Comparison and Characterization of Soybean and Sunflower

Feb 28, 2015 - The scarce availability of nongenetically modified soybeans on the world market represents a growing problem for food manufacturers. ...
6 downloads 15 Views 4MB Size
Download PDF
Article pubs.acs.org/JAFC

Comparison and Characterization of Soybean and Sunflower Lecithins Used for Chocolate Production by High-Performance ThinLayer Chromatography with Fluorescence Detection and Electrospray Mass Spectrometry Stephanie Krüger, Laura Bürmann, and Gertrud E. Morlock* Institute of Nutritional Science, Chair of Food Science, Justus Liebig University Giessen, Heinrich-Buff-Ring 26-32, 35392 Giessen, Germany S Supporting Information *

ABSTRACT: The scarce availability of nongenetically modified soybeans on the world market represents a growing problem for food manufacturers. Hence, in this study the effects of substituting soybean with sunflower lecithin were investigated with regard to chocolate production. The glycerophospholipid pattern of the different lecithin samples was investigated by high-performance thin-layer chromatography fluorescence detection (HPTLC-FLD) and by HPTLC-positive ion electrospray ionization mass spectrometry (ESI+-MS) via the TLC-MS Interface and by scanning HPTLC−matrix-assisted laser desorption ionization−timeof-flight mass spectrometry (MALDI-TOFMS). Especially, the contents of phosphatidylcholine (PC) and phosphatidylethanolamine (PE) were of interest due to the influencing effects of these two glycerophospholipids on the rheological parameters of chocolate production. The lecithin substitution led to only slight differences in the rheological parameters of milk and dark chocolate. Limits of detection (LODs) and limits of quantification (LOQs) of seven glycerophospholipids were studied for three detection modes. Mean LODs ranged from 8 to 40 mg/kg for HPTLC-FLD and, using a single-quadrupole MS, from 10 to 280 mg/kg for HPTLC-ESI+-MS as well as from 15 to 310 mg/kg for HPTLC-FLD-ESI+-MS recorded after derivatization with the primuline reagent. KEYWORDS: phosphatidylcholine, phosphatidylethanolamine, chocolate, rheology, high-performance thin-layer chromatography, mass spectrometry

■

INTRODUCTION

in chocolate production originate from soybeans. The liquid crude lecithin extract used for the production of chocolate is obtained as a byproduct of vegetable oil production.2,9 Although lecithin from soybeans is relatively cheap and easy to extract, it is difficult to purchase lecithin from nongenetically modified soybeans. Furthermore, soybean lecithin requires declaration on the packaging because of its intrinsic allergens.10,11 For these reasons, there are increasingly more efforts to move to other lecithin sources. Alternatively, lecithin extracted from sunflower seeds can be used. Although more expensive and more complex to extract, it has comparable emulsifying properties, but none of the above-mentioned negative aspects.7,12,13 The primary benefit of glycerophospholipids is the improvement of flow characteristics (rheology) of molten chocolate along with lipid crystallization and longer storage stability.12,14,15 Phosphatidylcholine (PC), phosphatidylethanolamine (PE), and phosphatidylinositol (PI) are the main glycerophospholipids found in commercially available lecithin products.13,16 The glycerophospholipids of lecithin affect the rheological parameters of molten chocolate mass in a different way. In the literature, the effects of PC and PE on the rheology of chocolate are described in detail.17 PC is known to

Glycerophospholipids are surface-active biomolecules and belong to the group of lipids.1−3 The amphiphilic molecules are composed of a glycerol backbone that is esterified with organic fatty acids of various lengths and numbers of double bonds in positions sn-1 and sn-2 and a characteristic polar phosphate ester headgroup (e.g., choline, ethanolamine, serine, and inositol) in position sn-3.1,3−5 Furthermore, they are the main components of lecithin, an important emulsifier in the food industry. The term “lecithin” stands for mixtures of polar and nonpolar lipids and, according to Commission Directive 2008/84/EC of the European Union, contains at least 60% of acetone-insoluble materials. Commercially available lecithin products are predominantly composed of glycerophospholipids (∼50%) and triacylglycerides (∼34%), with lower amounts of glyceroglycolipids and carbohydrates as well as various minor compounds (e.g., sterols, tocopherols, and free fatty acids).3 Although lecithin is an important food additive (E 322), it also has a wide range of other application possibilities such as emulsifier in feed products, pharmaceutical products, cosmetics, and the paint industry and as an industrial lubricant.2,3,6 In the chocolate industry, the emulsifying properties of glycerophospholipids are crucial. By reducing the interfacial tension between the continuous phase (cocoa butter) and the dispersive phase (particles such as sugars, cocoa, and milk powder), glycerophospholipids support the miscibility and stability of the product.7−9 The most frequently used lecithins © 2015 American Chemical Society

Received: Revised: Accepted: Published: 2893

September 12, 2014 February 28, 2015 February 28, 2015 February 28, 2015 DOI: 10.1021/jf506332f J. Agric. Food Chem. 2015, 63, 2893−2901

Article

Journal of Agricultural and Food Chemistry

Germany). Ammonia (25%, w/w) and methanol (gradient grade) were obtained from VWR (Darmstadt, Germany) and acetone (HPLC grade) from Carl Roth (Karlsruhe, Germany). Chocolate masses and lecithin samples were obtained from Stollwerck. Standard Solutions. The solvent used for the preparation of standard solutions was a mixture of chloroform and methanol (1:1, v/ v). About 0.5−1.0 mg of each glycerophospholipid standard was dissolved in 10 mL of solvent mixture (50−100 ng/μL). The standard solutions were stored in the dark at −20 °C. For limit of detection (LOD)/limit of quantification (LOQ) studies, the standard solutions were diluted 1:10 (5−10 ng/μL). Standard mixtures were generated via overspraying of the individual solutions onto the same starting zone. Lecithin Sample Preparation. About 50 mg of the different lecithin samples were extracted with a mixture of methanol and isopropyl acetate (2:3, v/v). After centrifugation at 2300g for 10 min, the supernatant was removed and stored in the dark at −20 °C. Sample Application. For quantitation, the solutions were sprayed with the Automatic TLC Sampler 4 (ATS4, CAMAG, Muttenz, Switzerland) as 8 mm bands allowing up to 20 tracks to be applied on one HPTLC plate of 20 × 10 cm (distance from lower edge 8 mm, distance from left edge 15 mm, automatic distance between bands). For calibration, 3.0−12.0 μL of the standard solutions was sprayed on the HPTLC plate (140−1200 ng/band), and sample volumes ranged between 4.0 and 12.0 μL. For LOD/LOQ determinations by HPTLCESI+-MS and HPTLC-FLD-ESI+-MS, 4 mm bands were applied to fit the dimensions of the oval elution head (4 × 2 mm). To determine the maximum sample application volume possible, volumes of the lecithin sample SN2 ranging between 60.0 and 320.0 μL were applied as rectangles (8 × 3 mm) on a cut HPTLC plate (smartCUT Plate Cutter, CAMAG). For HPTLC-MALDI-TOFMS, bands of 4 mm length were applied on HPTLC foils, later cut to the size of 5 × 7.5 cm (distances from lower and side edge as before, distance between bands 8 mm). The sample extracts (10 μL) and standard solutions of PC 34:1 and PE 34:1 (1 μg/band) were applied. Chromatography. Development was performed on HPTLC plates or foils silica gel 60 F254 with a mixture of chloroform/ methanol/water/ammonia (25%) (30:17:2:1, v/v/v/v)20 in the Automatic Developing Chamber (ADC2, CAMAG). Before development, the chamber was saturated for 20 min and the relative humidity was adjusted to 33% by a saturated magnesium chloride-6-hydrate solution for 3 min. The migration distance was 65 mm (from the lower edge of the plate). After application of the starting zones and after development, the plate was automatically dried for 0.5 and 2 min, respectively. Derivatization. For visualization of the glycerophospholipids, the developed plates were dipped into a primuline solution (100 mg of primuline in 200 mL of acetone/water 4:1, v/v) at a speed of 3 cm/s and an immersion time of 1 s using a TLC Chromatogram Immersion Device (CAMAG). The plate was dried in a stream of warm air (hair dryer at 30 cm distance from the plate) for 2 min. Densitometry and Documentation. The densitometric evaluation via fluorescence measurement at UV 366/>400 nm (mercury lamp, measurement slit size 6.0 mm × 0.2 mm, scanning speed 20 mm/s, K400 optical filter) was performed using a TLC Scanner 4 (CAMAG). Plate images were documented at 366 nm by a TLC Visualizer (CAMAG). All CAMAG instrumentation used so far as well as data obtained was processed with winCATS, version 1.4.6.2002 (CAMAG). Mass Spectrometry. The derivatized, at 366 nm, blue fluorescent glycerophospholipid bands (4 mm band length applied) were marked at UV 366 nm illumination with a soft pencil and directly eluted via the oval elution head (4 × 2 mm) of the TLC-MS Interface (CAMAG) using 100% methanol at a flow rate of 0.1 mL/min (pump of the HP 1100 ChemStation, Agilent, Waldbronn, Germany). In the tubing to the MS a C18 guard cartridge (SecurityGuard cartridge, Phenomex, Torrance, CA, USA) was integrated. A single-quadrupole MS (mass resolution of 1 unit) equipped with an ESI ion source was employed (ExpressIon CMS, Advion, Ithaca, NY, USA). For positive ionization, the capillary voltage was set to 170 V, the ESI voltage to

be the most effective glycerophospholipid with regard to the reduction of viscosity while having less of an effect on reducing the chocolate yield value. PE, on the other hand, is said to be more effective in reducing the chocolate yield value than the viscosity.17−19 Therefore, when sunflower lecithin is replaced with soybean lecithin, it is important to keep in mind that sunflower lecithin has slightly more PC and less PE. With regard to the analysis of lecithin’s glycerophospholipids, various methods have been used such as thin layer chromatography (TLC), high-performance thin-layer chromatography (HPTLC), gas chromatography (GC), and highperformance liquid chromatography (HPLC).1,20 GC was usually used for the identification of the individual fatty acids,1 wheras TLC, HPTLC, and HPLC methods were applied to separate and identify glycerophospholipid classes (assignment of headgroups by normal phase chromatography) or the fatty acid species (reversed phase (RP) chromatography, silver ion chromatography).6,21 HPLC was the most applied method because of its speed, resolution, high sensitivity, and specificity22,23 as well as hyphenation with MS. For the MS analysis soft ionization methods such as fast atom bombardment (FAB), nowadays rarely used, or electrospray ionization (ESI) and matrix-assisted laser desorption ionization (MALDI) were preferred because they allowed a direct analysis of the glycerophospholipid species.24 In recent studies, HPTLC separation was combined with MALDI−time of flight mass spectrometry (TOFMS).25−28 Recently, sialylated and sulfated glycosphingolipids as well as a total lipid extract of bovine brain were investigated by HPTLC-MS.29 Two different direct elution techniques (contactless surface sampling versus the elution head-based approach) were compared. Although it was mentioned that in the bovine brain glycerophospholipids such as PE, PI, and phosphotidylserine (PS) were detected, no further details were given. Hence, the aim of this study was to investigate the interchangeability of soybean and sunflower lecithin with regard to changes in the rheology of different chocolate masses. Occurring changes in rheology were compared with glycerophospholipid differences of the lecithin samples. The glycerophospholipid pattern and content of the different lecithin samples used in chocolate production were analyzed by HPTLC-FLD and further characterized by HPTLC-ESI+MS and HPTLC-FLD-ESI+-MS. LODs and LOQs were determined for seven glycerophospholipids by the three analytical techniques for the first time. HPTLC-MALDITOFMS was employed for confirmation.

■

MATERIALS AND METHODS

Materials and Reagents. Two sunflower lecithin samples from different suppliers (SN1 and SN2) and one soybean lecithin sample (SJ) were obtained from Stollwerck (Cologne, Germany). The standards PE 34:1, PC 34:1, lyso-phosphatidylcholine (LPC 18:0), lyso-phosphatidylethanolamine (LPE 18:0), phosphatidylserine (PS 36:4), and phosphatidic acid (PA 36:4) were all purchased from Avanti Polar Lipids (Alabasta, AL, USA), except PE 32:0, which was bought from Sigma-Aldrich (Steinheim, Germany). For HPTLC-FLD and HPTLC-ESI+-MS as well as HPTLC-FLD-ESI+-MS, HPTLC plates silica gel 60 F254 (20 × 10 cm), and for HPTLC-MALDI-TOFMS, HPTLC foils silica gel 60 F254 (20 × 20 cm) were used (all Merck, Darmstadt, Germany). 2,5-Dihydroxybenzoic acid (DHB) was also provided by Merck. Chloroform (>99.5%, stabilized with 1% ethanol), isopropyl acetate (>99.0%), acetonitrile (≥99.9%), trifluoroacetic acid (≥98.0%), ammonium phosphate (≥99.0%), and primuline (≥50%, as sodium salt) were purchased from Sigma-Aldrich (Steinheim, 2894

DOI: 10.1021/jf506332f J. Agric. Food Chem. 2015, 63, 2893−2901

Article

Journal of Agricultural and Food Chemistry

Figure 1. HPTLC chromatograms of (A) all glycerophospholipid standards available (tracks 1 and 2) and (B) PC 34:1 and PE 34:1 with different volumes applied (S2−S4) beside samples of soy lecithin (SJ) and sunflower lecithins (SN1 and SN2) as well as (C) an HPTLC 3 D densitogram. 3500 V, and the source voltage to 25 V. The nebulizer gas pressure was 60 psi, the temperatures of the capillary and the source gas were set to 250 °C, and the dissolvation gas flow rate was 4 L/min. The mass spectra were recorded in the full scan mode (total ion current (TIC) chronogram) m/z 100−1000 using a detection gain of 1100. Data processing for MS measurement was carried out with Mass Express 1.1.22.15, and evaluation was done with Data Express 1.1.22.15 (both Advion). A plate background at a migration distance comparable to the analyte zone was subtracted from the analyte spectrum. For LOD/ LOQ determination by HPTLC-FLD-ESI+-MS, the x,y-coordinate for elution was obtained after fluorescence measurement at UV 366/>400 nm, as such a low substance amount was not visible in the UV 366 nm image. For HPTLC-ESI+-MS of nonderivatized zones, the x,ycoordinate for elution was received by extrapolation from the respective derivatized band on a cut plate strip (y-coordinate) and from the respective starting zone (x-coordinate). For HPTLC-MALDI-TOFMS,28,30 the HPTLC foil was homogeneously coated with a DHB solution (8 g in 40 mL of a mixture of acetonitrile and 10 mM aqueous ammonium phosphate solution with 0.1% trifluoroacetic acid 9:1, v/v). Therefore, the foil was dipped into the DHB solution at a speed of 4.5 cm/s and an immersion time of 1 s using the TLC Chromatogram Immersion Device (CAMAG). For the first 90 s the solvent was allowed to evaporate in the air, followed by 90 s of drying with a hair dryer. All steps were repeated, but the last drying step was extended to 4 min. The foil’s reverse side was cleaned with a tissue, and the foil was cut to a size of 5 × 7.5 cm using a scissors. The cut foil part was mounted into a TLC-MALDI adapter target (Bruker Daltonics, Bremen, Germany). The HPTLC-MALDITOF mass spectra were acquired in the positive mode on an Ultraflex 1 TOF/TOFMS (Bruker Daltonics), which had a mass accuracy of about 75 ppm. Rheological Measurements of the Chocolate Masses. Three different chocolate masses (dark, milk, and white) were manufactured without emulsifiers. Of the regular soybean lecithin sample SJ the

original concentration according to the recipe (original-%) was added, whereas for the sunflower samples SN1 and SN2 three different concentrations (original-% as well as ±0.1%) were added to each of the three base chocolates by a defined manual mixing process.31 The rheology of the chocolate samples was measured at 40 °C with a RheolabQC viscosimeter (Anton Paar, Graz, Austria) using a DIN cup and bob geometry. After a defined conditioning step including 30 s of preshearing at 5 s−1, the viscosity measurements were carried out with increasing shear rates from 2 to 100 s−1 and recording of 29 data points. The rheological parameters viscosity and yield value were taken by fitting the Casson equation to the data.32,33 Each analysis was executed in triplicate (n = 3). Statistical Analysis. Statistical analysis was done with Microsoft Excel, version Office Standard 2010. The respective means, standard deviations, and relative standard deviations were calculated. The twosample t test (independent, two-sided, P = 95, 99, 99.9%) was applied to test the different lecithins for significant differences in the PC and PE contents.

■

RESULTS AND DISCUSSION The HPTLC-FLD method allowed a satisfying separation of seven glycerophospholipid standard substances (available in our institute), which were divided in two mixtures (Figure 1A). The standard compound PC 34:1 showed a hRF value of 30, whereas both of the PE standards (PE 32:0 and PE 34:1) had higher hRF values (hRF = 41 and 43). Comparison of the Lecithin Fingerprints. The separation of the lecithin samples’ glycerophospholipids showed equally good results (Figure 1B). The brilliance of the blue fluorescence increased with the concentration. The lecithins from both plants had a comparable glycerophospholipid pattern, with PC (hRF = 31) and PE (hRF = 43) as the main 2895

DOI: 10.1021/jf506332f J. Agric. Food Chem. 2015, 63, 2893−2901

Article

Journal of Agricultural and Food Chemistry

pure, an additional ion at m/z 319 was observed as base peak (S/N 50 in EIC chronogram). This ion was assumedly assigned as the desodiated fragment of primuline, named 2-(4aminophenyl)-6-methyl-1,3-benzothiazole-7-sulfonate [M− C7H3NS−Na]−). Because the primuline forms negative ions, the phospholipid measurement was performed in the positive ionization mode. Thus, the derivatization reagent did not impair the glycerophospholipid mass signal and eased its interpretation. Using a single-quadrupole MS, the glycerophospholipid bands in the standard mixture were clearly detected at 400 ng/band (S/N 110 and S/N 310 for the base peaks [M +Na]+ of PE 34:1/32:0 and PC 34:1, respectively). Thus, the positive ionization mode was found to be a good compromise, although the negative ionization might be more sensitive for some glycerophospholipids. Characterization by HPTLC-FLD-ESI+-MS. The preliminary assignment of the head groups (PC and PE) of the unknown glycerophospholipid bands in the lecithin samples was based on the hRF values. For characterization of their fatty acyl residues, the ESI+ mass spectra of the glycerophospholipid bands found were recorded. For comparison, the mass spectra of the PC and PE standards were recorded, too. The identities of the standards PE 32:0 and PE 34:1 as well as PC 34:1 were confirmed due to their characteristic sodium adducts [M+Na]+ and [M+2Na−H]+(Supporting Information Figure S1). The mass spectra of the respective glycerophospholipid bands preliminarily assigned as PC in the lecithin samples SJ, SN1, and SN2 (12.0 μL applied, Figure 3) showed a mass signal pattern comparable to the PC standard. Two molecular species with mass signals at m/z 782 [PC1]+ and m/z 758 [PC2]+ were evident. This suggested the presence of PC 36:4 (PC1) and PC 34:2 (PC2). These results corresponded with the literature, in which sn-1,2-dilinoleoyl-PC and sn-1palmitoyl-2-linoleoyl-PC were reported in soybean lecithin.27 The respective sodium adducts were observed at m/z 804 [PC1+Na]+ and m/z 780 [PC2+Na]+. Furthermore, the mass signal at m/z 564 was explained by fragmentation (fatty acid elimination), that is, elimination, of linoleic acid [PC1−C18:2]+ and palmitic acid [PC2−C16:0]+, respectively. With regard to the PE preliminarily assigned bands in the lecithin samples, all mass spectra showed a mass signal pattern (Figure 4) comparable to that of the PE standard. According to Stübinger et al.,27 the mass signals were assigned as PE 36:4 (PE1) and PE 34:2 (PE2). The obtained mass signals showed predominantly adducts, that is, the respective sodium adducts at m/z 762 [PC1+Na]+ and m/z 738 [PC2+Na]+ as well as the disodium adducts at m/z 784 [PC1+2Na−H]+ and m/z 760 [PC2+2Na − H]+. It can be concluded that the PE and PC present in the lecithin samples were most likely esterified with palmitic and linoleic acid. The latter is the main fatty acid in soybean and sunflower oil.13,38 In oils derived from plants, the unsaturated fatty acid (here, linoleic acid) is predominately located at the sn-2-position of the glycerophospholipid.3,27 For an unequivocal identification of the fatty acyl residues, additional MS/MS measurements and also measurements in the negative mode could be helpful, especially for samples in which the glycerophopholids and their fatty acyl pattern are not yet described in such detail. For the purpose of the current study, however, the simple identification in the positive ionization mode with a single-quadrupole MS and comparison with data from the literature were sufficient.

glycerophospholipids. Other glycerophospholipid bands visible in the samples (hRF = 12, 21, 36, 54, and 84) could not definitively be assigned to any standard compound available in our institute. Thus, the glycerophospholipid zones found were subjected to a subsequent characterization by MS (see Characterization by HPTLC-FLD-ESI+-MS). As PC and PE were identified to have the highest influence on rheology,17−19 quantitation was focused on these two compounds. Quantitation of PC and PE in the Lecithin Samples. With regard to the quantitative analysis of PC and PE in soybean and sunflower lecithins, the calibration was performed in the ranges of 140−550 ng/band for PC 34:1 and 300−900 ng/band for PE 34:1 (Figure 1C). If required, the working ranges can be expanded; however, they were sufficiently wide for the given analysis. The mean repeatabilities (%RSD, n = 2, instant repetition within a laboratory) for both compounds in the lecithin matrix were mostly ≤5.0%, whereas the intermediate precisions (%RSD, n = 2, repetitions within a laboratory over several weeks) were mostly ≤10.0% (Supporting Information Table S1). After statistical analysis, the mean PC and PE contents were found to be different in soybean and sunflower lecithins (Figure 2). Both sunflower lecithin samples

Figure 2. PC 34:1 and PE 34:1 contents (mean ± standard deviation) of one soybean (SJ) and two sunflower (SN1 and SN2) lecithin samples. (∗) Statistically, (∗∗) significantly, and (∗∗∗) highly significantly different compared to SJ (P = 95, 99, 99.9%).

showed significantly lower values for PE than the soybean lecithin, currently used for chocolate production. For both soybean and sunflower lecithin samples, the mean PC content was about 20 ± 2 mg/g (2.0%), whereas the results for PE showed a wider content range between 14 and 22 mg/g (1.4− 2.2%). This was reflected in the ratios PC/PE. The soybean lecithin sample (SJ) had a PC/PE ratio of 1.0. In contrast, the two sunflower samples (SN1 and SN2) showed higher ratios of 1.2 and 1.3, respectively (Table S1). According to the independent, two-sided t test, there were differences in the glycerophospholipid content between SJ and either SN1 (calculated t value t1 = 10.36 for PC and t1 = 4.76 for PE) or SN2 (calculated t value t2 = 5.55 for PC and t2 = 2.52 for PE) (Figure 2). Selection of the Ionization Mode of HPTLC-FLD-ESIMS. For HPTLC-FLD-ESI-MS, the derivatized, blue fluorescent zones of interest were eluted via the TLC-MS Interface as described.30 The derivatization reaction with the primuline reagent was based on a physisorption (only interaction with the analyte due to van der Waals forces), and during the MS measurement the primuline molecule was separated from the glycerophospholipid molecule. The primuline was observed in the negative ionization mode as a desodiated molecule at m/z 452 [M−Na]− (S/N 24 in the extracted ion current (EIC) chronogram). As the primuline substance bought was only 50% 2896

DOI: 10.1021/jf506332f J. Agric. Food Chem. 2015, 63, 2893−2901

Article

Journal of Agricultural and Food Chemistry

Figure 3. HPTLC-FLD-ESI+-MS: recorded mass spectrum of unknown zones at hRF 31 in the lecithin samples (SJ, SN1, and SN2, application volume 12 μL) and confirmation of the preliminary assignment as PC, specified as PC 36:4 (PC1) and PC 34:2 (PC2).

Figure 4. HPTLC-FLD-ESI+-MS: recorded mass spectrum of unknown zones at hRF 42 in the lecithin samples (SJ, SN1, and SN2, application volume 12 μL) and confirmation of the preliminary assignment as PE, specified as PE 36:4 (PE1) and PE 34:2 (PE2).

2897

DOI: 10.1021/jf506332f J. Agric. Food Chem. 2015, 63, 2893−2901

Article

Journal of Agricultural and Food Chemistry

Figure 5. HPTLC-FLD-ESI+-MS of six lower concentrated glycerophospholipid zones of the sunflower sample SN2 (zones 1−6, application volume 60 μL).

applied influences the calculation of the LOD/LOQ. As the spray-on application (as aerosol) includes a concentration step, the selected application volume determines the LOD/LOQ obtained. The rule is the higher the application volume, the lower the LOD. Thus, for the calculation of the LOD and LOQ, increased sample volumes (between 60 and 320 μL) were applied to get an impression of how the matrix influences the separation and how high a maximal application volume can be. Even at the highest investigated sample volume of 320 μL, the phospholipid zones were still clearly distinguishable (Supporting Information Figure S2). Therefore, it was possible to apply at least 320 μL to achieve low LODs. Thus, the highest experimentally proven sample volume (320 μL) was taken into account for the calculation of the LOD and LOQ values. However, it is important to keep in mind that for the LOD and LOQ analyses by HPTLC-ESI+-MS and HPTLC-FLD-ESI+MS, 4 mm instead of 8 mm bands were applied to ensure that the glycerophospholipid zones were eluted as completely as possible. Therefore, by halving the application band, the maximum application volume that was experimentally proven (320 μL) was also halved to be 160 μL. The mean LODs (S/N 3) of HPTLC-FLD-ESI+-MS were 60 and 45 mg/kg for PE 34:1 and PE 32:0, respectively. For PC 34:1, the mean LOD was 15 mg/kg. Mean LODs and LOQs of the seven glycerophospholipids investigated were mostly comparable, although slightly higher for most of the nonderivatized bands

Additionally, mass spectra were recorded from the SN2 sample’s lower concentrated fluorescent zones, clearly visible on the plate at higher application volumes (Supporting Information Figure S2). Several unknown minor glycerophospholipid zones were tentatively assigned, for example, LPC 18:2, PA 34:1, and 36:2, phosphatidylglycerol (PG) 34:2, and PI 36:3 (Figure 5).39 However, the zone 3 (hRF 36) just above PC could not be assigned unambiguously. Although the signal at m/z 963 would be consistent with both a triacylglycerol (C63H110O6) and PI (43:1) or PI (44:8), the zone’s position on the plate seemed to be implausible for a triacylglycerol, and the fatty acyl pattern of PI fits to neither soybean nor sunflower lecithins. Whereas for a final identification high-resolution MS would be the logical next step, this still demonstrated that mass spectra were obtainable also from minor glycerophospholipids and that it is a matter of the application volume in HPTLC to obtain clear mass signals (here, 60 μL applied). Mean LODs and LOQs of Seven Glycerophospholipids via HPTLC-FLD, HPTLC-ESI+-MS, and HPTLC-FLD-ESI+MS. So far, no literature has reported LOD and LOQ values for glycerophospholipids via HPTLC-FLD, HPTLC-ESI+-MS, and HPTLC-FLD-ESI+-MS, so it was unclear how sensitive potential phospholipids in samples could be detected. Thus, LOD and LOQ values34 were studied for three modes of detection, that is, via HPTLC-FLD, HPTLC-ESI+-MS, and HPTLC-FLD-ESI+-MS. In HPTLC, the application volume 2898

DOI: 10.1021/jf506332f J. Agric. Food Chem. 2015, 63, 2893−2901

Article

Journal of Agricultural and Food Chemistry

Table 1. Mean LOD and LOQ Valuesa (n = 3) of Glycerophospholipids for HPTLC-FLD after Derivatization with Primuline and for HPTLC-ESI+-MS with and without Derivatization (Each via Peak Height (H) and Area (A) for a Given Application Volume of 320 μL on an 8 mm Band or 160 μL on a 4 mm Band) ESI+-MSb

FLD derivatized (mg/kg)

PA 36:4 LPE 18:0 PC 34:1 PE 34:1 LPC 18:0 PS 36:4 PE 32:0

LOD

LOQ

H/A

H/A

8/