Elsevier

Phytochemistry

Volume 62, Issue 6, March 2003, Pages 859-863
Phytochemistry

Chemometric discrimination of unfractionated plant extracts analyzed by electrospray mass spectrometry

https://doi.org/10.1016/S0031-9422(02)00718-5Get rights and content

Abstract

Metabolic fingerprints were obtained from unfractionated Pharbitis nil leaf sap samples by direct infusion into an electrospray ionization mass spectrometer. Analyses took less than 30 s per sample and yielded complex mass spectra. Various chemometric methods, including discriminant function analysis and the machine-learning methods of artificial neural networks and genetic programming, could discriminate the metabolic fingerprints of plants subjected to different photoperiod treatments. This rapid automated analytical procedure could find use in a variety of phytochemical applications requiring high sample throughput.

Chemometric methods including discriminant function analysis, artificial neural networks, and genetic programming, could discriminate the metabolic fingerprints obtained from unfractionated Pharbitis nil leaf sap by direct infusion into an electrospray ionization MS.

  1. Download : Download full-size image

Introduction

Advances in analytical technology promise ever-expanding possibilities for metabolic characterization of: functions of new genes revealed by genome-sequencing programmes (Kell et al., 2001, Fiehn, 2002); physiological changes due to environmental stress (Fiehn et al., 2000b, Johnson et al., 2000); novel natural products (Wolfender et al., 1998); modes of action of bioactive compounds (Aranibar et al., 2001); or metabolic changes in transgenic crops (Noteborn et al., 2000). If plant metabolomics is to fulfil such aims, however, it needs to develop methodologies that encompass the diversity of phytochemicals, of which 200,000 have been estimated to occur in the plant kingdom (Fiehn, 2002), while achieving a sufficiently high analytical throughput to make the required experimentation feasible.

Hyphenated techniques, which couple GC or LC separations to MS, NMR or photodiode-array analysis, cope impressively with the complexity of minimally fractionated plant extracts (Wolfender et al., 1998, Fiehn et al., 2000a), but for high throughput the chromatographic step is a disadvantage, which is compounded when derivatization is required. In consequence, there is increasing interest in the analysis of plant extracts without chromatographic separation, using spectroscopic techniques such as NMR (Aranibar et al., 2001), FT-IR (Johnson et al., 2000) or pyrolysis MS (Goodacre et al., 1992). Such approaches may be termed ‘metabolic fingerprinting’ as they emphasize the pattern-recognition of metabolic phenotypes, rather than the cataloguing of specific compounds. Metabolic fingerprinting generally requires chemometric interpretation of the complexity resulting from simultaneous acquisition of analytical data on hundreds of metabolites (Goodacre et al., 1992, Beavis et al., 2000, Johnson et al., 2000, Aranibar et al., 2001).

Electrospray ionization (ESI)-MS, a soft-ionization technique that generates ‘molecular’ ions (Cole, 1997), has relatively unexplored potential for metabolic fingerprinting in plants. Schröder (1996) suggested that the ‘molecular’ ions in a complex sample may be sufficiently distinguished by their m/z values alone for the conventional LC column to be omitted and the unfractionated sample to be introduced directly into the ESI-MS, using flow-injection (Vaidyanathan et al., 2002) or direct-infusion (Zahn et al., 2001). ESI-MS has found recent application for the rapid characterization of microorganisms (Goodacre et al., 1999, Vaidyanathan et al., 2001), for rapid estimation of secondary metabolite expression in actinomycetes (Zahn et al., 2001), and for semi-quantitative determination of specific plant metabolites (Favretto et al., 2001).

The present study evaluates ESI-MS, in combination with a variety of post-analysis chemometric methods, as a possible new tool for rapid metabolic fingerprinting of complex phytochemical samples. Pharbitis nil was chosen as an experimental model for which advances in plant metabolomics might contribute to progress on a long-standing issue, the physiological nature of floral induction (Takeba and Takimoto, 1966, Durdan et al., 2000). We show that it is possible to discriminate metabolic fingerprints rapidly acquired by ESI-MS for P. nil leaves in different physiological states.

Section snippets

Results and discussion

The mean number of floral buds produced per P. nil cv Violet plant by 3 months after a single short day (SD) was 7.8±S.E. 2.1 (n=15), compared to only 0.4±0.24 on controls kept in long days (LDs). This confirmed that a single short day (SD) is sufficient to condition the transition from vegetative growth to flowering in this photoperiodically sensitive plant (Takeba and Takimoto, 1966, Durdan et al., 2000). Leaf sap, which is potentially appropriate material for flowering studies in view of the

Plant material

Pharbitis nil Chois cv Violet seed were obtained from Dr. R.J. Herbert, University College, Worcester, U.K. (Durdan et al., 2000). Plants were raised in John Innes No. 2 compost with weekly Chempak 3 feeds (Chempak, Hoddesdon, U.K), in a glasshouse (minimum 18 °C) under LDs of 16 h daylight supplemented when necessary with 400 W lights. After ca. 1 month, plants were either subjected to a single SD (8 h light/16 h dark) and then returned to LDs, or kept permanently in LDs as controls; thus, 6

Acknowledgements

We are grateful to Dr. M. Holland for technical assistance, Dr. R.J. Herbert for the plant material, and to the BBSRC Engineering and Biological Systems and Plant and Microbial Sciences Committees for financial support.

References (31)

  • S.F. Durdan et al.

    The determination time of the carpel whorl is differentially sensitive to carbohydrate supply in Pharbitis nil

    Plant Physiol.

    (2000)
  • D.I. Ellis et al.

    Rapid and quantitative detection of the microbial spoilage of meat using FT-IR spectroscopy and machine learning

    Appl. Env. Microbiol.

    (2002)
  • W. Eshuis et al.

    Some numerical aspects of reproducibility and specificity

  • B.S. Everitt

    Cluster Analysis

    (1993)
  • Cited by (114)

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

      2022, South African Journal of Botany
      Citation Excerpt :

      This highlights the drastic cost increase from in-house to academic to commercial rates. Furthermore, when comparing the academic cost of these methods to other plant differentiating methodologies such as high performance liquid chromatography (HPLC) (Viapiana et al., 2016), liquid chromatography-mass spectrometry (LC-MS) (Kharyuk et al., 2018), matrix-assisted laser desorption/ionization-mass spectrometry (MALDI-MS) (Ahmad et al., 2012), ultraviolet-visible spectroscopy (UV-Vis) (Philippidis et al., 2021) and electrospray ionisation-mass spectrometry (ESI-MS) (Goodacre et al., 2003), then it becomes apparent that LC-MS is the most expensive analysis and UV-Vis the least expensive. The average academic rates for analysing 10 samples at 30 min per sample using UV-Vis equates to ZAR 3880.60, with HPLC being 134% more expensive, followed by MALDI-MS (186%), ESI-MS (215%), DNA barcoding (230%), 1H-NMR (274%), GC-MS (314%), and LC-MS showing an increase of 334% in price.

    • A classification of liquid chromatography mass spectrometry techniques for evaluation of chemical composition and quality control of traditional medicines

      2020, Journal of Chromatography A
      Citation Excerpt :

      Direct MS and LC–UV–MS analysis may yield enough data to feed statistical models (PCA, PARFAC, PLS-DA, HCA, etc.), and the success of application of these models implies the possibility of machine learning application for distinguishing the samples. There were some attempts to develop such non-targeted annotation-free MS based analytical approaches [130–137]. Direct ESI–MS analysis was employed to discriminate metabolic fingerprints for P. nil leaves in different physiological states [130].

    • Metabolomics of medicinal and aromatic plants: Goldmines of secondary metabolites for herbal medicine research

      2020, Medicinal and Aromatic Plants: Expanding their Horizons through Omics
    View all citing articles on Scopus
    View full text