Introduction

Rice (Oryza sativa L.) is the most important staple food for a large part of the world’s human population and most of it is consumed after milling. Rice bran, which is a by-product after milling, is rich in oil (12–23%), proteins (14–16%), vitamins and crude fiber (8–10%). Rice bran oil (RBO) is typically an oleic–linoleic-type fatty acid, and its physico–chemical properties qualify it for good quality edible oil [1]. Moreover, the RBO has been considered superior in comparison to other vegetable oils because it is a rich source of ω-3 and ω-6 fatty acids, and high level of unsaponifiables [2]. In addition, consumption of RBO leads to hypocholestrolemic effects due to the presence of primarily phytosterols, tocols (tocopherols and tocotrienols), γ-Oryzanol, and triterpene alcohol in its unsaponifiable fraction [3, 4]. Although rice bran could be a valuable source of nutrition, its use is severely limited by the rapid degradation of rice bran oil and more rapid development of hydrolytic rancidity in comparison to other vegetable oils. Immediately after the completion of rice milling, triacylglycerols (TAGs) in rice bran rapidly decompose, and the level of free fatty acids (FFAs) increases. As a consequence, the bran becomes unsuitable for human consumption or for the production of edible oil with acceptable quality [5,6,7].

TAGs are the primary reserve lipids inside the phospholipid membrane-bounded oil bodies of rice bran. It has been reported that phospholipid-degrading enzymes viz. phospholipases, lipid-oxidizing enzymes, and acyl hydrolases, are important contributors to membrane degradation [8, 9]. Of all these enzymes, phospholipase D (PLD) initiates an early step in the membrane degradation and seed deterioration by breaking down phospholipids into phosphatidic acid [10, 11]. The treatment of rice bran fraction with PLD enzyme leads to disintegration of the rice bran oil bodies and cause decline in the levels of phosphatidylcholine, followed by the decomposition of TAGs into FFAs [12,13,14]. Further, interaction of the produced FFAs with endosperm starches reduces the edibility of the rice. Moreover, action of lipoxygenases (LOXs) on FFAs, having a 1, 4-pentadiene structure, such as linoleic and linolenic acids, and their subsequent oxidation and decomposition, convert them into low molecular-weight volatile products (creating a stale flavor) [15,16,17]. Thus, PLD is the enzyme which initiates lipid decomposition followed by deterioration of the rice grain and rice bran fractions.

A survey of the rice genome database indicated the presence of 17 PLD genes in the genome including eight isoforms of OsPLDα, two isoforms of OsPLDβ, three isoforms of OsPLDγ, two isoforms of OsPLD ξ, one isoform of OsPLD κ, and one isoform of OsPLD φ [18]. Among these PLD isoforms, OsPLDα1 has been found to be responsible for rice bran oil rancidity [19]. This gene has been cloned by Suzuki (2011) and is 6.28-kb in size including the promoter region, and is located on the antisense strand of rice chromosome 1 [20].

Although several methods are being used to suppress the OsPLDα1 activity in rice bran but all these bring only partial inactivation; are associated with negative effects on the nutritional value of rice bran; and add to the cost of oil production and time stringency for treatment [21,22,23]. Therefore, a cost effective alternative is required to reduce the susceptibility of rice bran to hydrolytic rancidity. Although decline in the activity of this enzyme is must for maintaining the rice bran quality, breeding for reduced activity of this enzyme in rice seeds has not been attempted so far. The use of breeding techniques by transferring the desired allele could be effective in increasing the shelf life of rice bran against lipid hydrolysis if genetic variations exist between cultivars for this trait.

The problem of reduced gene pool of cultivated germplasm is particularly relevant in self-pollinated crops such as rice, where the level of genetic variation among cultivars can be lower than 5% of the total variation in natural populations. The wild relatives of rice constitute a major gene pool for rice improvement and have been used as sources of agronomically important genes particularly those involved in tolerance to biotic and abiotic stresses [24,25,26]. However, there is no report in literature focusing on the evaluation of wild rice germplasm for variation in OsPLDα1 enzyme activity and the allelic variation that underpins such phenotype.

In plants, a large number of transcription factors are known to control the expression of genes encoding metabolic enzymes by binding to target regulatory sequences usually within the upstream regions [27]. In this study, we performed a parallel analysis of variation in the upstream sequences of OsPLDα1 orthologs and paralogs and their transcriptional activity during endosperm development across a panel of Oryza species and cultivars, in order to identify nucleotide variation in transcription factor binding motif (TBMF), number/frequency, and location of regulatory elements binding sites in promoter regions of allelic variants. The aim of present study was to dissect the molecular mechanism of OsPLDα1 gene expression and to identify efficient promoters to be used in genetic engineering to improve the rice bran quality in elite lines.

Materials and methods

Plant material

A total of 26 representative accessions belonging to 4 wild Oryza species viz. O. barthii (n = 2), O. nivara (n = 4), O. rufipogon (n = 14), O. longistaminata (n = 1), and African cultivated rice O. glaberrima (n = 5) were chosen for the study. The germplasm accessions were originally procured either from the International Rice Research Institute (IRRI), Philippines or from National Rice Research Institute (NRRI), Cuttack and being actively maintained at Punjab Agricultural University (PAU), Ludhiana. Selection of the wild species accessions for this study was done on the basis of country of their origin. The list of all the wild species accessions and their country of origin is given in the Table 1. The cultivars viz. Punjab Rice 114 (PR 114), PAU201, Nagina 22 (N22), IR64, Pusa 44, Minghui 63, Feng-Ai-Zhan, and Kitake were used as positive checks for the sequence analysis.

Table 1 List of wild Oryza species accessions selected for OsPLDα1 promoter analysis
Full size table

Primer designing, DNA extraction and PCR amplification

The 1.6-kb promoter sequence (1.2-kb upstream and 0.4-kb downstream of the Translation Start Codon) of the OsPLDα1 gene from Oryza sativa japonica cv. ‘Nipponbare’ was retrieved from NCBI (http://www.ncbi.nlm.nih.gov/nuccore/AB571657.1) and used as a reference for the present study. Three overlapping oligonucleotide primer pairs were designed (Supplementary Table S1) from this sequence using PerlPrimer (http://perlprimer.sourceforge.net/) software and re-checked by BLAST search to ensure that they matched uniquely with the expected positions in the rice genome. Leaf tissue from each selected wild species accession and cultivar was collected in the field and stored in − 80 °C deep freezer. The collected leaves were pulverized in liquid nitrogen using a sterilized and pre-chilled mortar and pestle to reduce the DNase activity. Further, the DNA extraction was conducted according to the improved CTAB (Cetyltrimethylammonium bromide) method modified from Saghai-Maroof et al. [28]. Quantity and quality of DNA from each sample was accessed on 0.8% agarose gel, and DNA samples were then diluted with 1X TE buffer and stored at − 20 °C for further use. PCR was performed in thermocycler (Eppendorf and Biometra) using 0.3 μl Phusion® high fidelity DNA polymerase, 3 μl of genomic DNA (20 ng/μl), 6 μl of 5 × HF buffer, 6 μl of dNTPs (1 mM), 3 μl of primer (5 μM), and Nuclease Free Water to make a total volume of 30 μl. Cycling conditions were an initial denaturation of 94 °C for 5 min followed by 35 cycles of 1 min at 94 °C, 1 min at 55 °C, 1 min at 72 °C; concluding with a final extension of 72 °C for 5 min. PCR products were analysed by electrophoresis in ethidium bromide stained 1.0% agarose gels. 1 kb plus ladder (Thermo Scientific Generular) was used to estimate PCR fragment size. The amplified fragments were excised and purified using Wizard® SV PCR Clean-Up System (Promega, Inc.) as per manufacturer’s protocol.

Sequencing of the PCR products and SNP detection

Sequencing was performed, in two replicates for each sample, on ABI Sequencer 3730xl using BigDye Terminator V3.1 cycle sequencing reaction kit at the School of Agricultural Biotechnology, Punjab Agricultural University, Ludhiana, India. The obtained OsPLDα1 promoter sequences were extracted from chromatogram files using CHROMAS Lite 2.1.1 (http://technelysium.com.au/). Contigs were generated using DNA Baser v4.23.0 (http://www.dnabaser.com/) software which also helps in automatic detection and trimming of low quality regions of the sequences. Alignment of the sequenced promoters against reference was carried out using ClustalX 2.1.1. Based on the alignment candidate SNPs (Single Nucleotide Polymorphisms) and InDels (Insertion-Deletions) were predicted. The candidate SNPs were then manually curated by analyzing and comparing chromatogram files with ClustalX alignment files. The bases which had Q-value below 20 were not considered as actual SNPs.

Phylogenetic analysis

A phylogenetic tree was generated using MEGA6 software (Tamura et al. 2013) using the multiple alignment file obtained earlier using ClustalX 2.1.1 [28]. The evolutionary distances were computed using the Maximum Composite Likelihood method with 1000 bootstrap (Tamura et al. 2004) and are in the units of the number of base substitutions per site [29].

In-silico analysis of the OsPLDα1 promoters

Cis-elements in the OsPLDα1 promoter regions of Nipponbare, wild species accessions and cultivated rice were identified using PLACE database (http://www.dna.affrc.go.jp/PLACE/).

OsPLDα1 gene expression analysis

To determine OsPLDα1 gene expression levels and assess their association with cis-elements in OsPLDα1 orthologs, total RNA was extracted 7 days after flowering from immature grains of selected wild species accessions. At that stage, plants have small grain size which leads to low amount of isolated RNA, if collected from the individual plant. Maintenance of wild species on a large area to obtain the required quantity of rice bran is highly tedious and time consuming. Moreover, cultivation of wild species on a large area may lead to the spreading of wild seeds into the broader area and can contaminate the cultivated rice land as these species have the weedy nature. Hence, only five plants per accession have been planted in the field each year to maintain the seed. Due to these limitations, from each selected wild species accession, nine panicles (3 panicles each from three different plants belonging to that accession) were collected after 7 days of flowering and harvested into liquid nitrogen to store at − 80 °C until RNA isolation. Further, same amount of tissue (immature grains) was taken from all the nine panicles and the collected tissue was weighed to get the equal amount (100 mg) for RNA extraction using the TRIzol® reagent (Thermo Fisher Scientific) according to the manufacturer’s protocol. Therefore, each experimental replicate in this study represents RNA from three biological replicates.

The 12 wild species accessions, for expression study at OsPLDα1 locus, were selected from the entire set of 26 wild species accessions on the basis of phylogenetic tree generated using their promoter sequences. From each phylogenetic group, one accession was selected to represent that group. In addition, maximum accessions were selected from O. rufipogon accessions as there was maximum variation obtained in the promoter region of this species. Moreover, RNA samples were collected from Punjab Agricultural University, Ludhiana, India and expression analysis was performed using lab facilities at the University of Maine, USA. Hence, a few wild species accessions were not included in the expression study as they remained in the vegetative state or flowering occurred very late after the sample collection, owing to their photoperiod sensitive nature.

The quantity of RNA in the samples was determined by using NanoDrop® ND-1000. RNA samples were reverse transcribed to obtain first stand cDNAs using an iScript cDNA kit (Bio-Rad laboratories, CA, USA). Total cDNA yields from the equimoler amounts of total RNA from each selected wild species accession were diluted to 1/10, and used as template for qRT-PCR performed with three replicates. Real-time PCR was performed in MyiQ™ thermal cycler (Bio-Rad Laboratories, CA, USA) using the iQ™ SYBR® Green Supermix (Bio-Rad) according to the manufacturers protocol. The thermal profile of the real-time system was one step at 95 °C for 30 s, followed by 40 cycles at 95 °C for 5 s (denaturation) and at 60 °C for 30 s (annealing and extension) followed by an added dissociation pattern.

OsPLDα1 genomic loci and transcript sequences (Locus ID Os01g0172400), were obtained from the RAP data base (http://rapdb.dna.affrc.go.jp/viewer/gbrowse/). Four exon-specific qRT-PCR primer pairs (PLDE1, PLDE2.1, PLDE2.2, and PLDE3) (Supplementary Table S2) were designed from the coding sequence of OsPLDα1 gene using Primer-BLAST tool (http://www.ncbi.nlm.nih.gov/tools/primer-blast/). In addition, a primer pair (sense primer 5′- CATGAAGATCAAGGTTGTCGCTCC-3′ and antisense primer 5′- CCAGATTCTTCATACTCAGCCCTTG-3′) was also designed for actin gene (Locus ID Os10g0510000). Each primer was dissolved in 1X TE buffer (stock solution) to have master stock of 100 µM.

Transcript abundance was normalized relative to the baseline constitutive expression of a reference gene i.e., actin-2 (ACT2). Relative expression levels were calculated using the \(2^{{ - \Delta {\text{C}}_{\text{T}} }}\) method [30]. According to this method, data have been normalized against the selected internal control (ACT2) to get ∆CT values.

$$\Delta {\text{C}}_{\text{T}} = \, \left[ {{\text{C}}_{\text{T}} \left( {{\text{gene}}\,{\text{of}}\,{\text{interest}}} \right) - {\text{C}}_{\text{T}} \left( {{\text{internal}}\,{\text{control}}} \right)} \right] \, \left( {\text{SampleA}} \right)$$

Calculated ∆CT values of each of the three replicates of a sample were then log transformed into \(2^{{ - \Delta {\text{C}}_{\text{T}} }}\) value. Mean of the \(2^{{ - \Delta {\text{C}}_{\text{T}} }}\) values was used to measure relative expression of different samples and to generate the heat map. Standard deviation for each sample was also calculated using the \(2^{{ - \Delta {\text{C}}_{\text{T}} }}\) value of the replicates of a sample.

Results

Detection of SNPs in OsPLDα1 promoter

The 1.6-kb upstream region of the OsPLDα1 promoter obtained from selected wild species accessions and cultivars was used for SNP/InDel detection. Multiple alignments of OsPLDα1 promoter sequences to the Nipponbare resulted in detection of 39 SNPs (Table 2). Of the detected SNPs, 32 were identified in the 1.2 kb region upstream of the translation start site. These SNPs represent base transitions (22) which were more frequent than base transversion (17) mutations. Interestingly, all the accessions of O. glaberrima were found to be sharing similar mutations in OsPLDα1 promoter, based on their alignment with the Nipponbare promoter. Moreover, SNPs C/T−1072, A/C−227, and A/C−226 were found specific to the O. glaberrima species. Out of the 9 SNPs identified in O. barthii accessions, C/G−1024 and G/C−1021 were specific to the IRGC106294 accession; C/A−1071 and G/A−588 were detected in the IRGC100117 accession and in O. glaberrima accessions as well; and the remaining five SNPs were detected in all the accessions of O. barthii and O. glaberrima. Among the 15 SNPs detected in five O. nivara accessions, only two SNPs including A/G−1108 and A/G−443, detected in the IRGC105880 accession, were found specific to this species while rest of the SNPs were present in the accessions of other species as well.

Table 2 The nucleotide variations observed in the OsPLDα1 promoter sequence of wild Oryza species accessions and Oryza cultivars in comparison to the Nipponbare OsPLDα1 promoter
Full size table

Of all the wild species and cultivars under study, the maximum number of SNPs (24) has been identified in the O. rufipogon species from which 11 were found specific to the O. rufipogon accessions (Table 2). In addition, two small insertions including CTC−253 and GCTT−128 have also been detected in the O. rufipogon accessions. The CTC insertion was found in IRGC89224, IRGC104425, and IRGC105902 accessions while the GCTT insertion was found in IRGC104308, IRGC103308, IRGC106162, and IRGC81976 accessions. Of the 11 SNPs identified in the eight cultivars under study, only 3 including C/A+37, A/C+39, and A/C+93 were found specific to the cultivar IR64 while rest of the SNPs were found in the wild species accessions as well. Kitake, in comparison to other cultivars, showed minimal variations as the SNPs C/T−1192, C/A−1064, G/T−858, and A/G−455 were detected in all the cultivars except Kitake.

Phylogenetic relationships among Oryza accessions based on OsPLDα1 sequence variation

Phylogenetic analysis of the OsPLDα1 promoter region revealed the divergence of 26 wild Oryza species accessions and 8 Oryza cultivars from the reference sequence of japonica cultivar Nipponbare (Fig. 1). Phylogenetic tree showed that all accessions of the wild Oryza species having African centre of origin, O. glaberrima, O. barthii, and O. longistamina, were more closely related to each other in comparison to the accessions of O. rufipogon and O. nivara species, which originated in South-Asian nations. Moreover, three O. rufipogon accessions (IRGC89224, IRGC104425, and IRGC105902) were found to be at the maximum distance from the reference, thus highlighting the presence of useful sequence variations which may further lead to the presence/absence of cis-elements involved in regulation of OsPLDα1 expression. Furthermore, differences observed between the indica and japonica cultivars were greater than within indica and japonica cultivars as japonica cvs. Nipponbare and Kitake were more similar to each other than to indica cvs. IR64, N22, Pusa44, PR114, PAU201, Minghui63, and Feng-Ai-Zhan.

Fig. 1
figure 1

Phylogenetic relationship across Nipponbare (Locus ID Os01g0172400), wild species accessions, and cultivars of rice based on OsPLDα1 promoter sequence variation. The neighbor-joining tree was drawn based on the nucleotide sequences of the OsPLDα1 promoters from 26 wild Oryza species accessions, 8 Oryza cultivars, and using Nipponbare OsPLDα1 promoter as the reference sequence. In all, 1,000 bootstrap replicas were applied. The tree indicates the three Oryza rufipogon accessions (IRGC105902, IRGC104425 and IRGC89224) at maximum distance from Nipponbare. Japonica cultivar Kitake shows close relatedness to the Nipponbare in comparison to the indica cultivars (IR64, N22, Pusa44, PR114, PAU201, Minghui63, and Feng-Ai-Zhan)

Full size image

Regulatory sequences in OsPLDα1 promoter

The analysis revealed the presence of 92 classes of sequence motifs representing known and novel regulatory cis-elements (341 including copy numbers) in the OsPLDα1 promoters of the wild species accessions and cultivars of rice. Further, comparison of the identified cis-elements with Nipponbare showed the copy number variations of 48 cis-elements (Table 3). In addition, 10 cis-elements including ARFAT, IBOXCORENT, MYB1LEPR, NTBBF1ARROLB, PYRIMIDINEBOXOSRAMY1A, ROOTMOTIFTAPOX1, SEBFCONSSTPR10A, SORLIP2AT, TBOXATGAPB, and TGTCACACMCUCUMISIN, which were detected in at least one of wild species accession, were found missing in Nipponbare. On the contrary, ACGTOSGLUB1 was detected in Nipponbare, but found missing in four wild species accession (CR100429, IRGC92713, IRGC105880, and IRGC80610). The role played by cis-elements, in regulating the gene expression, and their location on the promoter sequence was identified (Supplementary Table S3) and most of the regulatory elements (268 including copy number) were found to be located in the upstream region of the promoter. Furthermore, mapping and alignment of the cis-elements identified in the OsPLDalpha1 promoter of Nipponbare to the wild species accessions showed the relative position of all the detected cis-elements, variations in copy numbers, and presence/absence of cis-elements between as well as within the accessions of different species (Fig. 2).

Table 3 Presence/absence and copy number variations of cis-elements in the OsPLDα1 promoter region of wild species accessions and cultivars of rice in comparison to the reference promoter of Nipponbare
Full size table
Fig. 2
figure 2figure 2figure 2

Cis-regulatory elements maps in the OsPLDα1 promoter (1.2 kb upstream and 0.4 kb downstream of the Translation Start Codon) in the Nipponbare and wild Oryza species accessions. The colored shapes on the promoter sequence indicate the position of the respective cis-elements. This mapping and alignment of cis-elements indicates the presence/absence as well as copy number variations between as well as within the accessions of different species

Full size image

Differential expression of OsPLDα1

The OsPLDα1 gene expression was compared across the accessions selected on the basis of presence/absence and copy number variations of cis-elements in OsPLDα1 promoter. Expression profiling, generated using four exon-specific qRT-PCR primer pairs (Supplementary Table S2), indicated significant expression differences in the selected accessions (Fig. 3). The four qRT-PCR primer pairs were designed in a way to determine expression differences as well as abundance of different OsPLDα1 transcript forms (Fig. 5). Expression analysis using qRT-primer PLDE1, designed from the first exon (Fig. 3a), revealed the presence as well as variations in abundance of 5′ truncated OsPLDα1 mRNA in the accessions. Minimum expression for first exon was detected in the IRGC104425 accession of O. rufipogon while IRGC105279 accession of O. longistaminata had the highest expression levels. Further, expression levels for qRT-PCR primers designed from 5′ (PLDE2.1) and 3′ (PLDE2.2) ends of second exon, revealed differences in the expression levels within the same accession (Fig. 3b, c). It suggests the presence of truncations, within the second exon of OsPLDα1 gene, during the process of transcription. For the second exon, IRGC89224 had the lowest level for 5′end transcripts while for the 3′end transcript levels were lowest and comparable in IRGC89224 and IRGC105902 accessions. In addition, there were differences between all the selected accessions for the abundance of transcripts containing 3′end and 5′end of the second exon.

Fig. 3
figure 3

The expression patterns of OsPLDα1 gene in the wild Oryza species accessions. a The expression patterns for qRT-PCR primer (PLDE1) designed from first exon of the gene. b The expression patterns for qRT-PCR primer (PLDE2.1) designed from 5′end of second exon. c The expression patterns for qRT-PCR primer (PLDE2.2) designed from 3′end of second exon. d The expression patterns for qRT-PCR primer (PLDE3) designed from third exon of the gene. The actin gene (Locus ID Os10g0510000) was used as an internal control to normalize gene expression levels

Full size image

Lowest level of OsPLDα1 mRNA transcripts having third exon (Fig. 3d) were detected in the IRGC89224 accession of O. rufipogon while the highest levels were observed in O. barthii accession (IRGC106294). Among all the wild species accessions, IRGC105569 accession of O. rufipogon was found closely related to the Nipponbare cultivar, and thus it was used as control to compare the transcript levels in the other wild species accessions. The levels of cumulative relative expression for all the four exon specific primers were 0.20-, 0.26-, 0.41-, 0.44-, 0.61-, 0.62-, 0.75-, 1.37-, 1.91-, and 2.49-fold in IRGC89224 (O. rufipogon), IRGC105902 (O. rufipogon), IRGC104425 (O. rufipogon), IRGC106162 (O. rufipogon), IRGC104308 (O. rufipogon), CR100013 (O. rufipogon), IRGC103308 (O. rufipogon), CR100400 (O. nivara), IRGC106294 (O. barthii), IRGC102489 (O. glaberrima) accessions in comparison to the control accession (IRGC105569). Thus, the expression profiling showed lowest expression of OsPLDα1 gene in IRGC89224 followed by IRGC105902 and IRGC104425 accessions of O. rufipogon, while O. nivara, O. barthii, and O. glaberrima accessions had OsPLDα1 gene expressions higher than the control. Furthermore, hierarchical clustering dendrogram represents that expression values for all the primers in IRGC104425 and IRGC105902 accessions of O. rufipogon were falling in the green category (low expression values); however, the black color obtained in case of IRGC104425 was more skewed towards the lower side i.e., the green side (Fig. 4). For all the other accessions, at least one of the exon specific expression values was found falling in the red category and/or was more skewed towards the higher value (depicted in the red color). Thus, in addition to IRGC89224 and IRGC105902 accessions, IRGC104425 accession of O. rufipogon was concluded to have low level of OsPLDα1 gene expression.

Fig. 4
figure 4

Heatmap showing differential expression of OsPLDα1 transcripts between as well as within the accessions of wild Oryza species. PLDE1, PLDE2.1, PLDE2.2 and PLDE3 denotes the exon-specific qRT-PCR primers designed from different exons of OsPLDα1. Wild species accessions (horizontal) were hierarchially clustered (Pearson correlation, average linkage). Color patterns from green to red indicate low to high transcript levels, thus IRGC89224 had the lowest expression for all the four exon specific qRT primers

Full size image

Discussion

The PLD enzyme has been known responsible for the rapid degradation of rice bran oil, and consequently severely limits the use of rice bran for human consumption. Triacylglycerols, the main lipids of rice bran, occur in oil bodies with phospholipid membranes that are disintegrated by PLD [31]. Ueki et al. (1995) purified a PLD protein (OsPLDα1) from the rice bran fraction which plays an important role in the rice bran oil rancidity. However, there is no report in literature on detection of superior OsPLDα1 promoters in the wild germplasm of rice to improve the rice bran quality. Moreover, the molecular mechanisms of OsPLDα1 gene expression are still unclear. In this study, the OsPLDα1 promoter regions in wild Oryza species accessions and Oryza cultivars were screened to identify the novel cis elements and to describe the regulation of gene expression. We identified SNPs and InDels at OsPLDα1 promoters in the wild germplasm of rice in comparison to the reference sequence of Nipponbare. Interestingly, three accessions of O. rufipogon including IRGC89224, IRGC104425 and IRGC105902 were found to harbor useful variability which contributed towards lower expression of OsPLDα1 gene. The wild species of Oryza serve as an important reservoir of genetic variability for agronomic traits such as biotic and abiotic stresses and for improved yield potential [32,33,34,35]. Many useful genes viz. genes for resistance to bacterial blight (BB), blast, tungro virus, grassy stunt virus; tolerance to soil toxicity (acid sulphate, iron toxicity); and cytoplasmic male sterility have been transferred from wild species of rice [36, 37]. Among the most successful examples of utilizing wild species is the use of Oryza nivara genes to provide long-lasting resistance to grassy stunt virus [38]. Similarly, O. rufipogon has been the source of beneficial alleles for diverse traits including grain size, grain weight [39, 40], grain yield [41,42,43], grain quality [44], cold tolerance [45], aluminium tolerance [46], and flowering time [47]. In the similar fashion, O. rufipogon accessions showing variations in the OsPLDα1 promoter region could serve as the useful source to improve the rice bran quality in elite rice lines.

Further, our results on molecular diversity at OsPLDα1 promoter, revealed the presence of lower diversity between Nipponabre and japonica cultivar (Kitake) in comparison to the indica cultivars (N22, IR64, Pusa44, PR114, PAU201, Minghui63, and Feng-Ai-Zhan). The study conducted on the assessment of genetic diversity using both nuclear and cytoplasmic molecular markers also showed that indica and japonica are associated with different subgroups of O. rufipogon [48,49,50]. The accessions of African rice (O. glaberrima) and its wild progenitors viz. O. barthii and O. longistaminata were found closely related to each other. Similar results were obtained in the genetic diversity study conducted by using multiple gene sequences by Li et al. [51]. Similarly, the O. nivara accessions were found closely related to most of the O. rufipogon accessions. In a recent study, Kim et al. (2016) have grouped O. nivara and O. rufipogon together in a single O. rufipogon Griff. species complex (ORSC) on the basis of ease of cross-hybridization between the two species, supporting our results [52]. Our results also suggest the existence of lesser diversity among the O. rufipogon accessions belonging to Southeast Asian nations. Similar results were obtained during the several studies conducted for evaluating the genetic diversity of Asian wild rice using RFLP, microsatellite markers, SINEs, sequence based polymorphism, ISSRs, chloroplast, and low-copy nuclear markers [53,54,55].

Furthermore, to evaluate the expression patterns of OsPLDα1 gene in selected wild species accessions, the appropriate stage of plant development was selected by performing expression profiling of OsPLDα1 and its 16 other paralogs using the expression data from RiceXProv3.0 database. Expression profiling at 49 different time points and using different plant tissues including eight infloresence (I1–I8), four anther (A1–A4), three pistil (P1–P3), three lemma (L1–L3), three palea (PA1–PA3), ten ovary (O1–O10), twelve embryo (EM0–EM42), and six endosperm (EN07–EN42) stages revealed that OsPLDα1 keeps on expressing under all the stages of plant development in each selected tissue (Supplementary Fig. S1). Further, an experiment conducted by Suzuki (2011) demonstrated that the PLD content in embryos of Nipponbare increased during seed development up to 3 wk after flowering, becoming constant thereafter. Moreover, PLD-null rice mutant (03-s108), having < 0.01% PLD enzymatic activity in rice bran fraction when compared to Nipponbare, showed no PLD protein band in the seeds 1 week after the flowering. It concluded that the reproductive tissue at early stage of grain development could be used for RNA isolation for OsPLDα1 expression studies. Therefore, for the present study, immature seeds (1 week after flowering) were selected for RNA isolation to carry out the expression studies.

For expression profiling, four exon-spcific qRT-primer pairs (PLDE1, PLDE2.1, PLDE2.2, and PLDE3) were designed from the coding sequence of OsPLDα1 (Supplementary Table S2). The exon-specific primers were designed in a way to evaluate the selected accessions for differences in the gene expression as well as to reveal if the gene is alternatively spliced (Fig. 5). qRT-PCR generated expression profiles revealed the presence of interspecific and intraspecific differences in the OsPLDα1 transcript abundance (Fig. 3). In addition, it also indicated the existence of significant expression variations between the different accessions for the same splice form as well as within the same accession for different splice forms. The expression differences observed in this study could arise from cis-regulatory changes that affect transcription initiation, transcription rate and/or transcript stability in an allele-specific manner or from differences in trans-regulatory changes that modify the expression of factors that interact with cis-regulatory sequences [56, 57]. However, Wittkopp et al. (2004) reported that the expression differences are not caused by trans-regulatory changes with widespread effects, but rather by many cis-acting changes spread throughout the genome [58]. In this study, significant expression differences were observed for the transcripts obtained with primer designed from the 5′ (PLDE2.1) and 3′ (PLDE2.2) ends of the second exon, which reveals the presence of truncations, within the second exon of OsPLDα1 gene, during transcription. In addition, the IRGC106294 accession of O. barthi and IRGC105569 accession of O. rufipogon had high abundance of third exon containing transcripts, in comparison to the transcript abundance for the other two exons (Fig. 3d). The probable reason for this could be the presence of a new transcript form which consists of only third exon, in addition to the five already existing alternate splice forms of OsPLDα1. Alternate splicing is a widespread process used in higher eukaryotes to regulate the gene expression and functional diversification of proteins. It has been reported that alternative splicing of the Ca2+-independent phospholipase A2 (iPLA2) pre-mRNA in humans can result in the production of regulatory subunits that can modify iPLA2 in vivo activity [59].

Fig. 5
figure 5

Architecture of five transcipt variants of OsPLDα1 gene. It shows the location of four qRT-PCR primers (PLDE1, PLDE2.1, PLDE2.2, and PLDE3) which were designed from different exons, in a way to evaluate variations in relative abundance of different splice forms. The transcript form Os0t0172400-01 has the similar architecture as of the OsPLDα1 gene, and thus shows the position of exons/introns of the gene under study

Full size image

Also, it has been observed that primer PLDE1 shows low levels of expression of the gene for all the tested accessions (Fig. 4). The reason for this could be the presence of first exon only in three of the five OsPLDα1 transcript forms (Fig. 5), indicating that it might not play an important role towards determining the OsPLDα1 expression. Moreover, first exon is 108 base pairs in size which is the shortest one in comparison to the size of second (1897 base pairs) and third (243 base pairs) exons. PLDE1 denotes the exon specific primer designed on the first exon which covers the position from 7 to 101 nucleotide bases, and thus generates the amplicon of only 95 base pairs size. In contrast, the amplicon sizes obtained using the primers designed from 5′ end of second exon (PLDE2.1), 3′ end of second exon (PLDE2.2), and from third exon (PLDE3) were 116, 122, and 223 base pairs, respectively (Table S2). Due to the shorter amplicon length obtained with PLDE1, in comparison to the other three primers, it showed relatively lower expression for all the accessions under study.

The knowledge of the regulatory mechanisms for OsPLDα1 gene would help in the understanding of how the gene expression is controlled in plants. Unfortunately, however, there is no literature regarding promoter elements, transcription factors, or molecular mechanisms involved in the regulation of OsPLDα1 gene. To begin to explore these questions, we performed in silico analysis of the OsPLDα1 promoters, from wild species accessions and cultivars of rice. The results revealed copy number variations and presence/absence of cis-elements in comparison to the Nipponbare, and it is well known that the promoter activity of a gene is largely affected by the copy number, inter motif distance, position, deletion or mutation in the core sequences of the cis-regulatory regions [60]. In IRGC104425 accession, one copy of GT1CONSENSUS, IBOXCORE, NODCON2GM, OSE2ROOTNODULE, and SEF4MOTIFGM7S; two copies of SURECOREATSULTR11 elements were found missing. However, two copies of CGCGBOXAT and one copy of SORLIP1AT elements were found missing in IRGC105902 accession. The alteration in number of cis-elements is a consequence of segmental duplication over the course of evolution. Also, some conservative motifs lost their previous function and some specific elements appeared among genomes during evolution [61]. Owing to the changes in cis elements, importance of certain genes increased or decreased over the time, thereby, requiring an increase or decrease in its expression [62]. Moreover, the accumulated mutations in the promoter region of a gene have been reported to be the underlying cause for restricted expression pattern in cultivars as compared to the broad expression pattern of its wild type allele in the wild species of Oryza [63, 64]. Different cis-regulatory elements function differently when their copy number, distance between multiple copies, and distance from the TATA box are altered [65]. A study has already confirmed that GT elements play different functions depending on their copy number [66]. Therefore, variations in the copy number of GT element and other cis elements as detected in the current study, could be responsible for low OsPLDα1 gene expression in IRGC104425 and IRGC105902 accessions.

Furthermore, we report in this study presence of SEBF (TTGTCTC; Silencing Element Binding Factor) and ARFAT (TGTCTC; auxin response factor) elements within the OsPLDα1 promoter of O. rufipogon accession IRGC89224, which were absent in Nipponbare, rice cultivars, and other wild Oryza species. The identified elements shared the motifs and location in the upstream region of OsPLDα1 gene (Table S3). Auxin is recognized as the universal plant growth hormone because of its principal role in the regulation of cell expansion, division and differentiation [67]. Aux/IAA regulates its own transcription, and functions as transcriptional regulator of other auxin responsive genes in cooperation with Auxin Response Factors (ARF) [68]. In the absence of auxin, Aux/IAA inhibits ARF function and represses other auxin early responsive genes. Moreover, it has been estimated that between 2 and 5% of the total genes in Arabidopsis and rice are either upregulated or downregulated by auxin [69]. Many studies have revealed that the cis-acting elements function as activators or repressors in regulating tissue-specific gene expression which is governed by the combined action of cis-acting elements in its promoter region, as well as different nuclear proteins interacting with the elements [70]. There are several defense genes which have been known to be negatively regulated be auxin [71]. Also, PLD gene is reported to be involved in the plant’s response to pathogenic elicitation, seed germination, seedling growth, and leaf senescence, in addition to signaling by various plant hormones and stress responses [72,73,74,75,76]. However, identification, cloning, and expression of novel types of plant PLDs established that they are a family of heterogeneous enzymes that differ in catalytic and regulatory properties [77,78,79,80]. Through genome-wide analysis in rice, 17 PLD members have been found on different chromosomes including PLD α(8), β(2), γ(3), κ, ζ(2) and φ [17]. The function of plant PLDs has been studied using knockout, knockdown, and over expression of corresponding genes, resulting in the identification of the physiological functions of several PLDs. Many studies conclude that there are different isoforms of PLD which are responsible for similar kind of plant functions and defense responses [81,82,83]. Moreover, Suzuki (2011) identified OsPLDα1 null mutant plants, having no rice bran deterioration, which grew normally in a paddy field indicating that seed PLD (OsPLDα1) is not necessary for seed maturation and germination [20]. These studies indicate that down regulation of the OsPDLα1 activity, by the action of ARFAT and SEBF elements, will not compromise the other functions of the plant. Hence, ARFAT and SEBF elements which appeared to act as repressors in regulating the OsPLDα1 gene expression have been considered as ‘novel’ elements for maintaining the rice bran quality.

Conclusion

In summary, our efforts for identifying novel cis-elements in the OsPLDα1 promoter of wild Oryza species accessions revealed the presence of ARFAT and SEBF cis elements in the Oryza rufipogon accession (IRGC89224) which had the lowest expression of OsPLDα1 enzyme as well. Though, the further confirmation of low enzyme activity should be done by estimating the OsPLDα1 activity in the rice bran by biochemical means. For this purpose, maintenance and seed multiplication of the identified low expressing wild species accessions has been undertaken at the Punjab Agricultural University to produce the amount of rice bran (~ 10 g) needed to estimate the OsPLDα1 activity. In addition, the O. rufipogon accessions (IRGC89224, IRGC105902, and IRGC104425), with low OsPLDα1 expressions, have been deployed as the potential donors in the back-crossing programs to transfer the desired trait into elite cultivars of rice.