Ion dynamics of Supreme and Parish under normal and salt-treated conditions
Many studies have shown that seashore paspalum is among the most salinity-tolerant warm-season turfgrass species with a NaCl tolerance threshold of 474.0 mM [20]. To study the mechanisms underlying seashore paspalum’s high salt tolerance, two cultivars, Supreme and Parish were used for morphological, physiological and comparative transcriptomics studies (Figure 1A). Firstly, we compared their morphological differences in response to salt treatment. Supreme and Parish grown under the same conditions were exposed to 400 mM NaCl solution. After a 12-day treatment, chlorotic leaves were clearly observed in Parish while Supreme was not strongly affected, indicative of a more tolerant trait of Supreme than Parish (Figure 1B). Moreover, Supreme also has better recovery than Parish after salt treatment based on chlorosis in leaves (Figure 1C). To reveal possible physiological mechanisms of differential performance of Supreme and Parish under salt stress, we measured their leaf ion contents under normal and salt-stressed conditions. Supreme has significantly higher Na+ content than Parish under both conditions, whereas their K+ contents are similar, and remain the same even upon exposure to salinity (Figure 1D, 1E). In addition, Supreme has significantly higher Ca2+ content than Parish under normal conditions, but their Ca2+ contents are similar after treatment with salt (Figure 1F). The demonstration of higher salt tolerance of Supreme and its physiological characteristics implies the importance of the associated genetic underpinnings.
Transcriptome sequencing of Supreme and Parish under normal and salt-treated conditions
To characterize and compare the transcriptome response of Supreme and Parish under salt treatment, we treated plants with 400 mM NaCl for 1 hour. We use this condition because it was suggested that genes that rapidly changed expression upon salt stress should be important for salt tolerance [21]. Illumina sequencing of indexed and pooled RNA with polyA tails generated a total of 80.29 million and 78.88 million paired-end reads with a single read length about 101bp for Supreme and Parish, respectively. An overview of the sequencing and assembly results are represented in Supplemental table S1. Among these raw reads, 95.89% and 95.77% remained after trimming for Supreme and Parish, respectively, which were then de novo assembled into one reference transcriptome using Trinity. De novo assembly of mixed trimmed reads generated 342,165 Trinity transcripts (the individual assembled contig) with an average length of 784 bp and N50 value of 1,339 bp, and a total of 244,926 Trinity genes (the clustered Trinity transcripts based on shared sequence content) with average length of 580 bp and N50 value of 761 bp. GC content, which is an important indicator of the gene and genomic composition as well as DNA stability is 49.7% in seashore paspalum’s transcriptome, which is similar to the transcriptome GC composition of other monocot plants such as rice (51.1%) and Triticum aestivum (51.4%) [22, 23].
A total of 169,391 ORFs (49.5% of all Trinity transcripts) were identified among 342,165 Trinity transcript sequences using TransDecoder. Using CD-HIT software, the 169,391 ORFs were clustered into 82,608 unigenes. The length distribution of the unigenes is shown in Supplemental figure S1. Approximately 48.4% and 20.5% of the unigenes had a length >= 500bp and >= 1,000bp, respectively. To compare with the previously reported transcriptome with 32,603 reported Trinity genes assembled in another seashore paspalum cultivar “Adalady”, we conducted the Benchmarking Universal Single Copy Orthologs (BUSCO) analysis to check the quality and completeness of assembly. By searching 3,278 total BUSCO groups against our transcriptome, 3, 028 (92.3%) were “complete”, 174 (5.3%) were “fragmented”, and the remaining 76 (2.4%) were “missing”, indicating the high completeness of our assembled transcripts. As shown in Supplemental table S3, the transcriptome assembled in this study has a higher completeness and quality than the previously reported transcriptome, thus providing additional genomic resources that can be exploited for gene discovery and functional study [19].
Functional annotation of seashore paspalum’s transcriptome
Homology-based functional annotation of the seashore paspalum unigenes was then carried out. Distribution of the annotated unigenes in each database is shown in Supplemental table S2. 82,608 unigenes were blasted against the NCBI non-redundant (nr) protein database using Blastx. 65,540 (79.3%) out of the 82608 unigenes showed homology to the nr protein sequences. E-value distribution of blast results is shown in Supplemental figure S2. The best blastx hits against the nr database were then imported to Blast2GO software [24] for gene ontology (GO) classification and the result is shown in Supplemental figure S3. Among 82,608 unigenes, 36,387 unigenes (44%) were successfully annotated with 16 GO terms (level 2) and classified into three ontologies: biological process (BP, Supplemental figure S3A), cellular component (CC, Supplemental figure S3B), and molecular function (MF, Supplemental figure S3C). Within the BP category, genes involved in metabolic process (16946), cellular response (14342), single-organism process (8922) and biological regulation (3787) are highly represented. The CC category mainly comprises genes involved in membrane (10287), cell (10050), cell part (9904), membrane part (8528) and organelle (6716). Under MF, catalytic activity (15615) was the most abundant GO term, followed by binding (15411).
To compare the gene repertoire of seashore paspalum to other plant species, we aligned the unigenes against the nr protein database and performed the species distribution of the unigenes using Blast2GO software. As shown in Supplemental figure S4, the five top-hit species that best match the sequences of seashore paspalum unigenes are Setaria italica, Sorghum bicolor, Zea mays, Oryza sativa Japonica Group and Brachypodium distachyon, all of which belong to the Poaceae family.
Identification of transcription factors in seashore paspalum’s transcriptome
Transcription factors (TFs) play a vital role in regulating plant stress response as important regulatory elements. To identify potential TFs in the seashore paspalum’s transcriptome, 82,608 unigenes were searched against the PlantTFDB [25, 26] using Blastx. There are 3,250 transcripts that have at least one hit to the Arabidopsis and Oryza TFs, representing about 4% of the total unigenes and covering 68 putative TF families (Supplemental table S4). The TF gene families with ten or more unigenes identified in the seashore paspalum transcriptome are presented in Figure 2, among which the five most abundant categories are Myb (419), followed by WRKY (370), G2-like (268), bZIP (240), and bHLH (185).
Differentially expression analysis for Supreme and Parish under salt treatment
To compare gene expression levels in the control and salt-treated samples, the trimmed reads in each library were mapped to the 82,608 reference unigenes and the abundance of each unigene in different libraries was estimated using the RSEM software [27]. The expected count data produced by RSEM (Supplemental table S5) was used to identify DEGs with DEseq2 software [28]. To test reproducibility among two biological replicates, a Multi-Dimensional Scaling (MDS) plot (Figure 3) was generated for the control and salt-treated samples of Supreme and Parish. The fact that our biological replicates cluster so closely to each other on an ordination plot demonstrates their low inter-sample variability. Two comparisons were conducted: salt-treated Supreme versus untreated Supreme and salt-treated Parish versus untreated Parish. As shown in Figure 4A, a total of 828 unigenes were differentially expressed for salt-treated Supreme while 2,222 unigenes were differentially expressed for salt-treated Parish. 34 and 107 DEGs were identified to be potential transcription factors for Supreme and Parish, respectively (Figure 4B). Overlapping of two DEG lists generates 231 unigenes, out of which 12 unigenes are potential transcription factors (Figure 4A and 4B). The commonly regulated transcription factors in both cultivars under salt treatment are listed in Supplemental table S6.
Gene enrichment analysis of DEGs identified in Supreme and Parish under salt treatment
To inspect the biological relevance of DEGs, GO terms were assigned using Blast2GO. Five-hundred out of 828 DEGs (60.4%) were annotated for Supreme while 1,271 out of 2,222 DEGs (57.2%) were annotated for Parish (Figure 4A). GO enrichment analysis was then conducted to extract the over-represented GO terms that are significantly associated with the identified DEGs in Supreme and Parish under salt treatment, respectively. As shown in Figure 5A, genes that are up-regulated in salt-treated Supreme are involved in “oxidation-reduction process” and “nucleic acid binding” while genes that are down-regulated in salt-treated Supreme are involved in “regulation of transcription”, “transcription, DNA-templated”, “defense response” and “transcription factor activity”. GO functional enrichment analysis of DEGs in salt-treated Parish revealed that they are involved in much broader processes (Figure 5B). Many biological processes that are associated with salt response are induced in Parish, such as “oxidation-reduction process”, “cellular oxidant detoxification”, “response to oxidative stress”. Interestingly, “oxidation-reduction process” and “nucleic acid binding” are the most significantly enriched GO terms in the Biological Process (BP) category and Molecular Function (MF) category, respectively for up-regulated genes in both Supreme and Parish, implying their importance in salt tolerance in both cultivars. DEGs involved in “oxidation-reduction process” and “nucleic acid binding” are listed in Supplementary table S7 and S8, respectively.
Salt stress induced genes show higher expression in Supreme than in Parish under normal conditions
Although Supreme has fewer genes that are responsive to salt treatment than Parish, Supreme exhibits much higher tolerance than Parish. It is possible that Supreme may have a higher expression of salt stress induced genes than Parish under normal conditions that may or may not be induced upon salt treatment, and therefore may be more prepared when exposed to salinity. To test this hypothesis, we selected 202 genes based on the following criteria: 1) salt-induced genes in Parish; 2) higher expression in Supreme than in Parish under normal condition; 3) not changed or further induced in Supreme under salt treatment. To get insight into the biological meanings of these genes, we conducted GO enrichment analysis and found the following over-represented GO terms: “proline catabolic process”, “transcription factor activity”, “proline dehydrogenase activity” and “monooxygenase activity” (Figure 6). We then further examined genes with “transcription factor activity” (Table 1). It is interesting that many of these transcription factors have been associated with salt tolerance in the previous studies, such as dehydration-responsive element-binding (DREB) proteins, ethylene-responsive transcription factors (ERFs), and WRKY transcription factors [29].
Genes encoding for vacuolar Na+/H+ antiporters and proton pumps are differentially expressed between Supreme and Parish
As Supreme accumulated more Na+ and showed higher salt tolerance than Parish, we speculated that the former may have developed a strong capacity to sequestrate excessive Na+ into the vacuole through vacuolar Na+/H+ antiporters, thus maintaining high osmotic pressure to facilitate water uptake and protecting the cytoplasm from Na+ toxicity. To this end, we identified a total of seven candidate Na+/H+ antiporters (m.194123, m.133530, m.194121, m.194125, m.207121, m.28253, m.170234) in seashore paspalum’s transcriptome (Table 2). The differentially expressed Na+/H+ antiporter genes are highlighted in yellow background, one of which, m.194123 exhibits much higher expression in Supreme than in Parish under both normal and salt treated conditions. Interestingly, this gene is not induced by salt treatment in both Supreme and Parish. Among the remaining two differentially expressed candidate Na+/H+ antiporter genes, m.194121 has higher expression in Parish than in Supreme under salt treated conditions while m.170234 exhibits higher expression in Parish than in Supreme under normal conditions.
As vacuolar Na+/H+ antiporters are empowered by the electrochemical gradient created by H+-ATPases and H+-pyrophosphatases (H+-PPases) [30], we also identified eleven H+-ATPases and four H+-PPases in seashore paspalum’s transcriptome, which are shown in Table 3 and Table 4, respectively. None of the H+-ATPases showed differential expression (Table 3). Interestingly, all of the four vacuolar H+-PPases showed lower expression level in Supreme than in Parish under normal conditions, especially for one of the vacuolar H+-PPase m.112845 (Table 4). However, m.112845 was induced by about 1024 times (FC=210.28) in Supreme under salt treatment, suggesting a possible role in facilitating Na+ sequestration under high salinity and conferring salinity tolerance in Supreme (Table 4).