Transcriptome Analysis of Cocos nucifera L. Seedlings Having Contrasting Water-Use Efficiency (WUE) under Water-Deficit Stress: Molecular Insights and Genetic Markers for Drought Tolerance ^†

Abstract

Perennials utilize complex adaptive strategies and molecular mechanisms to cope with water-deficit conditions. Hence, in order to gain molecular insights regarding water-deficit stress, two-year-old coconut seedlings of the varieties Kalpa Sree and Kalpatharu were subjected to soil water-deficit regimes (25% of available of soil moisture and control). The biochemical, physiological and growth parameters underlying water-deficit stress revealed the differential enzymatic anti-oxidants, lipid peroxidation status and water use efficiency traits between the genotypes investigated. The entire plant water use efficiency at the control condition was significantly low in Kalpatharu (4.06) compared to in Kalpa Sree (4.74). Nevertheless, under severe stress (25% ASM), Kalpatharu exhibited the highest WUE (5.68) against dwarf variety Kalpa Sree (3.84). Furthermore, the leaf transcriptome profiles of the control and water-deficit stressed seedlings were examined by utilizing paired-end RNA-Seq. In total, ~7300 differentially expressed genes have been identified between the seedlings under water-deficit stress and control. Analysis of control and stressed Kalpasree leaf transcriptome showed significant upregulation of PHLOEM PROTEIN 2-LIKE A1-like, WRKY transcription factor 40 isoform X1 and downregulation of glycerol-3-phosphate acyltransferase 3 transcripts. On the other hand, the upregulation of transcripts encoding polyamine oxidase, arabinose 5-phosphate isomerase among others and downregulation of aquaporin PIP1-2 transcript was documented in Kalpatharu leaves. Moreover, long non-coding RNA and genic SSRs were also identified from the transcriptome data to further enrich the genomic resources of coconut palm, which could pave way for its utilization in developing climate-smart coconut crop. The implications of this study in molecular dissection of the adaptive response of coconut to the soil-water deficit are also discussed.

1. Introduction

Coconut palm (Cocos nucifera L.) is an important plantation crop of humid tropics. Coconut is grown in 13 Mha spread over 90 countries, including the Philippines, Indonesia, India, Brazil and Sri Lanka with an estimated production of 69,836.36 million nuts. In India, coconuts are grown in an area of 2.088 Mha with a production of 22,167.45 million nuts [1]. Among the various biotic and abiotic stresses that hinder the productivity of coconut, moisture-stress profoundly influences its annual yield. The amount of rainfall and dry spells preceding 4 years [2,3] and weather variables [4] also influence the productivity of coconuts. Furthermore, a dry spell of about 200 days could further reduce the yields of coconut palms [4]. It has also been documented that the coincidence of dry spell with sensitive stages such as inflorescence and nut development greatly influences the yield of palms. Eventhough the assessment of impact of climate change on coconut cultivation emphasizes that coconut plantations would benefit owing to increased CO₂, the caveat is that it requires appropriate quantity of moisture supply. Furthermore, the genetic improvement of local tall cultivars adapted for moisture-deficit stress and non-limiting supply of other inputs is also suggested as a long-term measure for managing water stress in coconuts. Considering the importance of drought/water stress in coconut plantation sector, ICAR-CPCRI has devised research programs for phenotype coconut varieties, to studies devising suitable drought mitigation systems and to studies developing climate smart coconut. In order to achieve this objective, rigorous screenings of tall and dwarf genotypes of coconuts are pertinent for water use efficiency (WUE) and for analyzing their molecular response.

2. Experiments

2.1. Coconut Genotypes

Seedlings of coconut varieties Kalpa Sree and Kalpatharu grown in large plastic buckets (64 × 49 cm) of 100 kg dry soil capacity were investigated for their water use efficiency (WUE) and the morphological, physiological, biochemical and molecular changes linked to water deficit stress.

2.2. Drought Experiment

Coconut seedlings of each variety were exposed to moisture regimes: 100% available soil moisture (ASM)-control; and 25% ASM-stressed. Briefly, after overnight saturation of soil, the amount of water that must be provided daily to maintain the 100% ASM (20% moisture content) (https://nrcca.cals.cornell.edu/soil/CA2/CA0212.1-3.php, accessed on 25 July 2017) was estimated. Soil moisture was measured by using a soil moisture probe (PR2 profile probe, delta T devices, UK) supplemented with a data logger. Every day, the soil moisture content was measured, and the amount of water required to bring the soil back to 100% ASM was determined and recorded. Stressed seedlings received 25% of the quantum of water as that of control.

2.3. Biochemical Characterisation

Epicuticular wax content was estimated following the method of Ebercon et al. (1977) [5]. Lipid peroxidation was estimated as malondialdehyde (MDA) equivalent following the methodology enumerated by Heath and Packer (1968) [6]. Assays for enzymatic antioxidant were performed [7,8].

2.4. RNA Extraction, Library Preparation and Sequencing

Based on the physiological data, coconut leaf samples subjected to moisture deficit stress and control were collected in the morning (around 8:00–9:00 a.m., IST) and flash frozen in liquid nitrogen. Total RNA was isolated following guanidium thiocynate method, and the quality of the isolated RNA was assessed on 1% formaldehyde denaturing agarose gel and quantified using Qubit^® 2.0 Fluorometer. The libraries were prepared with total RNA (~1 μg) following Illumina TruSeq Stranded mRNA Library Preparation Kit as per the manufacturer’s protocol. The amplified libraries were analyzed on Bioanalyzer 2100 (Agilent Technologies) using High Sensitivity (HS) DNA chip as per manufacturer’s instructions. The library was loaded into an Illumina platform for cluster generation and sequencing.

2.5. Transcriptome Analysis

A master de novo assembly was generated by using Trinity v2.1.1 (at default parameters, kmer 25). Transcripts were further processed for unigenes prediction using CD-HIT package v 4.6.1. The unigenes were subjected to similarity search against NCBI’s non-redundant (nr) database, Uniprot, KOG and Pfam databases by using the BLASTX algorithm v2.2.30 (e-value threshold of 1 × 10⁻⁵). Ortholog assignments and mappings of the unigene to the biological pathways were performed by using the KEGG automatic annotation server (KAAS). Simple sequence repeats (SSRs) were identified in unigene sequences of master assembly using the MISA perl script. For calculating the fold change in gene expression, the ratio of basemean value of the treated sample and basemean value of the control sample (i.e., FC = basemean of Treated/basemean of Control) was considered, and log 2 (FC) values and p-values were also calculated to infer the significance of genes. Differential gene expressions among the genotypes were assessed using DESeq (bioconductor). Plant related lncRNA sequences were obtained from ensemble plant database present in RNA central v-14, and the BLASTN homology search was performed taking 85,516 unigene sequences (having no blast hit against NR database) against the RNA central database with an e-value threshold of 1 × 10⁻⁵.

3. Results

3.1. Water Use Efficiency (WUE)

The whole plant WUE significantly differed amongst genotypes. At 100% ASM it was significantly low in KT (4.06) compared to Kalpa Sree (4.74). Both the biomass accumulation and water consumption were low in Kalpatharu compared to that of Kalpa Sree. However, under severe water stress (25% ASM), the tall variety kalpatharu had the highest WUE (5.68) as against dwarf variety Kalpa Sree (3.84). This suggested that, under water, limited condition tall genotypes are more efficient in utilizing water and are more adaptive than compared to dwarfs.

3.2. Physico-Chemical Parameters

Epicuticular wax (ECW) is known to act as a barrier for the loss of water through transpiration from the plants. Kalpatharu grown under moisture stress (25% ASM) exhibited the highest ECW (88.59 µg cm²), while the leaves of the Kalpa Sree showed ECW content of 54.63 µg cm². The mean value of MDA (Malondialdehyde), which is an indicator of lipid peroxidation, was high in stressed plants grown under 25% ASM (31.8854 mmol g⁻¹fr.wt), and the MDA content decreased with water status of the plants. Among the enzymatic antioxidants, SOD showed an increase in specific activity from 3.49 to 5.98 units/mg of protein in KT whereas, in the leaves of Kalpa Sree seedlings, the increase was minimal from 2.35 to 2.51 units/mg of protein. However, both the genotypes showed a huge increase in the specific activity of polyphenol oxidase (Kalpa Sree from 0.14 to 1.18; Kalpatharu 0.16 to 1.37), although kalpatharu show relatively a high increase in PPO activity.

3.3. Transcriptomic Features of Drought Tolerance

Four cDNA libraries yielded approximately 158.96 million bp reads with an average transcript length of 829 bp and transcript N50 value of 1225. The analysis of thetranscript length distribution revealed that a maximum of 68,482 transcripts belonged to class 300–400 bp. The transcripts were further processed for unigene prediction using CD-HIT package v 4.6.1, which predicted a total of 162,483 unigenes with a mean unigene length of 786 bp. A maximum of 60,676 unigenes predicted from the transcripts belonged to length class 300–400 bp. The master unigene analyzed for functional annotation and expectedly the top-hit species distribution revealed that the majority of the hits were found to be against the related species Elaeis guineensis followed by Phoenix dactylifera. EuKaryotic Orthologous Groups or KOGs analysis showed that the most enriched KOG categories for unigenes were (a) signal transduction mechanisms (T), followed by (b) general function prediction only and (c) post-translational modification, protein turnover and chaperones (O). In Pfam analysis, the most abundant domains identified represented “Pkinase”, followed by “Pkinase_Tyr” and “PPR_2”. The comparative counts of unigene annotatin using various databases are presented in Figure 1.

3.4. Differentially Expressed Transcripts

Analysis of transcripts derived from the leaves of stressed and control Kalpa Sree seedlings revealed that 2388 transcripts are significantly upregulated and 1278 are significantly downregulated, whereas similar analysis of Kalpatharu transcripts showed significant upregulation and downregulation of 2868 and 778 transcripts, respectively (Figure 2a,b). Moreover, significantly expressed genes (i.e., highly up and downregulated genes) are represented in the form of a heatmap using pheatmap package of R, following hierarchical clustering approach (Figure 3a,b). The analysis of control and stressed Kalpa Sree leaf transcriptome showed significant upregulation of PHLOEM PROTEIN 2-LIKE A1-like (Log2 FC 11.19) and WRKY transcription factor 40 isoform X1 (Log2 FC 9.59) and downregulation of glycerol-3-phosphate acyltransferase 3 transcripts. Significant upregulation of transcripts encoding polyamine oxidase (Log2 FC 9.78), arabinose 5-phosphate isomerase (Log2 FC 6.52) and WRKY transcription factor 40 isoform X1 (Log2 FC 5.57) among others was record. Furthermore, downregulation of aquaporin PIP1-2 transcript (Log2 FC −1.8) and ethylene responsive transcription factor ERF105 (Log2 FC −1.9) was documented in Kalpatharu leaves.

Pathway analysis of unigenes divulged that, in the primary metabolic pathways, carbohydrate and amino acids metabolism are enriched. In the genetic information processing pathways, protein translation and degradation or turnover pathways are highly enriched compared to transcriptional changes. In environmental information processing, signal transduction pathways are significantly enriched compared to membrane transport or signaling molecule interaction.

3.5. Genic Markers for Drought Tolerance

Genic simple sequence repeats (SSRs) or genic microsatellites or expressed sequence tag (EST)-SSRs are novel resources not only as molecular markers but also for providing invaluable information regarding the adaptation of a species towards external stressors such as drought or moisture stress. In this study, around 17,296 genic SSRs were identified from the transcriptome sequences of coconut, with potential for application in molecular breeding. The most common SSRs were di-nucleotide repeats (10,826) followed by tri-nucleotide repeats (5699); quadnucleotides, pentanucleotides and hexa nucleotides are of 633, 89 and 49, respectively. Among the di-nucleotide repeats, GA/TC (1904) was most abundant, followed by CG/CG (1849), AG/CT (1803) and TC/GA (1573), among others.

3.6. Differential Expression of Long Non-Coding RNAs

Analysis for expression of long non-coding RNAs (lncRNAs) in transcriptome data set revealed that a total of 63 Unigenes are potential lncRNAs. Differential expression analysis of those lncRNAs revealed that in Kalpa Sree, 32 lncRNAs are differentially regulated whereas, in Kalpatharu, 59 lncRNAs are differentially regulated.

4. Discussion

Climate change is expected to increase mean annual temperatures and to decrease the precipitation and sea water incursions in coastal and low-lying areas. Plantation crops that are perennials and possessing relatively long gestation period to yield economic produce are under major threat to water-deficit stress due to decline in precipitation [9,10]. In this context, the traditional coconut growing regions are expected to see a shift in area coverage due to reduced rainfall in future climate scenario [9]. Hence, it is pertinent to develop climate-smart coconut which can withstand water-deficit stress [11,12]. Unfortunately, the genomic resources of coconut—an economically important plantation crop—are relatively poor. Accordingly, this study was designed with an objective of identifying suitable genetic markers linked to water-deficit stress. We have chosen two coconut varieties Kalpa Sree (dwarf) and Kalpatharu (tall) of distinct morphoforms possessing differential WUE and biochemical or physiological response in order to analyze their molecular response during water deficit stress by utilizing RNA Seq approach. Our study has delineated that the transcriptome response of both the varieties varies significantly as Kalpa Sree showed upregulation of genes involved in phloem protein such as A1 and WRKY transcription factor 40 isoform X1 and downregulation of transcripts involved in metabolism. Kalpatharu, on the other hand, downregulated aquaporin PIP1-2 transcript and ethylene responsive transcription factor to tide over water deficit stress and to improve its WUE. The differential gene expression pattern is also corroborated by the differences in the enrichment of KEGG metabolic pathways and KOG analysis of the transcript data sets suggesting differential molecular and biochemical responses of the varieties.

5. Conclusions

This investigation elucidates the genotype-specific differential molecular responses of an economically important plantation crop during water deficit stress for which no prior genome-level response for water-deficit stress was available. Furthermore, the EST-derived SSRs uncovered in this study would add to the much needed molecular marker data set for developing climate-smart coconut. It also underscores the importance of lncRNAs’ repertoire in shaping up the adaptive response of coconut to water stress. To the best of our knowledge, this is the first report of whole transcriptome analysis of coconut subjected to water deficit stress.