Cassava cultivar TMS60444 was asexually propagated in a plastic pot with Murashige and Skoog (MS) media (pH5.8, KOH) containing 4.4 g L⁻¹ MS salts (Murashige and Skoog, 1962), 20 g L⁻¹ sucrose, 2 μM CuSO₄, and 3 g L⁻¹ Gelrite (Wako) under a light intensity of 40–80 μmol photons m⁻²s⁻¹ with a 12 h/12 h photoperiod at 28 ± 1°C. The 5-cm length shoot tops of 3-month-old cassava plantlets were cut and then transferred to liquid MS media without agar. After 1 week incubation, the plantlets showing root elongation were subjected to each analysis as follows.

To reveal to which extent SAHA treatment alleviates cassava plants from salinity stress, survival rate and biomass (fresh and dry weight) were measured as follows. At 1 week-incubation after the transfer of cut plantlets to liquid medium, the plantlets were incubated on liquid MS media supplemented with 100 μM SAHA (suberoylanilide hydroxamic acid: N-hydroxy-N′-phenyloctanediamide: vorinostat, from Tokyo Chemical Industry, code-H1388) for 24 h. SAHA was dissolved in dimethylsulfoxide (DMSO) and DMSO treatment served as the control for all analyses. Then, the plantlets were transferred to 200 mM NaCl containing liquid MS media for 10 days and subsequently transferred to normal liquid MS medium for 1 month to observe phenotype and count survival rate in each condition. At least 13 cassava plantlets were used for each condition. The survival rate was based on the percentage of surviving plants. Plants with regenerated leaves and green stems were counted as surviving plants, while the plants with white stems having no new leaves were counted as dead ones. The fresh weight (FW) of root and shoot samples was measured immediately after harvesting. To measure the dry weight (DW), cassava samples were incubated at 60°C for 1 week. FW and DW were measured from the regenerated plants that had 3 or more leaves. Three biological replicates were conducted for all analyses.

The cassava samples were washed for 5 s three times using distilled water. Stems and leaves were cut separately using mineral-free carbon knives (Feather 5) and kept at −80°C until use. Analytical measurement of ²³Na and ³⁹K content was conducted using an inductively coupled plasma mass spectrometer (ICP-MS; NexION300, Perkin Elmer). Firstly, the cassava samples were dried using a freeze dryer for 3 days. The cassava sample weight was measured. Then 5 mL of conc. HNO₃ was added to the cassava samples. The cassava samples were incubated with concentrated HNO₃ overnight. After that, the cassava samples were digested using a microwave sample preparation system (MultiWave-3000, Perkin Elmer). All samples were completely digested prior to running on the ICP-MS. The digested samples were adjusted to a volume of 50 mL with Milli-Q water, then filtered through 5B filter paper (Advantec). The ICP-MS analysis was carried out as described previously (Itouga et al., 2014). Three independent biological replicates were generated for each condition. Na⁺ and K⁺ content were analyzed statistically with a t-test.

The leaf and root samples were ground using a Multi-beads shocker (Yasui Kikai) and 200 μL of 2 × Laemmli sample buffer [2% SDS, 50.4% (w/v) glycerol, 0.02 M Tris-HCl (pH 6.8) with 2% 2-mercaptoethenol] was immediately added to the powder. The mixture was heated at 95°C for 3 min then placed on ice for 5 min, and centrifuged at 4°C for 10 min at 14,000 × g. The protein supernatant was transferred to a new tube, and separated by 12.5% SDS-PAGE gel electrophoresis. After gel electrophoresis, the protein was blotted on equilibrated Immobilon-P PVDF membrane (Millipore). Then the blotted membrane was blocked using 5% skim milk at room temperature for 1 h. Next, the membrane was incubated at 4°C overnight for the primary antigen–antibody interaction. Secondary antibodies were incubated with the membrane for 1 h at room temperature. Finally, detection was carried out using an ImageQuant LAS 4000 (GE). Primary antibodies were H3 (Abcam, 1791), H4 (Abcam, 10158), acH3 (Merck Millipore, 06-599), and acH4 (Merck Millipore, 06-866), diluted 1:5000, 1:3000, 1:2000, and 1:5000, respectively. The secondary antibody was goat anti-rabbit IgG (GE, NA931), diluted 1:25,000.

Total RNA was extracted using Plant RNA Isolation Reagent (Invitrogen) following spin-column RNA precipitation using RNeasy Plant Mini Kit (Qiagen). Briefly, the sample tissues were ground into fine powder using a Multi-beads shocker. The tissue powder was used for RNA extraction using Plant RNA Isolation Reagent according to the manufacturer's instruction (Invitrogen). To clean the RNA solution, after adding 20 μL of RNase-free water to the RNA pellet, the RNA solution was purified using spin-column RNA precipitation followed by addition of RNase-Free DNase (Qiagen). In order to remove contaminated DNA completely, the process for RNA purification using spin-column RNA precipitation with DNase treatment was repeated twice. All samples were stored at −80°C until use.

For microarray analysis, the plantlets were treated with 100 μM SAHA for 24 h, followed by treatment with 200 mM NaCl for 2 or 24 h. The quality of total RNA was evaluated using a Bioanalyzer system (Agilent). Microarray analysis was performed according to the procedure of Utsumi et al. (2012). Briefly, total RNA was labeled with cyanin-3 (Cy3) using the Quick Amp Labeling kit (Agilent Technologies). Next, the Cy3-labeled cRNA was fragmented and hybridized to the Agilent microarray (GPL14139; Utsumi et al., 2016). After hybridization, the microarrays were washed and immediately scanned on an Agilent DNA Microarray Scanner (G2505B) using one color scan setting for 8 × 60 K array slides. Feature Extraction software (Ver. 9.1, Agilent Technologies) was used to process the digital images. Three independent experiments were performed for each condition. After normalization, microarray probes were detected based on one-way ANOVA with Benjamini–Hochberg correction (FDR) < 0.0001. An AGI gene code is shown in Tables 1–6, if the protein encoded in each gene (probe ID) has the sequence similarity with those in Arabidopsis at E ≤ 1 × 10⁻⁵. The E-value for each probe ID is available in Tables S1–S5. Gene ontology (GO) analyses were carried out using agriGO at http://bioinfo.cau.edu.cn/agriGO/. The information from the microarray data is available on the GEO website (GEO ID: GSE84715).

First-strand cDNA was synthesized from 500 ng total RNA with random primers. ReverTra Ace (TOYOBO) was used for the reverse transcription reaction according to the manufacturer's instructions. Transcript levels were assayed using Fast SYBR Green Master Mix (Applied Biosystems) and a StepOnePlus Real-Time PCR System (Applied Biosystems) according to the manufacturer's protocols. Gene-specific primers were designed using the PrimerQuest tool (http://sg.idtdna.com/primerquest/Home/Index). Melting curve analysis was conducted to validate the specificity of the PCR amplification. For qRT-PCR, three biological replicates were conducted. Actin was used as reference gene to normalize data (cassava4.1_009807m). The relevant primers are listed in Table S6. Three independent biological replicates were generated for each condition. Changes in gene expression were analyzed statistically with a t-test.

An evolutionary tree was constructed using the maximum likelihood method in MEGA7 (Kumar et al., 2016). The tree was evaluated with 1000 bootstrap replicates.

To confirm which NaCl concentration inhibits growth in cassava, plantlets were incubated in liquid medium containing different NaCl concentrations. A concentration of 200 mM NaCl clearly inhibited the growth of cassava plantlets (Figure 1A). After transfer to 200 mM NaCl salt medium for 1 week, the plantlets' older leaves started to turn yellow. At 2 weeks after transfer to salt medium, leaves showed chlorosis at both 150 and 200 mM NaCl. After 14-d incubation with 0, 150, and 200 mM NaCl, plants showed root growth of 7.4 ± 2.1, 3.6 ± 0.8, and 1 ± 0.2 cm, respectively (Figure 1B). The 200 mM NaCl concentration was selected for further experiments.

To analyze whether SAHA treatment enhances tolerance to salinity stress in cassava, we counted the survival rate of SAHA-treated plants (Figure 2A). SAHA-treated plants showed a 30.6% higher survival rate than non-treated plants under salinity stress conditions (Figure 2B). Fresh and dry weight measurement revealed that SAHA treatment increased biomass in roots (Figures 2C,D). These results suggest that SAHA treatment improves tolerance to high salinity stress in cassava.

The concentration of sodium (²³Na) and potassium (³⁹K) in stem and leaf samples was quantitatively analyzed by ICP-MS to reveal the extent of SAHA treatment's influence on ion homeostasis under salinity stress conditions. SAHA-treated plants had reduced Na⁺ concentrations in both stems (2.9 mg g⁻¹ DW) and leaves (0.87 mg g⁻¹ DW) compared with those in control plants (stems and leaves: 4.36 and 1.52 mg g⁻¹ DW, respectively; Figure 3A). However, there was no obvious difference in K⁺ concentrations in SAHA-treated plants and control plants, resulting in higher K⁺/Na⁺ ratios in SAHA-treated plants than in control plants (Figures 3B,C). These results suggest that SAHA treatment allows cassava to control ion homeostasis under high salinity stress conditions.

It has so far been unclear whether SAHA functions as the HDAC inhibitor in plants. We performed immunoblotting analysis to determine whether SAHA has an effect on histone acetylation levels. Cassava plantlets treated with 100 μM SAHA for 12 or 24 h, and cassava plantlets treated with 100 μM SAHA for 24 h then subjected to 200 mM NaCl for 6 h, were used for detecting histone acetylation status. Hyperacetylation of histones H3 and H4 was detected in the SAHA-treated plants compared with the control plants, especially in root samples (Figures 4A,C). Levels of histone acetylation were significantly increased at 12 and 24 h of SAHA treatment and remained higher up to 30 h after additional NaCl treatment. In contrast to roots, the acetylation level of histone H3 in leaf samples was slightly induced at 12 h and returned to a level similar to that of non-SAHA treated plants at 24 h (Figure 4B). There was no difference in acetylation levels of histone H4 in leaves between SAHA-treated plants and control plants (Figure 4D). These results suggest that SAHA mainly induces hyperacetylation of histones H3 and H4 in the roots, which may lead to transcriptional changes that enhance tolerance to high salinity stress in cassava.

To understand responses to salt stress and unveil the mechanism underlying increased tolerance to salinity stress under SAHA treatment in cassava, the transcriptional changes in roots and leaves were analyzed. For root samples, there were 4389 genes with significant expression changes [one-way ANOVA with Benjamini–Hochberg correction (FDR) < 0.0001] in at least one condition (Figure 5). In contrast, only 53 genes showed significant differences in mRNA expression in leaves (Table S7). These data are consistent with the results of immunoblotting analysis.

The heatmap shows significant differences among treatment conditions in root samples. According to their expression profiles, genes differentially expressed in roots can be divided into five classes (Figures 6A–E). GO analysis was used to gain an overview of the functions of genes in each class. Class 1 contains SAHA-unaffected genes with functions assigned to GO:0000003 (reproduction), GO:0009791 (post-embryonic development), and GO:0008152 (metabolic process; Figure 6A). Class 2 contains genes up-regulated by SAHA treatment. GO enrichment analysis revealed that the majority of the genes in this group belong to GO:0050896 (response to stimulus), GO:0044237 (cellular metabolic process), and GO:0044238 (primary metabolic process; Figure 6B). Classes 3 and 4 contain genes down-regulated by the effect of SAHA treatment at 2 h and 24 h NaCl treatment. Salt stress-unresponsive genes and transient salt stress-responsive genes were enriched in class 3 (Figure 6C) and class 4 (Figure 6D), respectively. GO enrichment analysis indicated that GO:0051179 (localization), GO:0006810 (transport), GO:0016043 (cellular component organization), and GO:0048856 (anatomical structure development) are included in class 3, whereas GO:0051179 (localization) and GO:0006810 (transport) are included in class 4. Class 5 contains the most genes (1103 genes). Class 5 is a group of salt-responsive genes that are up-regulated at 2 h under both NaCl and SAHA treatment. Several gene ontology (GO) categories were enriched among these genes, such as GO:0006950 (response to stress), GO:0009414 (response to water deprivation), GO:0009651 (response to salt stress), GO:0009737 (response to abscisic acid stimulus), and GO:0006082 (organic acid metabolic process; Figure 6E).

As GO ontology analysis revealed, class 5 contains high salinity stress-responsive genes involved in ABA biosynthesis and signal transduction, including homologs of Arabidopsis genes NCED3 (MeNCED3: RknMes02_025528: cassava4.1_026283m, Tables 1, 2), EDL3 (MeEDL3: RknMes02_033569: cassava4.1_030886m, Tables 1, 2), ABI1 (MeABI1: RknMes02_033475: cassava4.1_020355m, Tables S1, S2), ABI2 (MeABI2: RknMes02_010484: cassava4.1_010060m, Tables 1, 2), and other PP2Cs, such as AHG1 (MeAHG1: RknMes02_003939: cassava4.1_013309m, Tables 1, 2), AHG3 (MeAHG3: RknMes02_013552: cassava4.1_008067m, Tables S1, S2), HAI1 (MeHAI1: RknMes02_002573: cassava4.1_007998m, Table 1), HAI2 (MeHAI2: RknMes02_006068: cassava4.1_007913m, Tables 1, 2), HAB1 (MeHAB1: RknMes02_048414: cassava4.1_005959m, Tables S1, S2) (Ma et al., 2006). In Arabidopsis, the salt-induced upregulation of several transcription factors including ATAF1, ATHB12, NAP, AZF2, HSF2, RD26, and ATERF4, has been reported (Ma et al., 2006; Matsui et al., 2008). In cassava, salt stress induced the expression of their homologs MeATAF1 (RknMes02_016900: cassava4.1_013132m, Tables S1, S2), MeATHB12 (RknMes02_016407: cassava4.1_015049m, Table 2), MeNAP (RknMes02_036224: cassava4.1_013467m, Tables S1, S2), and MeRD26 (RknMes02_031261: cassava4.1_010999m, Tables 1, 2). LEA proteins, whose overexpression gives tolerance to salinity stress, are regulated by ABA in land plants and thought to function as molecular shields to prevent aggregation caused by dehydration (Shinde et al., 2012). MeLEA (RknMes02_006505: cassava4.1_025947m) showed the highest the highest up-regulated gene by salt stress (Table 1, Table S1). These results suggest that ABA synthesis and its signal transduction pathway coordinate the basis of the response to salinity stress in cassava. This is consistent with previous data in Arabidopsis (Matsui et al., 2008).

In addition to ABA, the hormones JA and methyl jasmonate (MeJA) are also involved in salt stress response. Consistent with previous reports that MeJA induces salt responsive genes in roots (Ma et al., 2006), our transcriptome analysis from root sample detected the induced mRNA expression of genes for JA/MeJA biosynthesis (MeACX2: RknMes02_027394: cassava4.1_002855m, Table S1; MeACX3: RknMes02_012214: cassava4.1_002966m, Table S1; and MeJMT: RknMes02_028787, cassava4.1_010155m, Table 1), suggesting that the JA signaling pathway is also involved in response to salt stress in cassava.

The expression levels of osmoprotectant biosynthesis-related genes, such as Δ-1-pyrroline-5-carboxylate synthetase (P5CS) and raffinose synthase were induced under salinity stress condition. P5CS1 (MeP5CS1: RknMes02_003952: cassava4.1_002374m) catalyzes the rate-limiting step in the biosynthesis of proline. MeP5CS1 was up-regulated at 2 and 24 h NaCl treatment (Tables S1, S2). Soluble sugars of the raffinose family have been associated with plant response to abiotic stresses. Raffinose synthase (RS) is responsible for raffinose biosynthesis. It has been reported that high salinity stress increased RS5 transcription (Egert et al., 2013; ElSayed et al., 2014). The expression of MeRS5 (RknMes02_026274: cassava4.1_002019m) was upregulated at 2 h NaCl treatment (Table S1). Proline and raffinose are likely to be synthesized and function as osmoprotectants under salinity stress in cassava.

We identified 421 genes whose expression was upregulated by 24 h SAHA treatment (Table S3). After pre-treatment of SAHA, 745 and 268 genes were upregulated by 2 and 24 h salt treatment, respectively (Tables S4, S5).

Salt Overly Sensitive1 (SOS1) is a gene responsible for increased salinity tolerance in Arabidopsis plants treated with the HDAC inhibitor Ky-2. The HDAC inhibitor strongly induced the expression of AtSOS1 to 5-, 3.5-, and 1.67-fold during NaCl treatment for 0, 2 and 10 h, respectively (Sako et al., 2016). Overexpression of SOS1 enhances tolerance to salinity stress in Arabidopsis (Shi et al., 2003; Yang et al., 2009) and tobacco (Yue et al., 2012). SAHA treatment enhanced MeSOS1 gene expression to 1.35-fold compared with untreated plants (Figure 7A). Although, the upregulation of MeSOS1 gene by SAHA treatment was not so significant (Figure 7B), the slight enhanced expression of MeSOS1 might contribute to the increased salinity stress tolerance by SAHA pretreatment in cassava.

Transcriptome analysis revealed that the mRNA expression of genes, involved in phytohormone [abscisic acid (ABA), jasmonic acid (JA), ethylene, and gibberellin] biosynthesis pathways, was up-regulated after NaCl treatment in SAHA-pretreated roots (Table S8). Among them, the expression of an allene oxide cyclase gene (MeAOC4: RknMes02_051874: cassava4.1_022180m) was strongly induced by SAHA treatment (Tables 3–5). SAHA treatment enhanced MeAOC4 expression to 32.8-fold compared with untreated plants (Figure 7A). Previous studies have revealed that overexpression of AOCs can confer salinity stress tolerance to several crops (Yamada et al., 2002; Pi et al., 2009; Zhao et al., 2014). According to the current cassava genome database, AOCs constitute a small gene family (MeAOC3-1: RknMes02_049533: cassava4.1_014582m; MeAOC3-2: RknMes02_054684: cassava4.1_026961m; and MeAOC4: RknMes02_051874: cassava4.1_022180m) in cassava (Figure S1). Our transcriptome analysis revealed that MeAOC3-2 is a high salinity stress-responsive gene and the expression of MeAOC4 is not upregulated by high-salinity stress (Figure S1). SAHA treatment increased the expression of MeAOC4 to 35-fold (Figure 7B), suggesting that activation of the JA pathway mediated by overexpression of MeAOC4 might play a pivotal role in alleviating high-salinity stress in cassava.

In the case of the HDAC inhibitor Ky-2, 72.9% of the HDAC inhibitor-inducible genes are salt-responsive genes in Arabidopsis (Sako et al., 2016). In contrast, 28.3% of SAHA-upregulated genes are high salinity stress-responsive (Figure S2A), and only 27 SAHA-upregulated genes were upregulated by high-salinity stress (2 and/or 24 h NaCl treatment; Table 6). In mangrove plants, lignin accumulation functions in blocking metal-ion influx with higher suberization, suggesting that lignification can prohibit ions from flowing inside (Cheng et al., 2014). We analyzed the following three genes that are believed to play critical roles in regulating lignin accumulation: L-phenylalanine ammonia lyase (PAL), cinnamyl alcohol dehydrogenase (CAD), and caffeic acid 3-Omethyltransferase (COMT) genes (Whetten and Sederoff, 1995; Ma and Xu, 2008; Nguyen et al., 2016). Among them, the expression of only COMT was significantly enhanced by SAHA treatment (Figure S3), suggesting that the lignin accumulation might be changed by overexpression of COMT induced by SAHA treatment and block ion uptake.

SAHA treatment decreased the expression of 141 high salinity stress-responsive genes including cell wall-related genes such as beta-galactosidases (RknMes02_003559, cassava4.1_001503m; RknMes02_030412, cassava4.1_001733m; RknMes02_034218, cassava4.1_001885m), beta-glucosidase (RknMes02_058028, cassava4.1_032518m), pectin lyase (RknMes02_051310, cassava4.1_021247m), pectinesterase (RknMes02_057990, cassava4.1_032455m), and pectin methylesterase inhibitor 4 (RknMes02_055023, cassava4.1_027551m) (Figure S2, Table S9). The decreased expression of these genes might contribute to changes in cell-wall components caused by SAHA treatment (see Discussion).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.