Phytohormones regulate the non-redundant response of ω-3 fatty acid desaturases to low temperatures in Chorispora bungeana

To explore the contributions of ω-3 fatty acid desaturases (FADs) to cold stress response in a special cryophyte, Chorispora bungeana, two plastidial ω-3 desaturase genes (CbFAD7, CbFAD8) were cloned and verified in an Arabidopsis fad7fad8 mutant, before being compared with the microsomal ω-3 desaturase gene (CbFAD3). Though these genes were expressed in all tested tissues of C. bungeana, CbFAD7 and CbFAD8 have the highest expression in leaves, while CbFAD3 was mostly expressed in suspension-cultured cells. Low temperatures resulted in significant increases in trienoic fatty acids (TAs), corresponding to the cooperation of CbFAD3 and CbFAD8 in cultured cells, and the coordination of CbFAD7 and CbFAD8 in leaves. Furthermore, the cold induction of CbFAD8 in the two systems were increased with decreasing temperature and independently contributed to TAs accumulation at subfreezing temperature. A series of experiments revealed that jasmonie acid and brassinosteroids participated in the cold-responsive expression of ω-3 CbFAD genes in both C. bungeana cells and leaves, while the phytohormone regulation in leaves was complex with the participation of abscisic acid and gibberellin. These results point to the hormone-regulated non-redundant contributions of ω-3 CbFADs to maintain appropriate level of TAs under low temperatures, which help C. bungeana survive in cold environments.

Using the targetP prediction tool, two chloroplast targeting peptide of 50 and 60 aa were found in the N-terminal of the deduced CbFAD7 and CbFAD8 (Fig. 1a), respectively, predicting the subcellular localization of the proteins in chloroplasts. Amino acid alignment (Fig. 1a) showed that both of the deduced proteins contain three conserved histidine clusters (HDGCH, HXXXXXHRTHH and HHXXXXHVIHH) and four transmembrane domains (TMD), suggesting that they are chloroplast membrane-bound ω-3 FADs. The phylogenetic analysis (Fig. 1b) displayed that CbFAD7 (ARL62096) and CbFAD8 (ARL62097) were positioned in the group corresponding to plastidial ω-3 FADs, providing further evidence that CbFAD7 and CbFAD8 encode plastidial ω-3 FADs.
The functionality of CbFAD7 and CbFAD8 were verified in Arabidopsis mutant. To verify the functionality of CbFAD7 and CbFAD8, the ORF of the two genes were expressed in double fad7fad8 mutants under the CaMV 35S promoter of pBI121 vector, respectively. The fatty acids of leaf lipids showed that though the C18:3 contents in the complemented mutants F7 and F8 were still lower than that in WT plants, they were markedly higher than that in fad7fad8 mutants (Fig. 2a). Being exposed to 15 °C, the germination rates of F7 and F8 seeds were significantly higher than that of fad7fad8 mutant seeds, and close to that of WT seeds (Fig. 2b). These data confirmed that CbFAD7 and CbFAD8 were functional plastical ω-3 FAD genes.
The tissue-specific and cold-responsive expressions of ω-3 CbFAD genes in C. bungeana. The expression profiles of plastical and microsomal ω-3 CbFAD genes were analyzed in the suspension-cultured cells and the regenerated plants of C. bungeana (Fig. 3a). CbFAD7 and CbFAD8 have the highest expression in leaves and the lowest expression in roots, showing the characteristic of plastidial ω-3 FAD genes. CbFAD3 were mostly expressed in suspension-cultured cells, and lowly expressed in stems, exhibiting the feature of microsomal ω-3 FAD genes.
Considering the highest expression of microsomal and plastidial ω-3 CbFAD genes, the cold-responsive expressions of them were detected in suspension-cultured cells and the leaves from regenerated plants (Fig. 3b). In suspension-cultured cells, the expression of CbFAD3 was increased at 4 (6.3-fold) and 0 °C (5.7-fold), while that of CbFAD8 was increased at 0 (13.1-fold) and − 4 °C (27.7-fold). The increases in CbFAD3 and CbFAD8 mRNA all peaked at being treated for 3 h, and were accompanied by a decrease in CbFAD7 mRNA at different low temperatures. In C. bungeana leaves, the increase in CbFAD7 mRNA was found at 4 (4.1-fold) and 0 °C (2.6-fold), and the induction of CbFAD8 expression also occurred at 0 (3.4-fold) and − 4 °C (6.7-fold), like that found in cultured cells. Moreover, the cold-induced expression of CbFAD7 and CbFAD8 peaked at being treated for 3 and www.nature.com/scientificreports/ 6 h, respectively, which were accompanied by a decrease in CbFAD3 mRNA at tested low temperatures. In summary, the expression of ω-3 CbFAD genes presented a non-redundant pattern in response to low temperatures.
The hormone-and inhibitor-responsive expressions of ω-3 CbFAD genes in C. bungeana. To be in line with the cold-responsive experiments, the hormone-responsive experiments on ω-3 CbFAD genes were also studied in suspension-cultured cells and the leaves from regenerated plants. Data (Figs. 3b and 4) showed that though each of the tested hormones could affect the expression of ω-3 CbFAD genes, only the changes caused by certain hormones were similar to those induced by low temperatures, considering change trend and peak time. In respect to cell ω-3 CbFAD genes, the similar changes were all brought by JA and BRs. However, as for leaf ω-3 CbFAD genes, the regulation were more complex: the corresponding down-regulation of CbFAD3 expression were caused by JA, BRs, ABA and GA3; the comparable increases in CbFAD7 and CbFAD8 mRNA were induced by BRs and GA3 (1.6-and 1.9-fold, peaked at 3 h) as well as JA and ABA (4.3-and 3.6-fold, peaked at 6 h), respectively. The Pearson correlations between the cold-and the hormone-responsive expressions of ω-3 CbFAD genes (Tables 1 and 2) indicated that JA and BRs may participate in the low-temperature-responsive expressions of these genes in both suspension-cultured cells and plant leaves, while ABA and GA3 may only take part in the low-temperature-induced regulation in leaves.
To further confirm the participation of these hormones, the corresponding inhibitors were used. As showed by Fig. 5, in suspension-cultured cells, the chilling-induced increase in CbFAD3 mRNA was totally inhibited by Pcz (a synthetic inhibitor of BRs), but the inhibition could be partly relieved by DIECA (a synthetic inhibitor of JA); the cold-inhibited expression of CbFAD7 was partially eliminated by Pcz (32.9%) or DIECA (65.8%), and was completely eliminated by the combination of them; conversely, the cold-induced expression of CbFAD8 was mostly inhibited by Pcz (78.9%) or DIECA (79.9%), and was entirely inhibited by the cooperation of them. In C. bungeana leaves, the decrease in CbFAD3 mRNA caused by low temperatures was absolutely eliminated by the synergism of Pcz, DIECA, Pac (a synthetic inhibitor of GA3) and Flu (a synthetic inhibitor of ABA), and the synergistic effect of Pcz and DIECA play a major role (79.0%); Although the chilling induction of CbFAD7 expression could be suppressed by Pcz (66.0%) or Pac (33.9%) to some extent, the complete suppression need the joint action of them; likewise, the cold induction of CbFAD8 expression was incompletely inhibited by DIECA (73.1%) or Flu (29.9%), but was thoroughly inhibited by the combined effect of them.
Altogether, these results suggested that JA and BRs may result in the opposite changes in cell CbFAD3 expression, while the synergism of them may bring cell CbFAD7 inhibition and cell CbFAD8 induction, except the high level of BRs ( Supplementary Fig. S2). In C. bungeana leaves, JA and BRs may lead to decrease in CbFAD3 mRNA with the help of ABA and GA3, and the combined effect of BRs and GA3 may active the chilling-responsive induction of CbFAD7 expression, while the joint action of JA and ABA may trigger the cold-responsive induction of CbFAD8 expression.
The level of related phytohormones in C. bungeana during low-temperature treatments. To provide more evidence to the hormone-regulated cold response, the level of related phytohormones were www.nature.com/scientificreports/ detected in suspension-cultured cells and the leaves from regenerated plants at different low temperatures. Being exposed to 0 °C, the level of tested pyhtohormones presented a rapid and two-peaks increase in both analysis systems, however, the peak time and the peak value of them were various (Figs. 6a and 7a). In cultured cells, the level of BRs peaked at being treated for 1 (1.7-fold) and 3 h (1.5-fold), while that of JA peaked at being treated for 1.5 (2.6-fold) and 3 h (2.4-fold). Though the accumulations of BRs and JA in leaves were similar to those in cultured cells, the peak time of JA (2 and 4 h) was a little later. The change trends of GA3 (1.8-and 2.4-fold) and ABA (1.6-and 2.0-fold) resembled each other, but the peak time of GA3 (0.5 and 2.5 h) was half an hour earlier than that of ABA. Overall, the phytohormone increases induced by low temperatures, notably the first peak, preceded the temperature-responsive expression changes in corresponding ω-3 CbFAD genes; furthermore, the level of synergistic hormones, such as JA and ABA, or BRs and GA3, reached the peak value at staggered times to avoid redundant effect. Although the phytohormone increases could be induced by different low temperatures, the increments were varied with temperature (Figs. 6b and 7b). In suspension-cultured cells, the highest level of BRs and JA appeared at 4 (2.6-fold) and − 4 °C (3.1-fold), respectively. In C. bungeana leaves, the increased level of BRs was decreased with decreasing temperature, while that of JA was increased; meanwhile, the increase in the level of GA3 and ABA at 0 and − 4 °C presented an opposite trend. Together, the hormone increase were consistent with the dynamic expressions of corresponding ω-3 CbFAD genes at different low temperatures; moreover, the increased level of antagonistic hormones, for example GA3 and ABA, showed a reverse trend during temperature variation, which may be due to the trade-offs between plant growth and cold stress response.  www.nature.com/scientificreports/ To further confirm that the phytohormones affect the accumulation of TAs through gene regulation, the hormone-inhibitor treatments were performed at 0 °C (Fig. 8). In suspension-cultured cells, the cold-induced accumulation of C18:3 was completely inhibited by the synergism of DIECA and Pcz, but partially inhibited by either of them. In C. bungeana leaves, the cold-induced accumulation of C18:3 and C16:3 were both totally suppressed by the synergistic effect of DIECA, Pcz, Pac and Flu, but partly suppressed by either of them. These results were in accord with the inhibitor-responsive expression of ω-3 CbFAD genes, and verified that with or without the help of ABA and GA3 in distinct tissues, JA and BRs participated in maintaining appropriate level of TAs in C. bungeana through regulating ω-3 CbFAD genes in response to low temperatures.

Discussion
This work analyzed the behavior of plastidial and microsomal ω-3 FADs in C. bungeana at the level of gene expression and fatty acid content, to determine how low temperature affected the contribution of ω-3 FADs to the synthesis of TAs, especially C18:3. From the result, it can be seen that low temperatures resulted in significant transcriptional changes on each ω-3 CbFAD. Among these changes, the sharply increased CbFAD8 mRNA induced only by severe low temperatures (notably the subzero temperature) was observed both in the suspension-cultured cells and in the leaves from regenerated plants (Fig. 3b). Considering the difference between the two analysis systems, this common phenomenon shows the critical role of CbFAD8 in the synthesis of TAs under freezing and subfreezing temperatures, which is different with the way most plant species do. In most plant species [12][13][14][15]17,18,20,42 , the transcript level of FAD8 was increased in response to chilling temperatures ranging from 4 to 15 °C. FAD8 was originally identified as a cold-specific desaturase by phenotypic analysis of a fad3fad7 double mutant from Arabidopsis 43 that was capable of producing TAs only at low temperatures (15 °C). Given that FAD8 plays an important role in the biosynthesis of plastid TAs 44,45 required for the correct biogenesis and maintenance of chloroplasts 46 as well as the recovery from photo inhibition at low temperatures 47,48 , the induction of FAD8 expression shows a common choice of plants, namely protecting chloroplasts, particularly photosynthesis. In C. bungeana, the induction of CbFAD8 occurred at or below 0 °C may be due to its original membrane lipid unsaturation of chloroplasts brought by FAD8 is higher than that in most other plants, which can help the chloroplasts to get through chilling temperatures. This may also explain why the cold resistance of C. bungeana is much higher than that of normal plants. Of course, it still needs further researches.
In response to chilling temperatures, the expression of CbFAD3 and CbFAD7 presented a tissue-specific profile (Fig. 3b). The CbFAD3 mRNA was increased in suspension-cultured cells and decreased in plant leaves, while the expression profile of CbFAD7 was just the opposite. Combined with previous studies, we found that Table 1. Pearson correlation coefficients (two-tailed) between the cold-responsive and the hormone-regulated expressions of ω-3 CbFAD genes in C. bungeana cells. Three replications for each sample, n = 5 × 3. Significant positive correlation (R > 0.50, P < 0.05) is indicated in bold. *P < 0.05, **P < 0.01, ***P < 0.001. www.nature.com/scientificreports/ even in the same tissue under similar temperature conditions, the low-temperature-induced expression of FAD3 and FAD7 were varied with plant species 7,[12][13][14]16,17,[49][50][51] , which means some observations are consistent with ours but the others are not. Although we don't know why the contributions of microsomal and plastidial desaturases differ among plant species, it is not surprising that the contributions of CbFAD3 and CbFAD7 in C. bungeana were proportional to their transcript abundance in corresponding tissues (Fig. 3). Furthermore, this expression profile suggests that CbFAD3 would be more important in C. bungeana cultured cells while CbFAD7 might contribute to TAs production in C. bungeana leaves under chilling temperatures. When the temperature was further decreased, the increase in CbFAD3 or CbFAD7 mRNA was reduced and even turned into a decrease at -4 °C, which formed a non-redundant complementation with the gradual increase in CbFAD8 mRNA (Fig. 3b). The coordination of CbFAD3 and CbFAD8 indicates the distinct TA needs of C. bungeana cells at different low temperatures, for FAD3 affecting total TA levels and FAD8 affecting plastid TA www.nature.com/scientificreports/ levels 44,45 . The cooperation of FAD7 and FAD8 was not only found in C. bungeana leaves, but also found in Arabidopsis leaves 18 at different temperatures (8, 22 and 30 °C). Although both FAD7 and FAD8 affect the synthesis of TAs in chloroplastic lipids, the lipid specificity of them are not the same 27 : AtFAD7 prefers galactolipids, which are the major chloroplast lipids with higher TA content; AtFAD8 likes phosphatidylglycerol 18 , which has a specific role in the stability of photosynthetic complexes 52 . Hence, the trade-off between CbFAD7 and CbFAD8 in leaves reflects the strategy to maintain photosynthetic activity and stability during temperature variation. Nevertheless, the non-redundant expression of ω-3 CbFAD genes maintained appropriate level of TAs, especially C18:3, in response to low temperatures (Fig. 8a). Through a series of verification, including exogenous hormone application (Fig. 4), correlation analysis (Tables 1 and 2), inhibitor treatments (Figs. 5 and 8) and phytohormone detection (Figs. 6 and 7), we found that the low-temperature-responsive expression of ω-3 CbFAD genes were regulated by certain phytohormones, notably JA and BRs. JA was mainly responsible for FAD8 induction under freezing/subfreezing temperatures, with the tissue-specific assistance of BRs or ABA; BRs was in charge of the induction of CbFAD3 or CbFAD7 expression at chilling temperatures, with or without the help of GA3 in distinct tissues. Moreover, both JA and BRs took part in the inhibition of corresponding ω-3 CbFAD expression in response to low temperatures, with or without the participation of ABA and GA3 according to tissue specificity. These data not only provide new insights into our previous findings 17 that JA partially mediates the chilling-induced transcription of ω-3 FAD genes in Arabidopsis, but also agree with that JA may act as a core signal by interacting with other phytohormones to regulate the balance between plant growth and stress response 53 . As we know ( Supplemetary Fig. S3), ABA may interact synergistically with JA signaling to regulate the expression of cold-responsive genes [54][55][56] ; BRs acts in synergism or antagonism with JA, depending on BRs' concentration in response to stress 53 ; GA usually inhibits JA signaling, but in some cases, the synergistic effect of them is also exist 53,54,56 . In this study, the interactions between JA and the other tested hormones are in line with the reported findings. Notably, the joint action of JA and ABA regulated the cold-induced expression of leaf CbFAD8, and the combined effect of JA and BRs triggered the induction of CbFAD8 expression in cultured cells. Our previous study in Arabidopsis 17 , which confirmed the participation of JA in the chilling-induced expression of ω-3 FAD genes, also implied the existence of JA's partner. Besides that, a recent study reported that in Arabidopsis leaves 27 , AtFAD8 protein levels were increased upon cold or JA exposure, but did not respond to ABA. All these data prove the fact that though the partner of JA signaling may www.nature.com/scientificreports/ vary with tissues and plant species, JA did regulate the low temperature induction of FAD8 expression in both C. bungeana (cryophyte) and Arabidopsis (modal plant), which may be common in most plant species. In C. bungeana, BRs is another important pyhtohormone involved in the low-temperature regulation of ω-3 FAD genes. It is confirmed that BRs can improve frost tolerance by promoting GA biosynthesis and www.nature.com/scientificreports/ interplaying with GA at the signaling level 57 ( Supplemetary Fig. S3), which supports the observation that the low temperature response of CbFAD3 and CbFAD7 in C. bungeana leaves was mediated by the cooperation of BRs and GA3 (Fig. 5b). Recent studies also demonstrated that BRs may participate in drought or cold stress acclimation by three interconnected mechanisms, one of which is in communication with ABA signaling 58 ( Supplemetary  Fig. S3). This may explain the synergistic effect of BRs and ABA on the down-regulation of leaf CbFAD3 expression during low temperature exposure (Fig. 5b). To date, there is no report about the participation of BRs in the expression of ω-3 FAD genes in response to low temperatures, so we cannot predict that BRs also regulates the low temperature induction of FAD3 or FAD7 expression in other plant species before further research. It is worth noting that the interaction between different phytohormones demonstrates that each ω-3 CbFAD gene can respond to multiple hormones (Figs. 4, 5, 6, 7 and 8; Tables 1 and 2). A similar phenomenon was found in Arabidopsis, which showed the expression of AtFAD3 was regulated through the synergistic and antagonistic interactions of auxin, cytokinin and ABA during plant development 59 . These results reflect the existence of www.nature.com/scientificreports/ various promoter cis-elements combined with different transcription factors (TFs) from corresponding hormone signaling. A G-box-like motif required by JA-responsive expression was first found in the promoter of AtFAD7 from Arabidopsis roots 60 . After that, the SA-and ABA-responsive elements were found in the promoter of cabbage FAD8 61 . Recently, the analysis on AmFAD7 and AmFAD8 in Ammopiptanthus mongolicus confirmed that both of the promoters contain the elements for the response to ABA, JA, GA and MeJA 62 . Moreover, the multiple ABA-responsive elements were found in the promoter of microalgae ω-3 FAD genes 63 . All the findings support the idea that ω-3 FAD genes can be directly regulated by more than one hormone. As another necessary factor for the hormone regulation, TFs have been studied at recent years. On the one hand, some TFs related to ω-3 FAD expression were verified: for example, the expression of FAD3 was up-regulated by bZIP67 in Arabidopsis seeds 64 , but down-regulated by MaMYB4 in banana fruits 65 or by HD in soybean 66 ; WIPK was involved in wound-responsive expression of AtFAD7 gene in transgenic tobaccos 67 . On the other hand, it is confirmed in Arabidopsis that some TFs (such as WRKYs, MYCs, bHLHs and ICEs) occurred at the downstream of cold-induced JA signaling 54,56 , while some others (for example bZIPs) belong to the cold-responsive BR and/or ABA signaling pathways 57,68 . Additionally, a recent study in a mutant of Pyrus bretschneideri Rehd found that MYB1R1 and MYC2 regulate ω-3 FADs involved in ABA-mediated suberization 69 .
Though all the relevant information are not quite complete, if we piece them together, it is not difficult to speculate that through the common or distinct TFs combined with corresponding promoter elements, JA and BRs as well as ABA and GA3 achieve the synergistic or antagonistic regulation on ω-3 CbFAD genes, which result in the non-redundant cooperation on maintaining appropriate level of TAs and then help C. bungeana survive in cold environments. These results may provide valuable information to agricultural production, for example, multi-hormone application may help crops overcome the influence of low temperatures in the future.

Methods
Plant material and experiment treatments. The suspension-cultured cells and the regenerated plants of C. bungeana were prepared as described by Shi et al. 3 and Fu et al. 30 , respectively. Photomixotrophic cultured cells were initiated from wild C. bungeana leaves, while regenerated plants were originated from C. bungeana cotyledons. Though both of them were propagated under 25 °C with 12 h illuminations (1000 and 4000LX, respectively), the former were germinated in liquid modified MS medium (0.2 mg/L of 2,4-dichlorophenoxyacetic acid, 1-naphthaleneacetic acid, 6-benzylaminopurine and kinetin, respectively), and the latter were grown on gel modified MS medium (0.4 mg/L of gibberellin, 0.6 mg/L of kinetin and 3% glucose instead of sucrose). Regenerated plants having 3-5 cm long roots were used for experiments. Arabidopsis seeds of Col-0 (WT), fad7fad8 mutant (N8036, NASC, UK), and complemented fad7fad8 mutants (F7 and F8) were cultivated as our previous procedure 17 .
For cold treatments, C. bungeana cell suspensions and regenerated plants were exposed to 4, 0 or − 4 °C for 24 h, respectively. For low-temperature germination, Arabidopsis seeds planted on MS medium were exposed to 15 °C for 4 weeks. For exogenous hormone treatments, the cell suspensions and regenerated plants were moved to 1/2 MS medium with 100 µM JA, 0.5 µM BRs, 100 µM SA, 100 µM ABA, or 100 µM GA 3 for 24 h, respectively. For phytohormone inhibitions, the cell suspensions and regenerated plants were moved to 1/2 MS medium with 10 µM sodium diethyldithiocarbamate trihydrate (DIECA), 10 µM propiconazole (Pcz), 10 µM fluricbne (Flu), or 10 µM paclobutrazol (Pac) for 24 h, respectively. To ensure the hormone/inhibitor application on leaves, the over ground part of each plant was sprayed with corresponding solution after moving, and the excess liquid was blotted by cotton balls. The C. bungeana cells and leaves were collected at different time spots according to the experimental design in each treatment, while the leaves of different Arabidopsis lines were collected from 5 weeks old plants under normal growth conditions. All the collected samples were stored at − 80 °C until use.
RNA isolation and cDNA synthesis. Total RNAs of C. bungeana were isolated from suspension-cultured cells or the different tissues of regenerated plants (0.1 g each) according to the standard procedure of Plant RNA Kit (Omega, USA). After being treated with RQ1RNase-free DNase (Promega, USA), the purity and integrity of total RNAs were assessed by UV spectrophotometry and agarose gel electrophoresis. Then, the qualified total RNAs were employed in reverse transcription reaction by using PrimeScript™ Reverse Transcriptase (Takara, Japan) and Oligo(dT) 15 primer (Takara, Japan) following the manufacturer's instruction. Reverse-transcribed cDNA samples were stored at -20 °C until further use.

Cloning and bioinformatics analysis of CbFAD7 and CbFAD8. A 732-bp fragment of CbFAD7
and a 982-bp fragment of CbFAD8 were cloned from C. bungeana by using degenerate primers (P1 and P2 for CbFAD7; P3 and P4 for CbFAD8; Supplementary Table S1) designed basing on the conserved domain database from tobacco, Brassica napus and Arabidopsis. Amplification of 5' and 3' ends of CbFAD7 and CbFAD8 were accomplished using specific primers (P5-P8 for CbFAD7; P9-P12 for CbFAD8; Supplementary Table S1) and SMARTer™ RACE cDNA Amplification Kit (Clontech, Japan). Full-length cDNA of CbFAD7 and CbFAD8 were gotten by using specific primers (P13 and P14 for CbFAD7; P15 and P16 for CbFAD8; Supplementary Table S1). PCR products were cloned into the pMD-18T vector (Takara, Japan) and sequenced by GENEWIZ Inc. (Suzhou, China). Then these sequences were analyzed by DNAman 5.2.9, MEGA 6.06 and ClustalX 1.83 software. The prediction of transmembrane domain and transit peptide were analyzed by the online server program TMHMM (http:// www. cbs. dtu. dk/ servi ces/ TMHMM/) and TargetP (http:// www. cbs. dtu. dk/ servi ces/ Targe tP/), respectively. www.nature.com/scientificreports/ of the binary vector pBI121 to replace the GUS gene and construct the recombinant plasmid, pBI121-CbFAD7, under the control of CaMV 35S promoter. In the same way, the recombinant plasmid pBI121-CbFAD8 was constructed with the coding region of CbFAD8 amplified using specific primers (P19 and P20; Supplementary Table S1). Then, the two recombinant plasmids and the Arabidopsis seeds of fad7fad8 mutant were sent to Shanghai Weidi Biotechnology Co., Ltd (Shanghai, China) to get the complemented mutants, F7 and F8. The transgenic Arabidopsis plants were generated through the floral dip by Agrobacterium-mediated transformation. Positive transgenic lines of F7 and F8 exhibiting 3:1 segregation ratio were identified by PCR using specific primers (P17 and P18 for CbFAD7; P19 and P20 for CbFAD8; Supplementary Table S1 and Fig. S1). The homozygous lines were obtained by backcrossing or self-pollination, and the expression level of CbFAD7 and CbFAD8 were verified by qRT-PCR using specific primers (P23 and P24 for CbFAD7; P25 and P26 for CbFAD8; Supplementary  Table S1 and Fig. S1), respectively. IPT2 (AT2G27760) was taken as the house keeping gene 70  Extraction and analysis of total fatty acids. C. bungeana cell suspensions and leaves as well as Arabidopsis leaves (2 g each) were grinded with liquid N2, respectively. The lipids and the total fatty acids of each sample were prepared and analyzed as our previous reports 3, 17 . The fatty acid methyl esters of each sample were analyzed by GC-MS (6890N-5975C, Agilent, USA) fitted with a capillary column (Agilent DB-FFAP, 30 m × 0.25 mm × 0.5 µm) according to our previous procedure 70 . Fatty acid data were obtained from five independent biological samples for each experiment.

Complementation of
Extraction and analysis of phytohormones. C. bungeana suspension-cultured cells and leaves (0.5 g each) were sent to Genepioneer Biotechnologies (Nanjing, China) to detect the concentration of JA, BRs, ABA and GA3, respectively. Phytohormones were extracted through grinding and organic solvent extraction, and then analyzed using Plant JA ELISA Kit (Sinobestbio, China), Plant BRs ELISA Kit (Sinobestbio, China), Plant ABA ELISA Kit (Sinobestbio, China) and Plant GA3 ELISA Kit (Sinobestbio, China), respectively. During the double-antibody sandwich ELISA, the absorbance (OD value) was detected by microplate reader (Infinite F50, Tecan, SWIT), and the concentration was calculated through standard curve method. Phytohormone data were from five biological replicates for each experiment.
Statistical analysis. Statistical analysis was performed using the method of one-way ANOVA, followed by Duncan's multiple range test at the P < 0.05 or P < 0.01 levels.
Ethical approval. All procedures of this study, including the experimental research and field studies on C. bungeana and Arabidopsis as well as the collection of plant material, were conducted in accordance with the relevant institutional, national, and international guidelines and legislation. The wild C. bungeana plants were collected from an ice free cirque besides the Glacier No. 1 in Tianshan mountains (Xinjiang province, China), which is not a restricted area for researchers.

Data availability
The DNA and protein sequences generated during and/or analysed during the current study are available in the NCBI Genbank repository www.nature.com/scientificreports/