It is now common knowledge that nitrogen-rich industrial chemicals were used in China to fraudulently elevate the apparent protein content of foodstuffs. This practice first came to the public’s attention following the illness and deaths of domestic dogs and cats in North America during the spring of 2007. One of the largest forensic efforts ever undertaken by the U.S. Food and Drug Administration (FDA) traced the problem to the import, distribution and processing of grains that had been doped with chemicals to simulate high-protein gluten from wheat, rice or soy.

Once the animal sicknesses came to light, laboratory analyses identified the presence of melamine in certain feed ingredients. Also, kidneys of affected animals were found to contain crystals. It became apparent that these crystals—the actual cause of renal failure—were a virtually unknown health risk, consisting of macromolecular sheets of melamine complexed via strong hydrogen bonds with an oppositely charged, similarly shaped molecule called cyanuric acid.

The intent of the perpetrators was evidently to use industrial melamine to provide extra nitrogen to fool the Kjeldahl test for protein content. At first it was less clear how and why the cyanuric acid came to be present. Cyanuric acid, while an industrial commodity in its own right as a disinfectant, is also an impurity and a byproduct of melamine production. It could have been introduced into foodstuffs intentionally or otherwise if the violators used melamine production waste rather than high-purity melamine.

Hazard Detection
FDA scientists and others realized that the combination of melamine and cyanuric acid in the diet was a significant health hazard. A surveillance analytical method was developed in a matter of days based on gas chromatography/mass spectrometry (GC/MS) of derivatized feed extracts,[1] and this method was used to analyze thousands of field samples for melamine, cyanuric acid and the related compounds ammeline and ammelide.

In general, there are several goals of testing in the regulatory laboratory: (1) to detect residues of known chemicals, whether approved for some regulated purpose or not; (2) to create a deterrent against misuse of compounds; (3) to provide evidence of misconduct that can be used to support legal action via civil or criminal proceedings; and (4) to screen products and assure they are free of contaminants that pose a health risk. It is often difficult to combine all these features in a single method; sometimes separate screening and confirmatory methods are required. Although screening methods need to be relatively easy and widely applicable, full regulatory methods rely heavily on mass spectrometry, a capital-intensive approach that offers unsurpassed identification power and sensitivity.

The FDA’s GC/MS method served the immediate need to screen products and support regulatory action. However, this approach had some disadvantages as a general monitoring tool due to limitations of sensitivity and the need for chemical derivatization. An alternative method was sought, based on liquid chromatography/mass spectrometry (LC/MS) so that derivatization would not be required, with the capability for quantitation and confirmation for regulatory purposes. During 2007, the matrices of concern were understood to be a variety of grain-based ingredients and finished animal feeds.

Existing methods used solid-phase extraction (SPE) for sample cleanup. However, the different chemistries for melamine (cationic) and cyanuric acid (anionic) meant that two different SPE steps would be needed. On the other hand, it was very important to assay for both compounds because animal deaths were associated with the presence of both, not just melamine. It is the very tendency of these oppositely charged compounds to grip each other tightly that causes their co-precipitation in kidney tissues. However, co-precipitation in extract solutions or during analysis could become a significant impediment to method development.

New Method Development
At this point, there were several unanswered questions: Could melamine and cyanuric acid be extracted together and kept in solution together? Then, could both analytes be separated from co-extractants using good chromatography in one LC run?

The questions were answered with a final procedure that dovetailed several simple steps: (1) extract the sample with acidified water, (2) add a solvent to the water extract, (3) filter and (4) analyze with a specialized LC/MS procedure to detect melamine and cyanuric acid at separate times within a single analytical run. This type of method is sometimes referred to as “dilute-and-shoot.” The front end of the procedure requires minimal workup, while additional cleanup and detection are achieved with sophisticated instrumentation.

The first step of the procedure was to combine the extractant solution with solid sample in a 7:1 ratio. This ratio represents a balance between adding sufficient liquid to account for absorption by the sample and avoiding dilution of the extract below the instrument’s limit of detection. Water containing a strong acid was sufficient not only to dissolve both compounds but to completely disrupt any crystals of melamine cyanurate that might be present or might form in solution. These extract solutions were found to be stable for at least one month at refrigerator temperatures.

When an organic solvent (acetonitrile) was added to this extract, it caused other co-extracted material to precipitate, resulting in even cleaner extracts. Fortunately, it turned out that melamine and cyanuric acid remain dissolved in this acidified water-solvent mixture without forming the melamine cyanurate precipitate as long as their combined concentration was kept well below a critical level. Thus, a working calibration range was achieved that was still sufficient to measure the two compounds’ concentrations at levels of regulatory concern.

An important advance in the development of this method was the use of an LC column with a zwitterionic group covalently bound to silica particles. This type of column incorporates chemical groups with fixed positive and negative charges held closely together. This technique, known as ZIC-HILIC, has been found to be applicable to the analysis of multiple ionic compounds.

In this approach, analytes are introduced to the LC column in a solvent mixture that causes them to stick to the column. The injection solution is the same solvent mixture from the previous extraction step. When the sample is injected into an acidified solvent-water combination flowing through the column, cyanuric acid elutes smoothly in several minutes while melamine remains in place. The acidity of the solution passing through the LC column is carefully reduced, causing the melamine to be released and eluted from the column.

By causing the two compounds to be separated by their elution times, it enabled their detection at separate times by mass spectrometry. This was important because melamine requires detection in positive-ion mode, while cyanuric acid requires negative-ion mode. Some instruments are not designed for rapid polarity switching, which this LC separation rendered unnecessary.

Using the techniques described above, the two compounds could be fully separated and analyzed with two separate acquisition segments in a single run, enabling an efficient and sensitive analysis that met the regulatory goal.

New Method Validation
Various grain-based matrices were evaluated with the combined extraction and LC/MS method. Consistently good recovery was observed, but there was a disadvantage in that the “dilute-and-shoot” approach suffered some variability in the way each type of matrix influenced the mass spectrometer. Also, the salt content of the samples caused some changes in the ZIC-HILIC elution profile of melamine. Both problems could be compensated for by comparing the results from a test sample to results from a negative-control sample of the same product class.

The “dilute-and-shoot” method for animal feed was validated and described in a previous publication.[2] This method is routinely used to screen feed for animals used in the Center for Veterinary Medicine’s research into the pathology of melamine and cyanuric acid. More recently, the unfortunate news broke that melamine contamination had been found in powdered infant formula and other dairy products. The consequences for human health have been severe and are still under investigation worldwide.

Upon learning of the recurrence of melamine contamination, the FDA’s Denver District Laboratory and Animal Drug Research Center tested and validated the “dilute-and-shoot” method for powdered infant formula without modification.[3]

Summary
Analytical methods for melamine and cyanuric acid will continue to be developed and refined because there are many gaps in our understanding of how melamine and cyanuric acid behave once ingested. A recent World Health Organization report4 underscores the need for further research, and some of this work is already under way at the FDA Center for Veterinary Medicine. The “dilute-and-shoot” method based on ZIC-HILIC LC/MS is applicable in both regulatory and research environments. While this method serves an important regulatory function, it has become clear that new methods are also needed for routine testing of nutritional content. Tests for protein content need to be made specific for various proteinaceous compounds, rather than relying on a non-specific test for total nitrogen. These methods will need to be put into general use so that food producers, processors and distributors may regulate themselves. Furthermore, the current situation creates an impetus to search for other potential contaminants that might be used to circumvent the existing system of food safety controls.

David N. Heller is a residue chemist in the Office of Research, FDA Center for Veterinary Medicine (CVM). He has worked for CVM for over 19 years to develop, evaluate and apply residue methods to a variety of regulatory problems, primarily using mass spectrometry. He received a Bachelor’s degree in chemistry from Towson University in 1979. Prior to joining FDA/CVM, he worked at the Johns Hopkins University School of Medicine. Heller is an active member of the American Society for Mass Spectrometry.

Cristina Nochetto has developed and applied residue methods for more than 24 years. She has worked in the environmental field, for the U.S. Department of Agriculture and, since 2000, for FDA/CVM. She earned her Bachelor’s and Master’s degrees in biochemistry at the Universidad Nacional de Rosario, Argentina in 1979.

References
1. Litzau, J., G. Mercer, and K. Mulligan. “GC-MS Screen for the Presence of Melamine, Ammeline, Ammelide and Cyanuric Acid, Version 2.1.” FDA.gov, May 22, 2007. http://www.fda.gov/cvm/GCMSMelamine.htm (accessed December 17, 2008).
2. Heller, D. N. and C. B. Nochetto. 2008. “Simultaneous determination and confirmation of melamine and cyanuric acid in animal feed by zwitterionic hydrophilic interaction chromatography and tandem mass spectrometry.” Rapid Communications in Mass Spectrometry 22:3624-32.
3. Turnipseed, S., C. Casey, C. Nochetto, and D. N. Heller. “Determination of Melamine and Cyanuric Acid Residues in Infant Formula using LC-MS/MS.” CFSAN.FDA.gov, October 2008. http://www.cfsan.fda.gov/ ~frf/lib4421.html (accessed December 17, 2008).
4. World Health Organization. “Overall Conclusions and Recommendations: Expert Meeting to review toxicological aspects of melamine and cyanuric acid, held in collaboration with FAO and supported by Health Canada, Ottawa, Canada, 1-4 December 2008.” http://www.who.int/foodsafety/fs_management/conclusions_recommendations.pdf (accessed December 18, 2008).