Separation of Isomeric Metabolites and Gangliosides with High Performance (Drift Tube) Ion Mobility–Mass Spectrometry

The study of metabolites and gangliosides is increasingly important in drug discovery (1) and immunology (2). Accurate analysis of biologically relevant isomers is important because their structure affects their molecular properties. Typically, the isomers are separated using chromatography prior to mass spectrometry (MS) analysis. However, specialized chromatographic methods that distinguish isomers frequently require a complex setup and long runs. Techniques that allow accurate results to be acquired quickly and efficiently would be beneficial.

Ion Mobility Spectrometry Separations

Ion mobility spectrometry (IMS) offers a potential solution. It is faster than liquid chromatography (LC) and requires less effort because it separates molecules by their shape and size without needing complex sample preparation or setup protocols (3–12). Ion mobility filters for a set of isomers can be applied in a similar time frame to MS alone.

Several different IMS systems are available that can be coupled to mass spectrometers. Examples include drift tube ion mobility (DTIMS), (8) traveling wave IMS (TWIMS), (5,6) trapped ion mobility spectrometry (TIMS), differential mobility spectrometry (DMS) (13), and field asymmetric wave form IMS (FAIMS) (9,10).

Most IMS systems are combined with time-of-flight (TOF) mass analyzers. IMS-TOF techniques employ ion mobility at a reduced pressure to satisfy the vacuum regime of the TOF instrument, but the high resolving power of the IMS system is lost.

These hybrid quadrupole mass spectrometers and ion trap MS instruments offer an alternative. These analyzers feature ultrahigh resolution and high mass accuracy, but their longer acquisition times mean that they present challenges when coupling with IMS.

DMS and FAIMS are more compact IMS devices. Both forms of mobility analysis are based on nonlinear IMS and cannot be directly correlated to the collisional cross-section of a molecule (9), which limits their applicability for modeling molecular structures or analyzing conformation states because of the reliance on comparisons between collisional cross-sections.

Clearly, a range of IMS techniques can be coupled with mass analyzers to cover the majority of an analytical workflow. However, the missing component is a compact IMS device that does not sacrifice performance for size and can be coupled with a benchtop mass spectrometer for easy integration into the workflow.

High Performance IMS (HPIMS) Featuring Atmospheric Pressure Drift Tube

High performance ion mobility spectrometry (HPIMS) deploys atmospheric pressure short drift tube technology, enabling it to be coupled as a practical front-end to an MS device without breaking vacuum. High pressure is used to increase the number of ion collisions in a smaller space, boosting performance in a smaller footprint device. It offers a high degree of flexibility for separating isomers and other compounds (14).

Experimental

This study analyzed eight pairs of samples: a pair of gangliosides, GD1a and GD1b (Biovalley Nanterre), and seven metabolite pairs: corticosterone and 21-deoxycortisol; prednisolone and cortisone; dexamethasone and betamethasone; citric acid and isocitric acid; ursodeoxycholic acid and deoxycholic acid; and etiocholan-3-α-glucuronide and dihydrotestosterone-glucuronide (Sigma Aldrich). All samples were analyzed using an MA3100 Drift Tube High Performance IMS ion mobility spectrometer (Excellims) mounted on a Thermo Scientific Q Exactive Focus hybrid quadrupole Orbitrap MS instrument.

The MA3100 consisted of a 10.55 cm dual-gate linear drift tube operating at atmospheric pressure. It replaced the standard electrospray ionization (ESI) source on the front of most MS instruments (figure 1), enabling quick and easy transitioning between regular and IMS–MS use without breaking the vacuum. The ion mobility stage contained two Bradbury Nielsen-type ion gates that enabled operation in various IMS–MS modes.

Figure 1: Instrument layout details of the dual‐gate linear drift tube operated at atmospheric pressure prior to the mass spectrometer. The drift tube replaces the standard electrospray ionization (ESI) source on the front of the MS instrument.

Drift gas flow and temperature were under full software control. A manual mixing station consisting of two flow meters was used to apply different drift gases to study the potential improvement in mobility resolution.

Figure 2: Urso‐deoxycholic acid (UDCA) and deoxycholic acid (DCA) illustrating the separation of two bile acid isomers (391.2843 m/z), using 100% air as drift gas (individual standards and 50:50 mixture shown).

Figure 3: Two metabolite pairs separated in 100% air: (a) dihydrotestosterone‐glucuronide and etiocholan‐3‐α‐glucuronide ([M − H]⁻) and (b) prostaglandin A1 (only [M + Na]⁺) and prostaglandin B1 ([M + H]⁺ and [M + Na]+species shown).

A Directspray ESI source (Excellims) was used for ion generation. The ESI voltage ranged from 1400 V to 2400 V and optimized for each compound pair at a sample flow rate of 1.5 µL/min. The inlet gate of the IMS analyzer was pulsed open for 100 µs to start the ion mobility separation. The electric field applied across the mobility analyzer’s 10.55 cm drift region was set to 5530 V, resulting in a field strength of 524 V/cm. The drift gas temperature was controlled at 180 °C for all experiments, and the drift gas flow was set to 2.00–2.75 L/min, depending on the compound. The MS instrument was set to a mass range of 150–2000 m/z, with mass spectrometer resolution set to 35,000 or 70,000.

For each pair of samples, both instrument polarities were initially tried, with most compounds showing a clear preference for one polarity, typically yielding two to three orders of magnitude more signal than for the opposite polarity. The gangliosides ionized primarily in negative ion mode. For the metabolites, three pairs were found to ionize primarily in positive mode, and three others in negative mode. The prostaglandins showed good intensity in either ionization mode. For positive ion mode, both protonated and sodiated ions were initially considered, and the best species were chosen based on the intensity and separation. For negative ion mode, all experiments were carried out on the deprotonated [M − H]⁻ ion except for the gangliosides, which showed the highest peaks on the doubly deprotonated species [M − 2H]²⁻. Because drift times for the isomers were initially unknown, the MA3100 was run in scan mode with total acquisition times of 45–300 s, which were dictated by the length of drift time needed to see all ions of interest. This analysis time can be significantly shortened when setting only individual IMS windows to confirm or quantify isomers with previously determined drift times. For each sample, the extracted mass spectra (Thermo Xcalibur 4.0.27.19 Qual Browser) for the relevant masses were overlaid, and the retention time (tR) was converted to drift time. A 25 ppm sucrose solution was used for drift time calibration in both polarities to eliminate the influence of changes in atmospheric pressure. Resolving power was calculated as drift time divided by peak width (full width at half maximum).

Initially, all samples were analyzed with air as standard drift gas. If the separation with air was below 50%, carbon dioxide (CO₂) was attempted next, followed by mixtures of CO₂ and air using the flow meter apparatus described above, with a length of tubing for continuous flow blending. The percentage of CO₂ in the air was increased until maximum improvement was achieved. Percentages of gases were determined based on the readings of the flow meters. Residual CO₂ traces in the purified air were low enough to be ignored.

Results

Air as a Drift Gas

Half of the isomer pairs achieved sufficient separation (from 50% to baseline) using air as the drift gas (see figures 2, 3, and 4a).

Figure 4: Separation and resolving power for the ganglioside isomer pair (917.478 m/z) in (a) 100% air and (b) 100% CO₂, showing similar performance in both drift gases. No other metabolite pairs experienced any significant separation in air as a drift gas.

Polar Drift Gas

Changing to a more polar drift gas is expected to affect the drift separation of molecules that could be polarized (15). Because this may impact the mobility of two molecules differently, the net separation can be increased or decreased (or remain unchanged) when introducing a more polar drift gas, which is shown for corticosterone and 21-deoxycortisol in figure 5. Comparing the spectra in 100% air and in a 45:55 mixture of air and CO₂, the CO₂-rich drift gas shows longer drift times. More importantly, the minor separation of the two isomers in 100% air increased significantly in the gas mixture, most notably for the protonated species, but the effect is also visible for the sodiated species, which reflects a relatively stronger influence of CO₂ on the mobility of one isomer over the other.

Figure 5: [M + H]⁺ (347.220 m/z) and [M + Na]⁺ (369.202 m/z) species for corticosterone and 21‐deoxycortisol in 55:45 CO₂/air (a), showing significantly increased separation compared with 100% air (b). At 100% CO₂, the separated peaks had merged again (not shown).

The same approach was taken for the remaining isomer pairs. Prednisolone and cortisone increased in separation to 10% in the protonated species and 20% in the sodiated species for 25% CO₂. Dexamethasone and betamethasone could be separated 15% in their polymer state ([2M + Na]+, 807.403 m/z) with the optimum found at 50% CO₂. Although these pairs benefited from a mix with CO₂ 100% CO₂ yielded almost no benefit compared to air. The increased separation of the epimeric isomers by observing the sodiated dimer (as compared to protonated or sodiated monomer) has been shown for many biologically relevant pairs with a variety of drift gases, but until now has not been explored for gas mixtures.

Out of the eight isomer pairs analyzed by high-resolution atmospheric drift tube IMS, only the citric acid and isocitric acid pair remained unresolved, which is not surprising, given the minimal structural change between the two. A secondary peak was observed in CO₂, potentially pointing to different tautomers, but again with near identical drift times. The addition of a chemical modifier in the drift gas could be explored in the future (16).

Note that CO₂ mixtures were generally not tried if air alone gave good results. GD1a and GD1b were measured in 100% CO₂ to study the effect on resolution and separation. Although both molecules had a major change in drift time in CO₂, separation was only minimally improved (figure 4a). The resolving powers observed for the doubly charged gangliosides in CO₂ were 102 and 87, respectively, yielding a similar average resolving power to pure air.

Discussion

The results show that high performance ion mobility–orbital trap mass spectrometry (HPIM–OT-MS) is a powerful combination for the separation and analysis of biologically relevant isomers. The resolutions were generally found to exceed 70 and reach as high as 100, enabling the differentiation of comparatively small differences.

Although HPIM–OT-MS was capable of separating half of the isomers using air as drift gas, adding gas mixtures with CO₂ increased the number of at least partially separated isomer pairs to seven out of eight.

As neither pumping speed nor gas-specific pressure sensor calibration need to be considered, it is easy to switch between different drift gases, enabling optimization of compound separation based on polarity differences (17). The ability to adjust the gas mixture is more practical with HPIMS-orbital ion trap MS because of the ease of changing the drift gas. Other studies have shown biological epimeric isomers change their separation in different drift gases on a low-pressure DTIMS–MS instrument (18). For the epimeric isomer pair in this study, the increase in separation with gas mixing gives another dimension of the drift gas variable that is not necessarily available at low pressure, where changing the drift gas is a more involved process.

The HPIMS coupled with Thermo’s MS instrument offers other variables that might be useful in studying conformation. It has flexibility in temperature ranges and liquid gas modifiers, such as methanol (19) or propanol (20), can give further separation improvement. Alternatively, gas samples, solids, and swab samples can be introduced via an optional gas inlet or thermal desorber ion source.

Targeted Studies

In practice, HPIM–OT-MS lends itself to targeted studies, allowing the presence and relative abundance of multiple isomers to be determined through a quick series of mobility windows. Although complete 2D IMS–MS maps typically require acquisition times in minutes, individual confirmatory IM–MS scans can be acquired in a fraction of a second, thus making the application of several IM–MS windows per compound feasible.

Conclusion

HPIM–OT-MS is an interesting alternative for isomer separation compared to complex chromatographic methods or relying solely on MS2 fragmentation, especially for targeted applications. Once the drift time of potentially occurring isomers has been identified using standards, quick application of targeted drift time windows in gated mode enables different isomers to be distinguished.

Compared with multicolumn chromatographic setups, variables including drift gas composition, flow and temperature can be quickly optimized, giving flexibility and experimental convenience.

Employing ion mobility at atmospheric pressure without vacuum enables coupling to high resolution benchtop mass analyzers. The addition of the HPIMS to the mass analyzer, rather than the integration of IMS, allows existing MS users to expand their capability.

Drift gas variation also has potential in the analysis of other biomolecules, for example saccharide isomers. Combined with drift gas temperature variation and the addition of liquid modifiers, HPIMS on the orbital trap MS instrument offers a powerful addition to the toolbox in the investigation of a wide variety of samples (liquid, thermal desorption, or headspace [21]) and applications, including molecular characterization, isomer identification, analysis of protein conformation states, and characterization of complex mixtures like polymers and petrochemical samples (22).

Acknowledgments

The authors would like to thank Service de Pharmacologie et d’Immunoanalyse and Laboratoire d’Etude du Métabolisme des Médicaments for their contributions to this paper.

Original results were published in Rapid Communications in Mass Spectrometry.

References

1. C. Pereira Braga and J. Adamec, in Encyclopedia of Bioinformatics and Computational Biology (Elsevier, Amsterdam, The Netherlands, vol. 3, 2019).
2. I. Mocchetti, Cell Mol. Life Sci. 62, 2283–2294 (2005). https://pubmed.ncbi.nlm.nih.gov/16158191/
3. K. Michelmann, J.A. Silveira, M.E. Ridgeway, and M.A. Park, J. Am. Soc. Mass Spectrom. 26, 14–24 (2005).
4. M.E. Ridgeway, M. Lubeck, J. Jordens, M. Mann, and M.A. Park, Int. J. Mass Spectrom. 425(12), 22–35 (2018). https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6283298/
5. L.F.A. Santose, A.H. Iglesias, E.J. Pilau, A.F. Gomes, and G.C. Gozzo, J. Am. Soc. Mass Spectrom. 21, 2062–2069 (2010).
6. A.A. Shvartsburg and R.D. Smith, Anal. Chem. 80(24), 9689–9699 (2008). https://pubmed.ncbi.nlm.nih.gov/18986171
7. G.A. Eiceman and Z. Karpas, Ion Mobility Spectrometry (CRC Press, Boca Raton, FL, 2nd ed., 2005).
8. H.H. Hill Jr, W.F. Siems, R.H. St. Louis, and D.G. McMinn, Anal. Chem. 62(23), 1201A–1209 (1990). https://pubs.acs.org/doi/10.1021/ac00222a001
9. K.E. Swearingen and R.L. Moritz, Expert Rev. Proteomics 9(5), 505–517 (2012). https://pubmed.ncbi.nlm.nih.gov/23194268/
10. Y. Xuan, A.J. Creese, J.A. Horner, and H.J. Cooper, Rapid Commun. Mass Spectrom. 23, 963–1969 (2019).
11. S. Armenta, M. Guardia, A. Abad-Fuentes, A. Abad-Somovilla, and F.A. Esteve-Turrillas, Anal. Bioanal. Chem. 410(23), 5961-5967 (2018).
12. M. Hadener, M.Z. Kamrath, W. Weinmann, and M. Groessl, Anal Chem. 90(15), 8764–8768 (2018). https://pubs.acs.org/doi/abs/10.1021/acs.analchem.8b02180
13. C. Bylda, R. Thiele, U. Kobold, A. Bujotzek, and D.A. Volmer, Anal Chem. 87, 2121–2128 (2015).
14. J. Kaszycki, A. La Rotta, B. Colsch, F. Fenaille, C. Dauly, A. Kamleh, and C. Wu, Rapid Commun. Mass Spectrom. 33, 3–10 (2019). https://analyticalsciencejournals.onlinelibrary.wiley.com/doi/pdfdirect/10.1002/rcm.8414
15. R.T. Kurulugama, E. Darland, F. Khuhlmann, G. Stafford, and J. Fjeldsted, Analyst 140, 6834–6844 (2015).
16. F. Berthias, B. Maatoug, G.L. Glish, F. Moussa, and P. Maitre, J. Am. Soc. Mass Spectrom. 29, 752–760 (2018).
17. G.R. Asbury and H.H. Hill, Anal. Chem. 72, 580–584 (2018).
18. C.D. Chouinard, C.R. Beekman, R.H.J. Kemperman, H.M. King, and R.A. Yost, Int. J. Ion Mobil. Spectrom. 20, 31–39 (2017).
19. P.S. Chen, S.H. Che, J.H. Chen, et al, Anal. Chim. Acta. 946, 1–8 (2016).
20. P. Kwantwi-Barima, H. Ouyang, C.J. Hogan Jr., and B.H. Clowers, Anal. Chem. 89(22), 12416–12424 (2017). https://pubs.acs.org/doi/10.1021/acs.analchem.7b03518
21. N. Arroyo-Manzanare, A. Martin-Gomez, N. Jurado-Campos, R. Garrido-Delgado, C. Arce, and L. Arce, Food Chem. 578, 65–73 (2018).
22. A. La Rotta, J.K. Kaszycki, and C. Wu, “Separation of Saccharide Isomers Using High Performance Ion Mobility Spectrometry for an Orbitrap Mass Spectrometer,” ASMS Poster Presentation, TP 429 (2018).

Julia L. Kaszycki and Ching Wu are with Excellims. Claire Dauly and Anas Kamleh are with Thermo Fisher Scientific. Direct correspondence to: ching.wu@excellims.com.