Isolation and functional analysis of fatty acid desaturase genes from peanut (Arachis hypogaea L.)

Background Fatty acid desaturases are enzymes that introduce double bonds into fatty acyl chains. Extensive studies of fatty acid desaturases have been done in many plants. However, less is known about the diversity of this gene family in peanut (Arachis hypogaea L.), an important oilseed crop that is cultivated worldwide. Results In this study, twelve novel AhFADs genes were identified and isolated from peanut. Quantitative real-time PCR analysis indicated that the transcript abundances of AhFAB2-2 and AhFAD3-1 were higher in seeds than in other tissues examined, whereas the AhADS and AhFAD7-1 transcripts were more abundant in leaves. AhFAB2-3, AhFAD3-2, AhFAD4, AhSLD-4, and AhDES genes were highly expressed in flowers, whereas AhFAD7-2, AhSLD-2, and AhSLD-3 were expressed most strongly in stems. During seed development, the expressions of AhFAB2-2, AhFAD3-1, AhFAD7-1, and AhSLD-3 gradually increased in abundance, reached a maximum expression level, and then decreased. The AhFAB2-3, AhFAD3-2, AhFAD4, AhADS, and AhDES transcript levels remained relatively high at the initial stage of seed development, but decreased thereafter. The AhSLD-4 transcript level remained relatively low at the initial stage of seed development, but showed a dramatic increase in abundance at the final stage. The AhFAD7-2 and AhSLD-2 transcript levels remained relatively high at the initial stage of seed development, but then decreased, and finally increased again. The AhFAD transcripts were differentially expressed following exposure to abiotic stresses or abscisic acid. Moreover, the functions of one AhFAD6 and four AhSLD genes were confirmed by heterologous expression in Synechococcus elongates or Saccharomyces cerevisiae. Conclusions The present study provides valuable information that improves understanding of the biological roles of FAD genes in fatty acid synthesis, and will help peanut breeders improve the quality of peanut oil via molecular design breeding.

Introduction Peanut (Arachis hypogaea L.) is an allotetraploid species (2n = 4x = 40, AABB) cultivated widely in tropical, subtropical and temperate regions [1]. The peanut seed is comprised of around 50% oil, of which approximately 80% is oleic (36-67%) and linoleic (15-43%) acids [2]. A high O/L ratio (ratio of oleic and linoleic acid) is the most desired oil quality trait as it increases shelf life and the health benefits to manufacturers and consumers, respectively [3]. Improvements in peanut oil content and quality traits (high oleic and low linoleic acid) could be accelerated by investigating the fatty acid biosynthesis pathway.
Fatty acid desaturases are responsible for the insertion of double bonds into pre-formed fatty acid chains, and play an essential role in fatty acid metabolism and the maintenance of biological membranes in living organisms [4]. They belong to a large gene family that contains conserved histidine regions. Histidine-rich boxes are thought to form a part of the diiron center where oxygen activation and substrate oxidation occur [5,6].
To date, several types of fatty acid desaturase genes have been cloned and characterized from peanut, including AhFAB2-1, AhSLD-1, four microsomal AhFAD2 and two chloroplast AhFAD6 genes [17,18]. Among them, FAD2 is the most well-studied fatty acid desaturase gene. Two microsomal oleoyl-PC desaturase genes (AhFAD2-1A and AhFAD2-1B), each having its origin in different diploid progenitor species, have been isolated from cultivated peanut [19][20][21]. Reduction in transcript levels or inactivation of both genes is required to produce high O/L genotypes [22][23][24][25][26]. Different types of DNA markers from these two genes have been developed to facilitate marker-assisted selection for the high-oleate trait [27][28][29][30]. The functions of AhFAD2-2 and AhFAD6 were verified by heterologous expression in S. cerevisiae. Linoleic acid (18:2), normally not present in wild-type yeast cells, was detected in transformants of these two genes [17].
The cultivated peanut (Arachis hypogaea L.) is derived from two wild diploid species Arachis duranensis (A genome) and Arachis ipaensis (B genome) [31,32]. Now the genome sequences of Arachis duranensis and Arachis ipaensis were released. The availability of these genomes will lead to further advances in knowledge of genetic changes since the very recent polyploidization event that gave rise to cultivated peanut and to an expanding knowledge of understudied areas of plant biology [33,34]. In this study, we isolated twelve novel desaturase genes from cultivated peanut. The expression patterns of these genes were investigated in different tissues and at different stages of seed development. The expression of FAD genes in response to abiotic stress and abscisic acid (ABA) was also analyzed. Additionally, the functions of AhFAD6 and AhSLDs were confirmed by heterologous expression in Synechococcus elongatus (strain PCC 7942) or yeast (Saccharomyces cerevisiae). Our results indicated that these two types of genes are strong candidates for modifying fatty acid biosynthesis in peanut.

Isolation of FAD genes from peanut
In a previous study, four fatty acid desaturases were isolated from peanut. These were AhFAB2-1, AhFAD2-2, AhFAD6, and AhSLD-1 [17]. In this study, twelve new genes that probably encode FAD proteins were found using Bioedit software [35]. They were cloned and designated as AhFAB2-2, AhFAB2-3, AhFAD3-1, AhFAD3-2, AhFAD4, AhADS, AhFAD7-1, AhFAD7-2, AhDES, AhSLD-2, AhSLD-3, and AhSLD-4, according to the homologous genes identified in Arabidopsis (Table 1). Among the twelve genes, six had complete open reading frames in the peanut cDNA library and were cloned by conventional RT-PCR. However, six genes were cloned using the rapid amplification of cDNA ends (RACE) method. The open reading frames of these genes ranged from 924 bp to1371 bp in length, and encoded 307 to 456 amino acids (Table 1 and S1 Table). The sequence information for the twelve genes was submitted to Genbank along with their identification numbers (Table 1).
The twelve desaturase genes contained typical histidine-rich boxes (S1-S3 Figs), which was in accordance with the standard for different types of desaturase genes. For example, the two histidine-boxes of AhFAB2-1, AhFAB2-2, and AhFAB2-3 genes were consistent with those of plastidial stearoyl-ACP desaturases, which are represented as EENRHG and DEKRHE (S1 Fig). The three histidine-boxes of four microsomal Δ15 fatty acid desaturases (AhFAD3-1, AhFAD3-2A, AhFAD3-2B, and AhFAD3-2C) and three plastidial Δ15 fatty acid desaturases (AhFAD7-1A, AhFAD7-1B, and AhFAD7-2) genes matched the standards for Δ15 desaturase, i.e. LGHDCGH, HR(K)THH, and HVIHH. The third histidine box of AhSLD-1, AhSLD-2, AhSLD-3, and AhSLD-4 contained a His to Gln substitution at the third histidine residue, which is also found in several fatty acid desaturases, such as the plant and animal fatty acid Δ5and Δ6-desaturases [13,36]. Furthermore, in common with other desaturases of this type, four AhSLD genes encoded proteins with a cytochrome b 5 -like hem-binding domain at the N-terminus. The presence of this binding domain was characterized by the His-Pro-Gly-Gly motif, which indicated that this putative desaturase existed as a fusion protein. Sphingolipid Δ4 desaturase activity has evolved independently of sphingolipid Δ8-desaturase activity [16]. The sphingolipid Δ4 desaturases shared only limited similarity with other proteins characterized by the histidine box motifs (LAIHELSH, HLEHH, HNEHH), and they didn't contain a cytochrome b5 domain (S2 Fig). FAD4 encodes a predicted integral membrane protein that appears to be unrelated to classic membrane bound fatty acid desaturases, based on overall sequence conservation. Its inferred primary sequence has little resemblance to that of known fatty acid desaturases beyond the presence of histidine motifs and membrane-spanning domains. However, the FAD4 protein contains two histidine motifs resembling those of fatty acid desaturases [37]. The AhFAD4A and AhFAD4B protein sequences contained two histidine motifs, HAWAH and HSAHH (S3 Fig), whose sequence and spacing are reminiscent of, but not identical to, conserved motifs in membrane-bound desaturases [38]. While AhFAD4A and AhFAD4B contained a third histidine motif (QGHH; S3 Fig), its sequence diverged from the third histidine motif present in membrane-bound desaturases. Although glutamine substituted for histidine in these motifs [36], known membrane-bound desaturases typically have two-to-three amino acids between the glutamine and the histidines [39]. Another difference is that in characterized membranebound desaturases, two histidine motifs are located between the second and the third membrane spanning domains, while the third is located at the C-terminus of the protein. In AhFAD4A and AhFAD4B, only one proposed histidine motif is located between membrane spanning domains II and III, while the other two are located at the C-terminus (S3 Fig). Based on these sequence features, it seems likely that FAD4 is a metalloenzyme that evolved independently from the characterized desaturases [37].
The polyunsaturated fatty acids are synthesized by two distinct pathways in plants, known as the prokaryotic and eukaryotic pathways, which are located within the membrane of plastids and the endoplasmic reticulum (ER), respectively [40]. Therefore, plant desaturases fall into two major classes: soluble and membrane-bound desaturases. The soluble desaturases are analyzed separately from membrane-bound desaturases because they are restricted to higher plants and show no evolutionary relationship with the more widely distributed membrane desaturases [39].
The plant stearoyl-ACP desaturase is the only soluble desaturase identified to date. In contrast, all other desaturases identified in plants, algae, animals, and fungi are integral membrane proteins [4,41]. The phylogenetic tree indicated that three AhFAB2 genes were grouped with the stearoyl-ACP desaturases of higher plants and were distinct from those of the fern, the moss, and the green algae (Fig 1). This may suggest that stearoyl-ACP desaturases in the fern/ moss/green algae and higher plants arose by independent gene duplication events. The three AhFAB2 genes were clustered into two subgroups. The AhFAB2-2 and AhFAB2-3 proteins were grouped together with FAB2 enzymes from eudicots and were separate from those of monocotyledons, while AhFAB2-1 and AtFAB2-1 clustered together with genes from eudicots and were set apart from the monocotyledon genes.
FAD4 encodes a predicted integral membrane protein that evolved independently from classic membrane bound desaturases, which were analyzed separately from the remaining membrane-bound desaturases [37]. Moreover this protein class, originally designated Kua proteins [42], is highly conserved in organisms ranging from bacteria (but not cyanobacteria) to mammals. Unfortunately, definitive functional data are not available for these proteins. The phylogenetic tree showed that the AhFAD4 protein was more closely related to enzymes from  Gene sequences were shown by their strain names, accession numbers (locus tags), or labels. Bootstrapping with 1,000 replicates was used to establish the confidence limits of the tree branches. Colored branches indicated different groups of proteins. Red: Δ15 desaturase, blue: Δ7/Δ9 desaturase, purple: sphingolipid Δ4 desaturase, green: microsomal Δ12 desaturase, orange: plastidial Δ12 desaturase, pink: sphingolipid Δ8 desaturase.
desaturase subfamily, the sphingolipid Δ8 desaturase subfamily, and the sphingolipid Δ4 desaturase subfamily. Δ9 desaturase is assumed to be the ancestor of the remaining desaturases based on functional criteria and the position of the clade integrated by Δ9 desaturases [43]. The AhADS gene was grouped with eudicot Δ7 homologs and was set apart from Δ7 enzymes in the fern, the moss, and the green algae, while the ADS genes in green algae and cyanobacteria were placed in a basal position.
In the Δ12/ω3 desaturase subfamily, the AhFAD6 gene, grouped to eudicot chloroplastic Δ12 desaturase, was located along with the cyanobacterial Δ12 desaturases at the basal position of the tree. The higher plant microsomal Δ12 desaturases formed a group and were set apart from those of the fern and the green algae. The AhFAD2-1A and AhFAD2-1B genes clustered together and were separate from the AhFAD2-2 gene. The cyanobacterial ω3 desaturases were placed in a basal position and were grouped with both microsomal and chloroplastic ω3 desaturases from higher plants, the fern, the moss, and the green algae. Seven putative ω3 desaturases from peanut (four FAD3 and three FAD7) were grouped with their respective microsomal or chloroplastic ω3 desaturases from higher plants and were separate from the fern, moss, and green algae enzymes. Therefore, it can be speculated that the cyanobacterial Δ12 desaturase might be the origin of plant Δ12 and ω3 desaturases, including both chloroplast and ER isozymes.
The sphingolipid Δ8 desaturases formed a separate clade (Fig 3). The SLD genes of green algae were placed in a basal position. The four peanut sphingolipid Δ8 desaturase genes (AhSLD-1, AhSLD-2, AhSLD-3, and AhSLD-4) clustered with SLDs from eudicots and were separate from those of monocotyledons. The sphingolipid Δ4 desaturases integrated into one clade (Fig 3). The AhDES gene, grouped with the eudicot DES genes, was located along with DES desaturases from green algae at the basal position of the tree.

Tissue-specific expression patterns
Quantitative real-time PCR (qRT-PCR) was used to confirm the expression patterns of the seventeen FAD genes in different peanut tissues and at different stages of seed development. The actin 11 (AhACT11) gene was used as an internal reference control for total RNA input [44]. Figs 4 and 5 shows that these genes displayed specific temporal and spatial expression patterns across different tissues and developmental stages. AhFAB2-2 and AhFAD3-1 transcripts were more abundant in seeds than in any of the other tissues tested. The highest transcript accumulations of AhFAB2-3, AhFAD3-2, AhFAD4, AhSLD-4, and AhDES genes occurred in flowers. Among them, AhFAB2-3 and AhFAD3-2 had relatively higher expressions in stems. AhSLD-4 was expressed most strongly in flowers followed by seeds, whereas the expressions of the AhFAD4 and AhDES gene were largely restricted to flowers. AhADS and AhFAD7-1 had similar expression patterns, and were most abundant in leaves followed by flowers. The highest abundances of AhFAD7-2 and AhSLD-2 transcripts were in stems. AhSLD-3 transcript levels were highest in stems, followed by flowers, leaves, and roots, with the lowest levels being found in seeds.
The expression patterns of FAD genes across six seed developmental stages are shown in Fig 5. The AhSLD-3 transcript levels remained relatively low during the initial stage of seed development, but increased gradually during the later stages, peaking at 40 DAP, and decreased thereafter until 60 DAP. The expressions of AhFAB2-2 and AhFAD7-1 gradually increased in abundance during seed development, reached a maximum expression level at 30 DAP, and then decreased thereafter. The AhFAB2-3, AhFAD3-2, and AhADS transcript levels were highest at 10 DAP and decreased dramatically thereafter. The AhSLD-2 transcript level remained relatively high at the initial stage of seed development, but decreased gradually from 10 to 20 DAP, and then increased from 30 DAP. The AhFAD3-1 gene showed higher expression levels at 20 DAPs and much lower levels during the other stages. The expression levels of AhDES were highest at the initial stage of seed development, but gradually decreased in abundance during the later stages. The AhSLD-4 transcript levels remained relatively low at the initial stage of seed development, but showed a dramatic increase in abundance at 60 DAP. The AhFAD4 transcript levels remained relatively high in the initial five stages, but showed a  Fatty acid desaturase genes from peanut (Arachis hypogaea L.) dramatic decrease thereafter. The AhFAD7-2 transcript level remained relatively high at the initial stage of seed development, decreased gradually from 10 to 30 DAP, and then increased from 40 DAP, with the highest expression being seen at 60 DAP.
Many desaturase genes from different plant species have been studied and their expression levels are regulated in a tissue-specific manner. Our results indicated that the expression of the AhDES gene was largely restricted to flowers, which highly resembled the expression pattern of its closest ortholog AtDES in Arabidopsis with preferential expressions in pollen and floral tissues [45]. It has been reported that AtFAD2 (AT3G12120) and AtFAD6 (AT4G30950) mRNAs are present in various Arabidopsis tissues, including the roots, rosette leaves, cauline leaves, stems, flowers, and siliques [46,47]. The AhFAD2-1, AhFAD2-2, and AhFAD6 genes were expressed in all tissues surveyed, which was consistent with the Arabidopsis orthologs [17]. In Arabidopsis, AtSLD-1 (At3g61580) was ubiquitously expressed in all organs and most highly expressed in the flowers. However, AtSLD-2 (At2g46210) was expressed in flowers and siliques, but shows only low levels of expression in the leaves, stems, and roots [48]. The AhSLD-4 gene transcript accumulated to the greatest extent in flowers, which was consistent with the Arabidopsis orthologs. The AhSLD-1 transcript was more abundant in leaves, whereas AhSLD-2 and AhSLD-3 were most strongly expressed in stems. In Arabidopsis, AtFAB2-1 (At2g43710) and AtFAB2-7 (At5g16240) were expressed at high levels in the flower, stem and leaf tissues, but at low levels in roots and siliques. AtFAB2-4 (At3g02620) was relatively highly expressed in roots compared to its levels in the leaf, stem, flower and silique tissues, while AtFAB2-6 (At5g16230) was highly expressed in the leaves, but not in the roots. The AtFAB2-5 (At3g02630) isoform was expressed at high levels in the leaf, stem and flower tissues, but at low levels in roots and siliques. The AtFAB2-3 (At3g02610) isoform was only expressed at low levels in the roots and flowers, while AtFAB2-2 (At1g43800) was not detected in any of the tissues analyzed [49]. In peanut, AhFAB2-1 and AhFAB2-2 showed higher transcript abundances in seeds than in any of the other tissues tested [17], whereas the AhFAB2-3 gene accumulated in flowers. In Arabidopsis, the AtADS-1 (At1g06080) gene was expressed most strongly in inflorescence meristems followed by leaves and flowers, and very weakly in roots and seedpods. The AtADS-2 (At2g31360) gene was expressed strongly in all the analyzed organs, although the expression level was higher in flowers and roots than in other organs [50,51]. In peanut, AhADS transcript levels were higher in the leaves, followed by flowers, and the lowest levels occurred in the seeds and roots. The AtFAD3 gene was expressed in both the leaves and roots, while the AtFAD7 gene transcript was only observed in the photosynthetically active organs of Arabidopsis [52]. In peanut, the AhFAD3-1 transcript abundances were higher in seeds than in the other tissues examined, whereas the AhFAD3-2 transcripts were more abundant in flowers. The AhFAD7-1 gene transcript accumulation was highest in leaves, whereas AhFAD7-2 was expressed most strongly in stems. In Gossypium raimondii, GrFAD3.2, GrFAD8.1, and GrSLD5 were expressed at high levels in roots. GrFAD2.3, GrSLD3, GrSLD2, GrSLD1, GrFAD3.1, and GrDSD1 shared high expression levels in young stems. GrFAD5 and GrFAD6 displayed the highest transcript abundance in cotyledons. And GrFAD7 and GrFAD2.4 were predominantly expressed in leaves [53]. In Cucumber, three CsFAB2 genes were dominantly expressed in the seedling leaves. For CsFAD2.1, CsFAD3 and CsFAD6, the highest expression levels were detected in the leaves, whereas for CsFAD4, CsFAD5.1 and CsFAD7, the highest transcript abundances were detected in the cotyledons. In the roots and hypocotyls, only trace expression levels could be detected for any cucumber FAD gene except CsFAD2.1 and CsFAD3 [54]. In Perilla frutescens, PfrFAD2 and PfrFAD3 genes were expressed in leaves and during all stages of seed development, and their expression levels in 2-to 3-week-old developing seeds were 6.7and 25-fold higher than their expression in leaves, respectively. In contrast, although the expression of PfrFAD7-1 and PfrFAD7-2 was similar to that of PfrFAD2 and PfrFAD3 genes in leaves, their expressions were much lower in developing seeds [55]. Thus, the same type of FAD genes from different plants may have different spatial expression patterns, which requires further investigation.

Expression patterns of AhFADs in peanut under abiotic stress
We monitored changes to these transcripts in peanut leaves and roots to confirm the expression patterns of these FAD genes, including five previously cloned desaturase genes, under cold, salt, drought, and ABA stresses. Two microsomal oleoyl-PC desaturase genes (AhFAD2-1A and AhFAD2-1B) have been identified in peanut and their open reading frames (ORFs) were 99% identical [19,20]. Thus, the gene-specific primers used for the amplification of AhFAD2-1 in our analysis recognized and amplified both AhFAD2-1A and AhFAD2-1B genes. Fig 6 shows the expression patterns of these FAD genes in peanut leaves after cold treatment. There were no obvious changes in the abundances of the AhFAB2-1, AhFAB2-2, AhFAB2-3, AhFAD2-1, AhFAD2-2, AhFAD6, AhFAD3-1, AhFAD3-2, and AhDES transcripts in peanut leaves after cold treatment. The expressions of AhFAD4, AhFAD7-2, and AhSLD-2 increased slightly at 1 h after treatment, and then decreased from 3 h to 12 h. After 24 h, their transcript levels reached a maximum, which were approximately 93-, 5-and 3-fold, higher than the nontreated controls, respectively. The expression levels of AhADS increased under cold stress, peaking at 6 h, and then decreased. The greatest increase was about 3.6-fold. The expression of AhSLD-1 increased slightly after 1 h treatment with cold and then decreased. After 48 h, expression of the AhSLD-1 transcript reached its maximum level, with a nearly 2-fold increase. The transcript levels of AhFAD7-1, AhSLD-3, and AhSLD-4 gradually increased under cold stress, peaking at 48 h, 24 h, and 48 h, with approximately 4-, 6-and 2-fold increases, respectively, compared to the non-treated controls.
The expression patterns of AhFADs in peanut leaves and roots after treatment with 200 mM NaCl were also monitored (Fig 7 and S4 Fig). In leaves, there were no obvious changes in the abundances of the AhFAB2-1, AhFAB2-2, AhFAD2-2, AhFAD3-1, AhFAD4, AhSLD-1, AhSLD-4, and AhDES transcripts after salt treatment. The transcript levels of AhFAD6, AhADS, AhFAD7-1, AhFAD7-2, and AhSLD-2 gradually increased under salt stress, peaking at 3 h, 1 h, 1 h, 3 h, and 3 h, respectively, and showed approximately 3-, 2-, 2-, 13-, and 3-fold increases, respectively. The expressions of AhFAD2-1 and AhFAD3-2 gradually increased under salt stress, with peak levels that were about 2-and 7-fold higher, respectively, at 48 h. The expression of AhFAB2-3 increased slightly at 1 h after treatment, and then decreased from 3 h to 12 h. After 48 h, AhFAB2-3 transcripts reached a maximum level, with the greatest increase being approximately 3-fold. The expression of AhSLD-3 slightly increased in the leaves of seedlings subjected to salt stress, with about a 2-fold peak increase at 48 h.
Environmental factors compel organisms to acclimatize to the external conditions [56]. Poikilothermic organisms, such as cyanobacteria and plants, modulate the composition of their membrane lipids in response to changes in environmental conditions [57,58]. Unsaturated fatty acids are essential constituents of glycerolipids in biological membranes and the unsaturation level of membrane lipids is important in controlling the fluidity of membranes [59]. The extent of unsaturation is mainly determined by the activity of fatty acid desaturases, the enzymes that introduce double bonds into specific positions of lipid fatty-acyl chains [60]. Previous studies have revealed that FAD genes are crucial for the survival of plants faced with different environmental stresses. In Arabidopsis, the expression of AtFAD8 was strongly induced by cold temperatures [9]. AtFAD2 and AtFAD6 were found to be active in seedlings under salinity stress [46,47]. The ads2 mutant Arabidopsis plants showed increased sensitivity to chilling and freezing temperatures [61], and the mutants of AtSLD genes also showed enhanced sensitivity to prolonged low-temperature exposure [48]. In transgenic tobacco plants, over-expressing Arabidopsis AtFAD7 also enhanced cold tolerance [62], whereas antisense expression of AtFAD7 reduced salt and drought tolerance [63]. Tobacco plants overexpressing either Arabidopsis AtFAD3 or AtFAD8 gene also exhibited increased tolerance to drought and osmotic stress [64]. In Saussurea involucrate, the expression of sik-SACPD increased in leaves as the temperature decreased from 20 to -10˚C. The FAB2: SikSACPD transgenic plants showed a slightly more resistance to the freezing stress than the FAB2:FAB2 transgenic plants and the wild-type [65]. In Gossypium raimondii, GrFAD8.1, GrFAD2.2, GrFAD8.2, GrSLD2, GrSLD4, GrDSD1 and GrSLD5 were found to be significantly up-regulated in response to cold stress. Conversely, GrFAD5, GrFAD7, GrFAD2.3, GrSLD1 and GrSLD3 were heavily down-regulated after long periods of cold stress treatment [53]. In safflower, the transcription level of CtFAD3 remained constant at all different growth temperatures in the leaves; in contrast, the accumulation of CtFAD7 mRNA slightly increased at low temperature, while CtFAD8 mRNA decreased significantly. The expressions of CtFAD3, CtFAD7, and CtFAD8 in the roots significantly increased at low temperature [66]. In rice, OsFAD8 has been reported to have a functional role in stress tolerance at low temperatures [67]. In tomato, LeFAD3 overexpression enhanced the tolerance of tomato seedlings for salinity stress [68], whereas silencing the LeFAD7 gene alleviated high-temperature stress [69]. In soybean, the expression of FAD3 and FAD7 was tightly regulated in response to cold temperature [70].
Our results indicated that FAD transcripts from peanut were differentially expressed following exposure to abiotic stresses or a stress-induced plant hormone. The AhFAD7-2 transcript levels were considerably enhanced under all stress treatments. The expressions of AhFAD4, AhSLD-2, and AhSLD-3 increased in all materials after the stress treatments, except for salt-or ABA-treated leaves, whereas the transcript levels of AhFAD3-1 only increased in salt-stressed roots. The AhFAB2-2, AhFAD2-1, AhFAD6, and AhFAD2-3 transcript levels were distinctly enhanced after exposure to four kinds of stress separately, except for cold-, salt-treated leaves or cold-, ABA-treated leaves. The expressions of AhFAD7-1 increased in all materials after the stress treatments, except for drought-and ABA-treated leaves. The transcript levels of AhFAB2-1 and AhFAD2-2 increased in salt-, drought-, and ABA-treated roots, and droughttreated leaves, whereas the expression of AhADS increased in cold-, salt-treated leaves and salt-, drought-treated roots. The AhDES transcripts levels increased substantially in roots exposed to salt, drought, and ABA stresses, whereas the AhSLD-4 transcript levels were distinctly enhanced in salt-, drought-treated roots and cold-treated leaves. AhSLD-1 expression increased in salt-or ABA-treated roots and cold-treated leaves, whereas AhFAB2-3 transcript levels increased in salt-, drought-stressed leaves and salt-treated roots. Taken together, these results from qRT-PCR suggested that the expression of most peanut FAD genes was induced by stress treatment, consistent with the FAD genes from other plants [61][62][63][64][65][66][67][68][69][70]. To comprehensively decipher their functions involved in stress tolerance in peanut seedlings, some lipidomic and transcriptomic methods would be employed.

Heterologous expression of AhFAD6 in Synechococcus elongatus (strain PCC 7942)
Synechococcus elongatus (strain PCC 7942) is a freshwater unicellular cyanobacterium that only contains monounsaturated fatty acids. It is an excellent model system for studying fatty acid metabolism. When the Δ12 fatty acid desaturase gene is introduced into the genome of S. elongates, the resultant cells will produce considerable amounts of diunsaturated fatty acids.
Heterologous expression in S. elongates was used to confirm Δ12 regioselectivity and the function of AhFAD6 genes. The pYFAD6 and empty vector (pSyn_1, control) were transformed into the S. elongates. The total lipids of the transformants were analyzed using GC-MS. The results showed a novel fatty acid peak from pYFAD6, which was absent in the control. The novel fatty acid was designated as C18:2 by comparing the retention time to FAME standard mixtures (Sigma). No C16:2 was detected, which indicated that AhFAD6 recognized only one substrate (C18:1) in S. elongates ( Table 2). The C18:2 percentage was 3.3% for pYFAD6 transformants. These data showed that the activity of the AhFAD6 protein was significantly higher in the cyanobacterium than in yeast, where the percentage of C18:2 was 0.1% [17]. Yeast is known to be the model of choice for the functional characterization of microsomal FADs because it contains the short electron transport system required by these desaturases (i.e., cytochrome b5 and NADH-cytochrome b5 reductase) [71]. Nevertheless, the high desaturation level evident from Table 2 suggested that desaturases of plastidial origin, which usually require ferredoxin and NADPH-ferredoxin reductase, were supplied to some extent with reducing equivalents in yeast cells.
It is well known that the AhFAD2-1 gene plays a major role in the conversion of oleic to linoleic acid in seed storage oils [19,20,72]. Two other genes, AhFAD2-2 and AhFAD6 have been isolated by us [17]. They also contribute to the C18:2 pool, although a major portion of this pool reflects contributions from AhFAD2-1 activity. This may indicated that a switch from oleic acid to linoleic acid might involve more Δ12 desaturase genes and an intricate metabolic network that regulates linoleic acid biosynthesis between the endoplasmic reticulum and the chloroplast within peanut cells. The functional validation of these two novel members would facilitate the further genetic manipulation of the peanut oil quality trait that is based on a high O/L ratio [18].

Conclusions
In conclusion, twelve novel FAD-like genes from peanut were cloned, including two FAB2, two FAD3, two FAD7, one FAD4, one ADS, one DES, and three SLD genes. The functions of Fatty acid desaturase genes from peanut (Arachis hypogaea L.)  one AhFAD6 and four AhSLD genes were verified by heterologous expression in S. elongates or S. cerevisiae. Better understanding of this enzyme family will improve efforts to modify the content and composition of seed oils or improve abiotic stress resistance in plants. The results generated in our study provide new information that will increase our understanding of the evolution, functional diversity and gene expression of fatty acid desaturases in plants and opens the way to select candidate genes for functional validation studies in peanut.

Ethics statement
No specific permits were required for the described field studies. No specific permissions were required for these locations and activities. The location is not privately-owned or protected in any way and the field studies did not involve endangered or protected species. Identification of FAD family genes in a peanut cDNA library using Bioedit software
Using this method, we found twelve genes that may encode FAD proteins.

Isolation of full-length cDNA sequences
Total RNA was extracted using the RNeasy Plant Mini kit (Qiagen, Valencia, CA, USA). Contamination with genomic DNA was eliminated by treatment with recombinant DNase I (Qiagen), as recommended by the vendor. Only RNA preparations having an A260/A280 ratio of 1.8-2.0 and an A260/A230 ratio >2.0 were used for subsequent analysis. The integrity of RNA was verified by electrophoresis through 2% agarose gels, followed by SYBR Green staining. First-strand cDNA synthesis was carried out with 2 μg RNA using an RT-PCR kit (Promega, Madison, WI, USA) according to the manufacturer's procedure. We performed PCR with the LA PCR system (Takara, Dalian, China), using 2.5 μl of 10×PCR buffer with MgCl 2 , 1 μl of each primer (10 μM) (S2 Table), 4.0 μl of 10 mM dNTPs, 1 μl of cDNA sample, 0.5 μl of LA Taq™ DNA polymerase, and 15 μl of double-distilled water. The PCR products were separated by electrophoresis through a 1% agarose gel, and purified using a Gel Extraction Kit (Takara) according to the manufacturer's protocol. The purified products were then cloned into the pMD18-T Easy vector (Takara) and sequenced (Shangon, Shanghai, China).

Sequence analysis
The open reading frames (ORFs) and encoded amino acid sequences of all genes were deduced using BioXM 2.6. Physicochemical properties of the deduced proteins were predicted using Protparam (http://www.expasy.ch/tools/protparam.html). Active sites of the protein sequences were analyzed by comparisons against the PROSITE database. The putative subcellular localizations of the candidate proteins were estimated by TargetP (http://www.cbs.dtu.dk/services/ TargetP/) and Predotar (http://urgi.versailles.inra.fr/predotar/predotar.html).

Phylogenetic analysis
Homologs of each member of the Arabidopsis FAD family were identified by BLASTP searches with datasets from Phytozome v10.3 (www.phytozome.net) and Peanut Genome Project (http://peanutbase.org/home) [34]. Only those sequences with an e-value less than e− 50 were considered as members of the FAD family. In each tree, gene sequences were displayed using the nomenclature with the following abbreviations: At,  Table provides the pairwise comparison of the FADs from cultivated and wild peanut varieties. Amino acid sequences were aligned using the ClustalX program with the implanted BioEdit [76]. The neighbor-joining (NJ) method in MEGA4 [77] was used to construct the phylogenetic tree. Bootstrapping with 1,000 replicates was used to establish the confidence limits of the tree branches. Default program parameters were used.

Quantitative real-time RT-PCR
qRT-PCR analysis was performed using a LightCycler 2.0 instrument system (Roche, Germany). The actin 11 gene (AhACT11) was selected as the reference gene [44]. Seventeen pairs of gene-specific primers (S5 Table) were designed after analyses of the target genes' sequences. qRT-PCR reactions were performed using the SYBR Premix Ex Taq polymerase (Takara) according to the manufacturer's instructions. Each 20-μl reaction was comprised of 2 μl of template, 10 μl of 2× SYBR Premix, and 0.4 μl (200 nM) of each primer. The reactions were subjected to an initial denaturation step of 95˚C/10 s, followed by 40 cycles of 95˚C/5 s, 60˚C/ 30 s and 72˚C/10 s. A melting curve analysis was performed at the end of the PCR run over the 60-95˚C range, increasing the temperature stepwise by 0.5˚C every 10 s. The baseline and quantification cycle (CP) were automatically determined using the LightCycler Software. Zero template controls were included for each primer pair, and each PCR reaction was carried out in triplicate. The relative quantification method (delta-delta Cp) was used to evaluate quantitative variation [78].

Heterologous expression of AhFAD6 in Synechococcus elongatus (strain PCC 7942)
The AhFAD6 in the pSyn_1 plasmid was transformed into the freshwater unicellular cyanobacterium S. elongatus (strain PCC 7942), using the natural transformation method according to the manual (Invitrogen, Carlsbad, CA, USA). Transformants were selected by screening for resistance to 10 μg/mL spectinomycin in BG-11 solid medium. Colony PCR was performed to screen the transformed S. elongatus colonies for full integration of the promoter and the gene of interest. The positive colonies were transferred into BG-11 liquid medium with 10 μg/mL spectinomycin and grown at 28˚C for 30 days. Cells were harvested by centrifugation, washed three times with double-distilled water and used for the extraction of total fatty acids.

Fatty acid extraction and analysis
Total fatty acids were extracted and transmethylated with methanolic HCl from algae cells according to Browse et al (1986) [79]. All samples were analyzed using a 7890A/5975C gas chromatography (Agilent Technologies, California, USA) equipped with a 5975C single quadrupole GC/MSD detector and an HP-INNOWAX capillary column (30 m × 250 μm × 0.25 μm). High purity nitrogen was used as the carrier gas with flow rate of 40 mL/min. The injector and detector temperatures were both 250˚C, and the column temperatures were programmed from 150˚C to 230˚C. Measurements were performed using peak height area integrals expressed as a percentage of the total of all integrals. The experiment was carried out in triplicate.

Heterologous expression of AhSLDs in Saccharomyces cerevisiae
The AhSLDs in the pYES2 plasmid were transformed into the auxotrophic S. cerevisiae strain INVSc1 (MATa his3-Δ1 leu2 trp1-289 ura3-52), using the polyethylene glycol/lithium acetate method according to the manual (Invitrogen, Carlsbad, CA, USA). Yeast cells transformed with an empty pYES2 plasmid were used as the negative control. The AhSLDs-transformed yeasts were grown at 30˚C in SC-U containing 2% (w/v) glucose for 24 h, and expression was further induced by the addition of 2% (w/v) galactose and 1% (w/v) tergitol NP-40 (Sigma, Taufkirchen, Germany) for an additional 72 h at 20˚C. Yeast cells were collected by centrifugation for 10 min at 2000 g and dried at 50˚C.

Long chain base (LCB) analyses
Pellets of wild-type and transformed yeast cells (100 mg of dried weight) were used to prepare the LCBs for subsequent RP-HPLC analysis as previously described [80]. Phytosphinganine (t18:0) (Sigma) was used as the internal standard for yeast sample analyses. Briefly, induced yeast cells (100 mg, dried weight) were grounded into a fine powder, and subjected to strong alkaline hydrolysis in 2 ml of 10% (w/v) aqueous Ba(OH) 2 and 2 ml of dioxane for 16 h at 110˚C. After hydrolysis, 2 ml of 2% (w/v) ammonium sulphate was added, and the liberated sphingolipid long chain bases were extracted with 2 ml of diethylether. The upper phase was removed to a second tube, dried under nitrogen, and derivatized with o-phthaldialdehyde (OPA) (Invitrogen, Carlsbad, CA, USA) as previously described [81]. Individual LCBs were separated by RP-HPLC (Waters alliance 2695-2475 multi λ fluorescence detector, Waters, Milford, USA) with Penomenex C18 columns (250 mm 4.6 mm, 5 mm). Elution was performed at 1.2 mL/min with 20% solvent RA (5 mmol/L potassium phosphate, pH 7.0), 80% solvent RB (100% methanol) for 9 min, increasing to 90% solvent RB by 32 min, returning to 80% solvent RB and re-equilibrating for 2 min. Fluorescence was excited at 340 nm and detected at 455 nm.    Validation: Lijuan Pan.