Genome Engineering for Crop Improvement. Группа авторов

but often at the expense of energy‐rich and health‐promoting compounds such as anthocyanins.

      This chapter will focus on cereals and pseudo‐cereal plants that produce starchy grains since they are part of almost every meal we eat. Grains consists of the plant embryo – a miniature plant – and the seed coat that protects the embryo from abiotic and biotic stress. Storage compounds (starch, proteins, and lipids) that provide energy and building blocks for the young, developing plant after germination, are stored in the endosperm or in the cotyledons, while minerals are mainly stored in the aleurone layer or in the embryo. Nutrients and minerals are not evenly distributed within the grain (Pongrac et al. 2011; Singh et al. 2013, 2014; Vogel‐Mikuš et al. 2014; Mantouvalou et al. 2017). This is of high relevance for human nutrition, as cereals are usually processed before consumption, and certain nutritious compounds stored in the seed coat or embryonic tissue may be lost. There is, therefore, an urgent need to develop research tools that will provide better insights into the 2D and 3D spatial distribution of nutrients and minerals in cereals and the mechanisms behind these distribution patterns.

2.2 Exploring Nutrient Distribution in Grain

      The revolution in omics technology, particularly metabolomics, has been closely linked to technological developments in mass spectrometry (MS) (Boughton et al. 2016). MS is currently the most efficient technology for the characterization of biomolecular composition and has made significant progress in providing a comprehensive understanding of biological functions. Unfortunately, in MS‐based studies, where the analysis is performed on a tissue homogenate (e.g. liquid or gas chromatography), spatial information is lost (Dong et al. 2016a). Spatial information in mass spectrometric analysis can be preserved if the extraction process is limited to very small areas of the sample, which is then analyzed using new ionization techniques, such as Matrix Assisted Laser/Desorption/Ionization (MALDI) (Dong et al. 2016a). These techniques account for the majority of applications in biomolecular imaging. Techniques that can also be used for spatially resolved studies of molecular profiles are also Laser Desorption Ionization (LDI) (Lee et al. 2012), Secondary Ion Mass Spectroscopy (SIMS) or Desorption Electrospray Ionization (DESI) (Lee et al. 2012). In the aforementioned techniques, the analytes are raster‐scanned and ionized and desorbed directly from the surface, thus avoiding traditional liquid extraction. By automating the process, images can be generated that show the spatial distributions of all detected compounds (Dong et al. 2016a; Jenčič et al. 2016).

2.2.1 Matrix Assisted Laser/Desorption/Ionization

MALDI is mainly used in biomedicine to identify biomarkers associated with various diseases (Grassl et al. 2011), while applications in plant sciences remain limited (Boughton et al. 2016; Dong et al. 2016a). With this technique, large molecules such as proteins can be detected with a lateral resolution of 0 micrometers in novel, state‐of‐the‐art instruments (Römpp and Spengler 2013). The working principle of MALDI is based on the co‐crystallization of the analyte with a chemical matrix which absorbs the laser energy and releases the analyte into the gas phase in a process that leads to ionization (Bjarnholt et al. 2014). The addition of the matrix has a number of advantages: in particular, it enables the ionization of compounds that do not themselves absorb light. Moreover, since the energy of the laser light is absorbed by the matrix and not directly by the analyte, MALDI is a soft ionization technique that exhibits little fragmentation in the mass spectra (Bjarnholt et al. 2014). Unfortunately, when performing MALDI imaging with higher resolution (laser spot sizes smaller than 1 μm), the ion yield (and thus the sensitivity) decreases significantly. Further, the choice of the type of matrix becomes extremely critical, since the size of the matrix crystals incorporated with the analytes should be smaller than the required lateral resolution of the imaging experiment. Also, MALDI is unable to detect metabolites in the low mass range (m/z 1000) due to interference from the high matrix ion background (Bjarnholt et al. 2014).

      Recently, workflows were developed for MALDI imaging of plant carbohydrate, lipid, and protein as well as secondary metabolite localization (Dong et al. 2016b; Peukert et al. 2016; Gupta et al. 2019; Lim et al. 2020).

      Oligofructans, for example, represent one of the most important groups of sucrose‐derived water–soluble carbohydrates in the plant kingdom. In cereals, oligofructans accumulate in aboveground parts of the plants (stems, leaves, seeds), and their biosynthesis leads to the formation of both types of glycosidic bonds (ß (2,1); ß(2,6)‐fructans) or to mixed patterns. In recent studies, MALDI‐MS imaging has revealed tissue‐ and development‐specific distribution patterns of the different types of oligofructans in barley grains, which may be related to the different phases of grain development, such as cellular differentiation of grain tissue and accumulation of storage products (Peukert et al. 2016).

      In the cereal grain, sucrose is converted into storage carbohydrates; mainly starch, fructan, and mixed‐linkage (1,3;1,4)‐β‐glucan(MLG) are formed in ratios, that seem to be tissue‐dependent. Previously, endosperm specific overexpression of the HvCslF6 gene in hull‐less barley led to high MLG and low starch content in mature grains (Lim et al. 2020). Morphological changes included inwardly elongated aleurone cells, irregular cell shapes of the peripheral endosperm and smaller starch granules of the starch‐containing endosperm. Lim et al. (2020) also investigated the physiological basis for these defects by investigating how changes in the carbohydrate composition of the developing grain affect the morphology of mature grains. Visualization of oligosaccharides was performed by MALDI‐MS and complemented by histological studies pointing out the importance of the regulation of carbohydrate distribution for the maintenance of grain‐cell formation and filling processes. Augmented MLG coincided with increased levels of soluble carbohydrates in the cavity and endosperm during the storage phase. Transcript levels of genes related to cell wall, starch, sucrose, and fructan metabolism were disturbed in all tissues. The cell walls of endosperm transfer cells (ETC) in transgenic cereals were thinner and showed reduced mannan labeling compared to the wild type. In the early storage phase, a rupture of the unevenly developed ETC and disorganized neighboring endosperm cells was observed. Soluble sugars accumulated in the developing cereal cave, suggesting a disruption of carbohydrate flow from the cave toward the endosperm, resulted in a shrunken phenotype of the mature grain.

MALDI‐MS imaging was also performed to localize metabolites during the first 7 days of barley germination. Up to 100 mass signals were detected, of which, 85 signals were identified as 48 different metabolites with highly tissue‐specific localizations. Oligosaccharides were observed in the endosperm and in parts of the embryo. Lipids in the endosperm co‐localized, depending on their fatty acid composition, with changes in the distribution of diacyl phosphatidylcholines during germination. Also, 26 potentially antifungal hordatines were detected in the embryo with tissue‐specific localizations of their glycosylated, hydroxylated and O‐methylated derivatives (Gorzolka et al. 2016).

      MALDI‐MS also proved to be the technique of choice for studies in plant lipidomics. While profiles of lipids in seed extracts provide detailed measurements of the amounts and types of lipids present, the information about where these lipids are located in the seed is lost with sample preparation and solvent extraction. At present, mass spectrometric imaging allows the detailed identification and localization of lipids and other metabolites in situ, e.g. directly on tissue samples. The application of this technology to oil seeds has shown that lipid metabolites are unevenly distributed in the different tissues of the seed, indicating a spatial complexity of lipid metabolism that was previously unknown (Sturtevant et al. 2018).

2.2.2 Secondary Ion Mass Spectroscopy

      Although SIMS was the first of the imaging MS techniques presented (introduced for imaging in the1980s by Briggs 1983), its use for imaging of plant material remains limited. SIMS operates under high vacuum and uses high‐energetic (5–40 keV or MeV) primary ions for ionization (Jeromel et al. 2014). Its primary advantage over all other MS imaging techniques is very high spatial resolution (less than 100 nm) (Bjarnholt et al. 2014), while the major disadvantage lies in a high degree of molecular fragmentation. It is expected that high‐throughput analysis and construction of a large database of molecular fragments