Rice glycosyltransferase gene UGT2 functions in salt stress tolerance under the regulation of bZIP23 transcription factor

Rice glycosyltransferase gene UGT2 was identified to play a crucial role in salt tolerance. The transcription factor OsbZIP23 was demonstrated to regulate the UGT2 expression under stress conditions. UDP-glycosyltransferases (UGTs) play key roles in modulating plant responses to environmental challenges. In this study, we characterized a novel glycosyltransferase, UGT2, which plays an important role in salt stress responses in rice (Oryza sativa L). We found that seedlings overexpressing UGT2 exhibited better growth than wild type in shoot and root under hydroponic culture with salt stress treatments, while ugt2ko mutant lines suffered much more growth inhibition. When the soil-grown UGT2 transgenic plants were subjected to salt stress, we also found that ugt2ko mutant lines were severely withered and most of them died, while the overexpression lines grew well and had higher survival rate. Compared with wild-type plants, UGT2 overexpression greatly increased the expression levels of the reactive oxygen species scavenging genes and stress-responsive genes. Furthermore, the upstream regulatory mechanism of the UGT2 gene was identified and we found that a bZIP transcription factor, OsbZIP23, can bind to the UGT2 promoter and enhance the UGT2 transcription levels. This work reveals that OsbZIP23-UGT2 module may play a major role in regulating the salt stress tolerance in rice.


Introduction
As the global environment worsens, more and more farmlands are suffering from stressful stimuli such as salinity, drought, and other abiotic stresses. As a major and waterdependent crop in the world, rice growing and yield is greatly affected by these stresses (Boyer 1982). To cope with unfavorable environments, plants have evolved multiple mechanisms (Boyer 1982;Xiong and Zhu 2001;Agarwal et al. 2011). Hormones, especially abscisic acid (ABA), play an important role in response to abiotic stress and in mediating stress signals to regulate target gene expression (Sah et al. 2016). In addition, it is well known that many transcription factors play a crucial role in regulating ABAdependent and stress-related target genes.
Among these transcription factors, members of the basic leucine zipper (bZIP) family are paid more and more attention for their roles in stress responses. The bZIP transcription factor family is a large and well-studied family with multiple functions, such as light signaling, pathogen defense, flower development, seed maturation, and abiotic stress responses (Cutler et al. 2010;Hubbard et al. 2010;Peleg and Blumwald 2011). For example, rice OsbZIP58 is a key transcriptional regulator controlling starch synthesis in rice endosperm, and osbzip58 null mutants displayed abnormal seed morphology and reduced total starch and amylose content (Wang et al. 2013). The bZIP29 is involved in plant root growth in Arabidopsis (Leene et al. 2016). OsTGA10, as a bZIP transcription factor, was reported to be required for tapetum development (Chen et al. 2018). bZIP transcription factors also play important roles in coping with abiotic stresses. For example, the expression of OsbZIP71 was strongly induced by drought, salt, and ABA treatments. OsbZIP71 overexpression lines show better tolerance to drought and Communicated by Roger Thilmony. salt. In contrast, RNAi transgenic lines were more sensitive to salt, PEG osmotic stress and ABA treatment (Liu et al. 2014). Overexpression of OsbZIP62 enhanced the drought tolerance and oxidative stress tolerance of transgenic rice, while osbzip62 mutants exhibited sensitive phenotype (Yang et al. 2019). The bZIP transcription factor TabZIP15 could improves salt stress tolerance in wheat (Bi et al. 2021).
Four members of bZIP family (TRAB1, OsABI5, OsbZIP23, and OsbZIP46) got extensive studies for their involvement in ABA signal pathway. These studies reveal that most of them are involved in ABA or stress signaling. TRAB1 has been found to be activated through ABAdependent phosphorylation and to interact with the VP1 transcription factor (Hobo et al. 1999). OsABI5 is involved in ABA signal transduction and regulates stress responses (Yang et al. 2011). Overexpression of OsbZIP23 led to sensitivity to ABA and enhanced stress tolerance (Xiang et al. 2008). Thus, OsbZIP23 is believed to function as a transcriptional activator to positively regulate ABA responses and hence increase the stress resistance in rice.
In recent years, UDP-glycosyltransferases (UGTs) involved in abiotic stresses are drawing more and more attention and are extensively studied. UGTs have been reported to play a crucial role in metabolism modulation and stress adaptation. For instance, Arabidopsis UGT79B2/ B3 can improve the ROS-scavenging activity and abiotic stress tolerance by modulating anthocyanin metabolism (Li et al. 2017). UGT71C5 can promote the seed germination and increase the sensitivity to drought stress by dynamic change of ABA . The maize ufgt2 mutant showed hypersensitivity to salt and drought stresses via reducing quercetin and kaempferol glycosylation (Li et al. 2018). Recently, Dong et al. (2020) revealed that a rice UDP-glycosyltransferase GSA1 is required for the redirection of metabolic flux from lignin biosynthesis to flavonoid biosynthesis under abiotic stress and for the protection of rice against abiotic stress . Overexpression of UGT74E2, an Arabidopsis IBA glycosyltransferase, enhances seed germination and modulates stress tolerance via ABA signaling in rice .
In this study, an ABA and salt responsive rice glycosyltransferase gene UGT2 (Os11g0444000, nominated as UGT710E2 by international UDP Glucosyltransferase Nomenclature Committee) was characterized towards its function and regulatory mechanism in stress tolerance. Through mutant lines and overexpression lines, we demonstrated that the rice glycosyltransferase gene UGT2 is involved in the salt stress response and positively regulates salt tolerance. In addition, by analyzing the upstream regulatory elements of UGT2, we found that OsbZIP23 can directly bind to the UGT2 promoter and positively regulate UGT2 expression. Our data in this study reveal that OsbZIP23-UGT2 module plays an important role in regulating rice adaptation to salt stress.

Plant materials and growth conditions
Rice (Oryza sativa ssp. japonica) "Zhonghua 11" and Arabidopsis thaliana (Ecotype Columbia-0) were used as wildtype plants in this study. Rice seedlings were grown at 28 °C, 60% relative humidity, and a 14-h light/10-h dark cycle. Arabidopsis seedlings were grown at 22 °C, 60% relative humidity, and a 16-h light/8-h dark cycle.

Preparation for overexpression and knock-out lines of UGT2
The knock-out vector of UGT2 was constructed by CRISPR/ Cas9 strategy (Hangzhou Biogle Vector Construction Kit, http:// www. biogle. cn/). The selected target sequence is CCG CGT GGT CCT CTT CCC GATGC. To obtain UGT2 overexpression vector, the coding sequence of UGT2 gene (Os11g0444000) was amplified from cDNA of rice seedlings using TransStart® FastPfu DNA Polymerase (TransGen), and then cloned into pUN1301 vector driven by the maize Ubiquitin promoter. The Cas9 mutants and overexpression lines of rice were obtained by introducing knock-out vector or overexpression vector into rice through Agrobacteriummediated transgenic technology. The coding sequence of UGT2 gene was also cloned into pBI121 binary vector to generate CaMV 35S promoter:UGT2 construct. The constructed vector was introduced into Arabidopsis thaliana to obtain transgenic Arabidopsis by floral dipping method (Clough 2005). Primers used for plasmid construction are shown in Table S1.

Stress assays of the transgenic plants
The rice seeds of the transgenic line and the wild-type line were germinated in the water for 4 days, then transplanted to the water solution containing NaCl with indicated concentrations. The shoot length and root length were measured after 7 days.
For salt treatment in soil, rice plants were grown for 2 weeks at 28 °C, followed by treating with 200 mM NaCl for 2 weeks, and their survival rates were determined. Plant phenotypes were observed and photographed with a digital camera.
Arabidopsis seeds were sterilized and sown on MS medium for 3 days, and the seedlings with 1 cm primary roots were transferred to MS medium supplemented with 1 3 different concentrations of NaCl. The root length was measured after 2 weeks of stress treatment.
For salt treatment in soil, 2-week-old Arabidopsis plants were exposed to 200 mM NaCl solution for 15 days, and the phenotypes of these plants were photographed.
As for the stress treatments for gene expression analyses, 7-day-old rice plants were subjected to 200 mM NaCl treatments for 12 h and then the plants tissues were collected for RNA extraction from these samples. Plants grown in water were used as control.

Determination the content of proline and soluble sugar
For determination of the content of proline, the 4-week normal growth rice plants were treated with 200 mM NaCl for 24 h, or grown under normal conditions as control. The content of proline was measured as the method described by Bates and colleagues (Bates et al. 1973). The content of soluble sugar was measured as the method described by Zhang and Huang (Zhang and Huang 2013).

Detection of H 2 O 2 and superoxide by nitroblue tetrazolium (NBT) staining
Four-week-old plants were treated with 200 mM NaCl for 24 h. The plant leaves of each line were soaked in NBT solution for staining about 12 h. The method of NBT staining was described by Wang and colleagues (Wang et al. 2011).

RNA isolation and quantitative real-time PCR (qRT-PCR)
RNA was extracted from rice using TRIzol reagent (TaKaRa, Japan), and then total RNA was reverse transcribed using PrimeScript Reverse Transcription Kit (TaKaRa, Japan) and qRT-PCR was performed with TB Green PCR Master Mix kit (TaKaRa, Japan) on Bio-Rad Thermocycling System (CFX Connect, Hercules, CA). OsActin was used as an internal reference gene. Fold changes in gene expression level were calculated using the 2 −ΔΔCt method. The primers used are listed in Table S1.

Yeast one-hybrid assay
For the yeast one-hybrid assay, the yeast one-hybrid kit was used in accordance with the manufacturer's protocol (www. clont ech. com). The coding sequences of OsbZIP71, OsABI5, OsABF2, and OsbZIP23 were inserted into pGADT7, respectively (TaKaRa, Japan), while the promoter sequence of UGT2 was inserted into the cloning site of pAbAi. All constructs were transformed into yeast strain Y1HGold. Yeast cells were grown in SD/-Ura/-Leu medium Fig. 1 Analysis of UGT2 expression by qRT-PCR after treatments with NaCl, ABA, and PEG. Experiments were conducted for three biological replicates. Data shown are means ± SD. Student's t test was performed (*P < 0.05, **P < 0.01) in the presence of 400 ng/ml of Aureobasidin A allowing for a highly stringent interaction screening. The plates were incubated for 3 days at 30 °C.

ChIP-qPCR experiment
The ChIP assay were used with the EpiQuik™ Plant ChIP Kit (EpiGentek, https:// www. epige ntek. com). Seven-dayold rice seedlings were treated with 150 mMNaCl or water (control) for 12 h, respectively, and 3 g of rice seedlings were taken and fixed with 1% v/v formaldehyde. The rice chromatin was extracted and sheared by sonication to a size range of ~ 200-1000 bp and subjected to the ChIP assay. The specific primers were designed and used to detect the corresponding promoters in the ChIP products. The rabbit polyclonal antibody raised against synthetic peptide OsbZIP23 was ordered from Beijing protein innovation Co., Ltd. (Beijing, China; http:// www. prote omics. org. cn). Three biological repeats were conducted.

Dual-luciferase experiment
The OsZIP23 and OsABI5 coding sequences were amplified, and the resulting sequences were introduced into effector vectors under the control of the CaMV 35S promoter. The UGT2 promoter sequence was amplified and introduced into the pGreenII0800-LUCreporter vector to drive the expression of luciferase gene. Effector vector and reporter vector were simultaneously transferred into Arabidopsis thaliana protoplasts. Firefly luciferase (LUC) and Renilla luciferase (REN) activities were measured using the Dual-Luciferase Reporter Assay System (www. prome ga. com).

Determination of activities of ROS-scavenging enzymes
The rice leaves were detached from 7-day-old plants after 200 mM NaCl treatment for 12 h, and the activities of ROS-scavenging enzymes such as SOD, CAT, and APX were determined according to a previously described method (Cai et al. 2015).

Statistical analysis of experimental data
All experimental data were performed with at least three independent biological replicates, each containing three technical replicates. Significant differences of experimental data compared to wild type or control group were analyzed by Student's t test (*p < 0.05; **P < 0.01).

Expression patterns of UGT2 during salt and ABA treatments
The Genevestigator (https:// genev estig ator. com/ gv) was used in our initial screen for rice UDP-glycosyltransferase (UGT) genes responsive to abiotic stresses. The UGT2 gene (Os11g0444000) was found to be up-regulated by salt stress. To verify this response, the transcript levels of UGT2 in rice leaf tissues were analyzed during treatments with high salinity, ABA and polyethylene glycol (PEG) (Fig. 1). Our data showed that the UGT2 transcript was clearly induced by NaCl and ABA, but not by PEG. These data suggested that UGT2 was possibly involved in salt stress response in rice.

UGT2 positively regulates salt tolerance in rice
To further study the physiological role of UGT2 gene, we generated UGT2 knock-out mutants (ko lines) and overexpression lines (OE lines). Among those lines, the ugt2ko1 and ugt2ko2 are base deletion and insertion mutants, respectively, generated by the CRISPR/Cas9 system ( Fig. 2A, B). UGT2OE11 and UGT2OE12 are overexpression plants with much higher UGT2 transcript levels than those of wild-type (WT) plants (Fig. 2C).
To determine the sensitivity of mutants and overexpression plants to salt stress, seeds of WT and transgenic plants were germinated and then placed in water containing 0 mM, 50 mM or 100 mM NaCl. There were no differences in shoot and root lengths between WT and transgenic plants under normal conditions (0 mM NaCl). However, under NaCl treatment conditions, the average root and shoot lengths of UGT2OE lines were increased compared to that of WT plants, root and shoot lengths of ugt2ko lines were decreased (Fig. 2D, E).
Besides, the soil-grown UGT2 overexpression lines, ugt2ko mutant lines and WT plants were also subjected Fig. 2 Growth phenotypes of transgenic rice lines under salt stress. A Preparation of ugt2 mutants using CRISPR/Cas9 technology. A base deletion occurs in ugtko1, whereas a base insertion occurs in ugtko2. B The amino acid sequences of the mutants. After the red region, frameshift and premature termination occur in mutants. Symbol "*" indicates the location of premature termination. C Expression level of UGT2 in each overexpressing lines. D Phenotype of rice seedlings under salt stress (Scale bar, 1.5 cm). E Statistics of length of root and shoot of rice under salt stress. F, G Phenotypes of rice lines under salt treatment in soil (Scale bar, 15 cm). H Contents of proline and soluble sugar in transgenic rice lines after salt treatment. Experiments were conducted for three biological replicates. Data shown are means ± SD. Student's t test was performed (*P < 0.05, **P < 0.01) ◂ to salt stress (Fig. 2F, G). We observed that mutant plants exhibited earlier wilted phenotypes than WT after exposure to salt stress, and UGT2 overexpression lines showed increased tolerance than WT to NaCl treatments. Almost 75% of UGT2 overexpression plants and 60% of WT plants survived, whereas only 35% of the mutant plants survived under this treatment. Therefore, these results evidence that overexpression of UGT2 enhanced the salt stress tolerance in rice.
Additionally, the proline and soluble sugars, which serve as osmotic regulators, were also accumulated more in UGT2 overexpression plants and less in ugt2ko mutant plants after treatment with 200 mM NaCl (Fig. 2H). The results showed that UGT2OE can accumulate more osmoprotectants than wild-type under salt stress.

Ectopic expression of UGT2 in Arabidopsis enhances salt tolerance
Two homozygous Arabidopsis transgenic lines (OE1 and OE2) ectopically expressing UGT2 were used to investigate the salt tolerance. In the non-treatment group, root growth of Arabidopsis transgenic lines in vertical culture was similar with WT. However, a significant difference in root length was found between the transgenic plants and WT plants upon exposure to 150 or 200 mM NaCl after 7 days (Fig. 3A). Under salt stress conditions, the transgenic plants had longer roots compared to WT plants.
Soil-grown plants were also subjected to salt treatment experiments. Before salt treatment, transgenic Arabidopsis plants and WT showed uniform growth. After exposed to 200 mM NaCl for 15 days, the transgenic plants exhibited better growth and higher survival rate. However, the non-transgenic WT plants wilted (Fig. 3B). These results demonstrated that ectopic expression of the rice UGT2 gene enhanced salt tolerance of the transgenic Arabidopsis plants.

The physiological and biochemical changes in UGT2 overexpression and mutant plants after salt stress
It is well known that ROS produced in abiotic stresses can lead to oxidative damage. Thus, the ROS-scavenging capacity in plant cells was also an important index to indicate the stress tolerance. Here, we used NBT staining to investigate the accumulation of superoxide radicals (O2 − ·) in different transgenic plants under NaCl treatments. The results showed that no significant difference in the level of superoxide radicals between overexpression lines, mutant lines and wild-type lines under control conditions. After salt stress, however, UGT2 overexpression lines accumulated less superoxide radicals and ugt2ko mutant lines accumulated much more than WT in rice, which can be estimated from the coloration (Fig. 4A). Besides, we also examined the antioxidant enzyme activities of CAT, APX, and SOD in WT and transgenic plant under normal and salt stress conditions, and found that they were higher in overexpression lines than in WT in response to salt stress, but lower in the mutant/ lines than in WT (Fig. 4B). We determined the expression of antioxidant enzyme encoding genes, including OsAPX2, OsCATA , and OsCATB. Expression levels of antioxidant enzyme encoding genes in mutants and overexpression lines supported our observation from NBT staining and enzyme activity determination (Fig. 4C). These results indicated that the UGT2 overexpression lines had a stronger ability to remove active oxygen and accumulated less ROS under stress conditions.
To investigate the potential pathways affected by UGT2, we evaluated the expression of several abiotic stress-regulated genes using qRT-PCR. Although no significant difference in the expressions of OsDREB2A, OsLEA3, and OsP5CS was observed between overexpression lines, mutant lines and WT under control condition, the expressions of those abiotic stressrelated genes were higher in overexpression plants, while lower in the mutant lines than WT under NaCl treatment (Fig. 4D). The results imply that these abiotic stress-responsive genes might contribute to the salt stress tolerance of plants conferred by UGT2.

OsbZIP23 positively regulates UGT2 transcription by physically interacting with the UGT2 promoter
To know the upstream regulation mechanism of UGT2 gene, we scanned its promoter regions with the online tool PLACE (http:// www. dna. affrc. go. jp/ htdocs/ PLACE/), and found that UGT2 promoter bears a few stress-regulated cis-elements, including ABRE (CAC GTG ) and G-box (ACGT) elements (Noman et al. 2019;Ji et al. 2015), which are supposed to be bound by bZIP transcription factors and are involved in abiotic stress response (Fig. 5A).
We then selected several typical bZIP transcription factors involved in salt stress regulation, including OsbZIP23, OsABI5, OsABF2, and OsbZIP71, to explore their possible binding capability with UGT2 promoter. Yeast one-hybrid (Y1H) assays were performed to test the interaction between bZIP proteins and those cis-elements on UGT2 promoter. As shown in Fig. 5B, we found that OsbZIP23 had the strong capability of binding with the UGT2 promoter in Y1H assays. OsABI5 showed only a weak binding.
Previous studies had shown that ABRE (CAC GTG ) is the core sequence in the promoter of OsbZIP23-binding target genes (Zong et al. 2016). Here, a qRT-PCR-based chromatin immunoprecipitation (ChIP) assay was deployed to monitor the binding affinity of OsbZIP23 to the UGT2 gene promoter. One-week-old wild-type rice seedlings exposed to either 0 or 150 mM NaCl for 12 h were used as plant materials. Fragments of the UGT2 promoter, containing ABRE element (b) or far away ABRE element (a), were amplified by PCR to monitor the fragment enrichment after chromatin immunoprecipitation using the rabbit polyclonal antibody raised against synthetic peptide OsbZIP23 (Fig. 5C). Our results showed that the 'b' fragments were enriched much more than 'a' fragments under salt stress, but they were enriched to similar extents under control conditions (Fig. 5D), indicating that OsbZIP23 preferentially associated to the chromatin region containing a ABRE element of UGT2 promoter in vivo under salt stress.
To further confirm the transcription activating of UGT2 by OsbZIP23 or OsABI5, a dual-luciferase (LUC) reporter plasmid was constructed, which encodes a firefly LUC gene driven by the UGT2 promoter (− 2000 to 0 bp) and a Renilla luciferase (REN) gene driven by the constitutive 35S promoter. The constructs 35S:OsbZIP23 and 35S:OsABI5 were used as effectors, 35S:GFP was the empty plasmid control (Fig. 5E). Our experiment results showed that overexpression of OsbZIP23 can stimulate the luciferase activity of the UGT2 reporter as compared to the empty plasmid control. OsABI5 had only weak stimulation activity (Fig. 5F). These results suggested that OsbZIP23 is the major upstream regulator for transcription activity of UGT2.

UGT2 increases the plant sensitivity to exogenous ABA
It is reported that OsbZIP23 plays a critical role in mediating ABA sensitivity in rice and OsbZIP23-overexpressing lines were hypersensitive to ABA (Xiang et al. 2008). Given that UGT2 is directly regulated by the transcription factor OsbZIP23, it is necessary to test the ABA sensitivity of UGT2 transgenic plants. We examined the responses of UGT2 overexpression lines and mutant lines in terms of shoot and root growth upon ABA treatment. It was observed that the shoot and root growth of UGTOE11 and UGTOE12 seedlings were more sensitive to ABA treatment, whereas ugtko1 and ugtko2 were less sensitive to ABA treatment (Fig. 6A, B). These results further supported our conclusion that UGT2 is bound and regulated by bZIP23 transcription factor and is involved in abiotic stress response.

Discussion
Salinity is a major abiotic stress causing serious agricultural problem and yield reduction of crops worldwide (You et al. 2014). To grow and survive, plants have evolved a variety of mechanisms to resist abiotic stresses during long-term evolution (Kerr et al. 2018). Studying the molecular mechanisms of plant adaptation to abiotic stresses can help to solve crop production and yield problems.
Rice genome contains a big family of UGTs more than 200 members. However, most of them are not yet elucidated about their functions and their involvement in metabolism. In this study, a novel rice glycosyltransferase gene UGT2 was identified and demonstrated that it is involved in the notype of soil-grown UGT2 transgenic Arabidopsis. Experiments were conducted for three biological replicates. Data shown are means ± SD. Student's t test was performed (*P < 0.05, **P < 0.01) salt stress response of plants. The C-terminus of the UGT2encoded protein contains the conserved sequence (PSPG box) of family 1 glycosyltransferases, UGT2 is thus a member of glycosyltransferase family 1. qRT-PCR analysis showed that UGT2 was up-regulated by salt and ABA treatment, and the stress experiments also confirmed that UGT2 was involved in salt stress response. For example, overexpression of UGT2 can significantly enhance the tolerance of plants to salt stress, in contrast, the loss of the gene's function leads to the plant sensitivity to salt stress. These results fully proved that the rice glycosyltransferase gene UGT2 Fig. 4 Superoxide radical accumulation and expression level of genes related to antioxidant enzymes and stress response in UGT2 overexpression lines and mutant lines. A NBT staining of rice leaves under control and salt stress conditions. B Activities of ROS-scavenging enzymes were determined. C Expression level of genes related to antioxidant enzymes in mutants and overexpression lines. Transcript levels were normalized to the mRNA levels of OsActin. D Expression levels of genes related to stress response after salt treatment in mutants and overexpression lines. Experiments were conducted for three biological replicates. Data shown are means ± SD. Student's t test was performed (*P < 0.05, **P < 0.01) functions in the salt stress response of plants and positively regulates salt tolerance.
Overexpression and knockout of the UGT2 gene resulted in significant changes in ABA sensitivity and stress resistance. The UGT2-overexpressing rice plants showed significantly increased sensitivity to exogenous ABA. On the contrary, the mutants of this gene showed significantly decreased sensitivity to ABA. These observations are consistent with the OsbZIP23 function. Previous studies have found that the OsbZIP23 transcription factor plays an important role in the regulation of plant salt tolerance (Xiang et al. 2008). OsbZIP23 was up-regulated by salt stress, and the mutation of this gene reduced the tolerance to high salt, in contrast, the overexpression of this gene increased the tolerance to salt (Xiang et al. 2008). In this study, yeast one-hybrid experiment, ChIP-qPCR experiment, and dual-luciferase reporter experiment confirmed that OsbZIP23 directly binds to the UGT2 promoter and positively regulates UGT2 expression. This result indicated that UGT2 is directly regulated by the salt stress-related transcription factor OsbZIP23. Existing studies have learned that there are many downstream target genes of OsbZIP23 through RNA-seq and ChIP-seq analysis, including drought, high salt and ABA-related genes (Xiang et al. 2008). Some studies had found that the OsbZIP23 can be phosphorylated by OsSAPK2 and positively regulate OsPP2C49 and OsNCED4, thereby participating in the drought response of plants (Zong et al. 2016). Lou et al. found that OsSAPK2 significantly improved the tolerance of rice to salt and PEG stress (Lou et al. 2017). Therefore, we speculated that OsSAPK2 may positively regulate the expression of UGT2 gene by phosphorylating OsbZIP23 transcription factor. The upstream promoter sequence of UGT2 contains multiple stress-related cis-acting elements, including ABRE, DRE, MYBRS, and MYCRS. ABRE, MYBRS, and MYCRS can be recognized by AREB/ABF, MYB, and MYC transcription factors, respectively. These . Experiments were conducted for three biological replicates. Data shown are means ± SD. Student's t test was performed (*P < 0.05, **P < 0.01) cis-elements and corresponding transcription factors have important roles in abiotic stress responses (Yamaguchi-Shinozaki and Shinozaki 2005). OsbZIP23 was designated ABA-responsive element (ABRE)-binding factor and we have demonstrated that it can physically bind to the UGT2 promoter and activate UGT2 transcription. However, we cannot exclude the possibility that UGT2 may be regulated by other stress-relative transcription factors because of several different stress-related cis-acting elements on its promoter. Therefore, UGT2 gene may have a complex upstream regulatory mechanism, just reflecting the importance of UGT2 in response to abiotic stresses. Given that OsbZIP23 can bind to the promoter of UGT2 and enhance the transcription levels of UGT2, it would be better if we can investigate the growth phenotypes and survival rates of Osbzip23 ugt2 double mutant under NaCl treatment in rice. Unfortunately, we did not get the rice Osbzip23 ugt2 double mutant at the moments. We can speculate that Osbzip23 ugt2 double mutant would be less salt tolerant compared to Osbzip23 mutant and ugt2 mutant in rice.
UGT2 is a putative UDP-dependent glycosyltransferase, theoretically, which would catalyze the glycosylation of plant secondary metabolites including plant hormones. At present, the relevant substrate(s) has not been identified yet, although we examined a lot of secondary plant metabolites such as ABA, Auxins, Cytokinins, Gibberellins, Salicylic acid, and its derivatives, flavonoids, anthocyanins, phenylpropanoids etc. Thus, the precise metabolism pathway that UGT2 is involved in is not well known at the moment. But even so, this study has identified the function of the rice UGT2 and revealed that OsbZIP23-UGT2 module may play an important role in regulating the salt tolerance and downstream secondary metabolism in rice.

Fig. 6
The ABA sensitivity of UGT2 transgenic plants. A Growth performance under ABA treatments. B Measurements of shoot and root lengths under ABA treatments. (Scale bars: 2 cm). Experi-ments were conducted for three biological replicates. Data shown are means ± SD. Student's t test was performed (*P < 0.05, **P < 0.01)