Analytical Tools Developed for Nicotiana attenuata

Plants continuously adjust their metabolism to varying environmental conditions including the attack of herbivores and pathogens. Our group has substantial expertise in the identification and profiling of the volatile and non-volatile small molecules mediating some of these interactions. Picture 1 summarizes the techniques established in our group that we use to routinely analyze a large array of chemically diverse biologically active metabolites. These analytical screens are developed in close collaboration with the molecular biological platform as they play an essential role in the early stages of characterization of transgenic lines.

Overview of the analytical platform

Targeted trace analysis of hormones and low abundant molecules:

Temporal and spatial changes of phytohormone and oxylipin signatures, as well as of other low abundant small molecules, are routinely monitored by HPLC coupled to triple quadrupole mass spectrometers (1 Bruker EvoQ-Elite, 2 Varian 1200 LC-qqq-MS instruments) using the multiple reaction monitoring mode (MRM). Quantification is performed in most cases using isotopically labeled internal standards. These quantitative, high sensitive (fmol range) targeted analyses are a major element of our research and have recently helped to unravel some key regulatory steps in N. attenuata’s growth and defense regulation and herbivore interaction, e.g. the developmental control of phytohormone homeostasis in floral tissues [Sitz et al., (2014)].

Targets routinely screened in the group:

  • jasmonic acid (JA), JA derivatives (5; e.g. JA-amino acid conjugates, hydroxylated and carboxylated JA), and its precursors (e.g. 13-hydroperoxy-linolenic acid and 12-oxo-phytodienoic acid)
  • Salicylic acid (SA) and abscisic acid (ABA)
  • auxins (2), cytokinins (21) and gibberellins (15)
  • dodecenoic acid derivatives originated from green leaf volatile (GLV) biosynthesis (6)
  • free fatty acids, their oxygenated forms and their conjugates with amino acids (17)
  • 17-hydroxygeranyllinalool diterpene glycosides (11) Heiling et al., (2010)
  • nicotine metabolites and detoxification products (26)
  • amino acids (21)
  • phenolamines (9)
  • small peptides such as glutathione.
Bruker EvoQ-Elite QQQ-mass spectrometer for hormone analysis in the fmol range

Robust and simple high-throughput analysis of volatile organic compounds:

Headspace volatiles from leaves and flowers (e.g. GLVs, mono- and sesquiterpenes and aromatic substances, such as benzyl acetone) are routinely collected plants grown in the glasshouse and field. In the field (Ecological platform) and increasingly also in the glasshouse, headspace collections are conducted by passive adsorption using in-house-made pieces of silicone tubing (PDMS, polydimethylsiloxane), which are analyzed by GC-qMS (2 Shimadzu TDU-GC-MS instruments) after automated thermo-desorption (Figure 3). This technique facilitates the rapid, highly replicated and time-resolved screening of plant volatile bouquets involved in plant-plant and plant-animal communication [Kallenbach et al., (2014)] and even enables monitoring the volatile release from specific insect body parts [Kumar et al., (2014)].

Volatile analysis using silicone tubing

Targeted HPLC-UV/ELSD measurement of secondary metabolites:

Nicotine, phenol amines, diterpene glycosides and O-acyl sugars are the most abundant secondary metabolites of N. attenuata with established defensive functions against herbivores. Targeted screens based on HPLC-UV are routinely used in our group to quantify these metabolites. An evaporative light scattering detector (ELSD) is used for the in-line monitoring of compounds that lack a UV chromophore, such as O-acyl sugars and diterpene glycosides. Here is a list of targets routinely screened in our group by this procedure:

  • Nicotine, nornicotine, cotinine
  • Phenolamides
  • 17-hydroxygeranyllinalool diterpene glycosides
  • Rutin
  • Chlorogenic acid
SensorSense ETD-300 real time sub-ppb ethylene (C2H4) analyzer

Real time ethylene measurement:

We use a highly sensitive laser photo-acoustic detector with a detection limit of 0.3 ppb to measure ethylene emissions in real time from up to 6 plants or plant parts in parallel. This set-up allows for the quantification of ethylene production during pollination from single styles and has demonstrated the important role of ethylene during the mate selection process [Bhattacharya & Baldwin. (2012)].

Real-time, classical and non-targeted analysis of volatile organic compounds:

Real-time measurements of headspace volatiles in both, glasshouse and field conditions can be performed with a portable zNose [Bhattacharya & Baldwin. (2012)]. Volatile collections under constant flow and/or sample volume conditions are conducted using self-made super-Q traps with a constant flow push-pull technique. After elution with dichloromethane, volatiles collected with super-Q traps are routinely analyzed by GC-ion trap-MS (1 Varian 4000 GC-MS), GC-triple quad-MS (1 Bruker Scion GC-qqq-MS, Figure 4) and GC-FID (1 Varian 2000 GC-FID) [Schuman et al., (2012)].

Bruker Scion GC-triple quad-MS for routine volatile analysis

Two-dimensional gas chromatography GCxGC-ToF-MS (1 Leco Pegasus 4D GCxGC-ToF-MS, Figure 5) is used to obtain high chromatographic resolution which allows untargeted profiling of volatile bouquets [Garquerel et al., (2009)] or the analysis of specific volatile blends, e.g. alterations in cis-/trans-ratios caused by herbivore attack [Allman & Baldwin. (2010)] or food-derived body odors that tag herbivores for predation [Weinhold & Baldwin. (2011)]. Raw data are processed using software from the vendors or a combination of open-source software and proprietary multivariate statistics packages.

Leco Pegasus 4D GCxGC-ToF-MS

Sample preparation:

Agilent 7890A GC-FID connected to a Gerstel PFC

High throughput extraction procedures based on Fast-prep and Genogrinder instruments are routinely used for the extraction of the large arrays of metabolites found in leaves, roots, flowers and other plant parts. Extractions, as well as sample clean-up, can be conducted in 96-well-plate formats to increase sample capacity and decrease costs for solid phase extraction. For automated purification of volatile and non-volatile substances, we use an Agilent-HPLC connected to a fraction collector and an Agilent 7890A GC-FID connected to a Gerstel PFC [Stanton et al., (2016); Schuman et al., (2016)].

Analytical and in silico resources for metabolomics:

Bruker MicroToF for metabolomics analyses

Our group has developed a workflow for the metabolomics analysis of plant extracts. Most of these analyses are conducted using 2 ultra-high pressure liquid chromatography systems (UPLC) connected to either a Bruker micrOTOF or a Bruker micrOToF-QII. The UPLC systems provide highly reproducible chromatographic conditions that allow for the analysis of large data sets targeting a wide range of secondary metabolites.

Time-series analysis of multifactorial samples to evaluate transcriptional and metabolic dynamics and to construct gene-metabolite networks

We developed statistical methods to analyze metabolic profiling with complex experimental designs, such as multifactorial samples (time series, different tissues, treatments, accessions, etc.). We also use multivariate time-series data analysis to evaluate transcriptional and metabolic dynamics in N. attenuata by understanding the time of gene expression and the accumulation of metabolites for both pathway based and module based analyses that allow us to find assoications among induced genese and metabolites. These analyses allow us to construct gene-metabolite networks in which genes and metabolites from the same pathway are visualized as nearest neighbors in a network. These networks were generated after extracting specific modules from self-organized maps using bait-genes and correcting for the time lag behavior using temporal information. These analyses have allowed for the analysis of whole-plant molecular responses in time-course multivariate data sets and to simultaneously analyze stress responses in leaves and roots in response to leaf elicitations that occur when herbivores attack leaves[Gulati et al., (2014); Gulati et al., (2014); Gaquerel et al., (2014)].

Structural interpretation and quantification of high resolution ToF data:

For the structural interpretation of high resolution TOF data, we have adopted a versatile three-pronged approach integrating: 1) in vivo whole plant 15N isotope labeling; 2) a correlation-based grouping of the in-source fragmentation processes; 3) targeted and untargeted high resolution MS/MS measurements; 4) protein identification and quantification.

1) Nitrogen is an essential growth-limiting resource for plants, and despite being abundant in the post-fire environment in which N. attenuata grows, it is quickly depleted from the soil. In order to more precisely understand how N. attenuata allocates this limited resource among defensive metabolites (e.g., nicotine, phenolamides, etc.) and growth functions (proteins), we use labeling techniques by growing plant with a pulse of the stable nitrogen isotope 15N. This allows for the quantification of nitrogen incorporation into proteins using an LC-MSE-based method established in cooperation with the Mass Spectrometry/Proteomics Department of the MPI-CE [Ullman-Zeunert et al., (2012)]. The nitrogen incorporation into proteins can then be compared to that of defensive metabolites measured with a similar accuracy in our Bruker microToF UPLC-ToF-MS instrument, and, for example, we have shown that nitrogen used for biosynthesis of the rapidly induced phenolamides doesn’t come from degradation of photosynthetic proteins as previously thought [Ullman-Zeunert et al., (2013); Stanton et al., (2014)].

Spectral similarity networks of idMS/MS spectra used to mine mass spectral data of unknown metabolites

2) Indiscriminant MS/MS fragmentation analysis (idMS/MS) is routinely conducted in order to gain structural information on the overall metabolic profile detected by UHPLC-TOF-MS. idMS/MS assembly is achieved by correlational analysis between MS1 and idMS/MS mass signals for low and high collision energies. Furthermore, idMS/MS compound spectra are computationally assembled into spectral similarity networks, the biological information captured by this networking approach facilitates the mining of the mass spectral data of unknown metabolites (Li et al., 2016; Li et al., 2015).

3) Exploring the diversity of plant secondary metabolites is challenging and requires efficient methods to gain sufficient structural insight to quickly identify known compounds and enable identification of unknowns. De novo structure elucidation remains a major bottleneck, and improvements in the systematic analysis of MS fragmentation patterns of metabolite classes with complex decorations are needed to facilitate high-throughput compound annotation. As an examples, we are using the compound class of 17-hydroxygeranyllinalool diterpene glycosides (HGL-DTGs). HGL-DTGs are diverse plant secondary metabolites with various sugar and malonyl moieties. Containing multiple labile glycosidic bonds, HGL-DTGs exhibit extensive in-source fragmentation (IS-CID), which provide valuable information for structural elucidation. Using a deconvolution algorithm, we reconstructed IS-CID clusters and identified precursor ions. Those ions were selected to acquire high resolution MS/MS spectra and create an MS/MS database that formed the basis to generate a set of rules for a rapid dereplication of this compound class in various matrices (different species). In collaboration with Bruker Daltonics we establish a workflow that  is automatic and accelerates metabolite re-identification/annotation and is readily applicable to other compound classes with sufficient knowledge about fragmentation patterns. For the data analysis we use proprietary as well as open source software (Heiling et al., (2016)].

Work-flow for AMP protein quantification using nanoUPLC-MSE

4) In corporation with the Mass Spectrometry/Proteomics Department of the MPI-CE we conduct protein identification and quantification using a label-free nanoUPLC-MSE method. This method was developed for the determination of N incorporation into proteins (see above) and of small cationic peptides from crude apoplast fractions, where the accumulation of different heterologously expressed antimicrobial peptides (AMPs) could be quantified in transgenic N. attenuata plants [Weinhold et al., (2019], as well as to identify other regulatory proteins that interact with components of the plant’s clock.