Background

Cyanogenic glycosides (CNGs) are natural defense compounds identified in several plant species of different families, including economically significant fruit trees and crops (such as cassava, barley, apple, cherry, apricot, almond and wheat) [1,2,3,4,5,6]. Yet, the toxic hydrogen cyanide (HCN) is emitted from foods containing CNGs while being digested, and long-term consumption of edible plants with high level of CNGs may has a negative effect on human with adverse symptoms (such as headache, weakness, hyperventilation, vomiting, and respiratory system failure) at sub-acute cyanide dose, and even death at acute lethal dose [4, 7]. Hence, the cyanide poisoning in human has become a severe health problem in world. Several processing methods (such as peeling, fermenting, bacteria degradation, drying, boiling, soaking, and crushing) have been applied to reduce toxicity to safe level [4, 8], but all of which required a lot of time and cost, and caused serious environment pollution through release of dangerous cyanide gases or effluent water from processing factories [9,10,11,12], and thereby it is often impossible to remove all the CNGs through conventional processing. Yet, genetic engineering approach may offer an alternative method to reduce plant CNG content.

Siberian apricot (Prunus sibirica L.), an important fruit of Rosaceae family, is widely distributed in the mountainous areas of northeastern and northern of China, maritime territory of Russia, eastern Siberia regions, and eastern and southeastern parts of Mongolia [13]. P. sibirica is extensively researched in China owing to its good adaptability, rich resource, and significant economic and ecological importance [13,14,15]. The total area was approximately 1.7 million hectares, yielding around 192,500 metric tons of the seeds each year [14,15,14, 18, 20], which has been applied in various fields including medicinal product, cosmetic, surfactant, edible oil and biodiesel feedstock [14,15,16,17, 21]. Also, the seeds show a potential usefulness in human nutrition and traditional Chinese medicine [22,23,24,25,26]. However, the P. sibirica seeds contain phenylalanine-derived amygdalin and its hydrolysis can release the HCN, causing potential toxicity and illness [13, 16], which has intercepted the commercial application of P. sibirica seed oils and related foods. Therefore, how to effectively reduce seed amygdalin accumulation of P. sibirica has become one critical concern.

Amygdalin is one of the most wide-spread groups of CNGs in the Rosaceae family, and its metabolism is generally divided into biosynthesis, bioactivation and detoxification [3, 27]. Of these, amygdalin biosynthesis is delivered from the initially synthesized mandelonitrile from Phe by two cytochrome P450 (CYP) enzymes (CYP79/CYP71), and glucosylated by UDP-glucosyltransferase (UGT1) to produce prunasin, and finally converted into amygdalin by glucosyltransferase [3, 28,29,30]. In bioactivation process, amygdalin is hydrolyzed to form mandelonitrile by amygdalin hydrolase (AH) and prunasin hydrolase (PH) [27, 31,32,33], and then hydrolyzed by mandelonitrile lyase (MDL) to form benzaldehyde with release of HCN [34]. Several plants have evolved various different detoxification mechanisms, with the most effective pathway that is the conversion of HCN into β-cyanoalanine by β-cyanoalanine synthase (β-CAS), and then hydrated by nitrilases (NITs) to generate asparagine or aspartate along with ammonia [2, 27, 35]. All these indicated that mandelonitrile was one key intermediate for both biosynthesis and bioactivation of amygdalin involved in several regulatory enzymes. Another concern was that hydroxynitrile lyases (HNLs), one kind of important enzymes in CNG-rich plants, catalyze the final step in biodegradation process of CNGs in several plants (such as cassava, Arabidopsis, Japanese apricot, peach, almond, and loquat), resulting in the productions of HCN and carbonyl component [36,37,38,39,40,41,42,43,44], but almost of HNLs were identified with asymmetric synthesis of mandelonitrile. Of note, MDL, one group of the HNLs, can catalyze dissociation of mandelonitrile to HCN and benzaldehyde, which was only identified in almond and black cherry [45,46,47,48]. The MDLs in P. sibirica seeds remain enigmatic, and thus the question of which MDL family member required for mandelonitrile dissociation destined to low seed amygdalin accumulation of P. sibirica has become an imperative. Recently, our 454 RNA-seq analysis of P. sibirica (SRX339392) [49] should ensure to effectively annotate the genes encoding MDLs responsible for mandelonitrile cleavage destined for decrease seed amygdalin amount of P. sibirica.

This study was designated with aim to identify crucial MDL and to highlight function in governing seed mandelonitrile catabolism with low amygdalin accumulation. To this end, we conducted a genome-wide identification of mandelonitrile catabolism-related MDLs according to recent 454 transcriptome sequencing result of P. sibirica, and the association of transcriptional level of all annotated MDLs with accumulative amount of amygdalin and mandelonitrile was performed in the seeds across different germplasms and developmental periods as first attempt to identify key MDL for mandelonitrile catabolism of P. sibirica seeds. The resulting MDL2 with transcript abundance was identified, and the focus of following work was to highlight biological function of MDL2 in P. sibirica seeds. In this regard, MDL2 gene was cloned from the seeds of P. sibirica, and then its heterologous expression was performed in Pichia pastoris to identify catalytic property. Finally, an integrated assay of MDL2 overexpression and mdl2 mutant was conducted in Arabidopsis, and a subsequent detection was conducted in plant growth, seed development, mandelonitrile content, oil production, and some critical regulatory enzyme transcription for mandelonitrile catabolism, oil biosynthesis [including fatty acid (FA) biosynthesis and triacylglycerol (TAG) assembly] in different transgenic lines. All these works should help to unravel MDL2-mediate regulation in seed amygdalin and oil biosynthesis of P. sibirica. The obtained results may present potential molecular target (PsMDL2) for future engineering high oil production with lower amygdalin in other oil plants.

Result

Global identification of MDL2 as crucial MDL family member for seed mandelonitrile hydrolysis of P. sibirica

Mandelonitrile is known as one key intermediate of amygdalin metabolism (biosynthesis and dissociation), and the less of mandelonitrile accumulation can contribute to low content of amygdalin in P. sibirica seeds. Thus, one vital challenge is to determine which mandelonitrile lyase (MDL) is responsible for mandelonitrile hydrolysis of P. sibirica seeds. In the present study, our recent 454 RNA-seq assay of P. sibirica (SRX339392) [49] were applied to annotate 5 genes encoding the homologies of MDL1/2/3/4/5 with differential transcriptional levels, in which MDL2 displayed the richest transcript in P. sibirica seeds (Table 1), indicating that MDL2 may be as potential catabolic enzyme for seed mandelonitrile hydrolysis of P. sibirica.

Table 1 The annotated information of genes for MDLs in seeds of Prunus sibirica by 454 deep sequencing analysis
Full size table

To further determine MDL2 as key catabolic enzyme for mandelonitrile dissociation with low amygdalin production in P. sibirica seeds, five superior germplasms of P. Sibirica (accessions Ps-14/23/46/63/70) with high yield of the seeds were chosen as materials, and the comparative association of transcriptional level of annotated 5 MDLs (MDL1/2/3/4/5) by qRT-PCR assay with accumulative level of amygdalin and mandelonitrile was made in the seeds of different germplasms. A significant variation on the contents of mandelonitrile (0.004−0.317%) and amygdalin (0.051−5.314%) was detected in the mature seeds among different accessions. Of note, the maximum values of mandelonitrile (0.317%) and amygdalin (5.314%) were all detected for accession Ps-63, but the minimum value was recorded for Ps-70 (Fig. 1a), implying that high amygdalin accumulation may be related to high amount of mandelonitrile in the seeds across different accessions. Notably, MDL2 transcript level was negatively correlated with the amounts of mandelonitrile and amygdalin in mature seeds of all accessions (Fig. 1a, b), while MDL1/3/4/5 with less transcript exhibited no notable difference across different accessions (Fig. 1b). These results revealed an importance of MDL2 in regulating mandelonitrile and amygdalin accumulation of P. sibirica seeds, coincided with our sequencing result that MDL2 was identified with the most abundant transcription among our all-annotated MDLs (Table 1).

Fig. 1
figure 1

Variabilities on the contents of amygdalin and mandelonitrile and transcriptional level of MDL gene in the seeds of different accessions or develo** stages of Prunus sibirica. (a) Variabilities on the contents of mandelonitrile and amygdalin in mature seeds from 5 different accessions. The result was calculated and expressed as the percentage of dry seed weight (%, g/g). (b) Transcript analysis of our annotated MDLs in mature seeds from different accessions by qRT-PCR. The genes encoding for cyclophilin (CYP) and ubiquitin-conjugating enzyme (UBC) were used as the internal controls, and the expression level of MDL2 in mature seeds of accession Ps-70 with the lowest seed mandelonitrile and amygdalin content was set to 1.00 for standardization. (c) Dynamic content changes of amygdalin and mandelonitrile in develo** seeds of Ps-63 accession. The result was expressed as the percentage of dry seed weight (%, g/g). (d) Relative transcript level of MDL2 in develo** seeds of Ps-63 accession. The genes of CYP and UBC were used as the internal controls, and transcript level of MDL2 in seeds at 10 DAF was set to 1.00 for standardization. All analyses were performed in three biological replicates with three technical repetitions each

Full size image

In order to gain better insight into MDL2 in controlling mandelonitrile and amygdalin accumulation in P. sibirica seeds, we also explored whether the accumulative amounts of mandelonitrile and amygdalin were associated with MDL2 transcriptional level during seed development of P. sibirica. One accession Ps-63 with high amygdalin was selected to analyze dynamic patterns of amygdalin and mandelonitrile contents and MDL2 transcription in the seeds during the whole develo** period from 10 DAF (unmatured stage) to 70 DAF (full ripen stage). A gradual increase in amygdalin content (0.04−5.38%) was detected in develo** seeds with a rapid accumulation at 30–50 DAF (1.75−4.76%), which was the case for the content change of mandelonitrile during seed development (Fig. 1c). However, a seed development-dependently decreased pattern was detected for MDL2 transcript in seeds during development (Fig. 1d), and notably, its transcript profile was temporally associated with the increased amount of amygdalin and mandelonitrile during seed development (Fig. 1c, d). These results indicated that the down-regulated transcript of MDL2 may contribute likely to high accumulation of mandelonitrile and amygdalin in P. sibirica seeds during development. Hence, another vital challenge is to understand how MDL2 is targeted specifically to regulate mandelonitrile hydrolysis destined for low amygdalin accumulation in P. sibirica seeds.

Cloning and functional assay of PsMDL2 gene from the seeds of P. sibirica

In order to unravel the biological role of MDL2 in controlling mandelonitrile and amygdalin biosynthesis in the seeds of P. sibirica, we obtained 2,047-bp complete cDNA of MDL2 (referred as PsMDL2) from the seeds of accession Ps-70 according to our recent 454 RNA-seq result of P. sibirica seeds. This cDNA sequence encodes one protein containing 562 amino acids with a molecular mass of 61.1 kD (Fig. 2a and Additional file 1: Fig. S1). PsMDL2 analysis was involved in analyzing multiple sequence alignment, constructing a phylogenetic tree, identifying conserved motif and predicting 3D protein structure. The results showed that PsMDL2 had a N-terminal signal sequence, 4 N-glycosylation functional motifs (N-X-T/S) (Fig. 2b), and an active-site (His residue) at the position 523 (Fig. 2d), but without transmembrane helical region (Fig. 2c). Also, PsMDL2 had a typical FAD-binding motif (residues 60 to 65, GGGTSG) and conservative sequence (YWHYHGG) (Fig. 2e), and displayed a close correlation with the MDL2 proteins from the Prunus family, including Prunus dulcis (PdMDL2), P. mume (PmMDL2), P. avium (PaMDL2), P. persica (PpMDL2), and P. serotina (PseMDL2) (Fig. 2e, f).

Fig. 2
figure 2

Isolation and functional analysis of PsMDL2 gene from P. sibiricaseeds. (a) Cloning of PsMDL2 gene from P. sibirica seeds by PCR. M and line 1 signified 2000 DNA marker and PsMDL2 gene fragment, respectively. (b) Assay of secondary structure for PsMDL2 by PSIPRED. The signal peptide at N-terminus was marked in blue box, and 4 potential N-glycosylation sites (N-X-T/S) were marked in red box. (c) Analysis of transmembrane domain of PsMDL2 by TMHMM. (d) The predicted 3D model for PsMDL2 by AlphaFold2. The active site (His) at 523 position (His523) of PsMDL2 protein was predicted by 3D structure model, and the purple dashed box a was a close-up view on the region of His523 residue, in which His523 was highlighted in green. (e) Multiple alignment assay for amino acid sequences of MDL2 proteins from P. sibirica (PsMDL2), Prunus mume (PmMDL2, BAO51918.1), Prunus avium (PaMDL2, XP_021808893.1), Prunus persica (PpMDL2, XP_007222890.1), Prunus dulcis (PdMDL2, Q945K2.1), Prunus serotina (PseMDL2, O50048.1), and Arabidopsis thaliana (AtMDL2, At1g73050). The conservative sequences (YWHYHGG) were marked in yellow box, the FAD binding motif (GGGTSG) was marked in green box, and the letters with black and other colors showed identical and similar amino acids. (f) Phylogenetic assay of MDL2 protein from P. mume (PmMDL2, BAO51918.1), P. avium (PaMDL2, XP_021808893.1), P. persica (PpMDL2, XP_007222890.1), P. dulcis (PdMDL2, Q945K2.1), P. serotina (PseMDL2, O50048.1), Rhaphiolepis bibas (RbMDL2, ACY69988.1), Pyrus ussuriensis×Pyrus communis (PucMDL2, KAB2624723.1), and Malus sylvestris (MsyMDL2, XP_050112728.1) using neighbor-joining method

Full size image

Analysis of subcellular location and tissue-specific transcript of PsMDL2

To explore the role of PsMDL2 expression in governing mandelonitrile production of P. sibirica seeds, we examined whether PsMDL2 gene was specifically expressed in P. sibirica seeds. By qRT-PCR detection, the transcriptional level of PsMDL2 was notably increased in mature seeds (70 DAF), about 418 − 3,008 times higher than in the leaf, stem, the root or flesh (Fig. 3a), emphasizing a typical seed-specific transcript pattern of PsMDL2 in P. sibirica.

Fig. 3
figure 3

Analyses of tissue-specific expression and subcellular localization of PsMDL2. (a) Detection of differential transcript of PsMDL2 in different tissues of P. sibirica by qRT-PCR. The genes encoding for cyclophilin (CYP) and ubiquitin-conjugating enzyme (UBC) was used as the inner reference, the expression level in the roots was set to 1.00 for standardization, and all the analyses were performed in three biological replicates with three technical repetitions. (b) Subcellular localization of PsMDL2 protein in subepidermal cells of Nicotiana benthamiana leaves (Bar = 50 μm). 35S::GFP transient transformation was used as positive control. Dark was defined as the dark-field GFP-fluorescence image, bright was defined as the bright-field image, and the merged was defined as the merged image in both bright and dark fields

Full size image

Additionally, subcellular location of PsMDL2 protein was assayed using transient expression system in N. benthamiana. A fluorescence signal of 35S::PsMDL2-GFP fusion protein was marked within cell membrane, but the 35S::GFP control showed the fluorescence signal in cell membrane and nucleus (Fig. 3b), indicating that PsMDL2 was mainly localized in cell membrane, corresponded to the bioinformatics assay result (Fig. 2c).

Purification of recombinant PsMDL2 from Pichia pastoris and detection of enzymatic property

To investigate the biochemical property of PsMDL2, the enzyme was heterologously expressed in yeast (Pichia pastoris) (Additional file 2: Fig. S2), and the recombinant enzyme was purified by a combination of ammonium sulfate and His-tag Protein Purification Kit. The PsMDL2 activity (1.51 − 14.81 U/mg) for culture supernatant of recombinant P. pastoris GS115 gradually increased with an increase of induction times (12–96 h) using mandelonitrile as natural substrate, and notably, enzymatic activity of purified PsMDL2 was 170.25 U/mg after 96 h of induction, about 10.5 − fold higher than that of unpurified samples, but the negative control (empty vector) showed no any catalytic activity (Fig. 4a), revealing that PsMDL2 from P. sibirica seeds displayed catalytic property of mandelonitrile lyase.

Fig. 4
figure 4

Analysis of enzymatic kinetics property of recombinant PsMDL2 by heterologous expression in Pichia pastoris. (a) Enzymatic activity of PsMDL2 was detected for culture supernatant and purified sample of recombinant P. pastoris GS115. One activity unit (1U) was expressed as the amount of enzyme required for 1 µM of cleaved mandelonitrile per min under analysis condition, and the result was standardized and defined as the amount of PsMDL2 protein (U/mg protein). (b) Assay of enzyme kinetic characteristics of purified PsMDL2. Lineweaver-Burk plots for PsMDL2 activity in the presence of different concentrations of mandelonitrile (0.05 − 4.00 mM). l/V represented reciprocal of reaction rate, l/S represented reciprocal of the substrate (mM), Vmax represented the maximum reaction rate, and Km represented Michaelis constant. Intersection point of vertical coordinate was 1/Vmax, and Intersection point of abscissa coordinate was − 1/Km

Full size image

Both Km and Vmax are known as vital kinetic parameters of the enzyme. Here, a Lineweaver–Burk plot of recombinant PsMDL2 enzymatic activity was constructed by drawing the inverse of substrate concentration (1/[S]) against the inverse of enzyme catalytic velocity (1/v) (Fig. 4b), from which the values of Vmax and Km of PsMDL2 were calculated to be 178.57 U/mg and 0.22 mM, respectively. This further confirmed that PsMDL2 from P. sibirica seeds had high affinity and enzymatic activity for natural substrate (mandelonitrile).

Expression of PsMDL2 promotes on plant growth and seed development of transgenic lines

To assess whether PsMDL2 expression affected phenotype of transgenic plants, the traits (seed size and weight, plant height and root length) were analyzed on the seedlings from the WT (wild type, as the control), and different transgenic plants of PsMDL2 overexpression (35S::PsMDL2) and its mutant (mdl2) from T3 homozygous seeds (Additional file 3: Table S1 and Additional file 4: Fig. S3). The plant height (33.07–35.12 cm) and thousand-seed dry weight (20.12 − 20.33 mg) of transgenic 35S::PsMDL2 lines were greater than those of the control (27.53 cm and 17.80 mg, respectively), whereas transgenic lines showed a low root length (4.65–4.81 cm) compared with the control (5.82 cm) (Fig. 5), and thereby concluded that overexpression of PsMDL2 promoted plant growth and seed development of transgenic lines, but repressed root growth. However, plant height (24.47–25.31 cm), root length (3.05–3.13 cm) and thousand-seed dry weight (14.33 − 15.17 mg) of mdl2 mutant plants were lesser than those of both the control and PsMDL2 transgenic lines (Fig. 5), pointing to a negative contribution of MDL2 mutation to plant growth and seed development.

Fig. 5
figure 5

Effect of PsMDL2 expression on phenotype of transgenic Arabidopsis. (a) Plant growth status of wild type (WT) and different transgenic Arabidopsis at 48 d after transplanting. (b) The root growth status of WT and different transgenic Arabidopsis seedlings at 20 d after germination. (c) Seed phenotype of the WT and different transgenic lines. (d) Analysis of plant height of different transgenic lines at 48 d after transplanting. (e) Analysis of root length of different transgenic lines at 20 d after germination. (f) Detection of thousand seed dry weight of different transgenic lines. The analyses of root length, plant height and seed weight were conducted on six biological replicates, and the results were presented as the mean ± SD, and the different letters above columns indicate significant differences at P < 0.05

Full size image

Expression of PsMDL2 results in decreased mandelonitrile content of transgenic seeds

To ascertain whether PsMDL2 expression contributed functionally to seed mandelonitrile dissociation in transgenic plants, we measured mandelonitrile amount in ripen seeds of the transgenic plants (35S::PsMDL2 and mdl2 mutant) and control group. The mandelonitrile amount in ripen seeds of 3 independent PsMDL2-overexpressed lines was 0.31–0.41 µg/g DW, which was much lesser than that of the control (WT) (0.71 µg/g DW) (Fig. 6a), indicating that overexpression of PsMDL2 could reduce seed mandelonitrile level of transgenic lines. Also consistent with this fact, mdl2 mutant seed accumulated higher levels of mandelonitrile (1.03–1.10 µg/g DW) than both PsMDL2 overexpressing lines and WT plants (Fig. 6a).

Fig. 6
figure 6

Effect of PsMDL2 expression and its mutation on the contents of mandelonitrile and oil in the seeds of different transgenic lines. (a) Analyses of differential contents of mandelonitrile in mature seeds from different transgenic lines. (b) Assays of differential contents of seed oils from different transgenic lines. The content of extracted oil from each transgenic seed was calculated as the difference between the weights of seed sample before and after extraction, and the result was expressed as the percentage of the extracted oil weight to dry seed weight (%, g/g). (c) Detections of differential contents of various FA compositions in seed oils from different transgenic lines. The seed oils from each transgenic line were trans-esterified to detect the content of FA compositions, and the result was expressed as the percentage (%) of the obtained amount (g) of each FA methyl ester (FAME) to the amount (g) of raw seed oils. (d) Assays of differential levels of saturated, monounsaturated and polyunsaturated FAs in the seed oils from different transgenic lines. 35S::PsMDL2 represented overexpression of PsMDL2 driven by CaMV35S promoter, and mdl2 represented mutant. The results were presented as the mean ± SD by detecting three biological replicates with three technical repetitions, and the different letters above columns indicated significant difference at P < 0.05

Full size image

Expression of PsMDL2 increases oil accumulation of transgenic seeds with altered FA profiles

In higher plants, heterotrophic sink organs (such as seeds or fruits) are suppled carbon source mostly as sucrose, which is converted to some key precursors for oil or amygdalin biosynthesis via glycolysis and pentose phosphate pathway (PPP). This prompted us to gain insight into whether PsMDL2 expression caused the alteration of oil content of transgenic seeds. Compared with the WT (34.61%), the increased oil content (37.66 − 38.15%) was marked in mature seeds of 35S::PsMDL2 lines, but a decrease of seed oil content (31.89 − 31.95%) was detected for mdl2 mutants (Fig. 6b and Additional file 5: Table S2), revealing that PsMDL2 overexpression could increase oil production of transgenic seeds.

Another question was whether PsMDL2 expression could affect FA profiles in the seed oils of transgenic lines. Here, a total of 7 kinds of FAs were identified in the seed oils of all examined plants, including C16:0 (2.71 − 2.78%), C18:0 (2.65 − 2.79%), C18:1 (4.78 − 10.70%), C18:2 (5.68 − 9.91%), C18:3 (1.91 − 7.45%), C20:0 (1.58 − 1.69%) and C20:1 (6.11 − 6.63%) (Fig. 6c and Additional file 5: Table S2). Of these, the contents of C18:1 (8.36 − 8.70%) and C18:2 (9.83 − 9.91%) in the seed oils of 35S::PsMDL2 lines increased with a decline of C18:3 (5.71 − 5.81%) compared with the control (4.78%, 8.67% and 7.45% respectively), implying that PsMDL2 expression could prompt production of C18:2 and C18:1 in the seed oils of transgenic plants, but suppress C18:3 synthesis. Yet, a notable increase of C18:1 (10.36 − 10.70%) was detected in seed oils of mdl2 mutants with the decreased C18:2 (5.68 − 5.71%) and C18:3 (1.91 − 2.07%), indicating that mutation of mdl2 could lead to 18:1 accumulation. Also, the other FAs exhibited no significant difference in seed oils among all examined seeds. These results indicated that the expression or mutation of PsMDL2 could cause an alteration of unsaturated FAs in transgenic seed oils (Fig. 6d and Additional file 5: Table S2).

Expression of PsMDL2 responses specifically to seed development of transgenic plants

To highlight potential regulatory role of PsMDL2 in controlling mandelonitrile and oil accumulation of transgenic seeds, we analyzed spatiotemporal transcription pattern of PsMDL2 in various tissues and different develo** seeds of transgenic PsMDL2 lines. A development-dependently increased transcript profile of PsMDL2 was detected in transgenic seeds, of which a significant increase was recorded at 8–18 DAA (day after anthesis), and importantly, transcriptional amount of PsMDL2 in develo** seeds was greater than in vegetative organs (stems, roots, flowers, and leaves) (Fig. 7), and thus concluded that PsMDL2 expression responded specifically to seed development.

Fig. 7
figure 7

Detection of spatiotemporal expression of PsMDL2 in different tissues of transgenic Arabidopsis by qRT-PCR. 35S::PsMDL2 represented PsMDL2 overexpression driven by the CaMV35S promoter. The expression level from the root sample was arbitrarily set to 1.00 for standardization, and Arabidopsis EF1α gene was used as internal control. Error bars were SD of three biological replicates with three technical repetitions, and the different letters above columns indicated significant difference at P < 0.05

Full size image

Expression of PsMDL2 activates transcription of mandelonitrile catabolism-related genes in transgenic seeds

To unravel molecular mechanism by which how PsMDL2 expression was targeted specially to reduce seed mandelonitrile level of transgenic plants (Fig. 6a), we investigated whether PsMDL2 overexpression could influence transcription of endogenous genes in relation to mandelonitrile catabolism in transgenic seeds. Thus, the ripen transgenic seeds with the maximum transcription of PsMDL2 were chosen as materials (Fig. 7) to detect transcription of the genes for some critical enzymes of mandelonitrile catabolism [MDL1/2/3, β-cyanoalanine synthase (β-CAS) and nitrilase 1/2/3/4 (NIT1/2/3/4)] by qRT-PCR. Compared with the WT, the differentially up-regulated transcripts of endogenous Atβ-CAS, AtNIT1/2/3/4 and AtMDL1/2/3 were identified in PsMDL2 transgenic seeds (Fig. 8a), emphasizing that PsMDL2 overexpression effectively promoted transcription of the genes essential for mandelonitrile catabolism in transgenic seeds. Also consistent with this fact, the reduced transcription of them was marked in mature seeds of mdl2 mutant lines (Fig. 8a). Given that the seeds of PsMDL2 transgenic lines accumulated much lower amount of mandelonitrile than those of mdl2 mutant lines and WT plants (Fig. 6a), it seems clear that overexpression of PsMDL2 could lead to enhancement of mandelonitrile catabolism, destined for less accumulation of mandelonitrile in PsMDL2 transgenic seeds.

Fig. 8
figure 8

Effect of PsMDL2 on expression of the genes for the enzymes associated with mandelonitrile catabolism and oil accumulation in mature seeds of different transgenic lines of Arabidopsis by qRT-PCR. (a) Differential transcripts for mandelonitrile catabolism-related enzymes. (b) Differential transcripts for FA biosynthesis-related enzymes. (c) Differential transcripts for TAG assembly-associated enzymes. WT, 35S::PsMDL2 and mdl2 represented the control, overexpression and mutant, respectively. The mature seeds (at 20 DAA) were used as the materials, and Arabidopsis EF1α gene was used as the internal control. The relative expression values in heatmap were counted as 2−△△Ct, and the expression level of the genes in mature seeds from the WT was arbitrarily set to 1.00 for standardization. Abbreviations for the enzymes are as follows: CAS, β-cyanoalanine synthase; DGAT1, acyl-CoA: DAG acyltransferase 1; EAR, enoyl-ACP reductase; FAD2/3, FA desaturases 2/3; FATA/B, fatty acyl-ACP thioesterase A/B; GPAT9, acyl-CoA: G3P acyltransferase 9; KAS I/II/III, 3-ketoacyl ACP synthase I/II/III; LACS9, long-chain acyl CoA synthetase 9; LPAAT4, acyl-CoA: LPA acyltransferase 4; MAT, malonyl-CoA-ACP transferase; MDL, mandelonitrile lyase; NIT, nitrilase; PAP2, PA phosphatase 2; PDAT, diacylglycerol (DAG) acyltransferase; SAD6, 18:0-ACP desaturase 6; TAG, triacylglycerol

Full size image

Expression of PsMDL2 promotes transcription of oil synthesis-associated genes in transgenic seeds

The impact of PsMDL2 expression on seed oils and its FA profiles of transgenic plants (Fig. 6b-d) allowed to test transcript level of endogenous genes in relation to oil accumulation in the transgenic seeds, involving in FA biosynthesis [18:0-ACP desaturase 6 (SAD6), malonyl-CoA-ACP transferase (MAT), fatty acyl-ACP thioesterase A/B (FATA/B), FA desaturase (FAD2/3), enoyl-ACP reductase (EAR), 3-ketoacyl ACP synthase I/II/III (KAS I/II/III) and acetyl-CoA carboxylase (ACC)], and TAG assembly [PA phosphatase 2 (PAP2), acyl-CoA: DAG acyltransferase 1 (DGAT1), acyl-CoA: LPA acyltransferase 4 (LPAAT4), DAG acyltransferase 1 (PDAT1), long chain acyl-CoA synthetase 9 (LACS9) and acyl-CoA: G3P acyltransferase 9 (GPAT9)], almost of which (except FAD3) showed a differential up-regulation in PsMDL2 transgenic seeds compared with the WT, whereas the down-regulated transcript was marked in the seeds of mdl2 mutants (Fig. 8b, c), revealing that PsMDL2 overexpression effectively enhanced the transcriptional levels of multiple genes demand for oil biosynthesis, destined to an increase in seed oil production of transgenic plants. Another concerned was that the transcriptional level of DGAT1 in mature seeds was significantly higher than that of PDAT1 across all tested plants (Fig. 8b, c), reflecting that DGAT1 may be a crucial regulator for seed TAG assembly.

Discussion

MDL2 as crucial CYP member for seed low amygdalin accumulation of P. sibirica across different germplasms

Amygdalin is one of the richest groups of CNGs in the Rosaceae family (such as apricot, cherry and apple). Prunus sibirica (Siberian apricot), one widely distributed member of Prunus genus in the Rosaceae family in China, has been identified with rich seed oil content [14,15,52,53,54], and thus concluded that the oils from P. sibirica seeds had great potential as a source for edible oil and biodiesel. In this work, an obvious variation on the contents of amygdalin (0.051−5.314%) and mandelonitrile (0.004−0.317%) was detected in seeds across different accessions of P. sibirica (Fig. 1a), as was also noted in several Prunus species from different regions [55,3, 27, 31,32,33]. The release of toxic HCN from amygdalin hydrolysis has become a serious health issue [1a, c), it seems clear that the less of mandelonitrile biosynthesis may mostly contribute to low accumulation of amygdalin in P. sibirica seeds. Hence, exploration of the mechanism for controlling mandelonitrile accumulation destined to low amygdalin in seeds of P. sibiricas is one pivotal challenge.

Hydroxynitrile lyases (HNLs) are a group of critical enzymes in cyanogenic plants, which not only catalyze the cleavage of cyanohydrin into HCN and carbonyl compound, but also catalyze the reverse reaction for the synthesis of cyanohydrin in fine chemical or pharmaceutical industry [36, 59]. According to flavoprotein (FAD) content, HNLs is generally divided into two groups, FAD containing MDLs and non-FAD containing HNLs. Of note, almost HNLs have been identified with asymmetric synthesis of mandelonitrile in different families of cyanogenic plants [36,37,38,39,40,41,42,43,44], and MDLs have been reported only in black cherry and almond with specific activity for the dissociation of mandelonitrile [34, 45,46,47,48], which raised a vital question of which MDL was contributed specifically to decrease amygdalin accumulation of P. sibirica seeds. In this work, we applied a combined analysis of genome-wide annotation of mandelonitrile dissociation-associated MDL family members (Table 1) and the association with accumulative level of mandelonitrile and amygdalin and transcriptional amount of MDL family member in the seeds across different germplasms or develo** stages of P. sibirica (Fig. 1) to determine MDL2 as critical member for low accumulation of amygdalin and mandelonitrile of P. sibirica seeds. Also, the PsMDL2 from the seeds of P. sibirica had four N-glycosylation functional sites (N-X-T/S) and one typical FAD-binding motif (GGGTSG) (Fig. 2e), as the case for other MDLs (MDL1/2/3/4/5) [46], reflecting that PsMDL2 was a FAD-dependent HNL. In support of this fact, the purified recombinant PsMDL2 displayed a specific and high activity of 170.2 U/mg for dissociation of mandelonitrile into benzaldehyde (Fig. 4a) with the Km value (0.22 mM mandelonitrile) (Fig. 4b), which was lower than that (0.4 − 3.76 mM) reported for MDLs from Amygdalus pedunculata, Eriobotrya japonica and Prunus armeniaca [40, 42, 60], and thereby concluded that PsMDL2 had a high affinity of mandelonitrile. Also noteworthy was an involvement of Cys or Ser residue in catalytic mechanism of MDLs [46, 47]. Our finding that the catalytically active residue of His523 was identified for PsMDL2 (Fig. 2d), pointed to a diversity of MDL isoforms from different plant species.

PsMDL2 expression-mediated regulation for mandelonitrile and oil accumulation in of transgenic seeds

Almost FAD-independent HNLs in several plants (such as Arabidopsis, cassava, Japanese apricot, almond and peach) have been identified with asymmetric synthesis of mandelonitrile by the assay of overexpression, mutant, or catalytic activity [36,37,38,39,40,41,42,43,44]. To date, however, no MDL gene has been conducted on genetic transformation in plants to unravel its function. This work presented for the first time a combination of overexpression and mutation to identify PsMDL2 function. Given that PsMDL2 overexpression notably decreased mandelonitrile amount in the seeds of transgenic lines, but mdl2 mutant lines showed a great accumulation of mandelonitrile in comparison with the WT lines (Fig. 6a), it seems certain that PsMDL2 overexpression effectively reduced mandelonitrile accumulation in transgenic seeds, which was compatible with recent report showing that mandelonitrile amount was highly increased in hnl mutant plants than in the WT ones with a notable reduction in AtHNL overexpressing lines [36]. These, combined with a negative correlation of PsMDL2 transcriptional level with mandelonitrile accumulative amount detected in P. sibirica seeds across various germplasms or developmental periods (Fig. 1), revealed a key role of PsMDL2 for reduced mandelonitrile accumulation in P. sibirica seeds. This was in contrast to the ideal of MDL1 as crucial enzyme for the dissociation of mandelonitrile into benzaldehyde and HCN in the seeds of P. amygdalus and P. serotina [34, 46, 48]. It was also noteworthy that the plants have developed one effectively detoxicated route in which HCN can be converted by β-CAS to form β-cyanoalanine, and then hydrated by NITs to produce aspartate or asparagine and ammonia, all of which serves as reduced nitrogen source destined for plant growth [2, 3, 6]. Such pathway in our work could be clearly shown by the significantly up-regulated transcripts of endogenous AtCAS, AtNIT1/2/3/4 and AtMDL1/2/3 in the seeds of transgenic PsMDL2 lines compared with mdl2 mutant and the WT lines (Fig. 8a), and an obvious increase of plant height and seed weight of PsMDL2 transgenic lines (Fig. 5), indicating that the abundantly coordinated transcriptions of them may contribute mostly to plant growth and seed development of PsMDL2 transgenic plants. This coincided with the fact that expression of NIT4A/NIT4B2 or β-CAS could promote flower and grain development [3, 5].

Another concern was the effect of PsMDL2 overexpression or its mutation on the seed oils of transgenic plants. Given an opposite accumulation pattern between oil and mandelonitrile in the seeds across different transgenic lines (Fig. 6a, b), it seems likely that oil accumulative level might be affected by mandelonitrile amount in different transgenic seeds. It is known that the source of Phe required for phenylalanine formation destined to mandelonitrile and amygdalin synthesis is derived mostly from phenylpyruvate through glycolysis or erythrose 4-phosphate via oxidative pentose phosphate (OPP) pathway [61], both of which can also provide carbon source for acyl-CoA formation essential for FA biosynthesis and oil accumulation [14, 15, 52,53,54, 62,63,64]. Thus, the increased seed oil content in PsMDL2-overexpressed lines may be due to low funneling of carbon into amygdalin biosynthetic pathway. Yet, the question of how transgenic plant (overexpression or mutation) with altered amygdalin content is targeted specifically to regulate seed oil biosynthesis at the molecular level is still unknown. In oilseed plants, oil accumulation is known as a part of seed development, involving a series of gene expression. High transcripts of the enzymes required for FA biosynthesis and TAG assembly may play a key contribution to oil accumulation in the seeds of several oil plants [18, 52,53,54, 62, 64]. Here, the enzymes relevant for FA synthesis (EAR, KAS I/II/III, ACC, SAD6, FATA/B, MAT, and FAD2) and TAG assembly (LACS9, PAP2, GPAT9, LPAAT4, and DGAT1) were coordinated transcriptionally with PsMDL2 in the seeds of PsMDL2 transgenic lines (Figs. 7 and 8b and c), both of which displayed a close association with the increased seed oils (Fig. 6b), but the coordinatedly down-regulated transcripts of them with low oil content were noted in the seeds of mdl2 mutant (Figs. 6b and 8b and c), indicating that overexpression of PsMDL2 effectively activated transcription of multiple genes crucial for oil accumulation of transgenic seeds. This conclusion could be evidenced by the results of suppression, mutant or overexpression of some key oil biosynthesis-related genes (such as DGAT1, GPAT9, LACS4, FATB and LPAAT4) in several oil plants [65,66,67,68,69,70,71,72,73].

Altogether, low mandelonitrile level with high oil amount was identified in PsMDL2-overexpressed seeds, but an opposite content change pattern was marked in the seeds of mdl2 mutant, revealing that expression of PsMDL2 would offer high seed oil production with low amygdalin content. Hence, PsMDL2 was one key molecular target for engineering high seed oil production with low amygdalin in plants.

Conclusions

In this work, a combined analysis of genome-wide characterization of mandelonitrile dissociation-associated MDLs and the association of mandelonitrile/amygdalin accumulative amount with transcript level of MDL family members in seeds across different germplasms or developmental periods of P. sibirica was conducted to determine MDL2 as critical family member essential for seed mandelonitrile dissociation destined to low amygdalin accumulation. PsMDL2 showed specific and high transcription in the seeds, and had high enzymatic specificity for mandelonitrile dissociation. Overexpression of PsMDL2 from the seeds of P. sibirica in Arabidopsis greatly promoted seed development and oil production, reduced mandelonitrile accumulation, and activated transcription of genes essential for mandelonitrile dissociation (NIT1/2/3/4, CAS and HNL), FA biosynthesis (EAR, KAS I/II/III, ACC, SAD6, FATA/B, MAT, and FAD2) and TAG assembly (LACS9, PAP2, GPAT9, LPAAT4, and DGAT1), but all of which showed an opposite situation in mdl2 mutant. Our results revealed an importance of MDL2-mediated regulation in controlling seed amygdalin and oil biosynthesis. These results will be interest to study cyanogenic glycosides in plants, and the PsMDL2 gene as important molecular target has the potential for bioengineering high seed oil production with low amygdalin in oilseed plant.

Materials and methods

Plant materials

Five superior germplasms of P. Sibirica (accessions Ps-14/23/46/63/70) with high-yield of the seeds were marked by our recent investigated on different P. Sibirica germplasms located at Bei**g Chang** District of China (E116°23′, N40°22′) [14], and deposited in Forest and Flower Germplasm Resource Genebank of Bei**g Forestry University in China (Voucher No. 1111C0003307002308, 1111C0003307002309, 1111C0003307002347, 1111C0003307002008 and 1111C0003307002010, respectively). The different tissues (root, stem, leaf and mature fruit) were harvested from 12-year-old tree. About 150 fresh fruits with an average of 15 fruits per tree were collected respectively from 10 (unmatured period) to 70 DAF (fully ripen period) from the accessions Ps-63 and Ps-70. After the sarcocarps were removed, the plump fresh seeds were picked and kept at − 80 °C for amygdalin and mandelonitrile assay, gene cloning and qRT-PCR detection.

The seeds of wild type (WT) Arabidopsis (Col-0 ecotypes) were provided by our teams, and the seeds of Arabidopsis mdl2 mutant (CS926813) were obtained from Nottingham Arabidopsis Stock Center (http://arabidopsis.info). For the genetic transformation of PsMDL2, the sterilized WT seeds were germinated on 1/2 MS medium and subjected to a 4-day cold stratification at 4 °C in the darkness. Subsequently, they were moved into one controlled chamber under specific conditions (temperature alternating between 18/22°C night and day, a photoperiod of 8/16 h night and day, and light intensity of 200 µmol m− 2 s− 1). Following a 20-day period, the plants were relocated to the pots filled with a soil blend (perlite/vermiculite/humus-soil, 1:3:3, v/v/v) and then cultivated under specific conditions of 8 h of darkness (22 °C) and 16 h of light (24 °C) to facilitate the genetic transformation of gene and the collection of seeds.

Genome-wide annotation of the genes for MDL family members in the seeds of P. sibirica

The genes of the MDL family members were globally identified according to recent 454 RNA-seq result (SRX339392) of P. sibirica [49] by using BLASTX alignment against known protein databases of CDD, AP, SWISS-PROT, PFAM, TREMBL, COG, Arabidopsis proteome and NCBI nonredundant. Functional classification was analyzed by the GO terms using Blast2Go software, and KEGG pathway assignment was performed by the BLAST all against the Kyoto Encyclopedia of Genomes and Genes database.

Detections of amygdalin and mandelonitrile in P. sibirica seeds

Detections of amygdalin and mandelonitrile contents were performed in mature seeds of all examined 5 germplasms and different develo** seeds of two P. sibirica germplasms (Ps-63 and Ps-70) by HPLC. About 0.5 g of dry P. sibirica seeds (three samples, every germplasm or every developmental stage) was powdered and lyophilized using a freezing dryer, and then extracted for 30 min with ultrasonic generator (500 W, 40 kHz) by methanol as solute [56]. The supernatant obtained after centrifuging for at 4 °C, 10 min under 10,000 × g was clarified by filtration via microfiltration membrane (0.45 mm) for clarification.

HPLC analysis was conducted on Agilent 1290 Infinity II (USA) equipped with a chromatographic column (ZORBAX Eclipse Plus C18), and the temperature and pressure were respectively programmed at 40 °C and 317.0 bar. In the mobile phase of H2O: acetonitrile (90:10, v/v), the flow rate was 0.5 mL/min. A 10 µL of sample was injected, and the amygdalin and mandelonitrile were detected at 214 and 247 nm, respectively. The peaks were determined by comparison of known standards with their retention times, and the IBM SPSS Statistics 23 was applied to assay peak integration. The contents of amygdalin and mandelonitrile were calculated from their respective standard curves, and the obtained results were calculated as the percentage of dry seed weight (%, g/g). All analyses were presented in in three sets of biological replicates, each consisting of three technical repetitions.

Isolation and in silico analyses of PsMDL2 gene

RNA was isolated from seeds of P. sibirica and converted into complementary DNA using the EASYspin Plus plant RNA extraction kit and TRUEscript RT Kit (AidLab, China), respectively. Functional annotation and expression abundance were utilized to screen PsMDL2 gene from our transcriptome database [17, 49], and the obtained sequence was applied for designing the primers by using Primer Premier 5.0 (Additional file 6: Table S3) to meet PsMDL2 gene amplification by PCR. For the sequencing, the amplified products of PsMDL2 were inserted into plasmid (pTOPO-TA) and then delivered into DH5α (Escherichia coli).

To assay open reading frame of PsMDL2, the Online ORFfinder (https://www.ncbi.nlm.nih.gov/orffinder/) was utilized. A maximum likelihood method was applied to construct phylogenetic tree using RAxML-NG (https://raxmlng.vital-it.ch/) [74]. Multiple sequence alignment of PsMDL2 and homologous proteins of MDL family was made by DNAMAN7.0 via ClustalX program. Subcellular location of PsMDL2 was assayed by online tools of Plant-mPLoc and BUSCA. The predictions of secondary structure and membrane spanning domain of PsMDL2 were respectively used the websites of PSIPRED 4.0 and TMHMM2.0. 3D structure of PsMDL2 was conducted by online tool AlphaFold2 and potential regulatory site for protein activity was predicted by using homologous protein MDL2 from Prunus mume on the UniProt website (https://www.uniprot.org/).

Subcellular location of PsMDL2 protein

Transient expression analysis of fusion protein PsMDL2-GFP was applied to detect subcellular location of PsMDL2 in leaf cells of Nicotiana benthamiana [17, 18]. The complete ORF of PsMDL2 was cloned into pBI121/35S::GFP to generate vector of 35S::PsMDL2-GFP (Additional file 7: Fig. S4) and transformed A. tumefaciens GV3101 by using the freeze-thaw technique. The leaves of N. benthamiana (4-week-old) were applied for transient expression assay, and the fluorescent images were taken after 48 h of incubation at 22 °C in the dark, using positive fluorescence microscope SP8X (Leica TCS, Germany).

Expression and purification of PsMDL2 in Pichia pastoris

The complete ORF of PsMDL2 was amplified and seamlessly cloned into EcoR I site of plasmid vector of pPIC9 to obtain expression vector of pPIC9/AOX1::PsMDL2 (Additional file 6: Table S3) by Seamless Assembly Cloning Kit. The obtained construct and empty vector pPIC9/AOX1 were respectively transformed into P. pastoris GS115 cells, and PCR amplification was applied to verify authenticity of constructed vector. The transformant was cultured as previously described [38] with minor modification. The transformed single colony was cultivated in 50 mL of YPD medium (0.01 g/ mL yeast, 0.02 g/mL glucose, and 0.02 g/mL peptone) using a shaker under 250 rpm/min at 30 °C for 16 h, from which a portion (100.0 µL) of culture medium was introduced into 100.0 mL of BMGH medium (containing 1.0% glycerol, 0.04% biotin, and 1.34% yeast nitrogen base without amino acid). The resulting mixture was then cultured under the condition of 30 °C with a shake at 250 rpm/min to obtain the A600 value of 1.0 − 2.0 of the cultures. The yeast microsomes were collected after being centrifuged for 10 min at 4 °C under 4,000 × g, and then transferred to 100.0 mL of BMMY medium (0.5% methanol, 0.04% biotin and 1.34% yeast nitrogen base without amino acid) and cultured for 96 h at 28 °C under a shake of 250 rpm/min, during which culture medium was added with methanol at intervals of 24 h.

To effectively purify recombinant enzyme, the above culture medium was needed to undergo centrifugation at 4 °C for 10 min with a speed of 4,000 × g. Subsequently, the resulting supernatant was mixed with solid ammonium sulfate and continuously stirred until it reached saturation. The precipitated protein was resuspended using a buffer of sodium phosphate with 10 mM and pH 6.0 and purified with the His-tag Protein Purification Kit (Beyotime, China) after centrifuging under 10,000 × g for 10 min at 4 °C. To analysis enzymatic activity, the purified PsMDL2 from P. pastoris transformant was dissolved in 0.1 M buffer of sodium citrate (pH 5.0).

Assay of enzymatic catalytic specificity of PsMDL2

PsMDL2 catalytic activity was tested spectrophotometrically by detecting the kinetic slope at 280 nm [39] with minor modification. The above obtained enzyme (100 µL) was added to reaction mixture (1.0 mL) of citrate-phosphate buffer (0.1 M, pH 5.0) with the substrate (mandelonitrile, 8.4 mM), and then the dynamic absorbance slope was detected at 25 °C for 5 min. The slope generated by autonomous breakdown of mandelonitrile was used as the control. One activity unit (1U) was expressed as the amount of enzyme required for 1 µM of cleaved mandelonitrile per min under the analysis condition, and the result was standardized and defined as the amount of PsMDL2 protein (U/mg protein). The content of protein was analyzed by BCA Protein Assay Kit (102,536, Abcam).

Km (Michaelis constant) and Vmax (the maximum reaction rate) as two key kinetic parameters were measured with different substrate mandelonitrile concentrations (0.05 − 4.0 mM) using Lineweaver–Burk plot of the Michaelis–Menten equation. Initial reaction rate (v) at each substrate concentration was detected, and then a Lineweaver–Burk plot of enzyme activity was made by drawing the inverse of substrate concentration (1/[S], as abscissa) against inverse of velocity of enzyme reaction (1/v, as ordinate), from which the values of Vmax (U/mg protein) and Km (mM) were calculated.

Ectopic overexpression of PsMDL2 in Arabidopsis

The complete ORF of PsMDL2 was amplified and subcloned into pCAMBIA1301 directed by the promoter of CaMV35S to generate plant expression vector of pCAMBIA1301/35S::PsMDL2. The obtained construct vector was transformed into A. tumefaciens GV3101 using the freeze-thaw technique, and then delivered into WT Arabidopsis plant [75]. The T3 seeds were obtained by screening transgenic lines on 1/2MS medium (7 g/L agar, 20 g/L sucrose, and 50 mg/L kanamycin) and identifying them through PCR amplification, and then three independent-homogenous transgenic lines were obtained and examined by chi-square (χ2) test. Finally, T4 transgenic lines were applied for further analysis. All primers for constructing plant expression vector were displayed in Additional file 6: Table S3.

Identification of Arabidopsis mdl2 homozygous mutant

To effectively confirm T-DNA insertion (Columbia-0 background SALK_058105.13.60.X, ID: CS926813), the identifications of homozygous mutant plants of mdl2 were performed on T3 seeds by PCR detection by the specific primers (Additional file 6: Table S3) and the χ2 test.

Assays of phenotypic traits of transgenic plants

To examine the impact of PsMDL2 overexpression and its mutation on the phenotypes of transgenic plants, the traits (seed size and weight, plant height and root length) were assayed for the seedlings of T4 transgenic Arabidopsis lines (PsMDL2 overexpression and mdl2 mutant) and the WT. The root length was detected for the seedlings (10 days old) that were sprouted on 1/2MS medium. To detect plant height, the seedlings with 20 day old were transferred to pots and cultivated within the same chamber [53]. The weight of 1000 seeds (mg/1000 seeds) obtained from the T4 genetically modified plants and the control were determined. All determinations were conducted in triplicate.

Detections of seed oil amount and its FA compounds

The oil content and FA compounds were detected in mature seeds of WT and T4 generation of different transgenic lines. Approximately 20 seeds were weighted and subjected to trans-methylation in 300 µL of toluene and 1 mL of methanol (including 250 µg C17:0, 50 µg BHT and 50 µL H2SO4) at a temperature of 95 °C for 1.5 h, followed by the addition of 1.5 mL sodium chloride solution (0.9%, m/v). The oils were obtained through three consecutive extractions using 1.0 mL of hexane, and evaporated under a stream of nitrogen and re-dissolved in 50 µL hexane [76]. The obtained FA methyl esters (FAMEs) was applied to analyze FA profiles using a Agilent 6890 (California, USA) gas chromatograph equipped with flame ionization detector (GC-FID) [14]. A capillary column (HP-INNOWax) with split 1:20, film thickness 0.5 μm and inner diameter 0.32 mm was employed. The temperature was firstly conducted at 60 °C and raised at a speed of 4 °C/min until reaching 220 °C. Subsequently, it was further heated to 240 °C and maintained for 10 min. Helium was used as carrier gas at a flowing speed of 1.0 mL/min. The comparison of established standards with their retention times was applied to determine the peaks of FAMEs, and the HP3398A software was applied to assay peak integration. All analyses were presented in three sets of biological replicates, each consisting of three technical repetitions.

qRT-PCR assay

The isolated RNA from A. thaliana or P. sibirica by RNeasy Plant Kits (TaKaRa, Japan) was respectively quantified by a spectrophotometer (Nanodrop ND-1000, USA), and then reversely transcribed into cDNA using iScript cDNA Synthesis Kit (Bio-Rad, USA). The qRT-PCR analysis was conducted on 7500 Real-Time PCR System by using SYBR Premix Ex TaqTMII (TaKaRa, Japan), and all amplified primers used for amplification (Additional file 6: Table S3) were designed with an assistance of the PrimerQuest tool. The genes for ubiquitin-conjugating enzyme (UBC) and cyclophilin (CYP) were applied as control references [14, 15] for detecting transcript of mandelonitrile hydrolysis-related MDL family members in the seeds of P. sibirica. Additionally, gene expression in various tissues and different developmental seeds of transgenic plants was detected using EF1α gene of Arabidopsis as the control reference. For each qRT-PCR assay, three biological replicates with three technical repetitions each were performed.