Structural characterization of ligand-bound RNA hairpins on an Orbitrap Tribrid Mass Spectrometer

Significance of the topic

Oligonucleotide higher-order structure critically determines function, ligand recognition and therapeutic performance. Reliable, residue-level mapping of ligand binding sites and conformational stabilization is therefore essential for RNA-targeted drug discovery, quality control of therapeutic oligonucleotides and fundamental studies of RNA folding. Native top-down mass spectrometry (MSⁿ) that preserves noncovalent complexes offers a direct route to locate ligand interactions on intact RNA constructs without chemical derivatization or enzymatic digestion.

Objectives and study overview

This study developed and applied conformation-sensitive top-down MSⁿ workflows on Orbitrap-class mass spectrometers to characterize ligand binding to a 59-nucleotide hairpin RNA. The aims were to: (1) establish native MS conditions that preserve folded hairpin and RNA–dye complexes; (2) compare fragmentation strategies—low-q collision-induced dissociation (lqCID) and activated electron photodetachment combined with lqCID (aEPD–lqCID)—for generating informative product ions from low-charge, folded RNA; and (3) localize binding sites of two RNA-selective fluorescent dyes (SYTO RNASelect Red and Quant-iT RiboGreen) and evaluate ligand-induced conformational stabilization.

Methodology

Sample preparation and folded-state control:

• A 59-nt hairpin RNA containing four internal loops was buffer-exchanged into 50 mM ammonium acetate and adjusted to 2 μM. Annealing (95 °C for 5 min, slow cooling) was used to promote correct folding in the presence of MgCl₂ when required.
• Dyes (stock in DMSO) were mixed with RNA at molar ratios from 1:1 up to 1:20 and incubated 30 min at room temperature prior to analysis.

Mass spectrometry approaches:

• Native MS1 analyses used a Q Exactive UHMR Hybrid Quadrupole Orbitrap (nanoESI, spray voltage 1.2 kV, capillary 250 °C) at 200k resolving power to assess intact complexes and adduct distributions.
• Top-down MSⁿ experiments were performed on an Orbitrap Tribrid Apex equipped with a 213 nm UV laser. Activation and dissociation were executed in the linear ion trap; tandem spectra were recorded in the Orbitrap at 240k resolution. Quadrupole isolation windows of 2–5 Th were used; 50–150 transients were averaged per spectrum.
• Fragmentation modalities compared included lqCID alone and electron photodetachment followed by lqCID (aEPD–lqCID). Electron photodetachment produced radical precursor ions via 213 nm UV irradiation prior to supplemental collisional activation.
• Data analysis: MS1 deconvolution with BioPharma Finder and tandem fragment identification and mapping using OligoFinder with a 10 ppm mass tolerance.

Used instrumentation

• Thermo Scientific Q Exactive UHMR Hybrid Quadrupole Orbitrap mass spectrometer (native MS1).
• Thermo Scientific Orbitrap Tribrid Apex mass spectrometer equipped with a 213 nm UV laser for aEPD experiments and linear ion trap for lqCID/aEPD activation.
• Data analysis software: Thermo Scientific BioPharma Finder and OligoFinder.

Main results and discussion

Preservation of folded RNA and adducts:

• Native MS1 spectra showed the intact 59-mer with expected molecular weight after deconvolution and a distribution of metal adducts (1–8 Na+, K+ or Mg2+), consistent with folded and partially metal-bound states.
• Positive- and negative-mode analyses showed complementary charge distributions; negative mode produced higher absolute charge states, enabling selection of the −8 precursor at m/z ~2342 for top-down experiments.

Fragmentation behavior and method comparison:

• lqCID produced greater fragmentation overall than aEPD–lqCID under the tested conditions, likely because the low charge state and RNA size limit radical-driven fragmentation efficiency. However, combining aEPD with lqCID yielded complementary radical-derived d/w ions useful for localization when sequence coverage permitted.
• Fragment maps were generated with an intensity threshold; annealed versus non-annealed apoRNA showed differences in fragmentation efficiency, indicating conformation sensitivity of the top-down approach.

Ligand-binding modes and localization:

• Quant-iT RiboGreen complexes showed binding up to ~11 dye molecules per RNA at high ligand excess (1:20), but these complexes produced no sequence fragments in top-down MS, indicating strong structural stabilization. The lack of fragment ions is consistent with intercalative binding within the A-form stem that rigidifies base stacking and prevents backbone cleavage under the activation conditions.
• SYTO RNASelect Red bound distinctly: binding localized to the hairpin top loop rather than to internal loops or the 5′ bulge. Using optimized lqCID and aEPD–lqCID workflows and OligoFinder mapping, the study localized RNASelect Red to U31 in the 1:1 complex and to G28 and U31 in the 2:1 complex. Mapping results agreed with predicted 3D hairpin models (RNAComposer output), validating residue-level assignment by MSⁿ.

Benefits and practical applications of the method

• Residue-level localization of noncovalent ligands on intact RNA without crosslinking or digestion enables direct interrogation of binding sites for small molecules and dyes.
• Top-down native MSⁿ workflows provide a rapid orthogonal approach for confirming ligand binding modes (intercalation versus loop binding) and for assessing ligand-induced stabilization of secondary/tertiary structure.
• These approaches are directly applicable to QC in oligonucleotide therapeutics, screening of RNA-targeting small molecules, and mechanistic studies of RNA–ligand interactions where structural footprinting is required.

Future trends and potential applications

• Integration with ion mobility spectrometry to separate conformers prior to top-down fragmentation will increase sensitivity to conformational states and allow isolation of ligand-bound subpopulations.
• Higher-efficiency radical-based fragmentation methods and optimized charge-manipulation strategies could improve sequence coverage for larger RNA constructs and enable more routine residue-level mapping.
• Combining native MSⁿ with computational structure prediction and molecular modeling can refine binding-site assignments and support drug-design pipelines targeting structured RNA motifs.
• Automation of data-analysis workflows (improved scoring, false-discovery control for fragment assignment) will be important for adoption in regulated environments (QC/QA) and high-throughput screening.

Conclusions

The work demonstrates that native top-down MSⁿ on Orbitrap Tribrid platforms, using lqCID alone or combined with aEPD, can localize ligand binding on a folded 59-mer hairpin RNA and discriminate binding modes. Intercalative ligands (RiboGreen) markedly stabilized the stem and suppressed fragmentation, whereas loop-binding ligands (RNASelect Red) produced localized fragmentation patterns allowing assignment of binding residues (U31; G28+U31). These results validate conformation-sensitive top-down MS workflows as valuable tools for structural characterization of RNA–ligand interactions.

References

1. Hinkle J. D.; McGee W.; Salem J. P.; Mullen C.; Kronewitter S. R.; Yip P.; Syka J. E. P.; Stephenson J. L., Jr. Optimized Fragmentation of Oligonucleotides Suppresses Undesired Fragmentation Products and Enables Confident Sequence Assignment. ASMS Conference on Mass Spectrometry and Allied Topics; 2023; Houston, TX.
2. Purzycka K. J.; Popenda M.; Szachniuk M.; et al. Automated 3D RNA structure prediction using the RNAComposer method for riboswitches. Methods Enzymol. 2015;553:3-34.