J Am Oil Chem Soc (2014) 91:905–915 DOI 10.1007/s11746-014-2437-y
ORIGINAL PAPER
Optimization of Triacylglycerol-estolide Analysis by MatrixAssisted Laser Desorption/Ionization-Mass Spectrometry Haixia Zhang • Mark A. Smith • Randy W. Purves
Received: 25 July 2013 / Revised: 15 January 2014 / Accepted: 5 February 2014 / Published online: 5 March 2014 Ó Her Majesty the Queen in Right of Canada 2014
Abstract Acylglycerols containing more than three acyl groups (TAG-estolides) have been reported in plant seed oils and oil from ergot fungus. These TAG-estolides have considerable potential for industrial use, however, costs of producing synthetic TAG-estolides limits their use in largescale applications. Identification and structural characterization of additional natural sources of TAG-estolides has been restricted by their complexity and limitations of current analytical techniques. In this work, detection and characterization of TAG-estolides was optimized for use with MALDI-TOF-MS. Eight commonly used matrices were compared; 2,5-dihydroxybenzoic acid (DHB) and 2,4,6-trihydroxyacetophenone (THAP) gave good quality mass spectra. Matrix additives were examined and lithium was the most suitable, since MS/MS spectra of lithiated TAG-estolides provided the most informative fragmentation using an optimized method. The matrix solution pH was examined, and for THAP, replacing LiCl with 10–40 mM LiOH resulted in a slightly basic pH and significantly more intense TAG-estolide signals (up to eightfold higher). Since DHB is acidic, a larger amount of LiOH ([150 mM) was required for the matrix solution to Electronic supplementary material The online version of this article (doi:10.1007/s11746-014-2437-y) contains supplementary material, which is available to authorized users. H. Zhang M. A. Smith (&) R. W. Purves National Research Council of Canada, 110 Gymnasium Place, Saskatoon, SK S7N 0W9, Canada e-mail:
[email protected] R. W. Purves e-mail:
[email protected] R. W. Purves Department of Chemistry, University of Saskatchewan, 110 Science Place, Saskatoon, SK S7N5C9, Canada
become basic, leading to ion suppression and reduced signal intensities. Thus, for TAG-estolide analysis, THAP with *20 to 30 mM LiOH gives the highest quality spectra and the most informative MS/MS fragmentation. Keywords Triacylglycerol Estolide Ergot oil MALDI-MS DHB THAP Abbreviations AG Acylglycerol(s) CHCA a-Cyano-4-hydroxycinnamic acid DHB 2,5-Dihydroxybenzoic acid ESI Electrospray ionization FA Fatty acid(s) O Oleic acid (octadec-9-cis-enoic acid, 18:1) M Molecular ion MALDI Matrix-assisted laser desorption/ionization-mass spectrometry P Palmitic acid (hexadecanoic acid, 16:0) R Ricinoleic acid (12-hydroxyoctadec-9-cis-enoic acid, 18:1-OH) S/N Signal-to-noise ratio TAG Triacylglycerol(s) THAP 2,4,6-Trihydroxyacetophenone
Introduction The predominant storage lipids of plants, animals, and fungi are triacylglycerols (TAG), consisting of three fatty acids esterified to glycerol. In organisms that produce TAG containing fatty acids with hydroxyl- groups, secondary acylation, where additional fatty acids are esterified to the
123
906
Fig. 1 Example of a 5-acylglycerol (5-AG) TAG-estolide [RP?O? RP]
hydroxyl moieties, can result in the formation of TAGestolides. Figure 1 shows an example of a 5-acylglycerol (5-AG) TAG-estolide [RP?O?RP]. These unusual acylglycerols (AG), containing more than three fatty acids per glycerol, are abundant in the seed oil of certain species in the genus Lesquerella (now Physaria) [1], and are known components of oil from plants such as Trewia nudiflora L. [2] and Cardamine impatiens L. [3], and oil from the sclerotia of ergot fungus Claviceps purpurea (ergot oil) [4, 5]. The pathways of biosynthesis of natural TAG-estolides are unknown. Synthetic TAG-estolides have novel functionality compared with TAG-based products for use in applications such as biodegradable lubricants and functional fluids [6–8]. Due to their attractive potential for industrial applications, but high cost of production, there is considerable interest in finding a cost effective renewable source of these products. Potential routes include the identification of additional TAG-estolide rich oils, or the characterization of TAG-estolide biosynthesis and subsequent genetic engineering of estolide production in an existing oilseed crop. Detection and characterization of TAG-estolides is limited by the diversity of their fatty acid components, by their structural complexity, and by the existence of multiple isomers and isobaric molecular species. Furthermore, in many cases, only small amounts of material are available for investigation. Development of a rapid, simple, and robust method for TAG-estolide identification and characterization using a small amount of oil, ideally without time-consuming pre-separation, would greatly benefit their analysis. Various techniques have been reported for TAG-estolide analysis: these include separation by column chromatography [5], TLC [1], HPLC [6], GC [7], and supercritical fluid chromatography [9]; estolide linkage detection using NMR [1], refractive index [10], infrared [11], atmosphericpressure chemical ionization (APCI) [12], electrospray ionization-mass spectrometry (ESI–MS) [13], and matrixassisted laser desorption/ionization-mass spectrometry (MALDI-MS) [14, 15]. The use of mass spectrometry
123
J Am Oil Chem Soc (2014) 91:905–915
[16–18] offers advantages of high mass resolution, excellent detection sensitivity, and very low sample consumption, as low as 100 fmol of trilaurin has been detected with MALDI-MS [19]. TAG-estolides have similar chemical structures to TAG, and the application of mass spectrometry to the detection of TAG and phospholipids from extracts such as yeast [20], plants [21], or animal tissues [22] is a well established technique. However, the detection of TAG-estolides using MS has received little attention, with the exception of a report by Lin and co-workers [13] who used LC–ESI–MS to detect TAG-estolides present as a minor component of castor oil. We previously reported the identification and characterization of abundant TAGestolides from plant and ergot oils using MALDI-MS [14, 15]. Compared with ESI–MS, MALDI-MS has the advantages of a short analysis time, better salt/detergent tolerance, and the ability to re-examine the spot containing the analyte when required. Matrix selection is a critical step during MALDI-MS analysis. a-cyano-4-hydroxycinnamic acid (CHCA) [23, 24] or 2,5-dihydroxybenzoic acid (DHB) [25, 26] were reported by several groups for TAG analysis of different oils. It was reported that the protonated TAG molecular ions were thermally unstable during the MALDI process [27], and sodiated ions dominated the mass spectrum. Thus, the addition of alkali ions during matrix preparation is thought to stabilize TAG ions through the formation of alkali adducts. Asbury et al. [19] examined different MALDI matrices for TAG analysis, including DHB, CHCA, dithranol, and K4[Fe(CN)6]. They reported that DHB and CHCA in the presence of TFA have poor shot-toshot reproducibility; dithranol had very low sensitivity, whereas the use of K4[Fe(CN)6] matrix was not successful. We have also reported the analysis of TAG-estolides from fungal and plant oils using THAP [14] or DHB [15] with sodium chloride. Our results showed that both matrices had good performance for TAG-estolide analysis, with sodiated ions being the only TAG-estolide signals in the mass spectra, and informative MS/MS spectra were obtained for analyte determination. Besides conventional chemical matrices, pencil lead also has been reported for small molecule and peptide analysis [28, 29], as it showed less matrix interference compared with other chemical MALDI matrices. Pencil lead is made from graphite and clay with multiple metal ions present in it. Recently this matrix was employed for free fatty acid and TAG detection; sodiated ions were found to be dominant, and only minor potassiated adducts were observed [30, 31]. Compared with other chemical matrices, pencil lead had the cleanest matrix background at low m/z values (m/z \ 700), and offered higher signal intensity compared with DHB. Besides sodium salts, other alkali ions such as lithium and potassium also have been examined for MALDI-MS
J Am Oil Chem Soc (2014) 91:905–915
analysis of TAG standards. Allmaier and co-workers [32] reported the use of 2,4,6-trihydroxyacetophenone (THAP) as a matrix for TAG detection, with the addition of LiCl, NaCl, or KCl salts to prevent thermal degradation. They concluded that sodiated TAG exhibited the most informative fragmentation across the entire m/z range; lithiated TAG showed incomplete fragmentation in the low and high m/z region; and potassiated ions did not provide useful fragmentation. Besides TAG analysis, lithium salts are frequently used for MALDI imaging of tissue lipids or phospholipids [33]. For ESI–MS analysis, Cheng and Gross compared ammonium, lithium, sodium, and cesium salts for TAG structural analysis with a four-sector tandem ESI– MS [34]. They reported that alkali metal-cationized TAG fragment ions provided similar structural information and these ions were more informative compared with ammonium clustered ions. In contrast, Turk and coworkers [35] compared ammonium, lithium and sodium ions for TAG analysis using a linear ion-trap ESI–MS, and reported that lithiated TAG ions undergo complete fragmentation compared with sodiated or ammoniated TAG ions. As mentioned previously, prior reports concluded that protonated TAG ions are thermally labile during MALDIMS analysis [27]. TAG were detected exclusively as alkali adducts, especially sodiated ions; whereas protonated TAG ions were not detected. It was proposed that during ionization, protonated TAG ions experienced fragmentation very rapidly and completely resulting in very few being transferred into the gas phase. This matches the concept of charge delocalization proposed by Siuzdak [36]. In short, due to the covalent nature of proton binding, the protonated TAG ions can be destabilized leading to fragmentation [37]. Therefore the addition of a proton donor, such as an acid, to the matrix or sample solution should be avoided [27, 32]. Despite this, Addeo and coworkers [25] reported the use of a DHB matrix containing 0.1 % TFA to detect TAG, which were cationized by sodium chloride solution prior to MALDI-MS analysis. Sodiated TAG were detected as the dominant ions, however, the effect of TFA on TAG signal intensity was not discussed. Gidden et al. [27] investigated the effect of the MALDI matrix pH for TAG detection, by adding different concentrations of base to the matrix solution, such as sodium hydroxide (up to 200 mM) or ammonium bicarbonate. The authors claim that thermal degradation of TAG can be significantly reduced by preventing the formation of protonated TAG ions [27]. Although various MALDI-MS studies involving TAG analysis have been reported over the last decade, there has been no systematic matrix optimization study reported for TAG-estolide analysis. Consequently, the objective of this work is to systematically optimize different matrices and their additives for rapid TAG-estolide analysis.
907
Experimental Section Materials CHCA was purchased from Waters (Milford, MA). DHB was from Amersham Pharmacia Biotech (Uppsala, Sweden); sinapinic acid, THAP, dithranol, super-DHB, and 3-hydroxypicolinic acid (HPA) were obtained from SigmaAldrich (Milwaukee, WI). Pencil lead was purchased from the local store (General’s, 6B); potassium chloride, lithium hydroxide, acetonitrile (ACN) and chloroform were from Fisher Scientific (Fair town, NJ), whereas sodium chloride, lithium chloride and hexane were from EMD Chemicals (Gibbstown, NJ). Samples of ergot sclerotia (host is rye) were collected locally. The pH paper was from Micro Essential Laboratory (pH range 1–14, NY, USA). Lipid Extraction Ergot sclerotia were ground to a fine powder, and hexane was added to extract the lipids with further grinding for 30 s. The contents were transferred to a Pyrex tube and centrifuged for 3 min (2,000g). The hexane phase was removed to a clean tube and evaporated under a stream of nitrogen at room temperature. Samples were reconstituted in chloroform to a 1 lg/lL concentration before use. MALDI-MS for TAG-estolide Samples The MALDI matrices, DHB, CHCA, super-DHB, THAP, HPA, dithranol, and sinapinic acid were prepared at 20 mg/ mL as follows: sinapinic acid, CHCA, THAP and HPA were prepared in 1:1 ACN:water; DHB and super-DHB were each prepared in 1:1 water:methanol; and dithranol was in 1:1 chloroform:methanol. The matrix solution was then mixed 1:1 with each alkali salt solution to achieve the desired salt concentration. A volume of 0.8 lL of matrix solution was spotted onto the MALDI plate followed by 1.0 lL of sample deposition. Pencil lead was also selected as a matrix and it was gently applied to the MALDI plate without pretreatment of the plate surface. Afterwards the oil sample was directly spotted on it. An AB 4800 MALDI-TOF/TOF mass spectrometer (Applied Biosystems, Frederick, MD, USA) equipped with a Nd:YAG laser (355 nm wavelength) was used in the positive ion reflectron mode to acquire data. The laser pulse rate was 200 Hz with a duration of \500 ps per pulse. The maximum laser output is 20 lJ and the output was adjusted using the software. Instrument default calibrations in both MS and MS/MS modes were updated using a standard peptide mixture [14]. Both MS and MS/MS spectra were automatically collected by the software as an average of 800 laser shots, with laser irradiated randomly
123
908
J Am Oil Chem Soc (2014) 91:905–915
on the surface of the spot. In MS/MS mode, air was the collision gas and data were collected using a 2 kV MS/MS positive ion method, unless specified otherwise.
Results and Discussion Matrix Optimization for TAG-estolide Detection To gain a thorough understanding of the effect of different matrices on TAG-estolide detection, eight commonly used matrices were selected for comparison. These matrices are: CHCA, DHB, super-DHB, THAP, HPA, dithranol, sinapinic acid, and pencil lead. All matrices (except pencil lead) were prepared in 50 mM NaCl with appropriate solvent described in the experimental section to ensure that sodiated adducts were the major species detected from the sample. Since there is currently no commercial TAG-estolide standard available, an oil sample extracted from ergot sclerotia was selected here for matrix optimization. This oil, containing major TAG-estolides ranging from 4-AG to 6-AG, has been studied previously using THAP matrix [14]. Figure 2 shows mass spectra of ergot oil obtained from CHCA, pencil lead, DHB, and THAP, separately. The other four matrices, dithranol, super-DHB, HPA, and sinapinic acid only showed very weak signals, even with high energy laser irradiation, and therefore the results from those matrices are not shown. Signals from blank matrices were also collected to determine matrix interference peaks (data not shown). For all four matrices shown in Fig. 2, TAGestolides were detected in the form of alkali adduct peaks. The use of CHCA matrix resulted in TAG-estolide signals with low-energy laser irradiation (Fig. 2a, laser 9.1 lJ), however, the CHCA matrix itself showed intense matrixsodium cluster signals in the same mass range (m/z 500–2,000) as the TAG and TAG-estolides (peaks labeled with * in Fig. 2a). Although these matrix signals can be predicted with a program released from Li and coworkers [38], interpretation of the results is more challenging and time consuming, and the number of overlapping isobaric species is increased. In addition, CHCAsodium cluster ions dominate the mass spectrum and thus significantly suppress TAG-estolide signals. Therefore, CHCA was considered unsuitable for TAG-estolide analysis. For TAG-estolide analysis with pencil lead, shown in Fig. 2b, both potassium and sodium adduct species were observed with relatively strong potassium adduct peaks compared with the sodiated peaks. The formation of two different metal adducts from one target TAG species can be acceptable for a pure sample analysis, however in this study, the oil sample contains a complex mixture of TAGestolides having a variety of acyl chains and different
123
Fig. 2 Mass spectra obtained by MALDI-MS for 1 lg of ergot sclerotia oil using a matrix of: a CHCA; b pencil lead; c DHB; and d THAP. In all cases, the matrices were prepared in the presence of 50 mM NaCl, and each spectrum represents 800 laser shots. Note in a, asterisk indicates CHCA-alkali cluster peaks
degrees of unsaturation. Consequently, the detection of both potassium and sodium adducts greatly complicates data interpretation. Although the addition of a sodium salt layer can effectively reduce the formation of potassium adducts for common TAG (data not shown), with the addition of the sodium salt layer in our study, potassium adduct signals were still detectable for TAG-estolides in the mass spectra (data not shown). In addition, Fig. 2b shows that the high mass peaks (m/z [ 1,600) are virtually absent; increasing the laser energy had only a minor effect on improving their signal intensities, but resulted in the loss of the low mass signals (m/z \ 1,500). Thus, pencil lead was excluded from further study. For both DHB- and THAP-assisted spectra (Fig. 2c and 2d) sodiated ions are the dominant peaks. Matrix interference from both matrices was primarily below m/z 600,
J Am Oil Chem Soc (2014) 91:905–915
909
which is below the mass region of TAG-estolides (typically over m/z 900), thus matrix interference is not considered problematic. Since DHB and THAP require similar laser energy (laser was 12.7 lJ for THAP and 11.4 lJ for DHB), and both provide good quality mass spectra for ergot TAGestolides, these two matrices were selected for further comparison and optimization. Effect of Cations on MS and MS/MS Spectra of Ergot Estolides To understand the effect of different cations on MALDIMS of TAG-estolides, we selected LiCl, NaCl, KCl, and ammonium acetate as matrix modifiers for use with DHB or THAP, with each salt having a final concentration of 50 mM. Mass spectra collected using DHB matrix are shown in Fig. 3. Spectra collected using lithium (Fig. 3a) and potassium (Fig. 3c) as additives had a similar pattern to that with sodium (Fig. 3b). TAG-estolides were detected as the corresponding alkali adducts, [M?alkali]? (where the alkali is Li, Na and K, separately). In the presence of ammonium acetate (Fig. 3d), TAG-estolides were detected as weak sodiated ions rather than as ammonium-TAG-estolide adducts; consistent with findings reported by other groups [27, 39]. Ammonium adducts were not observed even when the ammonium concentration was increased to 1 M. Potassiated ions offered the highest signal-to-noise ratio (S/N) based on four of the intense TAG-estolide peaks selected (m/z 1,150, 1,176, 1,432 and 1,458 for potassiated ions), and signals for sodiated and lithiated ions were around 16 and 36 % less compared with the potassiated ions. Peaks acquired using an ammonium-mediated matrix exhibited the lowest signal intensity among all the matrix modifiers examined (S/N of the four selected peaks was around 6 % compared with potassium-mediated signals). In THAP-assisted mass spectra (data not shown), the peak profiles had a similar pattern to those collected with the DHB matrix. That is, TAG-estolide-alkali adducts were detected when lithium, sodium [14], or potassium was used; whereas weak sodiated adducts were observed when ammonium acetate was present in the matrix. The signals from different cation-mediated matrices also follows the trend observed with the DHB matrix, that is, potassiated TAG-estolides have the highest signal intensity, followed by sodium, lithium, and ammonium acetate. The MS/MS spectra of lithium-, sodium- and potassiumTAG-estolide ions in DHB or THAP were compared to determine which combination gave the best fragmentation data for TAG-estolide structural analysis. A 2 kV MS/MS method with CID gas (air) was used, and an abundant ergot estolide with structural composition [RP?O?RP], was selected for comparison. The corresponding MS/MS
Fig. 3 Mass spectra obtained using MALDI-MS for a sample of 1 lg of ergot sclerotia oil using DHB matrix in the presence of 50 mM of: a LiCl; b NaCl; c KCl; and d NH4Ac. Data were collected using 800 laser shots
spectra, collected using DHB matrix with different alkali adducts, are shown in Fig. 4a–c. Note that precursors [M?alkali]? are m/z 1,400.2, 1,416.2 and 1,432.2 for lithiated, sodiated and potassiated ions, respectively. Lithium adducts (Fig. 4a) give the most extensive fragmentation; the loss of a free fatty acid [M?Li–P]?, [M?Li–O]?, loss of multiple free fatty acids [M?Li–P-O]? (m/z 861.8), [M?Li–P–P–O]? (m/z 605.5), and the observation of free fatty acid lithiated ions such as [O?Li]? (m/z 289.3) and [P?Li]? (m/z 263.3) were recorded. In addition, the loss of estolides such as [M?Li-RP]? (m/z 863.8) and [M?Li-RPO]? (m/z 581.5), and the observation of lithiated estolide ions [RP?Li]? (m/z 534.5) and [RO?Li]? (m/z 569.5) were diagnostic of the estolide bond formation between
123
910
Fig. 4 Fragment ions of TAG-estolide [RP?O?RP], the structure of which was shown in Fig. 1, using a 2 kV MS/MS method with DHB matrix in the presence of 50 mM of: a LiCl, b NaCl, c KCl; and d LiCl. CID gas was on for (a–c) and was switched off for (d). Data were collected using 800 laser shots
two fatty acids. The observation of the [M?Li-RO]? (m/ z 837.8) and [RO?Li]? (m/z 569.5) peaks indicated that an isobaric species, [RP?P?RO], was present in the oil samples. Note that isobaric TAG positional isomers, have been separated on-line using HPLC containing either a C18 [12] or chiral [40] column coupled to an APCI source prior to MS analysis. A similar approach of using up-front separation (either on-line or off-line) to further simplify MALDI-MS spectra is currently under investigation. Interestingly, a strong signal at m/z 287.2 was observed in Fig. 4a. A proposed scenario for the formation of this ion is that, during collision-induced dissociation, the precursor ion loses an estolide-linked fatty acid as a free fatty
123
J Am Oil Chem Soc (2014) 91:905–915
acid (Fig. 5). This is mediated by a molecular re-arrangement involving the formation of an intermediate 6-element ring, as proposed by Al-Saad and coworkers [37] for TAG fragmentation, thus replacing the hydroxyl group of the ricinoleic acid (18:1-OH) with a double bond to form a diunsaturated fatty acid (18:2). Subsequent loss by rearrangement of this fatty acid from the glycerol backbone generates a lithiated ion with m/z 287 (note that in both steps shown in Fig. 5, the lithium ion can reside on either product and for simplicity, the corresponding neutral species were not shown). As ricinoleic acid is the only hydroxyl fatty acid in ergot estolides, the 287 peak was observed in all of the lithium-mediated MS/MS spectra examined. The MS/MS spectrum from sodiated TAG-estolides (Fig. 4b) also demonstrated the loss of fatty acids or estolides, such as [M?Na–P]? (m/z 1159.9), [M?Na–O]? (m/z 1133.9), [M?Na–P-O]? (m/z 877.7), [M?Na-RP]? (m/z 879.7) and [M?Na-RO]? (m/z 853.7), as well as the formation of sodiated estolides such as [RP?Na]? (m/z 559.4) and [RO?Na]? (m/z 585.4). These fragment ions offered sufficient information to derive the composition of the major TAG-estolide. The putative 18:2 fatty acid generated during fragmentation (as mentioned above, sodiated ion m/z 303) was also detected, but with much lower relative signal intensity. Compared with the lithiated fragment ions, the sodiated free fatty acid peaks [FA?Na]? were not observed, and some fragments resulting from the loss of multiple fatty acids were also absent. For potassiated TAG-estolide precursors (Fig. 4c), strong signals from the free potassium ion (m/z 39) dominated the MS/MS spectrum, whereas extremely weak TAG-estolide fragment ions were detected (high m/z region) as is shown in the inset. Similar MS/MS spectra for different TAG-estolide alkali adducts were observed for THAP matrix (data not shown). Since lithiated precursors provided the most informative fragmentation, the aforementioned lithiated TAG-estolide ([RP?O?RP?Li]?, m/z 1400.2), was selected to compare two MS/MS acquisition methods, 1 kV vs. 2 kV MS/MS methods for DHB and THAP matrices, with CID gas being turned on or off. For both matrices, all acquisition methods provided a similar pattern of estolide fragment ions (m/ z [ 250, see Fig. 4a). The intensities of the lower mass fragment ions (m/z \ 800) were relative weak when the CID gas was turned off (post-source decay mode, Fig. 4d, 2 kV MS/MS with CID off). This observation is consistent with our previous study in which a sodium-mediated THAP matrix [15] was used for TAG-estolide analysis with the CID gas off. When the collision gas was switched on (either 1 kV or 2 kV method), intense low mass fragment ions (m/z \ 200, Fig. 4a) were detected. Most of these low mass peaks have a m/z difference of 14 (difference of a CH2- unit), which indicates the fragmentation of the fatty
J Am Oil Chem Soc (2014) 91:905–915
911
Fig. 5 Proposed fragmentation pathway for producing an m/z 287 fragment from a lithiated TAG-estolide during collision-induced-dissociation (CID)
acid alkyl chain. In general, the 2 kV MS/MS method provided higher fragment ion signals compared with the 1 kV MS/MS method (up to twofold based on base peak S/N); and for the same acquisition method (same voltage), the S/N of the higher mass fragment ions (m/z [ 800) were almost doubled when the CID gas was turned off (Fig. 4d). Both matrices could provide similar MS/MS signal intensity by adjusting the laser energy, and similar to MS mode, in the MS/MS mode the laser energy required for THAP was around 1.2 lJ higher than DHB. The m/z 287 peak (lithiated 18:2 fatty acid), generated during TAG-estolide fragmentation (Fig. 5) as discussed in the previous section, was detected in all cases (CID gas on or off), with a relatively higher signal intensity when the CID gas was turned on. Based on this comparison, although lithiated TAG-estolide ions had relatively lower S/N compared with sodium or potassium clustered ions, the MS/MS data provided the most comprehensive information, especially when a 2 kV MS/MS method was used with the CID gas turned on. For
the subsequent study, lithium salt was selected for further matrix optimization with DHB and THAP. Effect of Salt Concentration on Ergot TAG-estolide Signals Different concentrations of alkali salts (Na?, Li?, or K? ions) have been reported for use in TAG analysis ranging from 5 mM to 500 mM [14, 39–41]. We therefore conducted a thorough study to examine whether LiCl concentration has an effect on TAG-estolide detection. A series of aqueous LiCl solutions were prepared, and mixed 1:1 with the DHB matrix solution to give final salt concentrations of 500, 100, 50, 25, 10, or 2 mM, respectively. Mass spectra were collected under the same laser energy irradiation for each matrix. Results demonstrated that strong signals were observed with a LiCl concentration between 10 to 50 mM (data not shown). When the lithium concentration was increased to over 100 mM, TAG-estolide signals decreased rapidly (S/N reduced to 64 % with
123
912
J Am Oil Chem Soc (2014) 91:905–915
100 mM LiCl, and 37 % with 500 mM LiCl, compared with signals collected in the presence of 10 mM LiCl), which is likely due to an ion suppression effect at high salt concentration. The effect of lithium chloride on THAP matrix is similar (data not shown), with the most intense signals being detected with a lithium concentration between 2 to 50 mM; 25 mM LiCl provided the highest signal intensity (up to 1.4-fold compared with signals collected from 2, 10 and 50 mM LiCl-mediated ones). When the lithium concentration was increased to over 100 mM, the signals decreased up to 65 % compared with signals collected at 25 mM LiCl. Thus, 25 mM LiCl is the optimized concentration for both matrices. Effect of Matrix pH The effect of acid (TFA) and base (LiOH) during matrix preparation on ergot TAG-estolide detection was investigated for both DHB and THAP. For ease in comparison, all mass spectra were collected using the same laser energy. A mass spectrum collected using a DHB matrix with 0.1 % TFA is shown in Figure S1a. Compared with spectra collected using an LiCl-mediated DHB matrix (Fig. 2c), the addition of TFA alone into the matrix complicated the mass spectrum (Figure S1a), as both sodiated and potassiated TAG-estolide ions were detected, whereas no protonated species were observed (overall signal intensities were reduced around 37 % compared with Fig. 2c). The addition of TFA to the LiCl-containing DHB matrix was also tested (data not shown), and in this case only lithiated TAG-estolide ions were detected. In addition, the overall lithiated TAG-estolide signals were similar to those collected without TFA addition (Fig. 2c), thereby indicating that lithium plays the major role during the ionization process. To determine the effect of base, solutions of LiOH were prepared ranging from 2 mM to 1 M and a mass spectrum acquired with the addition of 10 mM lithium hydroxide to the DHB matrix is shown in Figure S1b. Figure 6 shows a plot of the S/N of the base peak as a function of log[LiOH]; typical %RSD values were *15 %, however, for low S/N values, the %RSD was much larger (up to 45 % at a S/N of 33 for 1,000 mM LiOH) Note that for comparison, the S/N of spectra collected using the 25 mM LiCl-mediated matrix was also added to the plot (data point for LiCl is indicated on the figure). Note that no LiOH was present in the LiCl, but because of the log scale, for ease of comparison, the point was placed at a value of [LiOH] = 1 mM. The figure illustrates that at a low base concentration range ([LiOH] B100 mM), the S/N was reduced compared with the 25 mM LiCl-mediated matrix. However, by increasing the LiOH concentration to 200–400 mM, the S/N of the base peak increased compared with the lower concentrations,
123
Fig. 6 Relationship of base peak signal-to-noise (S/N), and pH of the DHB solution, as a function of log[LiOH]. For comparison, points representing the S/N and pH using a 25 mM LiCl-mediated matrix were added to the plot. Note that S/N is the average of six replicates
with overall signals of around 55 % compared with signals collected using 25 mM LiCl. This signal improvement with the addition of base is consistent with previous work [25]. Increasing LiOH concentration to over 500 mM resulted in a further decrease of the TAG-estolides signals. Changes to the TAG-estolide signal observed with the addition of base to the DHB matrix result largely because of a combined effect involving the matrix solution pH and the lithium ion concentration. To illustrate this point, the pH of the matrix solution, measured using pH paper, is also plotted on Fig. 6 (again a point for the 25 mM LiCl-mediated matrix was added for comparison and indicated on the plot). The stock DHB (20 mg/mL, pKa 3.01) solution has a low pH (pH 2–3) as DHB is an aromatic acid, and the addition of base forms a solution with a buffering capacity between pH *2 to 4. Figure 6 shows that with the addition of up to 100 mM LiOH to DHB, the matrix pH is still acidic and the TAG-estolide signals decrease within this range. However, further increasing the LiOH concentration up to 400 mM, results in a basic matrix solution (pH 8–10), and enhanced analyte signals were detected. Upon increasing LiOH concentration even higher (over 500 mM), the analyte signals decreased again and ion suppression is believed to be the major cause. Overall, compared with signals from LiCl-mediated matrix, the addition of base to DHB solution does have an optimal pH range, however this results in no advantage for improving TAG-estolide signal intensity. Therefore for the DHB matrix, LiCl remains the best choice for analyzing TAGestolides. The effect of TFA and LiOH on the THAP matrix was also investigated. Mass spectra of ergot oil using a THAP matrix in the presence of 0.1 % TFA, 25 mM LiOH, and
J Am Oil Chem Soc (2014) 91:905–915
913
Fig. 8 Relationship of base peak signal-to-noise (S/N), and pH of the DHB solution, as a function of [LiOH]. For comparison, points representing the S/N and pH using a 25 mM LiCl-mediated matrix were added to the plot. Note that S/N is the average of six replicates
Fig. 7 MALDI-MS spectra of ergot TAG-estolides obtained using a THAP matrix in the presence of: a 0.1 % TFA, b 25 mM LiOH and c 50 mM LiOH
50 mM LiOH are shown in Figs. 7a–c, respectively. The addition of TFA alone to the THAP matrix (Fig. 7a) is similar to DHB, with TAG-estolides detected in both sodiated and potassiated forms, and a lower signal was observed (S/N was 41 % less compared with signals obtained with 25 mM LiCl). The addition of TFA to the LiCl-containing THAP matrix resulted in the detection of only lithiated TAG-estolide ions, with the overall signals reduced by around 28 % (data not shown). The effect of base on the THAP matrix (Fig. 7b, c), however, is very different from DHB. Compared with signals obtained using 25 mM LiCl-mediated THAP, the overall TAG-estolide signals were similar at a LiOH concentration of 50 mM (Fig. 7c) but were increased *eightfold in the presence of 25 mM LiOH (Fig. 7b). A more detailed study of the S/N of the base peak as a function of the LiOH concentration is shown in Fig. 8. Again, for comparison, a data point representing the S/N of the base peak from spectra collected using the 25 mM LiCl-mediated matrix was also added to the figure. The figure shows that the use of LiOH with a concentration ranging from 10 to 40 mM (the optimum concentration is 20–30 mM) resulted in dramatically improved S/N, with 5
to eightfold higher compared with signals from 25 mM LiCl-containing THAP. Note that if a different concentration of THAP is desired, the lithium hydroxide concentration should be re-examined accordingly to achieve optimal results. The pH values of the corresponding THAP-base solutions (and the 25 mM LiCl-mediated matrix solution for comparison) were recorded and are also shown as a function of log[LiOH] in Fig. 8. The stock THAP (pKa 7.8) solution has a pH around 4–5, which is higher than DHB. When adding LiOH to THAP, the solution has a weak buffering capacity between pH *6.8 to 8.8. Addition of 10 mM LiOH resulted in a solution pH that was close to neutral, and higher LiOH concentrations up to 50 mM resulted in slightly basic pH values (pH 8 at 50 mM). The best TAG-estolide signals were detected from THAP in the presence of 10–40 mM LiOH, which indicates that neutral or slightly basic pH values represent the best conditions for TAG-estolide analysis. In comparison, the DHB matrix solution did not reach neutral pH until the LiOH concentration was over 100 mM (Fig. 6). Therefore, it is hypothesized that at this high concentration of base, the adverse effects of ion suppression exceeded the benefits of having a slightly basic pH effect and thus, the addition of base to DHB matrix showed no advantage for improving TAG-estolide signal.
Conclusions This study shows that MALDI-MS enables a rapid qualitative analysis for the identification of TAG-estolides using a limited amount of sample. The use of DHB and THAP, when used in combination with 25 mM LiCl, provided
123
914
equivalent signals for MS and MS/MS analysis of TAGestolides. However, THAP in combination with LiOH is the matrix of choice since THAP showed a significantly improved performance when a small amount of lithium hydroxide was added to neutralize the matrix solution. For 10 mg/mL of THAP, the addition of 20–30 mM LiOH is optimal for this purpose. Acknowledgments The authors are grateful for the technical support from Stephen Ambrose and Doug Olson. Thanks also to Dr. Patrick Covello and Dr. Jonathan Page for their internal review of this paper.
References 1. Hayes DG, Kleiman R, Phillips BS (1995) The triglyceride composition, structure, and presence of estolides in the oils of Lesquerella and related species. J Am Oil Chem Soc 72:559–569 2. Madrigal RV, Smith CR Jr (1982) Estolide triglycerides of Trewia nudiflora seed oil. Lipids 17:650–655 3. Mikolajczak KL, Smith CR Jr, Wolff IA (1967) Glyceride structure of Cardamine impatiens L. seed oil. Lipids 3:215–220 4. Morris LJ, Hall SW (1965) The structure of the glycerides of ergot oils. Lipids 1:188–196 5. Batrakov SG, Tolkachev ON (1997) The structures of triacylglycerols from sclerotia of the rye ergot Claviceps purpurea (Fries) Tul. Chem Phys Lipids 86:1–12 6. Isbell TA (2011) Chemistry and physical properties of estolides. Grasas Aceites 62:8–20 7. Cermak SC, Isbell TA (2003) Synthesis and physical properties of estolide-based functional fluids. Ind Crop Prod. 18:183–196 8. Cermak S, Isbell T (2004) Estolides-the next biobased functional fluid. Inform 15:515–517 9. Hayes DG, Kleiman R, Weisleder D, Adlof RO, Cuperus FP, Derksen JTP (1995) Occurrence of estolides in processed Dimorphotheca pluvialis seed oil. Ind Crop Prod 4:295–301 10. Aitzetmuller K, Xin Y, Werner G, Gronheim M (1992) Highperformance liquid chromatographic investigations of stillingia oil. J Chromatogr 603:165–173 11. Sprecher HW, Maier R, Barber M, Holman RT (1965) Structure of an optically active allene-containing tetraester triglyceride isolated from the seed oil of Sapium sebiferum. Biochemistry 13:1856–1863 12. Rˇezanka T, Sigler K (2007) The use of atmospheric pressure chemical ionization mass spectrometry with high performance liquid chromatography and other separation techniques for identification of triacylglycerols. Curr Anal Chem 3:252–271 13. Lin J-T, Arcinas A, Harden LR, Kfgerquist CK (2006) Identification of (12-ricinoleoylricinoleoyl)diricinoleoylglycerol, an acylglycerol containing four acyl chains, in castor (Ricinus communis L.) oil by LC-ESI-MS. J Agric Food Chem 54:3498–3504 14. Zhang H, Olson DJH, Van D, Purves RW, Smith MA (2012) Rapid identification of triacylglycerol-estolides in plant and fungal oils. Ind Crop Prod 37:186–194 15. Smith MA, Zhang H, Forseille L, Purves RW (2013) Characterization of novel triacylglycerol estolides from the seed oil of Mallotus philippensis and Trewia nudiflora. Lipids 48:75–85 16. Cozzolino R, De Giulio B (2011) Application of ESI and MALDI-TOF MS for triacylglycerols analysis in edible oils. Eur J Lipid Sci Technol 113:160–167
123
J Am Oil Chem Soc (2014) 91:905–915 17. Schiller J, Su¨ß R, Arnhold J, Fuchs B, Leßig J, Muller M et al (2004) Matrix-assisted laser desorption and ionization time-offlight (MALDI-TOF) mass spectrometry in lipid and phospholipid research. Prog Lipid Res 43:449–488 18. Han X, Yang K, Gross RW (2012) Multi-dimensional mass spectrometry-based shotgun lipidomics and novel strategies for lipidomic analyses. Mass Spectrom Rev 31:134–178 19. Asbury GR, Al-Saad K, Siems WF (1999) Analysis of triacylglycerols and whole oils by matrix-assisted laser desorption/ ionization time of flight mass spectrometry. J Am Soc Mass Spectrom 10:983–991 20. Ejsing CS, Sampaio JL, Surendranath V, Duchoslav E, Ekroos K, Klemm R et al (2009) Global analysis of the yeast lipidome by quantitative shotgun mass spectrometry. Proc Natl Acad Sci USA 106:2136–2141 21. Islam MN, Chambers JP, Ng CK-Y (2012) Lipid profiling of the model temperate grass, Brachypodium distachyon. Metabolomics 8:598–613 22. Ivanova PT, Milne SB, Byrne MO, Xiang Y, Brown HA (2007) Glycerophospholipid identification and quantitation by electrospray ionization mass spectrometry. Method Enzymol 432:21–57 23. Ayorinde FO, Elhilo E, Hlongwane C (1999) Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry of canola, castor and olive oils. Rapid Commun Mass Spectrom 13:737–739 24. Ayorinde FO, Eribo BE, Balan KV, Johnson JHJ, Wan LW (1999) Determination of major triacylglycerol components of polyunsaturated specialty oils using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Rapid Commun Mass Spectrom 13:937–942 25. Picariello G, Paduano A, Sacchi R, Addeo F (2009) MALDI-TOF mass spectrometry profiling of polar and nonpolar fractions in heated vegetable oils. J Agric Food Chem 57:5391–5400 26. Lay JO Jr, Liyanage R, Durham B, Brooks J (2006) Rapid characterization of edible oils by direct matrix-assisted laser desorption/ionization time-of-flight mass spectrometry analysis using triacylglycerols. Rapid Commun Mass Spectrom 20:952–958 27. Gidden J, Liyanage R, Durham B, Lay JO Jr (2007) Reducing fragmentation observed in the matrix-assisted laser desorption/ ionization time-of-flight mass spectrometric analysis of triacylglycerols in vegetable oils. Rapid Commun Mass Spectrom 21:1951–1957 28. Black C, Poile C, Langley J, Herniman J (2006) The use of pencil lead as a matrix and calibrant for matrix-assisted laser desorption/ ionization. Rapid Commun Mass Spectrom 20:1053–1060 29. Langley JG, Herniman JM, Townell MS (2007) 2B or not 2B, that is the question: further investigations into the use of pencil as a matrix for matrix-assisted laser desorption/ionization. Rapid Commun Mass Spectrom 21:180–190 30. Olson DJH, Smith MA, Dauk M, Reed DW, Abrams SR (2008) Negative ion pencil lead MALDI and MALDI MS/MS for the identification and structural analysis of free fatty acids. In: 56th ASMS conference on mass spectrometry and allied topics, Denver 31. Olson DJH, Giblin ME, Taylor DC, Ross ARS (2007) Enhanced production of cationized molecular ions by MALDI and structural elucidation of triacylglycerols by PSD using a graphite matrix. In: 55th ASMS conference on mass spectrometry and allied topics, Indianapolis 32. Pittenauer E, Allmaier G (2009) The renaissance of high-energy CID for structural elucidation of complex lipids: MALDI-TOF/ RTOF-MS of alkali cationized triacylglycerols. J Am Soc Mass Spectrom 20:1037–1047 33. Cerruti CD, Touboul D, Guerineau V, Petit VW, Laprevote O, Brunelle A (2011) MALDI imaging mass spectrometry of lipids
J Am Oil Chem Soc (2014) 91:905–915
34.
35.
36. 37.
by adding lithium salts to the matrix solution. Anal Bioanal Chem 401:75–87 Cheng C, Gross ML (1998) Complete structural elucidation of triacylglycerols by tandem sector mass spectrometry. Anal Chem 70:4417–4426 Hsu F–F, Turk J (2010) Electrospray ionization multiple-stage linear ion-trap mass spectrometry for structural elucidation of triacylglycerols: assignment of fatty acyl groups on the glycerol backbone and location of double bonds. J Am Soc Mass Spectrom 21:657–669 Siuzdak G (1996) Mass spectrometry for biotechnology. Academic Press, San Diego, pp 18–19 Al-Saad K, Zabrouskov V, Siems WF, Knowles NR, Hannan RM, Hill HH Jr (2003) Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry of lipids: ionization and prompt fragmentation patterns. Rapid Commun Mass Spectrom 17:87–96
915 38. Keller BO, Li L (2000) Discerning matrix-cluster peaks in matrix-assisted laser desorption/ionization time-of-flight mass spectra of dilute peptide mixtures. J Am Soc Mass Spectrom 11:88–93 39. Stu¨biger G, Belgacem O (2007) Analysis of lipids using 2,4,6trihydroxyacetophenone as a matrix for MALDI mass spectrometry. Anal Chem 79:3206–3213 40. Nagai T, Mozobe H, Otake I, Ichioka K, Kojima K, Matsumoto Y, Gotoh N, Kuroda I, Wada S (2011) Enantiomeric separation of asymmetric triacylglycerol by recycle high-performance liquid chromatography with chiral column. J Chromatogr A 1218: 2880–2886 41. Picariello G, Romano R, Addeo F (2010) Nitrocellulose film substrate minimizes fragmentation in matrix-assisted laser desorption ionization time-of-flight mass spectrometry analysis of triacylglycerols. Anal Chem 82:5783–5791
123