Transcript
Analytical, Nutritional and Clinical Methods Section Simultaneous HPLC analysis of fat-soluble vitamins in selectedanimal products after small-scale extraction P. Salo-Va Èa Èna Ènen a, *, V. Ollilainen a , P. Mattila b , K. Lehikoinen c ,E. Salmela-Mo Èlsa È a , V. Piironen a Department of Applied Chemistry and Microbiology, University of Helsinki, PO Box 27, FIN-00014, Helsinki, Finland b Agricultural Research Center, 31 600 Jokioinen, Finland c Raisio Group, PO Box 101, 21 201 Raisio, Finland Received 11 January 1999; received in revised form 3 May 2000; accepted 30 May 2000 Abstract A method for simultaneous determination of fat-soluble vitamins in animal products is presented. Milk, and two ®sh productscontaining many ingredients, were used as test materials. The vitamins determined were tocopherols, b -carotene, all- trans -retinoland in the case of ®sh, cholecalciferol. The sample preparation procedure, consisting of saponi®cation and extraction by n -hexane± ethyl acetate was carried out in small-scale. To cope with the highly dierent composition of the test materials, some modi®cationswere needed. Normal-phase HPLC separation using diisopropyl ether- n -hexane gradient elution with UV and ¯uorescencedetections for tocopherols, b -carotene and all- trans -retinol was used. Cholecalciferol was determined from the same extract byreverse-phase HPLC (with UV detection) after semi-preparative HPLC puri®cation. Recoveries of spiked samples varied from 80 to111% for all determined fat-soluble vitamins. The day-to-day repeatability was in most cases under 7%, and the within-day varia-tion of this method was small, under 5.5% for all vitamins. The described analytical method is eective and fast, enabling theprocessing of a large number of samples. # 2000 Elsevier Science Ltd. All rights reserved. Keywords: Fat-soluble vitamins; Foods; Simultaneous HPLC determination; Small-scale extraction 1. Introduction Vitamin A activity in food is related to the presence of both retinol and a large number of provitamin car-otenoids. In milk and in meat, vitamin A occurs mainlyas fatty acid retinyl esters. Typical retinoids in ®sh areall- trans -retinol and all -trans- 3,4-dehydroretinol. All- trans -3,4-dehydroretinol is a typical vitamin A com-pound in ®shes like pike ( Esox lucius ), pikepearch( Lucioperca lucioperca ), pearch ( Perca ¯uviatilis ) andwhite®sh ( Coregonus sp.) (Ollilainen, Heinonen, Lin-kola, Varo & Koivistoinen, 1989b). b -Carotene is thepredominant form of carotenoids in milk and meat.Other carotenes such as a -carotene are absent or presentin such low quantities that they may be ignored as asource of vitamin A (Indyk, 1988). The most commonpigments in ®shes are astaxanthin, tunaxanthin andcanthaxanthin, which do not possess vitamin activity(Simpson, Katayama & Chichester, 1981). Astaxanthinor its esters are important carotenoids in salmon,shell®sh and shrimps (Elmadfa & Majchrzak, 1998;Weedon, 1971). Naturally occuring vitamin E is alsocomposed of a number of tocopherol analogues ( a , b , g , d ) and their corresponding unsaturated tocotrienols.The major form of vitamin E in milk fat is a -tocopherol(Syva Èoja, Piironen, Varo, Koivistoinen & Salminen,1985). In marine animals a -tocopherol is the principaltocopherol but also minor amounts of b -, g - and d -ana-logues occur in some ®sh species (Syva Èoja, Salminem,Piironen, Varo, Koivistoinen & Salminen, 1985). Thepredominant vitamin D compound in ®sh is cholecalci-ferol (vitamin D 3 ), which occurs in highly varying con-centrations in dierent species (Mattila, 1995).Simultaneous determination of fat-soluble vitaminshas been widely used for serum and plasma samples. a -Tocopherol, retinoids and carotenoids can bedetermined simultaneously from these samples byreverse-phase chromatography using UV detection for 0308-8146/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.PII: S0308-8146(00)00155-2Food Chemistry 71 (2000) 535±543www.elsevier.com/locate/foodchem* Corresponding author. Tel.: +358-9-191-58225; fax: +358-9-191-58475. E-mail address: pirjo.salo-vaananen@helsinki.® (P. Salo-Va Èa Èna Ènen). carotenoids and retinoids and ¯uorescence detection fortocopherols (e.g. Chuang, Trosclair & Lopez, 1994;Hess, Keller, Oberlin, Bonfanti & Schu Èep, 1991; Manzi,Pan®li & Pizzoferrato, 1996; Yakushina & Tarakova,1995).Fat-soluble vitamins of foods are, however, usuallydetermined by time-consuming separate chromato-graphic methods. All four tocopherols and tocotrienolscan be separated and quanti®ed by normal-phasechromatography with silica stationary phases and¯uorescence detection. Furthermore, removing of nonpolar substances like triacylglycerides is not needed(Pan®li, Manzi & Pizzoferrato, 1994; Piironen, 1986).The use of reverse-phase chromatography for separat-ing tocopherols and tocotnenols is less frequent in foodanalysis due to its poorer separation eciency; a - and d -tocopherols can be separated but b - and g -analogues co-elute (Indyk, 1988; Manzi et al., 1996). In the case of vitamin D compounds, reverse-phase chromatographywith UV detection is preferred because this techniqueallows the use of vitamin D 2 as an internal standard forvitamin D 3 (Mattila, 1995). Because of weak retentionof carotenes on silica phases, reverse-phase chromato-graphy with UV detection is mainly applied forcarotenoids. On the other hand, normal-phase chroma-tography is required for retinoid analysis especially if cis- and trans- isomers of retinol have to be separated.Chromatographic conditions for the simultaneousmeasurement of tocopherols, retinols and b -carotenehave been reported for few foods only, namely dairyproducts and cheese (Hewavitharana, van Brakel &Harnet, 1996; Indyk, 1988; Manzi et al., 1996; Pan®li etal, 1994). To our knowledge, there are no studies onsimultaneous determination of fat-soluble vitamins incomplex animal products with varying concentrationlevels of vitamins.Saponi®cation followed by extraction is the mostwidely used sample preparation method for fat-solublevitamins in food samples. Saponi®cation removes thebulk of fat and facilitates extraction by releasing car-otenoids, retinoids, tocopherols and vitamin D com-pounds from the matrix. Factors to be optimised in asaponi®cation method are sample size, potassiumhydroxide concentration in the solution, saponi®cationtemperature and time, composition of the saponi®edsolution before extraction and extraction solvents.Selection of an antioxidant and removing of oxygen bynitrogen are also important. It is still quite common touse large-scale extraction methods when analysing fat-soluble vitamins in foods. These are, however, laboriousand reagents-consuming (Aminullah Bhuiyan, Raty-nayake & Ackman., 1993; Mattila, 1995; Ollilainen etal., 1989a, 1989b; Piironen, 1986; Rettenmaier &Schu Èep, 1992). There are few small-scale applicationsconcerning simultaneous extraction of b -carotene,tocopherols and all- trans -retinol from dairy productsand cheese (Indyk, 1988; Pan®li et al, 1994) but notfrom more complex animal products. Furthermore,vitamin D compounds have been determined from thesame extract as the other fat-soluble vitamins only afterlarge-scale extraction (Aminullah Bhuyian et al.).The aim of the present study was to develop a small-scale saponi®cation and extraction procedure andHPLC method for simultaneous determination of dierent fat-soluble vitamins from animal based foodmaterials. When developing the sample preparationprocedure, the aim was to decrease the amount of workand consumption of reagents compared with methodstraditionally used (Aminullah Bhuiyan et al., 1993;Mattila, 1995; Ollilainen et al., 1989a, 1989b; Piironen,1986; Rettenmaier & Schu Èep, 1992). Tocopherols, all- trans -retinol and b -carotene were determinedsimultaneously in the same run using normal-phasechromatography. The tested samples were two ®shproducts (fried Baltic herring ®llet and fried Balticherring burger) and standard milk. The sample typeswere chosen so that they varied extensively in theircompositions and vitamin levels. In the case of ®sh,which is the most important natural source of vitaminD 3 , our further aim was to determine this vitamin fromthe same extract. Reverse-phase chromatography wasused for its quanti®cation. 2. Materials and methods 2.1. Chemicals HPLC-grade n -hexane and ethyl acetate were pur-chased from Rathburn (Walkerburn, UK). Diisopropylether (J.T. Baker, Deventer, The Netherlands), 2-pro-panol, methanol, tetrahydrofuran (Rathburn) and MilliQ water were of HPLC-grade. Ethanol was eitherspectrophotometric grade (AAS, 99.5%) or 94% (A)(Primalco Oy, Finland). Ascorbic acid, EDTA andpyrogallol were purchased from Merck (Darmstadt,Germany) and were of pro-analysis grade. KOH pelletsfor saponi®cation solution were purchased from EKANobel (Bohus, Sweden). The saponi®cation solutionswere KOH (100) containing 100 g of KOH dissolved in100 ml of water and KOH (50) containing 50 g of KOHdissolved in 100 ml of water. DL - a -Tocopherol (Art. no. 8283, 98.0±102.0%) andtocopherol-isomers ( a -, b -, g -, and d -tocopherol fu Èrbiochemische Zwecke, Art no. 15496) were purchasedfrom Merck, all- trans - b -carotene (C-0126, type IV fromcarrots), vitamins D 2 and D 3 from Sigma (St Louis,USA). Vitamin-A-alcohol (all- trans -retinol, isomericpurity approximately 95% all- trans -retinol, 5% 13- cis -isomer, no. 95144) was purchased from Fluka (Buchs,Switzerland). Rye and barley ¯our were extracted toproduce a qualitative standard for tocotrienols. 536 P. Salo-Va È a È na È nen et al./Food Chemistry 71 (2000) 535±543 The standard stock solutions of the vitamins for ana-lysing milk samples were prepared to a concentration of approximately 5 mg/ml of a -tocopherol in ethanol(AAS), 400 m g/ml of all- trans -retinol in ethanol (AAS)and 20 m g/ml of b -carotene in n -hexane; they werestored at À 20 C in the dark. These solutions were dilu-ted in ethanol (AAS) or n -hexane as appropriate and theconcentrations were con®rmed spectrophotometricallyusing known absorption coecients of each vitamin (seebelow). The combined working solution was preparedby pooling suitable volumes of each stock solution anddiluting with n -hexane to obtain concentrations rangingfrom 0.4 to 1.4 m g/ml for each vitamin.The standard stock solutions of the vitamins for ana-lysing ®sh samples were prepared to a concentration of approximately 5 mg/ml, 500, 400 and 600 m g/ml of a -, b -, g -, and d -tocopherols (in ethanol, AAS), 400 m g/mlof all- trans -retinol (in ethanol, AAS), 20 m g/ml of b -carotene (in n -hexane), 200 m g/ml of vitamin D 2 (internal standard, in ethanol, AAS) and 200 m g/ml of vitamin D 3 (in ethanol, AAS). All others were stored at À 20 C but vitamin D stock solutions were stored at+4 C. These solutions were diluted in ethanol (AAS) or n -hexane as appropriate and the concentrations werecon®rmed spectrophotometrically using the knownabsorption coecients of each vitamin. The speci®cabsorption coecients ( E 1 7 1 m ) used were 75.8 for a -tocopherol, 89.4 for b -tocopherol, 91.4 for g -tocopheroland 87.3 for d -tocopherol at 292, 296, 298 and 298 nm,respectively, 1850 for all- trans -retinol at 325 nm (Frolik& Olson, 1984), 2592 for b -carotene at 450 nm (DeRitter & Purcell, 1981), 475 for D 2 and 480 for D 3 at265 nm (British Pharmaceutical Codex: Anon, 1979).The combined working solution was prepared bypooling suitable volumes of each stock solution anddiluting with n -hexane to obtain concentrations rangingfrom 0.2 to 75 m g/ml. 2.2. Samples When developing the method, the samples werestandard milk and more complex matrixes, a ®shproduct (fried Baltic herring ®llet) coated with rye ¯ourand fried in rapeseed oil. Suitability of the method foranalysis of a diering ®sh product (fried Baltic herringburger), coated with wheat ¯our and bread crumbs, andfried in rapeseed oil, was also tested. 2.3. Sample preparation All work was carried out under subdued lightconditions. 2.3.1. Milk samples For saponi®cation, 1 g of thermostated (40 C, inorder to stabilize the distribution of milk fat) milk wasweighed into a 10 ml Kimax tube. AAS ethanol (4 ml), aspatletip of pyrogallol and a spatletip of ascorbic acidwere added to the tube and the tube was capped, vor-texed and allowed to stand for 10 min. Nitrogen was ledinto the tube to remove oxygen and 0.5 ml of saturatedEDTA and 0.5 ml of KOH (KOH50) were added to thetube. The tube was capped, shaken, and transferred to aboiling water bath for 20 min. The tube was shakenonce after 10 min. After boiling, the tube was cooled inan ice-water bath for 10 min (Fig. 1).After saponi®cation and subsequent cooling of themilk sample, 2 ml of water and n -hexane±ethyl acetate(8+2, v/v) were added to extract the fat-solublevitamins. The tube was shaken with 500 strokes/min(Edmund Bu Èhrer Laborgera Ètebau, Glastechnik,Umwelttechnik) for 10 min prior to separating the lay-ers. The extraction was repeated with another 2 mlportion of n -hexane-ethyl acetate (8+2). The combinedorganic layer was evaporated with nitrogen (30 C) andthe dry residue was dissolved in 1 ml of n- hexane and®ltered (Whatman, Puradisc 25 TF, 0.45 m m) prior toHPLC analysis (Fig. 2). 2.3.2. Fish samples For saponi®cation of ®sh product, I g of homo-genized sample was weighed into a Pyrex tube, 5 ml of 2% ascorbic acid (in water), 10 ml of ethanol (A) and 1ml of internal standard (0.2 m g of vitamin D 2 ) were Fig. 1. Saponi®cation procedure for standard milk and ®sh productsamples. P. Salo-Va È a È na È nen et al./Food Chemistry 71 (2000) 535±543 537 added. The tube was vortexed and ¯ushed with nitro-gen, 4 ml of KOH (KOH 100) was added, the tube wascapped and transferred into a boiling water bath for 20min. The tube was vortexed after 10 min of boiling.After boiling, the tube was cooled in an ice-water bathfor 10 min. Ten ml of 10% NaCl was added to the tubeto avoid emulsion formation (Fig. 1).After saponi®cation of the ®sh sample, the fat-solublevitamins were extracted using three 20 ml portions of n -hexane-ethyl acetate (8+2). The tube was shaken with aDesaga mixer for 2 min after each extraction, and thephases were allowed to separate. The organic layerswere collected in a 100 ml tube where the combinedextract was washed with 20 ml of 5% NaCl. Afterwashing, the organic layer was transferred to a 100 mlbottle and evaporated with a Rotavapor (30 C). Five mlof ethanol (AAS) and 5 ml of n -hexane were added andthe solution was evaporated. The residue was quantita-tively transferred with a small amount of n -hexane to aKimax tube and evaporated to dryness with nitrogen.The residue was dissolved in 1 ml of n -hexane and®ltered (Whatman, Puradisc 25 TF, 0.45 m m) prior toHPLC analysis (Fig. 2). 2.4. Analytical and semipreparative HPLC Normal-phase chromatography with UV ( b -caroteneand all- trans -retinol) and ¯uorescence detections (toco-pherols) was used. When analysing vitamin D com-pounds, reverse-phase chromatography with UVdetection was used after the semipreparative clean-upprocedure. 2.4.1. Quanti®cation of tocopherols, -carotene, and all-trans-retinol with HPLC The analytical HPLC system consisted of a VarianVista 5500 liquid chromatograph equipped with a UVdetector (Waters 486), a ¯uorescence detector (Waters470), an autosampler with a cooling module (Waters700 Satellite WISP), and a m Porasil column (10 m m, 30cm  3.9mm,Waters)withasilicaguardcolumn (Guard-Pak Silica, Waters). The temperature of the columnoven was 30 C. Separation of vitamins was based onstep gradient elution (Table 1). The injection volumeused was 75 m l.Tocopherols, b -carotene, and all- trans -retinol werequanti®ed with an external standard method in whichquanti®cation was based on peak areas. Standardcurves (four concentration levels) were obtained dailyby standard injections. The variations of detectorresponse and the retention times were obtained bystandard injections, after every third sample injection. 2.4.2. Vitamin D: Clean-up with semipreparative HPLC Fish extracts were puri®ed by using the normal-phaseHPLC method. The semipreparative cleanup systemconsistedof aWaters510 HPLC pump,aMerck-Hitachi1-4200 UV±VIS detector set at 265 nm, and a m Porasilcolumn (10 m m, 30 cm  3.9 mm, Waters) with a silicaguard column (Guard-Pak Silica, Waters). The mobilephase consisted of n -hexane, tetrahydrofuran, and 2-propanol (98:1:1); the ¯ow rate was 1 ml/min, and theinjection volume was 300 m ml of the sample extract. Acollecting time of from 1.5 min before to 1.5 min afterthe retention time of co-eluting vitamin D 2 and D 3 standard peak was used. The collected fraction wasevaporated under nitrogen, and the residue was dis-solved in 150 m l of 7% water in methanol (Mattila et al.,1992). 2.4.3. Quanti®cation of vitamin D 3 with analytical HPLC The analytical HPLC system consisted of a Waters6000 pump, a UV detector set at 265 nm (Waters 486),an autosampler (Waters 712), and a Vydac TP 54column (5 m m, 25 cm  4.6 mm, The Separations group)with an ODS quard column (Nova-Pak C18, Waters).The temperature of the column oven was 25 C. Themobile phase contained 93% methanol and 7% water. Fig. 2. Extraction and HPLC determination of fat-soluble vitamins,from standard milk and ®sh product samples.538 P. Salo-Va È a È na È nen et al./Food Chemistry 71 (2000) 535±543