Quick UHPLC/MS analysis of and Its Analogues from Powdered Infant Milk Using Polymeric Solid Phase Extraction ASMS 2012  Daniel Tran, Agilent Technologies, 25200 Commercentre Drive, Lake Forest, CA 92630 Poster WP25- 506

Introduction Experimental Experimental Results and Discussion Results and Discussion Results and Discussion Conclusions

Since the infant milk scandal in 2007, several methods have Standards LC Conditions All compounds showed linearity from 0.5 µg/g - 10µg/g • This method couples generic polymeric SPE methods & been released for the analysis of melamine from dairy with a R2 above 0.995. A minimum of 5 levels were used for reversed phase LC to reduce the complexity involved products. However, a comprehensive method for the Standards were ordered from Chromadex and made into 1.0 mg/mL Column – Pursuit XRs Ultra 2.8 Diphenyl 100x2.0mm the calibration curves. with the current FDA method, employing complex stock solutions. Melamine and were prepared in water. analysis of the related compounds (ammeline, ammelide, Instrument: Agilent 1290/6460 QQQ Melamine sample preparation & HILIC LC. and cyanuric acid) is difficult to find. The US FDA has a Ammeline and ammelide were prepared in a 2% Average % Recovery ± RSD comprehensive method but the lengthy sample preparation hydroxide solution. All four compounds were mixed together to a • The two tiered approach using Plexa PCX & Plexa PAX concentration of 100µg/ml in 2% Ammonium Hydroxide. Ammonium Mobile Phase SPE Captiva steps may not be practical in a production laboratory where A: 0.1% Aqueous Formic Acid allows for the quick analysis of melamine and its Hydroxide is required to keep all compounds in solution. Cyanuric Acid low sample preparation time is critical. The hydrophilic B: MeOH Melamine 100 ± 7.9 94 ± 12.4 analogues down to and below regulatory levels without nature of the analytes, suggests a HILC LC column would Powdered infant milk was prefortified to concentrations ranging from Cyanuric Acid 117 ± 5.9 n/a sacrificing sensitivity and ease of use. be the most promising and most procedures use a HILIC LC 0.1 µg/g to 10 µg/g of melamine and its analogues. Pump Program Ammelide method. HILIC has many pitfalls concerned with Flow rate 400 µL/ min. Ammeline 114 ± 6.8 n/a • At regulatory levels melamine, ammeline, and reproducibility and high organic solvent usage. Ammelide 112 ± 5.8 n/a ammelide showed relative recoveries within 14% of t0 A: 98%, B: 2% %b Sample Preparation true value with RSDs below 8%. Cyanuric acid showed 100 Table 2. Recoveries of melamine and its analogues from fortified An alternative simple method was developed using Captiva t.0.5– 2.0 A: 95%, B: 5% a relative recovery within 17% of true value with a RSD ND plates, This quick and reliable method utilizes a Figure 2. baseline separation of the melamine-cyanuric acid pair powdered infant formula compared to Captiva ND method, t A: 20%, B: 80% 50 and the ammeline-ammelide pair at 10ng/ml of less than 6%. reversed phase LC column and limits the preparation time 1. Weigh out 1 ± 0.01 g powdered infant milk 2.01- 3:00 1.0 µg/g (n=6) to a minimum, while maintaining quantification of 2. The milk was then spiked with analytes to 1 µg/g t3.01- 4:00 A: 98%, B: 2% • Sensitivity was maintained below the regulatory level 0 Table 2 lists the relative recoveries of melamine and its melamine at regulatory levels. Using Captiva ND plates a 3. Add 20 mL H2O Run Time = 4:00 minutes. 0 1 2 3 4 Melamine - 9 Levels, 9 Levels Used, 9 Points, 9 Points Used, 0 QCs 4 of 1µg/g. Ammeline, ammelide, and cyanuric acid x10 y = 11.014853 * x + 98.762738 R^2 = 0.99318032 simple and quick analysis of melamine and its analogues 1.5 analogues following sample preparation with Plexa PCX & 4. Vortex or shake, there should be no remaining powder 1.4 maintained an SNR above 5:1 down to the 0.1µg/g 1.3 was developed. Good linearity was achieved for each 1.2 Plexa PAX. All analytes extracted with Plexa PCX showed 5. Transfer 2, 1mL aliquots of milk to test tubes 1.1 level. Melamine was capable of maintaining that SNR MS Conditions 1 analyte and as well as good accuracy and precision. A 0.9 relative recoveries within 14% of true value with RSDs 6. To one add 2mL 0.1N HCL (will be applied to Plexa PCX) 0.8 down to 0.5 µg/g. 0.7 below 8%. Cyanuric acid, extracted from Plexa PAX, comparison of diluted samples to samples processed MS Conditions 0.6 7. To the other sample add 2mL 0.1N (will be 0.5 0.4 showed relative recovery within 17% of true value with a through the Captiva ND was also investigated, 0.3 applied to Plexa PAX) Source Agilent JetStream ESI+ 0.2 2 0.1 r = 0.999 RSD of less than 6%. demonstrating better sensitivity is achieved by using the 0

Responses -0.1 8. Let sit for 30 mins. Gas Temp 300 °C -50 0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000 1050 Captiva ND plate Concentration (ng/ml) Gas Flow 5 L/min Analyte sensitivity increased across the board for all

Cyanuric Acid - 5 Levels, 5 Levels Used, 5 Points, 5 Points Used, 0 QCs Nebulizer 20 psi x10 2 y = 0.473833 * x + 2.233742 R^2 = 0.99950817 compounds. Figure 4 demonstrates a 3 fold sensitivity 4.75 4.5 4.25 increase for melamine compared to a similar sample Plexa PCX Plexa PAX 4 Sheath Gas Temp 275 °C 3.75 3.5 prepared via Captiva ND, both were prefortified to 1.0µg/g. 3.25 3 Sheath Gas Flow 7L/min 2.75 Condition 1. 3mL MeOH 1. 3mL MeOH 2.5 2.25 2 1.75 2. 3mL H2O 2. 3mL H2O Capillary +3500/-2000 1.5 1.25 1 0.75 Compound Precursor Ion Product Ion Fragment Collision Energy Polarity 0.5 2 Load Sample pretreated w/ Sample pretreated w/ 0.25 r = 0.997 0 -0.25 Responses -50 0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000 1050 0.1N HCL 0.1N NaOH Melamine 127 85.1 100 18 + Concentration (ng/ml) Cyanuric Acid 128 42.1 60 14 -

Wash 1. 3mL 2% formic acid in 1. 3mL H2O Ammeline - 9 Levels, 9 Levels Used, 9 Points, 9 Points Used, 0 QCs x10 2 y = 0.511477 * x + 8.378224 R^2 = 0.99555720

Ammeline 128 69.1 140 34 + 5.25 H2O 2. 3mL MeOH 5 References 4.75 4.5 4.25 2. 3mL 50:50 Ammelide 127 84 100 6 - 4 3.75 3.5 3.25 MeOH:ACN 3 2.75 1. Turnipseed SB, Casey C, Nochetto C, and Heller DN. 2.5 2.25 2 U.S. FDA Laboratory Information Bulletin, 2008: 24 : 1.75 Elute 3mL 5% Ammonium 5% Acetic Acid in MeOH 1.5 1.25 1 4421. Lib. No. 4421 0.75 2 Hydroxide in 50:50 Results and Discussion 0.5 r = 0.996 0.25 0 [http://www.fda.gov/Food/ScienceResearch/Laborato Responses 0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000 1050 MeOH:ACN Concentration (ng/ml) Although this method requires a separate preparation for cyanuric ryMethods/DrugChemicalResiduesMethodology/ucm0 Evaporate Reconstitute in 500ul 50:50 MeOH:ACN 71637.htm] accessed September 1, 2011. acid, a single reversed phase LC run can be used for all Ammelide - 9 Levels, 9 Levels Used, 9 Points, 9 Points Used, 0 QCs x10 3 y = 4.707766 * x + 10.793379 R^2 = 0.99776734

compounds. Figure 2 shows the separation achieved using a 5.5 2. Chinese Ministry of Health (2011). China sets limits of 5 Table 1. SPE Method 4.5 reversed phase column. The melamine-cyanuric acid pair as well 4 melamine levels tolerable in food products, 21 April, 3.5 Figure 4. Melamine extracted via Plexa PCX (top), Captiva ND as the ammeline-ammelide pair show baseline separation. 3 2011 (http://english.gov.cn/2011- 2.5 (bottom) at 1.0µg/g 2 04/21/content_1849392.htm) 1.5

1 The samples treated with HCL were applied to Plexa PCX 0.5 0 2 Quick analysis of melamine and its analogues were 3. Tran, DB, and Husdon, W. Fast and Quick Detection of cartridges. From those cartridges melamine, ammeline, and -0.5 r = 0.995 -1

Responses accomplished down to the regulatory levels of 1µg/g (2) Melamine and Its Analogues From Powdered Infant -50 0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000 1050 Figure 1. Clockwise from top left, melamine, ammeline, ammelide were extracted. Cyanuric acid could only be extracted Concentration (ng/ml) and below. The previous Captiva ND method was sensitivity Milk, 01 February, 2012. 5990-9591EN ammelide, and cyanuric acid respectively on an anion exchange cartridge. Those samples were pretreated enough only for melamine, maintaining a signal-to-noise (http://www.chem.agilent.com/Library/applications/5 Figure 3. calibration curves of melamine, cyanuric acid, ammeline, with NaOH and extracted on the Plexa PAX cartridges. All ratio > 5:1 required by the US FDA at the 1µg/g level (1). 990-9591EN.pdf) compounds were reliably detected down to 0.2 µg/g. and ammelide, from top to bottom.