Can biochemical phenotype, obtained from herbarium samples, help taxonomic decisions? – A case study using Gentianaceae

Abstract

With the emergence of metabolomics, small quantities of plant material can be used to generate chemical fingerprints with high throughput for comparative analyses without the need for tedious and in-depth classical phytochemical investigation of large amounts of plant material. Here, we tested the hypothesis that herbarium samples themselves could be used as suitable material for untargeted metabolomic analysis and taxon discrimination based on biochemical phenotypes. We employed mass-spectrometrybased metabolite profiling to classify given species of chemotaxonomical relevance from the Gentianaceae family based on their biochemical phenotypic differences. Our dataset consisted of 605 accessions each with 910 mass spectral features, representing 59 species from 23 genera that extend over 6 tribes of the family. We found that minute amounts of herbarium specimens (less than 50 mg) were sufficient to obtain comprehensive fingerprints for all investigated individuals. Morphologically or genetically distinct species were differentiated using metabolite profiles, confirming taxon distinctions and validating the utility of this method for identification and differentiation. Overall, our results suggest that metabolite profiling of herbarium specimens may be suitable for taxonomic studies and also opens new ground for detailed investigations of their specialised metabolite contents.

Navigating through chemical space and evolutionary time across the Australian continent in plant genus Eremophila

“…Morphologically, species in clade H14 have not previously been considered as closely related; however, strong phylogenetic support for this lineage, paired with the unique metabolome identified in the current study, has allowed these relationships to be revealed. This case underlines the utility of combining molecular phylogenetics and metabolite profiling: the observed distribution of particular chemical families supported unexpected phylogenetic relationships and furthermore raises questions about the potential interplay between specialized metabolism and species diversification for future investigation (Afzan et al., 2019 ).…”

“…Thereby, multiple public in silico spectral databases were used, i.e., GNPS, SUPNAT, NPAtlas, CHEBI, and DRUGBANK. Additionally, an Eremophila ‐specific in‐house in silico fragmentation database was used that contains 293 metabolite structures that have been characterized in Eremophila species during prior studies (Pedersen et al., 2019 , 2020 ; Singab et al., 2013 ; Zhao et al., 2019 ; Zhao et al., unpublished data). The in silico fragmentation‐based dereplication results were categorized by their reliability, with SMILES from Fusion, Consensus, and MetFrag algorithms corresponding to level 3a, 3b, and 3c identification, respectively.…”

Navigating through chemical space and evolutionary time across the Australian continent in plant genusEremophila

“…Morphologically, species in clade H14 have not previously been considered as closely related, however strong phylogenetic support for this lineage, paired with the unique metabolome identified in the current study, has allowed these relationships to be revealed. This case underlines the utility of combining molecular phylogenetics and metabolite profiling: the observed distribution of particular chemical families supported unexpected phylogenetic relationships and furthermore, raises questions about the potential interplay between specialised metabolism and species diversification for future investigation [55].…”

Integrating chemotaxonomic-based metabolomics data with DNA barcoding for plant identification: A case study on south-east African Erythroxylaceae species

“…These difficulties may be overcome by incorporating datasets from alternative identification techniques such as chemotaxonomy and DNA barcoding (Govaerts, 2001;Mishra et al, 2016;Mora et al, 2011;Scotland and Wortley, 2003). The effectiveness of chemotaxonomy as an identification and authentication tool has been shown (Afzan et al, 2019;Ahmad et al, 2010Ahmad et al, , 2009Endara et al, 2018;Heyman and Meyer, 2012;Mishra et al, 2016), with select studies investigating differentiating biomarkers related to specific Erythroxylum species (Johnson et al, 1998;Zanolari et al, 2005). However, to our knowledge, there have been no chemometric studies done on any of the south-east African coca species.…”