Applications of Time-Domain, HR-MAS, and Extraction-Based Liquid-State NMR Methods in the Compositional Analysis of Grass

Abstract

Poor nutrition in horses can lead to equine metabolic syndrome, which causes laminitis, a painful and life-threatening condition. It is therefore necessary to pursue greater metabolomic understanding of the grasses that horses eat. In this study, multiple NMR methods were explored as they relate to the compositional analysis of grass. Mahbub’s research in using time-domain NMR (TD-NMR) data and statistical software to predict grass compositions was continued. New grass samples were added to the existing calibration standard, but accuracy in predictions was maintained. 15 grass samples were selected for further metabolomics analysis by liquid-state and high-resolution magic angle spinning (HR-MAS) NMR. Grass extractions for liquid-state NMR were performed in D2O at room temperature, D2O at 90 °C, and in D2O:MeOD mixture at room temperature. One-dimensional 1H, two-dimensional total correlation spectroscopy (TOCSY), and 14N,1H heteronuclear single quantum coherence (HSQC) NMR experiments were conducted. Several metabolites were positively identified, including several key sugars and amino acids. Glucose and alanine absolute concentrations were calculated using the PULCON method and an alanine reference sample of known concentration. Positive correlations between glucose concentration and water-soluble carbohydrate content and between alanine concentration and crude protein content were observed. Lastly, qualitative comparisons between NMR spectra of different methods were made. It was found that some peaks appearing in extraction methods and some appearing in HR-MAS were mutually exclusive. No one method was determined to be optimal in future metabolomics studies, although the acquired knowledge will be instrumental in choosing one or the other for specific studies.