RAPID COMMUNICATIONS IN MASS SPECTROMETRY Rapid Commun. Mass Spectrom.

2007; 21: 19511957 Published online in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/rcm.3041

Reducing fragmentation observed in the matrix-assisted laser desorption/ionization time-of-ight mass spectrometric analysis of triacylglycerols in vegetable oils
Jennifer Gidden1, Rohana Liyanage1, Bill Durham2 and Jackson O. Lay, Jr.1*
1 2

Arkansas Statewide Mass Spectrometry Facility, University of Arkansas, Fayetteville, AR 72701, USA Department of Chemistry, University of Arkansas, Fayetteville, AR 772701, USA

Received 16 March 2007; Accepted 25 April 2007

Edible oils consist primarily of triacylglycerols (TAGs). Matrix-assisted laser desorption/ionization time-of-ight (MALDI-TOF) mass spectra of the oils are typically dominated by sodium adducts of these TAGs but also show prominent fragment ions (that do not contain sodium), which can interfere with analytical measurements of other components in oils. The fragments seemingly correspond to the loss of a fatty acid moiety from the sodiated TAGs as a sodium salt: RCOONa. However, a previous study suggested that the fragments actually arise from nearly complete fragmentation of unseen protonated TAGs. These authors suggested that the fragmentation occurs so rapidly and completely that protonated TAGs are not normally observed in the spectra of these oils. In this paper, we present evidence to support their theory and also demonstrate an approach to eliminate these interfering ions from the MALDI-TOF mass spectra via addition of a base to the matrix/sample mixture. The added base does not impede formation of the sodiated TAGs, but does signicantly reduce the amount of fragments observed. We propose that this occurs by depleting the HR ions from the matrix, thus preventing the formation of signicant numbers of protonated TAGs in the rst place. For measurements by MALDI-TOF, the relative abundances of the fragment ions are related to the strength of the base, and can be almost completely eliminated. However, in longer time-scale experiments such as in post-source decay and Fourier transform mass spectrometry, sodiated and non-sodiated diacylglycerol (DAG)-like fragments are present in spectra, regardless of whether or not a base is added to the sample. Copyright # 2007 John Wiley & Sons, Ltd.

Edible oils consist primarily of triacylglycerols (TAGs) but they also contain a variety of important diacylglycerols (DAGs), free fatty acids, esters, phenols, and other compounds. Identication of the individual components in complex mixtures such as edible oils can be challenging. However, the analysis of these components is essential for addressing a number of important issues such as determining whether oils have been adulterated or contaminated, examining the various roles lipids play in biological processes, or determining the usefulness of oils as energy sources.17 Even though the most abundant components of these oils are the TAGs, their routine analysis often involves acid or base hydrolysis, which breaks them down into free fatty acids that are then converted into fatty acid methyl esters (FAMES), often for analysis by gas chromatography/mass spectrometry (GC/MS).810 These methods, and others like them, focus on the determination of the total fatty acid composition of the oil and differentiate oils by comparing measured fatty acid ratios compared to known standards. However, other studies have shown that it actually might be
*Correspondence to: J. O. Lay, Jr., Arkansas Statewide Mass Spectrometry Facility, University of Arkansas, Fayetteville, AR 72701, USA. E-mail: jlay@uark.edu

more benecial to measure the intact TAG proles of oils rather than their hydrolyzed fatty acids. For example, the nutritional behavior of lipids has been reported to be dependent on how the fatty acid components are grouped together as DAGs or TAGs rather than on the overall fatty acid composition.11,12 Adulteration of oil can be detected due to the sudden appearance of a unique TAG species or an increase in the amount of a usually minor TAG species and TAG proles are not dependent on regional variations like fatty acid proles.13 Degradation of oils can also be easily monitored by following the extent of the decomposition of TAGs.1416 In recent years, mass spectrometry (MS) has been shown to be a viable option for the rapid characterization of TAGs in oils, lipids, and tissues. This is especially true because it can be accomplished without lengthy separation or derivatization of the oils and yet readily provides abundant characteristic information. For example, matrix-assisted laser desorption/ionization time-of-ight (MALDI TOF) studies have been successful in qualitatively and quantitatively characterizing the various TAGs present in oils,1722 and

Copyright # 2007 John Wiley & Sons, Ltd.

1952 J. Gidden et al.

tandem mass spectrometry studies using electrospray ionization (ESI) have been used to provide structural information about TAGs, such as the number and position of double bonds in the fatty acid chains.2327 However, one problem with MALDI TOF (and ESI) analysis is that even though the ionization process is soft, fragmentation does occur during the measurement process. Although fragmentation may be useful in obtaining structural information, it can also lead to signicant problems in the analysis of the relative amounts of the various components in oils for determining levels of adulteration or following the decomposition of oils. The most typical fragmentation observed in MALDI-TOF mass spectra of oils is the cleavage of one of the fatty acid residues from a TAG, resulting in a DAG-like ion (minus an oxygen). However, only a few papers have addressed this particular issue. A MALDI-TOF study on phospholipids indicated that fragmentation was strongly dependent on the nature of the matrix and the amount of water present in the matrix solvent.28 The author proposed that the fragmentation was caused by gas-phase hydrolysis. A second study focused on determining the origin and mechanism of the fragmentation by investigating post-source decay (PSD) spectra of individual TAGs.20 In the TOF spectra, TAGs were present as sodium adducts (NaTAG) along with non-sodiated DAG-

like fragments that had m/z values equivalent to the loss of an RCOONa moiety from the sodiated TAG. A given NaTAG ion was then isolated and allowed to undergo post-source decay. In the PSD experiments, however, two DAG fragments, one corresponding to the loss of RCOONa and another to the loss of RCOOH, were observed in equal proportions in the resulting spectra. Because only the rst fragment was observed in the TOF spectra, the authors argued that it most likely did not come from the NaTAG ion but rather originated from an unstable, protonated TAG species (HTAG) that readily fell apart in the ion source. The NaTAG ions could also fragment into a similar species, but did so over a much longer timescale. Protonated TAG ions are rarely observed in MALDI-TOF spectra of oils, making it difcult to directly measure their role in the fragmentation process. However, it is possible to perform positive ion MALDI and yet create an environment where protonated TAGs are not likely to form in the rst place and thus eliminate that particular fragmentation pathway. In this paper, we report MALDI-MS analysis of vegetable oils in which a base, such as NaOH or NH3, was added to the sample/matrix mixture to prevent the formation of protonated TAG ions and the ensuing fragmentation.

901.7

a) No base
Na+TAGs DAG-like fragments
599.5

877.7

575.5

600

650 903.7

700

m/z

750

800

850

900

901.7

b) 1 M NaOH

c) 14 M NH3

877.7 877.7

600

650

700 m/z

750

800

850

900

600

650

700

m/z

750

800

850

900

Figure 1. MALDI-TOF mass spectra of soybean oil in which (a) no base is added during sample preparation, (b) 1 M NaOH is added to the oil/matrix mixture, and (c) 14 M NH3 is added to the oil/matrix mixture.
Copyright # 2007 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2007; 21: 19511957 DOI: 10.1002/rcm

MALDI-TOFMS analysis of TAGs in vegetable oils

1953

EXPERIMENTAL Materials and sample preparation

All vegetable oils (canola, corn, olive, peanut, sesame, soybean, and sunower) were purchased from a local supermarket and used without further purication or modication. HPLC grade hexane, methanol, and water were used as solvents. 2,5-Dihydroxybenzoic acid (DHB) was used as the matrix. Each oil was dissolved in hexane to a concentration of approximately 5 mg/mL. DHB was dissolved in 90% methanol to a concentration of 1 M. Various concentrations (from 0.025 to 14 M) of NaOH, NH3, NaC2H3O2, and NH4HCO3 were prepared in water. The oil, base, and DHB were then mixed together in a ratio of 2:1:2 and 1 mL of the nal mixture was spotted onto a stainless steel MALDI target and dried.

TOF spectrometer. In the PSD experiments, the precursor ion of interest is isolated using a timed ion gate (width of $10 Da) and the resulting fragment ions are refocused onto the detector by stepwise reduction of the voltage applied to the reectron. MALDI Fourier transform mass spectrometric (FTMS) analysis was performed on a 9.4 Tesla Ion Spec (IonSpec Corporation, Lake Forest, CA, USA) Fourier transform mass spectrometer.

RESULTS AND DISCUSSION

MALDI-TOF mass spectra of soybean oil in which various bases were added during sample preparation are shown in Fig. 1. When no base is added (Fig. 1(a)), the mass spectrum is dominated by two series of peaks near m/z 900 and another two series near m/z 600. These two groups of peaks are commonly observed in MALDI-TOF mass spectra of vegetable oils and other lipids.18,2022 The peaks near m/z 900, shown in an expanded region of the spectra in Fig. 2, correspond to sodiated triacylglycerol (NaTAG) ions. The identities of each NaTAG in soybean oil (and other vegetable oils) have been previously reported22 and are given in Table 1. The peaks near m/z 600 are non-sodiated diacylglycerol (DAG)-like fragment ions that appear to correspond to the loss of a fatty acid salt residue: RCOONa.

Instrumentation
MALDI-TOF mass spectra were obtained on a Bruker Ultraex II (Bruker Daltonic GMBH, Bremen, Germany) time-of-ight mass spectrometer operated in the positive-ion reectron mode. The accelerating voltage, delayed extraction time, and laser power were adjusted to optimize sensitivity and resolution for ions between m/z 5001500. Post-source decay (PSD) studies were performed on a Bruker Reex III

a) No base

901.7 903.7

905.7

877.7 879.7 899.7 907.7 881.7

880

890

m/z

900

910

b) 1 M NaOH

903.7 901.7

c) 14 M NH3

901.7 903.7

905.7

905.7 877.7 879.7

877.7 879.7 899.7 881.7 880 890 m/z 900 910 907.7

899.7 907.7

881.7

880

890

m/z

900

910

Figure 2. Expansion of the TAG region in the MALDI-TOF mass spectra shown in Fig. 1.
Copyright # 2007 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2007; 21: 19511957 DOI: 10.1002/rcm

1954 J. Gidden et al.

Table 1. Mass values for the TAGs and DAG fragments observed in the MALDI-TOF mass spectra of various vegetable oils
Composition Na TAGs LLLn, (LnLnO) LLL, (LLnO), (LnLnS) LLO, (LnOO) LOO, (LLS) OOO, (LnSS) OOS, (LSS) OSS LLnP LLP, (LnOP) LOP, (LnSP) OOP, (LSP) OPS NaDAG fragments (FTMS) LnLn LLn LL, (LnO) LO, (LnS) OO, (LS) HDAG fragments LnP LP OP PS LnLn LLn LL LO OO

a) MALDI-TOF

907.7

Mass (obs) 899.7 901.7 903.7 905.7 907.8 909.8 911.8 875.7 877.7 879.7 881.7 883.7 619.5 621.5 623.5 625.5 627.5 573.4 575.4 577.4 579.5 597.4 599.4 601.5 603.5 605.5

603.5

881.7

575.5

550

600

650

700 m/z

750

800

850

900

Na+TAG
907.7

b) PSD

RCOOH RCOONa 625.5

603.5

550

600

650

700

m/z

750

800

850

900

L linoleic acid (18:2), Ln linolenic acid (18:3), O oleic acid (18:1), P palmitic acid (16:0), S stearic acid (18:0). The fatty acids shown in parentheses are expected to be minor components, if present at all. DAG fragments means loss of RCOOX, from XTAG, X H or Na.

Figure 3. (a) MALDI-TOF mass spectrum of olive oil and (b) post-source decay (PSD) spectrum of the NaTAGs in olive oil with m/z 903911. The peaks at lower mass are shown at higher gain power.

901.7

a) 1 M NH4HCO3

However, further experiments indicate that the fragments may arise from a different source. Figure 3 shows MALDI-TOF and PSD spectra of an olive oil sample to which no base was added. As with the soybean oil, the MALDI-TOF spectra of olive oil (Fig. 3(a)) shows a series of sodiated TAG ions near m/z 900 and a series of non-sodiated DAG-like fragments near m/z 600. In the PSD experiment, the NaTAG ions from m/z 903911 were isolated and their corresponding fragments were monitored. The resulting spectrum (Fig. 3(b)) shows two series of fragment ions near m/z 600 with nearly equal intensities. The series at m/z 599605 have the same m/z as the DAG-like fragments observed in the MALDI-TOF spectrum and correspond to the loss of an oleic acid sodium salt residue (RCOONa) from the corresponding NaTAG ion. The series of fragments at m/z 621627, however, are not observed in the TOF spectra and are consistent with the loss of the oleic acid residue, RCOOH. This same fragmentation pattern was observed in a previous PSD study on individual TAG species by Al-Saad et al.20 A fragment equivalent to [NaTAG RCOONa] is observed in the MALDI-TOF spectra while fragments equivalent to [NaTAG RCOONa] and [NaTAG RCOOH], in equal abundance, are observed in the PSD spectra. The
Copyright # 2007 John Wiley & Sons, Ltd.

877.7

600 599.5

650

700

m/z

750

800

850

900

b) 0.025 M NH4HCO3

901.7 575.5

877.7

600

650

700 m/z

750

800

850

900

Figure 4. MALDI-TOF mass spectra of soybean oil in which (a) 1 M NH4HCO3 was added to the oil/matrix mixture and (b) 0.025 M NH4HCO3 was added to the oil/matrix mixture.
Rapid Commun. Mass Spectrom. 2007; 21: 19511957 DOI: 10.1002/rcm

MALDI-TOFMS analysis of TAGs in vegetable oils

621.486

1955

a) No base

Na+DAG fragments

903.740

Na+TAGs

DAG fragments
597.490

877.728

550

600

650

700

m/z

750

800

850

900

b) 14 M NH3

621.487

Na+DAG fragments Na+TAGs

903.743

DAG fragments
597.490

877.730

550

600

650

700

m/z

750

800

850

900

Figure 5. FTMS mass spectra of soybean oil in which (a) no base was added to the sample and (b) 14 M NH3 was added to the oil/matrix mixture.

authors suggested that because [NaTAG RCOONa] is clearly observed in the MALDI-TOF spectra while [NaTAG RCOOH] is not, the former fragment most likely originated from something other than the NaTAG species. Citing work by Siudzak, who proposed that the covalent nature of proton binding destabilizes molecular ions,29 the authors suggested that the DAG fragments observed in the MALDI-TOF spectra actually originated from unstable protonated TAG species (HTAG), that quickly fragmented in the ion source. Because Na binds electrostatically rather than covalently, the NaTAG species were believed to be more stable and only underwent slow fragmentation, as evidenced in the PSD spectra. A similar argument for slow fragmentation was put forward in our discussion of TAG analysis by FTMS.22 In the FTMS spectra, DAG-like fragments corresponding to [NaTAG RCOONa] and [NaTAG RCOOH] are observed, just as in the PSD spectra, although with different relative intensities. The difference between the TOF and FTMS spectra was attributed to the difference in timescale of the measurements. The TOF analysis is carried out on the ms
Copyright # 2007 John Wiley & Sons, Ltd.

timescale while the FTMS experiments occur on the ms timescale, allowing plenty of time for the NaTAG species to fragment. If the DAG fragments observed in the MALDI-TOF mass spectra do indeed arise from the rapid fragmentation of unstable HTAG species, the addition of a base to the matrix/analyte mixture should help prevent this fragmentation by scavenging most of the free H ions donated by the matrix and thus lowering the amount of HTAG ions formed in the rst place. Figures 1(b) and 1(c) show MALDI-TOF mass spectra of soybean oil with 1 M NaOH and 14 M NH3 added to the sample. In each case, NaTAG ions are present in the spectra, with the expected relative abundances (see Fig. 2). However, no fragmentation peaks are observed. Similar results were obtained for canola, corn, olive, peanut, sesame, and sunower oil with NaOH, NH3, NaC2H3O2, or NH4HCO3 added to the sample. The extent of fragmentation appears to be dependent on the base and its concentration. Mass spectra of oils with 1 M NaOH added to the sample show no fragmentation at all but the spectra with 1 M NH3 or 1 M NH4HCO3 show small
Rapid Commun. Mass Spectrom. 2007; 21: 19511957 DOI: 10.1002/rcm

1956 J. Gidden et al.

amounts of DAG fragments. When 1 M NaCl is added to the sample (data not shown), the spectra look the same as when no base is added, i.e. there are signicant amounts of DAG fragments present. Figure 4 shows MALDI-TOF mass spectra of soybean oil with two different concentrations of NH4HCO3 added to the sample. At 1 M, almost no fragmentation is observed in the spectrum, but when the concentration of the base is lowered to 0.025 M, the DAG fragments dominate the spectrum. Taken together, all these results are in line with an effect based on the relative strengths of the bases. OH is a very strong base and should easily accept protons from the DHB matrix. NH3 and HCO , on the other hand, are weaker bases 3 and do not accept protons as easily, allowing for more of the TAG species to be protonated. However, the addition of a more concentrated solution of these bases will lead to a signicant decrease in fragmentation. Cl has essentially no base properties at all and has no effect in preventing fragmentation, regardless of its concentration in the sample. While the addition of a base effectively eliminates fragmentation in the MALDI-TOF spectra of oils, DAG fragments are still observed in the FTMS and PSD experiments. Figure 5 shows FTMS mass spectra of soybean oil with and without 14 M NH3 added to the sample. Although the relative abundances of the NaTAG ions are slightly higher when NH3 is added, sodiated and nonsodiated DAG-like ions are present in both spectra. This is most likely due to the slow fragmentation of the NaTAG ions during the measurement process since the addition of a base clearly eliminates fragmentation in the shorter timescale TOF experiments. PSD data shows the same thing: sodiated and non-sodiated DAG-like fragments still appear in the spectra even when NH3 is added to the sample (which is reasonable because the fragments originate from the NaTAG ion and, as a result, should not be inuenced by the presence of a base). It is interesting to note that in the PSD spectra the two DAG-like fragment ions, while small, have nearly identical intensities. However, the NaDAG-like fragments are much larger than their non-sodiated counterparts in the FTMS spectra. The reason for this is not clear at this time but it may, again, be a result of the difference in time scales of the measurements. One nal application of the effect of bases on the mass spectra of oils is shown in Fig. 6. If MALDI-TOF is to be of any use in characterizing oils, it is important to obtain mass spectra that provide accurate representations of the relative amounts of TAGs and DAGs in the oils. Figure 6 shows MALDI-TOF mass spectra of EnovaTM, a mixture of soybean and canola oil that is labeled as containing >80% DAGs. When no base is added to the sample (Fig. 6(a)), the spectrum consists primarily of sodiated DAG ions (m/z 630650) but non-sodiated DAG-like fragments, equivalent to the loss of RCOOH from HTAG species, are also present. However, no intact TAG species (protonated or sodiated) are observed in the mass spectrum. When 1 M NaOH is added to the sample (Fig. 6(b)), the spectrum is still dominated by NaDAG species but non-sodiated DAG-like fragments are not observed at all and a small amount of NaTAG ions are present. The same result is obtained when 14 M NH3 is added to the sample. In this case, not only does the addition of a
Copyright # 2007 John Wiley & Sons, Ltd.

Na+DAGs
a) No base

639.5

641.5

637.5

643.5

DAG fragments

550

600

650

700 m/z

750

800

850

900

Na+DAGs
b) 1 M NaOH
LL 639.5

(LnS) LO 641.5

(LS) OO 643.5 LLn 637.5

Na+TAGs
550 600 650 700 m/z 750 800 850 900

Figure 6. MALDI-TOF mass spectra of EnovaTM in which (a) no base was added during sample preparation and (b) 1 M NaOH was added to the oil/matrix mixture. The composition of each NaDAG ion is given in the inset. See Table 1 for nomenclature. The ones shown in parentheses are expected to be very minor components, if present at all.

base prevent the fragmentation of TAGs, it allows for their presence to actually be detected and thus the spectrum in Fig. 6(b) more directly reects the reported composition of the oil than the spectrum in Fig. 6(a).

CONCLUSIONS
The addition of a base signicantly reduces the amount of prompt fragmentation observed in the MALDI-TOF spectra of vegetable oils. A previous study proposed that the fragmentation was due to unstable protonated TAG ions and the presence of a base appears to block that fragmentation pathway, most likely by preventing the formation of protonated TAG ions in the MALDI plume. The extent of the prevention of fragmentation is directly proportional to the strength of the base. Over longer time periods, sodium adducts of the TAGs will fragment into sodiated and non-sodiated DAG species, irrespective of the presence of a base.

REFERENCES
1. Woodbury SE, Evershed RP, Rossell JB, Grifth RE, Farnell P. Anal. Chem. 1995; 67: 2685. 2. Aparicio R, Aparicio-Ruiz R. J. Chromatogr. A 2000; 881: 93. 3. Lamberto M, Saitta M. J. Am. Oil Chem. Soc. 1995; 72: 867.
Rapid Commun. Mass Spectrom. 2007; 21: 19511957 DOI: 10.1002/rcm

MALDI-TOFMS analysis of TAGs in vegetable oils 4. Cert A, Moreda W, Perez-Camino MC. J. Chromatogr. A 2000; 881: 131. 5. Ostlund RE Jr, Racette SB, Stenson WF. Nutr. Rev. 2002; 60: 349. 6. Hunter JE. Am. J. Clinical Nutr. 1990; 51: 809. 7. Ramadhas AS, Jayaraj S, Muraleedharan C. Renewable Energy 2004; 29: 727. 8. Holcapek M, Jandera P, Fischer J. Crit. Rev. Anal. Chem. 2001; 31: 53. 9. Marriott PJ, Shellie R, Cornwell C. J. Chromatogr. A 2001; 936: 1. 10. Dodds ED, McCoy MR, Rea LD, Kennish JM. Lipids 2005; 40: 419. 11. Murase T, Mizuno T, Omachi T, Onizawa K, Komine Y, Kondo H, Hase T, Tokimitsu I. J. Lipid Res. 2001; 42: 372. 12. Mu H, Porsgaard T. Prog. Lipid Res. 2005; 44: 430. 13. Ulbert F, Buchgraber M. Eur. J. Lipis Sci. Technol. 2000; 102: 687. 14. Schiller J, Suss R, Petkovic M, Hanke G, Vogel A, Klaus A. Eur. J. Lipid Sci. Technol. 2002; 104: 496. 15. Schiller J, Suss R, Petkovic M, Klaus A. Eur. Food Res. Technol. 2002; 215: 282. 16. van den Berg JDJ, Vermist ND, Carlyle , Holcapek M, Boon JJ. J. Sep. Sci. 2004; 27: 181.

1957

17. Ayorinde FO, Eribo BE, Balan KV, Johnson JH Jr, Wan LH. Rapid Commun. Mass Spectrom. 1999; 13: 937. 18. Asbury GR, Al-Saad K, Siems WF, Hannan RM, Hill HH Jr. J. Am. Soc. Mass Spectrom. 1999; 10: 983. 19. Jakab A, Nagy K, Heberger K, Vekey K, Forgacs E. Rapid Commun. Mass Spectrom. 2002; 16: 2291. 20. Al-Saad KA, Zabrouskov V, Siems WF, Knowles NR, Hannan RM, Hill HH Jr. Rapid Commun. Mass Spectrom. 2003; 17: 87. 21. Calvano CD, Palmisano F, Zambonin CG. Rapid Commun. Mass Spectrom. 2005; 19: 1315. 22. Lay JO Jr, Liyange R, Durham B, Brooks J. Rapid Commun. Mass Spectrom. 2006; 20: 952. 23. Cheng C, Gross ML, Pittenauer E. Anal. Chem. 1998; 70: 4417. 24. Hsu F-F, Turk J. J. Am. Soc. Mass Spectrom. 1999; 10: 587. 25. Hsu F-F, Turk J. J. Am. Soc. Mass Spectrom. 1999; 10: 600. 26. Byrdwell WC, Neff WE. Rapid Commun. Mass Spectrom. 2002; 16: 300. 27. McAnoy AM, Wu CC, Murphy RC. J. Am. Soc. Mass Spectrom. 2005; 16: 1498. 28. Harvey DJ. J. Mass Spectrom. 1995; 30: 1333. 29. Siuzdak G. Mass Spectrometry for Biotechnology. Academic Press: San Diego, 1996; 1819.

Copyright # 2007 John Wiley & Sons, Ltd.

Rapid Commun. Mass Spectrom. 2007; 21: 19511957 DOI: 10.1002/rcm