Visualizing the Spatial Distribution of Arctium lappa L. Root Components by MALDI-TOF Mass Spectrometry Imaging

This study is aimed at developing novel analytical methods to accurately visualize the spatial distribution of various endogenous components in Arctium lappa L. (A. lappa) roots, and to precisely guide the setting of pre-treatment operations during processing technologies and understand plant metabolism process. The matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) imaging technology was used for visual demonstration of the in situ spatial distribution in A. lappa roots. This work consisted of four steps: matrix selection, section preparation, matrix coating, and MALDI-TOF MS imaging analysis. Consequently, eight saccharides, four caffeoylquinic acids, four flavonoids, six amino acids, one choline, and one phospholipid were imaged and four unidentified components were found. Saccharides were distributed in the center, whereas caffeoylquinic acids and flavonoids were mainly present in the epidermis and cortex. Furthermore, amino acids were mainly detected in the phloem, and choline in the cambium, while phosphatidylserine was found in the secondary phloem and cambium. This study demonstrated that MALDI-TOF MS imaging technology could provide a technical support to understand the spatial distribution of components in A. lappa roots, which would promote the processing technologies for A. lappa roots and help us to understand the plant metabolism process.


Introduction
Arctium lappa L. (A. lappa), commonly named as burdock, belongs to the family Asteraceae. It has been widely used as a traditional medicinal herb and edible root vegetable for centuries [1]. The seeds, leaves, flowers, and roots of A. lappa are consumed, and its root is the main part used for nutritional and pharmacological purposes. A. lappa root exhibits excellent antibacterial and antifungal [2], antioxidant [3], anti-inflammatory [4], and anti-atherosclerosis [5] properties. These functional properties are attributed to various nutritional components, including polysaccharides, phenolic acids, flavonoids, vitamins, lignans, and polyacetylenes [6][7][8].
At present, there are various analytical methods that can be used to identify endogenous components in A. lappa roots, such as high-performance liquid chromatography-mass spectrometry (HPLC-MS) [9,10], gas chromatography-mass spectrometry [11] and nuclear magnetic resonance (NMR) [12]. However, these methods have some drawbacks, including complex pre-treatment processes and lengthy consumption, which may lead to the degradation and pollution of partial components. In particular, the spatial information regarding the locations of nutritional components in the roots of A. lappa was normally lost over the pre-treatment process, which would not guide its accurate extraction and precision processing.
Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) imaging technology is a powerful tool used to analyze spatial locations of macromolecules without any extraction, purification, or labelling technique [13]. MALDI-TOF MS imaging has been successfully applied to endogenous components' distribution profiling [14] and metabolic process analysis [15]. Recently, this technology has been introduced to the food science field to visualize spatial distributions of metabolites [16], explore complex defense mechanisms [17], and compare plant metabolite changes at different maturity stages [18]. In previous studies, MALDI-TOF MS has been applied to detect the components of A. lappa; however, so far there has been no report regarding mass imaging of A. lappa. Liu et al. [1] analyzed caffeoylquinic acids and lignans from roots and seeds in A. lappa using both HPLC-MS and MALDI-TOF MS methods, in which 80% methanol was used to extract the sample. Many lignans in extract seed samples were identified and no caffeoylquinic acids were detected in seeds and roots of A. lappa using MALDI-TOF MS technology, while their spatial information was destroyed. Furthermore, MALDI-TOF MS was also applied to detect the saccharides of A. lappa roots [5]. Three saccharides with different degrees of polymerization (DP2, DP3, and DP4, [M + Na] + ) in the crude extracts were determined by sonication with 0-60% (v/v) ratio of ethanol, while the spatial information of saccharides in A. lappa were destroyed as well. Most of these studies mentioned above focused on the crude extract of A. lappa and no spatial distribution information of endogenous components was provided. Here, it is proposed that the components could be located by the spatially resolved spectroscopy approach, which would provide important value for exploring novel analysis methods in complex food matrices.
This study is aimed at developing novel analytical methods to accurately visualize the spatial distribution of various endogenous components in A. lappa roots. This study is the first report on the composition and distribution of various endogenous components in A. lappa roots by using MALDI MS imaging. In this work, six traditional organic matrices assisting MALDI-MS analysis were compared for nine representative compounds (L-arginine, proline, phenylalanine, caffeic acid, quercetin, chlorogenic acid, ursolic acid, 1,5-dicaffeoylquinic acid, and 1,3,5-tricaffeoylquinic acid) abundant in A. lappa roots, and the optimal matrices were chosen. Specifically, using optimized MALDI-TOF MS matrices, the root sections were prepared by matrix coating and analyzed using MALDI-TOF MS imaging technology. The results are expected to analyze the spatial distribution information of different endogenous components in A. lappa, and to guide the food processing technology of A. lappa roots and promote the understanding of the plant metabolism process.

Samples and Reagents
Fresh A. lappa roots of Baiji variety (Chinese name) and Yanagawa-riso variety were purchased from Linyi city (Shandong, China).
The optimum cutting temperature (OCT) compound was supplied by SAKURA (Torrance, CA, USA).

Matrix Selection
The matrix plays important roles in separate components in samples, absorbs laser energy effectively, and provides reactive ions that ionize the samples [19]. Matrices suitable for MALDI MS imaging analysis of endogenous components should have several features, including strong optical absorption to ensure high ionization efficiency, minimum matrix background interference, and homogeneous co-crystals between the matrix and the test analytes [20]. For traditional organic matrices, the detection and imaging of low-molecularweight substances (m/z 100~600), such as caffeoylquinic acids, flavonoids, and amino acids, are more susceptible to producing background interference from the matrix. Therefore, selecting the appropriate matrix is important for the successful detection of low-molecularweight substances in real samples.
In order to improve the detection sensitivity, the performance of three traditional organic matrices (DHB, CHCA, SA) assisting MALDI-MS analysis were compared for three representative compounds (L-arginine, proline, phenylalanine) abundant in A. lappa roots in the positive ion reflector mode and the other three matrices (DHAP, 9-AA and NEDC) assisting MALDI-MS analysis were compared for six representative compounds (caffeic acid, quercetin, chlorogenic acid, ursolic acid, 1,5-dicaffeoylquinic acid, and 1,3,5-tricaffeoylquinic acid) abundant in A. lappa roots in the negative ion reflector mode. The optimal matrices were chosen according to the intensity of target analytes and the background interference of matrices in mass spectrometry.
The dried ground steel plate covered with samples and matrices was analyzed using a Rapiflex MALDI Tissuetyper™ TOF/TOF MS (Bruker Daltonik GmbH, Bremen, Germany) equipped with a smart beam Nd:YAG 355 nm laser. The mass spectrometry analysis followed our previous method, with minor modifications [22]. A mixture of DHB and CHCA standards were used for calibration of the mass analyzer, and mixed DHB/CHCA (1:1, w/w) solution was prepared in 0.1% TFA buffer (ACN/MeOH/water, 70:25:5, v/v/v) at 20 mg/mL [23]. The intensity and the global attenuator offset were set to 70% and 15%, respectively, and the repetition rate and frequency were 1000 Hz and 5000 Hz, respectively. The raw spectra were collected in positive reflector mode with the m/z range of 0-1200 and in negative reflector mode with the m/z range of 100-1300. The obtained mass spectra were analyzed using FlexAnalysis Version 4.0 (Build 9) software (Bruker Daltonik GmbH, Bremen, Germany).

Section Preparation of A. lappa Roots and MALDI-TOF MS Imaging
After washing, the fresh A. lappa roots were cut into 0.3 cm pieces and stored in the refrigerator at −80 • C. Prior to sectioning preparation, the A. lappa roots were placed in the refrigerator at −20 • C for 20 min to equilibrate temperature and mounted onto a cryostat sample holder using OCT compound. The mounted samples were then sectioned into 16 µm sections using CryoStar NX50 NOVPD cryostat microtome (Thermo Scientific, Bremen, Germany), and the sections were thaw-mounted onto ITO-coated glass slides. Finally, the slides coated with samples were dried in a vacuum desiccator for 30 min. Subsequently, dried A. lappa root tissue sections were transferred in a sprayer for matrix coating using a HTX TM -Sprayer TM (HTX Technologies, Carrboro, NC, USA). The spraying parameters were as follows: The flow rate of the sprayer was set to 0.03 mL/min at 60 • C, and the track speed was set to 1200 mm/min. The moving distance of the nozzle was 3 mm, the number of spraying cycles was 18, and the spraying density of the optimal matrix on the slice was approximately 0.0025 mg/mm 2 . After matrix spraying, the mixture of DHB/CHCA (1:1, w/w) solution was dropped on the ITO glass slides where there were no sections of A. lappa roots for molecular weight correction [23].
Finally, the ITO glass coated samples were dried in a vacuum desiccator for subsequent MALDI-TOF mass imaging analysis. MALDI MS imaging data were acquired in positive reflector mode with the m/z range of 80-3000 and in negative reflector mode with the m/z range of 80-2000 using Rapiflex MALDI Tissuetyper™ TOF/TOF MS imaging. The spatial resolution of the tissue sections was 200 µm, and the repetition rate, frequency, intensity, and global attenuator offset were 1000 Hz, 5000 Hz, 70%, and 15%, respectively. The obtained mass spectra were analyzed using FlexImaging 5.0 (Bruker Daltonics GmbH, Bremen, Germany) and SCiLS Lab2018b software (Bruker Daltonik GmbH, Bremen, Germany).

Statistical Analysis
Data were expressed as mean ± standard deviation (SD). A one-way analysis of variance was performed by ANOVA, followed by Duncan's test and Tukey's test using SPSS (Version 28.0, Armonk, NY, USA), where p < 0.05 was assumed to be statistically significant.

MALDI-TOF MS Matrix Selection
MALDI matrix selection is important for the desorption and ionization of the test analytes [24]. When MALDI-TOF mass spectrometry was performed, different matrices had distinct desorption and ionization efficiencies, and even matrices themselves may interfere with the detection of test analytes. Furthermore, different types of components might be detected in inverse (positive or negative) ion reflector mode. For example, phenolic acids and flavonoids are easily detected in the negative ion reflector mode, while alkaloids and amino acids are easily detected in the positive mode [1,23,25].
In this work, L-arginine, proline, and phenylalanine were analyzed using DHB, CHCA, and SA in positive ion reflector mode. As shown in Figure 1a Figure S2). Hence, DHB was selected as the most optimal matrix to analyze the spatial distribution of A. lappa roots in the positive ion reflector mode.
In previous studies, NEDC, 1,1 -binaphthyl-2,2 -diamine, 1,5-diaminonaphthalene, and 9-AA were used for the detection of flavones and flavone glycosides in the roots of Scutellaria baicalensis Georgi, and 9-AA was the optimum matrix due to their stronger ion intensities [25]. In this work, caffeic acid, quercetin, chlorogenic acid, ursolic acid, 1,5-dicaffeoylquinic acid, and 1,3,5-tricaffeoylquinic acid as representative components in A. lappa roots were analyzed using DHAP, 9-AA and NEDC in a negative ion reflector mode. All three matrices were able to detect caffeic acid (  Figure S3). Lastly, DHAP has few background interferences, and can be used as a matrix to assist laser desorption/ionization MS imaging to analyze samples ( Figure S1b). Therefore, DHAP was selected as the best matrix to analyze the spatial distribution of flavonoids and caffeoylquinic acids in the negative ion reflector mode of A. lappa roots.

Spatial Distributions of Components in A. lappa Roots
The diameter of A. lappa roots is about 1~2 cm with shirred surface, and the moisture content of the root is approximately 70% on wet basis [26,27]. Figure 2 shows fresh A. lappa roots, sections of A. lappa roots that have been cut, and the tissue structure of the roots. In the MALDI MS imaging process of A. lappa, thousands of individual mass spectra data with per shot (200 µm) of A. lappa tissue were transformed and reconstructed to present the fusion spectra ( Figure S4) and spatial distribution of components on the surface of tissue in the form of images (Figures 3-6).

Spatial Distributions of Saccharides in A. lappa Roots
Through multiple analyses and comparisons, several similar spatial distribution images in the large molecular weight range (m/z 500-1700) were found. Thus, it can be speculated that these components might be the same or similar compounds in root tissues. Additionally, there was a 162 Da difference between the adjacent peaks, which meant that these components were glycans, and the molecular weight of 162 Da corresponded to the mass of a hexose residue [28]. In order to verify the presence of saccharide components, 10-fold volumes of water were added to make homogenate, and MALDI MS was used to analyze the components in the homogenate of A. lappa with same experimental conditions of MALDI MS imaging. As shown in Figure 3a . This might be because of potassium ions were better able to promote the ionization of saccharides than sodium ions, which was related to the distribution and contents of potassium ions and sodium ions in A. lappa. Additionally, the peak intensities of these saccharides decreased with the increasing of molecular weight (m/z) (Figure 3a and Figure S4), which might be because the high-molecular-weight hindered an effective ionization in mass spectrum analysis without any separation or purification process [21]. Therefore, we speculated that the molecular weight of A. lappa saccharides could be larger than the present results (DP10) because the section was thin enough (not enough saccharide contents) to limit the ionization of the polysaccharide. Lastly, polysaccharides were the main ingredient of A. lappa root, and an inulin-type polysaccharide (2676 Da) was extracted with hot water and identified by high-performance anion-exchange chromatography, methylation analysis, NMR, and other analysis methods in our previous study [6]. Therefore, these peaks might be oligosaccharides or polysaccharides of A. lappa roots which were rich in the root center (cambium, pith, secondary xylem, secondary phloem, and cortex) instead of the epidermis (Figures 2 and 3c).
Saccharides are the major bioactive nutrients in food materials or products, including monosaccharides, disaccharides, oligosaccharides, and polysaccharides. They are the most important elicitors to alter physiological and biochemical response in plants, and influence secondary compound productions [29]. Recently, oligosaccharides and polysaccharides obtained from A. lappa have received extensive attention, and inulin-type fructan is the major component. Inulin-type fructan exhibited excellent bifidogenic effect, immunomodulatory activity, antioxidant activity, anti-inflammatory activity, and anticancer activity [26,30]. Therefore, it is very important to study saccharides including inulin-type fructan in A. lappa. At present, the spatial distribution of saccharides in plants has received increasing attention. It has been reported that hexose and sucrose were found in strawberries using MALDI-TOF MS imaging [18]. Furthermore, there have been some reports supporting that MALDI-TOF MS can be used for analyzing the molecular weight of polysaccharide components, which can provide a high sensitive and rapid analysis strategy for characterizing glycan structures [21]. Furthermore, there also have been some reports regarding the spatial distributions of polysaccharides or oligosaccharides in plant tissues. Feenstra et al. [31] reported that the binary MALDI matrix of organic DHB and inorganic Fe 3 O 4 could detect large oligosaccharides, and the MS images of hexose polysaccharide (DP 5~10, [M-H 2 O+K] + ) were obtained from maize seed sections. Sarabia et al. [32] revealed that the oligosaccharides (DP 3~7, [M + Na] + and [M + K] + ) were mostly in the cortex, while the disaccharides (DP 2, [M + Na] + and [M + K] + ) were distributed in the whole section of the roots of barley seedlings. This work was the first report regarding the spatial distributions of polysaccharides or oligosaccharides in A. lappa roots by MALDI-TOF MS imaging, which could provide ideas for the imaging of carbohydrate polymers and the high-value processing of polysaccharides in A. lappa roots.

Spatial Distributions of Caffeoylquinic Acids in A. lappa Roots
Caffeoylquinic acids are important acidic metabolites in plants, and their synthesis and amounts are related with several factors including genotype, harvest time, light, and water supplements [33]. Plant roots often secrete these organic acids to regulate the pH value in plants that improve their adaptability to adverse environments [34]. Additionally, caffeoylquinic acids exhibited excellent free radical scavenging ability and protected against oxidative damage, ulcerative colitis, and alleviated hepatic insulin resistance [35,36]. In this work, the caffeoylquinic acids were found successfully using MALDI mass spectrometry imaging analysis, and the image results of representative caffeoylquinic acids were shown in  (Figure 4d). Two varieties had similar spatial distribution in same components (caffeic acid and caffeic acid phenethyl). Many more types of caffeoylquinic acids could be detected in the Baiji variety, which might be related to composition differences due to different varieties, growing periods, and soil environment.
This work also enriched our previous study that ten caffeoylquinic acid derivatives were obtained by high-speed countercurrent chromatography combined with semipreparative HPLC from A. Lappa roots [8]. Although the MALDI-TOF MS method has been applied to the analysis of caffeoylquinic acid and lignans from the roots and seeds in A. lappa, the sample pretreatment using 80% methanol disrupts their spatial distribution and no caffeoylquinic acids were found using the MALDI-TOF MS [1]. Moreover, the detection method of caffeoylquinic acid was improved by using negative ion reflector mode instead of positive ion reflector and DHAP instead of DHB as matrix. This allowed us to obtain To identify more caffeoylquinic acids, the suitable matrix for caffeoylquinic acids will be further optimized and developed in future research.

Spatial Distributions of Flavonoids in A. lappa Roots
Flavonoids not only play important roles in mediating the response of plants to biological and non-biological environmental stressors as secondary metabolites [37], but also are important ingredients in plants with antioxidant, antibacterial, and hypoglycemic activities [38,39]. It has been reported that the tissues of A. lappa contain quercetin, luteolin, quercetin (quercetin and glycoside), flavanol, and other flavonoids [7,40]. However, there was no report related to detect flavonoids in A. lappa using MALDI-TOF MS imaging technology.
In this work, quercetin, luteolin, liquirtin, and quercetin-glucoside-rhamnoside were  (Figure 4h). From the distribution of flavonoids in A. lappa, the conversion of quercetin to quercetin-glucosiderhamnoside by UDP-glucose dependent glycosyltransferases might occur in the cortex of A. lappa roots [41,42].
Interestingly, the browning degree of the epidermis and cortex was higher than that of the center of the root tissue during our experiments (Figure 2). We observed that the epidermis and cortex were mainly composed of flavonoids and caffeoylquinic acids in the root of A. lappa. Therefore, the browning might be due to the reaction of phenolic acids (flavonoids and caffeoylquinic acids) in the root of A. lappa without epidermal protection with oxygen in the atmosphere.

Spatial Distributions of Amino Acids in A. lappa Roots
Previous studies have shown that the A. lappa root contains a variety of amino acids [43]. Using the DHB as a matrix in the positive ion mode, the spatial location of representative amino acid components in the root of A. lappa was successfully visualized (  (Figure 5d,e). Amino acids are the precursors of proteins and some nitrogenous compounds, which play an important role in plant cell signal transduction, root and bud structure regulation, flowering time regulation and stress defense response [44]. In conclusion, the phloem and pith of A. lappa might be the main parts for storing energy, coping with biological and abiotic stress, and synthesizing secondary metabolites of A. lappa root.

Spatial Distributions of Phospholipids and Choline in A. lappa Roots
Phospholipids are the main components of the cell membrane system, involved in cell energy supply, signaling and membrane trafficking in plant immunity [45]. Phospholipids are present in small amounts in most organisms, and the phosphate group of phospholipids is further esterified by serine to form phosphatidylserine [46]. As shown in Figure 5f Choline is a precursor participating in plant stress metabolites and is involved in the synthesis of phosphatidylcholine, and is a methyl donor in metabolism [47]. Choline is involved in a variety of metabolic pathways and can be interconverted with phosphatidylcholine, glycerophosphocholine, and phosphorylcholine [48]. In addition, choline is a quaternary amine substance which is extremely basic and suitable for mass spectrometry imaging analysis in positive ion detection mode [49]. The mass spectra imaging results of representative choline were shown in Figure 5h, which showed that choline ([M] + , m/z 104.1) was mainly distributed in the cambium of A. lappa root, with a small distribution in xylem, phloem and cortex, and this distribution feature might be related to the biological function of choline.

Spatial Distributions of Unidentified Compounds in A. lappa Roots
Four compounds could not be identified because their fragmentation pattern didn't match with the existing literature and related databases of A. lappa root. As shown in Figure 6 Root hairs play important roles in plant autotrophy and fitness, contributing to anchoring root tips into the soil, providing plants with access to water and nutrients from the soil, and interacting with soil microorganisms [50]. Interestingly, a unique component at m/z 303.1 ([M − H] − ) was found in the root hair in A. lappa. Previous studies have not investigated these unidentified compounds, which may be due to the differences in the variety, origin, planting conditions, climate, and picking time of A. lappa tissue, but they confirmed that the MALDI-MS imaging had the potential to explore new components and provide the spatial distribution of components in the field of food science.

Validation Assay
The A. lappa of the Yanagawa-riso variety was used for the intra-day validation experiment (the experiment was repeated twice, in negative ion reflection mode, with DHAP as the matrix) ( Figure S5). The A. lappa of the Baiji variety was used for the inter-day validation experiment (the experiment was repeated twice, in positive ion reflection mode, with DHB as the matrix) ( Figure S6). The results indicated that both slices of the same sample had similar fusion spectra of the endogenous components, demonstrating the good reproducibility of the method. Furthermore, the MALDI-TOF mass spectrometry imaging method in this study does not require complex sample preparation, and the matrix used is stable and can be stored for a long time, which also ensures its reproducibility.

Conclusions
In the present study, DHB and DHAP were selected as optimal matrices, respectively, in the positive and negative ion reflector mode to analyze components of A. lappa roots tissue. Oligosaccharides or polysaccharides, caffeoylquinic acids, flavonoids, amino acids, choline, and phospholipid were imaged by MALDI-TOF MS imaging in the Baiji variety and Yanagawa-riso variety. The same components were in a similar spatial distribution and different components were found in the root tissue between the Baiji variety and the Yanagawa-riso variety, which might be related to composition differences due to different varieties, growing periods and soil environment, picking time, etc. Among them, choline, phospholipid, and amino acid were easily detected in the Yanagawa-riso variety, while caffeoylquinic acids and flavonoids were easily detected in the Baiji variety. Additionally, some compounds observed by MALDI-TOF MS imaging in A. lappa were still unidentified and require further study. The increasing interest in using MALDI-TOF MS technologies to map the spatial distribution of components will greatly enrich the development of A. lappa roots tissue databases, providing references for developing food processing technologies and the understanding of the plant metabolism of A. lappa roots.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/foods11243957/s1, Figure S1: MALDI MS spectra of matrix background in positive (a) and negative (b) ion reflector mode; Figure Figure S4: MALDI MS imaging fusion spectra of endogenous components in A. lappa using DHB in positive ion reflector mode and using DHAP in negative ion reflector mode. Endogenous components of Baiji variety using DHB in positive ion reflector mode (a) and using DHAP in negative ion reflector mode (b); Endogenous components of Yanagawa-riso variety using DHB in positive ion reflector mode (c), and using DHAP in negative ion reflector mode (d); Figure S5: MALDI MS imaging fusion spectra of endogenous components in Yanagawa-riso variety of A. lappa using DHAP in negative ion reflector mode; Figure  S6: MALDI MS imaging fusion spectra of endogenous components in Baiji variety of A. lappa using DHB in positive ion reflector mode.
Funding: This research was funded by the Burdock Research Fund (Burdock2022001).

Conflicts of Interest:
Author Qiang Chen, Chaojie Zhang were employed by the company Shandong Burdock Medical Research Institute Co. Ltd. The Company had the following involvement with the study: Visualizing the spatial distribution of Arctium lappa L. root components by MALDI-TOF mass spectrometry imaging. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.