Multistep Mass Spectrometry Methodology for Direct Characterization

Nov 26, 2012 - tablet form and Nannochloropsis in paste form by mass spectrometry (MS). Tandem mass spectrometry (MS/MS) experiments using collision ...
0 downloads 0 Views 425KB Size
Article pubs.acs.org/ac

Multistep Mass Spectrometry Methodology for Direct Characterization of Polar Lipids in Green Microalgae using Paper Spray Ionization Sheran A. Oradu and R. Graham Cooks* Department of Chemistry, Purdue University, West Lafayette, Indiana 47906, United States S Supporting Information *

ABSTRACT: Paper spray ionization, an ambient ionization method, has been applied for the identification of polar lipids in green microalgae with no sample preparation. A multistep experimental protocol was employed to characterize the lipid species of two microalgae strains, Kyo-Chlorella in tablet form and Nannochloropsis in paste form by mass spectrometry (MS). Tandem mass spectrometry (MS/MS) experiments using collision induced dissociation (CID) were employed for initial characterization of the detected lipid species, which were dominated by polar glycolipids and phospholipids. Product ion scan experiments were performed to determine the lipid head groups and fatty acid composition. Precursor ion scan experiments using fragment ions such as m/z 184, which is characteristic of the phosphocholine headgroup, were then used to confirm the lipid classification. Lipid elemental compositions were determined by exact mass measurements using high resolution mass spectrometry. Finally, the position of unsaturation was determined using reactive paper spray ionization experiments with ozone used as a reagent to cleave double bonds. Ozone was produced in situ using dielectric barrier discharge from a low temperature plasma, and it reacted in ambient air with the spray of ions produced by paper spray ionization. Using the precursor ion scan experiment, the resulting ozone cleavage product ions were used to determine the position of unsaturation for some of these species. By applying this experimental protocol, the molecular formulas and key aspects of the structures of glycerophosphocholines (PCs) such as 9Z-16:1/9Z,12Z-16:2 PC and 6Z,9Z-18:2/6Z,9Z,12Z-18:3PC and monogalactosyldiacylglycerols (MGDGs) such as 18:3/16:3MGDG were identified in the positive ion mode, while glycerophosphoglycerols (PGs) such as 18:3/16:0 PG and sulfoquinovosyldiacylglycerols (SQDGs) such as 18:3/16:0 SQDG were identified in the negative ion mode.

M

chromatographic methods coupled to tandem mass spectrometry (MS/MS) and exact mass measurements.6 While these reliable methods have added to the available information on algal lipid composition, they depend on considerable sample preparation to isolate lipid fractions prior to analysis. The ability to distinguish high-lipid strains of microalgae from a pool of thousands of potential strains7 and to optimize processing conditions to the nature of the feedstock requires rapid analytical techniques. Such methods would enable rapid preprocessing characterization of algal lipid content. Analytical preprocessing techniques that are rapid, even if they provide incomplete molecular information, suffice for this application. Ambient ionization mass spectrometry methods address this need by simplifying the work flow in MS analysis. In ambient ionization MS analysis, samples are analyzed by forming ions outside the mass spectrometer without the need for chemical separation or extraction steps thus reducing the analysis time.8 These methods include desorption electrospray ionization

icroalgae have received considerable attention as feedstocks for biodiesel production because of their low land usage, rapid growth rate (1−3 doublings per day) based on their photosynthetic efficiency, and high biodiesel yield associated to their inherently high hydrocarbon content (20−50% by weight).1 It is estimated that 100 000 gallons of biodiesel can be produced from 1 acre of microalgae annually. Algal lipid composition is an important biochemical characteristic which affects biodiesel production. Full or partial preprocessing and characterization of lipid class and fatty acid composition should be useful for selecting the most suitable microalgae strain from a pool of potential strains.2 This information would also enable the development of suitable downstream processing strategies to maximize oil extraction, biodiesel conversion and biodiesel oxidative stability. Chromatographic methods coupled with mass spectrometry (MS), have been used in previous algal research to determine lipid and fatty acid compositions in different microalgae strains.3−5 In one report, the lipid and fatty acid composition difference between autotrophic and heterotrophic microalgae strains was described.3 Recently, molecular structures of lipids in particular microalgae strains were determined using © 2012 American Chemical Society

Received: June 19, 2012 Accepted: November 12, 2012 Published: November 26, 2012 10576

dx.doi.org/10.1021/ac301709r | Anal. Chem. 2012, 84, 10576−10585

Analytical Chemistry

Article

Figure 1. (a) Experimental arrangement for paper spray and reactive paper spray using ozone as the reagent (b) Illustration of MS/MS precursor ion scan and product ion scan experiments. (b) Adapted from ref 36.

(DESI),9 direct analysis in real time (DART),10 and laser ablation electrospray ionization (LAESI),11 among others. Successful applications of ambient ionization methods to the analysis of chemically complex samples in their native state have been reported for animal tissues,12,13 plant materials,14−17 single cells18,19 and other biological matrices.20−22 The molecular selectivity essential for the successful implementation of ambient analysis has been achieved by tandem MS,23,24 the use of in situ chemical reactions,25,26 and by high-resolution MS.27 High-resolution accurate mass measurements on complex matrices have facilitated the identification of lipids in rat brain28 and in microalgae extracts.6 In the relatively new method, paper spray ionization, a sample is loaded onto a triangular piece of chromatography paper.29 Compared to the related method of nano electrospray ionization, paper spray incorporates three analytical procedures: sample collection, analyte separation and analyte ionization,29 making it attractive for field applications in complex mixtures analysis. Successful applications include direct analysis and quantitation of pharmaceutical drugs and biomarkers in crude biological samples such as urine,30 dried serum and whole blood with acceptable reproducibility (RSD of ∼10%)29,31−33 and the analysis of natural products in plant tissue.34,35 The simplicity of paper spray ionization, the ability to analyze crude

samples without the need for sample preparation and its potential to be coupled to a portable mass spectrometer for field analysis, makes it attractive for use in algal biodiesel production plants for prescreening. In this paper, we report the use of paper spray ionization for the detection and molecular characterization of polar lipids from crude microalgae samples. To achieve more complete characterization, we employed a multistep experimental protocol involving tandem mass spectrometry, high resolution exact mass measurement and reactive ambient ionization experiments. We selected kyoChlorella and Nannochloropsis micro algae strains for these experiments. These strains are good candidates for potential large scale biodiesel production due to their high lipid content and fast growth rates. Tandem mass spectrometry using both product and precursor ion scans36−39 and high resolution exact mass measurements were employed to determine the lipid functional groups, molecular formulas and aspects of their structures. Additional reactive paper spray ionization experiments using ozone as the reagent enabled the determination of the position of double bonds in some of the detected lipid fatty acyl chains through characterization of the products of double bond cleavage in the precursor ion scan mass spectra. 10577

dx.doi.org/10.1021/ac301709r | Anal. Chem. 2012, 84, 10576−10585

Analytical Chemistry



Article

METHODS Chemicals and Materials. The chromatography paper used for paper spray ionization was from Whatman (Whatman International Ltd., Maidstone, England), methanol (HPLCgrade) from Mallinckrodt Baker Inc. (Phillipsburg, NJ), and Reserpine was purchased from Sigma-Aldrich (Milwaukee, WI). The chlorophyll catabolite, Cj-NCC-1, which was previously used in collaborative work with Dr. T. Mueller,17 was isolated and purified at the University of Innsbruck in Austria. KyoChlorella microalgae in tablet form were purchased from the Vitamin Shoppe Industries, Inc. (North Bergen, NJ). Nannochloropsis microalgae sample in paste form was purchased from Reed Mariculture, Inc. (Campbell, CA). Instrumentation. Mass analysis was performed using a Thermo Fisher LTQ mass spectrometer (Thermo Scientific Inc., San Jose, CA). The temperature of the MS capillary inlet was typically set at 150 °C. The tube lens voltage was set at 65 V and the capillary voltage kept at ±15 V. The voltage used for paper spray ionization was 4.5 kV in the positive mode and −3.5 kV in the negative mode. Tandem mass spectrometry for structural elucidation was carried out using collision induced ionization (CID). An isolation window of 1.5 Th (mass/charge units) and normalized collision energy of 20%−30% (manufacturer’s unit) was chosen. Exact mass measurement experiments were carried out using the Thermo Fisher Scientific LTQ-Orbitrap mass spectrometer at a mass resolution of 60 K. Reserpine (m/z 609.280657) and the isolated chlorophyll catabolite Cj-NCC-1 (m/z 643.276224) were used as the lock masses for the positive and negative ion modes respectively. All precursor ion scan and ozonolysis experiments were performed using a TSQ Quantum Access Max (Thermo Scientific, San Jose, CA). The precursor ion scan mode is used to detect specific parent ions producing a particular fragment (product) ion by collision induced dissociation. This was done by fixing conditions in Q3 to transmit a particular fragment ion while scanning the first mass analyzing quadrupole, Q1.36,37 The precursor ion scan parameters were as follows: Precursor ion m/z 184 in the positive ion mode and m/z 225 in the negative ion mode; Q2 offset (collision energy) 28 V; scan time 0.5s; Q1 peak width 0.7Th; Q2 CID pressure 1.0 Torr. Ozonolysis experiments were performed by intercepting the spray of ions from paper spray analysis with ozone as described later. Ozone was produced in situ from a low temperature plasma as described elsewhere.40 Compressed air/oxygen was substituted as the discharge gas for maximum ozone production, which was confirmed by its characteristic smell. The odor of ozone is distinctive at low concentrations (0.02−0.05 ppm). The permissible exposure limit for ozone is 0.1 ppm.41 As a safety measure, use of an enclosure around the ozone source is recommended to limit direct exposure. Paper Spray Ionization Mass Spectrometry. Prior to paper spray analysis, ∼0.5 μg of Nannochloropsis green algae paste was spotted onto a 1 cm × 1 cm × 1 cm triangle of chromatography paper. The kyo-Chlorella tablet was crashed to powder and a solution made by dissolving 1 mg of tablet powder in 1 mL of pure methanol. Approximately 0.5 μg (amount of algae) in the Chlorella tablet solution was then spotted onto the paper. Methanol/water solution (20 μL, 9:1, v/v) was used as solvent for paper spray unless otherwise noted. For a triangle of the above dimension, addition of 10 μL of solvent to the paper allowed analysis for up to 1 min, enough

time to perform MS and tandem MS experiments for several of the detected molecules. More solvent was added when additional experiments needed to be performed.



RESULTS AND DISCUSSION Characterization of the crude microalgae samples using paper spray ionization mass spectrometry was done using three types of experiments. Figure 1 illustrates the experimental set up for paper spray and reactive paper spray ionization, and the tandem mass spectrometry scan modes employed. From simple low resolution tandem mass spectrometry experiments using the product ion scan, information on the functional head groups and fatty acid composition was obtained. This enabled the classification of lipids as PG, PC, SQDG or MGDG. However, the complexity of the algae mixtures made it necessary to perform subsequent experiments using high resolution exact mass measurements and precursor ion scans to verify preliminary identifications and to distinguish isomers present in the lipid mixture. Tandem mass spectrometry also did not provide information on the position of unsaturation. Therefore, subsequent reactive paper spray experiments were done in attempts to do so. These experiments met with several difficulties related to the complexity of the algal matrix, resulting in only two successful identifications of positions of unsaturation, although these experiments were successful when lipid standards and simpler mixtures were used. Molecular Mass Profiles and Tandem Mass Spectrometry. First, the molecular mass profiles for both the kyoChlorella microalgae and the Nannochloropsis microalgae were recorded. The full mass spectra showed a cluster of peaks in the m/z 600−1000 range in both negative and positive ion modes. These peaks were suspected to be due to characteristic polar lipids such as phospholipids and lyso-phospholipids based on previous reports on lipid analysis by MS.42−48 Tandem mass spectrometry (MS/MS) was employed for initial structural characterization of the molecular species identified. The product ion scan mass spectra showed neutral losses (e.g., carboxylic acids) and product ions (e.g., m/z 184) in the positive ion mode, which are characteristic of monogalactosyldiacylglycerols (MGDG)6,49,50 and glycerophosphocholines (PC), respectively.42,43,51,52 In the negative ion mode, the product ion m/z 225 and the ions due to the loss of fatty acyl chains observed are characteristic of sulfoquinovosyldiacylglycerols (SQDG)6,50,53 and glycerophosphoglycerols (PG)42,43,46 respectively. Fragment ions which aid in structural elucidation, are typically characterized by losses of their polar head groups, lyso-lipid formation and fatty acid fragments.42−53 To further verify the classes to which the observed lipids belong, precursor ion scan experiments were used. In these experiments, the parent ions producing a particular product ion were recorded. Parent ions resulting from PC which form a characteristic m/z 184 product ion, and those from SQDG which produce a characteristic m/z 225 product ion, were recorded in the positive and negative ion modes respectively. This enabled the ready distinction between these lipid species and those associated with MGDG and PG. Although precursor ion scan experiments for MGDG and PG have been reported using precursor ions m/z 243 and m/z 227, respectively,53 these experiments were not employed in this study due to low intensities of these characteristic fragment ions. Additionally, precursor ion scan experiments of m/z 227 for PG detection have been shown to be less specific while the alternative characteristic fragment ion m/z 153 has been reported to be 10578

dx.doi.org/10.1021/ac301709r | Anal. Chem. 2012, 84, 10576−10585

Analytical Chemistry

Article

Figure 2. Full MS spectrum from the analysis of kyo-Chlorella in tablet form: (a) positive ion mode and (b) negative ion mode.

Figure 3. MS/MS spectra of negative and positive ions detected in kyo-Chlorella green microalgae. Characteristic fragment ions showing detection of (a) PC and (b) MGDG in the positive ion mode and (c) PG and (d) SQDG in the negative ion mode.

10579

dx.doi.org/10.1021/ac301709r | Anal. Chem. 2012, 84, 10576−10585

Analytical Chemistry

Article

nonspecific.53 The initial lipid class identifications made through MS/MS experiments are consistent with reports that sulfolipids, galactolipids, PC and PG are the predominant lipid components of the photosynthetic membrane in plants, algae and various bacteria.54−56 Kyo-Chlorella Green Microalgae. The MS/MS product ion spectra of the lipids detected in kyo-Chlorella microalgae produced two distinctly different fragmentation patterns. The product ion spectra of positive ions such as those at m/z 750, m/z 778 and m/z 800, showed fragment ions m/z 184, [M + H − 59] + and [M + H − 183] + which are characteristic PC fragment ions42,43,51,52 (Figure 3a). These are due to the losses of the monoprotonated dihydrogenphosphate choline polar headgroup [C5H15NO4P] + via cleavage of the sn-3 carbon− oxygen bond, the trimethylamine [N (CH3)3] headgroup and a neutral molecule of the dihydrogenphosphate choline polar headgroup [C5H14NO4P], respectively (Table 1). Low intensity

Table 2. Molecular Formulae Calculated using High Resolution Exact Mass Measurement Indicating the Identification of MGDG, PC, PG and SQDG kyo-Chlorella Microalgae in Table Form proposed formula

743.4704

C41H68O10 + Na

750.5055

C40H74NO8P + Na C43H70O10 + Na

769.4866 778.54 785.4598

Table 1. Main Classes of Polar Phospholipids Detected in Green Microalgaea

a

measured m/z

C42H78NO8P + Na C43H70O10 + K

800.5196

C44H76NO8P + Na

719.4858 721.5017 741.4700 743.48651 757.50208 793.51356 815.49801 837.48243

C38 H72O10P C38H74O10P C40 H74O10P C40H72O10P C41H72O10P C41H77O12S C43H75O12S C45H73O12S

error (ppm)

Positive ions −0.189 0.741 −0.038 0.868 0.004 0.220 Negative ions 0.096 0.483 −0.150 0.321 0.672 0.599 0.387 −0.053

identity

FA composition

32:5 MGDG 32:3 PC

16:2, 16:3

34:6 MGDG 34:2 PC

16:3, 18:3

34:6 MGDG 36:6 PC

16:3, 18:3

32:1 PG 32:0 PG 34:4 PG 34:3 PG 35:4 PG 32:0 SQDG 34:3SQDG 36:6SQDG

16:1, 16:0, 18:3, 18:3, 17:0, 16:0, 16:0, 18:3,

16:1, 16:2

16:1, 18:1

18:3, 18:3

16:0 16:0 16:0 16:1 18:4 18:0 18:3 18:3

detected in the positive ion mode is the diacylglycerols, particularly MGDG, which has also been reported to be found in significant amounts in green microalge.50 Similar fragmentation patterns showing significant losses of carboxylic acids in MGDG species have been reported previously.49 This study also suggested that the loss of MGDG fatty acyl chains is favored at the sn-1 position. This would suggest that for m/z 769 ion (Figure 3b), the 18:3 fatty acyl chain is at the sn-1 position and 16:3 is at the sn-2 position. Additionally, these species were not detected in the recorded m/z 184 precursor ion mass spectra, confirming that they belong to a different class of lipids. A search on the lipid map structure database revealed an MGDG lipid (18:3(9Z, 12Z, 15Z)/16:3(7Z, 10Z, 13Z MGDG) with molar mass 746.5 Da (C43H70O10).48,49 This lipid is structurally similar to the sodiated (m/z 769) and potassiated (m/z 785) species identified in Table 2 as 34:6 MGDG. Similarly, the MS/MS fragmentation patterns recorded in the negative ion mode suggested the detection of two major classes of lipids. The rich fragmentation pattern in the negative ion mode allowed for easier structural characterization of the detected lipids. The CID product ion spectrum of the negative ion m/z 741, for instance, was characterized by a major loss of one of the fatty acyl substituents as a ketene, detected as an ̀ abundant fragment ion at m/z 505, [M − H-R2CHCO] − (Figure 3c). Less abundant peaks associated with the loss of ̀ carboxylic acid m/z 487 [M − H-RCOOH]−, were also observed. These fragment ions are characteristic of PG.23,42,46 Carboxylate fragment ions m/z 253 and m/z 277 were observed as well. The difference between the mass of these two fatty acids and the [M − H] − ion m/z 741, is 209 Da. This is indicative of a glycerophosphoglycerol headgroup.46 Results obtained from experiments done using phospholipid standards in the negative ion mode have suggested that the carboxylic acid loss [− RCOOH] is more probable from the sn-2 position

Ionic lipids are in their sodiated forms.

fragment ions in the m/z 400−500 region were observed as well. These are presumably due to the formation of a lyso-PC ion, as a result of the loss of one of the fatty acyl groups from the sn-1 or sn-2 glycerol positions as a ketene [M-R1CHC O]+ or loss of a carboxylic acid from the sn-1 or sn-2 glycerol positions. It has been reported that fragment ions in the m/z 400−500 region, including the [M + H − 59] + and [M + H − 183] + ions are prevalent in sodiated or potassiated PC ions in contrast to the behavior of [M + H]+ PC ions, which only show the m/z 184 fragment ion.42,51 These data therefore suggest that the detected PC fragment ions are either [M + Na] + or [M + K] + species. Neutral losses of carboxylic acids were used to suggest the fatty acid chain compositions of these species as summarized in Table 2. Precursor ion experiments done by scanning all parent ions that fragmented to give the m/z 184 product ion, further confirmed that the species at m/z 750, m/z 778 and m/z 800 among others, are due to PC lipids. Summary of the identities of PC molecules are given in Table 2. The fragmentation pattern of ions such as m/z 743, m/z 769 and m/z 785, recorded in the positive ion mode, were structurally characteristic. The MS/MS spectra of m/z 769 for instance, only showed a loss of acetic acid (60 Da), and significant peaks at m/z 491 and m/z 519 resulting from the neutral losses of C18:3 and C16:3 carboxylic acid chains respectively (Figure 3b). MS/MS fragmentation patterns of these species indicated very clear fatty acid losses which were useful for determining their fatty acid compositions (Table 2). The spectra suggest that the other major group of lipids 10580

dx.doi.org/10.1021/ac301709r | Anal. Chem. 2012, 84, 10576−10585

Analytical Chemistry

Article

Figure 4. Paper spray MS and MS/MS spectra (inset) of Nannochloropsis microalgae confirming the identification of (a) MGDG in the positive ion mode (b) SQDG in the negative ion mode.

of the glycerol backbone than from sn-1.42 Since fatty acids anions are formed primarily by further decomposition of the [M − H-RCOOH] − ion, it has been suggested that the observed fatty acid anion from the sn-1 position is favored.57 Other quantitative studies done to analyze phospholipid positional isomers by MS3 experiments have suggested that neutral loss of fatty acids as ketene is the most position specific fragmentation pathway. This loss is suggested to occur predominantly from the sn-2 position.58 These reports in addition to our observations, therefore suggest that in the example of the m/z 741 ion, the sn-1 position is occupied by the C18:3 fatty acid, while the sn-2 position is occupied by the C16:1 fatty acid (Figure 3c). Ions detected at m/z 743, m/z 719, m/z 721 and m/z 757 were similarly characterized and classified as PG phospholipids and their fatty acid chain composition determined (Table 2). The negative ions m/z 793, m/z 815 and m/z 837 fragment with significant losses of the neutral carboxylic acids. The MS/ MS product ion spectrum of m/z 815 is shown in Figure 3d. The ratios of the carboxylic acid losses show that the loss of C18:3 fatty acid is more probable, suggesting but not establishing that the C18:3 fatty acid could be at the sn-2 position and the C16:0 fatty acid at the sn-1 position. Another distinctive ion observed in each of the MS/MS product ion spectra of the ions noted above was the characteristic fragment ion m/z 225. This is suspected to be the dehydrated sulfoquinovose species which is formed via cleavage of the sn3 carbon−oxygen bond (Table 1), and is a characteristic

fragment ion for the sulfoquinovosyldiacylglycerol (SQDG) class of lipids.6,40,50 This suggested that all three ions mentioned are those of SQDG lipids. Attempted precursor ion experiments in the negative ion mode were not as successful as those of the positive ion mode. Despite obtaining only low signal intensities in this experiment, we were able to distinguish from precursor ion scan mass spectra that m/z 793, m/z 815 and m/z 837 ions are those characteristic of SQDG (Supporting Information Figure 1). The MS/MS product ion spectrum of the m/z 793 suggested that it is due to 16:0/16:0 SQDG (Supporting Information Figure 2B). Further evidence for molecular composition comes from a Lipid Maps structure database search where a structurally similar neutral compound (1,2-dihexadecanoyl-3−96′-sulfo-α-D-quinovosyl)-sn-glycerol) with an exact mass of 794.5213 Da was found.46 The molecular identities of the SQDG molecules are summarized in Table 2. Preliminary classification of the polar lipids detected in the positive and negative ion modes was therefore successfully achieved using MS/MS fragmentation patterns, thus identifying each of the species detected as PC, MGDG, PG or SQDG. Previous studies on determining the position of fatty acyl chains of glycerophospholipids concluded that the ratio of the relative abundance of sn-1/sn-2 fatty acyl chains is dependent on the nature of the headgroup, the length and degree of saturation of the acyl chain, the fragmentation behavior of the compound and instrumental parameters.40 These conditions were not controlled in our experiments, and therefore the positions of fatty acyl chains as either sn-1 or sn-2 remain ambiguous. 10581

dx.doi.org/10.1021/ac301709r | Anal. Chem. 2012, 84, 10576−10585

Analytical Chemistry

Article

Nannochloropsis Green Microalgae. Crude Nannochloropsis green microalgae paste was similarly analyzed and characterized in both the positive and negative ion modes. The full mass spectra (Figure 4) showed fewer ions compared to the rich clusters observed from kyo-Chlorella mass spectra. This is most likely a result of the increased complexity of the microalgae paste matrix compared to the tablet form, from which unwanted materials had been removed. Additionally the cell walls of the kyo-Chlorella microalgae in tablet form are broken, perhaps making the detection of polar phospholipids easier. Nevertheless, the negative and positive ion mode MS/MS fragmentation patterns of detected Nannochloropsis ions were similar to those observed from the kyo-Chlorella, also revealing two distinctly different MS/MS spectral types (Figure 4). The MS/MS fragmentation patterns similarly confirmed that the polar head groups in the positive ion mode were those of PC, and MGDG (Figure 4a). Negative ions such as m/z 765 and m/z 791 fragmented with significant neutral losses of carboxylic acids in addition to the characteristic dehydrated sulfoquinovose fragment ion (m/z 225), and were thus identified as the SQDG species (Figure 4b). A neutral SQDG, 1-hexadecanoyl2-hexadecenoyl3-(6′-sulfo-α-D-quinovosyl)-sn-glycerol (exact mass 792.5057 Da) that has structural similarity to the m/z 791 ion was found in the Lipid Maps database, hence further supporting the identity of the constituent responsible for this ion.46 The negative ion of m/z 807 on the other hand, generated carboxylate anions (m/z 253 and m/z 277), lost a fatty acyl substituent as a ketene, and a carboxylic acid, which are characteristic fragment ions for PG species as discussed above. Analogously to the negative ion mode experiments, Nannochlropsis green microalgae were found to have the same classes of compounds as those detected in the green kyochlorella microalgae. The molecular identities of the Nannochloropsis lipids are summarized in Table 3.

MS experiments alone however, did not provide information on the isomer composition or the position of unsaturation. For more complete characterization additional experiments involving high resolution exact mass measurements and reactive paper spray were done. Exact Mass Measurements. MS/MS information was combined with high resolution exact mass measurements for further structural confirmation of the detected lipids. The use of 60000 mass resolution and lock masses proved to be sufficient to determine the molecular formulas with errors considerably below 1 ppm. The measured m/z values for these ions and the proposed chemical formulas and molecular identifications are shown in Table 2 (kyo-Chlorella) and Table 3 (Nannochloropsis). In both cases, the detection of PC, MGDG, PG, and SQDG was confirmed within reasonable error (always less than 1 ppm). The proposed molecular formulas based on exact mass measurement further confirmed that the detected PC and MGDG formed adducts with alkaline metals, as predicted earlier. Reactive Paper Spray: Double Bond Position Determination. Tables 2 and 3 suggest that the identified lipids are characterized by high degrees of unsaturation. This is consistent with literature reports that algae generally have high levels of polyunsaturated fatty acids.54,59,60 The positions of double bonds in fatty acid chains is important to their functionality and stability. The information on algal fatty acid unsaturation should facilitate the development of storage strategies that maximize biodiesel oxidative stability. Identification of the positions of unsaturation in lipids has previously been successfully reported using two triple quadruple methods, OzID61 and OzESI,62,63 and less fully using a simple ion trap low temperature plasma (LTP) experiment.40 OzESI involves sample preparation which ambient paper spray ionization avoids. OzID on the other hand, presents an additional requirement involving instrument modification making these experiments rather demanding even though highly successful. While LTP would be ideal for this work due to its simplicity, it is successful only with lower mass compounds and thus was not used here. To perform ozonolysis experiments, we performed reactive paper spray ionization experiments using ozone produced in situ by a separate low temperature plasma source, as the reagent molecule. Ozone was produced in an LTP discharge,40 using oxygen instead of helium as the discharge gas. The spray of ions from the paper was intercepted by a flow of ozone from the LTP discharge, positioned perpendicular to the ion spray. Using the precursor ion scan mode described earlier, the occurrence of the ambient ozonolysis reaction was confirmed by the detection of fragment ions (mainly aldehydes) due to double bond cleavage, previously reported as a characteristic neutral loss.64 Figure 5 shows the precursor ion scan mass spectra obtained before (Figure 5a) and after (Figure 5b) ozone was introduced. Peaks relating to the ozonolysis cleavage products of the ions m/z 752 and m/z 798 are indicated by their neutral losses using solid and dotted arrows, respectively (Figure 5b). These neutral losses suggest that the m/z 752 PC previously determined as 16:1/16:2 PC was cleaved at the n-4 (loss of 40 Da) and n-7 (loss of 80 Da) double bond positions64 therefore allowing its assignment as 9Z-16:1/9Z, 12Z-16:2 PC, where the double bond geometry was assumed to be Z. The 40 Da difference between m/z 712 and m/z 672 further suggested a skip-conjugate polysaturated fatty acid which confirmed the position of unsaturation in the 16:2 fatty

Table 3. Molecular Formulae Calculated using High Resolution Exact Mass Measurement Indicating the Identification of MGDG, PC, PG and SQDG Nannochloropsis Microalgae in Paste Form measured m/z

proposed formula

732.5747

C41H80O10

754.5358

C40H78O8NP + Na C42H78O8NP + Na C42H80O8NP + Na

778.5356 780.5514

765.48059 791.49789

C39H73O12S C41H75O12S

error (ppm)

Positive ions 0.191

identity

FA composition

32:0 MGDG 32:1 PC

16:0, 16:1

−0.136

34:2 PC

16:1, 18:1

0.140

34:3 PC

16:1, 18:2

30:0 SQDG 32:1 SQDG

15:0, 15:0 16:0, 16:1

0.111

Negative ions −1.600 0.651

16:0, 16:0

The identification of members of these four classes of lipids (about 20 individual lipid molecules) in these two green microalgae strains indicate that paper spray ionization allows detection and tentative characterization by a simple set of low resolution MS experiments without any sample preparation steps being required. In particular, paper spray ionization followed by product ion scan experiments, suffices as a prescreening method, especially when some preliminary information on the lipid content of a strain is available. MS/ 10582

dx.doi.org/10.1021/ac301709r | Anal. Chem. 2012, 84, 10576−10585

Analytical Chemistry

Article

Figure 5. Positive ion mode mass spectra of kyo-Chlorella microalgae produced using reactive paper spray ionization using m/z 184 precursor ion scan (a) before ozone was introduced showing the abundant peaks of 16:1/16:2 PC, 18:1/16:1 PC and a relatively lower intensity 18:2/18:3PC ion (b) after ozone was introduced showing additional peaks produced by ozonolysis cleavage products related to the m/z 752 (solid arrows) and m/z 798 (dotted arrows). Neutral losses indicate the presence of a 9Z-16:1/9Z, 12Z-16:2 PC and a 6Z, 9Z-18:2/6Z, 9Z, 12Z-18:3 PC.



acyl chain. The cleavage products of 18:2/18:3 PC with m/z 798 were observed as indicated by the dotted arrows. The observed neutral losses (68 Da, 108 Da, 148 Da) with 40 Da differences are indicative of n-6 class of fatty acids with double bond positions at 6Z, 9Z and 12Z64 leading to the assignment as 6Z,9Z-18:2/6Z,9Z,12Z-18:3 PC.65 The relative intensities of ozonolysis cleavage products obtained from reactive paper spray were comparable to those obtained using phospholipid standards and pure brain extract samples (Supporting Information Figure 3 and 4). These results therefore validate the viability of reactive paper spray coupled with MS/MS as a method for determining the position of unstaturation in lipid in both simple and complex matrices. As in the OzESI experiments, the main difficulty with this method for complex mixtures is the overlap of the ozonolysis cleavage products with pre-existing lipid ions and the fact that different lipid ions give ozonolysis cleavage products of the same m/z ratio. This may result in misinterpretation of the data. While this method successfully identified position of unsaturation for pure lipid mixtures and standards (Supporting Information Figures 3 and 4), it may not be effective for complex mixtures applications. Additional experiments on the reproducibility of this method with other complex mixtures need to be conducted.

CONCLUSIONS

A multistep mass spectrometry methodology for direct characterization of lipids in complex mixtures has been designed and implemented. The results obtained from the analysis of the two green microalgae species are consistent with reports in the literature on the presence of PC, MGDG, PG, and SQDG as the primary components of green micro algae cell walls. The use of paper spray ionization mass spectrometry coupled with tandem mass spectrometry and high resolution mass spectrometry as an analytical method for characterization of lipid compounds in complex mixtures was demonstrated. Also, the use of reactive paper spray as a potential tool for determining the position of unsaturation in lipids was implemented successfully. These MS methods collectively, should enable preprocessing screening of algal strains for lipid content, to enable identification of best strains for biodiesel production, and the development of strategies to maximize oil production and appropriate storage of oil products. The automation of the paper spray ionization method could extend the applicability of this methodology to larger sample pools. This multistep procedure provided characterization of these lipids essential for initial screening of unknown strains; the data obtained also show that paper spray ionization followed by low resolution tandem mass spectrometry can be sufficient for quick prescreening experiments, when some information on lipid content of a particular algal strain is available. 10583

dx.doi.org/10.1021/ac301709r | Anal. Chem. 2012, 84, 10576−10585

Analytical Chemistry



Article

(26) Wu, C. P.; Ifa, D. R.; Manicke, N. E.; Cooks, R. G. Anal. Chem. 2009, 81, 7618−7624. (27) Pol, J.; Vidova, V.; Kruppa, G.; Kobliha, V.; Novak, P.; Lemr, K.; Kotiaho, T.; Kostiainen, R.; Havlicek, V.; Volny, M. Anal. Chem. 2009, 81, 8479. (28) Bereman, S. M.; Nyadong, L.; Fernandez, M. F.; Muddiman., C. D. J. Mass Spectrom. 2010, 45, 223−226. (29) Manicke, E. N.; Dill, L. A.; Ifa, R. D.; Cooks, R. G. J. Mass Spectrom. 2010, 45, 223−226. (30) Wang, H.; Liu, J.; Cooks, R. G.; Ouyang, Z. Angew. Chem., Int. Ed. 2010, 49, 877−880. (31) Liu, J.; Wang, H.; Manicke, N. E; Lin, J.-M.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2010, 82, 2463−2471. (32) Manicke, N. E.; Yang, Q.; Wang, H.; Oradu, S.; Ouyang, Z. Int. J. Mass Spectrom. 2010, 300, 123−129. (33) Manicke, N. E.; Abu-Rabie, P.; Spooner, N.; Ouyang, Z.; Cooks, R. G. J. Am. Soc. Mass Spectrom. 2011, 22, 1501−1507. (34) Yang, Q.; Manicke, N. E.; Wang, H.; Petucci, C.; Cooks, R. G.; Ouyang, Z. Anal. Bioanal. Chem. 2012, 404, 1389−1397. (35) Liu, J.; Wang, H.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2011, 83, 7608−7613. (36) Tadjimukhamedoc, T. F.; Huang, G.; Ouyang, Z.; Cooks, R. G. Analyst 2012, DOI: 10.1039/c2an16077c. (37) Schwartz, C. J.; Wade, P. A.; Enke, G. C.; Cooks, R. G. Anal. Chem. 1990, 62, 1809−1818. (38) de Hoffman., E. J. Mass Spectrom. 1996, 31, 129. (39) Jackson, A. T.; Jennings, K. R.; Scrivens, J. H. Rapid Commun. Mass Spectrom. 1996, 10, 1449−1458. (40) Jackson, A. T.; Slade, S. E.; Scrivens, J. H. Int. J. Mass Spectrom. 2004, 238, 265−277. (41) Zhang, I. J.; Tao, A. W.; Cooks, R. G. Anal. Chem. 2011, 83, 4738−4744. (42) Limits for air contaminants-Occupational safety and Health Administration, http://www.osha.gov/dts/chemicalsampling/data/ CH_259300.html. (43) Manicke, E. N.; Wiseman, M. J.; Ifa, R. D.; Cooks, R. G. J. Am. Soc. Mass Spectrom. 2008, 19, 531−543. (44) Pulfer, M.; Murphy, C. R. Mass Spec. Rev. 2008, 22, 332−364. (45) Bruegger, B.; Erben, G.; Sandhoff, R.; Wieland, T. F.; Lehman, D,W. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 2339−2344. (46) Han, X.; Gross, W. R. J. Am. Chem. Soc. 1996, 118, 451−457. (47) Snyder, A. P.; Harden, S. C.; Smith, B. W. P. Anal. Chem. 1995, 67, 1824−1830. (48) Berry, A. Z. K..; Murphy, C. R. J. Am. Soc. Mass Spectrom. 2004, 15, 1499−1508. (49) The LIPID MAPS−Nature Lipidomics Gateway, http://www. lipidmaps.org/. (50) Guella, G..; Frassanito, R.; Mancini, I. Rapid Commun. Mass Spectrom. 2003, 17, 1982−1994. (51) Kim, H. Y.; Yoo, S. J.; Kim, S. M. J. Mass Spectrom. 1997, 32, 968−977. (52) Han, X.; Gross, W. R. J. Am. Chem. Soc. 1995, 6, 1202−1210. (53) Kerwin, L. J.; Tuininga, R. A.; Ericsson, H. L. J. Lipid Res. 2004, 35, 1102−1114. (54) Welti, R.; Wand, X.; Williams, D. T. Anal. Biochem. 2003, 314, 149−152. (55) Harwood, L. J. In Lipids in Photosynthesis: Structure, Function and Genetics; Siegenthaler, P.-A., Murata, N., Eds.; Kluwer Academic Publishers: The Netherlands, 1998; pp 53−64. (56) Guschina, A. I.; Harwood, L. J. Prog. Lipid Res. 2006, 45, 160− 186. (57) Heinz, E. In Lipids and Lipid Polymers in Higher Plants; Tevini, M., Lichtenthaler, K. H., Eds.; Springer: Berlin, 1977; p 102. (58) Hsu, F. F.; Turk, J. J. Am. Soc. Mass Spectrom. 2000, 11, 797− 803. (59) Ekroos, K.; Ejsing, C. S.; Bahr, U.; Karas, M.; Simons, K.; Schevchenko, A. J. Lipid Res. 2003, 44, 2181−2192. (60) Cojocarut, M.; Shlosberg, M.; Dubinsky, Z.; Finkel, A. Biomed. Environ. Mass Spectrom. 1988, 16, 477−480.

ASSOCIATED CONTENT

S Supporting Information *

Supplementary figures as noted. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Phone: +1(765) 494-5262. Fax: +1(765) 494-9421. E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS Financial support was provided by the National Science Foundation (0848650-CHE). The authors thank Dr. Kuangnan Qian, Dr. Chungpin Wu (both at ExxonMobil Corp), Dr. Nick Manicke and Dr. Christina Ferreira for valuable discussions.



REFERENCES

(1) Chisti, Y. Biotechnol. Adv. 2007, 25, 294−306. (2) Gong, Y.; Jiang, M. Biotechnol. Lett. 2011, 33, 1269−1284. (3) Wang, G.; Wang, T. J. Am. Oil Chem. Soc. 2011, 89, 135−143. (4) Kim, H. Y.; Yoo, S. J.; Kim, S. M. J. Mass Spectrom. 1997, 32, 968−977. (5) Vieler, A.; Wilhem, C.; Goss, R.; Sub, R.; Schiller, J. Chem. Phys. Lipids 2007, 150, 143−155. (6) He, H.; Rodgers, P. R.; Marshall, G. A.; Hsu, S. C. Energy Fuels 2011, 25, 4770−4775. (7) Hu, Q.; Sommerfield, M.; Jarvis, E.; Ghirardi, M.; Posewitz, M.; Seibert, M.; Darzins, A. Plant J. 2008, 54, 621−639. (8) Harris, A. G.; Galhena, S. A.; Fernandez, M. F. Anal. Chem. 2011, 83, 4508−4538. (9) Takats, Z.; Wiseman, M. J.; Gologan, B.; Cooks, R. G. Science 2004, 306, 471−473. (10) Cody, R. B.; Laramee, J. A.; Durst, H. D. Anal. Chem. 2005, 77, 2297−2302. (11) Nemes, P.; Vertes, A. Anal. Chem. 2007, 79, 8098−8106. (12) Eberlin, L. S.; Ifa, D. R.; Wu, C.; Cooks, R. G. Angew. Chem., Int. Ed. 2010, 49, 873−876. (13) Nemes, P.; Woods, S. A.; Vertes, A. Anal. Chem. 2010, 82, 982− 988. (14) Jackson, A. U.; Tata, A.; Wu, C.; Perry, R. H.; Haas, G.; West, L.; Cooks, R. G. Analyst 2009, 134, 867−74. (15) Lane, A. L.; Nyadong, L.; Galhena, A. S.; Shearer, T. L.; Stout, E. P.; Parry, R. M.; Kwasnik, M.; Wang, M. D.; Hay, M. E.; Fernandez, F. M.; Kubanek, J. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 7314−7319. (16) Nemes, P.; Barton, A. A.; Vertes, A. Anal. Chem. 2009, 81, 6668−6675. (17) Mueller, T.; Oradu, S.; Ifa, R. D.; Cooks, R. G.; Kraeutler, B. Anal. Chem. 2011, 83, 574−5761. (18) Urban, L. P.; Schmid, T.; Amantonico, A.; Zenobi, R. Anal. Chem. 2011, 83, 1843−1849. (19) Shrestha, B.; Vertes, A. Anal. Chem. 2009, 81, 8265−8271. (20) Watrous, J.; Hendricks, N.; Meehan, M.; Dorrestein, P. C. Anal. Chem. 2010, 82, 1598−1600. (21) Zhao, Y.; Lam, M.; Wu, D.; Mak, R. Rapid Commun. Mass Spectrom. 2008, 22, 3217−3224. (22) Kauppila, J. T.; Talaty, N.; Kuarane, T.; Kotiaho, T.; Kostiainen, R.; Cooks, R .G. Analyst 2007, 132, 868−875. (23) Zhang, I. J.; Talaty, N.; Costa, B. A.; Xia, Y.; Tao, A. W.; Cooks, R. G. Int. J. Mass Spectrom. 2011, 301, 37−44. (24) Williams, J. P.; Jennings, K. R.; Scrivens, J. H. Rapid Commun. Mass Spectrom. 2005, 19, 3643−3650. (25) Nyadong, L.; Green, M. D.; De Jesus, V. R.; Newton, P. N.; Fernandez, F. M. Anal. Chem. 2007, 79, 2150−2157. 10584

dx.doi.org/10.1021/ac301709r | Anal. Chem. 2012, 84, 10576−10585

Analytical Chemistry

Article

(61) Pettit, R. T.; Jones, L.; Harwood, L. J. Phytochemistry 1989, 28, 399−405. (62) Thomas, M. C.; Mitchell, T. W.; Harman, D. G.; Deeley, J. M.; Nealon, J. R.; Blanksby, S. J. Anal. Chem. 2008, 80, 303−311. (63) Thomas, M. C.; Mitchell, T. W.; Blanksby, S. J. J. Am. Chem. Soc. 2006, 128, 58−59. (64) Thomas, M. C.; Mitchell, T. W.; Harman, D. G.; Deeley, J. M.; Murphy, R. C.; Blanksby, S. J. Anal. Chem. 2007, 79, 5013−5022. (65) Thomas, M. C.; Mitchell, T. W.; Blanksby, S. J. Methods Mol. Biol. 2009, 579, 413−441.

10585

dx.doi.org/10.1021/ac301709r | Anal. Chem. 2012, 84, 10576−10585