Simultaneous Measurement of 1110 Note Application Vitamins A, D, E and K Along with CoQ10 and Carotenoids, in Multivitamin Tablets, Infant Formula and Milk Ian Acworth and Paul Gamache, Thermo Fisher Scientific, Chelmsford, MA, USA Key Words Carotenoids, Food Supplements, HPLC-ECD, Multivitamins, Retinoids, Tocopheryl Acetate Goal To develop a sensitive and selective HPLC-electrochemical detection (ECD) method for the simultaneous measurement of numerous fat-soluble nutrients Introduction Consumer interest and regulatory concerns for food This Application Note shows the use of the Thermo supplements, fortified food products and dietary nutrients Scientific™ Dionex™ CoulArray™ Coulometric Array have stimulated the development of improved analytical Detector for the measurement of a variety of lipid- methods. Fat-soluble nutrients are of particular interest soluble nutrients including vitamins A (retinol), D2 due to their potential health benefits and, in some cases, (ergocalciferol), D3 (cholecalciferol), E (a- and toxicity. Accurate, simple techniques for comprehensive γ-tocopherol), and K1 (phylloquinone), along with analysis is important for product labeling, nutritional carotenoids (a- and b-carotene), retinoids (retinyl research, development and quality control.1 palmitate), tocopheryl acetate, and coenzyme Q10 (ubiquinone) (Figure 1). The analysis of multivitamin Most laboratories use individual methods for a single tablets, infant formula and milk are given as examples. analyte class. Many methods, for example, include a Since extraction efficiency is highly dependent upon saponification step prior to extraction to convert multiple sample matrix the procedures described here are only forms of a compound to a single analyte and to exclude preliminary and need further investigation. potentially interfering substances.2-4 This may lead to rapid degradation and/or production of less stable forms of several analytes of interest.1,5 Coulometric array electrochemical detection (ECD) utilizes multiple sensors that can be optimized for more than one chemical class.6 Resolution is increased by electrochemical screening of possible interfering solutes based on differences in oxidation-reduction behavior. 2 CH 3 H C H C CH 3 CH 3 3 3 CH CH 3 3 H CH H CH 3 3 H H Ergocalciferol Cholecalciferol CH CH 2 2 HO HO CH 3 HO CH CH CH CH 3 3 3 3 H C O 3 CH 3 CH 3 α-Tocopheryl HO CH CH CH CH 3 3 3 3 H C O CH 3 3 CH 3 γ -Tocopheryl O O CH O CH CH 3 3 3 CH CH CH 3 3 3 CH O 3 (CH CH=CCH ) H CH CH=CCH (CH CH CHCH ) H 2 2 10 2 2 2 2 2 3 O O Phylloquinone Ubiquinone-10 CH CH H C CH 3 3 3 3 CH OH 2 CH 3 All-Trans Retinol CH 3 CH CH CH H C 3 3 3 3 CH 3 CH CH CH 3 3 3 CH 3 α -Carotene CH 3 CH CH CH H C 3 3 3 3 CH 3 CH CH CH 3 3 3 CH 3 ß-Carotene Figure 1. The structure of fat soluble vitamins and nutrients measured using this method. 3 Materials and Methods Infant Formula The gradient system consisted of two pumps, an The concentrate was diluted with water (according to autosampler, a thermal chamber and an eight channel label instructions) and extracted as for unsaponified CoulArray detector. milk samples below. Milk Samples (Unsaponified) LC Conditions A 1.0 mL volume, augmented with 10 µL of 1.0 µg/mL Column: C18 Keystone, 4.6 × .2150 mm, 5 µm D2 (internal standard), was thoroughly mixed with 3.0 mL Mobile Phase A: Acetonitrile-Water, 90:10 (v/v) containing diluent and 0.1 g magnesium sulfate. The resulting mixture 20 mM Sodium Perchlorate and 5.0 mM was extracted 2 times with 4.0 mL hexane. Combined Perchloric Acid hexane extracts were evaporated under a stream of nitrogen and residue was dissolved in 1.0 mL of diluent. Mobile Phase B: Acetonitrile-1-Propanol, 65:35 (v/v) containing 20 mM Sodium Perchlorate and The solution was centrifuged as above. 10 mM Perchloric Acid. Milk Samples (Saponified) Gradient Conditions: 20 minutes linear gradient from 10 to A 1.0 mL volume of milk was mixed with 1.75 mL 85% 100% B followed by a 5 minutes hold aqueous EtOH containing 75 mg/mL potassium hydroxide at 100% B before returning to initial and 0.25 mg/mL ascorbic acid. The sample was then conditions for 5 minutes. Total run time placed in a heated water bath for 45 minutes. at 95 °C. was 30 minutes. Saponified samples were then extracted as for Flow Rate: 1.5 mL/min unsaponified milk samples. Temperature: 32 °C Injection Volume: Sample dependent Results and Discussion Multi-component analysis using ECD has previously been Detector and Conditions limited by its poor compatibility with gradient elution Detector: Model 5600A, CoulArray chromatography. Furthermore, since the electrochemical Applied Potentials: -700, 100, 250, 400, 550, 750, 800, 850 properties of lipid soluble nutrients are very diverse, (vs. Pd) single channel detectors must be used at potential settings that are suitable for only a few analytes at the expense Standards of others.7 Coulometric array detection utilizes several high efficiency sensors in series and maintained at Standards were obtained from Sigma-Aldrich® different fixed potentials. This enables the concurrent (St. Louis, MO). Stock solutions were made by dissolving detection of different chemical classes, each at their approximately 10 mg in 10 mL of ethanol (EtOH) with optimal potential settings. Selectivity is increased through the exception of the carotenoids and coenzyme Q10. electrochemical screening of possible interfering These more lipophilic compounds (ca 1 mg) were dissolved compounds whose oxidation-reduction behaviors differ in 5.0 mL hexane followed by dilution with 15 mL ethanol. from the analytes of interest. Sensor, electronic and Solutions were assigned a concentration value, based on software design enables the routine use of gradient elution. molar absorbtivity, before addition of 10 mg/L butylated hydroxyanisole (BHA) as preservative. Stock solutions Chromatographic and detector conditions were optimized were stored at -20 °C for up to 6 months. Further dilutions for the measurement of a wide range of lipid soluble were made weekly in EtOH containing 10 mg/L BHA nutrients. Figure 2 shows a 12 component standard (diluent) and stored, protected from light at -20 °C. mixture. Sensor 1 (not shown) was maintained at -700 mV (vs. Pd) to reduce vitamin K1 and coenzyme Q10 which Sample Preparation were then measured oxidatively on sensors 2 and 3, Multivitamin Tablets respectively. Low oxidizing compounds (e.g. tocopherols, A single tablet was powdered and then sonicated in 10 mL retinoids) were oxidatively screened at upstream electrodes hexane for 10 mins. Following the addition of 90 mL and higher oxidizing compounds (calciferols, tocopheryl EtOH the sample was sonicated for 30 minutes. The esters) responded only at higher potentials. Using an solution was centrifuged at 3000 g for 10 minutes. array of potentials along the oxidative curve of each A 25 µL volume was analyzed. analyte allowed generation of response ratios for each peak. Comparison of ratios between authentic standard and unknown sample provided an estimate of the purity of each analyte peak in the samples. 4 Figure 2. Chromatogram showing the simultaneous measurement of fat-soluble vitamins and nutrients. Food supplements, fortified food and natural products The analysis of a multivitamin tablet is shown in Figure 3. contain a wide range of fat-soluble nutrients and within The simple dissolution technique enabled the measurement a single, often complex, matrix there exists a wide range of vitamins D and K well above their lower limit of of analyte levels. Multi-component analysis therefore detection. In the same analysis, higher level analytes requires high selectivity, sensitivity and a wide response (retinyl acetate, b-carotene and tocopherols) were detected range. Based on a signal/noise ratio of 5:1 the lod for well within their linear range. Response ratios obtained standard compounds was 4, 9, 8, 7, 5 and 12 pg (on by comparison of authentic standard and sample peaks column) for vitamins A, D2, D3, E, and retinyl palmitate, demonstrated high peak purity for all analytes. respectively. The assay showed excellent linearity over 4 orders of magnitude for all analytes (e.g., correlation coefficients [0 to 250 ng on column] for vitamins A, D2 and D3, were 0.9988, 0.9992 and 0.9983, respectively). 1.50 ocopheryl T ocopherol - T - Acetate α α -Carotene Acetate β D3 ocopherol itamin K1 T Retinyl - V γ 1.00 Retinol [800 mV] [750 mV] 0.50 [550 mV] Response (µA) [400 mV] [250 mV] 0.00 [100 mV] 0.0 5.0 10.0 15.0 20.0 25.0 Retention time (minutes) Figure 3. Analysis of a multivitamin tablet. 5 Infant formula is a complex sample matrix containing low itself). The feasibility of the analytical method for the levels of vitamins D and K. Current methods for vitamin measurement of multiple nutrients in infant formula D analysis are complex and include a saponification using a simplified extraction method that avoids step which can cause the degradation of lipid soluble saponification is illustrated in Figures 4 and 5. nutrients (e.g., vitamins E and K, and even vitamin D 8.0 D2 D3 ocopherol ocopheryl ocopherol T T - - -T γ α Acetate α 6.0 Retinyl Palmitate CoQ10 ocopherol T - δ [800 mV] 4.0 [700 mV] Response (µA) [550 mV] 2.0 [400 mV] [250 mV] 0.0 [100 mV] 0.0 5.0 10.0 15.0 20.0 25.0 Retention time (minutes) Figure 4. Analysis of infant formula presented at low amplification.