Perspectives on modern NMR spectroscopy for personalised nutrition

Since the pioneer discovery of nuclear magnetic resonance (NMR) spectroscopy by Isidor Rabi in 1938, it has become a central analytical technology in multiple scientific domains of chemistry, physics and biology. Uniquely suited to measure the spin properties of magnetically active nuclei, NMR has emerged as a very popular technique for both routine and research applications. The food industry for instance uses NMR to study food structure, composition and effects on the metabolism. We briefly review in the following some key features making NMR a successful analytical platform in modern food and nutrition industry. Emphasis is given to recent developments of high resolution NMR (HR-NMR) spectroscopy for food quality and authenticity and nutritional metabonomics.

The popularity of HR-NMR in food and nutrition research relies first with a series of technical advantages such as minimal sample prepara – tion, non-invasiveness, reduced matrix effects, detailed structural information, quantitative capacity within a broad dynamic range and high reproducibility. HR-NMR remains a technique of choice for establishing the structure of molecules as well as for analysing complex food matrices. Indeed, minimal structural modifications introduced by various stereo – chemistry, chiral centre and position of functional groups result in measurable changes of chemical shifts, signal multiplicity and couplings that can be exploited. Organic molecules such as proteins or small molecular weight compounds are mostly studied by proton (1H), carbon-13 (13C), nitrogen-15 (15N), and phosphorous-31 (31P) spectroscopy with many possibilities to register either mono – dimensional or multidimensional mononuclear or heteronuclear spectra. Other nuclei such as deuterium, lithium-6, oxygen-17, fluorine-19 and sodium-23 can also be studied.

During the last few decades, HR-NMR has undergone major technological breakthroughs with the development of electromagnetic technology and advanced electronics. The introduction of high magnetic field instruments, above 900 MHz, and improvements made in probe design and technology, i.e. cryogenic probes, have allowed us to reach unprece – dented levels of resolution and sensitivity. Hyphenated systems with liquid chromato – graphy and mass spectrometry (MS) have also been developed in particular to provide new technical solutions for compound identification and structural elucidation. Important progress has been achieved on improving analytical throughput with automated interfaces for sample preparation and analysis, and play a pivotal role in the popularity of NMR for industrial applications.

HR-NMR, mainly using 1H spectro – scopy, has been increasingly used for food product profiling. Based on a metabonomics-like approach, the goal is to leverage holistic profiles of small molecular weight constituents into food compositional analy – sis using a set of various statistical processing methods. Recent applica – tions include the study of foods and beverages such as fish, fruits, oils and fats, wine, milk, coffee and tea. Compositional variations from the gilthead sea bream fish were captured by 1H HR-NMR and modelled with multivariate statistical analysis to differentiate wild from farmed fish and to classify farmed specimen according to their areas of production1. A similar approach was used for authentication of the origin of salmon2 based on the support vector machine analysis of NMR spectra of oils extracted from the white muscle of fish samples. The method exhibited very good classification rates of wild and farmed salmon and their country or origins. 1H NMR fingerprinting combined with chemometric techniques was also successfully deployed to classify olive oils according to their geographical origins and year of production3. In this work, the authors introduce the application of a new method combining high throughput 1H NMR profiling of olive oils using an automated flow injection device. In addition, several works are reported on the application of both 1H and 31P NMR to detect adulteration of olive oils4,5. A recent application highlights the potential of 1H NMR and pattern recognition techniques to determine the geographical origin of roasted coffee6. Although rarely used in routine analysis, the technology of HR magic angle spinning NMR (HR-MAS NMR) was applied to studying the chemical composition of coarsely ground coffee beans (Arabica and Robusta) and its variation during roasting temperature7. This technique is particularly interesting for food quality control as no particular sample preparation is required, the NMR compositional profiles being generated on gel or suspensionlike food samples. Although not exhaustive, the previous list of applications of HR NMR spectroscopy provides evidence that this technique will increasingly become a central analytical platform for food quality and authenticity. The relatively high cost of NMR technology can be absorbed by high turnover of analyses, minimal sample preparation and automated systems.

It is interesting to highlight the high degree of similarity between food profiling, i.e. foodomics, and metabonomics approaches. Indeed, both of them are conceptually based on the same analytical principle that associates the holistic compositional profiling of complex matrices with pattern recognition techniques for data exploitation. Metabonomics has become a very well established system biology approach to study the metabolism of complex living organisms and its relation with both intrinsic and extrinsic biological determinants including genetic, environmental and diet factors. Metabonomic analysis refers at the global metabolic profiling of low molecular weight compounds (<1,500 Da) in biofluids (plasma/serum and urine), and tissue extracts or biopsies. Such complex biochemical fingerprints of hundreds of metabolites reflect the overall metabolic status of an individual as the result of a complex web of molecular interactions during physiological regulatory processes. Nutritional Metabonomics is mainly based on the use of both HR-NMR and MS. Nowadays and similarly to food metabolic profiling, HR-NMR analysis of a biofluid can be achieved under a high throughput mode (usually 10 – 15 minutes per sample). The acquired metabolic profile contains high density latent biological information on the metabolic status of an individual. Interestingly, this technique can also be used for the metabolic profiling of intact tissue samples by HR-MAS NMR. Nutritional Metabonomics provides the unique oppor tunity to elaborate mechanistic hypotheses of nutrients on human metabolism. Another key outcome is undoubtedly the opportunity to characterise the metabolic phenotypes, or metabotypes, of individuals through metabolic profiling. This phenotyping approach offers the advantage to capture the biological end product of the interactions between genetics and environ – mental conditions of both the constitutive elements of a complex mammalian system, the host eukaryote cells and prokaryote gut bacteria cells. For this reason, metabonomics is recognised as a pivotal systems biology approach. Nutritional Metabonomics also aims to identify so-called biomarkers of homeostasis loss8. Such bio markers would be used to predict the individual likelihood for developing a particular health risk factor or even disease later in life. This would mean that even in the absence of patho physiological symptoms, biomarkers would signal the need to proceed with nutrition and lifestyle adaptations for health maintenance and disease prevention.

Current research efforts are therefore deployed to understand what particular nutrition will suit for individuals or individual groups. For this, the development of so-called prognostic biomarkers, similar to the concept of pharmaco-metabolomics developed for personalised drug therapies9, is increasingly tackled by the scientific community. To date, commercialised whole genome sequencing is available and coming down rapidly in price. However, the type of advice that is provided based on single nucleotide polymorphism profiling is questionable, given the current lack of validated prognostic biomarkers and the difficulty to provide functional and efficient actions on the metabolism through nutrition. In contrast to exclusive genetic analysis, metabonomics may appear to be an ideal tool for personalised nutrition since it provides the molecular and functional linkage between the phenotype, nutrition, lifestyle and environment. Nevertheless, the development of these biomarkers is challenged by the lack of large size nutritional and epidemiological studies involving systems biology approaches. Another challenge is that unlike xenobiotics, nutrition-induced metabolic effects are of relative low amplitude and are often difficult to detect in the presence of population heterogeneity and highly variable dietary habits. Therefore, population-scale and metabonomic epidemiological studies linking nutrition with health outcomes will be increasingly needed for biomarker discovery and validation. This implies the need for internationally well standardised analytical protocols for nutritional metabonomics. It also highlights the need to opt for analytical techniques delivering high quality and robust data as well as high throughput capability, key features that are definitely well developed in NMR spectroscopy. Finally, new computational approaches to integrate Foodomics with metabonomics will pave the way to understanding what are the key pattern of ingredients that associate with a particular metabolic response to diet opening in turn new research avenues to develop for personalised nutrition solutions for health maintenance and disease prevention.

References

1. Rezzi S. et al., Classification of gilthead sea bream (Sparus aurata) from 1H NMR lipid profiling combined with principal component and linear discriminant analysis. J. Agric. Food Chem. 2007; 55 (24): 9963-9968

2. Masoum S. et al., Application of support vector machines to 1H NMR data of fish oils: methodology for the confirmation of wild and farmed salmon and their origins. Anal. Bioanal. Chem. 2007; 387 (4): 1499-1510

3. Rezzi S. et al., Classification of olive oils using high throughput flow 1H NMR fingerprinting with principal component analysis, linear discriminant analysis and probabilistic neural networks. Anal. Chim. Acta. 2005; 552 (1-2): 13-24

4. Vigli G., Classification of edible oils by employing 31P and 1H NMR spectroscopy in combination with multivariate statistical analysis. A proposal for the detection of seed oil adulteration in virgin olive oils. J. Agric. Food Chem. 2003; 51 (19), 5715-5722

5. Mannina L. et al., 1H NMR-based protocol for the detection of adulterations of refined olive oil with refined hazelnut oil. J. Agric. Food Chem., 2009; 57 (24): 11550-11556

6. Consonni R. et al., NMR based geographical characterization of roasted coffee. Talanta 2012; 88: 420– 426

7. Ciampa A. et al., Studies of coffee roasting process by means of nuclear magnetic resonance spectroscopy. Journal of Food Quality 2010; 33 (2): 199-211

8. Rezzi S. et al., Nutritional metabonomics: Applications and perspectives. J Proteome Res. 2007; 6(2):513-25

9. Chen R, et al., Personal omics profiling reveals dynamic molecular and medical phenotypes. Cell 2012; 148 (6): 1293-1307

About the author

Following a university program in Biology, Dr. Rezzi obtained a PhD in 2000 on the application of NMR spectroscopy for the profiling of secondary plant metabolites. He pursued his career as a scientist at the Joint Research Center of the European Commission where he set up a metabonomic platform. In 2005, he joined the Nestlé Research Center (NRC) in Lausanne (Switzerland) as an NMR scientist and project manager. Dr. Rezzi has more than 10 years’ experience on metabolic profiling of living organisms. Since 2009, he has led a research group on Metabonomics and Biomarkers at the NRC.