Isolation and functional analyses of PvFAD2 and PvFAD3 involved in the biosynthesis of polyunsaturated fatty acids from Sacha Inchi (Plukenetia volubilis)

Total RNA extraction and cDNA preparation

Sacha Inchi plants were grown under natural climate conditions at the Xishuangbanna Tropical Botanical Garden (XTBG), Chinese Academy of Sciences (CAS), Yunnan, China. The developing seeds were collected 45 days after pollination (DAP) from a three-year-old individual (at this stage of seed development, the lipids are accumulated rapidly). The developing seed tissues were dissected and stored in a freezer at −80 °C until use. Total RNA was extracted using RNAprep Pure Plant Kit (DP432; TianGen) following the manufacturer’s protocol. The cDNA was synthesized from total RNA using a PrimerScript 1st Strand cDNA Synthesis Kit (Takara, Dalian, China) according to the manufacturer’s instructions.

Isolation of two fatty acid desaturase genes

To obtain the full-length nucleotide sequence of PvFAD2 and PvFAD3, we downloaded transcriptome sequencing reads from Short Read Archive (accession number: SRP101395) produced from different tissues of Sacha Inchi (Hu et al., 2018) and assembled them using the Trinity software (Grabherr et al., 2011). The coding sequence (CDS) of AtFAD2 (AT3G12120) and AtFAD3 (AT2G29980) from Arabidopsis were respectively used as a query to BLAST against the assembled transcripts database using the BLAST software (version 2.2.24+: https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE_TYPE=BlastDocs&DOC_TYPE=Download). Based on amino acid sequence alignment and functional domain analysis, we found the PvFAD2 transcript contained a full-length coding sequence. The PvFAD2 gene was further comfirmed by high fidelity PCR using TransStart FastPfu DNA Polymerase (TransGen, Beijing, China) with the specific primers, PvFAD2-F and PvFAD2-R (Table S1). To obtain the full-length of PvFAD3 nucleotide sequences, we adopted a 5′ and 3′ RACE (rapid amplification of cDNA ends) method using a SMART RACE cDNA Amplification Kit (Clontech, Palo Alto, CA) according to the manufacturer’s protocol. All used primer sequences are listed in Table S1. The final PCR products with the full-length PvFAD3 nucleotide were purified and cloned into the pEASY-Blunt cloning Vector (TransGen, Beijing, China) for subsequent sequencing to confirm the full-length PvFAD3 CDS. The obtained full-length CDS of both PvFAD2 and PvFAD3 were submitted to GenBank with accession numbers MK121677 (PvFAD2) and MK121679 (PvFAD3).

Sequence alignment and phylogenetic analysis

Homologous proteins of both PvFAD2 and PvFAD3 were retrieved from NCBI by BLASTP program (https://blast.ncbi.nlm.nih.gov/Blast.cgi), and resulting sequence information were listed in Table S2. For the phylogenetic analysis of PvFAD2 and PvFAD3, we included the plastid-located Δ12/ω-6 (FAD6) and Δ15/ω-3(FAD7 and FAD8) fatty acid desaturases from representative plants (see Table S2). Multiple sequence alignments were performed by the DNAMAN program. Phylogenetic analysis was conducted using MEGA (version 5.0) with the Neighbor-Joining criteria (Tamura et al., 2011). Branch support of the phylogenetic tree was estimated on the base of 10,000 bootstrap replicates of the data.

Subcellular localization of PvFAD2 and PvFAD3

The full-length CDS of PvFAD2 and PvFAD3 were inserted into the binary vector pCambia35S::GFP to respectively generate 35S::GFP-PvFAD2 and 35S::GFP-PvFAD3 constructs. Subsequently, the constructed plasmids were transformed into the Agrobacterium tumefa ciens (EHA105). Positive clones were selected on an LB plate supplemented with rifampicin (50 mg/L) and kanamycin (50 mg/L), and cultured in liquid LB medium containing appropriate antibiotics at 28 °C. After centrifugation, the pellets were resuspended in the infiltration buffer (10 mM MES, 200 µM acetosyringgone and 10 mM MgCl₂, PH5.6) until the OD600 of 0.8 for the final cell density. After 2 hours standing at room temperature, the constructs were delivered into the lower epidermis of 10-week-old tobacco leaves by agro-infiltration. Fluorescence was observed at 2–4 days post-transfection using the confocal system (Olympus, Fluoview FV1000).

Heterologous transformation in yeast (Saccharomyces cerevisiae)

To dissect the function of PvFAD2 and PvFAD3 in yeast, we constructed yeast expression vectors using homologous recombination strategy. Briefly, the full-length CDS was amplified by PCR using Phanta Max Super-Fidelity DNA polymerase (Vazyme, Nanjing, China). The primers used in this step were listed in Table S1. The PCR products were purified and then ligated into P426-GAP vector (digested with BamH I and EcoR I) using ClonExpress Entry One Step Cloning Kit (Vazyme, Nanjing, China). Finally, the plasmid harbored the PvFAD2 or PvFAD3 was transformed into S. cerevisiae (INVSc1) using Frozen-EZ Yeast Transformation II Kit (Zymo, USA). An empty vector was also transformed into S. cerevisiae as a negative control. Positive clones were obtained via selection on plates of SD medium lacking uracil at 30 °C. Recombinant strains of S. cerevisiae were then transferred into lipid SD medium lacking uracil, and shaken at 30 °C/230 rpm. To test PvFAD3 function, we exogenously added methyl linoleate (Sigma, USA) into media for the supply of C18:2 fatty acid substrates. When cells reached the stationary phase, cells were collected for subsequent analysis.

Heterologous transformation in tobacco

To inspect the functions of PvFAD2 and PvFAD3 in plants, the two genes were heterologously transformed into tobacco, and specifically expressed in developing seeds. The full-length CDS of PvFAD2 and PvFAD3 were constructed into the binary vector pCambia2300 under the control of the seed-specific Napin promoter, and named Napin::PvFAD2 and Napin::PvFAD3. These vectors carry the NptII gene encoding neomycine phosphotransferase conferring kanamycin resistance. Constructs were next transformed into tobacco (Nicotiana tabacum L. cv. Honghua dajinyuan) using the Agrobacteria-mediated tobacco leaf-disk transformation method (Horsch et al., 1985). Transgenic plants (T₀) were selected on shoot induction and root induction medium containing 100 mg/L kanamycin. To test whether the targeted genes were successfully transformed into the selected individuals, the genomic DNA was isolated from young leaves of T₀ plants and PCR amplification was performed to confirm the presence of the targeted gene with specific primers. The expression levels of transformed FAD genes in transgenic plants were tested using quantitative real-time PCR (qRT-PCR). Total RNAs were isolated from developing seeds using RNAprep Pure Plant Kit (TianGen, DP432) and reversely transcripted using PrimerScrip™ RT reagent Kit with gDNA Erases (Takara, China). qRT-PCR was performed on the CFX96 machine (Bio-Rad, Hercules, USA) according to the following program: pre-cycling at 94 °C for 30 s followed by 42 cycles of 94 °C for 5 s, 56 °C for 15 s and 72 °C for 10 s. The NtActin gene was used as an internal control. All primers used in this study were listed in Table S1. The matured seeds of T₀ plants were harvested for the next analysis.

Lipid extraction and fatty acid composition analysis

Total lipids were extracted from transgenic yeast and transgenic mature seeds, and fatty acid methyl esters (FAMEs) were prepared according to our previous description (Xu et al., 2014; Yang et al., 2016). To analyse the lipid composition of the yeast, 20 mL of different yeast strain cells at the stationary growth phase were collected and dissolved in 4 M HCl. After boiling for 10 min, total lipids were extracted using 500 µL hexane/isopropanol (3:2, v/v). Tobacco seeds from each line were homogenised in 500 µL hexane/isopropanol (3:2, v/v). Then the total lipids were dissolved in 500 µL chloroform and transmethylated with 2 mL of methanol containing 5% H₂SO₄ (v/v) and then heated at 85 °C for 90 min. Finally, FAMEs were subjected to Gas Chromatograph (Agilent 6890N) and Gas Chromatography-Mass Spectrometer (Agilent 7890/5975C) for fatty acid species analysis.

Isolation and sequence analysis of two FA desaturase genes

The full-length of the coding sequences of PvFAD2 and PvFAD3 were 1152 bp (encoding 384 amino acid residues with a calculated molecular mass of 44.19kDa) and 1137 bp (encoding 379 amino acid residues with a calculated molecular mass of 43.53 kDa), respectively. To characterize the structural features of PvFAD2 and PvFAD3 proteins, we performed multiple sequence alignments together with other functionally characterized homologs from plants, such as Jatropha curcas, Olea europaea, Glycine max, Vernicia fordii, Brassica napus and Arabidopsis thaliana (Fig. 1) (Wu et al., 2013; Hernández, Mancha & Martínez-Rivas, 2005; Li et al., 2007; O’Quin et al., 2010; Reed, Schafer & Covello, 2000; Botella et al., 2016). The result showed that all FAD2 proteins contained four transmembrane domains (TMDs) and three histidine boxes (HXXXH, HXXHH and HXXHH) essential for desaturase activity (Fig. 1A) (Shanklin, Whittle & Fox, 1994). Similarly, all FAD3 proteins contained four TMDs and eight conserved histidine residues in three histidine boxes (Fig. 1B). In addition, in the C-terminal FAD2 and FAD3 contained a ER-DIR motif known as an ER location signal (Dehghan Nayeri & Yarizade, 2014; McCartney et al., 2004). It should be noted that the N-terminal sequences of FAD3 proteins exhibited high variation, notwithstanding the high similarities to other regions (Fig. 1B). Collectively, the alignment of the amino acid sequences of FAD showed that PvFAD2 and PvFAD3 shared the three conserved histidine rich motifs with other known functional homologs, indicative of the proper encoding of full-length cDNA of both isolated PvFAD2 and PvFAD3.

Phylogenetic analysis of FA desaturases

To elucidate the phylogenetic relationship of plant membrane fatty acid desaturase, an unrooted neighbor-joining (NJ) tree was constructed by the deduced amino acid sequences of FA desaturases from diverse flowering species, including plastidial ω-3 (FAD7/8), ER-location ω-3 (FAD3), plastidial ω-6 (FAD6) and ER-location ω-6 (FAD2) (Table S2; Fig. 2). The topology of this tree clearly showed that FA desaturases appear divided into four well-resolved monophyletic groups with well-supported bootstrap values (Fig. 2), suggesting the conserved role of FA desaturases in each group. These groups consist of the FAD6 group catalyzing the formation of Δ12/ω-6 fatty acids in plastid; the FAD2 group responsible for the formation of Δ12/ω-6 fatty acids in ER, the FAD7/8 group that is Δ15/ω-3 desaturases in plastid, and the FAD3 group that is Δ15/ω-3 desaturases in ER. Moreover, in each group, most dicot members formed a group relative to the monocots in the same clade, which may stem from the divergent evolution of dicots and monocots. Plant plastid-localized FAD6 showed a close relationship with ER-location FAD2, while plastid-localized FAD7/8 were clearly clustered with ER-location FAD3 (Fig. 2), suggesting a common evolutionary origin of Δ12/ω-6 desaturases or Δ15/ω-3 desaturases located in ER and plastids (Fig. 2). PvFAD2 and PvFAD3 were clustered with ER-located Δ12/ω-6 and Δ15/ω-3 desaturases, respectively, implying that they have a conserved functional role in catalyzing the formation of ω-6 or ω-3 fatty acids by operating in ER.

Subcellular localization of PvFAD2 and PvFAD3 proteins

The intracellular desaturation of fatty acid chains occurs widely in different compartments such as plastid for synthesizing the desaturated membrane lipids and ER for synthesizing the desaturated storage lipids (Tocher, Leaver & Hodgson, 1998). To further inspect the subcellular location of PvFAD2 and PvFAD3, that is, where they function in desaturating fatty acids, we transiently expressed their proteins fused with GFP in epidermal cells of tobacco leaves. As shown in Fig. 3, under confocal microscopy a distinct subcellular co-localization of PvFAD2 and PvFAD3 protein with a ER membrane marker CD3-959-mCherry were observed, indicating that PvFAD2 and PvFAD3 function to produce desaturated fatty acids in ER.

Functional analysis the PvFAD2 and PvFAD3 in yeast (S. cerevisiae)

The two full-length cDNAs PvFAD2 and PvFAD3 were cloned into the yeast expression vector P426 for the detection of their enzymatic activities by investigating their changes in their lipid content and fatty acid composition. As shown in Table 1A, the fatty acid compositions obtained from the S. cerevisiae with the empty vector (P426), grown under normal growth condition (30 °C), mainly included palmitic acid (C16:0), palmitoleic acid (C16:1), stearic acid (C18:0) and oleic acids (C18:1). When we heterologously expressed PvFAD2 in S. cerevisiae cells, we found substantial changes in FA species, resulting in two new peaks corresponding to hexadecadienoic acid (C16:2) and linoleic acid (C18:2, LA), respectively (also see Fig. S1A). Subsequently, the proportion of C16:1 or C18:1 fatty acids was significantly reduced in the lipids obtained from transformants, from 38.9% to 36.27% and from 35.44% to 24.34% respectively, while other FA species did not display a distinct change (Table 1A). These results suggested that PvFAD2 can desaturate both C16:1 and C18:1 FA and produce the C16:2 and C18:2.

Usually, ER-location ω-3 desaturase can directly catalyze the C18:2 into C18:3 by introducing third double bonds into the Δ15/ω-3 position of fatty acids. In order to investigate the function of PvFAD3, we expressed PvFAD3 in S. cerevisiae cells with exogenous C18:2 in the media, due to the yeast cells lack of C18:2 fatty acids (see Fig. S1B) (Reed, Schafer & Covello, 2000; Vrinten et al., 2005; Yin et al., 2018). As a result, under normal growth conditions (30 °C), yeast cells carrying PvFAD3 in culture, supplemented with C18:2, produced an additional peak, which we identified as C18:3 by GC-MS, while in yeast cells carrying the empty vector P426, the C18:3 was not detected (Fig. S1B; Table 1B). The amount of C18:3 in expressed PvFAD3 yeast cells reached by 1.58% of total lipids. Meanwhile, we noted that the yeast cells could easily absorb the exogenous C18:2, up to 59.43% of total lipids, but the content of C16:1 (8.02%) and C18:1 (8.21%) was significantly reduced as compared to the wild yeast cells (38.9% for C16:1 and 35.44% for C18:1), suggesting that the absorption of C18:2 followed a substantial inhibition of the biosynthesis pathway C16:1 and C18:1. In summary, this clearly showed that PvFAD2 and PvFAD3 genes from Sacha Inchi encode functional FA desaturase.

Many studies have found that the enzyme stability and catalytic efficiency of FAD2 and FAD3 in plants is often influenced by temperature (Li et al., 2007; O’Quin et al., 2010; Wang & Liu, 2014). To test this factor, we cultured yeast cells carrying the PvFAD2 or PvFAD3 gene under low (20 °C) and high temperature (30 °C) to compare changes in the proportion of FA composition. As shown in Table 1, when PvFAD2 transformed yeast cells were cultured under low temperature conditions, the content of C16:2 and C18:2 was changed but not significantly, decreasing from 4.25% to 3.53% for C16:2, and increasing from 9.34% to 10.18% for C18:2 relative to the 30 °C growth condition. In PvFAD3 transformed yeast cells, low temperatures promoted the biosynthesis of C18:3 in which the amount of C18:3 significantly increased by 48.7% from 1.58% to 2.35% when compared with high temperatures. These results showed that changes in temperature did not affect the activity of PvFAD2, whereas PvFAD3 protein were sensitive to temperature changes.

Functional analysis of PvFAD2 and PvFAD3 in tobacco seeds

To uncover the functional role of PvFAD2 and PvFAD3 in plant seeds, these two genes with the seed-specific promoters were transformed into tobacco via the leaf-disk transformation method. Transformants were selected by kanamycin, and further confirmed by PCR analysis as described in the ‘Materials and Methods’ section. In total, five independent transgenic lines of PvFAD2 and six independent transgenic plants of PvFAD3 were selected and cultivated in the greenhouse. Compared to wild type plants, the PvFAD2 and PvFAD3 were highly expressed in transgenic lines (Fig. S2). During plant growth, we observed no obvious phenotypic differences between transgenic and wild-type plants. At maturity, seeds from the T₀ generation of transgenic tobacco were collected to measure seed weight, oil content and fatty acid composition. As a result, we found no obvious change in seed weight and oil content in transgenic lines relative to wide-type plants, but the proportion of different fatty acids was significantly changed (Fig. 4). In transgenic PvFAD2 seeds, C18:2 content reached 70.5% of total lipids, and was significantly higher than that for wide-type plants (67.0% of total lipids), although the reduction in C18:1 content from 12.4% to 11.8% was not significant. We also observed that the proportion of C16:0 increased from 7.7% to 8.7%, whereas the proportion of C14:0 substantially decreased from 8.3% to 5.2%. In transgenic PvFAD3 seeds, the proportion of C18:3 was markedly increased by about 3.4 fold, from 0.7% in wild type to 3.1% in PvFAD3 expressing lines, while C18:2 content, a substrate of C18:3, was unchanged compared to wild-type lines (see Fig. 4). Similarly, the amount of C14:0 and C16:0 were substantially changed, with a significant decrease for C14:0 from 8.3 % to 5.1% and increase for C16:0 from 7.7% to 8.7% (Fig. 4). This result from transgenic tobacco seeds clearly showed that PvFAD2 and PvFAD3 make substantial contributions to the production of C18:2 and C18:3, respectively.