Transcriptome and metabolome reveal the primary and secondary metabolism changes in Larix gmelinii seedlings under abiotic stress Authors: Xuting Zhang, Xianling Gao, Bin Liu, Juan Wang, Jinyuan Shan, Jiaxiu Wang, Yanxia Zhang, Guojing Li, Yonghong Jia, Ruigang Wang Published: 2024 ABSTRACT BackgroundLarix gmelinii is an excellent stress resistant coniferous tree species with a wide distribution and important economic and ecological value. However, at seedling stage, L. gmelinii is extremely susceptible to abiotic stresses, and systematic research on the adaptation mechanisms of L. gmelinii seedlings to abiotic stress is still lacking.ResultsPhenotypic observation and physiological index detection showed that L. gmelinii seedlings wilted with needles withered and yellowish at later stages of drought and salt stress; Under low temperature, the seedlings grew slowly and turned red at later stage. Under all 3 abiotic stresses, the chlorophyll content in seedlings significantly decreased, while the MDA content significantly increased; The activity of SOD and CAT showed a trend of increasing first and then decreasing. Transcriptome analysis revealed that DEGs were mainly involved in carbohydrate and amino acid metabolism, phenylpropanoid biosynthesis, and flavonoid synthesis metabolism. Metabolomic analysis found unique DAMs under 3 stress treatments. The combined analysis of transcriptome and metabolome showed that the changing patterns of DEGs and DAMs in primary and secondary metabolism were consistent: carbohydrate were significantly accumulated under low temperature stress; amino acids showed the most significant changes under salt stress. The variation pattern of secondary metabolism was similar under both drought and salt stress, while anthocyanin accumulation was the most obvious only under low temperature stress.ConclusionOur study provides insightful information about the different mechanisms that L. gmelinii seedlings employ in response to drought, low temperature or salt stress.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12870-024-05831-w. Introduction Abiotic stresses are major environmental problems limiting the survival, secondary metabolism and productivity of plant species [1]. The molecular responses to abiotic stresses have been extensively studied in broadleaf species, such as Populus nigra [2], Betula platyphylla [3], Paeonia suffruticosa [4] and Ginkgo biloba [5, 6]. However, the understanding on stress tolerance of coniferous species remains elusive. Larix gmelinii, a representative member of Larix, acts as an afforestation species in north-eastern China, and is also one of the most drought-tolerant species of the Pinaceae family [7]. While the seedling stage of L. gmelinii is extremely vulnerable to abiotic stresses. Up to date, only a few studies addressed part of the mechnisms of abiotic stress response in larch. For example, Wang cloned four DHNs from L. kaempferi stems and found that they enhanced the tolerance of plant cells to osmotic stress [8]. Li developed a hypothetical model of drought tolerance mechanism of L. kaempferi through SMRT-seq and Illumina HiSeq [9]. A combined analysis of the transcriptome and metabolome revealed that L. olgensis prevented free radical-induced damage through glutathione metabolism and responded to drought through sugar accumulation [10]. So far, few studies have been conducted to screen candidate genes and metabolites from L. gmelinii seedlings under abiotic stress by transcriptome and metabolome approach. In this study, important candidate genes and metabolic regulatory substances were analyzed and mined, by measuring physiological and morphological indicators and performing combined transcriptome and metabolome analyses of L.gmelinii seedlings under drought, low temperature or salt stress. We hope that our results could shed light on the global view of stress response mechanisms of L. gmelinii seedlings and other conifers. Materials and methods Plant materials and abiotic stress treatment The seeds of L.gmelinii were provided by the forest farm in Songshan District, Chifeng City, Inner Mongolia Autonomous Region. Mature plump seeds were selected and grown on soil (a 1:3 mixture of peat soil and vermiculit), and cultured at normal condition (24 ℃, 16 h-day/8 h-night) for one month. Abiotic stress treatments were performed using the same batch of healthy and uniform seedlings. For drought treatment, 150 seedlings were well-watered the day before treatment start, then subjected to drought stress by halting water for 30 days. For cold treatment, 150 seedlings were placed in a 4℃ incubator (RUMED) for 30 days with normal water supply. For salt treatment, 150 seedlings were watered with 200mM NaCl every 3–4 d (usually about 300 mL of water or 200mM NaCl solution for each pot planted with 5 seedlings was put into a 80*60 cm tray, and then the water or NaCl solution was adsorbed through the holes in the bottom of the pots from the tray. After fully adsorbed for about 5 h, excess water or NaCl solution were removed from the tray). The treatment and lasted for 30 days. For the control group, 150 seedlings were growing in normal conditions without any treatment. Samples were taken every 5 days (i.e. 0, 5, 10, 15, 20, 25, 30d) until the L.gmelinii seedlings showed obvious apoptotic symptoms. Each sample includes a mixture of five seedlings. For each treatment, 3 biological replicates were performed. All samples were snap-frozen in liquid nitrogen for subsequent experiments. Determination of physiological indexes under abiotic stresses To measure chlorophyll content, approximately 0.1 g of L. gmelinii seedlings with various abiotic stress treatments were cut into pieces and submerged in extraction solution (ethanol and acetone, with a volume ratio of 1:1) overnight to extract the chlorophyll. Next, the extract was measured spectrophotometrically at 663 and 645 nm with a Microplate Reader (PerkinElmer VICTOR Nivo, USA), the total chlorophyll content (mg/g)= (20.2OD645 + 8.02OD663)×V/(1000×w) [11]. The same set of fresh samples were used for malondialdehyde (MDA), catalase (CAT) and superoxide dismutase (SOD) activity analysis, according to the manufacturer’s instructions (Shanghai Youxuan Biotechnology Co., Ltd.): breiefly, MDA was determined by thiobarbituric acid (TBA) method, CAT activity was determined by spectrophotometry, and SOD activity was determined by nitro-blue tetrazolium (NBT) method. Determination of marker gene expression level To assess the effects of drought, cold and salt treatments on L.gmelinii seedlings, the transcript of marker genes DREB1A, P5CS and SOS1 were detected. Total RNA was isolated from samples using a kit (Tiangen, China Cat.#DP441) according to the manufacturer’s instructions, and the RNA was used for reverse transcription with the FastQuant RT Kit with Dnase (TaKaRa, Japan) to synthesize the cDNA. The selected marker genes was validated via RT-qPCR with specific primers and SYBR Green I Master (Roche, Switzerland) and a LightCycler 480 Real Time PCR system (Roche, Switzerland). The thermal cycling program was 95 °C for 1 min, followed by 45 cycles of 95 °C for 5 s, 60 °C for 30 s and 72 °C for 10 s. The melting curves were analyzed at 60–95 °C after 45 cycles. All RT-qPCR reactions were carried out with three technical replicates. The relative RNA transcript level was calculated according to the 2−ΔΔCT method [12]. Tublin gene of L. gmelinii was used as the internal reference to normalize the samples. All primers used for RT-qPCR are listed in Table S1. Transcriptome analysis Based on the analysis of phenotypes, physiological indicators and marker genes expression determination, 7 samples (namely CK, D15, D30, C15, C30, S15 and S30) were selected and subjected for RNA-seq analyses. Each sample included 3 biological replicates. Illumina RNA-Seq was performed by Applied Protein Technology Co., Ltd. (Shanghai, China). After the total RNA passes the quality inspection, mRNA was enriched with magnetic beads with Oligo(dT), and then mRNA was randomly interrupted by adding fragmentation buffer. The first strand of cDNA was synthesised using a six-base random primer using the enriched mRNA as template. The second strand of cDNA was synthesized by adding buffer, dNTPs and DNA polymerase I. The double-stranded cDNA was subsequently purified using AMPure XP beads. The purified double-stranded cDNA was end-repaired, A-tailed and ligated to sequencing junctions, followed by fragment size selection using AMPure XP beads and finally PCR amplification to enrich the final cDNA library. The resulting cDNA libraries were then sequenced via an Illumina HiSeq platform. The raw sequencing data was sorted and filtered, and the clean reads were obtained by removing low quality sequence fragments and those containing adaptor, reads containing ploy-N and low quality reads from raw data. At the same time, Q20, Q30 and GC content in the clean data were calculated. The clean reads assembly was performed using Trinity software [13]. Then unigenes function annotation information was searched in Nt, Pfam, Swiss-Prot, GO and KEGG databases. The fragments per kilobase of exon per million fragments mapped (FPKM) values were then used to normalize the reads from RNA-seq [14]. The DESeq2 was used for differentially expressed genes (DEGs) analysis. The p-value obtained from the original hypothesis test is corrected, which with q ≤ 0.05 and |log2foldchange|≥1 are identified as DEGs [15]. Twenty-seven DEGs were randomly selected for RT-qPCR assays to validate the reliability of RNA-seq analysis. All primers used for RT-qPCR validation are listed in Table S1. Metabolome analysis The same set of samples used for RNA-seq were also subjected to metabolome analyses. Each sample included 6 biological replicates. The non-target metabolome analysis was performed by Applied Protein Technology Co., Ltd. (Shanghai, China). For each sample, 80 mg tissue were grounded in liquid nitrogen into fine powder with a mortar and pestle quickly. Then 1000µL methanol/acetonitrile/H2O (2:2:1, v/v/v) were added to homogenized solution for metabolites extraction. The mixture was centrifuged for 15 min (14000 g, 4 °C). The supernatant was dried in a vacuum centrifuge. The samples were re-dissolved in 100µL acetonitrile/water (1:1, v/v) solvent. The quality control (QC) sample was prepared by mixing an equal aliquot of the supernatants from all of the samples. LC-MS/MS Analyses were performed using an UHPLC (1290 Infinity LC, Agilent Technologies) coupled to a quadrupole time-of-flight (AB Sciex TripleTOF 6600). Raw data were peak aligned, retention time corrected and peak areas extracted using XCMS software [16]. The extracted data were firstly subjected to metabolite structure identification, data pre-processing, and then to quality evaluation, and finally to data analysis [17]. Multivariate statistical analysis of metabolites, including principal component analysis (PCA) and partial least squares analysis (PLS-DA), was carried out to reveal the differences in the metabolic composition of different alignment groups. The variable importance in the projection (VIP) value of each variable in the OPLS-DA model was calculated to indicate its contribution to the classification. Metabolites with VIP > 1 and P < 0.05 were considered differential accumulation metabolites (DAMs). Results Phenotypic and physiological changes of L. gmelinii seedling under abiotic stresses The seedlings grew significantly taller under normal conditions, while the overall growth was slow down under drought, low temperature and salt stress, with mild wilting under drought and salt stress at 15d, and much slower growth under low temperature stress. Wilting was severe under drought and salt stress at 30d, while trunk and needles of seedlings turned red under low temperature stress (Fig. 1 and Figure S1). Fig. 1Phenotype of L. gmelinii seedlings under CK, drought, low temperature and salt stress Phenotype of L. gmelinii seedlings under CK, drought, low temperature and salt stress Measurement of physiological responses of L. gmelinii seedlings showed that total chlorophyll content has little change during 0-10d of stress, and a significant decrease was detected by 15d under stress. The MDA content increased to different degrees, indicating that the damage of membrane lipid peroxidation became greater with prolonged stress treatment, and the change in MDA was more obvious after 15d of stress. The pattern of changes in CAT and SOD activity was similar, both showed a trend of increasing and followed by decreasing as the duration of stress treatment increased, with the enzyme activity reaching its highest value during 10-15d of stress (Fig. 2). Fig. 2Physiological indicators of L. gmelinii seedlings under drought, low temperature and salt stress. Within each set of experiments, the different letters indicate significant differences (P < 0.05) Physiological indicators of L. gmelinii seedlings under drought, low temperature and salt stress. Within each set of experiments, the different letters indicate significant differences (P < 0.05) Detection of stress marker genes expression To further verify the effectiveness of drought, low temperature and salt stress on L. gmelinii seedlings, the expression of stress marker genes P5CS, SOS1 and DREB1A were examined under each treatment. The expression of P5CS changed significantly under drought and salt stress, with a significant increase after 15d of drought treatment, and up to over 8-fold at 30d, while up to 10-fold after 20d of salt stress. The expression of SOS1 was up-regulated under all three stresses, with a significant change after 15d, especially under salt stress. The expression of DREB1A was significantly up-regulated under all three stresses, especially after 15d of low-temperature stress, where the expression was over 30-fold. After 25d of drought or salt stress, the transcript of DREB1A also increased more than 10-fold, compared to the control (Fig. 3). Fig. 3The transcript level of 3 marker genes of L. gmelinii seedlings under drought, low temperature and salt stress. Within each set of experiments, the different letters indicate significant differences (P < 0.05) The transcript level of 3 marker genes of L. gmelinii seedlings under drought, low temperature and salt stress. Within each set of experiments, the different letters indicate significant differences (P < 0.05) The above results showed that drought, low temperature and salt stress were effective for L. gmelinii seedlings, and the three marker genes changed significantly after 15d of stresses. Meanwhile, in combination with the phenotype and physiological indicators, we decided to choose samples of the control, and those with 15 and 30d of stress treatments for subsequent transcriptome and metabolome analysis. Changes in transcriptome of L. gmelinii seedling under abiotic stresses Totally 15G clean data were obtained after filtration, each with a Q30 base from 92.99 to 93.75% and a GC content from 44.52 to 45.32% (Table S2). A total of 568,660 unigenes were obtained, of which the average and N50 lengths were 713 and 1,201 bp, respectively. The transcripts were then sequentially annotated using several databases, annotation information was obtained for only 118,901 of 568,660 assembled unigenes, occupied 20.91%. Annotation analysis showed that the number of annotated genes was the largest in the NR database, the next large number were from Swiss-Prot and PFAM, and the GO and KEGG databases contained the least annotation information (Table S3). The results of PCA analysis showed good reproducibility within sample groups from different treatments, and there were significant changes between the different groups. The distribution of samples treated for 30d differs greatly among three treatments. The distribution of samples under drought and salt stress for 15d was similar to that of the control group, while the distribution of samples under low temperature stress for 15d and 30d were similar, indicating that L. gmelinii seedlings were more sensitive to low temperature, consistent with the observed phenotype (Fig. 4). Fig. 4Correlation scatter plots of different treatment samples of L. gmelinii seedlings Correlation scatter plots of different treatment samples of L. gmelinii seedlings The expressions of DEGs under CK, D15 and D30; CK, C15 and C30; CK, S15 and S30 abiotic stresses were analyze. In the nine comparative datasets, the most DEGs were detected in S30 vs. CK pair, and fewer DEGs were found in C30 vs. C15 pair (Figure S2). Based on the KEGG enrichment analysis of these DEGs, the ten pathways were significantly enriched in 3 abiotic stress treatments (P < 0.05), including plant hormone signal transduction, phenylpropanoid biosynthesis, photosynthesis, flavonoid biosynthesis, glutathione metabolism, and starch and sucrose metabolism (Table 1). Table 1Significant enrichment analysis of DEGs pathways (P value < 0.05)PathwayPathway IDD15vsCKD30vsCKD30vsD15C15vsCKC30vsCKC30vsC15S15vsCKS30vsCKS30vsS15Plant hormone signal transductionko040756786517769335410796Phenylpropanoid biosynthesisko00940616256606743549796Photosynthesisko001954634355034456450Flavonoid biosynthesisko009413832392729225451Glutathione metabolismko004802741414523406561Starch and sucrose metabolismko0050021363032365049Amino sugar and nucleotide sugar metabolismko0052026524756303767Cutin, suberine and wax biosynthesisko00073282823262734MAPK signaling pathway - plantko040163244332640Arginine and proline metabolismko0033029143635 Table 1Significant enrichment analysis of DEGs pathways (P value < 0.05)PathwayPathway IDD15vsCKD30vsCKD30vsD15C15vsCKC30vsCKC30vsC15S15vsCKS30vsCKS30vsS15Plant hormone signal transductionko040756786517769335410796Phenylpropanoid biosynthesisko00940616256606743549796Photosynthesisko001954634355034456450Flavonoid biosynthesisko009413832392729225451Glutathione metabolismko004802741414523406561Starch and sucrose metabolismko0050021363032365049Amino sugar and nucleotide sugar metabolismko0052026524756303767Cutin, suberine and wax biosynthesisko00073282823262734MAPK signaling pathway - plantko040163244332640Arginine and proline metabolismko0033029143635 Significant enrichment analysis of DEGs pathways (P value < 0.05) To verify the accuracy of the transcriptome data, we assayed the expression of 27 DEGs randomly selected by RT-qPCR In general, the results were consistent with the RNA-seq data, despite some differences in expression levels (Figure S3). This means that the transcriptome data can be further applied to the study of drought, salt and low temperature stress of L.gmelinii seedling. Metabolites of L. gmelinii seedlings changed significantly under abiotic stresses A total of 428 metabolites were identified by non-targeted metabolomics using LC-MS from L.gmelinii seedlings. Utilizing OPLS-DA analysis, it can be seen that the two sets of data for the three comparative groups under drought, low temperature, and salt stress are clearly distinguished (Figure S4), supporting that selection of time points of 15 and 30 d after stresses is reasonable for metabolomic analysis. The DAMs of 15d and 30d under 3 abiotic stresses vs. CK were analyzed statistically in a thermogram according to different classifications. The results showed that carbohydrates accumulation changed most significantly under low temperature. Most lipids were significantly reduced, and most amino acids remarkably changed under 3 abiotic stresses. Among the nucleosides, the content of adenine, cytosine, cytidine and uracil, with the exception of adenosine, were significantly reduced under drought and salt stress. Hormones changed obviously under 3 abiotic stresses. Flavonoids showed the most significant changes among the secondary metabolites (Fig. 5). Fig. 5Heat map of DAMs in different classifications L. gmelinii seedlings under different stress treatments compared with the control. DAMs (P-value < 0.05 and VIP ≥ 1). Different colors indicate the magnitude of the metabolite in a log2 value in different comparisons, with purple indicating a larger value and green indicating a smaller value Heat map of DAMs in different classifications L. gmelinii seedlings under different stress treatments compared with the control. DAMs (P-value < 0.05 and VIP ≥ 1). Different colors indicate the magnitude of the metabolite in a log2 value in different comparisons, with purple indicating a larger value and green indicating a smaller value In order to find marker metabolites of early and late stress duration, further statistics were conducted on DAMs under 3 abiotic stresses for 30d vs. 15d. Among carbohydrates, the accumulation of raffinose, stachyose, and melibiose was most significant at C30; and fucose accumulated remarkably at S30. Stearidonic acid accumulated highly at S30. Among amino acids, proline accumulated significantly at D30 and S30; arginine, isoleucine and phenylalanine accumulated obviously at D30. Among flavonoids, anthocyanins increased most significantly at C30; taxifolin increased strikingly at D30; catechin increased remarkably at S30. Other metabolites, such as 2-Oxadipic acid, increased significantly at S30 (Fig. 6). Fig. 6Heat map of DAMs in different classifications in L. gmelinii seedlings under different stress for 30d vs. 15d. DAMs (P-value < 0.05 and VIP ≥ 1). Different colors indicate the magnitude of the metabolite in a log2 value in different comparison pairs, with purple indicating a larger value and green indicating a smaller value Heat map of DAMs in different classifications in L. gmelinii seedlings under different stress for 30d vs. 15d. DAMs (P-value < 0.05 and VIP ≥ 1). Different colors indicate the magnitude of the metabolite in a log2 value in different comparison pairs, with purple indicating a larger value and green indicating a smaller value Therefore, samples under different stress can be distinguished based on changes in specific DAMs. For example, with the prolongation of low temperature stress, the accumulation of carbohydrates such as raffinose, melibiose, and stachyose became more obvious, and the accumulation of cyanidin 3-glucoside cation and peonidin 3-galactoside cation became more obvious at the late stage of low temperature stress. At the early stage of drought stress, carbohydrate changes were not significant, but at the late stage of drought stress, the accumulation of trehalose, maltose, fructose, and mannose increased; some amino acids such as pyroglutamic acid, arginine and phenylalanine decreased first and then increased during drought stress; while taxifolin increased strikingly at the late stage of drought stress. The accumulation of amino acids are the most evident under salt stress, such as proline and histidine, while fucose, stearidonic acid, and 2-oxyadipic acid were more distinguishable at the late stage of salt stress (Fig. 6). Carbohydrate and amino acid metabolism in L. gmelinii seedling under abiotic stress Carbohydrates are important energy sources for plant growth and metabolism, play an important role in environmental adaptation by integrating internal and external information to balance nutrition and regulate development in response to abiotic stresses [18]. In carbohydrate metabolism, DEGs and DAMs exhibit close consistency, and are significantly enriched in starch and sucrose metabolism, fructose and mannose metabolism, and galactose metabolism pathways (Fig. 7 and Figure S5). Fig. 7Metabolic pathway changes of main differential carbohydrates in L. gmelinii seedlings under different stress treatments. The asterisk represents the DAMs of VIP ≥ 1 in the three stress treatments for 15d and 30d, with * representing P < 0.05 and * * representing P < 0.01. Different colors indicate the magnitude of the metabolite in a log2 value in different comparison pairs, with purple indicating a larger value and green indicating a smaller value Metabolic pathway changes of main differential carbohydrates in L. gmelinii seedlings under different stress treatments. The asterisk represents the DAMs of VIP ≥ 1 in the three stress treatments for 15d and 30d, with * representing P < 0.05 and * * representing P < 0.01. Different colors indicate the magnitude of the metabolite in a log2 value in different comparison pairs, with purple indicating a larger value and green indicating a smaller value In starch and sucrose metabolism, UDP-glucose, trehalose and maltose accumulated most significantly under low temperature stress, especially trehalose, which increased more than 16-fold at C30 (Figure S5). Compared with low temperature stress, sucrose accumulated strikingly under drought and salt stress; while the content of UDP-glucose decreased under drought and salt stress. The possible reason is that UDP-glucose, as an intermediate product of both sucrose-trehalose and sucrose-starch pathways, increased sucrose decomposition under low temperature stress, and significantly accumulated trehalose and maltose in a consistent pattern with UDP-glucose. DEGs involved in sucrose-trehalose metabolism were significantly upregulated under three stresses, such as those encoding sucrose phosphate synthase (SPS), sucrose synthase (SUS), trehalose 6-phosphate synthase (TPS), and trehalose 6-phosphate phosphatase (TPP); The expression of DEGs encoding two amylases (AMY and β-AMY) and 4-alpha-glucanotransferase (malQ) involved in starch degradation were also upregulated, which was closely related to the accumulation of sucrose, trehalose, and maltose to varying degrees under 3 abiotic stresses (Fig. 7). Similarly, changes in the galactose metabolism were the most significant under low temperature stress, with lactose, galactinol, raffinose, stachyose, and melibiose increasing 40, 15, 50, 30 and 140 folds respectively at C30 (Figure S5). Galactinol decreased under drought and salt stress. Most DEGs encoding inositol 3-alpha-galactosyltransferase (GOLS) and raffinose synthase (RFS) were upregulated under 3 abiotic stresses, among them, TR11716_c1_g1 encoding stachyose synthase (SCS) was only significantly upregulated under low temperature stress, which was closely related to the striking accumulation of raffinose family. In the fructose and mannose metabolism, sucrose can be hydrolyzed into fructose by invertase (Fructofuranosidase, INV), DEGs encoding INV were upregulated under drought and salt stress, while downregulated under low temperature stress. It is hypothesized that this is closely related to fructose and mannose metabolism, which did not change remarkably under low temperature stress. TR3468_c0_g1 encoding hexokinase (HK) was upregulated under S30, which was presumably related to the obvious accumulation of mannose-6 phosphate in S30 closely (Fig. 7). Glucose, as the junction of metabolism, was accumulated most significantly under low temperature stress. Other carbohydrates, such as talose, accumulated significantly under low temperature stress either, increasing by 10 and 20 folds at C15 and C30 compared to the CK, respectively. The accumulation of fucose was significant at S30, reaching more than 9 folds compared to that of the CK. The accumulation of maltitol was striking under drought and salt stress, and increased by 8 and 15 folds at D30 and S30 compared to that of the CK, respectively (Figure S5). The composition and accumulation of amino acids in plants are crucial for their adaptation to adversity [19]. The organic acids produced by the carbohydrate EMP pathway and TCA cycle provide a carbon framework for the synthesis of amino acids, which can be divided into five categories based on the source of the synthetic carbon framework [20, 21]. Aromatic amino acids, such as tryptophan, phenylalanine, and tyrosine, all increased under 3 abiotic stresses, and their synthesis requires the shikimic acid pathway, providing important precursors for secondary metabolism. Drought stress has significantly reduced glutamate of the α-ketoglutaric acid group, aspartate in the oxaloacetic acid group, and alanine in the pyruvate pathway; while significantly increased proline in the α-ketoglutaric acid group, asparagine and isoleucine in the oxaloacetic acid group, and leucine in the pyruvate pathway. Low temperature stress strikingly increased glutamate, proline, aminobutyric acid (GABA), and citrulline in the α-ketoglutaric acid group, and aspartate in the oxaloacetic acid group. The down-regulated DAMs under both salt stress and drought stress were consistent. However, most amino acids changed significantly, including glutamine, histidine, proline, and arginine of the α-ketoglutaric acid group, and asparagine of the oxaloacetic acid group, under salt stress (Fig. 8 and Figure S6). Fig. 8Carbon and nitrogen metabolism network of L. gmelinii seedlings under different stress treatments. The asterisk represents the DAMs of VIP ≥ 1 in the three stress treatments for 15d and 30d, with * representing P < 0.05 and * * representing P < 0.01; different colors indicate the magnitude of the metabolite in a log2 value in different comparison pairs, with purple indicating a larger value and green indicating a smaller value Carbon and nitrogen metabolism network of L. gmelinii seedlings under different stress treatments. The asterisk represents the DAMs of VIP ≥ 1 in the three stress treatments for 15d and 30d, with * representing P < 0.05 and * * representing P < 0.01; different colors indicate the magnitude of the metabolite in a log2 value in different comparison pairs, with purple indicating a larger value and green indicating a smaller value It is presumed that histidine may have been degraded to glutamate, which accumulated the most and degraded the least under drought and salt stress. Glutamate, an important substrate for synthesis of glutamine, arginine, proline and GABA, etc., decreased under drought and salt stress, and was converted to synthesize other amino acids, especially under salt stress. A large amount of glutamine, proline and arginine were detected to accumulate under salt stress, while glutamate, GABA and citrulline accumulated greatly under low temperature stress. Aspartate, an important substrate for synthesis of amino acids such as asparagine, threonine and isoleucine, decreased significantly under both drought and salt stress; among them, asparagine and isoleucine accumulated highly under drought stress. In contrast, alanine, which is involved in the TCA cycle, was significantly reduced under drought and salt stress (Fig. 8 and Figure S6). In summary, significant changes in carbohydrate and amino acid metabolism occurred in L. gmelinii seedlings under drought, low temperature, and salt stress, including the EMP pathway, TCA cycle, shikimic acid pathway, and amino acid metabolism. The accumulation of glucose and glucose 6-phosphate increased, indicating that the EMP pathway was activated under 3 abiotic stresses. The content of citric acid, succinic acid, and malic acid in the TCA cycle decreased. Striking changes were also observed in the five major groups of amino acids using the intermediates from EMP pathway and TCA cycle as synthetic precursors. The mechanisms of coordinated regulation of carbohydrate and amino acid metabolism in L.gmelinii seedlings to adapt to abiotic stresses are complex. Secondary metabolism in L. gmelinii seedling under abiotic stresses Statistical analysis of differential accumulation of secondary metabolites under 3 abiotic stresses were carried out (Figure S7). The anthocyanins were the most pronounced metabolites increased under the low temperature stress, for example, cyanidin 3-glucoside cation and peonidin 3-galactoside significantly elevated 170-fold and over 20-fold at C30 respectively, and procyanidin increased over 10-fold at C30. The flavonoids, such as kaempferol strikingly increased by more than 2-fold at C30, and decreased by 5-fold and 2-fold at D30 and S30, respectively. Vitexin changed significantly under low temperature stress, rising more than 3-fold at C30, and almost unchanged under drought and salt stress. Taxifolin accumulated significantly under drought and salt stress, with an obvious increase of over 8-fold at D30. Catechin accumulated highly under low temperature and salt stress, with a significant increase of more than 5-fold at S30. Phloridzin was significantly decreased by about 2-fold at D30, while myricetin was significantly reduced by 5-fold at S30. Hyperoside and kaempferol 3-O-rutinoside increased greatly under low temperature stress, by more than 6-fold and 3-fold at C30 respectively, but with little change under drought and salt stress. The isoflavones, such as genistein, accumulated strikingly under drought and salt stress, and increased more than 5-fold at D30. In lignin metabolism, 4-hydroxycinnamic acid was the most obvious metabolite at S30. Ferulic acid has certain accumulation at D30 and C30; while matairesinol with more than 6-fold and 8-fold elevations at D30 and S30 respectively. In astragalus metabolism, resveratrol showed an over 7-fold and 6-fold elevation at D30 and S30 respectively (Figure S7). The DEGs and DAMs in phenylpropane and flavonoid metabolism exhibited close consistency, forming a secondary metabolic regulatory network for the adaptation of L. gmelinii seedlings to abiotic stresses (Fig. 9). Phenylalanine ammonialyase (PAL) is the first key enzyme in the phenylpropane biosynthesis, catalyzing the deamination of phenylalanine to produce cinnamic acid, and is a key and rate-limiting enzyme in the synthesis of lignin, flavonoids and astragalus [22]. Most of the DEGs encoding PAL were up-regulated under 3 abiotic stresses, especially TR376_c0_g1. 15, which was detected to be significantly correlated with lignin biosynthesis. The DEGs encoding 4-coumarate–CoA ligase (4CL) were significantly upregulated under drought and salt stress. The DEGs encoding trans-cinnamate 4-monooxygenase (CYP73A) were strikingly upregulated under 3 abiotic stresses, especially TR3888_c0_g1 and TR3598_c0_g1. TR3872_c0_g2 and TR3872_c0_g1 encoding caffeic acid 3-O-methyltransferase (COMT) were also significantly upregulated under 3 abiotic stresses; while TR26559_c0_g1 encoding cinnamoyl CoA reductase (CCR) was upregulated under drought and salt stress. In addition, TR769_c1_g3 encoding cinnamyl alcohol dehydrogenase (CAD) was upregulated under 3 abiotic stresses, and the significant up-regulation of some DEGs under drought and salt stress may be an important reason for matairesinol accumulation under drought and salt stress. Key genes for flavonoid biosynthesis were up-regulated in response to 3 abiotic stresses, such as DEGs encoding chalcone synthase (CHS), chalcone isomerase (CHI), naringenin 3-dioxygenase (F3H), flavonol synthase (FLS) and flavonoid 3’-monooxygenase (CYP75B1); and the DEGs encoding flavanone 4-reductase (DFR), anthocyanidin synthase (ANS) and leucoanthocyanidin reductase (LAR), which are related to anthocyanin synthesis, were up-regulated under drought and salt stress. The DEGs encoding anthocyanidin reductase (ANR) were significantly up-regulated under 3 abiotic stresses. These results may be closely related to the great changes in taxifolin, catechin, anthocyanin, and genistein under 3 abiotic stresses (Fig. 9). Fig. 9Combined analysis of the secondary metabolites content and related gene expression changes in L. gmelinii seedlings. The asterisk represents the DAMs of VIP ≥ 1 under the three stress treatments for 15d and 30d, with * representing P < 0.05 and * * representing P < 0.01; different colors indicate the magnitude of the metabolite in a log2 value in different comparison pairs, with purple indicating a larger value and green indicating a smaller value Combined analysis of the secondary metabolites content and related gene expression changes in L. gmelinii seedlings. The asterisk represents the DAMs of VIP ≥ 1 under the three stress treatments for 15d and 30d, with * representing P < 0.05 and * * representing P < 0.01; different colors indicate the magnitude of the metabolite in a log2 value in different comparison pairs, with purple indicating a larger value and green indicating a smaller value Fig. 10Regulatory network for drought, low temperature and salt stress of L. gmelinii seedlings. The black box represents the common changes under three stresses; The red, gray, and blue boxes represent changes under drought, low temperature, and salt stress, respectively Regulatory network for drought, low temperature and salt stress of L. gmelinii seedlings. The black box represents the common changes under three stresses; The red, gray, and blue boxes represent changes under drought, low temperature, and salt stress, respectively Flavonoids, lignans and astragals share the same precursor synthesis pathway and changes in flavonoid content are inextricably linked to changes in lignans and astragals [22]. The regulatory network of secondary metabolism in L. gmelinii seedlings under 3 abiotic stresses revealed significant changes in the content of secondary metabolites related genes. The high accumulation of anthocyanins was one of the main reasons for the purplish-red colour of stem segments and needles of L. gmelinii seedlings under low temperature stress. The patterns of change of the most secondary metabolites were similar with their genes regulation under drought and salt stress, for example, matairesinol and resveratrol increased under drought and salt stress. Discussion Transcriptomics and metabolomics are effective methods to analyze the relationship between key metabolites and genes in biosynthetic pathways. In this study, we used integrated transcriptome and metabolome analyses to clarify the molecular mechanisms underlying L. gmelinii seedlings under drought, low temperature and salt stress. Changes in carbohydrate metabolism The present study found that some carbohydrates accumulated more significantly under low temperature stress than drought and salt stress. For example, Trehalose increased more than 15-fold at C30 (Figure S5). Trehalose, one of the most chemically stable sugars, can be used as an energy source, osmolyte and protein protector [23]. It has been shown that A. thaliana had an 8-fold increase in trehalose content after 4d of treatment at 4 °C [23]. Oryza sativa also showed a rapid increase in trehalose content under low temperature [24]. In addition, exogenous trehalose pre-treatment increased cold tolerance of pepper-fruit; while exogenous trehalose significantly improved the survival rate of cold-stressed Brassica napus seedlings [25]. Therefore, trehalose is an important osmotic regulator involved in the response of L. gmelinii seedlings to low temperature stress. Lactose, galactitol, raffinose, stachyose and melibiose increased more than 40, 15, 50, 30, and 140-fold at C30, respectively (Figure S5). It has been shown that the raffinose family, including raffinose and stachyose, plays a role in carbon storage, signal transduction, and osmotic protection [26]. The glucose, raffinose and stachyose of Distichlis spicata increased continuously with decreasing temperature during the overwintering period, reaching the highest level during the coldest period of winter [27]. Raffinose content increased by 49% after 2 weeks of 4 °C treatment in Triticum aestivum compared to the control [28]. A study on the mechanism of cold resistance in Picea obovata found that raffinose content was the most variable sugar [29]. In raffinose synthesis, TR9810_c0_g1, which encodes GOLS, increased most strikingly under low temperature stress; and TR11716_c1_g1, which encodes SCS, was significantly up-regulated only under low-temperature stress (Fig. 7). Hence, it is speculated that changes in the content of raffinose family could be one of the important indicators of cold tolerance of larch species. It has been shown that under drought stress, sugar alcohols accumulated in cells and increased osmotic stress tolerance [30]. In our results, maltitol was found to increase highly under drought and salt stress, but did not change obviously under low temperature stress. We also found that fucose accumulated significantly under salt stress treatment, increasing more than 9-fold at S30 (Figure S5), which has not been reported in other studies so far. And talose was found to accumulate highly under low temperature stress for the first time too (Figure S4). It was worthy to explore further to uncover their roles and regulation mechanism in abiotic stress response of L. gmelinii seedlings and other plant species. Changes in amino acid metabolism Glutamate is involved in response and adaptation to abiotic stresses, and glutamate metabolism was activated under cold stress to improve freezing resistance in (A) thaliana, Hordeum vulgare and T. aestivum through production of proline and GABA [31, 32]. Exogenous glutamate application enhanced proline synthesis and play an essential role in improving drought stress in (B) napus [33]; and exogenous glutamate also reduced the inhibitory effect of salt stress on Cucumis sativus seeds [34]. In addition, exogenous glutamate increased cold tolerance in A. thaliana and barley, but showed contrary effect in maize, rice and tomato, with certain differences in cold tolerance of glutamate metabolism among different species [35]. Glutamine and arginine are important amino acids for the storage and transport of nitrogen [36, 37]. Glutamine is an effective source of nitrogen and its application promoted nitrogen and biomass accumulation in maize seedlings [38]; exogenous glutamine as a single source of nitrogen benefited poplar growth [39]. Exogenous arginine improved cold tolerance in papaya fruit [40], and significantly increased the proline content of wheat seedlings under drought [41]. Proline and GABA are involved in a variety of abiotic stresses and act as important osmoregulators, maintaining cellular osmotic balance, reducing water dissipation in plants and scavenging free radicals [42]. We found that the amino acids of the α-ketoglutarate group in L. gmelinii seedlings changed most significantly under 3 abiotic stresses, for example, proline accumulated significantly under 3 abiotic stresses, especially under salt stress. Glutamate was decreased under drought and salt stress, which may provide a framework for proline synthesis, then enhance drought and salt stress resistance through significant accumulation of proline. Under low temperature stress, glutamate accumulated slightly, while GABA accumulated most significantly. Glutamine and arginine increased most strikingly under salt stress. Histidine increased greatly under drought and salt stress, possibly due to less degradation of histidine to glutamate (Figure S6). Aspartate is involved in biosynthesis of many amino acids and plays a critical role in plant resistance to abiotic stresses [43]. Previous studies have found that the content of aspartate in Caragana korshinskii decreased with drought stress [44]. However, under low temperature stress (2 ℃), the content of aspartate in strawberry leaves rapidly increased [45]; and the levels of aspartate and glutamate were up-regulated in fig fruit during cold storage, while the levels of most other free amino acids decreased [46]. In this study, aspartate was found to accumulate in the same pattern as glutamate, with a slight accumulation under low temperature stress, but decreased under drought and salt stress (Figure S6). Asparagine has a high nitrogen to carbon ratio and acts as an important amino acid for long-distance nitrogen transport in plants [47]. It has been found that rice accumulated a large amount of asparagine under drought and salt stress [48]. Exogenous asparagine as a single nitrogen source could greatly improve nitrogen-deficient growth and promote growth in poplar [49]. Our study also found that the accumulation of asparagine in L. gmelinii seedlings was most significant under salt treatment (Figure S6). Isoleucine and leucine are important branched chain amino acids that can be converted into each other, the branched chain amino acids in Lathyrus sylvestris increased significantly under drought stress; the changes in isoleucine, valine, and leucine in tomato leaves were most significant after drought stress [50]. Exogenous leucine improved the efficiency of plant nitrogen metabolism through reduction of ROS and regulation of the osmotic system, mitigating the damage caused by metal stress to peach seedlings [51]. However, exogenous application of branched-chain amino acids was not significant in improving low-temperature stress tolerance in maize [35]. Here we found that both isoleucine and leucine increased most significantly under drought stress (Figure S6). In brief, the accumulation of proline and GABA were important osmo-regulators to adapt to 3 abiotic stresses. The accumulation of branched chain amino acids isoleucine and leucine were most significant under drought stress. Low temperature stress is significantly different from drought and salt stress, with aspartate and glutamate only accumulating slightly under low temperature stress. Under salt stress, the accumulation of glutamine, arginine, and asparagine, the three amino acids that store and transport nitrogen, was most significant. Exogenous glutamine, arginine, and asparagine can improve the symptoms of nitrogen deficiency in plants [36, 37]. Therefore, it is speculated that the ability of nitrogen transport increases under salt stress of L. gmelinii seedlings. Changes in secondary metabolism As secondary metabolites, flavonoids play key roles in many processes of plant growth and development, such as color formation, stress resistance, etc., and the accumulation of flavonoids enhanced plant stress tolerance [52]. It has been found that taxifolin was increased in callus culture of Taxus baccata under drought stress [53]. The catechins in tea could remove excess ROS during drought [54]. Exogenous catechins improved salt tolerance of Capsicum annuum seedlings [55]. The application of genistein under salt stress improved the yield of soybean [56]. In this study, the accumulation of taxifolin and genistein were increased under drought stress, and catechin was increased under salt stress (Figure S7). Anthocyanins accumulate in the leaves, flowers, fruits and other tissues and organs of plants showing different colors from orange to deep red, and are the end products of flavonoid metabolism, they usually located downstream of phenylpropane metabolism [57]. Anthocyanins play a positive role in plant resistance to external low temperature stress, and low temperature promotes anthocyanin synthesis. It has been found that red mango fruits were more cold tolerant, and red fruits with high anthocyanin and flavonoid content were more tolerant to biotic and abiotic stresses [58]. Anthocyanins play a considerable role in cold stress tolerance of tartary buckwheat sprout [59]. In blood orange, storage at 4 °C significantly induced anthocyanin accumulation in the pulp [60]. Low temperature promoted anthocyanin biosynthesis in grape pericarp [61]. When plants are exposed to external low temperature stress, genes related to anthocyanin metabolism are changed accordingly in plants. ANS and DFR in B. rapa showed a response to low temperature and freezing stress, promoting anthocyanin accumulation and cold resistance [62]. In grapes, low temperature promoted the up-regulating expression of anthocyanin biosynthetic genes F3H, PAL and CHS, thereby promoting anthocyanin synthesis [63]. In woody plants, continuous low temperature treatment (4 ℃ for 20 h) up-regulated the expression of flavonoids biosynthesis genes and the accumulation of flavonoids and anthocyanins in Hevea brasiliensis [64]. Camellia sinensis is a low temperature resistant tree species, but when placed at 10 ℃/8 ℃ for 2 days, its leaves color changed from green to reddish brown with curly edges; and after being placed at 4 ℃/2 ℃ for 2 days, the leaves turned dark brown and the leaf tips withered, accompanied by a significant increase in the total anthocyanin content in the leaves [65]. In this study, we also found that anthocyanins accumulated significantly in L. gmelinii seedlings under low temperature stress, causing the trunk and needles to appear reddish (Fig. 1), and cyanidin 3-glucoside cation and peonidin 3-galactoside cation were the major components of anthocyanins in L. gmelinii seedlings (Figure S7). Furthermore, the anthocyanin synthesis-related genes, ANS and ANR, were also significantly up-regulated under low temperature stress. It is therefore hypothesized that low temperature induced the accumulation of anthocyanins in L. gmelinii and the significant accumulation of anthocyanins also improved the cold tolerance of L. gmelinii seedlings. Lignin is the main component of the cell wall in plants, and lignin deposition is a common response in plants when encountering abiotic stresses [66], CCR and CAD are the main reductases in the lignin biosynthesis pathway. Studies have shown that CCR was induced by drought stress, and high expression of CAD enhanced drought tolerance through lignin biosynthesis [67, 68]. Ferulic acid can be combined with cell wall carbohydrates to adapt to drought conditions in response to a lack of water in leaves; and drought-resistant varieties of winter triticale have significantly higher ferulic acid content than the sensitive varieties [69, 70]. Cinnamic acid, caffeic acid, gallic acid and ferulic acid accumulated in plants under saline conditions [71]. Lignin accumulation in cell wall under cold stress helped to enhance the resistance of cold stress [72]. In this study, CCR were significantly up-regulated under drought and salt stress, and other lignin metabolism-related genes such as 4CL, CYP73A, COMT and CAD were also up-regulated under 3 abiotic stresses (Fig. 9). Among them, 4-Hydroxycinnamic acid accumulated most significantly at S30; ferulic acid showed some increase under drought and low temperature stress; matairesinol increased under drought and salt stress (Figure S6). Thus, lignin metabolism is involved in abiotic stress response of L. gmelinii seedlings, especially to drought and salt stress. Resveratrol, as a plant antitoxin, is a beneficial astragalloid compound that plays an important role in the response of plants to adversity stresses. Resveratrol in plants is mainly synthesized through the phenylpropane pathway [73]. It was found that resveratrol was accumulated under oxidative burst caused by several abiotic stresses, especially salinity and UV radiation, which have similar effects on enhancing biosynthesis of non-enzymatic antioxidants [74]. We found a significant increase in resveratrol under drought and salt stress (Figure S6). It is hypothesized that changes in resveratrol metabolism are mainly involved in drought and salt stress responses of L. gmelinii seedlings. Conclusions Taken together, we identified DEGs and DAMs involved in primary and secondary metabolism of L. gmelinii seedlings responding to abiotic stresses. A regulatory network has been constructed: low temperature stress caused significantly changes compared with drought and salt stress; carbohydrates were significantly accumulated under low temperature stress; amino acids showed the most significant changes under salt stress; variation pattern of secondary metabolism was similar under both drought and salt stress, while anthocyanin accumulation was the most obvious only under low temperature stress (Fig. 10). Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1: Fig. S1: The height of L. gmelinii seedlings under different stress treatments and normal conditions. Fig. S2: Statistics on the number of DEGs in L. gmelinii seedlings under different stress treatments. Fig. S3: RT-qPCR verifies the expression patterns of 27 DEGs in L. gmelinii seedlings under different stress treatments. Fig. S4: OPLS-DA score map of the metabolites of L. gmelinii seedling samples under 3 stress treatments. Fig. S5: Variation patterns of main carbohydrates in L. gmelinii seedlings under different stress treatments. * representing P < 0.05 and * * representing P < 0.01. Fig. S6: Variation patterns of main amino acids in L. gmelinii seedlings under different stress treatments. * representing P < 0.05 and * * representing P < 0.01. Fig. S7: Change patterns of DAMs in secondary metabolism in L. gmelinii seedlings under different stress treatments. * representing P < 0.05 and * * representing P < 0.01. Table S1: The sequence of primers used in the second part. Table S2: Quality of the sequencing results of the seedling samples of L. gmelinii under different treatments. Table S3: Unigenes annotation of the seedling samples of L. gmelinii under different treatments. Supplementary Material 1: Fig. S1: The height of L. gmelinii seedlings under different stress treatments and normal conditions. Fig. S2: Statistics on the number of DEGs in L. gmelinii seedlings under different stress treatments. Fig. S3: RT-qPCR verifies the expression patterns of 27 DEGs in L. gmelinii seedlings under different stress treatments. Fig. S4: OPLS-DA score map of the metabolites of L. gmelinii seedling samples under 3 stress treatments. Fig. S5: Variation patterns of main carbohydrates in L. gmelinii seedlings under different stress treatments. * representing P < 0.05 and * * representing P < 0.01. Fig. S6: Variation patterns of main amino acids in L. gmelinii seedlings under different stress treatments. * representing P < 0.05 and * * representing P < 0.01. Fig. S7: Change patterns of DAMs in secondary metabolism in L. gmelinii seedlings under different stress treatments. * representing P < 0.05 and * * representing P < 0.01. Table S1: The sequence of primers used in the second part. Table S2: Quality of the sequencing results of the seedling samples of L. gmelinii under different treatments. Table S3: Unigenes annotation of the seedling samples of L. gmelinii under different treatments.