quantitative calculations.

Data processing approach

Insight Biologics was used to construct the target sequence and recognize variant compositions. The software consolidates multi-valent and isotopic ion signals into single component chromatograms. Quantitation combined UV peak area information with MS-derived component proportions: UV peak area scaled by a factor (reported in the study) and apportioned among components according to their summed MS peak areas within the UV peak. Linearity and impurity quantitation were therefore calculated using both orthogonal detectors (UV for total chromatographic signal and MS for composition within overlapping peaks).

Main results and discussion

• Linearity: FLP peak area vs concentration was linear across 1–20 µmol/L with R2 ≥ 0.999. For impurity spikes relative to 10 µmol/L FLP, component peak areas showed excellent linearity (R2 ≥ 0.999) across added relative concentrations from 2% to 50% for both n-1 and n-3 variants.
• Charge state handling: Electrospray produced multiple charge-state distributions (e.g., +3 to +7 or higher depending on variant); Insight Biologics successfully combined these to allow mass-based identification. Mass identification errors for FLP and variants were within approximately 0.3 Da.
• Sensitivity and detection limits: Using HILIC with ammonium acetate, the method detected and identified truncated variants down to 2% relative to FLP (0.2 µmol/L when FLP = 10 µmol/L). However, MS sensitivity with HILIC was substantially lower than with reverse-phase ion-pair methods (reported reductions on the order of one-fifth to one-tenth), indicating challenges for very low-concentration analyses.
• Mobile-phase comparison: Ammonium acetate provided better MS sensitivity and more stable retention times than ammonium bicarbonate. The bicarbonate-containing mobile phase (pH 8) exhibited progressive shortening of retention times over repeated injections due to CO2 dissolution effects and yielded roughly half the MS sensitivity observed with ammonium acetate; raising pH to ~10 with added ammonia did not recover sensitivity.
• Practical separation: HILIC achieved separation of FLP and the two truncated variants under the tested gradients without using ion-pair reagents, and component chromatograms (MS-derived) matched UV chromatographic peaks, enabling quantitation and identity confirmation.

Contributions and practical applications

The study demonstrates that a single-quadrupole LC-MS instrument paired with HILIC and dedicated oligonucleotide processing software can:

• identify and quantify common truncation impurities in synthetic oligonucleotides without ion-pair reagents;
• provide a non-dedicated analytical workflow (HILIC avoids the need for triethylamine-type additives that contaminate LC systems), facilitating multi-use instrument deployment in QC labs;
• use combined UV and MS information to apportion overlapping chromatographic peaks to individual sequences, improving confidence in impurity assignment.

Limitations include reduced MS sensitivity relative to ion-pair reverse-phase methods and retention-time instability with volatile bicarbonate buffers, which constrain detection at very low absolute concentrations.

Future trends and potential uses

HILIC-LC-MS workflows for oligonucleotides are promising for routine impurity profiling where ion-pair reagents are undesirable. Future developments and areas of application include:

• improvements in ionization efficiency for HILIC conditions (source optimization, additives compatible with MS) to close the sensitivity gap versus ion-pair methods;
• software advances for more automated charge-state deconvolution and quantitative isotope-profile modeling on single-quadrupole platforms;
• application to broader oligonucleotide chemistries (modified bases, varied backbone chemistries) and higher-throughput QC environments;
• hybrid strategies combining HILIC separations with higher-resolution MS when ultra-low-level impurities or isobaric variants must be resolved.

Conclusion

HILIC coupled to a single-quadrupole LC-MS and specialized processing software can successfully separate, identify and quantify full-length synthetic oligonucleotides and truncated impurities without ion-pair reagents. The approach yields excellent linearity and reliable mass-based identifications, but currently shows lower MS sensitivity than ion-pair reverse-phase LC-MS and practical constraints with volatile bicarbonate buffers. Despite limitations at very low concentrations, this methodology offers a viable, instrument-friendly alternative for oligonucleotide impurity profiling in many routine analytical settings.

References

1. Takamine et al., Journal of the Mass Spectrometry Society of Japan, 2023, 71(2), 51–54.
2. Guimaraes et al., Journal of the American Society for Mass Spectrometry, 2023, 34.