Construction of an informative transcriptome dataset for R. trigyna
For many non-model species, there is no background genomic information available for researchers to conduct comprehensive investigations into the genetic mechanisms underlying their unique features. Therefore, the newly developed NGS technique has been used widely to explore genomic solutions to important physiological questions. Gene annotation is an important element of NGS in which biological information is attached to the predicted genes or unigenes. A high proportion of unigenes with high-confidence BLASTX similarity to protein sequences from annotated gene catalogs of other plant species is considered to indicate the integrity of transcript sequences assembled from Illumina short-read data . This assumption was also verified by the results generated in the present study. In R. trigyna, 54.27% of 65340 unigenes were annotated by BLAST analysis and functional bioinformatics analyses (e.g., GO, Swiss-Prot, and KEGG). Overall, the top five species with BLAST hits to annotated unigenes were Arabidopsis thaliana, Oryza sativa, Arabidopsis lyrata, Populus trichocarpa, and Vitis vinifera, species for which the annotations of their genomes are comprehensive and largely accepted. This suggested that the sequences of the R. trigyna unigenes generated in the present study were assembled and annotated correctly.
The lengths of the assembled sequences are crucial in determining the level of significance of a BLAST match . Out of all the assembled unigenes, 66.79% had lengths between 300 and 500 bp, among which 37.1% had significant BLAST hits in the public databases. There were 1224 unigenes showing strong homology with the sequences hit in the database (E-value < 1.0E-50). Among 21697 unigenes with sequence lengths > 500 bp, 87.0% had significant BLAST scores and 59.2% showed E-values less than 1.0E-50. Out of all the unannotated unigenes (28991), 70.65% were longer than 500 bp, demonstrating that the lack of annotation for these unigenes was not because of a shorter sequence length but because of a genuine lack of hits to sequences in the database. Therefore, we can speculate that these unannotated unigenes represent a specific genetic resource of R. trigyna, which warrants further investigation.
In this study, we also used other strategies to enhance the effectiveness of the short reads assembly, apart from the use of bioinformatics methods. These included preparation of high-quality RNA samples, removal of dirty raw reads, BLAST assembly of unigenes using multiple databases, and the large sample population for sequencing [20, 24]. First, RNA was isolated from sterile R. trigyna seedlings to minimize the risk of contamination by foreign RNase. Second, three seedlings were used to extract RNA samples, not only to reduce sample bias, but also to ensure comprehensive coverage of the R. trigyna transcriptome. Last, two sequencing libraries (C21 and T43) were merged to generate longer sequences and to increase the sequencing depth. These strategies were an effective way to improve the quality of assembly and annotation of assembled unigenes .
In summary, an extensive and diverse expressed gene catalog, representing a large proportion of the transcribed genes in R. trigyna, was successfully sampled in the present study. The gene catalog provided a comprehensive understanding of the gene transcription profiles of R. trigyna, and laid a solid foundation for further study of salt-tolerance mechanisms and identification of novel genes in this species.
Transcriptome comparison identified genes related to salt-stress in R. trigyna
Gene transcription and/or expression is often compared among different developmental stages, among different plant organs, or among plants under different growth conditions [20–22]. In this study, many genes showing transcriptional changes under salt stress were identified by comparing NaCl-treated seedlings with untreated controls. There were 64694 unigenes showing differences in transcript abundance, and 5032 were defined as DEGs using the thresholds of false discovery rate (FDR) ≤ 0.001 and |log2Ratio| ≥ 1. GO clustering analysis suggested the potential biological functions of these DEGs. For example, many DEGs were enriched in GO terms such as “oxidoreductase activity”, “catalytic activity” and “response to stress”. This information will be useful to elucidate salt-tolerance mechanisms and to find new salt-stress-related genes specific to R. trigyna. KEGG enrichment analysis identified significantly enriched metabolic pathways or signal transduction pathways involving DEGs. Surprisingly, genes in the phenylpropanoid biosynthesis and flavonoid biosynthesis pathways were transcribed at much high levels under salt stress than in the control in R. trigyna, which was not expected from the results of previous studies. This result suggested that genes in these two pathways may play vital roles in resisting oxidation and maintaining membrane integrity.
An outstanding advantage of NGS is the detection of transcripts present in low copy numbers. Among the 5032 identified DEGs, almost half were detected as low-abundance transcripts that were up-regulated after exposure to salt-stress. For example, only 14 reads from C21 could be mapped to unigene24841, encoding flavanone 3-hydroxylase (F3H), compared with 470 reads from T43. Nearly 30-fold up-regulation suggested that this gene may act as a key element in the salt-stress response. Overall, transcriptome comparison analysis provided sufficient information to study the salt-responsive mechanisms and genes related to salt-tolerance related in R. trigyna.
Ion transport genes are important for salt tolerance in R. trigyna
Ion transport is a crucial element in response to salt stress in plants. This is particularly true in halophytes . It has been shown that halophytes are able to grow under extreme salinity conditions because of their anatomical and morphological adaptations and/or their avoidance mechanisms [25, 26]. R. trigyna is a representative recretohalophyte with unique morphological characteristics, such as salt glands and succulent leaves, allowing it to adapt to the salinized conditions of the Ordos desert. Our previous study showed that the salt glands of this species functionally excrete salt ions under normal and salt-stressed conditions. However, the amount of Na+ and Cl- excreted significantly increased under salt treatment. In addition, R. trigyna shows a strong ability to uptake K+ from barren soil. Therefore, it is possible that the ion transport mechanisms and their gene expression patterns in this plant differ from those of other plant species to some extent.
Once Na+ is taken up into the cell, ATPases (PM-H+-ATPases, V-H+-ATPases, and V-H+-PPases) are induced to create the driving force for Na+ transport. This results in not only extrusion of Na+ into the external environment by Na+/H+ antiporters in the plasma membrane, but also in compartmentalization into vacuoles by tonoplast Na+/H+ antiporters, which are essential for reestablishing cellular ion homeostasis in salt-stressed plants [25, 27–30]. In this study, five V-H+-PPases were detected in R. trigyna, and were proposed to generate a proton electrochemical gradient. Taking the morphological characteristics of this plant into account, the largest share of the driving force generated by such enzymes is likely to be consumed by the succulent leaves and the salt glands, which include two mesophyll-like collecting cells with obvious vacuoles, and six secretory cells . All mesophyll cells in leaves need V-H+-ATPases and V-H+-PPases to generate an electrochemical gradient to compartmentalize excess Na+ ions into the tonoplast, as a mechanism to cope with physiological drought conditions and ion toxicity [3, 25]. When Na+ is excreted via secretory cells, both V-H+-ATPases and V-H+-PPases are required to produce an H+ electrochemical gradient to pump Na+ into collecting cells. PM-H+-ATPases generate sufficient driving force to exchange Na+ ions out of the cells. In addition, higher H+-pumping activity may be indispensable to maintain the high concentration of K+ in salt-stressed R. trigyna. The activation of PM-H+-ATPases in salinized Populus euphratica led to repolarization or hyperpolarization of the plasma membrane and thus decreased NaCl-induced K+ loss through depolarization-activated outward rectifying K+ channels (DA-KOCs) [32, 33]. In R. trigyna, 10 out of 17 RtKOCs showed high homology with those of Populus, demonstrating similar functions of these genes. The analysis of transcripts related to K+ uptake showed that the outstanding potassium uptake capacity of R. trigyna was probably conferred by enhanced KUP/HAK (potassium transporter) and AKT K+ uptake systems, among which the number of up-regulated unigenes (39) far exceeded that of down-regulated unigenes (16) under salt stress. In summary, proton electrochemical gradient-dependent K+ maintenance, ions secretion, compartmentalization, and enhanced K+ uptake systems may represent important components in reestablishing ion homeostasis in R. trigyna.
On the other hand, we detected 12 NHXs that were probably responsible for Na+ sequestration. The significantly up-regulated unigene20634 and 5272 showed high homology to the Citrus reticulata NHX1 and a Salsola komarovii Na+/H+ antiporter, respectively. The slightly down-regulated unigene8445 and 752 had a highly significant BLAST hit to the Na+/H+ antiporter of M. crystallinum and the NHX1 of Tetragonia tetragonioides, respectively. These four unigenes also showed strong homology to the NHX2 of A. thaliana, suggesting that these AtNHX2-like proteins play an important role in avoiding or mitigating the deleterious effects of high Na+ levels in the cytosol or in regulating intravacuolar K+ and pH, which has been demonstrated in Arabidopsis. Our prediction is also supported to some degree by the studies of succulent leaves of the halophytes Suaeda salsa and Halostachys caspica.
In R. trigyna, we identified only one moderately expressed and up-regulated PM Na+/H+ antiporter (SOS1B, unigene798), named RtSOS1. The transcript profile of this gene was consistent with those of ThSOS1 and PeSOS1, which remained relatively constant or were slightly up-regulated under salt-tress [36, 37]. Escherichia coli complementation experiments with PeNhaD1 rescued salt-exposed bacteria by lowering salt accumulation , suggesting that the P. euphratica gene PeNhaD1 also functions as an Na+⁄H+ antiporter. Interestingly, a highly abundant transcript (unigene16859, RtNha1) encoding an Nha protein was identified in our dataset. This may play a role in Na+ exclusion in R. trigyna. At present, it is still unknown how the moderately up-regulated RtSOS1 achieves effective efflux of excess salt ions under extreme salt-stress, and whether any other gene products are also involved in this process. This requires further investigations in the near future.
ROS scavenging plays a key role in salt-stress response in R. trigyna
Salt stress causes a rapid increase in ROS, including superoxide radicals (·O2-), hydrogen peroxide (H2O2), and hydroxyl radicals (·OH), which perturb cellular redox homeostasis and result in oxidative damage to many cellular components and structures [39, 40]. According to our previous studies, ·O2- increased rapidly in R. trigyna seedlings treated with 400 mM NaCl, as did the concentrations of other antioxidants, for example, GSH. At the same time, increased activities of antioxidant enzymes (SOD, POD) were also detected. This suggested that R. trigyna probably has a similar ROS scavenging system to that of other plant species, and that this system can be enhanced to increase its antioxidative ability. As the first line of defense against oxidative damage, SODs are usually induced by salinity to rapidly dismutate ·O2- into oxygen and H2O2, which is subsequently removed through various pathways . In this study, 14 SODs were identified. Two of them (unigene23472, unigene8340) were transcribed at high levels, and showed up-regulated transcription under salt stress. Therefore, they were probably associated with enhanced SOD activity and responsible for converting the increased ·O2- into H2O2. On the other hand, increased GSH could be consumed in the ascorbate-GSH cycle and the GPX cycle . In the ascorbate-GSH cycle, the oxidized ascorbate (i.e., monodehydroascorbate) can be converted into dehydroascorbate (DHA); DHA is then reduced to ascorbate at the expense of GSH, yielding oxidized GSH (i.e., glutathione disulfide). In the GPX pathway, GPX can reduce H2O2 to the corresponding hydroxyl compounds using GSH. However, to fulfill its function as an antioxidant, GSH must be catalyzed by GSTs, which have GPX activity and can use GSH to reduce organic hydroperoxides of fatty acids and nucleic acids to the corresponding monohydroxy alcohols . Interestingly, in our dataset, there were 35 salt-induced GSTs genes that likely participated in such catalytic processes, especially the 11 Tau family GSTs . This finding was similar to the results of other studies on salt-stressed plants [44, 45]. Therefore, the interaction between GSH and GST may be an important factor in the ROS scavenging system of R. trigyna. In addition, we identified 84 genes encoding components of the PrxR/Trx defense system, which employs a thiol-based catalytic mechanism to reduce H2O2 and is regenerated using Trxs as electron donors .