Analysis of plant secondary metabolism using stable isotope-labelled precursors

Introduction: Analysis of biochemical pathways typically involves feeding a labelled precursor to an organism, and then monitoring the metabolic fate of the label. Initial studies used radioisotopes as a label and then monitored radioactivity in the metabolic products. As analytical equipment improved and became more widely available, preference shifted the use stable ‘ heavy ’ isotopes like deuterium ( 2 H)-, carbon-13 ( 13 C)- and nitrogen-15 ( 15 N)-atoms as labels. Incorporation of the labels could be monitored by mass spectrometry (MS), as part of a hyphenated tool kits, e.g. Liquid chromatography (LC) – MS, gas chromatography (GC) – MS, LC – MS/MS. MS offers great sensitivity but the exact location of an isotope label in a given metabolite cannot always be unambiguously established. Nuclear magnetic resonance (NMR) can also be used to pick up signals of stable isotopes, and can give information on the precise location of incorporated label in the metabolites. However, the detection limit for NMR is quite a bit higher than that for MS. Objectives: A number of experiments involving feeding stable isotope-labelled precursors followed by NMR analysis of the metabolites is presented. The aim is to highlight the use of NMR analysis in identifying the precise fate of isotope labels after precursor feeding experiments. As more powerful NMR equipment becomes available, applications as described in this review may become more commonplace in pathway analysis. Conclusion and Prospects: NMR is a widely accepted tool for chemical structure elucidation and is now increasingly used in metabolomic studies. In addition, NMR, combined with stable isotope feeding, should be considered as a tool for metabolic flux analyses.


| INTRODUCTION
Analysis of biochemical pathways typically involves feeding a labelled precursor to an organism, and then monitoring the metabolic fate of the label. Initial studies used radioisotopes as a label and then monitored radioactivity in the metabolic products. 1,2 As analytical equipment improved and became more widely available, preference shifted the use stable 'heavy' isotopes like deuterium ( 2 H)-, carbon-13 ( 13 C)and nitrogen-15 ( 15 N)-atoms as labels. 3,4 Incorporation of the labels could be monitored by mass spectrometry (MS), as part of a hyphenated tool kits, e.g. liquid chromatography (LC)-MS, gas chromatography (GC)-MS, LC-MS/MS. MS offers great sensitivity but the exact location of an isotope label in a given metabolite cannot always be unambiguously established. Although the detection limit for nuclear magnetic resonance (NMR) is quite a bit higher than that for MS, this technique can also be used to pick up signals of stable isotopes, and give information on the precise location of incorporated label in the metabolites. As the sensitivity and spectral resolution of NMR increase, this analytical tool can be expected to play an everincreasing important role in studying the regulation of biosynthetic pathways.
It was a fact universally acknowledged that the formation of IPP was via mevalonate (MVA), which in turn is synthesised by the condensation of two acetyl-coenzyme A (acetyl-CoA) units, giving acetoacetyl-CoA, and subsequent addition of a third acetyl-CoA resulting in the formation of β-hydroxy-β-methylglutaryl-CoA ( Figure 2A). This pathway had been established in the 1950s based on feeding experiments with 2 H-and 13 C-labelled precursors followed by careful mass spectrometric analysis of the formed products. 5 In the 1980s, the wider availability of NMR instruments allowed assigning the signals in 13 C-NMR spectra of hopanoids to the carbon atoms by reference to published spectra, and homonuclear 1 H-1 H correlation (COSY) and by heteronuclear 1 H-13 C correlation using 1 J coupling constants. It was then found that feeding [1-13 C]acetate to various bacterial species resulted in the formation of hopanoids with 13 C labelling patterns that differed completely from those that would be expected if these triterpenoids were made via the classical MVAbiosynthetic route. 6 13 C-label if the precursors were glycolytic products derived from [1,2-13 C 2 ]-glucose after, the newly discovered biosynthetic pathway was also found to be active in green algae, 9 and in higher plants. 10 It became clear that in plants two pathways are active that lead to the formation of IPP and DMAPP, and thus to all terpenoids, i.e. the MVA pathway and the deoxyxylulose phosphate (DXP) or methylerythritol phosphate (MEP) pathway. Both pathways start with precursors derived from glycolysis; the DXP/MEP pathwaywhich is localised in the plastidsstarts with pyruvate and D-glyceraldehyde-3-phosphate, whereas the MVA pathwaylocalised in the cytosolstarts with acetyl-CoA units which are formed by oxidation of pyruvate in the mitochondria. Although the pathways are compartmentalised, substrate exchange across membranes does occur. 11 Acetyl-CoA is a common precursor for a range of biosynthetic pathways, but any acetyl-CoA units that enter the plastids are used for fatty acid synthesis rather than terpenoid biosynthesis. 12 In order to assess whether a particular terpenoid is formed via the MVA route or via the DXP/MEP route, feeding experiments can be done using either [1-13 C]-glucose, [6-13 C]-glucose, or [1,6-13 C 2 ]glucose. 13 Glycolysis of these labelled glucose molecules will result in the formation of When [1,2-13 C 2 ]-acetate was fed to cell cultures of Taxus chinensis, 13 C label could be traced in the four acetyl groups of taxuyunnanin C, but not in the taxane ring system. Label from [1-13 C]glucose was incorporated into both the taxane ring system and the acetyl groups. These results indicate that the taxane carbon skeleton in this case was derived of the DXP/MEP pathway. 14

| BIOSYNTHESIS OF TROPANE ALKALOIDS
The tropane alkaloids are all derived from the amino acid ornithine.
Attempts to elucidate the exact individual steps leading from ornithine to hyoscyamine date back to the early 1950s. Initially, biosynthetic studies were done by feeding 14 C-or 3 H-labelled precursors to plants followed by isolation and careful chemical degradation of the metabolic products as a means to locate the position of the radioactive label. 26 Although sporadic reports mention NMR spectroscopic analysis locate the position of 2 H, 13 C, or 15 N stable isotopes after precursor feeding before 1990, NMR spectrometers only became widely available as a tool from the 1990s onwards. Then, feeding experiments with 13 C-labelled precursors, followed by 13 C-NMR analysis, confirmed previous findings and added further detail, e.g. feeding of  27 Whereas spin-spin coupling is a common occurrence in 1 H-NMR spectra, it is normally not seen in an 13 C-NMR spectrum due to the low abundance of 13 C atoms (1.1%); the chance that two naturally occurring 13 C-atoms in a molecule are contiguous is too small.
What remained a puzzle was that, though hyoscyamine is an ester of tropine and tropic acid, feeding experiments with radioisotope labelled precursors had shown that labelled phenylalanine and phenyllactic acid are efficiently incorporated into hyoscyamine, feeding The rearrangement could be made visible in the 13 C-NMR spectrum since the 13 C-1 0 and 13 C-3 0 atoms of littorine were rearranged to form the contiguous 13 C-1 0 and 13 C-2 0 in hyoscyamine ( Figure 5).
The rearranged contiguous 13 C-labels again resulted in the rarely seen split peak pattern in the 13 C-NMR spectrum due to spin-spin coupling. 3,31 The process by which littorine rearranges to hyoscyamine is still a matter of debate. 32-34 A cytochrome P450 enzyme CYP80F1, has been identified from Hyoscyamus niger. Expression of the gene in yeast confirmed that CYP80F1 catalyses the oxidation of (R)-littorine with rearrangement to form hyoscyamine aldehyde, a putative precursor to hyoscyamine. 35 The enzyme essays use GC-MS as a detection tool since this analytical technique has a much lower detection limit than NMR.
Another stable isotope that can be used to monitor bioconversion of alkaloids in plants is 15 N. This stable isotope has a low natural abundance (0.37%), and it is usually necessary to enrich the sample with 15 N before NMR analysis, e.g. by supplying plants with 15

| METABOLOMICS -FLUX ANALYSIS -PATHWAY ANALYSIS
Over the past few decades, more powerful NMR spectrometers have become available and increasingly higher resolution spectra could be analysed for identification and structure elucidation purposes. The increasing resolution has also made possible progress in the field of metabolomics and allowed it to be integrated into the other OMICS technologies that rely on shared massive databases.

| FUTURE DEVELOPMENTS IN CONTEXT OF PLANT RESEARCH
Any attempt to modify production bioactive plant secondary metabolites requires structural identification of individual precursors and endproducts, but also an in-depth analysis of the specific branches of the metabolic pathways of interest, and their integration within overall plant metabolism. 46 With the advent of functional genomics and metabolic engineering, analysis of secondary plant metabolism by NMR, using precursors labelled with stable isotopes becomes even more important as it can help in these analyses.  Plants, as multicellular eukaryotic organisms, have a more complex spatial distribution of metabolic pathways than bacteria, 52