The research team focused on two cultivars with very different personalities: A. tricolor L. cv. Valentina, a red-leaved variety, and A. cruentus L. cv. Krepysh, a green-leaved type. Plants were grown outdoors at the Federal Horticultural Research Center for Breeding, Agrotechnology and Nursery in Moscow across the 2020–2021 seasons, with drought imposed after two months of growth by withholding water until substrate humidity fell to 20–25 percent, a condition maintained for seven days before sampling. Ten biological and ten technical replicates per treatment gave the researchers a statistically robust foundation for comparing stressed and well-watered leaves.

What they found was a dramatic metabolic reprogramming. In the leaves of cv. Valentina, drought triggered an approximately 70-fold surge in mannonic acid, a 40-fold rise in myo-inositol, a 23-fold increase in caffeic acid, a 15-fold jump in tartaric acid, and a seven-fold elevation of glycerol. The amino acids L-proline and serine climbed four-fold, while glycolic, oxalic, and lactic acids rose two to three times. These molecules are classic osmoprotectants and compatible solutes: by accumulating inside cells, they lower osmotic potential, helping leaves retain turgor and water even as the soil dries out around their roots.

Each of these metabolites plays a distinct protective role. Myo-inositol, for instance, is far more than a simple osmolyte; it is a core precursor in the phosphoinositide signaling pathway, which mediates intracellular signal transduction and orchestrates the downstream transcriptional activation of drought-responsive genes. Caffeic acid, a key intermediate in the phenylpropanoid pathway, boosts the plant’s non-enzymatic antioxidant capacity by scavenging the reactive oxygen species that accumulate when drought inhibits photosynthesis, thereby preventing lipid peroxidation and cellular damage. Flavonoids such as apigenin add another layer of defense, neutralizing ROS and maintaining cellular redox homeostasis.

In total, the team annotated 41 distinct substances across the leaf extracts, with 40 detected in cv. Krepysh and 34 in cv. Valentina. Nine of these compounds carry documented antimicrobial properties, including lactic, pyruvic, glyoxylic, malic, acrylic, and tartaric acids, along with glycerol, acetamide, and benzoic acid. Under baseline conditions, lactic, benzoic, malic, and mannonic acids were 40, 6, 2, and 1.5 times more abundant, respectively, in cv. Valentina than in cv. Krepysh. Glyoxylic acid appeared exclusively in the green-leaved cultivar, while acetamide was unique to the red-leaved one, underscoring the distinct biochemical identities of the two varieties.

Yet the study’s most provocative contribution lies not in the biology but in the analytical chemistry. Conventional high-throughput metabolomics suffers from a sobering limitation: typically only 5 to 15 percent of the spectral features in a complex mass spectrometry dataset can be unambiguously assigned using existing database-searching algorithms. Incomplete derivatization, analyte degradation, matrix effects, and co-eluting metabolites all conspire to blur the picture. Structurally similar compounds and stereoisomers often produce virtually identical mass spectra, and library matching scores below 750 or correlation coefficients under 0.98 leave annotators guessing, with real risks of structural misidentification.

To break through this bottleneck, the authors deployed a framework they call stochastic dynamics mass spectrometry, or SD-MS. Rather than treating mass spectra as static fingerprints to be matched against libraries, SD-MS models ion fragmentation and peak formation as stochastic processes governed by the electronic structures, ground-state energies, and vibrational frequencies of isolated ionic species. The mathematical equations link experimental measurands such as mass-to-charge ratios and ion peak intensities directly to intrinsic molecular properties and three-dimensional structural parameters, explicitly accounting for the random fluctuations and signal variances that conventional deterministic models ignore.

The framework proved its mettle on notoriously difficult cases. Fumaric acid, for example, yielded a low library score of 710 that could not distinguish its E- and Z-isomers, yet SD-MS correlation analysis achieved coefficients as high as 0.9965 for the E-isomer by comparing theoretical ion intensities with experimental data. For the flavonoid apigenin, whose spectrum is nearly indistinguishable from its structural isomer genistein, the method delivered a correlation of 0.99977, clearly separating the two. Silylated dicarboxylic acids and even bulky synthetic fluorinated derivatives of trifluoroacetic acid reached correlations up to 0.99985, far exceeding the performance of standard omics-matching algorithms.

The quantum-chemical engine behind these calculations relied on density functional theory with the hybrid meta-GGA M06-2× functional and basis sets including SDD, LANL2DZ, and aug-cc-pVDZ, executed in Gaussian, Dalton, and GAMESS-US packages. Geometry optimizations, transition-state validations, and Born–Oppenheimer molecular dynamics simulations provided the theoretical ion intensities that were then cross-correlated with experimental measurements using rigorous chemometrics, including Shapiro–Wilk normality tests and two-way ANOVA to account for random error sources across scan intervals.

Beyond the technical triumph, the work carries real agricultural weight. The enriched antioxidant and nutraceutical profile of cv. Valentina, with its superior adaptive potential under water deficit, marks it as a promising genetic source for breeding programs aimed at climate-resilient crops. By identifying specific biochemical markers of drought tolerance and pairing them with an analytical tool capable of definitive three-dimensional molecular identification, the study offers both a blueprint for understanding plant resilience and a practical foundation for engineering or breeding the drought-proof vegetables of the future.

Subject of Research: Drought-induced metabolite profiling of Amaranthus tricolor and Amaranthus cruentus leaves by GC–MS and stochastic dynamics mass spectrometry

Article Title: Gas chromatography–mass spectrometry molecular structural analysis of Amaranth leaf metabolites under drought stress

Article References: Feng, J., Gins, E., Ivanova, B., Motyleva, S., Gins, V., Gins, M., & Zhang, H. (2026). Gas chromatography–mass spectrometry molecular structural analysis of Amaranth leaf metabolites under drought stress. Results in Chemistry, 30, Article 103816. https://doi.org/10.1016/j.rechem.2026.103816

Image Credits: AI Generated

DOI: 10.1016/j.rechem.2026.103816

Keywords: amaranth, drought stress, metabolomics, gas chromatography–mass spectrometry, stochastic dynamics mass spectrometry, osmoprotectants, antioxidants, myo-inositol, caffeic acid, apigenin, quantum chemistry, plant resilience