Revolutionizing Food Safety: The Role of NMR spectroscopy in Food Science

NMR can efficiently and unbiasedly monitor food, leading to surprising new findings. The food industry is one of the most important industries for us as a species. With an increasing population, finding new ways of growing crops and improving food production is essential for continued survival. However, with new emerging food types and a larger yield of production than ever, new methods to monitor food production and quality are needed to ensure that what we eat is safe and healthy.
Søren Balling Engelsen and his group at UCPH Food are dedicated to analyzing and monitoring food production using different spectroscopic methods, with NMR spectroscopy being a key technology. NMR can investigate metabolites in an effective and unbiased manner, leading to all sorts of interesting findings relevant to food and health.

Søren Balling Engelsen master’s degree at DTU focused on vibrational spectroscopy and the development of molecular mechanical force fields. He also developed software programs for structural modeling of molecules.

In his PhD, also conducted at DTU, he further developed this technique to include modeling of lipids and carbohydrates, spending a year at Cornell University, USA, where he worked on molecular dynamics of carbohydrates in aqueous solutions.

Shortly after his PhD defense, Serge Pérez invited him to Nantes, France, where he spent three years as a post doc, working on computational models of polysaccharides.
It was in Nantes that Søren started working with NMR since the only detailed source of experimental structural and dynamical information was obtained through NMR spectroscopy. For example, NMR allows differentiation between the different conformers of the structures, which were otherwise very hard to decipher with other methods. Since Søren’s computer models mostly revolved around the equilibria between conformers, NMR spectroscopy ended up being essential for his work.

UCHP Food

While Søren was in France, his former supervisor Lars Munk from The Royal Veterinary and Agricultural University traveled to France to visit him together with Lars Nørgaard.

They told Søren about the new group they were starting at UCPH Food, focusing on chemometrics and multivariate data analysis, a research field combining principles from chemistry, numerical analysis and statistics, which turned out to use many of the same algorithms as Søren had used for his modelling work. Søren decided to accept the position and from 1995 onwards, he worked under Lars Munk taking over as group leader in 2004, a position he has held ever since.

Søren’s research at UCPH Food has revolved more around quantitative spectroscopic methods than the computational models themselves. So now, much of their work concerns quality control and monitoring of food production processes.

At UCPH Food, they have had low-field NMR instruments for a long time, and in 2004, they acquired their first high-field NMR instrument, which they are still using to this day. It has over the time been supplemented with one Bruker 500 MHz with sample robot and one Bruker iVDR 600 MHz w. SampleJet and SamplePro for metabolomics/foodomics work. The three spectrometers are physically collected in UCPH Plant Science Center NMR Laboratory at Frederiksberg.

Unique NMR methods

In Søren’s group, they primarily focus on quantitative ¹H NMR, which is quite unique in Denmark, as most other groups focus more on NMR for studying structure and dynamics of molecules. What they are investigating are the chemical components in complex mixtures, which is mostly a mixture of low molecular weight metabolites.

When Søren bought his first NMR instrument back in 2004, Bruker told him that he would never be able to obtain an uncertainty of less than +/- 10%. While this has been improved significantly by implementing strict standard operating procedures for quantitative ¹H NMR, getting accurate quantitative ¹H NMR data is still not trivial.

“Increasing the quantitative accuracy and precision of the NMR instruments is something we have spent a great deal of time trying to do. This has become increasingly important worldwide today as our need for absolute quantification and quality control is increasing.” Søren explains and during this development process, Søren’s group has discovered some surprising findings.

 In Søren’s research group, they utilize Gas Chromatography (GC) and Liquid Chromatography (LC), which are methods that have certain biases, which are mostly related to the interaction of some substances with the columns used in these techniques, thereby affecting their sensitivity. In essence, when using GC or LC, you need to have a clear idea of what you’re looking for to effectively measure it (targeted analysis). In contrast, with NMR, you have the advantage of being able to detect a wide range of compounds, including those you might not even be actively searching for (untargeted analysis).

“If you only measure what you think is relevant, you will confirm what you already know is important, thereby reducing your chances for discovery and innovation.” Søren emphasizes.

Lipoproteins and their importance in food and health

Lipoproteins play a crucial role in transporting triglycerides and cholesterol within the body, making them a significant focus of research in the food industry. Søren’s expertise in vibrational spectroscopy (VS), which he acquired during his Ph.D., has been instrumental in his work at FOOD.
This is because VS operates in a manner like NMR spectroscopy, involving the acquisition of a spectrum and its correlation to a reference area. However, these areas are typically not baseline separated in VS, wherefore machine learning and regression methods are required to decipher the results. They have developed the so-called “prediction methods”, where relevant parameters are predicted from the spectral data. This analogy has proven highly beneficial in his research within the field of metabolomics.

When working with plasma lipoproteins quantification, they focus on a small region in the ¹H NMR spectrum, around 0.6 to 1.4 ppm. This is where most lipid signals appear, thus enabling a correlation of the ¹H NMR signals to the lipoprotein-profile, which is classically measured by ultra-centrifugation.

“The fascinating thing about using these machine learning methods, is that we can predict anywhere from 50 to 100 lipoprotein fractions from this small spectral area, highlighting the impressive amount of information in contained in this spectral area and the sensitivity of the approach.” Søren explains.

Lipoprotein fractions refer to the various main fractions (VLDL, HDL, IDL and LDL) and their subfractions, all complex molecules responsible for transporting cholesterol and triglycerides in the bloodstream. The result of these efforts is that in any future plasma metabolomics studies the lipoprotein profile is provided as bonus information.

Sometimes Søren and his group have philosophical considerations of whether it really can be correct that these lipoprotein subfractions are fundamentally different or/and if they are only strongly intercorrelated. They call it the biological “cage of covariance”.  

In Søren’s group, they are also trying to apply the results from their NMR lipoprotein studies to diet and health interventions, where they are working together with the Institute of Sports and Nutrition (NEXS). Here they have conducted intervention studies to investigate how an “new Nordic diet” compared to “an average Danish diet, can affect people’s health. This was done, for example, by investigating all the small molecules in the human body, which are known as the metabolome. The dietary intervention lasted six months.

“From the results it was evident that there was a huge individual variation, which made it hard to draw any fundamental conclusions. However, they found biomarkers for a more plant-based diet, also including a few surprises.” Søren tells.

One example was 2,6-diisopropylnaphtalene, which they found to be more prevalent in plasma from people who ate a normal Danish diet than in those who did not. The only possible explanation was that this substance, which is used as an anti-sprouting agent for long term storage of potatoes, derived from the diet, specifically from French fries in the average Danish diet. Although surprising and very interesting, theses finding may not be very useful, but they show the utility of unbiased and non-targeted metabolomics, and its potential to uncover such surprising correlations in nature.

Norwegian Salmon

In Søren’s group, they look at a range of different food-related projects. A notable one has been a large Norwegian research project, which aimed to map the metabolome of the Atlantic salmon using  ¹H NMR spectroscopy.

Salmon is an important commodity in Norway., Recent studies have highlighted that the increased amount of plant-based feed in the diets of Atlantic salmon, a notoriously carnivorous species, has negative impact on their health, thereby affecting the quality of the final fish product.

By using ¹H NMR, Søren’s group has investigated whether there were differences in the metabolome of salmon collected from different aquaculture sites along the Norwegian coastline, which were possibly related to the fish feed and to the onset of gut disease.  They looked at plasma and the intestinal contents, both analyzed by ¹H NMR metabolomics. The results contributed to define and understand the dynamics of the onset of the salmon’s gut conditions and has ultimately helped the Norwegians aquafarmers to improve the quality of their salmon’s health.

What the future holds

With quality control of products becoming more and more important, the future of the food industry demands much more monitoring of the processes and their products. In this framework, Quality by Design and Process Analytical Technology can be helpful tools to optimize the products of the food industry with respect to quality and sustainability.

In the Sustainability Era, cellular agriculture stands as another important technology in the future of the food industry.  Cellular agriculture and biotechnology entail growing of foods and ingredients inside a bioreactor. Although Søren is convinced that ordinary agriculture will exist in the future, he also believes that a large part of our food will eventually be grown in bioreactors, as it is a much more efficient and sustainable way of producing foods.

“The best way to monitor bioreactors is by NMR, and therefore NMR could end up being an even more important tool in the future.” Søren explains.

Søren currently has a PhD student together with Novo Nordisk, where they used ¹H NMR to look at a protein production by E. Coli bacteria grown in bioreactors. In the project, they have investigated, which factors were affecting the product yield, including how the substrate was mixed, the temperature, and the way the substrate was supplied into the bioreactor.

They set up a small-scale stop-flow experiment where samples were drawn from the bioreactor and subsequently analyzed by ¹H NMR. The results highlighted metabolite formation and utilization as a function of cellular activity and ingredient production, and the relevance to increase the yield.

Søren is certain that these methods are going to be even more important as he explains “In the future, we are going to eat more plant-based foods, including many new foods we are not used to. In the past, we only ate around 20-30 different kinds of staple crops, so it is important to investigate the new crops and their processing as we are not entirely sure what if and how they will impact our health.“

“NMR is an essential tool for monitoring new food products and their related waste stream, but also to monitor the human metabolome after consumption of the food. Another reason NMR is such a great tool because it is unbiased and provides a good overview of the mass; there is almost nothing that it can’t do, the only issue is its relative low sensitivity.” Søren adds

Collaborations

Over the years, Søren has had numerus collaborations with important industrial partners including Novo Nordisk and Arla, the latter also financed his first NMR instrument.

Some other notable collaborations are FOSS, CP Kelco, Novozymes, and Christian Hansen, as well as many universities and institutes that present him with exciting projects. An example is the salmon project, where Søren collaborated with several universities/institutes in Norway, including the Norwegian University of Life Science (NMBU) and Nofima, and even some Danish fish feed factories such as the BioMar Group.

Strengths of Danish NMR

“The yearly DANNMR meeting, where all the groups gather to share ideas and discuss the future of NMR, is one of the great things about the Danish NMR society. Unfortunately, I am rarely able to attend the meeting since it often collides with my busy teaching schedule in January. I always make sure that people from my group attend.” Søren explains.

“Another important aspect of Danish NMR is that it is widely distributed but also built around centers (Danish Center for Ultrahigh Field NMR Spectroscopy in Aarhus and the SBiNLab: Structural Biology and NMR Laboratory in Copenhagen) with many resources, and it is great that these centers are very accessible, which makes it easy to get a sample tested.
I doesn’t use the centers much myself, but I have recently considered doing so as it could be useful to run some of our samples at a higher field.” Søren ends.

Written by: Jonatan Emil Svendsen