A Multi-Analyte LC–ESI-MS/MS Method to Analyze BPA, BADGE, and Related Analytes

LCGC Europe interviewed Stefan van Leeuwen from Wageningen Food Safety Research (WFSR), in Wageningen, The Netherlands, on his novel multi-analyte approach to investigate bisphenol A (BPA), bisphenol A diglycidyl ether (BADGE), and their analogues using liquid chromatography–electrospray ionization tandem mass spectrometry (LC–ESI-MS/MS).

Q. You recently developed a technique to analyze bisphenol A (BPA), bisphenol A diglycidyl ether (BADGE), and their analogues in food and beverages (1). How did this project come about and why are these analytes being investigated?

A: BPA is used as a building block to create polycarbonate plastics. These plastics are widely used in the automotive industry and many applications, including construction, electronics, CDs and DVDs, packaging material and thermal paper, but there is an intensive ongoing debate about the safety of BPA between scientists, national authorities, industries, and food safety authorities.

We have learned from other cases, such as perfluorooctane sulfonate (PFOS) or perfluoroctanoate (PFOA), that when a substance is under discussion, chemical manufacturers may move to produce alternative substances that have similar chemical structures. There are numerous examples of these so-called BPA-analogues, including, bisphenol-B (BPB), bisphenol-F (BPF), and bisphenol-S (BPS).

Several studies were published that reported the occurrence of these replacement chemicals in foods and human samples showing that we are being exposed to these chemicals (2,3,4). Our group wanted to study the situation for these chemicals in The Netherlands, and we needed to design a new analytical approach.

Q. You have developed a new multi-analyte method using liquid chromatography–electrospray ionization tandem mass spectrometry (LC–ESI-MS/MS).What is novel about this approach?

A: The method needed to be applicable to complex matrices found in foods and beverages. We successfully created a true multi-analyte method to analyze 23 BPA and BADGE analogues, a substantially higher number than in earlier published studies. The method also demonstrated excellent sensitivity because we choose to use alkaline MS ionization conditions rather than acidic conditions.

Q. What were the main analytical challenges you had to overcome?

A: First of all, we wanted to obtain a very sensitive method in the low‑ppb range in foods and low‑ppt range in beverages. We also wanted to resolve some important structural isomers. We learned that the structural isomers 4,4’-BPA, 2,2’‑BPA, and 2,4’-BPA could not be separated by MS/MS because they fragmented similarly, giving the same product ions. The same was true for 2,2’-BPF and 4,4’-BPF isomers. Therefore we experimented with three different ultrahigh-pressure liquid chromatography (UHPLC) columns with C18 stationary phases, modifiers (acetonitrile and methanol), and ionization aids (ammonium formate and ammonium hydroxide). The acetonitrile–water gradient containing ammonium hydroxide gave the best separation of 2,2-BPF and 4,4-BPF on all three columns investigated, whereas the separation of these compounds in methanol–water gradient was poor. The ionization with the ammonium hydroxide produced the best responses, and we ended up with excellent sensitivity of approx 1–10 pg on-column for most compounds included in our study.

Another difficulty we discovered is that if BADGE was present in an extract of, for example, a canned beverage, the in-source fragmentation leads to the transformation of BADGE into BPA. This BPA entered the mass analyzer and fragmented into the typical BPA fragments. We figured out that this “virtual” BPA (resulting from in-source BADGE fragmentation) eluted closely to the true BPA peak, meaning that potential misidentification was possible if no proper attention was paid to this issue. We did not need to adapt our methods in this case, but caution is needed to prevent misidentification.

Finally, all laboratories that work in the area of omnipresent environmental contaminants, such as perfluoroalkyl substances (PFASs), flame retardants, mineral oil saturated hydrocarbons-mineral oil aromatic hydrocarbons (MOSH-MOAH), and chlorinated paraffins are familiar with a major challenge, which is to keep the blanks low. Because of the wide application of these chemicals in many products, they are present everywhere, including in the laboratory environment. Dust particles contaminating your extract during sample preparation may alter the levels of your target analytes substantially, and one should take good care to work clean and avoid contamination of the sample during sample processing and analysis.

Q. What were your main findings?

A: Once we had the LC–MS/MS method on track, we wanted to screen a couple of real food and beverage samples. We designed a sample preparation strategy based on acetonitrile extraction (for solid samples), and sequential clean-up by mixed-mode solid-phase extraction (SPE) and silica SPE. The resulting purified extract was analyzed by our LC–MS/MS method. We did a small survey with the purpose of getting a first hint on which BPA analogues and BADGE analogues we would encounter in these food and beverage samples. 4,4’-BPA was detected in several samples, but also BPS, 2,2’-BPF, 4,4’-BPF, BADGE, and some BADGE analogues were detected. This shows that several analogues may be present in food.

It should be noted that detection of these substances does not automatically imply that there is a risk, but I would recommend researchers modify their methods to include more of these analogues. Our small-scale survey findings fit with data reported in other peer-reviewed studies on these analogues (1).

Q. Are you planning to use this multi-analyte approach for other applications?

A: The benefit of multi-analyte approaches is that it saves resources because you get more data out of the same analytical run. We therefore aim to design methods that can accommodate multiple compounds, or compound classes. In the area of environmental contaminants that enter the food chain we also use a multi-analyte approach (with tandem MS) for per- and polyfluoroalkyl substances. We analyze approximately 20 different PFASs in a single method, and continuously look at expanding this number.

A true multi-analyte approach is the LC-high resolution (HR) MS method we are currently designing for chloroparaffins. This environmental contaminant class consists of thousands of individual homologues and isomers that we would like to capture in a single method. Such complex mixtures are extremely challenging, and so far, no laboratory has been able to design an approach to detect the individual homologues and isomers needed to study the contamination patterns in foods and to support toxicological studies. It really is like finding a needle in a haystack, or worse! We currently use a modified method originally published by Bogdal et al. (5) and are now able to analyze different chain lengths and chlorination degree. But even then, we do not know which positional isomers we are looking at, so more development work is highly needed in that area. Also, the MOSH-MOAH mixture originating from, for example, printing ink residues in recycled food packaging have a complex nature.

Q. Have you used hyphenated tandem mass spectrometry for other areas of food and beverage analysis?

A: Most of the food control-related sample analysis in our institute is performed on gas chromatography (GC)–MS/MS and LC–MS/MS systems. These machines routinely run thousands of samples on pesticides, natural toxins, veterinary drugs, environmental contaminants, and many more compound classes. Next to that we employ the magnetic sector HRMS for dioxins and PCBs and we use the orbital ion trap mass spectrometers to analyze more complex matrices, or to work on identification of unknown compounds encountered in food or environmental samples. Hyphenation in our case also means automating sample preparation with on-line introduction of the sample into the GC–MS/MS or LC–MS/MS system.

Q. Do you have any practical advice for chromatographers who have not used tandem mass spectrometry before?

A: Tandem mass spectrometry combines excellent selectivity and unsurpassed sensitivity and is a very versatile technique. If you are looking for these characteristics when you design your analytical approach for targeted analysis, I would recommend tandem mass spectrometry. It is relatively easy to use and you can obtain data from multiple analytes in a single run.

Q. What other areas of food analysis are you currently investigating?

A: We are investigating a broad suite of environmental contaminants. I already mentioned the PFASs and MOSH-MOAH, and we routinely look at dioxins, PCBs, and brominated flame retardants. In recent years I have also become interested in the field of heat‑induced processing contaminants. For processing contaminants (acrylamide, AGEs, 3-MCPD, and furan), we study the effect of heating of foods in relation to the production of these contaminants. Obviously, a reliable analytical approach is instrumental to that aim.

Q. What is the future for the analysis of environmental contaminants?

A: We have seen large developments in detecting and identifying new environmental contaminants since I started working in this field 20 years ago. I look forward to the next 20 years, and I think several challenges are still ahead of us. There are approximately 5000 PFASs compounds that can potentially enter the environment and we need to find ways to resolve complex mixtures, such as chloroparaffins. That will keep me busy for a while!

References

1. S.P.J. van Leeuwen, T.F.H. Bovee, M. Awchi, M.D. Klijnstra, A.R.M. Hamers, R.L.A.P. Hoogenboom, L. Portier, and A. Gerssen, Chemosphere221, 246–253 (2019).
2. X.-L. Cao and S. Popovic, J. Food Protect. 78,1402–1407 (2015).
3. C. Liao and K. Kannan, J. Agric. Food Chem.61, 4655–4662 (2013).
4. H.J. Lehmler, B. Liu, M. Gadogbe, and W. Bao, ACS Omega3, 6523–6532 (2018).
5. C. Bogdal, T. Alsberg, P.S. Diefenbacher, M. MacLeod, and U. Berger, Analytical Chemistry87(5), 2852–2860 (2015).

Stefan van Leeuwen is a senior scientist at Wageningen Food Safety Research (WFSR, Wageningen, The Netherlands). During his Ph.D. on environmental analytical chemistry at the VU University in Amsterdam, he worked on method development for new persistent organic pollutants (POPs), such as brominated flame retardants and perfluoroalkyl substances (PFASs). Over the years he has worked at NIZO food research, Wageningen Marine Research, VU University Amsterdam, and in 2011 he joined RIKILT (current name Wageningen Food Safety Research, WFSR). Research on new environmental contaminants has been the main thread throughout his career, focusing on method development and food analysis. In recent years he has started to work on processing contaminants formed when food is heated.