Development of marker-free transgenic Jatrophaplants with increased levels of seed oleic acid

Identification of FAD2genes

The first step towards generating high oleic acid Jatropha was to determine the gene(s) which encode putative microsomal delta 12 fatty acid desaturase. For this, we isolated three cDNAs which possess extensive similarity to extant FAD2 genes from a J. curcas seed cDNA library (ZC Yin et al., unpublished data). One of the cDNA had a mutation that generated a premature stop codon, rendering it non-functional. The other two cDNAs encoded proteins of 383 and 387 amino acids that were 74% identical to each other and 77.3% and 72.1% identical to Arabidopsis FAD2, respectively (Additional file 1). We further confirmed that there are at least three FAD2 genes in the Jatropha genome through the data mining of both deep sequencing data of Temasek Life Sciences Laboratory (H Yan, unpublished data) and the Jatropha genome database available online [11]. The cDNA which encoded a higher amino acid sequence identity to the FAD2 enzyme family was designated as JcFAD2-1 [GenBank: JN544421] and another one was named JcFAD2-2 [GenBank: JN544422]. The non-functional one shared higher similarity with JcFAD2-2 and was named as JcFAD2-2m. JcFAD2-1 has identical amino acid sequences to Arabidopsis thaliana FAD2 at its enzyme active center in three conserved His-rich boxes (data not shown), while JcFAD2-2 has a variation on a key residue Ala in active site His-rich box 3 (Thr in JcFAD2-2; Additional file 2). The change of small hydrophobic Ala substituted with polar Thr could potentially alter FAD2-2 substrate specificity and enzyme activity because the hydrophobic core environment is crucial for its activity [12].

To investigate the gene expression patterns of FAD2-1 and FAD2-2, RNA was extracted from different stages of seed development (three weeks, five weeks, seven weeks and eight weeks after fertilization, corresponding to the early, middle, late and mature stages of Jatropha seed) and used in RT-PCR reactions containing primers specific for each cDNA. As shown in Additional file 2, the expression of FAD2-1 gene is high in seeds and weaker in vegetative tissues, while the FAD2-2 gene is expressed highly in seeds and not detectable in leaf tissue. The expression pattern of these two FAD2 genes in Jatropha is very similar to others in the Euphorbiaceae: FAD2 and Fatty acid hydroxylase 12 (FAH12) in castor bean [13], FAD2 and FADX in the tung tree (Aleurites fordii) [12]. All the above data suggest that JcFAD2-2, similar to FADX and FAH12, may function more like an unusual fatty acid enzyme than a desaturase. We have previously identified that JcFAD2-1 mediates the conversion of oleic acid to linoleic acid in Jatropha leaf by virus-induced gene silencing (VIGS). Therefore, in this study we chose FAD2-1 as our target to knockdown for the development of high oleic acid composition.

Generation of marker-free FAD2-1i RNAi transgenic Jatrophausing a constitutive promoter and analysis of oleic content in T0 leaves

Considering the possible large-scale commercial planting of such transgenics as Jatropha, we aimed to obtain transgenic plants free of an antibiotic selection marker. To accomplish this, we used a chemical inducible Cre-lox-mediated site-specific recombination system which had been shown to work in Arabidopsis [14, 15]. To produce transgenic Jatropha plants with high oleic acid content we used RNAi technology to silence JcFAD2-1 expression by targeting an 862 bp coding region that shares 72.9% identity with JcFAD2-2. We first obtained hygromycin-resistant regenerated shoots. Once the small shoots (2 to 3 mm) came out, they were immediately transferred to β-estradiol induction medium without hygromycin to induce marker excision. After two weeks of induction, shoots were transferred to regeneration medium II without hygromycin. Figure 1A shows the chemical inducible RNAi construct used to silence the JcFAD2-1 gene. Upon induction, Cre-lox mediated recombination excises the hygromycin gene-containing DNA fragment. Only this recombination event happened; the JcFAD2-1 hairpin structure is situated immediately downstream of the G10-90 promoter (pX7-FAD2-1i). The G10-90 promoter drives constitutive expression and has higher expression levels than does the cauliflower mosaic virus 35S promoter in Arabidopsis. It drives the constitutive expression of JcFAD2-1 hairpin RNA, which is further processed into JcFAD2-1 specific siRNA. Finally, these siRNA guide the degradation of JcFAD2-1 transcripts.

We randomly selected ten regenerated shoots of pX7-FAD2-1i and extracted their genomic DNAs for genotypic analysis. PCR analysis revealed a smaller fragment of expected size in two out of ten regenerated shoots consistent with successful marker excision (Figure 1B). One of the lines, #1-1, had a band corresponding to the hygromycin phosphotransferase gene (HPT) suggesting that the shoot was chimeric marker free. By contrast, no band corresponding to the HPT gene was detected in the other line, #1-2, suggesting that line #1-2 might be a pure marker-free transgenic Jatropha (Figure 1B).

In Arabidopsis, FAD2 encodes the desaturase responsible for the introduction of a second double (Δ¹²) bond in oleate. To confirm that RNAi-mediated suppression of FAD2-1 should block the conversion of 18:1 to 18:2 fatty acids in Jatropha, we used gas chromatography (GC) of fatty acid methyl ester (FAME) to analyze leaf fatty acid profiles. As predicted, there was a higher oleic acid content in line #1-1 and a much higher level of oleic acid in line #1-2 compared with regenerated shoots from control (untransformed) cotyledons (Figure 1C). The linoleic acid level was significantly reduced in lines #1-1 and #1-2, suggesting that suppression of FAD2-1 increased the oleic acid level at the expense of linoleic acid and that the JcFAD2-1 encoded enzyme mediates conversion of oleic acid to linoleic acid.

Characterization of oleic acid content in seeds and leaves of T1 pX7-FAD2-1 transgenic RNAi lines

Using PCR analysis, we identified 20 more putative pure marker-free X7-FAD2-1 RNAi lines (partial genotyping results shown in Additional file 3). These transgenic lines were grown to maturity in a greenhouse. T1 seeds were collected and endosperms were carefully separated from the embryos germinated on hormone-free medium. GC analysis showed that oleic acid content in the best two lines, X7#79 and X7#170, increased to 50% to 60% of the total fatty acid in contrast to the 36.7% in control (35S:GFP) endosperms (Figure 2A). Correspondingly, the linoleic acid levels reduced to less than 25% from the original 41% in control endosperms (Figure 2A). We noted that the increase of the oleic acid level was moderate and not as high as that found in the initial T0 leaves (#1-2 in Figure 1C) or in Jatropha leaves treated with Tobacco Rattle Virus-induced FAD2-1 RNAi during our previous VIGS studies [8]. One possibility could be the weak expression of the G10-90 promoter in Jatropha seeds. This assumption was confirmed by RNA analysis using quantitative RT-PCR, which showed that the two transgenic lines still expressed FAD2-1 transcripts at 20% of the control level (Figure 2B). It also showed that the FAD2-1 suppression was gene-specific as there was no effect on the expression of FAD2-2 in the endosperms of these two transgenic lines (Figure 2B). We also analyzed the FAD2-1 RNAi effect in vegetative organs of T1 plants. There was considerable reduction of FAD2-1 expression levels in T1 leaves as well (Figure 2C).

High oleic acid transgenic lines with seed-specific promoter

Since the G10-90 promoter was not very effective in seeds, we replaced it in the pX7 vector with the soybean (Glycine max) 7S (which encodes a seed storage protein) gene promoter which displays seed-specific expression [16]. The new vector with the soybean 7S promoter was named pX8-FAD2-1i (Figure 1A). We generated 30 X8-FAD2-1i lines which were confirmed to be pure marker-free or partial marker-free (partial genotyping results shown in Additional file 3). Quantitative RT-PCR analysis showed that line X8#34 and X8#291 contained only 0.7% and 1.1% JcFAD2-1 transcript in the endosperm compared to that of 35S: GFP control lines (Figure 3A). Since soybean 7S promoter also has high activity in seed cotyledon of soybean, we checked the JcFAD2-1 transcript level in the cotyledons of T1 transgenic Jatropha. Quantitative RT-PCR results suggested that the soybean 7S promoter was also active in Jatropha cotyledons as much lower levels of FAD2-1 transcript accumulated in T1 cotyledons compared with that of the control (Figure 3B). As expected, there was no significant change of FAD2-1 transcript levels in vegetative organs such as leaves (Figure 3C).

GC analysis showed that oleic acid accumulation was as high as 77.4% and 74.7% of the total fatty acid content in the T1 endosperm of lines #34 and #291, respectively (Figure 3D; Additional file 4). The linoleic acid levels reduced to less than 5% of the total fatty acid content in these lines. Moreover, the stearic acid level also slightly reduced from 7.7% to 5.4% to 5.7% of the total fatty acid in these transgenic Jatropha. There was no marked difference in C16 fatty acids composition between transgenic lines and control plants. The total unsaturated fatty acids (oleic and linoleic) in control and transgenic Jatropha endosperms was estimated to be about 78% to 79% of the total fatty acid. In lines #34 and #291, almost all of the unsaturated fatty acids was stored as oleic acid. There was no obvious difference in the total oil content between line #34 endosperm and control endosperm (Figure 3E). Consistent with no changes on gene expression level in leaf (Figure 3C), there was no significant difference in fatty acid profile of true leaves between X8-FAD2-1i lines and controls (Figure 3F,Additional file 4). These data further confirmed the seed-specific high oleic acid accumulation in transgenic lines.

To further analyze the possible side-effect of the FAD2-1 RNAi on the fatty acid biosynthesis pathway, we determined endosperm transcript levels of several key fatty acid genes such as KASII, FATB and SAD. No significant difference between transgenic lines (#34 & #291) and control was found with respect to the expression of these genes implicated in fatty acid biosynthesis (data not shown). This was in agreement with the similar level of oil content between transgenic and control endosperms (Figure 3E).

We performed Southern blot analysis on line #34 to determine the complexity of the transgenic locus. Using PCR-based genotypic analysis, we observed that line #34 was a chimeric marker-free plant (Additional file 3). There is only one XhoI site in the pX8-FAD2-1i vector (Figure 1A). In case of a single copy insertion in line #34, we would expect two bands with a size difference of around 5.7 kb due to the partial Cre-lox recombination event. Therefore, the total genomic DNAs of T0 and T1 plants were digested with XhoI and probed with soybean 7S promoter. The Southern blot data (Figure 4A) showed two bands with a size difference of around 5 to 6 kb in #34 T0 plants and these two bands segregated in T1 plants (lines #34-1 to #34-4). T1 plants #34-2 and #34-4 contained a single copy of the transgene and should have been marker free, whereas #34-1 was a chimera and #34-3 still contained the marker. To analyze whether #34-2 and #34-4 were indeed marker free, we further treated the total genomic DNA of these two T1 plants with EcoRV and XbaI, and hybridized it with FAD2-1 probe. Beside the endogenous 5 kb band from the JcFAD2-1 genomic locus according to its genomic DNA sequence, an extra band was found in plants #34-2 and #34-4 (Figure 4B) which was absent in Jatropha curcas MD isolate (Jc-MD) wild-type control plant. We stripped the membrane and hybridized it with an HPT probe. No signal was detected in the DNA derived from these two transgenic plants (Figure 4B). These results confirmed that #34-2 and #34-4 plants were indeed marker-free transgenic Jatropha.

Effect of oleic acid manipulation on agronomic traits of the plant

Explant material for transformation

Seeds were obtained from the J. curcas (Jc-MD) elite plants which were pre-selected by Y Hong and C Yi [17]. The seeds were germinated on^1/2Murashige and Skoog salt medium. Cotyledons were harvested from five- to seven-day-old seedlings, cut into small pieces (5 × 5 mm) and used as explants.

Jatrophatransformation procedure

The detailed transformation protocol has previously been described [28]. In brief, the protocol consisted of four steps; co-cultivation, shoot regeneration, shoot elongation and rooting.

Transgenic plasmid construction and materials

The pCAMBIA1300-derived vector which carried a 35S-GFP gene cassette was used for Jatropha transformation. Transgenic lines (called 35S:GFP lines) were confirmed using gene markers, fluorescence and protein gel analysis.

To generate the β-estradiol chemical-regulated inducible JcFAD2-1 RNAi lines, we used a gene-specific 862-bp fragment corresponding to the coding region (nt 85 to 946) of the JcFAD2-1 cDNA. This cDNA fragment was PCR-amplified with forward primer 5'-ATCACTCGAGCCACCATTCACACTTGGTCAG-3' and reverse primer 5'-GTATAAGCTTCATGAGTGTCTGTAATGTTATG-3'. The PCR fragment was inserted in the sense orientation into the XhoI/HindIIII sites of a pSK-int vector as described previously [15]. Another fragment, amplified with forward primer 5'-CAATAACTAGTACCATGGGTGCCGGTGGCAGAATG-3' and reverse primer 5'-TATTGGATCCGGAAACTTGTTTTTGTACCAGAACAC-3', was subsequently placed in the antisense orientation into the BamHI/SpeI sites of pSK-int-sense FAD2-1 to form pSK-int-FAD2-1 RNAi. Finally, the entire RNAi cassette comprising the sense and antisense fragments interspersed by the actin II intron was excised from pSK-int using the flanking XhoI/SpeI sites and inserted into the XhoI/SpeI site of pX7-GFP vector, yielding the construct pX7-FAD2-1i.

To generate seed-specific JcFAD2-1 RNAi lines, we replaced the G10-90 promoter in pX7-GFP with the soybean seed storage protein 7S gene promoter. The resulting vector was called pX8-GFP. The entire FAD2-1 RNAi cassette in pSK-int vector was inserted into pX8-GFP to substitute the GFP coding region and hence the construct pX8-FAD2-1i was formed.

Transformants were selected using PCR-based genotypic analysis with either P1 and P4 primers (for X7-FAD2-1 RNAi lines) or P7S and P4 primers (for X8-FAD2-1 RNAi lines), together with primer pairs for the HPT gene. Additional file 5 shows the sequence information of these primers. Independent lines (X7#79, X7#170 from X7-FAD2-1i; X8#34, X8#291 from X8-FAD2-1i) were confirmed by the analysis of the expression of FAD2-1 and fatty acid composition in endosperms of individual seeds. Plants were grown in a greenhouse under natural photoperiods and ambient temperature (ranged from 25 to 35°C) in Singapore.

Fatty acid analysis

Total lipid was extracted from 100 mg fresh Jatropha leaves as previously described [8]. The outer seed coat was removed from dried Jatropha seeds. The seeds were surface sterilized for 1 minute with 75% (v/v) ethanol, immersed in 10% (v/v) H₂O₂ for 1 hour, rinsed with sterile water two times, and finally immersed in sterile water overnight at 28°C in darkness for 24 hours. The seed endosperm was carefully separated from the embryo. The dry endosperm part was ground to fine powder and the lipids were extracted with hexane three times. The combined supernatant was transferred to a glass vial and the hexane was evaporated with a flow of dry nitrogen gas at 50°C. The weight of the raw oil was determined and the oil content was recorded as the ratio of raw oil to dried endosperm weight.

About 10 mg of lipid was transmethylated with 3N methanolic-HCl (Sigma, St. Louis, MO, USA) plus 400 μL 2,2, dimethoxypropane (Sigma). The resultant FAMEs were separated by GC and detected using GC Agilent 6890 (Agilent, Santa Clara, CA, USA) employing helium as the carrier gas and DB-23 columns for components separation. The GC analysis was performed at 140°C for 50 seconds and 30°C per minute ramp to 240°C, and the final temperature was maintained for 50 seconds. Peaks were identified based on their retention times compared with a FAME reference mixture (Sigma). The fatty acid composition value included in the analyses was calculated based on the peak area percentage of total fatty acids in three biological replicates. The data were presented as means ± standard deviations.

RNA extraction and analysis

The 100-mg leaf or endosperm samples were ground to fine powder in liquid nitorigen and extracted with plant RNA purification reagent (Invitrogen, Carlsbad, CA, USA). RNA concentration was measured by Nanodrop (Thermo Scientific, Wilmington, DE, USA). Moloney Murine Leukemia Virus Reverse Transcriptase (Promega, Madison, WI, USA) was used for RT reactions. Real-time PCR was performed with Power SYBR ^® Green PCR Master mix (Applied Biosystems, Foster City, CA, USA) and run in ABI7900HT. All samples were run in triplicate and the data were analyzed with RQ manager at a pre-set threshold cycle value (Applied Biosystems). The Jatropha rbcL transcript served as an internal control for leaf RNA samples while the Jatropha α-tubulin transcript served as an internal control for seed RNA samples. Threshold cycle values included in the analyses were based on three biological replicates, with three technical replicates for each biological sample. Standard deviation was calculated based on the three biological replicates. See Additional file 5 for PCR primer sequences.

Southern blot analysis

Total genomic DNA was isolated from the leaves of glasshouse-grown transgenic or control plants by the cetyltrimethylammonium bromide method [29]. Genomic DNA was digested with restriction enzymes and separated on 0.8% agarose gels. The gels were processed and transferred to a nylon Hybond-N⁺ membrane (GE Biosciences, Buckinghamshire, UK) following standard procedures [30]. Membranes were hybridized with 7S promoter, HPT or FAD2-1 open reading frame probes. The probes were labeled with [α-32P]-deoxycytidine triphosphate by random prime synthesis using Amersham Rediprime II Random Primer Labelling System (GE Biosciences). Hybridization was performed overnight at 42°C using the ULTRAHyb-Oligo hybridization buffer (Ambion, Austin, TX, USA) and signals were detected by autoradiography.

Plant growth condition, agronomic traits collection and statistical analysis

All the transgenic or control plants were grown in a biosafety level 2 greenhouse according to standard practices. The transgenic and control plants were transplanted into pots (diameter 30 cm) and randomly placed in a greenhouse with a space of 1 m × 1 m for each tree in Temasek Life Sciences Laboratory, Singapore. Plant management, such as fertilization, pesticides spraying, watering and artificial fertilization, was carried out according to normal practices. Half a year after transplanting, the plants were pruned at the height of 50 cm from the ground. The tree height, diameter of main stem and number of primary branches were measured and recorded.

Every fruit and seed of these plants was collected and counted and their weight was taken over the whole experiment. Dry T1 seeds were weighed and endosperms were further analyzed for oil content and oil profile. Four T0 plants with > 70% oleic acid content were grouped as high-oleic-acid plants and the other 16 T0 plants with normal oleic acid content (37% to 42%) were grouped as non-high-oleic-acid plants. The data were analyzed with Student's t-test and presented in Table 1 as means ± standard errors.