Metabolomics in Fundamental Plant Research

Abstract

This chapter overviews the latest developments in plant metabolomics, explicitly focusing on novel gene identifications, plant species identification, plant root metabolome, plant biomarkers, and plant single-cell metabolomics. This chapter explores the importance of identifying novel genes in plants and their role in enhancing crop yield and tolerance to environmental stresses. It also discusses the challenges of identifying plant species and the various techniques used to overcome them, including metabolite quantitative trait loci and genome-wide association studies. Furthermore, this chapter examines the plant root metabolome and its significance in plant growth and development. It highlights the role of metabolomics in plant biology research, particularly in identifying potential biomarkers for monitoring plant health and productivity. This chapter also covers the emerging field of plant single-cell metabolomics and its potential for advancing our understanding of plant metabolism at the cellular level. In addition, this chapter discusses the application of mass spectrometry imaging (MSI) techniques in plant biology research, particularly in visualizing and mapping the distribution of metabolites in plant tissues. It provides an overview of the different MSI techniques and their advantages and limitations in plant biology research. The information presented in this chapter is valuable to researchers and practitioners in plant biology. It can aid in developing new strategies for improving plant health and productivity.

1 Introduction

Plant metabolomics is an emerging research field focusing on the comprehensive analysis of metabolites, small molecules (<1500 Da) that play a critical role in plant growth, development, and response to environmental changes. By applying advanced analytical techniques, such as mass spectrometry and nuclear magnetic resonance spectroscopy, researchers can identify and quantify thousands of metabolites in a single sample, providing a holistic view of the metabolic pathways and processes that occur in plants. The insights gained from plant metabolomics research have broad implications for plant biology, agriculture, and biotechnology, as well as for human health and the environment.

Metabolomics is crucial to studying abiotic stress tolerance, pathogen resistance, robust ecotypes, and metabolic-assisted breeding of crops. The plant kingdom contains a considerable diversity of metabolites of approximately 200,000 compounds; the majority are still unknown. It is estimated that around 10,000 secondary metabolites have been discovered in different plant species [1], since metabolites have a wide range of physicochemical properties and functions. This makes it challenging for metabolomics techniques to study their diversity and gain insights into plant biology.

Since the word “metabolomics” was mentioned for the first time in the literature, it has evolved and been applied to many disciplines, such as plant biology. Plant metabolomics has become hugely modernized in the last decade. This chapter describes novel applications of mass spectrometry-based metabolomics approaches in recent fundamental plant research. We first commented on different metabolomics techniques, mass spectrometry, and nuclear magnetic resonance-based metabolomics. Mass spectrometry-based metabolomics is the widest in use; the main approaches are targeted or quantitative metabolomics and untargeted, nontargeted, or discovery metabolomics, also known as global metabolomics. We then explored contemporary literature on gene identification and their functional characterization for crop improvement enabled by applying nontargeted and targeted metabolomic analysis in combination with genome-wide association studies, metabolite quantitative trait loci, and transcriptomics.

From there, we explored the current application of metabolomics methods for plant species identification, a trendy topic that is enabling researchers to support many aspects of food authentication, food quality control, and traceability; plant species identification is an essential factor in understanding biodiversity, the discovery of bioactives from herbal medicines and correlating chemical components from plants with chemical markers of patients who intakes herbal medication, in the same manner monitoring food intake in foodomics studies.

We exemplified by commenting on the use of molecular networking analysis and its application to classify plant species; by providing an example of the Malpighiaceae family, chemotaxonomic studies guided by metabolomics methods become hugely in use as it allows rapid classification of plant samples based on the endogenous chemical content. We also discussed pioneering work on classifying and discriminating cinnamon, vanilla, and coffee plant species using different metabolomics techniques.

We then explored the impact of climate change on the root metabolome and the differences in root abiotic and metabolic changes associated with other biotic factors interactions, highlighting key metabolites involved in root exudates when exposed to these types of stresses. In addition, we reviewed recent literature on plant biomarker discovery outlining different application areas, such as the food industry, where the identification of biomarkers has also worked in quality processes for food authenticity and food traceability matrices of plant origin. We outlined a list of key metabolites identified in various plant species using different analytical techniques including metabolites detected in transgenic plants.

Finally, we briefly explored the emerging field of single-cell metabolomics methods. We describe the latest development in mass spectrometry imaging, including different approaches for collecting single-cell from plant tissues and a revision of some essential techniques on mass spectrometry imaging. Mention has also been made on the challenging and future needs for plant metabolomics research.

2 Novel Gene Identifications and Their Functional Characterization for Crop Improvement

Metabolomics studies all small molecules – metabolites – content of a cell or whole organism. Plant metabolomics refers to comprehensive, nonbiased, high-throughput analyses of complex metabolite mixtures typical of plant extracts. The role of metabolomics in such studies is twofold: (1) to identify the spatial and temporal distribution of the target compounds as influenced by plant development and environmental cues and (2) to identify related phytochemicals, which may be considered as either intermediate of biosynthesis or alternative or alternative products of promiscuous enzymes that support the biosynthesis of the target phytochemical [1].

This chapter describes the latest development and application of plant metabolomics in combination with metabolite genome-wide association studies (mGWAS), metabolite quantitative trait loci (mQTL), and transcriptomics for the discovery and characterization of genes and enzymes associated with the biosynthesis of specialized metabolites in significant crops such as maize, rice, and tomato (Fig. 1).

Fig. 1

Schematic presentation of metabolomics applications in crop improvement programs: (1) A representative sampling source of vegetable crop plants (tomato) as a biological source from which cellular metabolome can be extracted from almost all the plant parts and the rhizosphere under varying experimental environmental conditions; (2) data acquisition approach in metabolomics to be applied whether unbiased nontargeted fingerprinting is required or the analysis and quantification of a few selected target molecules is the need of the experiment; (3) study outcomes which needs biological interpretation for hypothesis questions; (4) possible answers to the hypothesis questions in the cellular chemistry and its entwining relations with the environmental impacts; (5) functional targets that could be achieved through metabolomics analysis of the vegetable plants; (6) result-oriented applications of the data outcomes in crop improvement practices; and (7) the “end product” of the experimental metabolomics exercise in vegetable crops. (Reproduced from Ref. [2])

Different metabolomics techniques have been developed in the last two decades, including mass spectrometry-based metabolomics (MS); nuclear magnetic resonance spectroscopy (NMR); gas-chromatography-mass spectrometry (GC-MS); capillary electrophoresis-mass spectrometry (CE-MS); liquid chromatography-mass spectrometry (LC-MS); and more recently the implementation of high-resolution metabolomics with the aid of Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS). Mass spectrometry-based metabolomics is one of the most used in plant metabolomics, where researchers can undertake two main approaches: targeted and untargeted metabolomics (Fig. 2). Targeted metabolomics is a hypothesis-driven approach focusing on a specific set of metabolites. This approach is often used when the researcher has prior knowledge about the metabolites of interest or when a specific metabolic pathway is under investigation. In targeted metabolomics, a set of known metabolites is selected, and the mass spectrometer is set up to detect these specific metabolites. Targeted metabolomics can provide more accurate and quantitative information about specific metabolites, but it requires prior knowledge of the metabolites under investigation; on the other hand, untargeted metabolomics is an exploratory approach that aims to identify all metabolites in every biological sample. This approach does not require prior knowledge of the metabolites under investigation, and it allows for detecting unexpected or unknown metabolites. In untargeted or nontargeted metabolomics, the mass spectrometer is set up to detect all metabolites within a certain mass range. The resulting data is analyzed using statistical methods and bioinformatics tools to identify metabolites. Untargeted metabolomics can provide a more comprehensive view of the metabolic profile of a biological sample. Still, it may miss some metabolites outside the mass spectrometer’s detection range.

Fig. 2

Illustrative diagram of possible plant environmental interactions, which are supposed to influence the metabolic status of the crop plants. Analyzing the metabolome of plants exposed to such challenges using metabolomics approaches can yield competitive vegetables crops with better yield, high level of defense and stress-mitigating capabilities. (Reproduced from Ref. [2])

The combination of several omics has been recently implemented on gene discovery and their functional characterization to investigate gene relationship to metabolites supporting and accelerating crop improvements. For example, Wu et al. established a high-quality chromosome-level genome assembly of Melilotus albus, by resequencing 94 Melilotus accessions to characterize their phylogenetic relationships, the genetic exchange between M. albus and M. officinalis, and the differentiation of flower color and coumarin content. In addition, transcriptomics, metabolomics, and bulked segregant analysis (BSA) have been used to investigate M. albus near-isogenic lines segregating at the coumarin level to identify the key metabolites and enzymes in the coumarin biosynthesis pathway [3]. Similarly, Li et al., based on the integrative analysis of the transcriptomics and targeted carotenoid research, found that differentially expressed genes (DEGs) related to carotenoid metabolism had a stronger correlation with the critical carotenoid metabolite content in the panicle of foxtail millet. Correlation and weighted gene coexpression network analysis (WGCNA) identified and predicted the gene regulation network related to carotenoid metabolism [4].

Zheng et al. used a combination of metabolome and transcriptome of 11 tea cultivars and then a WGCNA-based biological system strategy to interpret metabolomic flux, predicted gene functions, and mined critical regulators involved in the flavonoid biosynthesis pathway; in this manner, they revealed new uncharacterized transcription factors (TFs) such as MADS, WRKYs, and SBP; and microRNAs (including 17 conserved and 15 novel microRNAs) that are potentially implicated in different steps of the catechin biosynthesis. In addition, they applied the metabolic-signature-based association method to capture additional critical regulators involved in the catechin pathway. This provides important clues for the functional characterization of five SCPL1A acyltransferase family members, which might be implicated in the production balance of anthocyanins, galloylated catechins, and proanthocyanins in tea cultivars [5].

Another approach has been implementing metabolite genome-wide association studies (mGWAS) and metabolite quantitative trait loci (mQTL), especially in cereal grains such as wheat. Chen et al. developed an approach that has also been applied in other major crops such as maize, barley, tomato, and blueberry, while the mQTL approach has been used in crops sunch as soybean, rice, and carrot [6]. Traditionally the output of mQTL/mGWAS was merely linkages/associations between chromosomal locations and metabolite contents, and this was basically due to the lack of genomic information for some crops until very recently, a wheat mGWAS successfully identified 26 candidate genes with high confidence, among them two were validated to be involved in flavonoid metabolism pathway of wheat [6, 7].

This way, by combining interval mapping and genome-wide association studies (GWAS), the genetic determinants of tocochromanol accumulation in tomato (Solanum lycopersicum) fruits have been unveiled. Specifically, the content of vitamin E has been enhanced in tomato plants by identifying the genes involved in the chorismate-tyrosine pathway [8]. With this approach, Alseekh et al. reported a large-scale metabolic quantitative trait loci (mQTL) analysis on the well-characterized Solanum pennellii introgression lines to investigate the genomic regions associated with secondary metabolism in tomato fruit pericarp [9]. In total, 679 mQTLs were detected across the 76 introgression lines. Heritability analyses revealed that the environment affected mQTLs of secondary metabolism less than mQTLs of primary metabolism. Network analysis allowed to assess the interconnectivity of primary and secondary metabolism and compare their respective associations with morphological traits. Additionally, a real-time quantitative PCR platform demonstrated a transcriptional control mechanism of a subset of the mQTLs, including those for hydroxycinnamates, acyl-sugar, naringenin chalcone, and a range of glycoalkaloids. Intriguingly, many of these compounds displayed a dominant-negative mode of inheritance, contrary to the conventional wisdom that secondary metabolite contents decreased on domestication. Additionally, two candidate genes for glycoalkaloid mQTLs via virus-induced gene silencing were also validated [9].

More recently, Alseekh et al. identified several metabolite quantitative trait loci that reduce variability for both primary and secondary metabolites (phenylalanine, glucose-6-P, fructose-6-P, and maltose), which they named canalization metabolite quantitative trait loci (cmQTL); on their study nine cmQTL were validated using an independent population of backcross inbred lines, derived from the same parents, which allows increased resolution in mapping the QTL previously identified in the introgression lines. These cmQTL showed little overlap with QTL for the metabolite levels themselves. Moreover, the intervals they mapped harbor few metabolism-associated genes, suggesting that regulatory genes largely control the canalization of metabolism [10].

Maize has also been favored by the combination of GWAS and metabolomics profiling to highlight genes involved in the biosynthesis of several metabolites. Among these compounds, Liang et al. reported the identification of metabolites biomarkers for the tolerance to salt-induced osmotic stress; a citrate synthase, a glucosyltransferase, and a cytochrome P450 were found to be responsible for controlling the associations between the genotype and metabolites that induced the tolerance [11]. Owens et al. reported essential genes controlling maize grain carotenoid composition by using GWAS of quantified seed carotenoids across a panel of maize inbreeds ranging from light yellow to dark orange in grain color; significant associations at the genome-wide level were detected within the coding region of zep1 and lut1, carotenoid biosynthetic genes not previously shown to impact grain carotenoid composition in association studies, as well as within previously associated lcyE and crtRB1 genes [12]. Similarly, Li et al. examined the genetic architecture of maize oil biosynthesis in a genome-wide association study using 1.03 million SNPs characterized in 368 inbred maize lines, including “high-oil” lines. Seventy-four loci were significantly associated with kernel oil concentration and fatty acid composition [4]. Another combination of maize metabolic profiling with GWAS has been reported by Riedelsheimer et al. who outlined an association of genetic variants and concentration of 118 metabolites in leaves of 289 diverse maize inbred lines from worldwide sources; genome-wide association mapping with correction for population structure and cryptic relatedness identified for 26 distinct metabolites with strong associations with SNPs explains up to 32.0% of the observed genetic variance [13]. Similarly, Chen et al. described a comprehensive profiling of 840 metabolites and a further metabolic genome-wide association study based on ~6.4 million SNPs obtained from 529 diverse accessions of Oryza sativa. They identified hundreds of common variants influencing numerous secondary metabolites with significant effects at high resolution. They observed substantial heterogeneity in the natural variation of metabolites and their underlying genetic architectures among different rice subspecies [13, 14]. Data mining identified 36 candidate genes modulating metabolite levels that are potentially physiological and nutritionally important. As a proof of concept, they functionally identified (annotated) five candidate genes influencing metabolic traits; the study provides first-time insights into the genetic and biochemical bases of rice metabolome variation and can be used as a powerful complementary tool to classical phenotypic trait mapping for rice improvement. Besides, Dong et al. reported a comprehensive metabolic profiling and natural variation analysis of phenolamides in rice using a liquid chromatography (LC)-mass spectrometry (MS)-based targeted metabolomics method; spatiotemporal controlled accumulations were observed for most phenolamides, together with their differential accumulations between the two major subspecies of rice.

Further GWAS on rice leaves identified Os12g27220 and Os12g27254 as spermidine hydroxycinnamoyl transferases that might underlie the natural variation of spermidine conjugate levels in rice [15]. Likewise, Chen et al. identified 32 candidate genes underlying metabolic traits in rice grains; 8 candidate genes were involved in the biosynthesis and transportation of amino acids and their derivatives. Three candidates were assigned to the choline levels and its lysophosphatidyl derivatives. Precise signals for trigonelline, a bioactive compound implicated in cell cycle control, resulted in the assignment of seven candidate genes for this metabolite. Furthermore, mGWAS in rice grains revealed 40 candidates (both regulatory and structural genes) involved in the biosynthesis, modification, and transportation of phenylpropanoids, including the C-glycosyl flavones, the primary class of flavonoids in cereals [7]. More recently, Yang et al. reported the identification of a gene s07g32020 (UGT707A3) that encodes a glucosyltransferase that converts naringenin and uridine diphosphate-glucose to naringenin-7-O-β-D-glucoside; the function of Os07g32020 was verified with CRISPR/Cas9 mutant lines, which accumulated more naringenin and less naringenin-7-O-β-D-glucoside and apigenin-7-O-β-D-glucoside than wild-type Nipponbare [16].

It is evident that when working with a crop with a reference genome sequence, such as maize, rice, and wheat, to mention some of the global agricultural importance, enormous advances have been made in elucidating genes and their functional annotation for their improvements; sadly this is not the case for minor crops, medicinal plants, and other regional staple foods. Therefore, considerable effort must be made to genome sequence minor crops and develop a collection of segregating crop populations. On the other hand, plant metabolome lacks a free online accessible metabolite database for support in the annotation of unknown metabolites, and most of the studies carried out under a nontargeted metabolomic approach remain with several novel metabolites, making association studies such as transcriptomics, mQTL, and GWAS difficult when searching for gene annotation and their functional characterization. A summary of mQTL and mGWAS is presented in Table 1.

Table 1 Summary of mQTL and mGWAS studies in plants [17]

3 Plant Species Identification

The plant metabolome constitutes an enormous reservoir of bioactive compounds; many of these are products of secondary or specialized metabolism. Their taxonomic distribution is in relatively narrow phylogenetic clades within Plantae [1]. While the identification of plant species based on morphological characteristics is a well-established practice in botany, being able to identify plant species is an essential factor in understanding biodiversity, the discovery of bioactives from herbal medicines, and correlating chemical components from plants with chemical markers of patients who intake herbal medication, as well as monitoring food intake in foodomics studies, food authentication, and fraud detection, among other applications. Since the small-molecule profile of an organism ultimately reflects the genes that distinguish it, the information content of the metabolome might be just as well suited to genomic fingerprinting and assessment of genetic relatedness between species as the genomes themselves [43].

Despite its practical importance, the establishment of phylogenetic diversification and distribution patterns of secondary plant metabolites is still in its early steps, and several plant families have not been deeply explored to date in this context; Mannochio-Russo et al. described a strategy for chemotaxonomic investigations using the Malpighiaceae botanical family as a model; their workflow (Fig. 3) was based on MS/MS untargeted metabolomics, spectral searches, and recently described in silico classification tools, which were mapped into the latest molecular phylogeny accepted for this family, the workflow combines several approaches to perform a comprehensive evolutionary chemical study. It is expected to be used in further chemotaxonomic investigations [44].

Fig. 3

Experimental workflow followed for the metabolomics and chemosystematics analyses of Malpighiaceae samples. (1) The samples were initially collected. (2) The extracts were prepared with different solvents (EtOH:H,0 (4:1. v/v) or EtOAc) and then (3) subjected to LC-ESI-MS/MS analysis in positive and negative ionization modes in an untargeted method. (4) The data acquired were processed for feature finding, and the exported data were used for multivariate analysis. The clustering groups observed were merged to the phylogeny using the maximum likelihood estimation (MLE) for preliminary chemotaxonomic investigations. (5) The data were also used for feature-based molecular networking and library searches workflows to observe clade-specific molecular families. (6) A chemical hierarchy analysis and in silico classifications were obtained and finally (7) merged to the currently accepted Malpighiaceae phylogeny to determine the ubiquitous and the taxa-specific in silico classes. (Reproduced from Ref. [44])

The metabolomic analysis revealed that different ionization modes and extraction protocols significantly impacted the chemical profiles, influencing the chemotaxonomic results. In addition to the library searches for metabolite annotation, the MS/MS data generated were visualized by molecular networking analysis (Fig. 4). Molecular families constructed by such analysis represent the similarity of fragmentation patterns obtained by tandem mass spectrometry (MS/MS) analysis. These molecular families consisted of nodes (representing MS/MS spectra) and edges of connecting these nodes (representing the cosine similarity between two nodes, which measure the relatedness in MS/MS spectra). The library matches retrieved from the analysis obtained in the positive ionization mode showed the presence of a high diversity of compound classes, including C-glycosylated and O-glycosylated flavonoids, lipids, alkaloids, quinic acid derivatives, amides, triterpenes, iridoids, and lignans [44].

Fig. 4

Molecular families obtained from the feature-based molecular networking workflow and annotated based on spectral matches within the GNPS platform: (a) phenolic compounds, (b) alkaloids, and (c) lipids and terpenoids. Each node represents a tandem mass spectrometry spectra (MS/MS), while the edges that connect them represent the MS/MS fragmentation similarity (cosine >0.7). Pie charts indicate the relative abundance of ion features in each Malpighiaceae phylogenetic clade (A–J). Node sizes are relative to the summed peak areas of the precursor ion in MS1 scans. These are level 2 or 3 annotations according to the 2007 metabolomics standards initiative [45]. (Reproduced from Ref. [44])

Cinnamon is one of the oldest spices used in the world. A growing number of studies have illustrated varied phytochemical compositions among cinnamon species. Primary cinnamon metabolites, such as coumarin, cinnamaldehyde, cinnamic acid, cinnamyl alcohols, and proanthocyanidins, are shown to be differentially produced among various species; in this context, Zhang et al. developed a metabolomic ratio rule-based classification method for the automated metabolite profiling and differentiation of four cinnamon species using ultra-performance liquid chromatography-high-resolution mass spectrometry. Among the species studied were Cinnamomum cassia (Chinese cinnamon), C. loureiroi (Vietnamese cinnamon), C. verum (Ceylon cinnamon), and C. burmannii (Korintje cinnamon); proanthocyanidins, coumarin, and cinnamaldehyde were the preselected metabolites allowing the classification [46].

The genus Vanilla, a source of the most appreciated flavor worldwide, comprise over 110 species. Currently, only three species have commercial relevance, Vanilla planifolia Andrews, V. tahitensis J. W. Moore, and V. pompona Schiede. V. planifolia are preferred by industry due to its higher content of vanillin, the main flavor component; more attention needs to be made to other Vanilla species. Leyva et al. developed a nuclear magnetic resonance (NMR) metabolomic platform to profile for the first-time leaves that are known to accumulate putative vanillin precursors of V. planifolia and those of Peruvian V. pompona, V. palmarum, and V. ribeiro to determine metabolite difference among them. Their NMR analysis identified 36 metabolites, and multivariate analysis identified malic and homocitric acids, together with 2 vanillin precursors (glucoside A and B), as relevant markers for species identification [47].

Coffee is appreciated worldwide for its aroma, flavor, and stimulant properties. Souard et al. examined leaves of nine Coffea species grown in the same environmental conditions by an untargeted liquid chromatography high-resolution mass spectrometry (LC-HRMS) approach, with the primary objective of identifying metabolites that significantly contribute to the classification between Coffea species. Based on their results of multivariate analyses, 1637 variables (metabolites) were analyzed, from which 92% (1505 metabolites) were significantly different overall taxa. Among the species studied, when two well-known C. arabica and C. canephora were compared, a feature with an m/z = 195.0870 corresponding to [M + H]⁺ of caffeine came out as the main discriminant compound. Caffeine concentration was approximately 800 times higher in C. arabica leaves than in C. canephora. Another feature observed at an m/z value of 247.0598 had much higher intensities in C. arabica than in C. canephora, but unfortunately, this feature was not identified. This metabolic fingerprinting study aimed to determine the specific differences between the metabolomes. All nine clusters of each species studied were observed on both PCA and PLS-DA score plots, with good discrimination between the eight Coffea species [48].

Several studies have described the use of metabolomics to distinguish herbal medicinal plants. For example, Lesiak et al., using seeds of the genus Datura plants, outlined direct analysis real-time mass spectrometry (DART-MS), which could provide diagnostic fingerprint profiles of nine Datura species seeds and whether chemometric processing of the observed profiles could enable species-level identification and differentiation. They confirmed that the seeds could be analyzed by DART-MS directly in a high-throughput manner without using a solvent extract. Each species exhibits a distinct chemical signature, and the processing of this data by multivariate statistical methods enables species-level differentiation. In addition, they observed that while intraspecies chemical signatures are similar, interspecies fingerprints are distinct enough to be discriminated against using multivariate statistical analysis tools [49]. Another example using seed samples of Polygonatum species was reported by Qi et al.; Polygonatum plant species have properties that make them sound like medicine and food in China. There were almost no differences in the contents of the metabolites in the amino acids and derivatives, nucleotides and products, and others (e.g., saccharides, alcohols, and vitamins) classes among the seed samples. In contrast, the seed samples had some diversity in the contents of lipids, phenolic acids, lignans and coumarins, tannins, and quinones. The flavonoid, steroid, and terpenoid classes and contents varied among the Polygonatum seed samples; these compounds have relatively strong pharmacological effects; their findings indicate that different Polygonatum seeds differ in terms of their medicinal and nutritional value [50].

Other plant species that have been classified and identified using metabolomic approaches are Mentha species [51]; Acorus [52] – plants in Acorus have been used as herbal medicine by various linguistic groups for thousands of years; and Phyllanthus species [53]; among other medicinal plant species.

4 Plant Root Metabolome and Climate Change

Climate change is a relevant issue due to its adverse and high-impact consequences it can cause directly and indirectly at social, ecological, biological, and health levels [54, 55]. The main factors promoting climate change comprise natural and anthropogenic activity; agriculture can produce 30–40% of the total greenhouse gas emissions [56]. In addition, due to these issues, an increase in the use of pesticides is expected, which will significantly affect global crop production, as well as pathogen diseases, abiotic stress, and the decrease in the production of the major crops worldwide [57]. Quality and crop productivity are negatively affected by global warming. It is expected that dramatic increase in the following years due to the increasing annual temperature, solar radiation, changes in precipitation, and high CO₂ levels [54, 56, 57]. Some other factors which reduce crop production quality are floods and droughts [58, 59] and are affected by alterations in rainy seasons, pest invasions, crop disease, water supplies, price of products for agricultural processes, and premature consumption of fertilizers [54].

Humanity’s well-being and economics depend strongly on the agricultural sector, which simultaneously depends on the ability to adapt crops to environmental conditions and is therefore considered a climate-dependent industry [58]. To improve productivity, nutrient quality, and crop resilience, it is necessary to adapt strategies and design technologies to help mitigate climate change’s effects [55, 60]. The leading crop breeding technologies for adaptation comprise biotechnology techniques, such as next-generation sequencing and RNA-mediated silencing [58, 61, 62]. Another strategy to improve crops and their resistance to biotic and abiotic stress is the engineering of the root microbiota, which represents a promising technology in the future to face climate change [63]. This strategy arises from analyzing the plant microbiome in the rhizosphere and endosphere, in which several interactions occur between the plant, microorganisms, their metabolites, and the metabolites of the surrounding plants. Once the relevance of these interactions on plant development is understood, the rhizosphere microbiome should be manipulated and thus increase production and their well-being; reduce the need for farmland, pesticides, and fertilizers; and thus reduce the intrinsic carbon footprint associated with these activities [64,65,66].

Plant evolution gave way to the adaptation through metabolite excretion to the soil (exudates) to interact with rhizosphere composition. The exudates may alter the composition and activity of the microbiome around, changing the pH, soil structure, and availability of oxygen or supplying organic compounds as a source of energy [67]. Some of the compounds exudated can also act as chemotrophic signals, which can attract pathogenic microorganisms to the plant, nematodes, or herbivorous arthropods [68]; however, these compounds can also work in the recruitment of beneficial microorganisms which can aid the plant in defense against pathogens, diseases, biotic stress conditions or enhance nutrient absorption [63, 64]. Under stress conditions, plants trigger the production of many secondary metabolites, with defense signals that promote the cope against pathogen organisms. Synthesis of secondary metabolites can relieve stress by modifying root microbiota to further degradation of different types of pollutants to carry out bioremediation [63, 69].

Among the strategies proposed for manipulating the rhizosphere microbiome is the direct inoculation of microorganisms in the soil. However, one of the biggest challenges to achieve is the determination of the species that act on the mechanisms involved, the competitive behavior with the native microbiota, and the effects it has on agricultural conditions. Alternatively, the metabolites observed in rhizosphere exudates should be used and applied in specific areas to stimulate native microorganisms [67]. Therefore, it is necessary to determine qualitatively and quantitatively the composition of the exudates under particular conditions, as well as the metabolomic analysis of the prominent participants in the interactions described and reasonably link the production of metabolites and their primary function in the face of possible types of stress [65, 70].

A metabolomic analysis is an excellent choice to study plant-rhizosphere interactions due to the complexity and quantity of compounds involved in the metabolic relation between roots and microorganisms. Instrumentation such as LC-MS, GC-MS, and NMR are the most widespread techniques and powerful tools used for identifying the compounds present in the rhizosphere [68, 71]. On the other hand, the exometabolomics concept consists of analyzing the metabolic traces of microorganisms present in certain conditions to understand the underlayer mechanisms that exist in the rhizosphere and the determination of substrate consumption by microorganisms [64]. In this context, we reviewed recent literature in metabolomics and exometabolomics, the findings in primary metabolites, and their function on specific rhizosphere interactions.

4.1 Metabolome Changes in Roots by Abiotic Stress

Metabolome alterations in the rhizosphere due to interactions result in significant interest in setting the defense mechanisms that play plants against external agents and expanding the landscape about key metabolites to degrade specific contaminants. Therefore, Wang et al. explored changes in chemical composition in root exudates of the urban greening trees exposed to phenanthrene. In Loropetalum chinense, Gardenia ellis, and Rhododendron simsii, carbohydrate levels increased in the presence of phenanthrene, indicating a regulatory function mediated by the biopolymer’s degradation, whereas in Osmanthus fragrans levels decreased, suggesting different behavior between species. Phenolic compounds increased in the presence of phenanthrene in Ligustrum japonicum, R. simsii, O. fragrans, Gardenia jasminoides, and Camellia sasanqua, implying an adaptation to attract rhizobia bacteria, with the aim of cope with exposure to phenanthrene [72]. Regarding heavy metals stress, Lu et al. analyzed two wheat genotypes with different tolerance to Cd: Aikang 58 with low accumulation and Zhenmai 10 with high accumulation. Both phenotypes showed an increase in phenylalanine and tyrosine in the presence of Cd, relating these changes in the shikimate-phenylpropanoid pathway. The rise in acetylglycine and histidine indicates a chelating activity to chelate Cd in vacuoles. At the same time, glutamate, glutamine, aspartic acid, asparagine, and lysine perform an osmotic balance to detoxify heavy metals. In the presence of Cd, an increase in maltose, isomaltose, sorbose, tagatose, and polyols assists the cell wall’s structure. In contrast, the addition of glyceric acid, cis-aconitate, malic acid, salicylic acid, and citrate indicates a deterioration in the activity of the tricarboxylic acid cycle activity to assimilate carbon under stress conditions. These alterations in the metabolism promote a high ability to take out and defend the plant against reactive oxygen species, inducing molecular signaling and antioxidant enzymes [73].

A report studying the effect of acid drainage contamination on Phragmites australis by Kalu et al. analyzed root and rhizosphere metabolome. The main compounds found in roots at contaminated sites were adenosine, inosine, methionine, carnitine, and dimethylglycine. On the other hand, uridine, dopa, asymmetric dimethylglycine, adenosine, and phenylalanine had a lower abundance in contaminated sites. This alteration has the purpose of recruiting microorganisms that promote the growth of the plant while at the same time attracting microorganisms specialized in heavy metal detoxification. As for bacterial communities, the main phylum in samples grown at contaminated sites was proteobacteria, β-proteobacteria, and the Methylocystis, Rhizobium, and Delftias genera [74]. In salinity stress, Wang et al. analyzed the canola roots metabolome in the presence of NaCl. The abundance of proline and soluble sugars increased in canola roots under saline stress. However, the metabolites with the most significant difference between groups were lipids, primarily fatty acids, which increased compared to controls. In saline stress, lipids affect membrane permeability, fluidity, integrity, and protein transport activity; therefore, the reconstitution of lipids in cells becomes transcendental. This stress triggers the production of polyunsaturated fatty acids, which can help the activation of membrane ATP-loop activity, which is responsible for maintaining homeostasis, facilitating the pumping of Na⁺ from the cytosol to the external medium, and blocking K⁺ channels [70].

Finally, it is essential also to consider UV radiation stress. Mannucci et al. analyzed the effect of UV exposure on tomato plants and their metabolomic changes in roots and leaves. In the roots of plants exposed to UV radiation, terpenoids and phenylpropanoid derivatives pathway synthesis increased compared to controls. For carbohydrates, degradation processes increased in radiation treatments, suggesting a wide degree of mobilization of reserve compounds to produce necessary precursors for secondary metabolites synthesis, such as flavonoids. As an effect of a rearrangement of lipid membrane composition, monogalactosylglycerol was found in the plants exposed, while 4α-carboxy-5α-cholesta-7,24-dien-3β-ol levels decreased. Finally, phenolic compounds and p-coumaroyl glycolic acid decreased in the UV treatments, a compound with anti-inflammatory properties [75].

4.2 Metabolic Changes Associated with Root and Other Biotic Factor Interactions

The most relevant biotic factors that transcend the metabolome analysis include plant pathogenic microorganisms; it is essential to decipher the compounds produced by plants to counteract the conditions and even the mutual organisms that help defend against these agents. Another biotic stress to consider is the neighborhood of other plant species competing for nutrients. In addition, the importance of the interaction of mycorrhizal fungi and the benefits they have in association with root plants is known. Both metabolisms are relevant to develop strategies that can help to improve crops.

To find out how rye competition affects Vicia villosa Roth, Hazrati et al. analyzed the metabolic effect on the roots of these plants. Kaempferol-Rha-Xyl-Gal was the main compound found. This compound decreased in V. villosa when it was grown in the presence of rye. Thus, it is estimated that competition produces a deficit of several nutrients in hairy vetch, decreasing the production of flavonoids [76]. Phytophthora sojae is known to cause phytophthora root rot disease in soybean; Zhang et al. analyzed the rhizosphere of Glycine max inoculated with P. sojae. The post-inoculated rhizosphere of a resistant species had a greater abundance of metabolites related to cutin biosynthesis, suberin, wax, arginine, ansamycins, pyrimidine, galactose, linoleic acid, ABC transporters metabolism, and lysine degradation. Most of the metabolites in the post-inoculated rhizosphere include antibiotics, which are responsible for conferring plant resistance to pathogens. On the other hand, some compounds in the control rhizosphere contained compounds to attract possible pathogenic microorganisms to the plant, such as daidzein and genistein. Although some flavones and isoflavones repel zoospores, others have the opposite effect, each specific to the conditions the plants were exposed to. Besides, cutin, suberin, and wax biosynthesis are inferred to provide drought tolerance by preventing water loss and insect tolerance [77].

Finally, interaction with symbiotic microorganisms focuses on knowing the benefits they bring and how they achieve them. Therefore, a view with a metabolomic approach is interesting to know the main compounds in the rhizosphere used to improve crop capacity. In this context, Zhang et al. analyzed the interaction between Medicago truncatula and rhizobium bacteria to get the metabolites in the formed nodules. Oxylipin-9-HODE decreased during the application of rhizobia bacteria, indicating decreased jasmonic acid precursors and inhibition due to the interaction [78]. Oxylipins are critical signaling molecules in the defensive response of plants to protect their tissues from attacks by herbivores or pathogens, and some contain antimicrobial properties.

Similarly, Sebastiana et al. analyzed cork oak roots colonized with Pisolithus tinctorius, an ectomycorrhizal fungus. In the study, the inoculated roots showed higher levels of γ-aminobutyric acid, alanine, β-glucose, and citrate. It was also demonstrated that the inoculated samples decreased quercitol, glycine-betaine, α-glucose, fructose, malate, and lactate levels. Inoculated samples influenced alkaloids, terpenoids, oxylipins, lipids, carbohydrates, amino acids, nucleic acids, and vitamins. In addition, a decrease in isomers of glucose, sucrose, sorbitol, and mannosyl glycerate was observed, as well as a reduction of isomers of fatty acids and compounds involved in the metabolism of tyrosine and histidine. The decrease of organic acids and glycine-betaine is related to apoplastic protective barriers, indicating the transfer of these metabolites to fungi, in the case of lipids, including monoacylglycerols, which are the main components of suberin and bark. This layer accumulates on the most exposed face of tree stems and roots and protects against drying and pathogen attack [79].

Another example was shown by Hernández et al. who analyzed the beneficial activity of Rhizobium tropici in Phaseolus vulgaris growth under phosphorous deficiency. It was observed that some organic acids, polyhydroxy acids, sugars, and polyols increased significantly in nodules with phosphorus (P) deficiency. In contrast, some amino acids and nitrogenous compounds decreased, reducing N fixation in P-deficient plants. In addition, they presented sugar accumulation, indicating demand for root photosynthate due to the decrease in the photosynthesis network. On the other hand, changes in carbohydrate content mean glycolysis/C binding pathways are induced in nodules under P deficiency stress [80]. P-deficient roots showed a decrease in the organic acid concentration, suggesting their exudation toward the rhizosphere; this has also been recently demonstrated in a study by Gomez-Zepeda et al. when using mass spectrometry imaging to locate organic acid exudates in P-deficient Arabidopsis plants. Organic acid exudation by roots is considered a core response to different types of abiotic stress and the interaction of roots with soil microbes. For decades it has been a target trait to produce plant varieties with increased capacities of inorganic orthophosphate uptake and aluminum tolerance [81].

Microbiota may vary according to the plant growing zone. In this way, Li et al. analyzed metabolome and microbiota in the roots of Aconitum vilmorinianum grown in two different sites in China (Luquan and Weixi). The difference observed in the metabolites was an increase in yunnaconitin and vilmorranine A in Weixi and a decrease in amino acids and some derivatives in Luquan. A correlation was found between 137 bacteria and 17 fungi with 75 differential metabolites in the 2 regions, among which the fungus on Cladosporium stands out, with a high probability conditional on aconitine, demonstrating the appearance of this metabolite in Weixi samples. Regarding Luquan, three bacterial and six fungal biomarkers were found, while Weixi showed the presence of five bacterial and five fungal biomarkers. This finding in the microbiome may be due to the environmental temperature, while in Weixi usually snows, and Luquan rarely occurs [82].

Knowledge about metabolome interaction between roots and their environment is crucial to identify the relevant metabolites produced in this medium. With this, it is intended to know which metabolic pathways are altered in plants to produce detoxifying compounds, antibiotics, or those that recruit beneficial microorganisms for plants. On the other hand, the compounds produced by microorganisms and their identification are also convincing to analyze the possibilities of the inoculation of different bacterial species in scenarios of various types of stress. However, it is important to specifically identify these conditions since not all species will behave similarly under the same state or stress. Herewith, it is essential to trace how crops can be treated to face biotic and abiotic stress caused mainly by climate change and improve their production and development.

5 Plant Biomarker Identification

Metabolomics is one of the most recent powerful tools for studying plants and other organisms and is becoming a complementary technique to genomics, transcriptomics, and proteomics [83]. Metabolomics addresses the activity of small molecules (<1500 Da) produced by cells during their life cycles, that is, products of primary or secondary – specialized – metabolism, found in various biological systems, studying how metabolic profiles change within an organism in response to some situations, such as disease or stress [84].

Therefore, unlike other “omics,” metabolomics best describes phenotypes, can give instantaneous information on the physiological state of cells and thus provide a broader view of the biochemical state of an organism, and can track the metabolic network of a biological system and its perturbation in response to stimuli. Metabolomics aims not to identify every metabolite observed but to compare patterns of metabolites that change in each biological system. When these analyses are performed on enough biological replicate samples, it allows researchers to discriminate and classify samples and gain insight into changes in metabolome composition related to a particular physiological state, influence of stress or stimulus, genetic modification, or interaction with other organisms.

Metabolites provide a “fingerprint” of the complex interaction between the genome and the environment. They can generally be divided into two groups: primary metabolites, essential for maintaining processes directly involved in plant survival, growth, and reproduction, and secondary metabolites, which contribute to specialized processes in each organism synthesized to fulfill a nonessential function in the plant.

Due to the structural heterogeneity of metabolites and their different ranges of magnitude and concentration, their identification and measurement present a considerable challenge. For the plant kingdom alone, researchers have reported more than 400,000 plant species worldwide [85]. As for structurally distinct secondary metabolites, there are approximately 200,000 to 1,000,000 [86], which is why the field of plant metabolomics is the most advanced with a wide range of applications [1, 85].

All this information and understanding of the metabolome as it is affected by factors including environmental changes, physical changes, biotic stresses, abiotic stresses, and even internal changes in the plant as a function of its developmental stage can be used to monitor significant variations in metabolites and in the search for metabolites that can act as biomarkers.

The study of metabolomic biomarkers is one of the least explored areas in metabolomics. By 2022 only 16% of the publications examined the discovery or discrimination of biomarkers, while 46% of the publications refer to metabolic mechanisms and 33% examined metabolic profiles. However, many of these publications on metabolic profiles include preliminary and descriptive findings for more detailed analysis of the machine tool and discovery of biomarkers [87].

A metabolomic biomarker differs greatly from a protein biomarker and transcriptomic biomarkers because of the close relationship between individual metabolites. Factors measured in other “omics” technologies are independent, although there may be patterns of abundance that reflect a disease state. A metabolomic biomarker is not just a chain of changes in individual metabolites. Instead, it is composed of co-related metabolites that change together [88].

For the discovery and characterization of a metabolomic biomarker, validation based on the environment and research design is necessary to determine whether the proposed biomarker can distinguish between the changes to which plants are subjected [88], that is, for a biomarker to be classified as such, it must meet specific characteristics: be measurable, reflect the qualitative or quantitative interaction of the plant with the chemical of interest, be precise and sensitive, and be commonly shared among individuals in a population and plant species.

In this way, and through preliminary findings of metabolic profiling, some biomarkers have been identified in plants in response to exposure to stress factors: For example, a study on biomarker discovery [89] demonstrated that volatile organic compound (VOC) profiles could be used as diagnostic markers of stress in grapevine; this study shows that VOC emission can be considered as a universal response of grapevine to plant defense elicitors, given that the elicitors evaluated induced the emission of a standard set of VOCs encompassing chemically different compounds, including the sesquiterpenes α-farnesene and β-caryophyllene and that such a response is analogous with the induction of stilbene phytoalexins.

Similarly, plant metabolomics can help to identify resistant metabolites in plants that are subjected to stress conditions [17]; the selected biomarker can be used as a diagnostic metabolite for plant stress, as in the case of the study of wheat and barley resistance against F. graminearum infection where they point to various plant hormones that respond to this infection [90]. Such is also the case of phenylpropanoids and organic acids, metabolites identified as biomarkers of nitrogen deficiency in leaves and roots of tea plants (Camellia sinensis) that are elevated when there is a nitrogen deficit [91]. For example, hexadecanoic acid and dotriacontane, highly expressed metabolites, were identified as potential biomarkers in rice seeds infected with Rhizoctonia solani toxin, metabolites involved in several important rice biosynthetic pathways, such as the biosynthesis of saturated fatty acids and the unsaturated fatty acids cutin, suberin, and wax [92].

However, the identification of metabolites not only corresponds to stress responses, but the detection of metabolic changes at different developmental stages also contributes to finding metabolites characteristic of each stage (biomarkers), as in the case of metabolomic analysis of rice where developmentally controlled phenolamide accumulation patterns are observed [15] or in Arabidopsis where patterns of glucosinolate, raffinose, and galactinol accumulation are present at the base of leaves during the senescence stage [90]. Analysis of the spatiotemporal metabolic profile of plant development also allows the identification of potential biomarkers to capture intrinsic genetic features of plant development, as in the study of rice tillering (branching), in which 21 metabolites captured nearly 83% of the metabolic variation [93], and the developmental phase of soybean during the transition from vegetative to reproductive stage, in which eight flavonoid kaempferol glycosides were identified as potential growth markers [94].

In the food industry, the identification of biomarkers has also worked in quality processes for food authenticity and food traceability matrices of plant origin, especially in the field of aromatic herbs and spices, which are very susceptible to food fraud, as in the case of thyme, an aromatic herb traditionally used for food purposes due to its organoleptic characteristics and medicinal properties. In this particular case, it was possible to determine the geographical traceability of thyme based on different origins (Spain, Poland, and Morocco), as well as to evaluate its processing by comparing sterilized thyme with non-sterilized thyme, where 24 differential markers belonging to different classes were identified: among monoterpenoids, diterpenoids, sesquiterpenoids, alkylbenzenes, and other diverse compounds for its authenticity [95].

Another example of this application that helps to detect adulterants in plants that are used commercially is observed in the study by Ivanovic et al. using wild garlic (Allium ursinum) and poisonous adulterants Convallaria majalis and Arum maculatum as a model for the detection of adulterants in edible plants; the metabolites isovitexin, vicenin II, azetidine-2-carboxylic acid, and trigonelline were elucidated as biomarkers of adulteration [96].

On the other hand, metabolomics approaches have also been used to characterize and diagnose plant diseases and thus crop improvement, for example, during the study of the interaction maize-Fusarium graminearum-Bacillus amyloliquefaciens or soybean-Rhizoctonia-B. amyloliquefaciens, a better understanding of the metabolic regulation of all interacting systems has been achieved, providing valuable insights potentially useful in plant breeding and metabolic bioengineering [97].

Metabolite markers against drought stress (malonate, leucine, 5-oxo-L-proline, saccharic acid, trans-cinnamate, succinate, and glyceric acid) have been reported by Khan et al. who identified biomarkers in the metabolome of chickpea (Cicer arietinum L.) when treated with plant growth regulators (salicylic acid and putrescine) and the PGPR growth consortium (B. thuringiensis, Bacillus subtilis, and Bacillus megaterium). Deliberative metabolic reprogramming of chickpea targeting biomarker synthesis pathways resulted in drought-tolerant chickpea varieties [98].

Biomarker identification can also be applied to predict phenotypic traits and provide early detection tools to identify and use them in plant breeding development [99]. In China, for example, hybrid rice combinations have been created using sterile lines and restorer lines to reduce seed deterioration during storage and establish galactose and gluconic acid as metabolic biomarkers that reflect the degree of seed aging [100]. Also, in understanding the functioning of plants growing under extreme conditions, the identification of biomarkers in these plants could provide information that would benefit crop improvement; for example, it was possible to identify associated metabolic biomarkers in an alpine medicinal plant (Rhodiola crenulata) that can survive in extreme altitude conditions where the shikimic acid-phenylalanine-phenylpropanoid flavonoid pathway was enhanced with phenylpropanoids upregulating much more than flavonoids [101].

Surveillance for potential pathogens is critical to plant innate immunity, so plants depend on the perception of pathogen-derived molecules to activate defense-related signaling cascades and specialized metabolites in response; in studies of the tomato plant (Solanum lycopersicum), by monitoring metabolic profiles of signaling cascades in response to pathogens, significant biomarkers were noted for several classes of metabolites including amino acid derivatives, lipid species, steroidal glycoalkaloids, hydroxybenzoic acids, hydroxycinnamic acids, and products, as well as flavonoids [102].

Other metabolites identified as biomarkers in the plant defense response to pathogens are hydroxycinnamic acids; the conjugation of these acids with amide groups contributes to the regulation of the dynamic metabolic pool of hydroxycinnamates; a wide range of biogenic amine compounds found in most plant cells and these conjugates can scavenge radicals, confer antimicrobial activity, and can be deposited in the cell wall; so finding the activity of these metabolites is indicative of the plant-pathogen response [103]. A summary of primary and secondary or specialized metabolites identified in various plant species is presented in Table 2.

Table 2 Identifying key metabolites in various plant species using different analytical methods [104]

The identification of biomarkers in plants can have diverse applications, as described above; however, to reach the validation of these metabolites, metabolic profiling studies are necessary; metabolomics has been widely applied in the study of plants showing a breakthrough in understanding how the phenotype is related to the metabolome and therefore the function of metabolites under normal conditions, stress, and during their development.

Understanding the adaptative physiology and biochemistry of plants, as well as the underlying metabolic events, is relevant to have a global perception of the metabolomic status of plants, with the identification of biomarkers providing helpful information on metabolites involved in resistance responses, stress, growth, a better understanding of intra- and interspecific microbial interactions occurring at different heterogeneous levels within the plant habitat, identification of systemic responses of various crops to pathogen stress, and pathogens and their biological control, would allow crop scientists to identify unique metabolic markers that can be applied to early detection of a plant pathogen as well as to the development of bio fungicides, for example, for use during pre-harvest, post-harvest, and harvest storage and large-scale storage of crops. Identifying and applying metabolic biomarkers could favor controlled and semi-controlled planting systems shortly, and if properly integrated into crop protection strategies, food security could be mitigated. However, the applications of these biomarkers could be helpful in various areas, such as the food and pharmaceutical industry in food quality and safety processes, diagnosis and treatment of plant diseases, crop improvement, and analyzing genetic-modified crops. Still, the work done so far is relatively new, and efforts should continue to cover the tremendous potential presented by identifying metabolic biomarkers. Table 3 summarizes the identification of some metabolites in transgenic plants using different analytical techniques.

Table 3 Identification of important metabolites in transgenic plants using different analytical tools [104]

6 Plant Single-Cell Metabolomics

In the past, genomics, transcriptomics, and metabolomics techniques have been applied in bulk plant samples consisting of many cells; in such experiments, the biological process leading to cell heterogeneity is often considered not to be biologically relevant. However, cell heterogeneity has been shown to play important biological roles in many situations for which averaging would mask relevant metabolic processes [130]. Plants contain several cell types and exhibit complex regulatory mechanisms. Studies at the single-cell level have gradually become more common in plant science. Single-cell transcriptomics, spatial transcriptomics, and spatial metabolomics techniques have been combined to analyze plant development. These techniques have been used to study the transcriptomes and metabolomes of plant tissues at the single-cell level, enabling the systematic investigation of gene expression and metabolism in specific tissues and cell types during defined developmental stages [131]. However, single-cell technologies require labor-intensive protocols for plant cell isolation. On that respect several attempts have been developed; these strategies can be classified into three main groups: those that attempt to isolate material of specific cell type to perform the analysis on platforms used for regular metabolomics, which we will refer to as single-cell type metabolomics [132]; those based on micromanipulation of single cells; and those based on mass spectrometry imaging [130].

Several methods for harvesting cells have been developed for single-cell and single-cell type metabolomics, whereby cells can be obtained or extracted in situ. The in situ techniques include micropipetting for isolating the contents of specific cells, laser microdissection (LMD), laser microdissection and pressure catapulting (LMPC), laser capture microdissection (LCM), and fluorescence-activated cell sorting (FACS). Laser microdissection and pressure catapulting and laser capture microdissection use laser to excise single cells or microareas from fixed or frozen intact tissues and are becoming very popular for plant cell and tissue sampling. FACS, on the other hand, is used to obtain specific cell types; for example, those identified from root developmental zones by transgene-labelled nuclei or by immunolabelled-based collection and microfluidic sorting-based methods that exploit intrinsic cell properties [129]. Figure 5 summarizes different approaches for cell-specific metabolomics.

Fig. 5

Overview of experimental steps and data structure from the different approaches for cell-specific metabolomics. (Reproduced from Ref. [130])

However, to obtain single-cell suspension, it is a very challenging activity and deeply laborious. In addition, plant cells are rigid cells when compared to animal cells; rigid cell walls remain the main obstacle for single-cell technologies in plants. Since protoplasts must remain alive and be subjected to a minimal level of disturbance during isolation, for example, for protoplast isolation, the cell wall digestion procedure requires optimization for suitability for the tissue under study [133].

Mass spectrometry imaging (MSI) technique can provide spatially resolved information on the structure and content of metabolites including know and unknow endogenous metabolites, and it thus produces tissue molecular imaging maps. Three MSI techniques have been developed based on different ions sources: secondary ions MS (SIMS), desorption electrospray ionization (DESI), and matrix-assisted laser desorption/ionization (MALDI) [131]. Figure 6 summarizes different ionization techniques used for MSI. Among them, MALDI is the most popular ionization technique for MSI experiments. In MALDI mode a matrix applied to the sample is excited by a laser; this energy is further transferred to the sample resulting in the desorption/ionization event. Preparation for MALDI usually comprises cryo-sectioning and lyophilizing a frozen sample before applying the matrix by either a sprayer or a special device, as well solvent free sublimation [130]. However, MALDI remains a technique that still lacks significant improvement, for example, matrix selection and the choice of matrix method application, tissue sectioning technique, embedding protocols, sample preparation, and mounting. In other words, MALDI imaging technique requires optimization for every tissue and metabolites of interest; for example, Pérez-López et al. developed a protocol of MALDI imaging by sample imprinting in nylon membranes to locate fructans in stem and rhizome tissues of Agave tequilana plants [134]; in addition, the combination of ion mobility spectrometry allowed the detection of fructan isomers even if these have not been mapped on their images obtained; Fig. 7 outlines the protocol developed.

Fig. 6

Schematic representation of the different ionization strategies used for mass spectrometry imaging (MSI). (a) MALDI, (b) secondary ion mass spectrometry (SIMS), (c) desorption electrospray ionization (DESI), (d) laser-ablation electrospray ionization (LAESI). (Reproduced from Ref. [130])

Fig. 7

Outline of tissue printing technique. (a) Agave tequilana plant showing the crown region (leaf/stem to root transition). Dotted red line indicates the longitudinal axis; dotted blue line indicates the transversal axis. (b) A. tequilana plant as in (a) with leaves and most roots removed and dissected longitudinally. (c) Longitudinal stem section. (d) Transverse stem section. (e) Representation of tissue printing process. (f) Tissue-printed transverse section mounted on MALDI Imaging plate. (g) PAS staining of a tissue-printed longitudinal section. (h) MALDI-ToF-MSl of a tissue-printed longitudinal section obtained, using a sprayer for matrix application and a QTOF SYNAPT G1 spectrometer with a spatial resolution of 100 pm per pixel. (Reproduced from Ref. [134])

More recently, DESI have become the newest development for mass spectrometry imaging to visualize plant metabolites; DESI offers a great advantage being matrix-free ionization alternative to MALDI. In DESI, a solvent stream originating from an electrospray probe is directed at an angle (most important parameter) toward the sample at ambient pressure, propelling secondary ions to the mass analyzer, enable direct analysis of unprocessed frozen samples sections which simplify samples preparation [135, 136]. Very recently, some metabolites detected with DESI source ranged from monoterpenoid alkaloids, which were localized in several of the major parts of the Rauvolfia tetraphylla plant when analyzed by MALDI and DESI-MSI [137], alkaloids in the leaves of Gelsemium elegans [138] through cannabinoids and flavonoids in the leaves of Cannabis sativa [139], among other recent applications of DESI.

7 Concluding Remarks

The utilization of a model plant like Arabidopsis or a crop with a reference genome sequence, such as - maize, rice, tomato, and wheat, to mention some of the global agricultural importance - offers a unique opportunity, where approaches such as mQTL, GWAS, mGWAS, and transcriptomics can effectively provide a vast potential to reveal gene annotation and their functional characterization. On the contrary, enormous efforts must be made for other minor crops, medicinal plants, and regional staple foods. However, plant metabolomics research needs to be focused on developing strategies to develop confident metabolite annotation through implementing a free online accessible database for metabolite identification. Plant metabolomics showed, as reviewed here, great potential to assist crop improvements, supporting exploring species identification for diversity and botanical purposes, food authentication, fraud, and traceability.

Without forgetting the knowledge of metabolome interaction between roots and their environment to identify the relevant metabolites produced, on the other hand, metabolomics approaches can play a crucial role in studying the interaction between plants with biotic and abiotic stresses. Knowing the compounds produced by microorganisms and their identification in climate change conditions that enhance crop production and development is crucial.

Identifying and applying metabolic biomarkers could favor controlled and semi-controlled planting systems shortly, and if properly integrated into crop protection strategies, food security could be mitigated. However, the applications of these biomarkers could be helpful in various areas, such as the food and pharmaceutical industry in food quality and safety processes, diagnosis and treatment of plant diseases, and crop improvement. Still, the work done so far is relatively new, and efforts should continue to cover the tremendous potential presented by identifying metabolic biomarkers.

Plant metabolomics could benefit in developing new strategies to face challenges and demands in crop improvement. Spatiotemporal metabolomics can effectively support plant-soil interactions studies; the implementation of mass spectrometry imaging in combination with ion mobility spectroscopy could potentially reveal metabolites’ location in plant tissue without the need for extraction, in addition, to providing isomer identification without forgetting the need to develop metabolite databases that can support full plant metabolome coverage.