Analysis of Melamine in Food

[Bruce Hamilton]

The problem appears to be that there were several sources of adulterant cheaply available in China - melamine, melamine scrap,
cyanuric acid ( which is cheaper than melamine ), cyanuric acid scrap, and a mixture of melamine and cyanuric acid.

Tragically, the high global price of grain ( biofuel and other reasons ), meant that farmers weren't able to feed their cows properly - even though they were paying more for grain.. That meant the milk quality was poor, hence the adulteration with the above chemicals to produce high protein readings.

As the regulators don't know which of the possible materials was used, and cyanuric acid was cheaper than melamine in China, they need to test for both.

Many of the above methods, including the Agilent HILIC method, will measure both melamine and cyanuric acid. Other related compounds
( eg ammelide, ammeline, ureidomelamine, methylmelamine ) may be present usually at lower concentrations, but I suspect they are mainly included to prevent future criminal adulteration.

The relevant World Health Organisation risk assessment is also available at:-
http://www.who.int/foodsafety/fs_manage ... lamine.pdf

Most methods should work with any suitable HILIC column.

Bruce Hamilton

[SIELC_Tech]

Bruce is right any HILIC column will work for this application, you can get a good HILIC column for about $300-400 (compare it to $1300 for Sequant HILIC).

Melamine can also be retained by mixed-mode cation exchange mechanism. It is basic in nature and slightly hydrophobic http://www.sielc.com/compound_404.html

Cyanuric acid will not retain too much on mixed-mode cation exchange. It is reverse order of elution in mixed-mode anion-exchange (Dionex and Sielc columns). Question is how critical it is to test both. My guess this problem might go away pretty soon and we will be looking for another chemical (triazine?) used to alter protein content. Another approach is to test milk producers on early stages before mixing stocks from tens of thousands of farms. Same story with any other food, milk ingredients. you need a significant amount of melamine to change nitrogen content for protein reading, so if somebody is adding melamine intentionally it is not on the level of ppm but rather %. Also melamine can come from some commercial and house hold plastics which are made with melamine based polymers.

We at SIELC developed melamine analyzer which will cost below 10K and will allow you to monitor melamine down to 1 ppm level. Approach will allow you do do direct injection of your milk products into analyzer:
http://www.sielc.com/Products_Phlex_Ana ... amine.html
....but there is always GC/MS, LC/MS and ELISA approaches to monitor melamine and cyanuric acid. Approach depends on the task and budget of the "testing parties"

[Einar Ponten]

Last night FDA posted two method updates

Laboratory Information Bulletin No. 4421

“Determination of Melamine and Cyanuric Acid Residues in Infant Formulaâ€

[Bryan Evans]

Imtakt just released new data using Unison UK-Amino:

LC-MS/MS of Melamine and Related Compounds:
http://www.imtakt.com/TecInfo/TI438E.pdf

It's true that many HILIC columns will work for melamine

My understanding is that this is the first time these 4 compounds
(melamine, ammeline, cyanuric acid, ammelide) have been
separated for LC-MS analysis.

[Einar Ponten]

This is discussed in a number of papers issued by the FDA, for summary see www.sequant.com/melamine

[thohry]

Hi all,

I am running melamine on Hilic column according to Agilentt method (QQQ). The mobile phase is : 5% A (H2O+5mM amoniac acetat) and 95%B (ACN+5mM amonium acetate) isocratic.

The problem is melamine peak is tailing quite severely while the peak in the ref is quite symmetrical.

Can anyone help me to overcome this problem.

Thanks for any ideas.

[Bryan Evans]

Maybe increase the buffer concentration in mobile phase,
sample solvent, or both.

[thohry]

Maybe increase the buffer concentration in mobile phase,
sample solvent, or both.

thank you, I will try it.

[Einar Ponten]

The plain silica column you are using actually acts as a cation exchanger at the close neutral pH you are using. This means that you mainly use the column in ion-exchange mode with these conditions.

Partly I agree with Bryan, but first you also need to use acidic conditions in the mobile phase. This is a characteristic, in my opinion, drawback with all plain silica columns ("HILIC columns").

At acidic conditions melamine will be charged and even more hydrophilic, but the silica will be less active as cation exchanger and therefore you will get a separation more based on hydrophilic interaction. Since the mechanism then mainly depends on partitioning the peaks will improve considerably.

Still, why invent the wheel again? The US FDA has recommended methods and we try to keep relevant information updated at www.sequant.com/melamine

[Bruce Hamilton]

I am running melamine on Hilic column according to Agilentt method (QQQ). The mobile phase is : 5% A (H2O+5mM amoniac acetat) and 95%B (ACN+5mM amonium acetate) isocratic.

The problem is melamine peak is tailing quite severely while the peak in the ref is quite symmetrical.

Can I just clarify my understanding. The melamine peak is tailing only in the samples, and not the reference standards?.

If that's the case, then maybe you should look closely at the SPE sample preparation process, especially the ultrasonication in the extraction step, and taking care with the water and methanol washes.

I'd also ensure that the sample eluates were taken to complete dryness under N2 before dissolving and filtering.

Whilst alternative ( eg FDA ) procedures are available, I assume Agilent procedures are also suitable, as Agilent are currently performing training in China using their methods.

Note that on 15 Oct, Agilent updated the SPE procedure listed in the article on their website, so make sure you are following the latest version. I suggest that you also contact the relevant people in Agilent, and ask for guidance.

Please keep having fun,

Bruce Hamilton

[Nalizer]

"SeQuant's invented wheel" is about $1,200. I get the same results on Atlantis HILIC which is 3 times cheaper. It is like you have a Chevy and put golden expensive rims on it instead of getting a regular one. (“They spinning, they spinning, they spinningâ€

[Bruce Hamilton]

I should note that Waters have just released a UPLC/MS/MS HILIC method with LLOQ of 20 ppb, that measures Melamine in infant formula. It doesn't seem to explicitly measure cyanuric acid, as the Agilent method does, but that may just be a simple omission.
Library Number: APNT10087429
Part Number:720002823EN

http://www.waters.com/waters/library.ht ... &xcid=9186

Please keep having fun,

Bruce Hamilton

[Nest Group]

I am running melamine on Hilic column according to Agilentt method (QQQ). The mobile phase is : 5% A (H2O+5mM amoniac acetat) and 95%B (ACN+5mM amonium acetate) isocratic.

The problem is melamine peak is tailing quite severely while the peak in the ref is quite symmetrical.

I'd also ensure that the sample eluates were taken to complete dryness under N2 before dissolving and filtering.

I suspect the problem with your real samples, thohry, is buffer exchange on the Agilent column. Bruce is correct to encourage going to complete dryness after sample prep, except that you also need to make sure the counter ions are the same in the sample as in the solvent, especially when the analyte elutes so early, as noted previously.

This is much less of a problem when the k' is 3. The value of the ZIC-HILIC column's stronger polarity is that it allows more water in the sample and mobile phase, and is less susceptible to this buffer exchange problem.

While you can do the melamine and cyanuric acid separations on lots of other HILIC columns, it is the composition of real samples that makes the case for why the FDA chose the ZIC-HILIC column. After all they have tighter budgets than most, so it seems that they would have looked at "cheaper" alternatives.

[SIELC_Tech]

Wow, intense discussion here (SeQuant and Nest Group double teaming people).
Anyway, in the last few weeks I spoke with chemists in different government organizations and here what they use:
1. FERN (Food Emergency Response Network) which is partially responsible for analysis of food. They are using the following equipment/columns
Waters Alliance 2695/Waters Micromass Quatro micro API with Phenomenex Synergy Polar RP (150x4.6, 4 um) with acetontrile/water/ammonium acetate.
2. FDA Labs are using three methods/approaches:
2.1 HILIC SeQuant (described here numerous times)
2.2 LC/MS (different one) with GL Sciences Inersil HILIC column (150x3.2, 5 um) wit ACN/water/ammonium acetate gradient. Retention time for cyanuric acid 3.6 min, melamine 5 min (void is 1.2 min for 0.5 ml/min flow rate).
2.3 ELISA method (non chromatographic)

3. Departments of Health for various states across US don't have a specified single method and use various approaches.

My personal opinion is, that it is insane to spend $1,300 of tax payers money (FDA) on something besides chiral, SEC or prep column if you can do the same analysis with $400...but this is just my opinion, although it is shared by one FDA scientists I talked to last week.

P.S. I represent another column manufacturer and I might be bias in this case.

[mfiligenzi]

We developed an early method on the Phenomenex Polar-RP that allows for the detection of all 4 melamine analogs at sub-ppm levels in feed, body fluids, and tissue samples. I'm not sure, but I think it was the basis for the FERN method. The conditions are in J Agric Food Chem. 2008 Sep 10;56(17):7593-9. Only problem with that method was a low level of background melamine (100 - 200 ppb in samples) that wouldn't go away. We're going currently looking at the HILIC columns to try to get around this.

The Polar-RP method worked quite well, though. We ran a lot of samples that way last year. It's fast and quite straightforward.

Mike