Control of glycerol biosynthesis under high salt stress in Arabidopsis

Additional keywords: abiotic stress, osmoprotection, T-DNA insertion.

A plant’s life depends on its ability to sense and positively respond to environmental stresses (Padamsee et al. 2012). Under saline condition, cells respond to osmolarity changes through multiple signalling pathways (Hohmann 2009). Eukaryotic cells are able to respond to environmental stresses by changing their protein repertoire (Melamed et al. 2008). Several adaptations are required in the cell in order to survive osmotic stress. These adaptations involve osmoregulation, ion homeostasis, accumulation of solutes, and possibly modifications of cell morphology (Gunde-Cimerman et al. 2009; Kralj Kunčič et al. 2010).

There is reported evidence for the existence of a specific signal transduction pathway that senses the osmotic situation of the cell and eventually induces physiological changes to adapt to the new growth condition (Brewster et al. 1993). A candidate element of this pathway is the product of the high osmolarity glycerol 1 (HOG1) gene, which is a mitogen-activated protein kinase or MAPK (Lawrence et al. 2004). This protein may be involved in signal transduction after the change in the osmotic environment has been sensed.

The most prominent physiological effect in osmotically-stressed cells is the enhanced production of intracellular compatible solutes to counterbalance the osmotic pressure. Glycerol seems to be an important compatible solute produced in yeast, and mechanisms exist to specifically retain glycerol in the cytosol under abiotic stress (Albertyn et al. 1994). Glycerol 3-phosphate (G3P) generated by the glycerol pathway can be transported between the cytosol and plastidial stroma. The plastidial ACT1 gene product catalyses the acylation of oleic acid (18 : 1) by the action of G3P acyltransferase (GPAT) on the generated G3P backbone (Kunst et al. 1988). This is a shift of G3P utilisation in the direction of lipid biosynthetic pathway. Plants impaired in utilisation of plastidial G3P (act1 mutants) are expected to accumulate elevated levels of G3P (Venugopal et al. 2009). The pathway for glycerol production starting from dihydroxyacetone phosphate (DHAP) is shown in Fig. S1, available as Supplementary Material to this paper.

In conjunction with the induction of glycerol 3-phosphate dehydrogenase (G3PDH), the activities of ADH and ALDH enzymes – involved in NADH : NAD⁺ recycling – also change, resulting in an enhanced production of NADH, and consequently production of G3P and glycerol (Blomberg and Adler 1989). G3PDH is a NAD⁺-dependent key enzyme of cellular glycerol biosynthesis essential for growth of cells under abiotic stresses in which ADH is repressed, and ALDH is induced, but actions of both enzymes result in enhanced production of NADH, and consequently G3P and glycerol. The pyridine nucleotides NAD⁺ and NADH are the primary redox carriers involved in cell metabolism. Analysis of mutants in Arabidopsis deficient in G3PDH (namely gly1) showed an elevated NADH : NAD⁺ ratio under standard growth conditions resulting in constitutively increased level of reactive oxygen species (ROS) (Shen et al. 2006). A yeast mutant lacking GPP2 gene activity (homologue of AtGPP2), which encodes glycerol 3-phosphatase (G3PP) for glycerol production in yeast, is devoid of G3PP activity, and produces only a small amount of glycerol. However, overproduction of G3PP in yeast did not significantly enhance glycerol production, indicating that the enzyme is not rate-limiting for glycerol production (Påhlman et al. 2001). Expression of GPP2 gene was shown to be induced under hyperosmotic stress and this induction partially depends on the high osmolarity glycerol (HOG) pathway.

There are at least two avenues for glycerol metabolism. The first is initiated upon conversion (phosphorylation) of glycerol to glycerol 3-phosphate (G3P) via the action of glycerol kinase (GK) (Chanda et al. 2008). The second is initiated upon conversion (oxidation) of glycerol to dihydroxyacetone (DHA) (Mandel et al. 2011). The gene GLI (analogue of GUT gene in yeast) in Arabidopsis is known for its effect in glycerol utilisation and a knockout mutant of this gene (namely gli) is glycerol-insensitive (Eastmond 2004). The latter was reported for the overproduction of glycerol and consequently resistance to abiotic stresses. Glycerol dehydrogenase (GLYDH) is known for its effect on catalysing the oxidation of glycerol to DHA with concomitant reduction of NAD⁺ to NADH (Ruzheinikov et al. 2001; Rawls et al. 2011). This confirms that the combined action of the enzymes ADH, ALDH and GLYDH is recycling of NADH : NAD⁺ to allow continuous operation of glycerol production and metabolism.

In previous work (Edris et al. 2012), we investigated the possible production and subsequent role of glycerol in osmostress response by expressing the yeast (Saccharomyces cerevisiae) GPD1 gene (homologue of AtGLY1) in Escherichia coli and analysing the response of the modified E. coli to osmostress. The results led us to examine expression of this gene in Arabidopsis thaliana (L. Heynh.) to determine its effect on salt tolerance and to study the genetic control of the glycerol pathway. G3P was shown lately to be a regulator of plant defence signalling (Chanda et al. 2008; Venugopal et al. 2009), but its role in plant salt tolerance was not detected. Here we report observations of a reduction in glycerol production when salt concentration reached 200 mM. As the activity of the enzymes in the corresponding pathway remained essentially unchanged, we speculate that a salt-dependent change in the permeability of membranes for glycerol may occur at this salt level. This result is in accordance with the finding that the facilitators (e.g. MIPs) that transport glycerol across the membrane (Luyten et al. 1995) change as a function of osmotic stress.

The coding nucleotide sequence of the yeast (Saccharomyces cerevisiae) glycerol 3-phosphate dehydrogenase (GPD1) cDNA was isolated using the GeneRacer kit (Invitrogen Life Tech, Grand Island, NY, USA) and cloned into pCR4TOPO vector (Invitrogen Life Tech). The complete cDNA sequence (861 bp) contained 105 bp of a 5′ untranslated region, 645 bp of an open reading frame (ORF), and 111 bp of a 3′ untranslated region. An 881 bp EcoRI fragment containing the full-length cDNA was filled using Klenow fragment and blunt-end ligated to the previously digested and filled BamHI site of pAHC17 (Christensen and Quail 1996). A 3.16-kb GPD1 gene cassette was recovered via partial digestion with HinDIII/EcoRI, filled using Klenow fragment and blunt-end ligated to the previously digested and filled NcoI/BglII sites of the binary vector pCAMBIA1302 (~10.55 kb, CAMBIA, Brisbane, Qld, Australia) to obtain pGPD1 (13.71 kb, Fig. S2). The plasmid was mobilised into Agrobacterium tumefaciens strain GV3101 and used for genetic transformation of Arabidopsis thaliana (L. Heynh.) wild type A. thaliana (accession Columbia) and the three T-DNA insertion mutants (Schubert et al. 2004), namely gly1, gli1 and act1 with codes of At2 g41540 (NACS ID: N520437), At1 g80460 (NACS ID: N567205) and At1 g32200 (NACS ID: N627295), respectively, were obtained from the Nottingham Arabidopsis Stock Centre (NASC, the European Arabidopsis Stock Center, University of Nottingham, Loughborough, UK). A single-copy T₂ GPD1-transgenic plant was selected for the salt stress experiments. This selection was based on the plant’s ability to set seeds, its overall morphological similarity to wild-type, transgene copy number and level of transgene expression. The WT plant was used as a negative control. The gly1, gli1 and act1 mutants were also used to examine elements of the glycerol pathway and possible participation of G3P and/or glycerol in conferring salt tolerance in Arabidopsis.

WT A. thaliana (accession Columbia) plants (5 weeks old) of were exposed to an A. tumefaciens strain harbouring the pGPD1 by using the vacuum infiltration method (Bechtold et al. 1993) and grown in the greenhouse. Seeds were collected and transformants were screened in MS medium (Murashige and Skoog 1962) supplemented with 20 μg mL^–1 hygromycin. Transgenic plants were transferred to soil with mixture of vermiculite : peat moss : perlite with equal volumes and left to grow until seed set in the greenhouse under 16 h light/8 h dark cycle at 21°C. Three GPD1 homozygous lines at the second generation (T₂) whose overall morphology looked like WT plants were selected for subsequent analyses based on the ability to set viable seeds.

Plant genomic DNA (30 ug) was extracted from leaves of WT plant and the three selected T₂ GPD1-transgenics as described by Dellaporta et al. (1983). Genomic Southern hybridisation was conducted using the non-radioactive digoxigenin (DIG) High Prime DNA Labelling and Detection Starter kit (Roche Applied Science, Penzberg, Germany). Hybridisation was conducted to demonstrate the integration and copy number of the transgene and to confirm independence of the different transgenic events. Genomic DNA was digested with SalI to release a 1.41 kb fragment involving a portion of the P-ubi promoter, the ubi intron and the 5′ untranslated region of GPD1 gene and probed with a PstI/EcoRI fragment (1.38 kb) of the P-ubi promoter and portion of the ubi intron of pGPD1. A genomic Southern blot was hybridised and processed according to manufacturer’s instructions. The three T₂ GPD1 homozygous lines were selected for real-time PCR to detect the event with the highest expression level of GPD1 transgene.

qRT-PCR was conducted for the WT and the three selected T₂ transgenic plants using the Agilent Mx3000P QPCR Systems (Agilent Technologies, Santa Clara, CA, USA). All cDNA-synthesised samples were diluted (1 : 10) before amplification. The reaction (25 μL) components were 12.5 μL Maxima SYBR Green/ROX qPCR master mix, 0.2 μM of each gene forward (5′ TGTCTCGGTGGTGTCCCTAT 3′) and reverse (5′ ACCATGGCTGATGGAAGACT 3′) primers, and PCR-grade water was added up to 22.5 μL. Primers were designed using the Universal ProbeLibrary Assay Design Center (Roche, www.roche-applied-science.com). Finally, 2.5 μL of diluted cDNA template was added to the reaction mix. Amplification for each sample was conducted in three replicates along with a no-template control (NTC, PCR-grade water). The thermal cycling conditions consisted of 1 cycle of pre-incubation at 95°C for 5 min (hot-start), denaturation at 95°C for 10 s, annealing at 59°C for 10 s and extension at 72°C for 10 s followed by 40 cycles of denaturation at 95°C for 15 s, annealing at 59°C for 30 s and extension at 72°C for 2 min. Data were collected and amplification plots of ΔR_n versus cycle number were generated for analysis. Amplicons were further analysed by agarose gel electrophoresis. Actin gene was used as a reference to show that equal amounts of RNA were used in the analysis. To avoid false positives due to DNA contamination, PCR reaction was carried out for all RNA samples and results were negative (data not shown).

Real-time PCR data were analysed and amplification plots were generated using the software program MxPro QPCR (Agilent Mx3000P QPCR Systems, Santa Clara, CA, USA). The standard curves were generated using both the software programs followed by Microsoft Excel (Microsoft, Redmond, WA, USA). Calculations of ΔΔC_T for the relative quantification of target DNA were done as follows: (ΔΔC_T =(C_T, target gene – C_T, Actin) x – (C_T, target gene – C_T, Actin) y), where x is the treated sample and y is the control sample. Validation and normalisation to cr-Actin relative to the copy number of the target gene in control were done following work by Livak and Schmittgen (2001). Two experiments were carried out for each sample in which individual samples were run in triplicates and the average value was accepted when the s.d. was lower than 0.38 (Pfaffl 2001).

Experiments were conducted to observe the performances of different genotypes (WT, GPD1-transgenic plant, and gly1, gli1 and Act1 mutants) at 0, 100, 150 or 200 mM NaCl. We determined levels of glycerol, as well as G3P at different NaCl concentrations and activities of four enzymes involved directly in the glycerol pathway (namely G3PDH, G3PP, GK and G3PDH) and two for NAD+/NADH recycling (ADH and ALDH).

For germination studies, 50 seeds/replicate/treatment/genotype were surface sterilised, sown in Petri dishes containing either MS medium (Murashige and Skoog 1962) or MS medium with different concentrations of NaCl (100, 150 or 200 mM). Experiments were conducted in four replicates and analysed using a randomised complete block design. The plates were kept in the dark at 4°C for 2 days and then shifted to 21 ± 2°C (day/night) under a 16-h-light/8-h-dark cycle. Germination was recorded up to day 6. Emergence of the radicle from seed coat was considered as completion of germination.

Two-week-old plants of different genotypes, grown at 0 mM NaCl under a 16-h-light/8-h-dark cycle at 21 ± 2°C (day/night), were transferred to MS agar plates with different concentrations of NaCl (0, 100, 150 or 200 mM). Light intensity was ~175 μmol m^–2 s^–1 for the 16-h photoperiod. Plants were allowed to grow on these media for 2 more weeks. Then, individual plant root length, number of leaves per plant and the rosette area were scored. The experiment was conducted using a randomised complete block design with three replicates in which 50 plants were utilised for each replicate, genotype and treatment.

For analysis of glycerol and G3P levels and enzyme activities at different salt concentrations, 5 g of 4-week-old Arabidopsis leaves of genotypes grown at different salt concentrations for 2 weeks were homogenised in liquid nitrogen to a fine powder or collected as leaf extracts. Protein content was determined as described by Bradford (1976) using bovine serum albumin as a standard.

Cellular glycerol and G3P levels were determined enzymatically as described by Wei et al. (2004). Glycerol and G3P concentrations were given in μmol per gram leaf tissue (FW) of different genotypes grown at different salt concentrations (0, 100, 150 or 200 mM NaCl). Assays of G3PDH, G3PP and GLYDH enzyme activities were done according to work by Chen et al. (2012). Glycerol kinase assay was performed using the method described by Liu et al. (1994), ADH assay was performed according to work by Benz et al. (2007), and ALDH assay was performed according to Li et al. (2004). Statistical analyses of different experiments were performed following the procedure outlined by Gomez and Gomez (1984).

A total of 20 putatively transgenic T_o plants were generated from transformation experiments, as indicated by hygromycin selection and PCR analyses for the presence of GPD1 gene (data not shown). Of the 20 plant lines, 10 with morphological similarities to WT were selected for further characterisation. Among the T₁ transgenic lines, only three (lines 1, 5 and 6) segregated for the presence of GPD1 gene in a ratio of 3 : 1 indicating the insertion of a single copy of the GPD1 gene cassette. This result was supported by genomic Southern hybridisation of T₂ transgenics, which also demonstrated independence of the three transgenic events (Fig. S3).

In the pGPD1 plasmid (Fig. S2), the restriction enzyme SalI has several cleavage sites: one of them is located within the ubi1 promoter and the second is located at the end of the 5′ untranslated region of GPD1 gene. This SalI cleavage results in the liberation of a 1.41 kb fragment in the three transgenic lines (Fig. S3). One other hybridising high-molecular-weight fragment of different size was observed in each of the three lines suggesting unique transgene insertions for each event and confirming the insertion of a single copy of the transgene. The results for the positive control (plasmid DNA) indicate the presence of the 1.41 kb fragment as well as a fragment for the rest of the plasmid (~11.65 kb), except for the 645 bp of the GPD1 ORF that was cleaved with SalI at several sites and consequently, not hybridised to the probe (Fig. S3).

T₂ families of the three GPD1-transgenic events were further characterised for overexpression of the transgene as detected by real-time PCR (Fig. S4). The results indicated that the transgene was most highly expressed in line 6 followed by lines 5 and 1. Line 6 was chosen as the GPD1-transgenic event for further analyses, along with the three T-DNA insertion mutants gly1, gli1 and act1 representing genes involved in the glycerol pathway. These genotypes were utilised in evaluating salt tolerance in conjunction with the overproduction of glycerol and G3P as well as for determining the activities of six enzymes in the glycerol pathway.

We first examined the effects of NaCl, at different concentrations, on germination of seeds of T₂ GPD1-transgenic family of line 6 and the three mutants versus the WT as shown in Figs 1 and 2. We observed that seed germination initiated at day 4 under the control condition (0 mM NaCl), whereas it began at day 3 in all of the salt stress treatments (Fig. 1). Thus, NaCl treatments stimulated germination of seeds compared with the controls. Essentially, no further changes in germination efficiency across genotypes and salt conditions were observed after day 5. Although germination rates of the lines varied somewhat, no statistically significant differences were observed at day 6 across genotypes under different conditions except at 150 mM NaCl (Figs 1 and 2). In that condition, germination percentages for the GPD1-transgenic line and the two mutants gli1 and act1 were significantly higher than those of the WT and gly1 mutant. NaCl at the highest concentration (200 mM) severely affected seed germination percentages and rates for all genotypes and was thus excluded. Maintenance of the plants for an additional week at 150 mM NaCl, revealed the transgenic line and the two gli1 and act1 mutants continued to grow better and appear more vigorous than the WT and gly1 mutant (Fig. S5).

Fig. 1.  Effects of salt stress on germination percentages – up to day 6 – of T₂ GPD1-transgenic Arabidopsis as well as WT and mutant (gly1, gli1 and act1) seeds germinated in MS medium supplemented with different concentrations of NaCl (0 mM (r² = 0.80), 100 mM (r² = 0.73), 150 mM (r² = 0.71) or 200 mM (r² = 0.58)).

Fig. 2.  Effect of salt stress on germination percentages at day 6 of T₂ GPD1-transgenic Arabidopsis as well as WT and mutant (gly1, gli1 and act1) seeds germinated in MS medium supplemented with different concentrations of NaCl (0, 100, 150 or 200 mM). Concurrently, seeds of different genotypes were germinated at 0 mM NaCl, left for 2 weeks, transferred to MS agar plates with different concentrations of NaCl (0, 100, 150 or 200 mM) and allowed to grow for 2 more weeks. Then, growth parameters (root length (mm), no. leaves per plant and rosette area (cm²)) were scored.

After 2 weeks of growth at different salt stress levels, the mean root length (mm), number of leaves per plant and rosette areas (cm²) were measured for the different genotypes (Fig. 2). The results for these three parameters indicated better performance of transgenic line GPD1 and the two mutants gli1 and act1 at 100 and 150 mM NaCl as compared with the WT and the mutant gly1. The latter showed absence of greening under salt stress and started dying soon after being transferred to the soil, whereas the other three genotypes were still green at 100 and 150 mM NaCl and were able to continue growth up to maturity (data not shown). However, none of the genotypes, including those expected to overproduce glycerol, were able to continue growth up to maturity at 200 mM NaCl. It is worth noting that the morphological and visual growth characteristics of transgenic and mutant plants were similar to the WT plants under normal growth conditions.

The different lines utilised in salt stress experiment, were next subjected to biochemical analysis. The abundance of glycerol in the genotypes at different NaCl concentrations correlated with the growth performances under the same conditions (Fig. 3). Glycerol levels across treatments were lowest for the gly1 mutant followed by WT, and highest for mutant gli1 followed by GPD1-transgenic and act1 mutant. For the last two genotypes, glycerol levels significantly increased as NaCl concentration increased until 150 mM, and subsequently dropped at 200 mM.

Fig. 3.  Glycerol and G3P levels as well as activities of six enzymes in the glycerol pathway of T₂ GPD1-transgenic Arabidopsis as well as WT and mutant (gly1, gli1 and act1) plants grown at different concentrations of NaCl (0, 100, 150 or 200 mM).

The act1 mutant, followed by the transgenic plants, accumulated higher G3P levels at 100 mM NaCl as compared with the other genotypes. However, no statistically significant difference in G3P levels were scored across genotypes and treatments, indicating that G3P level might be insensitive to salt stress. The overall results of glycerol measurements indicate its association with tolerance to salt stress in Arabidopsis. Eastmond (2004) indicated that overproduction of glycerol in gli mutants resulted in higher resistance to hyperosmotic and salt stresses in Arabidopsis. Mandel et al. (2011) reported that G3P contributed to resistance to biotic stress and to play a role in inducing systemic acquired resistance (SAR), however, our data do not suggest a possible role for G3P during salt stress.

Despite a knockout mutation in the gly1 gene, this mutant still produces a low level of glycerol across treatments indicating the possible presence of another G3PDH-encoding gene(s) or the occurrence of other avenue(s) for glycerol production in Arabidopsis. Another possibility is the occurrence of incomplete loss-of-function in this particular mutant. This may also indicate the presence of a closely related gene, analogous to the yeast GPD2 gene (Albertyn et al. 1994; Hubmann et al. 2011), in Arabidopsis for glycerol production. The act1 mutant overproduced G3P, but glycerol levels were not higher than in the WT. This result indicates that the G3P pool might not be directly accessible to glycerol production due to compartmentalisation. Glycerol levels decreased for gli1, act1 and GPD1 lines when salt concentration increased from 150 to 200 mM. As the levels of the precursor G3P did not change significantly this suggests that mechanism(s) of glycerol retention in the cell might be impaired at 200 mM NaCl in Arabidopsis. During osmotic stress, Bohnert et al. (1999) indicated that salt-tolerant Zygosaccharomyces rouxii accumulated glycerol by increasing its retention within the cell. They suggested the occurrence of active uptake of glycerol, rather than increased production of glycerol, during this stress condition. Further, they reported that in Saccharomyces cerevisiae, glycerol biosynthesis increases during osmotic stress, although cells do not alter membrane permeability. They hypothesise that holding high glycerol concentration in the cell is a difficult task as it requires high energy cost that limits further increases in salt tolerance. The latter hypothesis seems to match ours in which mechanism(s) of glycerol retention in the cell of Arabidopsis might be impaired at high NaCl concentration.

Activities of enzymes of the glycerol pathway during salt stress were analysed in order to detect the performance of enzymes in this pathway at different NaCl concentrations (Fig. 3). The activities of ADH and ALDH in the Gly1 mutant indicate that both enzymes are insensitive to salt concentrations whereas ADH activity decreased and ALDH activity increased with increasing salt concentration in the other lines. This result suggests that G3PDH might have a role in decreasing activity of ADH and increasing activity of ALDH. The latter response results in enhanced production of NADH and, consequently, of glycerol (Blomberg and Adler 1989). Shen et al. (2006) indicated that the specific activity of the NADH-dependent G3PDH is induced by osmotic stress up to 8-fold. The results of the present study indicate no substantial increase in enzyme activity of the GPD1 transgenic across different growth conditions.

Chen et al. (2012) conducted a comparative analysis in Dunaliella salina, a halotolerant green algae, on four key enzymes of the glycerol cycle metabolic pathway, e.g. G3PDH, G3PP, GLYDH (or DHAR) and dihydroxyacetone kinase (DHAK) under different salt stress conditions. Their results showed that NaCl could stimulate the activities of G3PDH and G3PP and consequently glycerol production in agreement with the results of the present finding up to 150 mM NaCl. Chen et al. (2012) also emphasised the role of G3PDH in the regulation of glycerol metabolism by directing its flow under different salt stresses. They found that the G3PDH activity in D. salina reached highest at continuous 3.5 M NaCl exposure, but decreased at higher concentrations. Upon glycerol excessive accumulation due to increased salinities, G3PDH activity tends opposite to the direction of glycerol synthesis. In other words, it is the net rate of glycerol synthesis rather than the net accumulation that is responding to the external osmoticum. These results confirm our finding on the trilateral relationships comprising the activity of this enzyme, glycerol production and salt tolerance.

The results of G3PP and GK enzyme activities (both are functioning in opposite directions in the glycerol pathway) indicated that G3P (substrate) is a limiting factor for the direction of the pathway (Fig. 3). GK activity in gly1 mutant (G3P-deficient) was high across treatments, while low in gli1 mutant (for G3P overproduction). The results of G3PP activities for the same two mutants were controversial. The results of GLYDH activities for different genotypes during salt stress treatments were arbitrary - with no specific pattern. The results of the present study indicate the possible combined action of the enzymes ADH, ALDH and GLYDH in recycling NADH : NAD⁺ to allow continuous operation of glycerol production and metabolism. However, scoring NADH : NAD⁺ levels during salt stress is required to support this hypothesis. The results also indicate that GLYDH is insensitive to salt stress, to the activities of the five upstream enzymes in the pathway or to the level of glycerol or G3P in different genotypes. Furthermore, the results of enzyme activities during salt stress do not justify the drop of glycerol level at 200 mM NaCl. The results also indicate that this drop is not due to its conversion to DHA via the action of GLYDH or to G3P via the action of GK. The activities of the four enzymes, directly related to glycerol pathway, at 150 and 200 mM NaCl were not different.

The permeability of plant membranes is reported to be mainly controlled by the regulation of major intrinsic proteins (MIPs) accumulation (Gerbeau et al. 1999; Lee et al. 2009). MIP gating (opening and closing of the pore) is reported to respond to a wide range of stimuli such as pH, phosphorylation, programmed cell death, hydrogen peroxide and abiotic stresses (Cabello-Hurtado and Ramos 2004; Horie et al. 2011). However, Klepek et al. (2005) indicated that other types of transporters (e.g. Arabidopsis POLYOL TRANSPORTER5) can mediate the transport of several substrates including glycerol. It is also possible that the drop in glycerol level under severe salt stress is due to glycerol leakage across membranes that were damaged by the exposure to high NaCl concentration. The hypothesis for the possible impairment of glycerol retention or gain of glycerol efflux mechanism(s) in the cell of Arabidopsis under severe salt stress requires further attention.