Extending the breadth of metabolite profiling by gas chromatography coupled to mass spectrometry

Abstract

Gas chromatography coupled to mass spectrometry (GC-MS) is one of the most frequently used tools for profiling primary metabolites. Instruments are mature enough to run large sequences of samples; novel advancements increase the breadth of compounds that can be analyzed, and improved algorithms and databases are employed to capture and utilize biologically relevant information. Around half the published reports on metabolite profiling by GC-MS focus on biological problems rather than on methodological advances. Applications span from comprehensive analysis of volatiles to assessment of metabolic fluxes for bioengineering. Method improvements emphasize extraction procedures, evaluations of quality control of GC-MS in comparison to other techniques and approaches to data processing. Two major challenges remain: rapid annotation of unknown peaks; and, integration of biological background knowledge aiding data interpretation.

Figure 1

Number of publications published until 2006, querying the ISI database with the key words (metabon* OR metabolom* OR “metabol* profil*”) AND (gas chromatogr* OR GC).

Figure 2

Processing of GC-TOF metabolite profiles using two different software packages. Left panel: Depending on the parameter settings, processing GC-TOF netCDF files with the freely available AMDIS software may yield a high number of false-positive peak detections (indicated by black and orange stars) and a high number of false-negative (undetected) peaks. Right panel: Processing the same chromatogram with the instrument-specific ChromaTOF 2.32 version software does not miss peaks or report false deconvolutions in the retention-time window exemplified here. Note: The AMDIS report is visualized by a log-scaled intensity axis whereas the ChromaTOF ion traces are multiplied by factors ranging from 1x to 100x.

Figure 3

Substructure information (left panel) generated from mass spectrum (right lower panel) for bis(trimethylsilyl)-cytosine using the NIST substructure-recognition algorithm.

Figure 4. Example of identification of uncommon plant disaccharides inulobiose and levanbiose in transgenic potato tubers using derivatization and GC-MS

Left panel: Three unidentified peaks (1–3) were detected under methoximation and subsequent trimethylsilylation (TMS) with retention times close to that of sucrose (red ion trace, m/z 217). When using ethoxyamine instead of methoxyamine for the first derivatization step, compounds bearing keto- or aldehyde carbonyl moieties shifted to longer retention times (blue ion trace, peaks 1′–3′, m/z 217), whereas metabolites without such groups (such as sucrose) do not shift. Middle panel: Identical EI mass spectra were observed for peaks 1 and 2 (red spectrum), indicating that these may represent the syn/antiforms of a single methoximated compound. Best library hits and substructure recognition pointed to carbohydrates, specifically, fructose (black spectrum). Under ethoximation/TMS, some ion fragments shifted for 14 amu (blue spectrum), but the most abundant generic carbohydrate ions remained unaltered. Right panel: After database query, diverse authentic standards for fructosyl-fructoses were donated from plant researchers. Peaks 1 and 2 matched retention times and mass spectra for inulobiose (chemical structure in non-derivatized form), peaks 3 and 4 matched levanbiose.