2.3. Sample Analysis
The instrumentation used was similar to that of Jäpelt et al. [
25]. Extraction procedures were derived from published methodology [
25,
41]. An equivalent deuterated internal standard was added for each vitamin D analogue under investigation: 100 µL of a mixed internal standard solution was added to each sample. This contained 100 ng/mL each of vitamin D
3 [
2H
3], vitamin D
2 [
2H
3], 25(OH)D
3 [
2H
3], and 25(OH)D
2[
2H
3] (Iso Sciences/PM Separations).
The samples were homogenized with 1 g ascorbic acid, 10 mL deionized water, 30 mL absolute ethanol, 2 g potassium hydroxide pellets, and 100 µL of 100 ng/mL deuterated internal standard mix and made to 50 mL with deionized water. The headspace was flushed with nitrogen gas, capped, and placed in a shaker for saponification overnight. The samples underwent centrifugation and 10 mL of the ethanol layer was extracted onto diatomaceous earth Solid Phase Extraction (SPE) cartridges (ChemElute Agilent). The organic soluble compounds were washed off with two 30 mL aliquots of petroleum spirits. The washes were collected into 80 mL glass EPA vials and then evaporated to dryness under high purity nitrogen gas. The residue was reconstituted into 400 µL heptane and transferred to a LC vial containing a 400 µL glass insert. The prepared extracts were stored at −20 °C.
The reagent, PTAD (4-phenyl-1,2,4-triazoline-3,5-dione) reacts non-specifically with dienes under a reaction mechanism called the Diels–Alder Reaction. While in theory, excess PTAD is added to derivatize the vitamin D analogues, samples with high diene content may limit the amount of PTAD available for derivatization. Therefore, where samples were determined to have high diene content, extract clean-up via normal phase chromatography fraction collection was performed. The extracts were inspected for cold precipitate: if present, the liquid extract was transferred to a fresh 400 µL glass insert. Of the remaining liquid extract, 200 µL were injected onto a normal phase chromatographic system with a silica column, 1 mL/min 2% isopropyl alcohol in heptane mobile phase, and a photodiode array detector set to 265 nm. Vitamin D and 25(OH)D fractions were collected.
Fractions of vitamin D and 25(OH)D were combined and evaporated under high purity nitrogen gas. The dry material was reconstituted in 200 µL of dry acetonitrile containing 1 mg/mL of 4-phenyl-1,2,4-triazole-3,5-dione (PTAD) and transferred to a fresh LC vial. Two hours were allowed to complete derivatization. The sample was evaporated under high purity nitrogen gas. The dry material was reconstituted in 100 µL of methanol and water (70:30), transferred to a fresh 400 µL glass insert, and placed into an LC vial. A limit of quantitation was conservatively set at 0.05 µg/100g, which is half of the ‘spiked’ recovery level of 0.1 µg/100g.
The recoveries were determined for each sample analyzed. The recoveries at the 0.1 µg/100 g ‘spiked’ recovery level were as follows: vitamin D3, 86–104%; vitamin D2, 76–105%; 25(OH)D3, 85–114%; and 25(OH)D2, 90–114%. The recovery for 25-hydroxyvitamin D3 in wattleseed (roasted/milled/ground seed) could not be determined due to a matrix interference.
The samples were analyzed for vitamin D2, vitamin D3, 25(OH)D2, and 25(OH)D3 using LC-QQQ (Agilent, San Jose, CA, USA). The calibration samples of vitamin D2, vitamin D3, 25(OH)D2, and 25(OH)D3 were prepared. The calibration concentrations (in ng/mL) were 0, 2.5, 5, 7.5, 10, 15, 25, 50, 75, and 100. Each calibration sample also contained 10 ng/mL of deuterated internal standard for each vitamer (vitamin D analogue) tested. The calibrations and samples were analyzed using 1290 Infinity Series LC System/6460 Triple Quad liquid chromatography–tandem mass spectrometry (LC–MS/MS; Agilent Technologies Mulgrave, Victoria, Australia) fitted with a Jet Stream ESI source in positive ion mode using a Supelco Ascentis Express C18 10 cm × 2.1 mm, 2.7 µm LC chromatographic column (Sigma-Aldrich, Sydney Australia).
For each vitamer analyzed, each sample was tested in duplicate, and duplicate values were averaged to obtain one mean value for each sample. A third sample, spiked with the same vitamer, was analyzed for each sample tested to provide quality control data. The mean percentage recovery and mean relative percentage difference were calculated for each vitamer. At the time of writing, the expected limit of detection, post validation study, is expected to be 0.05 µg/100 g (N. Strobel, email communication, 10 October 2017).
The mean recovery percentage across all samples for vitamin D2, vitamin D3, 25(OH)D2, and 25(OH)D3 was 96%, 98%, 101%, and 94%, respectively. Across all samples, the mean relative percentage difference between duplicate samples for vitamin D2, vitamin D3, 25(OH)D2, and 25(OH)D3 was 71%, 15%, 50%, and 56%, respectively.