Transcriptome sequencing, de novo assembly and functional annotation
Transcriptome sequencing generated more than 119 and 118 million raw reads from the roots and shoots of P. cornutum, respectively (Supplementary Table S1). After exclusion of the low-quality reads, including empty reads, adapter reads and reads with unknown nucleotide “N” or only one copy number, 117 and 116 million clean reads with GC percentage of 45% and 46% were obtained from the roots and shoots, respectively (Supplementary Table S1). The total clean reads/total raw reads percentages were more than 98% for both roots and shoots. Only a small number of reads was excluded by filtering, indicating that the data were highly reliable.
The de novo assembly of total clean reads generated 64,978 unigenes from shoots and 80,307 unigenes from roots, with an average mean length of 1091 and 997 bp, respectively (Table 1). After further assembly and redundancy elimination of these unigenes using the CAP Assembler, 72,068 gap-free unigenes with an average length of 1243 bp were obtained (Table 1). The length distribution of all unigenes showed that the length of 34,108 unigenes (more than 50%) was greater than 1000 bp (Supplementary Fig. S1).
In total, 63,396 unigenes were matched to known homologs from other plant species by being aligned against the protein databases including NR, NT, Swiss-Prot, KEGG, COG, GO, accounting for 88% of the total unigenes (Supplementary Table S2). Specifically, 57,695, 60,628 39,893, 35,136, 24,608 and 53,435 unigenes had functional annotations in the above six databases, which accounted for 80, 84, 55, 49, 34 and 74% of the total unigenes, respectively (Supplementary Table S2).
Identification of DEGs in the roots and shootsof P. cornutum treated with 50 mM NaCl
After treatment with 50 mM NaCl for 6 h, 7356 and 7088 differentially expressed genes (DEGs) were identified in P. cornutum roots and shoots, respectively, among which 4020 and 3673 were upregulated, and 3336 and 3415 were downregulated, respectively (Fig. 1). After treatment with 50 mM NaCl for 24 h, 12,648 and 4905 DEGs were identified in the P. cornutum roots and shoots, among which 4064 and 1972 were upregulated and 8584 and 2933 were downregulated, respectively (Fig. 1).
DEGs related to ion transport
The absorption and accumulation of inorganic osmoticums such as Cl-, Na+, SO42- and PO43-, as well as maintaining homeostasis of the inorganic macronutrients K+ and NO3-, are primary physiological mechanisms in the salt tolerance of P. cornutum [19, 20]. Therefore, we firstly analyzed the expression changes in DEGs related to ion transport in the roots and shoots of P. cornutum treated with 50 mM NaCl for 6 and 24 h.
After treatment with 50 mM NaCl for 6 and 24 h, 49 and 41 DEGs associated with Cl-, NO3-, Na+, K+, Ca2+, SO42-, PO43-, Zn2+, Cu2+, NH4+, Mn2+, Mg2+, Fe2+, BO33- and H+ transport were respectively upregulated in roots, and 18 and 31 DEGs were respectively downregulated in roots (Fig. 2A and B). Among the upregulated DEGs, a large proportion was Cl- and/or NO3- transporter-encoding genes, including SLAH1, CLCs and NPFs. It was noteworthy that the transcript levels of SLAH1, CLCg, NPF6.4 were upregulated in roots after treatment with 50 mM NaCl for both 6 and 24 h (Table 2, Supplementary Table S3 and S4), indicating that these genes might be closely involved in Cl- absorption and accumulation, as well as NO3- homeostasis of P. cornutum under saline conditions. Furthermore, it was observed that SLAH1 was almost completely unexpressed in roots under the control condition (the RPKM value of SLAH1 was 0.01 under the control condition for both 6 and 24 h), while its expression level was increased by over 8-fold after salt treatment for both 6 and 24 h (Supplementary Table S3 and S4), indicative of a vital role of SLAH1 in salt response of P. cornutum. Among the downregulated DEGs related to Cl- and/or NO3- transport, the transcript abundance of CCC1, which encodes a Cl- transporter facilitating the retrieval from xylem sap into roots [23], was substantially decreased in roots under salt treatment for 24 h (Supplementary Table S4). Na+/H+ antiporters (NHXs) are widely recognized to play important roles in Na+ and/or K+ homeostasis in plants under saline conditions [24–26]. Our results showed that the transcript abundance of plasma membrane-located Na+/H+ antiporters-encoding gene SOS1 (also named NHX7) in roots was upregulated after salt treatment for 6 h but downregulated after salt treatment for 24 h (Fig. 2A and B, Supplementary Table S3 and S4), indicating that SOS1 in roots might positively regulate salt tolerance in P. cornutum during relatively short-term salt treatment. The transcript abundance of the vacuole-located Na+/H+ antiporters-encoding gene NHX1 was upregulated in roots after 24 h of salt treatment, and the golgi-located Na+/H+ antiporters-encoding genes NHX5 and NHX6 were upregulated after both 6 and 24 h of salt treatment (Fig. 2A and B, Supplementary Table S3 and S4); therefore, these three genes very likely confer Na+/K+ homeostasis in roots of P. cornutum under saline conditions. Furthermore, several K+ transporter-encoding genes that also contribute to Na+/K+ homeostasis, such as KT2, KT3, KUP2, HAK3, KEA1, KEA4, KEA5 and KEA6, were upregulated under salt treatment for either 6 or 24 h (Fig. 2A and B, Supplementary Table S3 and S4). CNGCs are important plasma membrane-located cation channels mainly permeable to K+, Na+ and Ca2+; GLRs as glutamate (Glu) receptors can provide Glu to activate plasma membrane-located nonspecific cation channels in root cells and, therefore, trigger large and rapid rises in the concentrations of cytosolic cations, especially Ca2+ [27–29]. Our results also characterized many upregulated CNGC and GLR genes in roots of P. cornutum under salt treatment (Fig. 2A and B), indicating that P. cornutum is also likely to maintain root cationic homeostasis under saline conditions by increasing the expression of these CNGC and GLR genes in roots to control K+, Na+ and Ca2+ transport across the cell membrane. Many upregulated DEGs related to the transport of other inorganic ions, including SO42- (e.g., Sultr1.2, Sultr2.1; Sultr3.1),, PO43- (e.g., PhT1.3, PhT2.1, PhT4.3, PHO1),, Zn2+ (e.g., ZnT1, ZnT8, ZnT12),, Cu2+ (e.g., CTR2, CTR5, CCH, PAA1),, Mn2+ (e.g., PDR2),, Fe2+ (e.g., IRT1),, Mg2+ (e.g., MgT),, NH4+ (e.g., AMT1.1, AMT1.2, AMT2),, and BO33- (e.g., BOR3, BOR4),, were also identified in roots of P. cornutum under either 6 or 24 h salt treatment (Fig. 2A and B, Supplementary Table S3 and S4). Additionally, as the transport of inorganic ions across cell membranes generally couples H+ as a proton motive force [30], the upregulated P-H+ ATPase and V-H+ ATPase in roots under salt treatment identified by our data (Fig. 2A and B) likely play essential roles in ion transport across the root cell plasma membrane and tonoplast, in addition to maintaining the cellular membrane electrochemical gradient.
After salt treatment for 6 and 24 h, 46 and 30 DEGs in shoots associated with inorganic ion transport were respectively upregulated and 23 and 26 DEGs in shoots were respectively downregulated (Fig. 2C and D). In contrast to the observation in roots, the expression of CCC1 in shoots was considerably upregulated under both 6 and 24 h of salt treatment (Fig. 2C and D, Supplementary Table S5 and S6), which might be beneficial for the exclusion of more Cl- from shoot vascular tissues to facilitate Cl- uptake by mesophyll cells. Cl- can only be largely accumulated in plant shoots by compartmentalization into vacuoles of mesophyll cells [31, 32]. This process in higher plants is dominated by the tonoplast-located chloride channel CLCg [33]. Our results showed that the transcript abundance of CLCg in shoots was upregulated under salt treatment for both 6 and 24 h (Table 3). Hence, CLCg should play a key role in the Cl--accumulating characteristic of P. cornutum under saline conditions. In addition to CLCg, CLCb, which encodes tonoplast NO3- transporter [34], and several NPFs such as NPF6.3 and NPF6.4 that encode NO3- and/or Cl- transporters [35], were also upregulated under salt treatment (Supplementary Table S5 and S6), suggesting that these genes might also be essential for Cl- accumulation and NO3- homeostasis in shoots of P. cornutum under saline conditions. Certain DEGs found in roots encoding Na+/H+ antiporters such as SOS1 and NHX1; K+ transporters such as KT2, KUP2, KEA1, KEA4, KEA6; cation channels such as CNGC1 and CNGC6; and glutamate receptors such as GLR3.3, were also upregulated in shoots after salt treatment for either 6 or 24 h (Supplementary Table S5 and S6), indicating that these genes might also regulate Na+/K+ homeostasis in shoots of P. cornutum under saline conditions. Furthermore, several DEGs related to Na+ or K+ transport, including HKT1, SKOR and GORK, were only found in shoots but not in roots after salt treatment (Fig. 2). HKT1-type proteins facilitate Na+ absorption and transport in the roots of model plants A. thaliana [36, 37]. But, in P. cornutum, HKT1 is likely to be distinctively responsive to salt stress in shoots. As an outward rectifying K+ channel, GORK has been proven to participate in stomatal movement [38]; therefore, the observed upregulation of GORK in shoots after salt treatment (Fig. 2D) are likely essential for stomatal movement and, consequently, affect salt tolerance in P. cornutum. Many DEGs related to the transport of other inorganic nutrients or osmoticums, such as SO42-, Zn2+, Cu2+, Mn2+, Fe2+, Mg2+, and NH4+, were identified in shoots after salt treatment for either 6 or 24 h (Fig. 2C and D), indicating that these genes might also regulate shoot osmotic adjustment of P. cornutum under saline conditions. Among the DEGs encoding V-H+ ATPase, P-H+ ATPase and P-Ca2+ ATPase in shoots, V-H+ ATPase c2, V-H+ ATPase e1 and V-H+ ATPase h were not expressed under the control condition but were highly expressed under salt treatment (Supplementary Table S5 and S6), suggesting that these three genes might play vital roles in vacuole compartmentation of inorganic osmoticums for facilitating shoot osmotic adjustment.
DEGs related to ROS-scavenging system
The ROS-scavenging system in higher plants mainly consists of the ascorbate-glutathione (AsA-GSH) cycle, glutathione peroxidase (GPX) pathway, catalase (CAT) pathway and peroxiredoxin/thioredoxin (PrxR/Trx) pathway [9]. After salt treatment for 6 and 24 h, a total of 19 and 15 DEGs categorized in the abovementioned ROS-scavenging pathways in roots of P. cornutum were respectively upregulated and 5 and 29 DEGs were respectively downregulated (Fig. 3A and B). Among the upregulated DEGs in roots after 6 h of salt treatment, most were GSTs (involved in both the ASA-GSH and GPX pathways) and Trxs (involved in the PrxR/Trx pathway) (Fig. 3A). Hence, the ASA-GSH, GPX and PrxR/Trx pathways might be the major components of the root ROS-scavenging system of P. cornutum in adaption to salinity. After 24 h of salt treatment, although many DEGs associated with ROS scavenging were downregulated, the expression of GST-U10 (no expression under the control condition) and TrxH8 was still obviously upregulated (Fig. 3B, Supplementary Table S8), indicating that these two genes might play vital roles in ROS scavenging in roots of P. cornutum under saline conditions. Furthermore, after salt treatment for 24 h, 2 upregulated PODs, 1 upregulated CAT and 1 upregulated SOD emerged (Fig. 3B), indicating that these genes might mainly facilitate root ROS-scavenging of P. cornutum during relatively long-term salt treatment.
Following 6 h of 50 mM NaCl treatment, the expression levels of 3 GST genes and 2 GPX genes in the GPX pathway were upregulated in the P. cornutum shoots (Fig. 3C). In particular, GST-F9 showed no expression under the control condition, with high expression detected in the shoots after 6 h of 50 mM NaCl treatment (Supplementary Table S9). Despite the finding that fewer POD genes were upregulated than downregulated, one POD (POD43) was not expressed under the control condition, whereas its expression was substantially induced in the shoots after 6 h of 50 mM NaCl treatment (Fig. 3C, Supplementary Table S9). Almost all the differentially expressed PEX, Trx and PrxR genes in shoots after 6 h of salt treatment were upregulated (including 1 PEX and 2 Trxs, which were not expressed under the control condition but were highly expressed under salt treatment) (Fig. 3C, Supplementary Table S9), indicating that the CAT and PrxR/Trx pathways should mainly function in the shoot ROS-scavenging system of P. cornutum under salt treatment. Following 24 h of salt treatment, only a few genes associated with ROS scavenging were differentially expressed in shoots, and most of them were downregulated (Fig. 3D). However, PEX3–1a showed no expression under the control condition but upregulated expression under both 6 and 24 h of salt treatment (Supplementary Table S9 and S10), suggesting that PEX3–1a might play a vital role in ROS scavenging in shoots of P. cornutum under saline conditions.
DEGs related to photosynthesis
The maintenance of high carbon assimilation efficiency is an important strategy by which P. cornutum adapts to soil salinity [19]. The oxygenic photosynthesis of higher plants consists of photosynthetic electron transport with many components such as chlorophyll, photosystem II-light harvesting complex (PS II), photosystem I-light harvesting complex (PS I), cytochrome b6f complex, ferredoxin, ATP synthase, and carbon fixation using various enzymes [39]. Thus, we analyzed the expression of DEGs related to the abovementioned processes in P. cornutum treated with 50 mM NaCl.
Under 50 mM NaCl treatment for 6 h, the majority of DEGs in shoots of P. cornutum related to the components of PS II complex and chlorophyll biosynthesis were upregulated and all of the DEGs related to the components of PS I complex, cytochrome b6f complex and ATP synthase were upregulated (Fig. 4A), suggesting that the expression of these genes in shoots might be essential for light energy absorption and photosynthetic electron transport to generate more ATP for carbon fixation. Furthermore, it was observed that nearly 20 DEGs encoding various enzymes (mainly malate dehydrogenase and phosphoglycerate kinase) in carbon fixation process were upregulated (Fig. 4A, Supplementary Table S11), indicative of a vital role of these genes in maintaining high carbon assimilation efficiency in P. cornutum under salt stress.
Under 50 mM NaCl treatment for 24 h, the numbers of upregulated DEGs in shoots related to the components of PS II complex, PS I complex, chlorophyll biosynthesis and ATP synthase were much lower than that under 50 mM NaCl treatment for 6 h, but the expression of one cytochrome f-encoding gene, one ferredoxin-dependent glutamate synthase-encoding gene and one uroporphyrinogen III methyltransferase-encoding gene was still upregulated (Fig. 4B, Supplementary Table S12), suggesting that these three genes might play a crucial role in light energy absorption and photosynthetic electron transport in P. cornutum under saline conditions. Furthermore, the majority of the upregulated DEGs involved in carbon fixation after 6 h of salt treatment were also upregulated after 24 h of salt treatment (Fig. 4, Supplementary Table S11 and S12).
DEGs related to transcription factors
Along with functional genes, regulatory genes also participate in the responses of plants to environmental stresses by regulating signal transduction or functional gene expression [40]. Transcription factors (TFs) are important and abundant regulatory genes in higher plants. The major classes of TFs include NAC (NAM/ATAF/CUC), AP2/ERF (APETALA2 and ethylene-responsive element binding proteins), bHLH (basic helix-loop-helix), MYB (myeloblastosis), WRKY (WRKY-domain), bZIP/HD-ZIP (basic region-leucine zipper/homeodomain-leucine zipper), ZF (zinc finger) and HSP (heat shock protein), some of which have been validated to confer salt and drought tolerance in various plant species by transcriptional regulation of downstream target stress-responsive genes [9]. As TFs generally show rapid responses to abiotic stresses, we only analyzed the differentially expressed TF genes in roots and shoots after 6 h of salt treatment (the number of the differentially expressed TF genes after salt treatment for 24 h was much lower than that after salt treatment for 6 h in our results; data not shown).
After 6 h of 50 mM NaCl treatment, more than 100 TF genes were differentially expressed in roots; approximately two-thirds were upregulated, among which the majority were WRKYs, MYBs, ZFs, or bZIPs/HD-ZIPs, including WRKY33, WRKY54, and MYB3, CCCH-type ZFs and HD-ZIP6 (Fig. 5A, Supplementary Table S13), which have been validated to play important roles in salt tolerance and drought resistance in other plants [41–45]. Thus, the expression of these TF genes in roots should also be essential for salt tolerance of P. cornutum. Although less research have shown that MADS-box genes are closely related to the responses of plants to salinity, all 5 differentially expressed MADS-box genes in roots after salt treatment were considerably upregulated, with 3 members (AGL16, AGL27 and AGL29) showing no expression under the control condition but high expression under salt treatment (Fig. 5A, Supplementary Table S13). Up to 6 upregulated AP2/ERF-encoding genes were also found in roots. Only a few NAC, bHLH and HSP-encoding genes were differentially expressed after salt treatment (Fig. 5A).
After 6 h of 50 mM NaCl treatment, nearly 120 TF genes were differentially expressed in shoots (Fig. 5B). The number of upregulated WRKY and AP2/ERF-encoding genes in shoots was much lower than that in roots (Fig. 5), suggesting that WRKYs and AP2/ERFs might mainly regulate stress-responsible functional genes in roots to confer salt tolerance in P. cornutum. Similarly, upregulated MADS-box-encoding genes in shoots were less abundant than in roots after salt treatment, but the expression of MADS-box gene AGL30 was upregulated more than 5-fold in both shoots and roots (Fig. 5, Supplementary Table S13 and S14), suggesting that this TF is distinctively implicated in salt tolerance in P. cornutum. Many upregulated MYBs, including six members that showed no expression under the control condition, and ZFs, including seven members that showed no expression under the control condition, were also observed in shoots after salt treatment (Fig. 5B, Supplementary Table S14). Notably, upregulated HSP-encoding genes were only found in shoots but not in roots (Fig. 5), suggesting that HSPs might mainly regulate stress-responsible functional genes in shoots of P. cornutum under saline conditions.
Real-time PCR validation of differential gene expression
A real-time PCR validation was performed using 40 DEGs (20 randomly selected from roots and other 20 randomly selected from shoots). As shown in Supplementary Fig. S2, the fold changes of the selected DEGs measured by qRT-PCR were highly consistent with the results obtained from RNA-seq data. The correlation coefficient R2 of the DEGs from roots between qRT-PCR and RNA-seq results under salt treatment for 6 and 24 h was 0.93 and 0.92 respectively (Supplementary Fig. S2A and B). The correlation coefficient R2 of the DEGs from shoots between qRT-PCR and RNA-seq results under salt treatment for 6 and 24 h was 0.86 and 0.91 respectively (Supplementary Fig. S2C and D). Hence, the RNA-Seq results in the present study were highly reliable.