Fast quantitative determination of melamine and its derivatives by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry

Research Article

Received: 8 April 2011 Revised: 31 May 2011 Accepted: 30 June 2011 Published online in Wiley Online Library

Rapid Commun. Mass Spectrom. 2011, 25, 2844–2850

2844

Fast quantitative determination of melamine and its derivativesby matrix-assisted laser desorption/ionization time-of-flight massspectrometry

Anne Arnold, Tabiwang N. Arrey, Michael Karas and Markus Persike*LOEWECluster ’AmbiProbe’, Cluster of Excellence ’Macromolecular Complexes’, Institute of Pharmaceutical Chemistry, GoetheUniversity Frankfurt, Germany

A simple, sensitive and fast method for the determination of melamine and its derivatives in milk powder usingmatrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) was developed.Neither time-consuming sample preparation, nor special target plates, or other extra equipment are necessary. Thecommon matrix sinapinic acid (SA) was used with a dried-droplet preparation. Detection limits (signal-to-noise(S/N) ratio = 3) for standard solutions of melamine, ammeline and cyanuric acid were 10, 25 and 10 mg/L, respectively.The limit of quantification (LOQ) for melamine was 25 mg/L and excellent linearity (R2: 0.9990) was maintained overthe range of 10–2000 mg/L. Ammeline and cyanuric acid were analyzed with an LOQ of 50 mg/L and also excellentlinearity (R2: 0.9997 and R2: 0.9998). Good accuracy and precision were obtained for all concentrations within therange of the standard curve. The developed method was successfully used for the determination of melamine,ammeline and cyanuric acid in milk powder samples with a simple sample preparation. The LOQ of melamine was0.5 mg/g. Ammeline and cyanuric acid were detectable at 0.5 and 5 mg/g. This method showed excellent accuracy,precision and linearity and significantly reduces the needed analysis time, as only approximately 10 s/samplemeasuring time is required. To the authors’ knowledge, this is the first published method to quantify melamineand derivatives by MALDI-TOF-MS. Copyright © 2011 John Wiley & Sons, Ltd.

(wileyonlinelibrary.com) DOI: 10.1002/rcm.5170

In recent years, melamine (2,4,6-triamino-1,3,5-triazine) hasgained dubious fame. This organic base is primarily used forthe synthesis of melamine formaldehyde resins, for the fabrica-tion of plastics, laminates, coatings or glues. The structures ofmelamine, ammeline and cyanuric acid are shown in Fig. 1.Melamine attracted worldwide attention in 2008 because ofits presence in milk powder. Due to the fact that melamineis enriched in nitrogen, it was illegally used to falsify the pro-tein content of powder milk. This resulted in six reporteddeaths and the hospitalization of approximately 294 000 chil-dren.[1] The ingestion of melamine can induce the formationof kidney stones, crystalluria and acute renal failure.[2,3] Thetoxicity of this substance is further increased if combined withcyanuric acid.[4] Various studies have shown that kidneystones and renal failure only occurred in combination of bothsubstances.[5–7] The increased toxicity is probably a result ofthe formation of an insoluble melamine–cyanuric acid complexin the kidney.[5] Based on these data, safety limits of 2.5 μg/g(1 μg/g in infant formula) for melamine have been deter-mined by the World Health Organization (WHO).Due to this global food crisis, manymethodswere developed

for the analysis ofmelamine and related substances. Numeroustechniques such as gas chromatography (GC),[8,9] liquid

* Correspondence to: M. Persike, Institute of PharmaceuticalChemistry, Goethe University, Max-von-Laue-Str. 9, 60438Frankfurt, Germany.E-mail: [email protected]

Rapid Commun. Mass Spectrom. 2011, 25, 2844–2850

chromatography (LC),[10–12] capillary electrophoresis (CE),[13]

desorption electrospray-ionization (DESI),[14] desorptionatmospheric pressure chemical ionization (DAPCI),[14] directanalysis in real time (DART),[15] extractive electrospray-ionization (EESI)[16] and matrix-assisted laser desorption/ionization (MALDI)[17] can be used to analyze melamine. Inaddition, immunological assays have been developed forthe detection of melamine.[18]

Despite these numerous techniques for melamine analysis,the usually used methods are LC-based. These methodsuse different column materials, together with a UV or massspectrometer as detector. Depending on the matrix and samplepreparation, the limits of quantification (LOQs) for HPLC-UV/DAD have been reported to be within 0.05–65 μg/g.[10]

However, when using a UV detector a special sample pre-paration and a good chromatographic separation are needed,due to difficulties in quantification because of matrix effects.This limitation is not encountered in LC/MS-based methodsdue to high selectivity using the tandem mass spectrometry(MS/MS) mode.[11] Different LC/MS-based methods withLOQs between 0.01–0.25 μg/g for melamine and cyanuricacid were developed.[11,19–22] Nevertheless, LC/MS-basedmethods often require an intensive sample preparation andlong analysis time.

New ionization techniques such as DAPCI, DESI, DARTor EESI have a very simple or no sample preparation andtherefore enable a high sample throughput. Thus, thesemethods can only be used as a screening method, becausea valid quantification is not possible.

Copyright © 2011 John Wiley & Sons, Ltd.

Figure 1. Structures ofmelamine, ammeline and cyanuric acid.

Determination of melamine and its derivatives by MALDI-TOF-MS

In contrast, by using matrix-assisted laser desorption/ionization (MALDI), a high sample throughput as well asa valid quantification were achieved. Due to the high salt andbuffer tolerance of MALDI, only a simple sample preparationis necessary.[23] Direct quantification using MALDI hasalready been shown for several small molecules.[24–29]

The success of a MALDI experiment is highly dependenton the sample preparation. Different dedicated samplepreparation protocols, such as fast evaporation, multi-component matrices and thin layer preparations, wereapplied to minimize the effects of inhomogeneous crystalliza-tion and co-crystallization of the matrix and the analyte.[30–32]

This inhomogeneous crystallization can cause a high fluctua-tion of ion signal intensity in single laser-shot MALDImass spectra.[23] Therefore, it is necessary to accummulate asufficiently high number of laser-shots. In order to compensatefor the remaining variations, an internal standard (IS) is anecessity for quantitative analysis.[23]

Also, the choice of the organic MALDI matrix is critical forsample preparation. Early studies compared differentmatrices such as a-cyano-4-hydroxycinnamic acid (CHCA),sinapinic acid (SA) and 2,5-dihydroxybenzoic acid (DHB).[33]

CHCA yielded the best results in terms of sensitivity andreproducibility in quantification of small molecules in thepositive ion detection mode. Therefore, CHCA is mostly usedfor the analysis of small molecules.The aim of this study was to develop a simple, sensitive

and fast matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) method forthe determination of melamine and related compounds in milkpowder. Moreover, we also want to show that MALDI-MScan generally be used for the fast qualitative and quantitativeanalysis of small molecules.

284

EXPERIMENTAL

Materials

The following chemicals were purchased from Sigma-Aldrich Laborchemikalien GmbH, Seelze Germany: atrazinePESTANALW, cyanuric acid-13C3 VETRANALW, melamine-triamine-15N3 (80–90 atom % 15N (triamine), 10–20 atom %15N (triazine)), ammeline PESTANALW, cyanuric acid (≥98%)and melamine (≥99%). For solvent preparations formic acid,trifluoroacetic acid (TFA) (both Carl Roth GmbH & Co KG,Karlsruhe Germany), acetonitrile (ACN) (VWR International,Darmstadt, Germany) and Milli-Q organic free water(Millipore, Purelab, ELGA Lab Water, UK) were used. Sinapinicacid and a-cyano-4-hydroxycinnamic acid were purchasedfrom Bruker Daltonik, GmbH, Leipzig, Germany. Milk powderwas obtained from Sucofin, TSI GmbH & Co. KG, Zeven,Germany. All chemicals and solvents were of high purityif not otherwise mentioned.

Copyright © 2011Rapid Commun. Mass Spectrom. 2011, 25, 2844–2850

Sample preparation

Method development

CHCA: The first experiments were carried out using CHCA,known as standard matrix for small molecule analysis. A dilu-tion series of melamine was prepared in water in the range12.6 ng/L to 630 μg/L. The internal standard for the quantifi-cation was atrazine at a constant concentration of 43 μg/L.Sinapinic acid: For further experiments, the use of SA asMALDImatrix was more practical. The solvent for all solutionswas 80% ACN (v/v) with 0.01% TFA. The solvent also servedas blank sample. The sample preparation using SA was asfollows. Stock solutions of melamine, ammeline, cyanuric acid,cyanuric acid-13C3 (IS) and melamine-triamine-15N3 (IS) witha concentration of 1 g/L were prepared. Two separatedilution series were prepared from these stock solutions.The first contained a 1:1 mixture of melamine and ammelinein a concentration range of 10–2000 μg/L. Melamine-triamine-15N3 was used as IS with a concentration of100 μg/L. The second dilution series contained cyanuric acidin a concentration range of 10–2000 μg/L. Cyanuric acid-13C3

with a concentration of 1000 μg/L was used as IS.

Milk powder

CHCA: The sample preparation of the milk powder is basedon a simplified preparation protocol according to FDAstandards.[19] Milk powder (0.1 g) was spiked with melaminesolution in the concentration range between 0.25–50 μg/g.After addition of atrazine (IS), the sample was filled up to atotal volume of 1 mL with 2.5% formic acid (v/v). The samplewas sonicated for 15 min and centrifuged for 5 min at14000 rpm. Then 50 mL of the supernatant was diluted to avolume of 1 mL with 80% ACN (v/v) and sonicated againfor 15 min. An aliquot of 1 mL of this solution was spottedon the MALDI target plate.Sinapinic acid: Stock solutions of melamine, ammeline,cyanuric acid, cyanuric acid-13C3 (IS) and melamine-triamine-15N3 (IS) with a concentration of 1 g/L were preparedin 2.5% formic acid (v/v). From these stock solutions twomelamine standards (10 and 1 mg/L), two ammeline stan-dards (10 and 1 mg/L), two cyanuric acid standards (100and 10 mg/L) and one melamine-triamine-15N3 standard(10 mg/L) were also prepared in 2.5% formic acid (v/v).

The milk powder sample was prepared according to aFDA protocol.[19] Milk powder (0.1 g) was spiked with thestandard solutions to obtain the following concentrations:melamine (0.1, 0.5, 1, 2, 4, 6, 8, 10 and 25 μg/g), ammeline(0.5, 1, 2, 4, 6, 8, 10, 20 and 40 μg/g), cyanuric acid (5, 15,25, 50, 75 and 100 μg/g). The internal standards melamine-triamine-15N3 and cyanuric acid-13C3 were spiked at concen-trations of 4 and 200 μg/g referring to the milk powder. Thevials were filled up to a volume of 700 mL with formic acid2.5% (v/v). Non-spiked milk powder served as blank sample.The sample was mixed and sonicated for 15 min and centri-fuged at 14000 rpm for 1 h. Then 100 mL of the supernatantwas mixed with 200 mL of ACN and centrifuged againfor 15 min at 14000 rpm and 100 mL of the supernatant wastransferred to a new vial. An aliquot of 1 mL of this solutionwas spotted on the MALDI target plate as described below.These final sample solutions were stored at �20 �C andremained stable for at least 2 weeks.

wileyonlinelibrary.com/journal/rcmJohn Wiley & Sons, Ltd.

5

A. Arnold et al.

2846

Matrix preparation and MALDI Spotting

The MALDI matrices CHCA and SAwere prepared at a con-centration of 1 g/L in 80% ACN (v/v) with 0.01% TFA. Allsamples were spotted using the dried-droplet preparationtechnique on a 384-well insert Opti-TOF-stainless steelMALDI plate (AB SCIEX, USA). First, 1 mL of the samplewas spotted and left to dry completely. Afterwards, 1 mL ofthe matrix solution was spotted on the dried sample andallowed to dry.[24]

MALDI-MS experiments

All MALDI experiments were carried out on a 4800 MALDITOF/TOF™ Analyzer (AB SCIEX, USA) which is equippedwith a 200-Hz Nd:YAG laser (355 nm). For data acquisitionand processing, the 4000 Series Explorer version 3.5.3 andthe Data Explorer version 4.8 software (AB SCIEX, USA) wereused. MALDI spectra of melamine and ammeline wereanalyzed in positive reflector mode in the m/z range 50–300with optimized fixed laser intensity. The MALDI massspectra of cyanuric acid were acquired in the negativereflector mode in the m/z range 100–800. 2000 shots perspectrum were accumulated randomly from different spotpositions. Spectra with signal-to-noise (S/N) values below200 (for each signal) were automatically discarded. Toimprove the mass accuracy of the instrument (<20 ppm)all spectra were internally calibrated using matrix signalsand the signal of the internal standard. The intensity of themonoisotopic protonated analyte signals [A + H]+ were usedfor data analysis.

Method validation

All measurements were performed in triplicate. Linearity,accuracy and relative standard deviation (RSD) were calcu-lated using MicrosoftW ExcelW 2007 SP1. The linearity wasdetermined by plotting the signal intensities of the analyterelative to the signal intensity of the internal standard, versusthe concentration of the analyte. The limit of detection (LOD)and the LOQ were determined in two different ways. First,the LOD and LOQ were determined experimentally basedon their S/N ratios. Therefore peaks with S/N ratios of atleast 3 and 10 were used for determination of LOD andLOQ. Additionally, data analysis according to the GermanDIN standard 32645 was made.[34] The LOD and LOQ werecalculated based on the calibration curve and the followingparameters: three replicates, uncertainty 33.3% and probabilityof error (alpha) 0.05. The calculated RSD and accuracy valuesalso had to fulfill FDA � 15/20 criteria.[35]

RESULTS AND DISCUSSION

The analysis of molecules with low molecular masses (below800 Da) by MALDI-MS is generally not as established as thedetermination of larger biological molecules such as peptidesor proteins. One major challenge is the intense background ofmatrix-ion signals (ma), which span the range from matrix-molecule fragment ions to cluster ions up to at least m/z 800.On the one hand, these signals and their strong intensitiesclearly pose a problem. On the other hand, they can be usedfor the internal calibration of the instrument, yielding a mass

wileyonlinelibrary.com/journal/rcm Copyright © 2011 John Wile

accuracy better than 20 ppm. As a result of this accurate massmeasurement and high resolution, the analytes were clearlyseparated from the interfering matrix background.

In addition, the application of MALDI-MS in the analysisof small molecules is quickly gaining confidence amongstresearchers, due to its tolerance to salts, detergents andother contaminants. Hence, time-consuming sample prep-aration steps can be avoided. Furthermore, with measure-ment times of about 10 s/sample, MALDI lends itself tohigh-throughput analysis. The identification of the analyteis assured by the high mass accuracy, resolution and theselectivity of MS/MS with modern MALDI-TOF/TOFinstruments.

However, a too low reproducibility has been argued tobe the main limitation in quantification studies usingMALDI-MS. To avoid this drawback, an optimized MALDIspotting procedure and the use of an internal standard arenecessary. To improve the precision, it is also helpful toaccumulate a sufficiently high number of laser shots permass spectrum.

Method development

CHCA

CHCA is a well-known standard matrix for the analysisof small molecules and has been used in different studiesfor the quantification of small analytes.[23–25,33] This matrixpossesses a better reproducibility and sensitivity than otherMALDI matrices and was used for our first experiments.The method development was carried out in a dilution seriesof melamine in water in a concentration range of 12.6 ng/Lto 630 μg/L. Atrazine was added as IS at a constant concen-tration of 43 μg/L. The LOD and LOQ of melamine weredetermined at 12.6 ng/L and 630 ng/L. The analyte wasquantifiable with an excellent linearity (R2: 0.9996), acceptableprecision (mean RSD 12.1%) and accuracy (mean error 9.7%)in the range between 630 ng/L and 630 μg/L. Also themeasurements in milk powder showed good results witha LOQ at 0.25 μg/g (data not shown).[36]

Sinapinic acid

Despite the advantages of CHCA in the analysis of smallmolecules, it was not suitable for the sensitive detection ofcyanuric acid in the positive or in the negative mode. Thisproblem could be solved by using sinapinic acid (SA) asMALDI matrix. The use of SA improved the analysis ofcyanuric acid significantly and also enabled the detection ofmelamine and ammeline in the same spot. Due to the lesshomogenous crystallization of SA compared to CHCA, it wasnecessary to verify whether the MALDI matrix SA is suitablefor a valid quantification of small molecules.

Therefore, standard solutions were analyzed first. Figure 2shows a mass spectrum obtained from a mixture of 50 μg/Lmelamine and 50 μg/L ammeline. Melamine (m/z 127.07) andammeline (m/z 128.06) are clearly visible as protonated singlycharged ions. Beside the analyte signals, the matrix signals[ma + H]+ (m/z 225.07) and [ma–H2O + H]+ (m/z 207.07) arevisible and can be used for internal calibration of the instrument.

To reduce the inhomogeneous co-crystallization of theanalyte with SA, a fast drying protocol was used, thatyielded a more homogeneous crystal pattern.[24] The

y & Sons, Ltd. Rapid Commun. Mass Spectrom. 2011, 25, 2844–2850

Figure 2. Mass spectrum of a sample containing melamine,ammeline (both 50 μg/L) and the IS melamine-triamine-15N3(100 μg/L) in 80% ACN (v/v), measured in positive iondetection mode. MALDI-matrix signals are labeled as ma.


combination of homogenous crystallization and the use ofmelamine-triamine-15N3 as IS achieved a good linearity (mel-amine R2: 0.9990, ammeline R2: 0.9997), accuracy and preci-sion. These results are summarized in Table 1.Melamine was quantified according to the FDA � 15/20

criteria between 25 and 250 μg/L. Ammeline was quantifiedin the range of 50–2000 μg/L. The experimentally determinedLOD and LOQ of melamine were 10 μg/L and 25 μg/L,respectively. Ammeline (LOD 25 μg/L, LOQ 50 μg/L) wasquantified in the same spot. All LODs and LOQs wereadditionally calculated according to the DIN standard

Table 1. Linearity, RSD and accuracy (as % Error) formelamine, ammeline and cyanuric acid all measured in80% ACN (v/v)

Conc.(μg/L)

Accuracy(% Error) % RSD R2

Melamine 10 43.35 6.72 0.999025 7.52 2.5550 2.81 3.27100 �6.09 0.64250 �0.87 0.611000 �21.79 2.302000 �20.39 1.45

Ammeline 10 n.d. - 0.999725 �53.03 3.4350 �4.53 11.98100 1.32 6.96250 �1.29 3.381000 1.33 4.412000 0.65 1.47

Cyanuric acid 10 72.72 1.11 0.999850 �2.40 2.19100 4.15 1.66500 �3.30 0.341000 �6.04 1.332000 1.16 0.13

n.d. not detectable.


32645.[34] This German Industrial Standard is widely usedin analytical method development and validation. The resultsare summarized in Table 2.

Because the combination of melamine and cyanuric acidincreases the toxicity, it is important to analyze both substancesin the sample. Figure 3 shows a mass spectrum of cyanuricacid (m/z 128.01) with the concentration 50 μg/L, measuredin negative mode. It was possible to quantify cyanuric acidin a concentration range of 50–2000 μg/L with an excellentlinearity (R2: 0.9998), accuracy (mean error between 2.4–6.4%)and precision (RSD between 0.13–2.19%). These results are alsosummarized in Table 1.

All in all, these results illustrate the excellent linearity,precision, and accuracy of this method, as well as the generalability of MALDI-TOF systems to serve in the identificationand quantification of small molecules. At this point, wewanted to verify the high sensitivity, reproducibility, and reli-ability of our newly developed MALDI-TOF-MS method, byinvestigating spiked milk powder samples.

Milk powder

Milk powder, a product with an extended shelf life, is easy tostore and transport. The nutritional composition of milkwidelyremains intact during the reprocessing to the powdered form.Non-fat milk powder contains 36% protein, 51% lactose anddiverse vitamins andminerals. Usually, such a complex samplematrix demands an exhaustive sample preparation, such assolid-phase extraction (SPE) or liquid-liquid extraction (LLE).

In our case, the sample preparation included only two dilu-tion and centrifugation steps, and is a simplified version of aFDA preparation protocol used for a LC/MS/MS method.[19]

A time-consuming LC separation after sample preparationwas not necessary for MALDI-MS analysis.

Milk powder was spiked with melamine, ammeline andcyanuric acid in the following concentration ranges: mela-mine 0.1–25 μg/g, ammeline 0.5–40 μg/g and cyanuric acid5–100 μg/g. Two different ISs were used, because melamineand ammeline were analyzed in positive mode, while cyanuricacid had to be analyzed in negative mode. As IS in positivemode, melamine-triamine-15N3 with a constant concentrationof 4 μg/g was used. In the negative ion mode cyanuricacid-13C3 with a constant concentration of 200 μg/g was used.In spite of the complex background of the sample and only asimple sample preparation, the analysis with MALDI-MSwas still possible. Figure 4(A) shows a mass spectrum of amilk powder sample spiked with a mixture of melamine(0.5 μg/g), ammeline (1 μg/g) and cyanuric acid (50 μg/g)measured in the positive ion mode. The most intense signalswere the MALDI matrix signals [ma–H2O + H]+ (m/z 207.07),

Table 2. LOQs and LODs of melamine, ammeline and cya-nuric acid measured in 80% ACN, determined using DINstandard 32645

LOD(μg/L)

LOQ(μg/L) R2

Melamine 7.31 27.37 0.9994Ammeline 25.80 98.76 0.9998Cyanuric acid 22.59 86.61 0.9998


2847

Figure 3. Mass spectrum of cyanuric acid (50 μg/L) andthe IS cyanuric acid-13C3 (1000 μg/L) in 80% ACN (v/v),measured in negative ion detection mode. MALDI-matrixsignals are labeled as ma.

A. Arnold et al.

2848

[ma + H]+ (m/z 225.08), [ma]+ (m/z 224.07), and [ma + K]+

(m/z 263.03). The inset shows the signals of melamine,ammeline and the IS. The good mass resolution and accuracyfor these signals underlines the potential of MALDI-TOF-MSfor the analysis of small molecules. The experimentally

Figure 4. (A) Mass spectrum of milk powder sample spikedwith melamine (0.5 μg/g), ammeline (1 μg/g), cyanuric acid(50 μg/g) and the IS melamine-triamine-15N3 (4 μg/g),measured in positive ion detection mode. (B) Mass spectrumof milk powder sample spiked with melamine (1 μg/g),ammeline (1 μg/g), cyanuric acid (12 μg/g) and the IS cyanuricacid-13C3 (200 μg/g), measured in negative ion detectionmode. MALDI-matrix signals are labeled as ma.


determined LOQ of melamine was 0.5 μg/g and fulfills theWHO guidelines, which set a limit of 2.5 μg/g in generaland 1 μg/g for infant formula. Melamine was detectabledown to a concentration of 0.2 μg/g and could be quanti-fied in the concentration range of 0.5–10 μg/g with excel-lent linearity (R2: 0.9994), accuracy (mean error between0.38–5.83%) and precision (RSD between 0.64–10.84%).Ammeline was also quantified (LOQ: 1 μg/g) with good lin-earity (R2: 0.9994), accuracy (mean error between 0.74–7.69%)and precision (RSD between 0.30–8.28%). The detection limitwas 0.5 μg/g. These results are listed in Table 3.

The low sample consumption in MALDI experimentsenables a second measurement of the same sample spots inthe negative mode. Figure 4(B) shows a spectrum of a milkpowder sample spiked with 12 μg/g cyanuric acid, 1 μg/gmelamine and 1 μg/g ammeline measured in negative ionmode. Cyanuric acid was analyzed in the concentration rangeof 5–100 μg/g with a good linearity (R2: 0.9985), accuracy(mean error between 3.48–8.47%) and precision (RSDbetween 0.11–13.26%). The detection limit was 5 μg/g, theLOQ was 12 μg/g.

All LODs and LOQs were additionally calculated usingDIN standard 32645. The results are shown in Table 4. Theexperimental and calculated LODs and LOQs are comparable.For example, the experimental determined LOD and LOQfor melamine are slightly better than the calculated values.For cyanuric acid the calculated LOD is slightly betterthan the experimentally determined LOD.

Table 3. Linearity, RSD and accuracy (as % Error) formelamine, ammeline and cyanuric acid all measured inmilk powder

Conc.(μg/g)

Accuracy(% Error) % RSD R2

Melamine 0.1 n.d. 0.99940.5 �3.04 1.211 5.83 0.642 0.38 1.364 2.18 1.986 �1.61 2.038 �1.99 10.8410 �1.96 5.1925 �48.24 8.63

Ammeline 0.5 53.77 1.86 0.99941 7.69 1.412 �0.74 4.204 �7.04 5.806 �2.94 0.308 �1.61 8.2810 2.67 6.2520 �2.80 4.4240 1.09 3.17

Cyanuricacid

5 �4.69 1.98 0.998515 �3.48 12.1125 �6.63 13.2650 5.99 1.1575 8.47 0.11100 �4.04 3.94

n.d. not detectable.


Table 4. LOQs and LODs of melamine, ammeline and cya-nuric acid measured in milk powder, determined usingDIN standard 32645

LOD(μg/g)

LOQ(μg/g) R2

Melamine 0.27 0.92 0.9997Ammeline 0.55 2.01 0.9996Cyanuric acid 3.33 12.02 0.9992


Melamine and derivatives have already been determinedby MALDI-MS, nevertheless only standard solutions withhigh concentrations (10 μg/mL) were analyzed.[17] MALDIwas also used for the analysis of melamine and melaminecyanurate in urine and urine stones.[37,38] However, bothmethods were only used for identification and not forquantification. In contrast to these methods, our developedMALDI-TOF-MS method permits the identification andalso the quantification of low concentrations of melamine,ammeline and cyanuric acid. The advantage of thismethod is that only a simple preparation step is required.In addition, only low amounts of chemicals and solventare used and the measurements are completed withinseconds.This study shows that MALDI-TOF-MS is a method with

promising potential for the high-throughput detection andquantification of melamine and its derivatives.

CONCLUSIONS

The present study describes a quick and easy method for theanalysis of melamine and its derivatives in standard solutionand in milk powder using MALDI-TOF-MS. A dried-dropletmethod with sinapinic acid as MALDI matrix was used. Itwas possible to quantify melamine in milk powder withinthe concentration range defined by the WHO guidelines.Furthermore, the quantification of ammeline and cyanuricacid in the same spot was demonstrated. The LOQ ofmelamine was 0.5 μg/g. Ammeline and cyanuric acidwere detectable at 0.5 μg/g and 5 μg/g. The LOQs ofammeline and cyanuric acid were 1 μg/g and 12 μg/g,respectively. The method showed excellent accuracy, pre-cision and linearity.The advantage of this method is that only a simple sample

preparation is necessary. No additional sample preparationsteps such as SPE or LLE are necessary. Also, a LC methodis not needed. Therefore, this method saves time, costs andreduces waste of chemicals. Furthermore, these experimentsdemonstrate the high impurity tolerance of MALDI-MS.Since the measurement time is short (~10 s/sample), thismethod lends itself to high-throughput analysis. Moreover,this study demonstrates the capabilities of MALDI-MS insmall molecule identification and quantification.

284

AcknowledgementThe work was funded by Hessisches Ministerium fürWissenschaft und Kunst (LOEWE Schwerpunkt ’AmbiProbe’).


REFERENCES

[1] J.R. Ingelfinger. N. Engl. J. Med. 2008, 359, 2745.[2] W.I. Jeong, S.H. Do, H. Jeong da, J.Y. Chung, H.J. Yang, D.W.

Yuan, I.H. Hong, J.K. Park, M.J. Goo, K.S. Jeong. J. Vet. Sci.2006, 7, 299.

[3] H. Ogasawara, K. Imaida, H. Ishiwata, K. Toyoda, T.Kawanishi, C. Uneyama, S. Hayashi, M. Takahashi, Y.Hayashi. Carcinogenesis 1995, 16, 2773.

[4] K. Burns. J. Am. Vet. Med. Assoc. 2007, 231, 1636.[5] R.L. Dobson, S. Motlagh, M. Quijano, R.T. Cambron, T.R.

Baker, A.M. Pullen, B.T. Regg, A.S. Bigalow-Kern, T.Vennard, A. Fix, R. Reimschuessel, G. Overmann, Y. Shan,G.P. Daston. Toxicol. Sci. 2008, 106, 251.

[6] B. Puschner, R.H. Poppenga, L.J. Lowenstine, M.S. Filigenzi,P.A. Pesavento. J. Vet. Diagn. Invest. 2007, 19, 616.

[7] R.E. Cianciolo, K. Bischoff, J.G. Ebel, T.J. Van Winkle, R.E.Goldstein, L.M. Serfilippi. J. Am. Vet.Med. Assoc. 2008, 233, 729.

[8] J.J. Litzau, G.E. Mercer, K.J. Mulligan. FDA-LIB 4423Melamine and Related Compounds, 2008.

[9] H. Miao, S. Fan, Y.N. Wu, L. Zhang, P.P. Zhou, J.G. Li, H.J.Chen, Y.F. Zhao. Biomed. Environ. Sci: BES 2009, 22, 87.

[10] World Health Organization. Ottawa, Canada, 2008.[11] A. Desmarchelier, M. Guillamon Cuadra, T. Delatour, P.

Mottier. J. Agric. Food Chem. 2009.[12] W.C. Andersen, S.B. Turnipseed, C.M. Karbiwnyk, S.B. Clark,

M.R. Madson, C.M. Gieseker, R.A. Miller, N.G. Rummel, R.Reimschuessel. J. Agric. Food Chem. 2008, 56, 4340.

[13] H.A. Cook, C.W. Klampfl, W. Buchberger. Electrophoresis2005, 26, 1576.

[14] S. Yang, J. Ding, J. Zheng, B. Hu, J. Li, H. Chen, Z. Zhou, X.Qiao. Anal. Chem. 2009, 81, 2426.

[15] T. Vail, P.R. Jones, O.D. Sparkman. J. Anal. Toxicol. 2007, 31,304.

[16] L. Zhu, G. Gamez, H. Chen, K. Chingin, R. Zenobi. Chem.Commun. (Camb.) 2009, 559.

[17] J.A. Campbell, D.S. Wunschel, C.E. Petersen. Anal. Lett.2007, 40, 3107.

[18] E.A. Garber. J. Food Prot. 2008, 71, 590.[19] S.B. Turnipseed, C. Casey, C. Nochetto, D.N. Heller. FDA-

LIB 4421 Melamine and Cyanuric Acid Residues in InfantFormula, 2008.

[20] M. Smoker, A.J. Krynitsky. FDA-LIB 4422 Melamine andCyanuric Acid Residues in Foods, 2008.

[21] M. Ibanez, J.V. Sancho, F. Hernandez. Anal. Chim. Acta 2009,649, 91.

[22] L. Yan, M. Wu, Z. Zhang, Y. Zhou, L. Lin, E. Fang, D. Xu, L.Chen. Se Pu 2008, 26, 759.

[23] L.H. Cohen, A.I. Gusev. Anal. Bioanal. Chem. 2002, 373, 571.[24] M. Persike, M. Karas. Rapid Commun. Mass Spectrom. 2009,

23, 3555.[25] M. Persike, M. Zimmermann, J. Klein, M. Karas. Anal. Chem.

2010, 82, 922.[26] M. Wagner, E. Varesio, G. Hopfgartner. J. Chromatogr. B Ana-

lyt. Technol. Biomed. Life Sci. 2008, 872, 68.[27] R. Rathore, J. Corr, G. Scott, P. Vollmerhaus, K.D. Greis. J.

Biomol. Screen. 2008, 13, 1007.[28] J. Gobey, M. Cole, J. Janiszewski, T. Covey, T. Chau, P.

Kovarik, J. Corr. Anal. Chem. 2005, 77, 5643.[29] P. Hatsis, S. Brombacher, J. Corr, P. Kovarik, D.A. Volmer.

Rapid Commun. Mass Spectrom. 2003, 17, 2303.[30] O. Vorm, P. Roepstorff, M. Mann. Anal. Chem. 1994, 66, 3281.[31] A.I. Gusev, W.R. Wilkinson, A. Proctor, D.M. Hercules. Anal.

Chem. 1995, 67, 1034.[32] T. Miliotis, S. Kjellstrom, J. Nilsson, T. Laurell, L.E. Edholm,

G. Marko-Varga. Rapid Commun. Mass Spectrom. 2002, 16,117.


9

A. Arnold et al.

2850

[33] L. Sleno, D.A. Volmer. Rapid Commun. Mass Spectrom. 2005,19, 1928.

[34] Deutsche Norm. Normenausschuss Materialprüfung (NMP) imDIN, 2008, ICS 71.040.01, 1.

[35] U.S. Department of Health and Human Services, Food andDrug Administration (FDA), Center for Drug Evaluationand Research (CDER), Center for Veterinary Medicine(CVM), 2001.


[36] M. Persike. Goethe University Frankfurt (Frankfurt/Main),2010. Available: http://publikationen.ub.uni-frankfurt.de/volltexte/2010/7875/

[37] H.W. Tang, K.M. Ng, S.S. Chui, C.M. Che, C.W. Lam, K.Y.Yuen, T.S. Siu, L.C. Lan, X. Che. Anal. Chem. 2009, 81, 3676.

[38] C.W. Kim, J.W. Yun, I.H. Bae, J.S. Lee, H.J. Kang, K.M. Joo,H.J. Jeong, J.H. Chung, Y.H. Park, K.M. Lim. Chem. Res.Toxicol. 2010, 23, 220.


Documents

Fast quantitative determination of melamine and its derivatives by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry