Physaria fendleri FAD3-1 overexpression increases ɑ-linolenic acid content in Camelina sativa seeds

Camelina (Camelina sativa) is an oil crop with a short growing period, resistance to drought and cold, low fertilizer requirements, and can be transformed using floral dipping. Seeds have a high content of polyunsaturated fatty acids, especially ɑ-linolenic acid (ALA), at 32–38%. ALA is an omega-3 fatty acid that is a substrate for eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) in the human body. In this study, ALA content was further enhanced by the seed-specific expression of Physaria fendleri FAD3-1 (PfFAD3-1) in camelina. The ALA content increased up to 48% in T2 seeds and 50% in T3 seeds. Additionally, size of the seeds increased. The expression of fatty acid metabolism-related genes in PfFAD3-1 OE transgenic lines was different from that in the wild type, where the expression of CsFAD2 decreased and CsFAD3 increased. In summary, we developed a high omega-3 fatty acid-containing camelina with up to 50% ALA content by introducing PfFAD3-1. This line can be used for genetic engineering to obtain EPA and DHA from seeds.


Results
FA composition analysis in seeds across T1 to T3 generations of PfFAD3-1 overexpressed camelina. PfFAD3-1 was overexpressed in camelina to enhance ALA production in seeds. The pBinGlyRed3 vector was used for seed-specific expression 41 . This vector contains a glycinin promoter, a seed-specific promoter, and DsRed (a red fluorescent protein) as a selection marker. PfFAD3-1 complementary DNA (cDNA) was cloned into pBinGlyRed3 vector (Fig. 2). The FA composition of the T2 seeds of 22 plants was analyzed and compared with that of the WT to observe changes in FA content in PfFAD3-1 OE lines (Fig. 3, Table S1). Transgenic plants are denoted as PfFAD3-1 OE numbers in the following description. In most transgenic plants, 18:3 content was increased compared to that in the WT, leading to a decrease in 18:2 content, which is a substrate of FAD3. The average 18:3 content in the WT was 41.4% and 45.5% in the transformant, indicating an increase of approximately 4%. Furthermore, the average 18:2 content in the WT was 15.8% and 8.7% in the transformant, showing a reduction of approximately 7%. The average 18:1 content was 12% in the WT but increased to an average of 16.9% in the transformant. The contents of other FAs did not significantly differ between the WT and transgenic plants (Fig. 3).
The total oil content of the seeds was measured as the total fatty acid methyl ester (FAME). The FAME content of WT was 272.5 μg/seed, and that of PfFAD3-1 OE #6-9, #19-5, and #22-1 lines was 251.8, 263.8, and 268.9 μg/ seed, respectively (Fig. 4b). In the PfFAD3-1 OE #19-5 and #22-1 lines, the total FAME (μg/seed) content was similar to that in the WT; however, the total FAME in the PfFAD3-1 OE #6-9 line decreased by 8% compared to that in the WT (Fig. 4b). www.nature.com/scientificreports/ In summary, the results of the FA analysis up to the T3 generation revealed that seed-specific overexpression of PfFAD3-1 enhanced 18:3 content in seeds and decreased 18:2 content. Moreover, the effect of PfFAD3-1 was confirmed to be maintained until the T3 generation.
Seed phenotype of PfFAD3-1 OE lines in T3 generation. Seed weight and size were analyzed to investigate the effect of overexpressing PfFAD3-1 on camelina seeds. Seed width, length, and size were measured in triplicate for 50 seeds per line (Fig. 5a). There was no significant difference in seed weight between the WT and PfFAD3-1 OE lines (Fig. 5b). However, the seed width and length of all three transgenic lines were significantly greater than those of the WT (Fig. 5c,d). The average seed width of the WT was 0.83 mm, whereas that of PfFAD3-1 OE #6-9, #19-5, and #22-1 lines was 0.85, 0.88, and 0.87 mm, respectively (Fig. 5c). The average seed length of the WT was 1.62 mm, whereas that of PfFAD3-1 OE #6-9, #19-5, and #22-1 lines was 1.73, 1.74, and 1.66 mm, respectively (Fig. 5d). Seed size was calculated by multiplying the width and length, which showed an increase in the transgenic lines compared with that in the WT (Fig. 5e).   www.nature.com/scientificreports/ selected to investigate whether the introduction of PfFAD3-1 affected the expression of endogenous FA synthesis genes in camelina (Fig. 6). Seed development was divided into three stages (S1, S2, and S3), and FA composition and expression of FA biosynthesis genes in the seeds were analyzed for the PfFAD3-1 OE #6-9-3, #19-5-2, and #22-1-3 lines of the T4 generation ( Fig. 6a,b).

Analysis of FA composition and
In the WT and three PfFAD3-1 OE lines, 16:0 FA and 18:0 FA continued to decrease from S1 to S3, showing no significant differences. The 18:1 content decreased from S1 to S2 and increased in the S3 stage, and the three transgenic plants showed higher content than the WT plants at all stages. The opposite trend was observed between the 18:2 and 18:3 contents; the 18:2 content decreased from S1 to S3, whereas the 18:3 content increased. The decrease and increase in 18:2 and 18:3 content, respectively, in the transformants gradually differed from those in the WT as seed development progressed. The 20:1 and 22:1 contents tended to increase from the S1 to S2 stages, and then slightly decreased at the S3 stage. However, there was no significant difference between the WT and transformants for either 20:1 FA or 22:1 FA (Fig. 6a).
The expression of FA biosynthesis genes in camelina was analyzed in WT and PfFAD3-1 OE plants at each stage of seed development (Fig. 6b). The expression of PfFAD3-1 was not detected in the developing seeds of the WT plants; however, PfFAD3-1 was highly expressed in the developing seeds of all three transgenic lines. CsFAD3, which has the same function as PfFAD3-1, showed different expression patterns between the WT and transgenic plants. At the S2 and S3 stages, the transgenic plants showed a 2.4-fold upregulation of CsFAD3 compared to the that in the WT. The expression patterns of CsSAD, CsFATB, and CsFAD2 showed the same tendencies in the WT and the transformants. However, the expression levels of CsSAD and CsFAD2 were 0.7-fold lower in transgenic plants than in WT plants at S2 stage. During the S2 stage, CsFATB expression was elevated 1.6-fold in the PfFAD3-1 OE #6-9-3 line compared to the WT; however, it showed a 0.4-fold decrease in the PfFAD3-1 OE #19-5-2 and #22-1-3 lines compared to that in the WT. CsKASII, CsFATA , and CsFAE1 showed different expression patterns in the WT and transgenic plants (Fig. 6b). In summary, 18:3 content increased and 18:2 content decreased in the developing seeds of transgenic plants compared to that in WT plants. This was due to

Discussion
In this study, PfFAD3-1, which converts 18:2 FA to 18:3 FA, was transformed into camelina (Fig. 1). Consequently, the ALA (18:3) content increased by up to 48% in the T2 seeds and up to 50% in the T3 seeds of the transgenic camelina plants (Figs. 3, 4). The PfFAD3-1 OE lines showed an average increase of 12.3-13.7% in 18:3 content and a decrease of 13.0-13.7% in 18:2 content (Table S3). This similarity in the increase of 18:3 and decrease of 18:2 suggests that the conversion of 18:2 FA to 18:3 FA in camelina by PfFAD3-1 is both specific and efficient. In a previous study, when PfFAD3-1 was transformed into soybeans under the control of the seed-specific promoter, β-conglycinin, the ALA content increased by up to 52% in T1 seeds and up to 42% in T2 seeds 26 . In soybean, PfFAD3-1 expression increased ALA content under the control of the phaseolin promoter and 35S promoter; however, the increase in ALA content was the greatest under the control of β-conglycinin 26 . PfFAD3-1 expression in soybeans did not result in significantly different oil content from that of the WT; however, it increased both seed size and weight 26 . Similarly, in this study, the seed oil content of PfFAD3-1 OE #6-9 line was slightly lower than that of the WT, whereas that of PfFAD3-1 OE #19-5 and #22-1 lines was similar to that of the WT (Fig. 4b).
In the PfFAD3-1 OE #6, #19, and #22 lines, some plants in the next generation exhibited lower ALA content than those in the previous generation, as indicated in Table S2. This may be due to the presence of non-homozygous plants or variation in transgene expression among individuals. Nonetheless, the average ALA content of the progeny of these lines ranged from 46.1 to 49.8%, which was similar or slightly higher than that of the previous generation (Table S2). These findings suggest that the transmission and expression of the PfFAD3-1 transgene www.nature.com/scientificreports/ is stable and heritable in subsequent generations. Seed weight did not differ between the WT and the three transgenic lines. However, seed width, length, and size increased compared with those of the WT (Fig. 5b-e). This showed a seed phenotype similar to that observed when PfFAD3-1 was expressed in soybean 26 . Hence, the seed-specific expression of PfFAD3-1 in camelina did not significantly affect seed weight but increased seed size. Since seed size increased without a significant effect on seed oil content, further research is required to determine whether ALA content affects seed size. Camelina seed oil can be used for various purposes such as baking, frying, and salad dressings 42 . The camelina developed in this study is beneficial for health because it contains up to 10% or more omega-3 FA compared to that in the WT. Camelina seeds contain eicosenoic acid (20:1) and erucic acid (22:1) at 12-17% and 2-3%, respectively 33 . The 20:1 and 22:1 FAs were synthesized from 18:1 by FAE1 43 . Camelina was identified as having three copies of FAE1 genes 39 . When three copies of FAE1 were knocked out in camelina using the CRISPR/Cas9 technology, 20:1 and 22:1 synthesis was blocked, thus increasing the 18:1 content and continuously increasing the ALA content up to 50% 44 . In this study, PfFAD3-1 expression increased 18:3 content by up to 50% (Fig. 4). Therefore, knocking out CsFAE1 using CRISPR/Cas9 in the PfFAD3-1 overexpression line will further increase the ALA content by over 50%. In addition, under PfFAD3-1 expression, the 18:2 content was 2.6-3.4%, indicating that PfFAD3-1 sufficiently synthesized 18:3 using 18:2 as a substrate. Therefore, if the FAD2 gene is additionally overexpressed in the PfFAD3-1 overexpression line, the flux from 18:1 to 18:2 would increase; thus, PfFAD3-1 could use 18:2 to further enhance ALA content.
Since humans cannot synthesize ALA, it must be obtained through dietary sources, such as rapeseed or walnuts, which have high ALA content 45 . This study successfully developed a high-ALA camelina that could be used as a source of plant oil for ALA intake. From a clinical standpoint, increasing the ALA intake is important because it has been shown to reduce the risk of incident stroke 46 . While there are suggestions that ALA may have positive effects on cardiovascular health, the results are conflicting, indicating the need for further research in this area 47 . In addition to its direct health benefits, ALA can be converted into EPA and DHA in the human body, which also have beneficial effects. However, the efficiency of this conversion process is typically low 8 . To address this, research has been conducted to develop crops that can produce EPA and DHA directly from ALA through the expression of the desaturase and elongase enzymes involved in the synthesis of these FAs [48][49][50] . Recently, a transgenic camelina strain was developed using multigene metabolic engineering to produce 12% DHA in seeds 49 . In addition, this line was crossed with CsFAE1 knockout camelina containing high ALA content in seeds, further increasing omega-3 FA content (ETA, EPA, DPA, and DHA) from 27 to 33% 51 . Therefore, the high-ALA camelina developed in this study can be used to produce omega-3 FAs such as EPA and DHA. www.nature.com/scientificreports/

Materials and methods
Vector construction of PfFAD3-1 gene. cDNA was synthesized using mRNA from developing seeds of P. fendleri using a PrimeScript 1st strand cDNA synthesis kit (Takara, Japan). P. fendleri seed WCL-LY1 (PI596362) was provide by the U.S. National Plant Germplasm System (NPGS). The synthesized cDNA was used Error bars represent SD. Statistical analysis was performed using one-way ANOVA with Tukey's multiple comparison test (*p < 0.05, **p < 0.01, ***p < 0.001). www.nature.com/scientificreports/ as a template, and polymerase chain reaction (PCR) was performed using primers targeting the PfFAD3-1 open reading frame (Table S4). The PfFAD3-1 sequence was obtained from a previous study 14 . The PCR product was eluted using a PCR purification kit (Cosmogenetech, Korea) and cloned into the pGEMT-Easy vector (Promega, USA). Thereafter, we checked whether there was a PCR error using Sanger sequencing and proceeded to the next step. There is an EcoRI enzyme site on both sides of the pGEMT-Easy vector and a multi-cloning site, including the EcoRI enzyme site, in the pBinGlyRed3 vector equipped with a seed-specific glycine promoter. Therefore, it was cut with EcoRI enzyme, and the PCR product was ligated into the pBinGlyRed3 vector.
Selection of PfFAD3-1 transformed camelina. Camelina sativa cv. Suneson was used for all the experiments in this study. Camelina seeds were provided by Dr. Edgar Cahoon (University of Nebraska-Lincoln, Lincoln, Nebraska, USA). Camelina seeds were germinated on filter paper containing water in a culture chamber at 16 °C under a 16 h light/8 h dark photoperiod. After one week, the seeds from which the cotyledons emerged were transplanted into the soil. Eighteen plants were grown in soil, and transformation was performed using the floral dipping method 34 . The Agrobacterium strain GV3101 was used. Agrobacterium cells (500 μl) and kanamycin antibiotics (500 μl) were added to 500 ml LB medium and incubated overnight at 28 °C until an optical density value of 1 was obtained. This solution was centrifuged at 3000 × g for 10 min and resuspended in 500 ml solution containing 1% sucrose and 0.05% Silwet L-77. The camelina flowers were then immersed in this solution for 1 min. After floral dipping, plants were placed in a growth chamber for 24 h in the dark. This procedure was performed three times at intervals of 5 d. After harvesting the transformed plants, 22 fluorescent seeds were selected using a green flashlight and a red filter. After germination, these 22 T1 lines and the WT were grown in growth chamber at 20 °C under a 16 h light/8 h dark photoperiod to obtain the next generation. All procedures, including the collection of plant materials and experimental research, were conducted in accordance with institutional, national, and international guidelines.

Analysis of gene expression in PfFAD3-1 OE lines. Reverse transcription real-time PCR (RT-qPCR)
was performed on RNA extracted using the method described in the reference 52 . Developing seeds were taken at the beginning of the desiccation phase (30,40, and 50 days after flowering) based on reference 53 . After RNA isolation, 2 μg cDNA was synthesized using a cDNA synthesis kit (Takara, Japan). RT-qPCR analysis was performed using the synthesized cDNA as a template and SYBR Green Master Mix (TOYOBO, Japan) in a StepO-nePlus Real-Time PCR System (Thermo Fisher Scientific, USA). Camelina ACTIN2 was used as an endogenous gene control, and primers capable of targeting all three copies of the genes were designed and used in this study (Table S4).
Fatty acid analysis. Seven camelina seeds were subjected to GC. Seeds were placed 500 μl of toluene and 1 ml of 5% H 2 SO 4 including pentadecanoic acid (15:0) standard (100 μg/ml) and allowed to react for 2 h in an 85 °C water bath. When the reaction was complete, 1 ml of 0.9% NaCl and 1 ml of hexane were added, mixed thoroughly, and centrifuged at 330 × g for 2 min. The supernatant was transferred to a 6 ml tube. The procedure of adding hexane and centrifuging was repeated three times. A total of 3 ml of the supernatant was purged with nitrogen gas, dissolved in 200 µl hexane, and transferred to a GC vial. A DB-23 column (30 m × 0.25 mm, 0.25 μm film, Agilent, USA) was used, and analysis was performed using a GC-2030 instrument (Shimadzu, Japan). The GC oven temperature was increased from 190 to 230 °C at a rate of 5 °C per min.
Analysis of seed phenotype. Using the WT as a control, seed weight, length, width, and size were measured for the transgenic line. Seeds from each line were analyzed in triplicate to determine the seed phenotype. Seed photographs were captured using an optical microscope (NIKON, Japan). Seed weight was measured using an electronic scale (OHAUS, USA). Seed length and width were measured using ImageJ software, and seed size was calculated by multiplying the seed length and width.
Statistical analysis for data. Statistical analyses were performed using one-way analysis of variance (ANOVA) and Tukey's multiple comparison test using GraphPad Prism to determine the significance of the differences. Different letters indicate significant changes compared to the WT plants in the FA and seed phenotype analyses. Significant differences in FA composition and gene expression patterns at different seed developmental stages are indicated by *p < 0.05, **p < 0.01, and ***p < 0.001.

Data availability
All data generated or analyzed in this study are included in this published article (and its supplementary information files). www.nature.com/scientificreports/