Article pubs.acs.org/JAFC
Improving Method Reliability in Carotenoid Analysis through Selective Removal of Glycerolipid Interferences by Lipase Treatment Sebastiaan Bijttebier,*,†,§ Els D’Hondt,† Bart Noten,† Nina Hermans,§ Sandra Apers,§ and Stefan Voorspoels† †
Business Unit Separation and Conversion Technology (SCT), Flemish Institute for Technological Research (VITO), Boeretang 200, 2400 Mol, Belgium § NatuRA, University of Antwerp, Universiteitsplein 1, 2610 Antwerp, Belgium ABSTRACT: Saponification is most often used to hydrolyze glycerolipid interferences during carotenoid analysis. Ester bonds of other plant metabolites such as carotenoids are, however, also hydrolyzed during saponification, thus altering the natural carotenoid composition. A straightforward and selective cleanup procedure was therefore developed involving the enzymatic hydrolysis of matrix glycerolipids. The optimized procedure (100 μL of extracted vegetable or algal oil in 20 mL of 50:50 phosphate buffer/methanol with 25 μL of sodium n-octyl sulfate, 30 mg of bile salts, and 250 μL of NaCl solution (5 mM), magnetic stirring for 2 h at 40 °C with 1 mL of Lipozyme TL 100 L and 1 mL of Lipozyme CALB L) removed the greater part of triglycerides (94.8−100.0%) and diglycerides (88.2−99.8%) while preserving the natural carotenoid composition. KEYWORDS: carotenoid LC-MS analysis, ionization suppression, lipase treatment, selective glycerolipid hydrolysis, esterified carotenoids
■
INTRODUCTION Carotenoids consist of a large group of natural compounds that are found predominantly in photosynthesizing organisms such as green plants, algae, and some bacteria. Carotenoids play a key role in photosynthesis and are identified as potentially important natural antioxidants that might aid in the prevention of several human chronic degenerative diseases, such as cancer, cardiovascular diseases, and age-related eye diseases.1−9 Characterization of carotenoids in biological matrices is a crucial step in any research valorization trajectory. Carotenoids can be divided into two main groups, namely, the carotenes, which consist of only carbon and hydrogen, and the xanthophylls, which contain functional groups. In nature, these functional groups can occur esterified with fatty acids, which leads to different physicochemical properties.10 Several feeding studies of chicken have shown that best absorption is observed with free xanthophylls.10 The solubility is also different between free and esterified compounds. Free plant stanols are, for instance, esterified with fatty acids to increase their solubility in margarine emulsions so that more plant stanols can be added per unit of margarine.11 Knowledge of the natural occurrence of carotenoids (free, esterified, glycosylated) can therefore contribute to related health research. However, accurate analytical methods are needed that allow differentiation between these different forms. Such a method would also allow differentiation between natural and synthetic carotenoids, for example, in food supplements. Astaxanthin obtained from a natural resource such as Haematococcus pluvialis contains a high percentage of astaxanthin esterified with fatty acids, whereas free astaxanthin can be obtained by synthesis.12,13 There are still concerns about the biological functions and food safety of synthetic astaxanthin as synthetic astaxanthin is different in isomerism and chemical structure from natural astaxanthin.14 © 2014 American Chemical Society
Analysis of complex carotenoid mixtures is not straightforward. To date, even the most state-of-the-art detectors are often not capable of distinguishing between different analytes present in a sample. Therefore, compound separation has to be maximized before detection. Reversed-phase high-performance liquid chromatography (RP-HPLC) is most often used as a separation technique for carotenoids because of its ease of use and high selectivity.15 Carotenoids are usually detected by photodiode array (PDA) or ultraviolet (UV) detectors because their polyene backbone enables their detection at very selective wavelengths (typically 400−500 nm).15 However, not much structural information can be obtained by using PDA detectors, and it can thus be very difficult or even impossible to identify an unknown compound solely on the basis of its PDA spectrum. Mass spectrometric (MS) detectors are much more selective than PDA detectors and can therefore facilitate the identification of carotenoids. Since a few years, orbital trap or Orbitrap MS detectors have become commercially available. Orbitrap detectors are screening detectors that can routinely generate mass spectra with a resolving power up to 100,000 at full width half-maximum (fwhm) and mass accuracies within 2 ppm (ppm). High resolution and exact mass MS screening detectors enable the calculation of the most probable molecular formulas of the generated ions, thus facilitating identification of unknown compounds. Several soft ionization techniques are used for carotenoid detection, most often atmospheric pressure chemical ionization (APCI) and electrospray ionization (ESI) because of their ease of use and high dynamic range.16,17 High levels of triglycerides can however induce ionization Received: Revised: Accepted: Published: 3114
December March 14, March 17, March 17,
9, 2013 2014 2014 2014
dx.doi.org/10.1021/jf405477s | J. Agric. Food Chem. 2014, 62, 3114−3124
Journal of Agricultural and Food Chemistry
Article
chromatography, sodium hydrogen carbonate, potassium phosphate, and sodium chloride for analysis were purchased from Merck (Darmstadt, Germany). Ammonium acetate, (D-Ala2)-leucin enkephalin, bile salts, and butylated hydroxytoluene (BHT) were purchased from Sigma-Aldrich (Bornem, Belgium). Commercially available mixtures to calibrate the mass spectrometer, namely, MSCAL5-1EA (caffeine, tetrapeptide “Met-Arg-Phe-Ala”, Ultramark) for positive ion mode and MSCAL6-1EA (sodium dodecyl sulfate, taurocholic acid sodium salt, Ultramark) for negative ion mode, were purchased from Supelco (Bellefonte, PA, USA). Capsanthin, astaxanthin dipalmitate, and astaxanthin were purchased from Carotenature (Ostermundigen, Switzerland). β-Carotene and β-sitosterol were purchased from Sigma-Aldrich (Bornem, Belgium). Paprika oleoresin 65000 CU (900 g) was obtained from Natural Spices (Mijdrecht, The Netherlands), and Asta Krill 8 (60 softgels of approximately 1 g) was obtained from Swanson vitamins (Fargo, ND, USA). The content of the Asta Krill 8 soft gels consists of Euphausia superba Antarctic krill oil combined with an astaxanthin-rich extract from the microalga Haematococcus (as stated by the supplier). The food grade enzyme formulations Lipozyme TL 100 L (100 KLU/ g) and Lipozyme CALB L (5 KLU/g) were kindly provided by Novozymes (Bagsvaerd, Denmark). Sodium lauryl sulfate, glycerol dioleate, and glycerol trioleate were obtained from Chem Service (West Chester, PA, USA). Extra virgin olive oil was obtained from a local store. Preparation of Standard Solutions. Standard stock solutions and working solutions were prepared for each analyte separately at a concentration of approximately 200 μg/mL. Stock solutions for glycerolipids were prepared in methanol. The stock solution of βsitosterol was prepared in methanol + 0.1% BHT. Stock solutions for carotenoids were prepared in DCM + 0.1% BHT. Standard stock and working solutions were stored at −25 °C in the dark under an inert atmosphere (nitrogen). Dilutions of the stock solutions were prepared in DCM + 0.1% BHT for analysis. Instrumental Analysis. All analyses were performed with an ultrahigh-performance liquid chromatography−photodiode array− accurate mass mass spectrometry (UHPLC-PDA-am-MS) configuration. The LC system consisted of an Accela quaternary solvent manager, a “Hot Pocket” column oven (Thermo Fisher Scientific, Bremen, Germany), and a CTC PAL autosampler (CTC Analytics, Zwingen, Switzerland). The LC system was hyphenated to an Accela PDA detector and an Orbitrap mass spectrometer (Exactive, Thermo Fisher Scientific, Bremen, Germany). The method of analysis was previously described by Bijttebier et al.17 Briefly, 1.25 μL was injected on a thermostatically heated (35 °C) Acquity UPLC HSS C18 SB column (2.1 mm × 100 mm, 1.8 μm; Waters, Milford, MA) and chromatographically separated by gradient elution. The mobile phase solvents consisted of 50:22.5:22.5:5 (v/v/v/ v) water + 5 mM ammonium acetate/methanol/acetonitrile/ethyl acetate (A) and 50:50 (v/v) acetonitrile/ethyl acetate (B), and the gradient was set as follows (min/%A): 0.0/90, 0.1/90, 0.8/70, 20.0/9, 20.1/0, 20.4/0, 20.5/90, 23.0/90. APCI was used as MS ionization technique, and polarity switching was applied. The scanning range was set to m/z 90−1400 and a mass resolution of 50,000 fwhm was applied. The corona discharge current was set at 5 μA, the vaporizer and capillary temperatures were set at 450 °C, and a capillary voltage of ±25 V was used for both positive and negative APCI. Lock mass correction with (D-Ala)2-leucin enkephalin was applied, and mass deviations of no more than 2 ppm were observed. The PDA detector was set to scan from 190 to 800 nm. Xcalibur software (Thermo Fisher Scientific) was used to process the data. Sample Preparation. Red paprika fruits (Capsicum annuum) were bought in a local store. After removing the stems and kernels, they were cut into pieces, frozen with liquid nitrogen and immediately freeze-dried. After drying they were milled and homogenized with a Grindomix GM 200 from Retsch (Haan, Germany) at 10,000 rpm and stored in the dark under nitrogen at −25 °C. Before sample preparation, the sample was allowed to equilibrate to room temperature. To 20 g of freeze-dried paprika were added 20 g of sodium hydrogen carbonate and 330 mL of 70:30 acetone/methanol +
suppression in positive ionization mode. Esterified xanthophylls tend to elute together with the bulk of triglycerides during RPLC (e.g., on a C18 column), and this can hamper correct quantification or even identification. Triglycerides do not ionize in negative APCI mode and one could therefore suggest using negative APCI mode instead. However, some carotenoids do not ionize in negative ionization mode.17 For proper identification and quantification of esterified carotenoids in samples with high lipid content, it is therefore essential that these lipids are removed before detection. Saponification is most often used to hydrolyze glycerolipid interferences during carotenoid analysis. Ester bonds of other plant metabolites such as carotenoids are however also hydrolyzed during saponification, thus altering the natural carotenoid composition.18 Preparative column chromatography has been used to separate lipids from carotenoids.18 This procedure is, however, tedious if performed manually and expensive if performed automatically. Beyond that, this separation procedure has to be optimized for every different matrix as the composition of glycerolipids and carotenoids is matrix dependent. Triacylglycerol hydrolases or lipases are an important group of biotechnologically relevant enzymes that find immense applications in food, dairy, detergent, and pharmaceutical industries. Lipases act under aqueous conditions on the carboxyl ester bonds present in triacylglycerols to liberate fatty acids and glycerol. The reaction is catalyzed at the lipid− water interface.19 Lipase enzymes have, however, not only been tested for the cleavage of triacylglycerides. Several publications have been written on the use of lipase enzymes for the hydrolysis of carotenoid esters to their free form.10,20 Conversion rates from esterified to free carotenoids ranged from 0 to 89% and are largely dependent on the type of enzyme. As free carotenoids have been reported to have a better bioavailability, the aim of these studies was to create more elegant alternatives for chemical saponification of marigold flower extracts and red paprika extracts on an industrial scale. Breithaupt et al. were the first to report the use of lipase enzymes for the selective removal of triacylglycerides for the identification of carotenoid esters in potatoes.21 During the first cleanup step, open bed column chromatography was used to obtain a fraction containing the carotenoid esters. This fraction was then further purified by the selective enzymatic cleavage of residual triacylglycerides with two commercial lipase enzymes. To our knowledge this is the only peer-reviewed publication written on the selective removal of triacylglycerides with the emphasis on the preservation of the natural carotenoid ester composition. As this procedure involves a fractionation step, the full natural carotenoid profile is, however, lost. Furthermore, open bed column chromatography is often laborious and tedious. Therefore, the aim of this study was to optimize this method to an efficient one-step enzymatic cleanup method without prefractionation to simplify the method and to preserve the natural carotenoid profile of the samples.
■
MATERIALS AND METHODS
Chemicals. LC-MS grade methanol, acetonitrile, and ethyl acetate were purchased from Biosolve (Valkenswaard, The Netherlands). Ultrapure water with a resistivity of 18.2 MΩ·cm at 25 °C was generated with a Millipore system. Dichloromethane (DCM) for gas chromatography, n-hexane for gas chromatography, acetone for gas 3115
dx.doi.org/10.1021/jf405477s | J. Agric. Food Chem. 2014, 62, 3114−3124
Journal of Agricultural and Food Chemistry
Article
Table 1. Summary of the Accurate Masses Obtained for the MS Protonated and Deprotonated Molecules and Product Ions in Positive and Negative APCI Mode and the Maximum Absorbances, Used for the Identification of the Compounds Described Throughout the Paper positive APCI mode sample matrixa
component
molecular formula
retention time (min)
1
β-caroteneb
C40H56
16.5
1
C64H100O5
1 1
capsanthin diester (C12:0, C12:0) capsanthin diester (C12:0, C14:0) capsanthin diester (C14:0, C14:0) capsanthin diester (C14:0, C16:0) astaxanthinb astaxanthin monoester (C18:4) astaxanthin monoester (C18:3) astaxanthin monoester (C18:2) astaxanthin monoester (C18:1) astaxanthin diester (C18:2, C18:3) diglyceride (C18:3, C18:3) diglyceride (C18:2, C18:3)
1, 3 1, 3
1 1 1 2 2 2 2 2 2
UVmax
detected protonated molecules
in source fragments
537.445
445.382
536.439
19.29
949.764
749.587; 549.409
948.757
C66H104O5
19.94
463
977.795
749.587; 549.409
976.788
C68H108O5
20.53
473
1005.830
777.617
1004.820
776.611; 227.202
C70H112O5
21.05
462
1033.860
1032.850
776.611; 548.400
C40H52O4 C58H78O5
7.85 14.25
481 478
597.393 855.591
805.648; 777.617; 549.409 579.383 579.383; 837.581
596.386 854.584
504.324 578.376
C58H80O5
14.93
481
857.606
579.383; 839.597
856.600
578.376
C58H82O5
15.68
481
859.622
579.383; 841.612
858.616
578.376
C58H84O5
16.55
481
861.638
579.383; 835.623
860.631
578.376
C76H110O6
18.75
480
1119.840
1118.830
c
613.482 615.498
C39H64O5 C39H66O5
11.5 12.4
nd nd
diglyceride (C18:2, C18:2) diglyceride (C18:1, C18:2)
C39H68O5 C39H70O5
13.31 14.36
nd nd
617.513 619.529
3 2
diglyceride (C18:1, C18:1) diglyceride (C16:1, C14:0)
C39H72O5 C33H62O5
15.58 13.08
nd nd
621.544 539.467
1
diglyceride (C18:3, C16:0)
C37H66O5
13.49
nd
591.498
3
diglyceride (C18:2, C16:0)
C37H68O5
14.43
nd
593.513
2
diglyceride (C18:1, C14:0)
C35H66O5
14.46
nd
567.498
2, 3
diglyceride (C18:1, C16:0)
C37H70O5
15.83
nd
595.529
1
triglyceride (C18:2, C18:2, C18:2) triglyceride (C18:1, C18:2, C18:2) triglyceride (C18:1, C18:1, C18:2) triglyceride (C18:1, C18:1, C18:1) triglyceride (C18:1, C18:1, C18:0) triglyceride (C18:2, C18:2, C16:0) triglyceride (C18:1, C18:2, C16:0) triglyceride (C18:1, C18:1, C16:0) triglyceride (C18:1, C14:0, C16:0) triglyceride (C14:0, C16:1, C16:0) β-sitosterolb β-sitosteryl glucoside β-sitosteryl linolenoyl glucoside (C18:3) β-sitosteryl linoleoyl glucoside (C18:2)
C57H98O6
19.7
nd
879.743
C57H100O6
20.36
nd
881.758
C57H102O6
21.0
nd
C57H104O6
21.53
C57H106O6
1, 3 3 3 1 1, 2, 3 3 2 2 1 1 1 1
in source fragments
430; 455; 482 460
839.597; 841.612; 560.364 595.471; 335.258 597.487; 337.273; 335.258 599.503; 337.274 601.518; 337.273; 339.289 603.534; 339.289 521.456; 311.258; 285.242 573.487; 313.273; 335.258 575.503; 313.273; 337.274 549.487; 339.289; 285.243 577.518; 339.289; 313.273 599.502
1
negatve APCI mode detected radical ions or adducts
840.606; 838.591
nd nd
277.217 279.233; 277.217
nd nd
279.233 279.232; 281.248
nd nd
339.289 227.201; 253.217
nd
277.217; 255.233
nd
nmd
nd
281.248; 227.201
nd
255.233; 281.248
nd
279.233
601.518; 599.503
nd
279.233; 281.248
883.774
603.534; 601.518
nd
279.233; 281.248
nd
885.790
603.534
nd
nm
22.16
nd
887.806
605.549; 603.534
nd
nm
C55H98O6
20.48
nd
855.743
599.502; 575.503
nd
279.233; 255.233
C55H100O6
21.3
nd
857.759
nd
C55H102O6
21.66
nd
859.774
577.519; 575.503; 601.519 577.518
255.233; 281.248; 279.233 nm
C51H96O6
21.45
nd
805.728
nd
C49H92O6
20.79
nd
777.696
C29H50O C35H60O6 C53H88O7
15.36 8.13 17.0
nd nd nd
nd nd nd
577.519; 549.488; 523.472 549.488; 523.472; 521.457 397.383 397.383 397.383
nd 635.453e 895.667e
C53H90O7
18.09
nd
nd
397.383
897.682e
3116
nd
nd
281.248; 255.233; 227.201 255.233; 227.201; 253.217 575.432
dx.doi.org/10.1021/jf405477s | J. Agric. Food Chem. 2014, 62, 3114−3124
Journal of Agricultural and Food Chemistry
Article
Table 1. continued positive APCI mode sample matrixa 1 1
component β-sitosteryl oleoyl glucoside (C18:1) β-sitosteryl palmitoyl glucoside (C16:0)
detected protonated molecules
negatve APCI mode detected radical ions or adducts
molecular formula
retention time (min)
C53H92O7
19.42
nd
nd
397.383
899.698e
C51H90O7
20.06
nd
nd
397.383
873.682e
UVmax
in source fragments
in source fragments
a 1, paprika oleoresin; 2, Asta Krill 8 soft gels; 3, in-house paprika extract with olive oil spike. bIdentification with standard. cnd, not detected. dnm, not measured. eAcetate adduct.
Figure 1. Representative structures of the different compound groups. 0.1% (v/v) BHT. The bottle was flushed with N2 gas and closed. After shaking for 30 min and subsequent centrifugation for 5 min at 5000 rpm, the solvent was transferred into a round-bottom flask. This extraction procedure was repeated two more times. The three supernatants were pooled, evaporated, and dissolved in 20 mL of DCM + 0.1% BHT. The extract was stored in the dark under nitrogen at −25 °C. For the lipase treatment, an aliquot of the extract was evaporated and dissolved in the solvent of the treatment. A commercially available paprika oleoresin was bought and stored at 4 °C. Before hydrolysis with lipase enzymes, the bottle was shaken vigorously by hand for 30 s and an aliquot was taken for enzymatic hydrolysis. The Asta Krill 8 soft gels were bought and stored at 4 °C. The oil present in the Asta Krill 8 soft gels was obtained by cutting the soft gels open. This oil was subsequently used for lipase assays. Lipase Efficiency and Selectivity Determination. A UV wavelength of 440 nm was chosen to compare carotenoid profiles in the samples before and after lipase treatment without interferences from glycerolipids because di- and triacylglycerides do not absorb light in this region. To compare the lipid profile before and after enzymatic hydrolysis, the total ion current (TIC) signal obtained by the Orbitrap MS detector in positive APCI mode was used. Due to the high lipid content of the samples used during this study, lipids dominated the TIC chromatograms before the enzyme assay. Identification of glycerolipids was achieved with the protonated molecules and insource product ions generated during am-MS detection. The order of the fatty acids bound to glycerol (i.e. sn-1, sn-2, or sn-3 position of the glycerol backbone) could not be deduced from the spectra during the experiments, and the lipids were therefore noted as “TG(C18:2, C18:2, C16:0)” instead of “TG(C18:2/C18:2/C16:0)”. The decline of the level of individual lipids by lipase treatment was monitored with the extracted ion chromatograms (EIC) of selected ions. Loss in ionization efficiency during the elution of high lipid concentrations can lead to an underestimation of the percentage of lipids hydrolyzed.22
Lipid hydrolysis rates are therefore most probably higher in practice than described in this paper. The spectral and chromatographic data used for compound identification are listed in Table 1. Mass accuracies of
