Pseudotargeted Metabolomics Reveals the Effects of High Root-Zone Temperature Stress on Leaves and Roots of Medicago sativa

Alfalfa (Medicago sativa L.) is the most widely planted forage, which have higer economically valuable and better ecological benefits in the world. However, it’s sensitive to high root-zone temperature in tropical and subtropical areas. In order to explore the changes of metabolites in alfalfa under heat stress, the study applied the unique alfalfa landrace ‘Deqin’ to explored the mechanisms on the basis of pseudotargeted metabolomics. The results showed that the phenotype of leaves and roots were changed remarkly, leaf chlorophyll content, electrolyte leakage, H2O2 content and soluble sugar also changed at different treatments. Meanwhile, heat stress may stimulate the germination of new alfalfa roots in response to adverse environment. Metabonomics analysis showed that a total of 809 and 749 metabolites were detected in leaves and roots respectively, which treated with four different ways. Of these, 343 and 383 were identified as differential metabolites in the leaves and roots, respectively; they mainly belongs to phenylpropanoids, polyketides, organic acids and derivatives, organic acids and derivatives. Among them, the differential metabolites were mainly enriched in glyoxylate and dicarboxylate metabolism, citrate cycle, aminoacyl-tRNA biosynthesis, flavone and flavonol biosynthesis, and phenylpropanoid biosynthesis. Meanwhile, some metabolites such as phenylpropanoids, polyketides, organic acids and derivatives may play an important role in response to heat stress in M. sativa ‘Deqin’. Overall, this study reports the differential metabolites and pathways in the roots and leaves of ‘Deqin’ in response to high root-zone temperature, which provides new insights into the heat stress mechanisms of alfalfa.


Introduction
High temperature is one of the major environmental stresses affecting plant growth and development (Aditya and Aryadeep 2020). Many plants have developed different strategies, including physiological and metabolic responses, to ensure their survival and reproduction under high temperature stress (Jin et al. 2016). Previous research has shown that high temperature adversely affects photosynthesis, respiration, water relations, and membrane stability in plants, leading to physiological changes Wang et al. 2013). Recent studies have showed that amino acids, carbohydrates, nitrogen metabolisms and metabolites involved in the biosynthesis and catalyzing actions which can help plants to deal with heat stress (Raza, 2020). Moreover, secondary metabolites that play a key role in stress resistance respond rapidly, before the physiological changes are apparent (Akula and Aswathanarayana 2011).
Omics approaches provide important technical support for exploring the physiological and molecular mechanisms of plants under temperature stress (Raza et al, 2021). Among them,the wide application of metabolomics techniques provides a powerful research tool to investigate the mechanisms of plant response to heat (Ravi et al. 2020). Such as nuclear magnetic resonance spectroscopy (NMR), gas chromatography-mass spectrometry (GC-MS), liquid chromatography-mass spectrometry (LC-MS), fourier-transform infrared spectroscopy (FT-IR), matrixassisted laser desorption (MALDI)-ToF-MS, capillary 1 3 electrophoresis-mass spectrometry (CE-MS) (Raza 2020). Those techniques have been applied to different species, for example, several studies on Panicum maximum (Wedow et al. 2019), Sargassum fusiforme (Liu and Lin 2020), and Glycine max (Das et al. 2017) have revealed the vital role of saccharides and amino acids under high temperature stress. Furthermore, it cannot be disregarded that soil temperature, which is more critical than shoot temperature for plant growth, is strongly associated with air temperature (Hou et al. 2015;Miransari 2016). High root-zone temperature is detrimental to the uptake of nutrients and water by plant roots, which in turn influences shoot growth (Shahid and Aamer 2011). Although many studies have investigated the effects of air temperature on the growth of plant shoots (Hatfiled and Prueger 2015), less attention has been paid to the harm of high root-zone temperature.
Alfalfa (Medicago sativa L.) is a perennial forage grass of the Eurasian genus Medicago (family: Leguminosae). Medicago sativa is an excellent plant resource for ecological restoration because of its deep and rapidly growing roots, profuse leaves, short growth cycle, and N 2 fixation ability (Small 2011). In addition, M. sativa is of economic value for its outstanding nutritional quality (Wassie et al. 2019). This plant species is mainly distributed in the arid and semi-arid regions of Northwest and North China. However, high temperature is the major environmental factor that limits M. sativa cultivation in the tropical and subtropical areas of South China (Mo et al. 2011;Song et al. 2019). In such areas with high temperature stress and acid soil infertility, and environmental management using plants with ecological benefits is of practical significance.
Medicago sativa 'Deqin' is a unique alfalfa landrace with strong heat tolerance (Zhao et al. 2010), which can grow and propagate very well in xerothermic valleys of Jinsha River in Yunnan Province, China (Bi et al. 2007). It was used as forage by local villagers with highly digestible and rich in proteins (Bi et al. 2005). Previous research has shown that high root-zone temperature influences the shoot and root growth of M. sativa 'Deqin', with remarkably enhanced root elongation and regeneration at 30 °C (Li et al. 2018a, b). However, the mechanisms underpinning the response of M. sativa 'Deqin' to high root-zone temperature is largely unclear.
In this study, we aimed to identify the heat stress mechanisms of M. sativa 'Deqin' on the basis of pseudotargeted metabolomics. We compared the morphological, physiological, and metabolite changes in whole alfalfa plants in response to high root-zone temperature stress (30 °C) in a short period. The results will provide new insights into the heat stress mechanisms of alfalfa and contribute to its ecological value in South China.

Plant Material, Growth Conditions, and Stress Treatment
Alfalfa (M. sativa 'Deqin') seeds were collected at Deqin county field breeding farm without contamination of alien pollen. These seeds were kindly provided by Professor Yufen Bi from the Faculty of Animal Science and Technology, Yunnan Agriculture University (Kunming, China). The seeds were disinfected with concentrated sulfuric acid for 30 min and rinsed with distilled water. Seed germination was conducted on moist filter paper in an artificial climate box at 25 °C ambient temperature. Fourteen days later, germinated seeds were transferred to polyethylene pots (5 cm × 5 cm × 7.5 cm) containing vermiculite and perlite (v:v = 1:2). The pots were placed in an artificial climate box with 22 °C/20 °C (day/night) ambient temperature, 65% relative humidity, a 16 h photoperiod, and photosynthetically active radiation of 400 µmol m -2 . Alfalfa seedlings were watered with Hoagland's nutrient solution (Barker et al. 2006) every 2 days and maintained in an artificial climate box.
Eight-week-old seedings showing consistent growth characteristics were selected and transferred to a hydroponic box, which was covered with aluminum foil to protect the roots from light. The seedlings were maintained in the hydroponic box for 7 days to allow for adaptation to the root environment and then treated at 30 °C root-zone temperature using a heating rod. On the 8th day, plant samples were collected at different time points for morphological observation, physiological measurement, and metabolomic analysis. In addition, quantitative reverse transcription PCR (qRT-PCR) analysis was performed to verify the related genes associated with metabolites.

Morphological Observation
Leaf and root morphological changes were observed and photographed at 0, 6, 12, 24, 48, and 72 h and 7 days of stress treatment (30 °C, rhizosphere). Then, the temperature was restored to 22 °C for 6 h (H6 h). The angle among leaf, main stem and the width of the third fully expanded leaves at 0, 0.5, 6, 12, and H6 h of stress treatment were measured using a protractor and a straight ruler, respectively. The measurement in all treatment groups was repeated three times.

Physiological Measurement
The chlorophyll content of leaf samples and the electrolyte leakage (EL), H 2 O 2 content, and soluble sugar content of leaf and root samples were examined at 0, 0.5, 6, 12, and H6 h of stress treatment. Specifically, chlorophyll a, chlorophyll b, and carotenoid content was measured according to the method of Li (2000). Electrolyte leakage was measured following the method of Li (2000) with slight modifications. In brief, leaf and root samples with a fresh weight of 0.2 g each were cut into 1 cm pieces and placed into test tubes containing 20 mL of distilled water. The tubes were left to stand at room temperature for 30 min, and the EL was then measured using a conductivity meter (recorded as E 1 ). Subsequently, the tubes were incubated in a boiling water bath for 10 min. After cooling, the solution was mixed by shaking, and the EL was measured again (recorded as E 2 ). In addition, the EL of distilled water was measured (E 0 ). The EL of a sample was calculated as follows: EL (%) = (E 1 −E 0 )/(E 2 −E 0 ) × 100%. H 2 O 2 content was measured using spectrophotometry (Li 2000). Sucrose, fructose, and glucose content was measured according to the method of Ye (2004) and Zhang (1977).

Statistical Data Analysis
The physiological data were expressed as mean and standard error of the mean. One-way analysis of variance was performed to determine significant differences between group means, and statistical significance was assessed by Duncan's test using SPSS (v18.0; SPSS Inc., Chicago, IL, USA). Graphs of physiological parameters were generated using SigmaPlot (v14.0; Systat Software Inc., San Jose, CA, USA).

Metabolite Extraction
Fresh leaf and root samples (200 mg each) were collected at 0, 0.5, 6 h of 30 °C stress treatment, and restored to 22 °C for 6 h (H6) (leaf designated as L1, L2, L3, L4 and the root as R1 R2, R3, R4, respectively). Each treatment group had six biological replicates. The samples were immediately frozen in liquid nitrogen and stored at -80 °C until further processing. The extraction procedure of metabolites was modified from Vos et al. (2007) and Chen et al. (2013). In brief, freeze-dried samples were crushed to powder with a mixer mill for 4 min at 45 Hz and extracted with 75% methanol (10 µL mg -1 ) in Eppendorf tubes kept on a shaker overnight. Following centrifugation (10,000×g, 15 min, 4 °C), the resulting supernatants were filtered through a 0.22 µm filter membrane and transferred to sample vials (2 mL) for liquid chromatography-mass spectrometry (LC-MS) analysis. In addition, the supernatants of all samples were mixed in equal proportions as the quality control (QC) sample for monitoring the analytical process. The QC sample was analyzed with every six samples throughout the LC-MS analytical process.
The parent ions and daughter ions in MS 2 were analyzed to generate a multiple reaction monitoring (MRM) database, and MRM-based analysis of all samples was performed using a triple-quadrupole mass spectrometer (QqQ-MS; Agilent 6490 Series, Agilent) coupled to Triple-TOF-MS, equipped with an Agilent JetStream electrospray ionization interface. The operation parameters were as follows: capillary voltage, + 3500/-3000 V; gas (N 2 ) temperature, 220 °C; gas (N 2 ) flow, 16 L/min; sheath gas (N 2 ) temperature, 350 °C; sheath gas flow, 12 L/min; and nebulizer pressure, 40 psi.

Raw Data Pre-processing
High-resolution MS data were converted to the mzXML format using the ProteoWizard software and then pre-processed using MAPS software. The pre-processed results generated a data matrix consisting of retention time (RT), mass-tocharge ratio (m/z), and peak intensity. Then, the mass spectrum peaks of the same metabolites in different samples were subject to retention time correction, integration correction, area integration using MAPS software to ensure the accuracy of the qualitative and quantitative analyses. Qualitative analysis and classification of metabolites were performed in HMDB (https:// hmdb. ca/). Furthermore, a series of data management steps, including filtering of single deviation mass spectrum peaks, simulating of missing values in the original data and normalizing data according total ion current (TIC), were performed to accuracy of the quantitation results.

Orthogonal Projections to Latent Structure-Discriminant Analysis
The pre-processed data were log transformed and UV scaled using the SIMCA software (v14.1; Sartorius Stedim Data Analytics AB, Umea, Sweden), followed by automated modeling for orthogonal projections to latent structurediscriminant analysis (OPLS-DA). First, the first principal component was subjected to OPLS-DA modeling, and the model quality was tested by sevenfold cross validation. Then, the resulting R 2 Y (the interpretability of the model on the categorical variable Y) was used to evaluate the validity of the model by 200-fold cross validation. Finally, the permutation test of Y variables was performed to generate different random Q 2 values, which were then used to further test model validity.

Screening of Differential Metabolites
Multivariate analysis was used to screen differential metabolites using the following criteria: Log 2 |fold change| > 1, P value < 0.05 and the variable importance in the projection (VIP) of the first principal component in the OPLS-DA model > 1. The classification results and volcano plots of differential metabolites were obtained using SigmaPlot (v14.0) and R (v3.6.5; http:// www.r-proje ct. org/), respectively.

Pathway Annotation and Analysis of Differential Metabolites
The differential metabolites were annotated in the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database (https:// www. kegg. jp/). Then, according to a high impact value and P value, enrichment and topological analysis of the pathways in which the differential metabolites were present enabled further identification of the key pathways that showed the highest correlation with metabolite differences. The bubble plots of differential pathways were produced using R (v3.6.5).
RNA Isolation, cDNA Synthesis, and qRT-PCR qRT-PCR was performed to verify the genes associated with metabolites through the analysis of changes in gene expression of nine key enzymes involved in the significantly enriched pathways of differential metabolites. Eastep® Super Total RNA Extraction Kit and Eastep® RT Master Mix Kit (Promega, Shanghai, China) were used to isolate total RNA and synthesize cDNA, respectively. The sequences of target genes were obtained from the National Center for Biotechnology Information database (NCBI; https:// www. ncbi. nlm. nih. gov/ gene/). Optimal forward and reverse primers (Supplementary Table S1) were designed using the online tool Primer-BLAST (https:// www. ncbi. nlm. nih. gov/ tools/ primer-blast/). The β-actin gene (F: 5′-CGA AAT CGT GAG AGA CGT GA-3′, R: 5′-ACC TGT CCA TCA GGC AAC TC-3′) was used as the internal control (Zhao 2014). qRT-PCR experiments were three technical repeats. The relative expression levels of target genes were estimated using the 2 −ΔΔCT method (Livak and Schmittgen 2001). The results were processed and analyzed in Excel 2007 (Microsoft Corp., Redmond, WA, USA). Graphs were produced using SigmaPlot (v14.0).

Morphological Response to High Root-Zone Temperature
The high root-zone temperature stress directly affected the morphological traits of M. sativa 'Deqin' plants ( Fig. 1a) and reduced their root system (Fig. 1b). Under stress treatment, the shoot and root morphology showed no evident changes at 12 h compared with 0 h, except a significant (85.2%) increase in the angle between leaf and main stem (P < 0.05; Fig. 1c). At 24 h, the bottom leaves showed yellowing, and the number of roots was reduced slightly (Fig. 1a). At 48-72 h, most leaves were dehydrated and rolled back (Fig. 1a), whereas the roots were sparse and soft (Fig. 1b). At 96 h, a small fraction of leaves was etiolated and fell off; however, a few new roots emerged (Fig. 1b). After 7 d, the vast majority of leaves became etiolated, and leaf margins were shriveled, and several new roots were developed (Fig. 1b).

Physiological Response to High Root-Zone Temperature
High root-zone temperature affected the physiological status of M. sativa 'Deqin' plants. Leaf chlorophyll content showed an overall downward tendency with increasing treatment time. Chlorophyll a, chlorophyll b, and carotenoid content first decreased by 26.4%, 41.5%, and 29.9%, respectively, from 0 to 6 h (P < 0.05) and then increased by 10.6%, 13.4%, and 19.6%, respectively, from 6 to 12 h (P < 0.05; Fig. 2a). After restoring the root-zone temperature to 22 °C, the chlorophyll a and carotenoid content at H6 h was 12.7% and 9.8% lower than that at 12 h (P < 0.05), respectively; however, no significant change was observed in the chlorophyll b content. The electrolyte leakage (EL) of leaves and roots increased by 20.6% and 30.0%, respectively, from 0 to 12 h of stress treatment (P < 0.05; Fig. 2b). A larger increase was observed in the corresponding H 2 O 2 content-72.5% and 84.5%, respectively (P < 0.05; Fig. 2c). At H6 h, the EL and H 2 O 2 content were slightly-but not significantly-lower and higher, respectively, than at 12 h. With increasing treatment time, the sucrose content of leaves and roots first increased by 45.2% and 58.3%, respectively, from 0 to 0.5 h (P < 0.05) and then sharply decreased by 76.8% and 85.8%, respectively, from 0.5 to 12 h (P < 0.05; Fig. 2d). The glucose content of leaves and roots also first increased from 0 to 6 h, by 38.1% and 79.7%, respectively (P < 0.05), and markedly decreased at 12 h, by 59.0% and 69.6%, respectively (P < 0.05; Fig. 2e). The fructose content of leaves and roots continuously increased by 66.0% and 85.8%, respectively, from 0 to 12 h (P < 0.05; Fig. 2f). The sucrose content of roots was 35.9% higher at H6 h than at 12 h (P < 0.05), whereas the fructose and glucose content of both leaves and roots was significantly lower at H6 h than at 12 h (P < 0.05).

Qualitative and Quantitative Analyses of Metabolites
The total ion chromatograms of leaf (L1-L4), root (R1-R4), and QC samples in positive and negative modes are summarized in Supplementary Table S2. Among the replicate samples, the area of most chromatographic peaks overlapped, and the appearance time of the peaks was consistent, indicating that the sample quality was satisfactory, and the analytical instrument was stable, and the obtained data were reliable for further analysis. In total, 809 and 749 peaks were detected in the leaves and roots by MS. After comparing the leaf and root metabolites, 809 and 747 peaks were retained. On the basis of the qualitative and quantitative analyses of the peaks, the leaf and root metabolites were classified into 13 and 12 categories, respectively (Supplementary Table S3).

Establishment of OPLS-DA Models
The OPLS-DA of leaf and root metabolite data was designed for modeling two classes of data to increase the class separation and find differential metabolites (Fig. 3). The score plot of OPLS-DA shows that L1, L2, and L3 (Fig. 3a-c) and R1, R2, and R3 ( Fig. 3g-i) were distinct from each other in the leaf group and root group, respectively. All samples were located in the 95% confidence intervals (Hotelling's T-squared ellipse).
In the permutation test, the R 2 Y of each stochastic model was close to but less than 1 (Fig. 3d-f, j-l), indicating that the established models were stable, reliable, and accurate descriptions of the real situation in the leaf and root metabolite data. Moreover, the Q 2 in each stochastic model was less than that in the original model, and the Q 2 in the stochastic model gradually decreased with decreasing retention degree of permutation ( Fig. 3d-f, j-l). These results indicated that the original models were simulated well in statistical significance and not overfitted.

Classification and Volcano Plots of Differential Metabolites
Through screening of metabolites in the leaf and root groups, we identified 343 and 383 differential metabolites, respectively, in the plants under high root-zone temperature stress (P value < 0.05, VIP > 1). The differential and j-l Validation plots of the OPLS-DA models corresponding to the score plots, in which the abscissa (correlation coefficient) represents the retention degree in the permutation test, the ordinate represents the R 2 Y or Q 2 value in the permutation test, and the dotted lines represent the regression lines of R 2 Y or Q 2 metabolites in leaf (L1-L4) comprised unknown metabolites (73.1%); phenylpropanoids and polyketides (6.7%); lipids and lipid-like molecules (5.5%); organoheterocyclic compounds (4.4%); organic acids and derivatives (3.8%); Fig. 4 The classification and volcano plots of differential metabolites in leaves and roots of M. sativa 'Deqin' plants under high root-zone temperature stress. a differential metabolites in leaves; b L2 vs. L1; c L3 vs. L2; d L4 vs. L3; e differential metabolites in roots; f R2 vs. R1; g R3 vs. R2; h R4 vs. R3. L1 to L4 and R1 to R4 indicate leaf and root samples collected at 0, 0.5, 6, and H6 h of stress treatment, respectively. The abscissa of the volcano plot indicates the log 2 foldchange of each metabolite in the sample group, and the ordinate indicates the −log 10 P value of Student's t-test. Each point represents a differential metabolite and the size of the point represents the variable importance in the projection (VIP) of the OPLS-DA model benzenoids (2.3%); organic oxygen compounds (1.5%); organooxygen compounds (1.2%); organic nitrogen compounds (0.6%); nucleosides, nucleotides, and analogues (0.6%); and lignans, neolignans, and related compounds (0.3%; Fig. 4a). The volcano plots show that there were 102 upregulated metabolites in L2 vs. L1 (Fig. 4b) and 118 upregulated metabolites in L3 vs. L2 (Fig. 4c), whereas the corresponding 62 and 63 downregulated metabolites. Meanwhile, there are only 48 upregulated and 33 downregulated metabolites in L4 vs. L3 (Fig. 4d).

Changes of Differential Metabolite Content in Significantly Enriched Metabolic Pathways
According to the pathway analysis, we analyzed the significantly enriched pathways with a high impact value and P value, and the relative content of differential metabolites in the metabolome (Fig. 6). In leaf samples, the relative content of quercetin (Fig. 6a) increased in the beginning and then decreased from 0 to 6 h of stress treatment. The relative content of l-glutamine (Fig. 6b), l-tyrosine (Fig. 6c), l-methionine (Fig. 6d), and l-lysine (Fig. 6e) was higher at 6 h than at 0.5 h. The relative content of naringenin (Fig. 6f), sphinganine (Fig. 6g), and naringenin chalcone (Fig. 6h) changed little from 0 to 6 h of stress treatment and markedly increased at H6 h.
In root samples, the relative content of citric acid (Fig. 6i) and l-malic acid (Fig. 6j) first increased and then decreased from 0 to 6 h of stress treatment. The relative content of riboflavin (Fig. 6k) decreased at 0.5 h compared with 0 h, whereas the relative content of succinic acid (Fig. 6l) decreased at 6 h compared with 0.5 h and continuously decreased at H6 h. The relative content of l-phenylalanine (Fig. 6m) was higher at H6 h than at 6 h, whereas the relative content of 5-hydroxyferulate (Fig. 6n) exhibited an inverse trend.
In order to detect the expression of genes of key metabolite-related enzymes under different treatments, the realtime quantitative PCR were carried out. The results were essentially consistent with the relative RNA expression levels detected by qRT-PCR of key enzymes related to citric acid (CS and ACO) and l-malic acid (MDH1 and aceB) in glyoxylate and dicarboxylate metabolism and l-glutamate (GLT and P5CDH) in alanine, aspartate, and glutamate metabolism (Fig. 7).

Morphological and Physiological Responses of M. sativa 'Deqin' to High Root-Zone Temperature
In this study, we found that high root-zone temperature stress (30 °C) caused evident morphological changes in the shoots and roots of M. sativa 'Deqin' plants. Despite no visable changes in root and leaf morphology within 12 h of heat stress, a considerable increase in the angle between leaf and main stem was observed. This increase in the angle between leaf and main stem is likely to be related to specific phytohormones, such as those playing a vital role in the regulation of leaf inclination in rice (Oryza sativa) (Zhang et al. 2015;Zhao et al. 2010;Zhao et al. 2013a, b). After 12 h of stress treatment, the roots of the plants were reduced substantially and became sparse, whereas the leaves showed yellowing, dehydration, and back-rolling. These changes are similar to previous findings that high root-zone temperature directly inhibited the root vigor of different grain and vegetable crops, such as O. sativa (Wu 2015), tom a to (Lycopersicon esculentum) (Han et al. 2015), lettuce (Lactuca sativa) (Gao and Li 1996), and cucumber (Cucumis sativus) (Moon et al. 2007). As a result, the uptake of water and nutrients by plant roots was inhibited, and the leaves therefore showed yellowing and dehydration.
In the present study, we observed that a few new roots emerged at 96 h, whereas a large number of new roots were  (Guo and Yu 2003) in which root growth almost stopped under high root-zone temperature. The new roots that emerged after 96 h of stress treatment, heat stress may stimulate the germination of new alfalfa roots in response to adverse environment. which is beneficial to recover the water and nutrient uptake and thus promotes shoot growth under high root-zone temperature stress.
Although their morphological changes were not remarkable until 12 h of stress treatment, M. sativa 'Deqin' plants had responded to high root-zone temperature in terms of physiological parameters within 12 h of stress treatment. Photosynthesis is an important pathway to provide energy for plant growth and stress resistance (Yagut et al. 2015). Here, we found that the chlorophyll a, chlorophyll b, and carotenoid content of leaves overall decreased from 0 to 12 h of stress treatment. This decreasing trend of chlorophyll content is similar to the reduction of photosynthesis in Zoysia japonica (Huang 2009), Poa pratensis (He et al. 1997), Rhododendron simsi (Shen and Zhao 2018), and Fargesia yunnanensis (Wang et al. 2014) under high temperature stress. The continued decrease in leaf chlorophyll content observed after the temperature was restored to 22 °C for 6 h may be related to the damage or destruction of photosystems due to high root-zone temperature stress, which led to the decreased energy supply and yellowing of leaves.
High temperature stress can induce oxidative stress, such as the generation of reactive oxygen species (ROS) including singlet oxygen ( 1 O 2 ), superoxide radical (O 2− ), hydrogen peroxide (H 2 O 2 ), and hydroxyl radical (OH − ), which in turn leads to the modification of cell membrane semi-permeability and functions ). Maintaining the relative integrity of cell membranes is an essential approach for plants to resist high temperature. Both EL and H 2 O 2 content are often used to assess the permeability and degree of peroxidation of cell membranes (Sakae et al. 1995;Smirnoff and Arnaud 2019). In the present study, the EL and H 2 O 2 content of M. sativa 'Deqin' leaves and roots increased continuously from 0 to 12 h of stress treatment. Similar increasing trends have been observed for EL in Rhododendron spp. (Li et al. 2018a, b) and Malus domestica (Ma et al. 2010) and for H 2 O 2 content in Cicer arietinum (Kumar et al. 2013), Lotus corniculatus (Zhao et al. 2013a, b), and Lolium perenne (Wang and Xiong 2016). The increase content of EL and H 2 O 2 indicated that M. sativa 'Deqin' plants had suffered damage to cell membranes caused by high root-zone temperature.
Soluble sugars such as sucrose, glucose, and fructose are a class of crucial regulators of plant growth and development, and they play a vital role in plant growth, reproduction, metabolism, and repair after stress-induced damage (Dietze et al. 2014;Kozlowski 1992). A previous study reported that the glucose and fructose content of Pinus ponderosa and P. menziesii (Marias et al. 2017) increased under heat stress, which is in line with our results obtained from the leaves and roots of M. sativa 'Deqin' under high Fig. 6 Changes in the relative content of differential metabolites significantly enriched in metabolic pathways with a high impact value and P value in leaves (gray columns) and roots (black columns) of M. sativa 'Deqin' plants under high root-zone temperature stress. Data are the means of six replicates root-zone temperature stress for 0.5 h. After 0.5 h of stress treatment, sucrose content markedly decreased, whereas glucose and fructose content both increased. A plausible explanation is that sucrose was converted to fructose and glucose to provide energy metabolism and maintain osmotic pressure to reduce cell damage under high temperature stress (Zhao et al. 2006). Between 6 and 12 h of stress treatment, sucrose and glucose content decreased, whereas fructose content continuously increased in M. sativa 'Deqin'. We speculate that M. sativa 'Deqin' responded to high root-zone temperature by converting sucrose to fructose. In summary, the physiological changes in M. sativa 'Deqin' indicate that high root-zone temperature causes damage within the first 0.5 h of stress treatment.

UHPLC-Triple-TOF-MS/QqQ-MS for Pseudotargeted Metabolomics Study in M. sativa
Stability remains a common issue in LC-MS, which limits its application in metabolic studies (Luo et al. 2016). In the present study, an integrated method based on UHPLC combined with two MS techniques (Triple-TOF-MS and QqQ-MS) was used for the first time to our knowledge to perform a pseudotargeted metabolomics study on M. sativa. After UHPLC separation, high-resolution Triple-TOF-MS was used to perform secondary MS scanning with ultra-high coverage to obtain MS 2 spectra of "all metabolites" from the leaves and roots of M. sativa 'Deqin'. Parent ions and daughter ions in MS 2 spectra were analyzed to generate a specific MRM database. Then, QqQ-MS was used to perform accurate qualitative and quantitative analyses of metabolites in the MRM database. Using the integrated method, we obtained high-stability results from M. sativa 'Deqin'. The total ion chromatograms and OPLS-DA results showed that UHPLC-Triple-TOF-MS/QqQ-MS used in this study was a reliable method for pseudotargeted metabolomics.
The total ion chromatograms of metabolites in the leaves and roots of M. sativa 'Deqin' showed a high stability in terms of wider polarity range, higher sensitivity, and better repeatability than other UPLC-Q-TOF-MS and UHPLC-QqQ-MS studies (Xu et al. 2021;Yan et al. 2019). Furthermore, the column filter used in the UHPLC technique was < 2.0 µm, which improved the column voltage to over 15,000 psi, which is considerably higher than that in traditional UPLC (5800 psi). These conditions are beneficial to increase the width of peaks, maintain the separation ability, and reduce the LC analyses time during sample analysis (Kenneth 2011). Given the advantages of UHPLC, this technique is well suitable for combined application with high-resolution Triple-TOF-MS. Thus, pseudotargeted metabolomics based on UHPLC-Triple-TOF-MS/QqQ-MS provides a fundamental solution for the analysis of heat resistance mechanisms of M. sativa 'Deqin'.

Phenylpropanoids and Polyketides in M. sativa 'Deqin' Leaves and Roots in Response to High Root-Zone Temperature
Phenylpropanoids and flavonoids are two major classes of plant metabolites that play a key role in inhibiting the production of ROS against abiotic stress (Commisso et al. 2016). Under high root-zone temperature stress, the relative content of kaempferol and quercetin (involved in flavone and flavonol biosynthesis) in M. sativa 'Deqin' roots decreased from 0 to 0.5 h of stress treatment, whereas the relative content of quercetin in the leaves increased simultaneously. Flavonoids are one of the most important plant pigments that can potentially combat oxidative stress to protect the plants under heat stress (Austen et al. 2019;Brenda 2002;Heim et al. 2002). A previous study showed that high temperature stress induced the flavonoid production in sugarcane (Saccharum officinarum) (Wahid and Close 2007) for thermotolerance, which supports our result of decreased kaempferol and quercetin content in M. sativa 'Deqin' roots. We speculate that kaempferol and quercetin were used to eliminate ROS as the corresponding H 2 O 2 content increased from 0 to 0.5 h of stress treatment.
After restoring the root-zone temperature to 22 °C for 6 h, the relative content of 5-hydroxyferulate and l-phenylalanine (involved in phenylpropanoid biosynthesis) in M. sativa 'Deqin' roots decreased and increased, respectively, whereas the relative content of naringenin, naringenin chalcone, and quercetin in the leaves increased. It is known that phenylpropanoid biosynthesis is activated in plants under abiotic stress, resulting in the accumulation of phenolic compounds that have the potential to restore ROS-induced damage (Sharma et al. 2019). We speculate that high rootzone temperature activated l-phenylalanine biosynthesis in the roots and naringenin, naringenin chalcone, and quercetin biosynthesis in the leaves of M. sativa 'Deqin'.

Organic acids and Derivatives in M. sativa 'Deqin' Leaves and Roots in Response to High Root-Zone Temperature
Organic acids and derivatives can reduce abiotic stressinduced damage to plants by changing antioxidant enzyme activity, osmotic regulator content, and photosynthesis (Bao 2017). Here, we found that under high root-zone temperature stress, the relative content of citric acid and L-malic acid (in glyoxylate and dicarboxylate metabolism) in the roots of M. sativa 'Deqin' increased from 0 to 0.5 h of treatment, followed by a decrease from 0.5 to 6 h of treatment (in glyoxylate and dicarboxylate metabolism and citrate cycle). A similar increase in the relative content of citric acid and L-malic acid has been reported in Cynodon x magennisii under heat stress (Peng 2014). The citrate cycle plays a key role in producing ATP and providing carbon skeletons for various biosynthetic processes (Akram 2014). Our results indicate that M. sativa 'Deqin' responded to high root-zone temperature by rapidly increasing the biosynthesis of citric acid and L-malic acid in glyoxylate and dicarboxylate metabolism within 0.5 h, whereas glyoxylate and dicarboxylate metabolism and citrate cycle were regulated between 0.5 and 6 h, and the period of ATP supplementation was short.
Aminoacyl-tRNA biosynthesis, which involves 20 different enzymes, establishes the rules of genetic code and thus plays a key role in translation and gene expression (Naeem et al. 2019). In the present study, we observed an increase in the relative content of l-glutamine, l-lysine, l-tyrosine, and l-methionine (involved in aminoacyl-tRNA biosynthesis) in the leaves of M. sativa 'Deqin' from 0.5 to 6 h of stress treatment. As aminoacyl-tRNA biosynthesis was shown to be upregulated in Saccharina latissimi (Sandra et al. 2012) under abiotic stress, the increase in the l-glutamine, l-lysine, l-tyrosine, and l-methionine content may be related to the equilibrium of the expression of stress-and growth-related genes to the removal and replacement of damaged proteins in the leaves (Sandra et al. 2012).

Conclusions
The present study reveals the whole-plant morphological, physiological, and metabolic changes of M. sativa 'Deqin' under high root-zone temperature. We identified several pathways that are crucial to resist the high root-zone temperature stress in M. sativa 'Deqin', such as glyoxylate and dicarboxylate metabolism, citrate cycle, aminoacyl-tRNA biosynthesis, flavone and flavonol biosynthesis, and phenylpropanoid biosynthesis. They mainly included phenylpropanoids and polyketides such as quercetin, kaempferol, and luteolin; organic acids and derivatives such as citric acid, L-malic acid, L-glutamine, and L-lysine; organoheterocyclic compounds such as riboflavin; and organic acids and derivatives such as succinic acid. The differential metabolites were primarily enriched in the seven metabolic pathways such as flavone and flavonol biosynthesis. Overall, this study reports the differential metabolites and pathways in the roots and leaves of 'Deqin' in response to high root-zone temperature, These findings expand the knowledge of heat stress mechanisms in alfalfa.