Peanut plants (A. hypogaea L. cultivar Huayu 19) were grown in a growth chamber with a 16 h light/8 h dark photoperiod at 26 °C/22 °C day/night temperatures. Leaves, stems, cotyledons, hypocotyls, and roots were sampled from the seedlings at the trefoil leaf stage. Seeds were sampled at 10, 20, 30, 40, 50, and 60 days after pegging (DAP). Flowers were collected when the seedlings were in the flowering phase. For the cold treatment, seedlings in the soil at the trefoil leaf stage were kept at 4 °C, and leaves were sampled separately either before cold treatment (0 h) or after continuous exposure to 4 °C for 1, 3, 6, 12, 24, 48, or 72 h. For stress treatments, the roots of seedlings grown in soil were flushed carefully with tap water to remove all soil, and then submerged in solutions of 200 mM NaCl, 20% PEG-6000, or 100 μM ABA. Leaves and roots were sampled separately after treatment for 0, 1, 3, 6, 12, 24, 48, or 72 h. All samples were immediately frozen in liquid nitrogen and stored at −80 °C until required.

The cDNA sequences used in this study came from three cDNA libraries from three institutes (data not shown): Shandong Peanut Research Institute, Oil Crops Research Institute of The Chinese Academy of Agricultural Sciences, and Crops Research Institute of Guangdong Academy of Agricultural Sciences. All expressed sequence tags (ESTs) of the 36,741 cDNA sequences were saved as FASTA format. The amino acid sequences of glycerol-3-phosphate acyltransferase genes of Arabidopsis, AtGPAT2 (NP_563651), AtGPAT9 (NP_568925) and AtATS1 (NP_174499) were used to search for homogeneous genes from the peanut cDNA library. Before searching for members of the GPAT gene family, a local nucleotide database file was created using Bioedi software. A local BLAST procedure was then run to find the homologous genes of the GPAT family. Using this method, we found six genes that may encode GPAT proteins.

The 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, WI, USA) according to the manufacturer's procedure.

We performed PCR with the LA PCR system (TaKaRa) using 2.5 μL of 10×PCR buffer with MgCl₂, 1 μL of each primer (10 μM), 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).

The open reading frames (ORFs) and encoded amino acid sequences of all the genes were deduced using BioXM 2.6. The physicochemical properties of the deduced protein were predicted using Protparam (http://www.expasy.ch/tools/protparam.html). Active sites of the protein sequence were analyzed by comparison against the PROSITE database. Predicted transmembrane domain (TMDs) in GPAT proteins were identified using the TMHMM Server (version 2.0) (http://www.cbs.dtu.dk/services/TMHMM) and visual inspection. 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).

The gene structure display server (GSDS) program (Guo et al., 2007) was used to illustrate exon/intron organization for individual desaturase genes by comparison of the cDNAs with their corresponding genomic DNA sequences. To identify the conserved motifs, MEME (Multiple Expectation Maximization for Motif Elicitation) version 4.9.1 (http://meme.nbcr.net/meme/cgi-bin/meme.cgi) was employed with a set of parameters as follows: number of repetitions – any, maximum number of motifs –20, optimum motif width set to ≥6 and ≤200 (Bailey and Elkan 1995). The motifs obtained were recorded using the SMART (http://smart.embl-heidelberg.de/) and NCBI-CDD (National Center for Biotechnology Information Conserved Domain Database) search programs.

Homologs of each member of the Arabidopsis GPAT family were identified by BLASTP searches with datasets from Phytozome v9.1 (www.phytozome.net). Only those sequences with an e-value less than e⁻⁵⁰ were considered as members of the GPAT family. In each tree, gene sequences other than Arabidopsis and peanut GPATs were displayed using the nomenclature with the following abbreviations: Ah, Arachis hypogaea; At, Arabidopsis thaliana; Glyma, Glycine max; Medtr, Medicago truncatula; Pp, Physcomitrella patens; Cre, Chlamydomonas reinhardtii; Vocar, Volvox carteri. Table 1 provides a detailed description of the proteins used and the corresponding accession numbers. Amino acid sequences were aligned using the ClustalX program with the implanted BioEdit (Thompson et al., 1994). The neighbor-joining (NJ) method in MEGA4 (Tamura et al., 2007) 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. Bootstrap values from the neighbor-joining analyses were listed to the left of each node, and values higher than 50 were shown.

A quantitative Real-time RT-PCR (qRT-PCR) analysis was performed using a LightCycler 2.0 instrument system (Roche, Germany). The alpha tubulin 5 gene (AhTUA5) was taken as a reference gene (Chi et al., 2012). Seven pairs of gene-specific primers (Table 2) were designed after analysis of the sequences of target genes. qRT-PCR reactions were performed using the SYBR Premix Ex Taq polymerase (TaKaRa, Japan) according to the manufacturer's instructions. Each 20-μL reaction comprised 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⁻¹, followed by 40 cycles of 95 °C·5s⁻¹, 60 °C·30s⁻¹ and 72 °C·10s⁻¹. A melting curve analysis was performed at the end of the PCR run over the range 60–95 °C, increasing the temperature stepwise by 0.5 °C every 10 s. The baseline and quantification cycle (CP) were automatically determined using the Light Cycler 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 (Livak and Schmittgen, 2001).

Lipid content in seed was determined by a standard Soxhlet extraction method (Harwood, 1984). From each cultivar, 1 g sample was ground and then extracted with petroleum ether in a Soxhlet apparatus for 8 h. Petroleum ether was then volatilized in the draft. The experiment was carried out in triplicate. Lipid content was expressed as % of seed dry weight.

Six genes that likely encode glycerol-3-phosphate acyltransferase (GPAT) proteins were found using Bioedit software. They were cloned and designated as AhATS1, AhGPAT1, AhGPAT2, AhGPAT6, AhGPAT8, and AhGAPT9 according to the homologous genes identified in Arabidopsis. Among the six genes, two genes have the complete open reading frame (ORF) in the peanut cDNA library and cloned by conventional RT-PCR, whereas four genes were cloned using the rapid amplification of cDNA ends (RACE) method. The ORFs of the five genes were 1,356 bp, 1,668 bp, 1,635 bp, 1,500 bp, 1,518 bp, and 1,131bp in length, encoding 451, 555, 544, 499, 505, and 376 amino acids, respectively. The genomic sequences were 5,766 bp, 2,146 bp, 2,209 bp, 3,176 bp, 4,474bp, 4,970 bp in length, respectively (Table 3). The sequence information of six genes was submitted to Genbank, with the Genbank identification numbers KC762933, JN032676, HQ589243, HQ589244, JX843442, and JX843441, respectively.

A search using NCBI BLAST revealed that six GPAT proteins have high sequence similarities with GPATs in Arabidopsis. AhATS1 shares 55.1% sequence similarity with AtATS1. AhGPAT1 shows 55.3% sequence similarity with AtGPAT1, AhGPAT6 shares 78.2% similarity with AtGPAT6. The AhGPAT2 protein shares 52.8% and 52.0% sequence similarity with AtGPAT2 and AtGPAT3, respectively, and AhGPAT9 shows 79% similarity with AtGPAT9. The AhGPAT8 protein is most similar to AtGPAT4 (77.9%), and AtGPAT8 (78.4%), both of which have been implicated in the synthesis of cutin polymers (Li et al., 2007b).

As shown in Figures 1 and 2, alignment of the deduced polypeptide sequences of six GPAT proteins demonstrates that they are similar in length and share several features that are characteristic of other plastidial and membrane-bound GPATs from evolutionarily diverse organisms. These features include the presence of four conserved amino acid motifs (AT-I to AT-IV) which are important for acyltransferase activity (Mañas-Fernández et al., 2010). The typical acyltransferase (AT) domain is localized within the C-terminal half of the molecule. Residues implicated in catalysis, such as histidine and aspartic acid residues in AT-I, glycine residues in AT-III, and a proline residue in AT-IV are all present in peanut GPATs, as are the arginine (AT-II) and glutamic (serine) (AT-III) residues involved in binding to the G3P substrate (Gonzalez-Baró et al., 2007). In addition to the AT region, a haloacid dehalogenase (HAD)-like domain is found in the N-terminal half of AhGPAT6 and AhGPAT8. This conserved domain is present in a super-family of proteins, most of which are phosphohydrolases. Close inspection of this region in AhGPAT6 and AhGPAT8 and their putative orthologues reveal the presence of highly conserved motifs, named HAD-I through HAD-IV (Figure 1), which have been described in HAD-like proteins (Burroughs et al., 2006). They include the typical DXD signature (which contains critical aspartic acid residues that act as a nucleophile during catalysis), the extremely conserved threorine and lysin residues in HAD-II and HAD-III boxes, respectively (both of which contribute to the stability of the reaction intermediates), and a GDXXXD motif in HAT-IV that contains acidic residues required for coordination to the Mg²⁺ ion in the active site.

The AhATS1 protein is probably located in chloroplast, as predicted using the TargetP Server and Predotar tools. The N-terminal end of AhATS1 had a high proportion of hydroxylated and small, hydrophobic amino acids, which is typical of a chloroplast transit peptide. The AhGPAT1 and AhGPAT2 proteins also possess an extended N-terminal region that exhibits characteristics of a mitochondrial targeting peptide. All of the other GPAT proteins lack any recognizable N-terminal intracellular targeting signal motifs, but do contain putative C-terminal ER retrieval signals.

Genes in the same clade had more similar exon/intron structures than those genes in the other clades (Figure 3). Both ATS1 and GPAT9 clade members had twelve exons, whereas the GPAT4/8 clade members possessed four exons. All of the remaining six GPAT clade members had two exons, except for AhGPAT1, which possessed three exons.

The MEME motif search tool was employed to identify the conserved motifs present in peanuts and Arabidopsis GPAT proteins (Figure 3), and 20 distinct motifs were identified. Most of the motifs belonged to the regions that represented the typical domains of acyltransferase. The motif 4 was found in all the members of the GPAT family proteins. Both of the ATS1 proteins had the motifs 4, 11, 13, and 19, whereas the GPAT9 proteins all possessed the motifs 4, 9, and 12. The conserved motifs 1–8 and 14 were present in all of the remaining eight GPAT clade members. Both of the GPAT1 proteins had the motifs 1–8, 10, 14, and 20, whereas another motif 17 was present in AtGPAT1. All proteins belonging to GPAT2/3 clade had the motifs 1–8, 10, and 14, except for AtGPAT2, which had another motif 17. All GPAT4–GPAT8 clade members possessed the motifs 1–8, 14, and 15. The motifs 10 and 16 were present in GPAT4/6/8 proteins, whereas the motifs 17 and 18 existed in GPAT5/7 proteins.

To examine the relationships among different sources of GPAT genes, the neighbor-joining method was used to construct phylogenetic trees and all tree topologies were highly congruent (Figure 4). As shown in the phylogenetic tree, all of the GPATs fell into three distinct clades: the ATS1 clade, GPAT9 clade, and GPAT1–GPAT8 clades.

Searches against prokaryote and non-photosynthetic eukaryotic sequences, and of the fully sequenced genomes of Chlamydomonas, Volvox and other algae do not identify any GPATs with significant similarities (BlastX E <10⁻⁵) to the sn-2 GPATs found in land plants (Yang et al., 2012). In contrast, plastid-localized ATS1 and GPAT9 were found in the algal genomes. Thus the sn-2 GPAT family clearly belongs to a lineage specific to land plants and evolved to provide pathways for functions not present in other organisms. The AhATS1 protein was grouped with ATS1 enzymes from higher plants and green algae, and lie apart from membrane-bound GPAT clades. AhGPAT9 clustered with GPAT9 from higher plants and green algae, apart from the subgroup comprised of GPAT1–GPAT8 from higher plants. The sn-2 GPAT family also fell into three distinct conserved subfamilies. It was assumed that the GPAT4/6/8 clade is the most ancient and arose early during the evolution of land plants (bryophytes), which is involved in the assembly of cutin or cutin-like polymers in the first land plants (Yang et al., 2012). In contrast, the phosphatase-minus GPAT1-3 and 5/7 clades diverged later with the appearance of tracheophytes (Yang et al., 2012). Whereas AhGPAT8 was grouped with GPAT4 and GPAT8 from higher plants, AhGPAT6 fell into the GPAT6 subfamily. Sequences of the GPAT1–GPAT3 clade were more divergent compared with the GPAT4/6/8 and GPAT5/7 clades. AhGPAT1 and AhGPAT2 were grouped with their respective GPAT1 or GPAT2/3 enzymes from higher plants, and lie apart from the GPAT4–GPAT8 clades.

Quantitative real-time RT-PCR (qRT-PCR) was employed to confirm the expression patterns of the six novel genes in different peanut tissues and at different stages of seed development. The alpha tubulin 5 (AhTUA5) gene was used as an internal reference control for total RNA input (Chi et al., 2012). As shown in Figure 5, these six genes displayed specific temporal and spatial expression patterns across different tissues and developmental stages. AhATS1 showed higher transcript abundance in flowers and leaves than in any of the other tissues tested. The highest abundance of AhGPAT1 transcript was in leaves and the lowest was in stems and flowers. Levels of AhGPAT2 transcript were highest in leaves, followed by stems and seeds, with the lowest levels in roots and flowers. AhGPAT6 and AhGPAT8 had similar expression patterns, showing higher transcript abundance in leaves and roots. AhGPAT9 exhibited its highest transcript accumulation in stems followed by flowers and seeds.

The expression patterns of six GPAT genes across six developmental stages of seeds are also shown in Figure 5. Levels of AhGPAT1 transcript were maximal at 10 days after pegging (DAP) and decreased gradually thereafter. The expression patterns of AhGPAT2 and AhGPAT6 were similar over the course of seed development, with higher levels of AhGPAT2 and AhGPAT8 transcripts seen at 10 DAP and 40 DAP. The expression levels of AhATS1 and AhGPAT6 were highest at the initial stage of seed development but dramatically decreased in abundance during later stages. The AhGPAT9 transcript remained relatively low at the initial stage of seed development but increased gradually during later stages of seed development. In peanut cultivar Huayu19, seed lipid content was low in the first period of lipid accumulation, but was characterized by a drastic increase during the initial four stages after pegging (Figure 5). The seed lipid content reached a maximum value of 49.75% at 50 DAP and decreased thereafter at 60 DAP. The expressions of the AhGPAT9 gene coincided with the lipid accumulation rate in peanut seed, whereas the expressions of other AhGPAT genes were not in complete agreement with seed lipid accumulation rate, especially in the earlier stages of seed development like the period from 10 to 30 DAP. These results indicated that AhGPAT9 may be an important component in the lipid biosynthesis process.

To confirm the expression patterns of six GPAT genes under cold, salt, drought and ABA stress, we monitored the changes in these transcripts in peanut leaves and roots. Figure 6 shows the expression patterns of six GPAT genes in peanut leaves upon cold treatment. Transcript levels of AhGPAT1 in the leaves decreased distinctly and rapidly between 1 h and 6 h after cold treatment, and increased thereafter. The levels of AhGPAT2 transcript gradually accumulated between 1 h and 24 h after cold treatment, and then decreased drastically, with a peak level of about 4-fold increase at 24 h. The expressions of AhGPAT6 and AhGPAT8 were slightly increased under cold stress, with a peak level at 1 h, and then decreased gradually. The expression of AhATS1 and AhGPAT9 gradually decreased under cold stress, while the lowest level was detected at 72 h.

The expression patterns of AhGPATs in peanut leaves and roots after treatment with 200 mM NaCl were also monitored (Figure 6). The expression patterns of AhGPAT1, AhGPAT9 and AhATS1 were different in leaves and roots. Transcript levels of AhGPAT1 decreased distinctly and rapidly from 1 h to 48 h in the leaves of seedlings subjected to salt treatment, but increased obviously in roots after 12 h treatment, with a nearly 4-fold increase 48 h after the salt treatment. The levels of AhGPAT9 transcript decreased gradually in leaves under salt stress, but increased obviously after 24 h treatment and exhibited nearly a 7-fold increase after the roots were treated for 48 h. The transcript levels of AhATS1 decreased distinctly and rapidly from 1 h to 48 h in salt-treated leaves, but increased obviously in roots after 1 h treatment, with a peak level of about 20-fold observed at 3 h. The expressions of AhGPAT2 were increased under salt stress, with a peak level at 3 h in both leaves and roots, where the greatest increases were about 3-fold and 41-fold, respectively. In leaves, the expressions of AhGPAT6 and AhGPAT8 decreased rapidly after 1h treatment and then increased to the peak level at 3 h. After 3 h, the levels of AhGPAT6 and AhGPAT8 transcripts decreased distinctly. In the roots, the expressions of AhGPAT6 and AhGPAT8 gradually increased under salt stress, with a maximum increase of about 4-fold observed at 3 h, and then decreased substantially.

A 20% solution of PEG-6000 was used to mimic drought stress to monitor the expression patterns of AhGPATs in peanut leaves and roots (Figure 7). In the leaves, the expressions of AhGPAT1 slightly increased 3 h after treatment, and then decreased from 6 h to 72 h. In PEG-treated roots, the levels of AhGPAT1 transcript were distinctly enhanced relative to the peak level (an approximately 9-fold increase) which was observed at 12 h. The transcript levels of AhGPAT2 were obviously increased in both leaves and roots under drought stress, with peak expression levels at 1 h in leaves and 24 h in roots. The greatest increase was about 16-fold in leaves and 5,537-fold in roots. Within 6 h after treatment, AhGPAT6 and AhGPAT8 genes were slightly down-regulated in leaves and obviously up-regulated in the roots of peanut seedlings subjected to drought stress. The expressions of AhGPAT9 in leaves increased rapidly with a peak level of about 6-fold increase at 3 h under drought treatment, whereas in roots the expressions increased slightly after 6 h treatment and then decreased from 6 h to 72 h. The expressions of AhATS1 were obviously increased in both leaves and roots under drought stress, with peak levels for both at 3 h. The greatest increase was about 3-fold in leaves and 12-fold in roots.

We also examined the response of AhGPAT genes to exogenously applied ABA, which is a plant signaling molecule involved in plant defense signaling pathways (Figure 7). There was no obvious change in the levels of AhGPAT1 transcript in peanut leaves following ABA treatment, although the levels of AhGPAT1 transcript in roots were obviously higher 48 h after initial exposure to exogenous ABA. In leaves, the expressions of AhGPAT2 increased slightly after 1 h treatment with ABA and then decreased from 3 to 12 h. After 24 h, the levels of AhGPAT2 transcript remained slightly higher than in untreated leaves. The Levels of AhGPAT2 transcript were higher in ABA-treated roots than in untreated roots observed 6 h after treatment, with a maximum increase of approximately 40-fold. There were no obvious changes in the abundances of AhGPAT6, AhGPAT8 and AhATS1 transcripts in peanut leaves after ABA treatment. However, the levels of three transcripts increased in roots, where they reached maximum levels 6 h after ABA treatment, with the greatest increases observed being about 6-, 5- and 25-fold, respectively. The expressions of AhGPAT9 were slightly increased in both the leaves and roots of seedlings subjected to ABA stress, with peak levels at 48 and 72 h, respectively.

The above results indicate that GPAT transcripts from peanuts are differentially expressed following exposure to abiotic stresses or abscisic acid. The levels of AhGPAT2 transcript were distinctly enhanced after exposure to all four kinds of stress treatments except for ABA-treated leaves. The transcripts of AhGPAT1, AhGPAT6, AhGPAT8 and AhATS1 increased substantially in roots exposed to salt, drought, and ABA stress. The expressions of AhGPAT6, AhGPAT8, AhGPAT9 and AhATS1 were slightly higher in leaves under certain stress conditions than under normal conditions. These results suggest that these genes may play an important role in enhancing peanut resistance to abiotic stress. Some genes were obviously down-regulated after stress treatments, such as AhGPAT1 and AhGPAT9 transcripts in cold- and salt-stressed leaves. This indicates that these genes may have a negative function in peanut abiotic stress regulation.