An Inositol 1,3,4,5,6-Pentakisphosphate 2-Kinase 1 Mutant with a 33-nt Deletion Showed Enhanced Tolerance to Salt and Drought Stress in Rice

OsIPK1 encodes inositol 1,3,4,5,6-pentakisphosphate 2-kinase, which catalyzes the conversion of myo-inositol-1,3,4,5,6-pentakisphosphate to myo-inositol-1,2,3,4,5,6-hexakisphosphate (IP6) in rice. By clustered regularly interspaced short palindromic repeats (CRISPR) and CRISPR-associated protein (Cas9)-mediated mutagenesis in the 3rd exon of the gene, three OsIPK1 mutations, i.e., osipk1_1 (a 33-nt deletion), osipk1_2 (a 1-nt deletion), and osipk1_3 (a 2-nt deletion) were identified in T0 plants of the rice line Xidao #1 (wild type, WT). A transfer DNA free line with the homozygous osipk1_1 mutation was developed; however, no homozygous mutant lines could be developed for the other two mutations. The comparative assay showed that the osipk1_1 mutant line had a significantly lower level of phytic acid (PA, IP6; −19.5%) in rice grain and agronomic traits comparable to the WT. However, the osipk1_1 mutant was more tolerant to salt and drought stresses than the WT, with significantly lower levels of inositol triphosphate (IP3), reactive oxygen species (ROS) and induced IP6, and higher activities of antioxidant enzymes in seedlings subjected to these stresses. Further analyses showed that the transcription of stress response genes was significantly upregulated in the osipk1_1 mutant under stress. Thus, the low phytic acid mutant osipk1_1 should have potential applications in rice breeding and production.


Introduction
Phytic acid (PA), also identified as Myo-inositol-1,2,3,4,5,6-hexakisphosphate (IP 6 ), is thought to be the main storage form of nutrient phosphorous (P) (~80%) constituting 1.6% of the dry biomass in crop grains [1]. In seeds of most cereal, IP 6 occurs as phytates in the protein bodies, which chelate with several divalent metal ions e.g., Ca 2+ , Zn 2+ , Mg 2+ , and Fe 2+ [2]. During seed germination, endogenous phytase activity is induced and hydrolyzes the PA, releasing bound mineral cations, stored phosphorus (P), and inorganic phosphorus (Pi) that are utilized for seedling growth [3]. Due to the lack of phytase, monogastric animals cannot utilize PA and the micronutrients in phytates as a nutrient source [4]. Furthermore, the undigested phytic acid phosphorus (PA-P) in animal wastes has gradually become one of the main causes of P pollution [5]. Because of these PArelated environmental and nutritional adverse effects, mutagenesis and biotechnological approaches have been used to produce low phytic acid (lpa) mutant lines in cereals [6] such as maize [3,7], wheat [8], and rice [9][10][11][12][13]. Twelve genes predicted to encode enzymes in the PA metabolism pathway have been identified in rice seeds [14]. Mutations of these genes are recognized to create the phenotypes of LPA in cereals [5]. Inositol 1,3,4,5,6-pentakisphosphate 2-kinase 1 (IPK1) is reportedly to catalyze the final step in the biosynthesis of IP 6 , by which the inositol 1,3,4,5,6-pentakisphosphate (IP 5 ) is further phosphorylated at position 2 to form IP 6 [14][15][16][17]. The pathway of IP 6 biosynthesis reported in Saccharomyces cerevisiae [17] and Dictyostelium discoideum share the last step: the phosphorylation of IP 5 to IP 6 by IPK1 designated as a 2-kinase enzyme. The IPK1 mutant in S. cerevisiae presented a reduction in the capability to transfer mRNA from the nucleus and exhibited comprehensive incapability to synthesize PA. The AtIpk1-1 gene was examined in Arabidopsis using T-DNA insertion mutant, and the IP 6 level was decreased in the AtIpk1-1 mutant by 83% [18]. Ali et al. [19] produced transgenic rice by down-regulating the OsIPK1 expression using an RNAi-mediated technique. The reduction in IP 6 content and concomitant increase in phosphate (Pi) level were found in transgenic rice plants with a seed-specific decrease in the gene expression of OsIPK1. The seed germination and some agronomic characteristics of transgenic rice were similar to the wild type (WT) [19]. These findings are consistent with earlier research indicating that the mutations of IPK1 gene had no adverse effects on seed viability and agronomic performances in soybean [20,21].
The association of inositol phosphates in molecular and cellular responses of plants to abiotic stresses has been investigated. For instance, IP 3 was found to be transiently induced in plants under abscisic acid and salt stress [22,23]. The physiological and molecular proof of the role of IPK1 in the response to abiotic stresses is vague, although inositol phosphates are recognized as signaling molecules in the response to stresses. The clustered regularly interspaced short palindromic repeats (CRISPR) and CRISPR-associated protein 9 (Cas9), which is a precise and efficient approach for genome editing [24][25][26], has been used in crop breeding, particularly for generating low phytic acid (lpa) rice. In this study, we generated osipk1 mutants by CRISPR/Cas9-mediated mutagenesis and analyzed seed accumulation of phosphorus (total phosphorus, TP; phytic acid phosphorus, PA-P; inorganic phosphorus, Pi), agronomic traits, seed germination, and stress tolerance with the aim to evaluate and explore the probability of producing the LPA rice with no adverse effect on agronomic traits and seed viability, and with better tolerance to salt and drought stresses.

Mutations of OsIPK1 and Development of Homozygous Transgene-Free Mutant Lines
A homozygous osipk1_1 mutant line with a 33-nt deletion in the third exon of the gene was generated by CRISPR/Cas9-mediated mutagenesis. Twenty-one hygromycin phosphotransferase (HPT)-positive T 0 rice seedlings were selected from transformation with the vector of CRISPR/Cas9, pH_ipk1, and eight rice seedlings were finally verified mutated at the position of target. This signifies an editing efficiency of 38.1%, which manifests that the pH_ipk1 is efficient. Eventually, one homozygous transgene-free osipk1 mutant (osipk1_1) and two heterozygous transgene-free osipk1 mutants (osipk1_2 and os-ipk1_3) ( Figure 1A) were selected in T 0 plants. Rice seeds of the mutated T 0 seedlings were harvested and planted as lines of T 1 plants. Interestingly, we found that no presence of homozygous transgene-free in T 1 plants of osipk1_2 and osipk1_3, whereas we can test some homozygous osipk1_2 and osipk1_3 mutations in the seeds which were not germinated. The same findings also appeared in the T 2 and advanced-generation plants, so the heterozygous osipk1_2 and osipk1_3 mutants could not be used for further analysis. This may be due to a serious decline in phytic acid content in these IPK1 knock-out plants, which prevented seeds from germinating. Approximately ten surviving seedlings of the T 1 line of osipk1_1 were verified for the presence of target mutations and T-DNA at the seedling stage after bentazon treatment. All T 1 seedlings of the selected line were further confirmed for both the presence of mutation and T-DNA. Rice seeds from the confirmed homozygous, transgene-free mutant seedlings selected as the mutant line osipk1_1, were harvested and used for further evaluation. The mutation of osipk1_1 (a 33-nt deletion) would cause an 11-amino acids fragment missing from amino acid positions 65 to 75 ( Figure S1). Investigation of IPK1 proteins in nine organisms implied that the loss of 11-amino acids in the osipk1_1 mutant was positioned in a highly conserved segment ( Figure S2), signifying that the osipk1_1 mutation could have potential functional consequences. In contrast, other mutations, osipk1_3 (a 2nt deletion) and osipk1_2 (a 1-nt deletion) would produce a premature stop codon almost right after the site of mutation ( Figure S1). Consequently, the mutant alleles of osipk1_3 and osipk1_2 were predicted to generate proteins with only 75 and 89 amino acids ( Figures  1B and S1). This may be the major reason why the homozygous transgene-free osipk1_2 and osipk1_3 mutants could not germinate.

Plant Growth and Phosphorus Content of Osipk1_1 Mutant
Plant growth of osipk1_1 was comparable with WT. No significant differences in the plant height, panicle length, tiller numbers per plant, seed-set, and 1000-grain weight were observed between Xidao 1 and osipk1_1 (Figure 2A-E). The seed germination rate of osipk1_1 was lower in the first four days, but caught up with that of the WT after five days, and almost the same on the 7th day ( Figure 2F). The mutation of osipk1_1 (a 33-nt deletion) would cause an 11-amino acids fragment missing from amino acid positions 65 to 75 ( Figure S1). Investigation of IPK1 proteins in nine organisms implied that the loss of 11-amino acids in the osipk1_1 mutant was positioned in a highly conserved segment ( Figure S2), signifying that the osipk1_1 mutation could have potential functional consequences. In contrast, other mutations, osipk1_3 (a 2-nt deletion) and osipk1_2 (a 1-nt deletion) would produce a premature stop codon almost right after the site of mutation ( Figure S1). Consequently, the mutant alleles of osipk1_3 and osipk1_2 were predicted to generate proteins with only 75 and 89 amino acids ( Figure 1B and Figure S1). This may be the major reason why the homozygous transgene-free osipk1_2 and osipk1_3 mutants could not germinate.

Plant Growth and Phosphorus Content of Osipk1_1 Mutant
Plant growth of osipk1_1 was comparable with WT. No significant differences in the plant height, panicle length, tiller numbers per plant, seed-set, and 1000-grain weight were observed between Xidao 1 and osipk1_1 (Figure 2A-E). The seed germination rate of osipk1_1 was lower in the first four days, but caught up with that of the WT after five days, and almost the same on the 7th day ( Figure 2F). The colorimetric assay showed that osipk1_1 had a similar Pi level to the WT control ( Figure 3A). To measure the mutational effects of osipk1_1, the contents of Pi, PA-P, and TP were evaluated in seeds between the mutant line and WT. The osipk1_1 mutant line had greatly lower PA-P and TP content compared with the control, while the content of Pi was not greatly different from that of WT ( Figure 3B-D). The seeds of WT had a Pi level of 0.29 mg/g, which was not significantly different from the osipk1_1 mutant line ( Figure  3B). osipk1_1 had a PA-P level of 1.88 mg/g, which was 19.5% lower than the control (2.34 mg/g) ( Figure 3C). osipk1_1 had TP content of 2.60 mg/g, which were 17.6% lower than the control (3.16 mg/g) ( Figure 3D). The colorimetric assay showed that osipk1_1 had a similar Pi level to the WT control ( Figure 3A). To measure the mutational effects of osipk1_1, the contents of Pi, PA-P, and TP were evaluated in seeds between the mutant line and WT. The osipk1_1 mutant line had greatly lower PA-P and TP content compared with the control, while the content of Pi was not greatly different from that of WT ( Figure 3B-D). The seeds of WT had a Pi level of 0.29 mg/g, which was not significantly different from the osipk1_1 mutant line ( Figure  3B). osipk1_1 had a PA-P level of 1.88 mg/g, which was 19.5% lower than the control (2.34 mg/g) ( Figure 3C). osipk1_1 had TP content of 2.60 mg/g, which were 17.6% lower than the control (3.16 mg/g) ( Figure 3D).

Mutant of Osipk1_1 Has a Better Tolerance against Drought or Salt Stress than Wild Type (WT)
To verify whether the mutant of osipk1_1 also has any effect on stress tolerance, we tested the osipk1_1 and WT seedlings to salt and drought stress treatment. The growth of osipk1_1 seedling was similar to that of the WT when grown under normal conditions, while better than the control under either salt (100 mM NaCl) or drought (20 mM mannitol) stress conditions ( Figure 4A).
After 7-day treatment by 20 mM mannitol or 100 mM NaCl, the plant height, root length, and dry biomass of WT seedlings were greatly decreased ( Figure 4B-D). In contrast, the plant height, root length, and dry biomass of osipk1_1 were significantly higher than that of WT plants after stress treatments ( Figure 4B-D), signifying that the mutant of osipk1_1 is much more tolerant against drought or salt stress than WT seedlings.

Mutant of Osipk1_1 Has a Better Tolerance against Drought or Salt Stress than Wild Type (WT)
To verify whether the mutant of osipk1_1 also has any effect on stress tolerance, we tested the osipk1_1 and WT seedlings to salt and drought stress treatment. The growth of osipk1_1 seedling was similar to that of the WT when grown under normal conditions, while better than the control under either salt (100 mM NaCl) or drought (20 mM mannitol) stress conditions ( Figure 4A).
After 7-day treatment by 20 mM mannitol or 100 mM NaCl, the plant height, root length, and dry biomass of WT seedlings were greatly decreased ( Figure 4B-D). In contrast, the plant height, root length, and dry biomass of osipk1_1 were significantly higher than that of WT plants after stress treatments ( Figure 4B-D), signifying that the mutant of osipk1_1 is much more tolerant against drought or salt stress than WT seedlings.

The Osipk1_1 Mutant Accumulated Less Reactive Oxygen Species (ROS) than WT under Salt and Drought Stresses
To explore how mutation of osipk1_1 alleviated salt and drought stress, we measured the contents of IP 3 , IP 6 , stress-related free amino acid proline (Pro), ROS (H 2 O 2 ), malondialdehyde (MDA), and anti-oxidant enzyme activities (peroxidase (POD), catalase (CAT), and superoxide dismutase (SOD)) in seedlings subjected to stresses (Figures 5-7). Without treatment, all these parameters were not significantly different between WT and osipk1_1.  Thirdly, we found that seven days of stress treatment greatly decreased the anti-oxidant enzymes (POD, CAT, and SOD) activities, relative to control in WT (Figure 7). Following seven days of stress treatment, the anti-oxidant enzymes activities of CAT, POD, and SOD were also decreased in the mutant of osipk1_1, while with less magnitude compared those of WT (Figure 7).
These findings exhibited that osipk1_1 decreased the ROS level and mitigated oxidative stresses with induced antioxidant enzyme activities and decreased level of IP3, compared to WT, under drought and salt stresses.  Thirdly, we found that seven days of stress treatment greatly decreased the anti-oxidant enzymes (POD, CAT, and SOD) activities, relative to control in WT (Figure 7). Following seven days of stress treatment, the anti-oxidant enzymes activities of CAT, POD, and SOD were also decreased in the mutant of osipk1_1, while with less magnitude compared those of WT (Figure 7).
These findings exhibited that osipk1_1 decreased the ROS level and mitigated oxidative stresses with induced antioxidant enzyme activities and decreased level of IP3, compared to WT, under drought and salt stresses. The different letters (a, b, c, d, and e) show the significant difference at a probability of p ˂ 0.05.

Transcription of Stress Response Genes Was Significantly Upregulated in the Osipk1_1 Mutant under Stress
To explore the molecular mechanism of how osipk1_1 enhanced stress tolerance, the analysis of gene expression was conducted for phytic acid biosynthetic genes and stressrelated genes in rice plants.
Mutant of osipk1_1 had a significantly higher abundance of phytic acid biosynthesis genes (except for OsIPK1) in the absence of stress treatment ( Figure S3). Gene expression of OsIPK2, OsITPK1, OsITPK2, OsITPK3, and OsITPK6 were significantly induced, while OsIPK1, OsITPK4, and OsITPK5 expression were significantly decreased in WT after treatment ( Figure S3). In osipk1_1, the abundance of OsIPK2, OsITPK1, OsITPK2, OsITPK3, and OsITPK6 were significantly lower and OsIPK1, OsITPK4, and OsITPK5 were significantly higher compared to WT seedlings after treatment ( Figure S3).
The expression of stress responsive genes was not significantly different between WT and osipk1_1 without treatment (Figure 8). Exposure to salt and drought stress greatly induced the transcription of all eight tested genes (OsPOX8.1, OsPOX8.1, OsP5CS, OsRab16D, OsGDSL, OsbZIP23, OsSNAC1, and OsDRB1A) in WT ( Figure 8). Furthermore, osipk1_1 had a significantly increased abundance of these stress-response genes with salt and drought treatment compared to WT seedlings ( Figure 8).
These results demonstrated that the expression of phytic acid biosynthetic genes was deregulated and stress-response genes were significantly up-regulated in osipk1_1 mutant under salt and drought stress conditions. Firstly, we found that the level of IP 3 was significantly increased after salt or drought stress treatment in WT, i.e., by 1.62-fold with salt stress, and by 1.69-fold with drought stress, respectively, relative to control after treatment for seven days ( Figure 5A). The content of IP 3 was also significantly increased in osipk1_1, but was still lower than that in WT after one-week treatment of salt or drought stress ( Figure 5A). We also observed that seven days of stress treatment greatly reduced the accumulation of IP 6 in WT ( Figure 5B). In contrast, a significantly increased level of IP 6 was observed in osipk1_1 compared to WT ( Figure 5B).
Secondly, we found that the accumulation of Pro, H 2 O 2 , and MDA were significantly increased after salt or drought stress treatment in WT (Figure 6). Following a one-week salt or drought stress treatment, the contents of Pro, H 2 O 2 , and MDA were also elevated in the mutant of osipk1_1, although with less magnitude as those of WT ( Figure 6).
Thirdly, we found that seven days of stress treatment greatly decreased the antioxidant enzymes (POD, CAT, and SOD) activities, relative to control in WT (Figure 7). Following seven days of stress treatment, the anti-oxidant enzymes activities of CAT, POD, and SOD were also decreased in the mutant of osipk1_1, while with less magnitude compared those of WT (Figure 7).
These findings exhibited that osipk1_1 decreased the ROS level and mitigated oxidative stresses with induced antioxidant enzyme activities and decreased level of IP 3 , compared to WT, under drought and salt stresses.

Transcription of Stress Response Genes Was Significantly Upregulated in the Osipk1_1 Mutant under Stress
To explore the molecular mechanism of how osipk1_1 enhanced stress tolerance, the analysis of gene expression was conducted for phytic acid biosynthetic genes and stressrelated genes in rice plants.
Mutant of osipk1_1 had a significantly higher abundance of phytic acid biosynthesis genes (except for OsIPK1) in the absence of stress treatment ( Figure S3). Gene expression of OsIPK2, OsITPK1, OsITPK2, OsITPK3, and OsITPK6 were significantly induced, while OsIPK1, OsITPK4, and OsITPK5 expression were significantly decreased in WT after treatment ( Figure S3). In osipk1_1, the abundance of OsIPK2, OsITPK1, OsITPK2, OsITPK3, and OsITPK6 were significantly lower and OsIPK1, OsITPK4, and OsITPK5 were significantly higher compared to WT seedlings after treatment ( Figure S3).
The expression of stress responsive genes was not significantly different between WT and osipk1_1 without treatment (Figure 8). Exposure to salt and drought stress greatly induced the transcription of all eight tested genes (OsPOX8.1, OsPOX8.1, OsP5CS, OsRab16D, OsGDSL, OsbZIP23, OsSNAC1, and OsDRB1A) in WT (Figure 8). Furthermore, osipk1_1 had a significantly increased abundance of these stress-response genes with salt and drought treatment compared to WT seedlings (Figure 8).

Discussion
The CRISPR/Cas9 system has been broadly utilized in plant functional genomics studies. Our study intended to offer such an instance on the manipulation of PA content in rice grains by targeted mutagenesis of a PA synthesis gene. To develop rice genetic resources with agronomically competitive LPA, we targeted mutagenesis of osipk1 using CRISPR/Cas9-mediated genome editing technology in this research. The homozygous and transgene-free mutant, named as the mutant line of osipk1_1, was evaluated for its seed P level together with the WT. The mutant line had significantly lower PA-P and TP contents compared with the control, while the Pi had no significant difference compared to the WT ( Figure 3). IPK1 is believed to catalyze the last step in inositol metabolism by using the inositol 1,3,4,5,6-pentaphosphate as the substrate [14][15][16][17]. This process only adds one phosphate group, which may result in a significantly decreased level of phytic acid, while the Pi content of osipk1_1 does not significantly change. By contrast, the Pi level of transgenic rice was increased and the TP was similar to the WT [19]. In our present study, the agronomic traits and seed germination rate of osipk1_1 were not significantly different from those of the WT, which was similar to the results of Ali et al. [19]. It was reported that seed-specific downregulation of OsIPK1 did not negatively affect agronomic traits, but increased tolerance to artificial aging and seed viability [19]. The findings implied that the editing of the OsIPK1 gene in rice seeds might be a valid way to generate LPA rice.
Terrestrial plants are frequently faced with numerous abiotic stresses throughout their whole life. Physiological, biochemical, and molecular responses to such stresses are mediated by a group of signal transduction pathways. Former studies have found that IP3 can be a vital second messenger, which always displays a transient induction upon exposure to exogenous stimuli, e.g., pathogens, reactive oxygen, light, osmotic, and salt stresses [23,27,28]. The relationship between abiotic stresses and metabolism of IP3 has been observed in mammalian cells and yeast (Saccharomyces cerevisiae) [22,23,29], but very little research has been described regarding the association between abiotic stresses and These results demonstrated that the expression of phytic acid biosynthetic genes was deregulated and stress-response genes were significantly up-regulated in osipk1_1 mutant under salt and drought stress conditions.

Discussion
The CRISPR/Cas9 system has been broadly utilized in plant functional genomics studies. Our study intended to offer such an instance on the manipulation of PA content in rice grains by targeted mutagenesis of a PA synthesis gene. To develop rice genetic resources with agronomically competitive LPA, we targeted mutagenesis of osipk1 using CRISPR/Cas9-mediated genome editing technology in this research. The homozygous and transgene-free mutant, named as the mutant line of osipk1_1, was evaluated for its seed P level together with the WT. The mutant line had significantly lower PA-P and TP contents compared with the control, while the Pi had no significant difference compared to the WT (Figure 3). IPK1 is believed to catalyze the last step in inositol metabolism by using the inositol 1,3,4,5,6-pentaphosphate as the substrate [14][15][16][17]. This process only adds one phosphate group, which may result in a significantly decreased level of phytic acid, while the Pi content of osipk1_1 does not significantly change. By contrast, the Pi level of transgenic rice was increased and the TP was similar to the WT [19]. In our present study, the agronomic traits and seed germination rate of osipk1_1 were not significantly different from those of the WT, which was similar to the results of Ali et al. [19]. It was reported that seed-specific downregulation of OsIPK1 did not negatively affect agronomic traits, but increased tolerance to artificial aging and seed viability [19]. The findings implied that the editing of the OsIPK1 gene in rice seeds might be a valid way to generate LPA rice.
Terrestrial plants are frequently faced with numerous abiotic stresses throughout their whole life. Physiological, biochemical, and molecular responses to such stresses are mediated by a group of signal transduction pathways. Former studies have found that IP 3 can be a vital second messenger, which always displays a transient induction upon exposure to exogenous stimuli, e.g., pathogens, reactive oxygen, light, osmotic, and salt stresses [23,27,28]. The relationship between abiotic stresses and metabolism of IP 3 has been observed in mammalian cells and yeast (Saccharomyces cerevisiae) [22,23,29], but very little research has been described regarding the association between abiotic stresses and inositol phosphates in terrestrial plants. In tomato plants, genetically decreasing IP 3 by increasing InsP 3 hydrolysis caused enhanced drought stress tolerance [30]. The content of IP 3 in osipk1_1 was increased by salt and drought stresses in WT and showed a reduced level in osipk1_1 ( Figure 5A). Delineation of the stress-insensitive phenotype of osipk1_1 would help understand the stress signaling pathway leading to less accumulation of IP 3 and a non-standard level of a secondary messenger.
Free proline has been considered to contribute to the protection of macromolecules during dehydration [31] and the osmotic adjustment [32] as well as to be an important scavenger of hydroxyl radical [33]. The free proline content was induced following stress treatment but significantly lower in osipk1_1 ( Figure 6C). A similar change was observed for MDA ( Figure  6B), which is an important intermediate in ROS scavenging under abiotic stress, and is toxic to the plant cells if over accumulated [34,35]. The activities of antioxidant enzymes can be inducible by oxidative stresses produced by salt and drought stresses. In this study, however, POD, CAT, and SOD activities were lower after treatment of 100 mM NaCl and 20 mM mannitol (Figure 7). Phytic acid is a natural antioxidant in plants, constituting 1-5% of most legumes, nuts, oilseeds, cereals, spores, and pollen [36,37]. Although IP 6 was significantly reduced by salt or drought stress in both WT and mutant, a significantly higher level occurred in osipk1_1 ( Figure 5B). It is inferred that more accumulation of IP 6 will alleviate the stresses of salinity and drought. The decreased activities of antioxidant enzymes and phytic acid may have resulted from the oxidative burst in plants under salt and drought stresses as revealed by the ROS level, e.g., H 2 O 2 ( Figure 6A) in stressed plants. These oxidative stresses appeared to be greatly alleviated in osipk1_1 (Figure 6), signifying that improvement of the capacity of anti-oxidant defense is the main mechanism for mitigation of abiotic stress toxicity in rice seedlings of osipk1_1. Our findings propose that osipk1_1 had improved capacity for ROS scavenging and osmotic adjustment.
We further examine whether the expression of OsIPK2 and the other six ITPK members in rice have any changes in osipk1_1 under salt or drought stress. The OsITPK genes' exonintron organization structures suggest that three ITPKs in rice (OsITPK1, OsITPK2, and OsITPK3) belong to subgroup I, with each including 9 introns and 10 exons with similar phases [38]. OsITPK4 and OsITPK5 included no intron and belong to subgroup II. OsITPK6, belonging to subgroup III, have 11 introns and 12 exons. The analysis of the expression level proposes that the strictly related OsITPKs have small differences in the patterns of gene expression in their responses to various stresses.
Previous studies showed that OsPOX22.3 and OsPOX8.1, both encoding plastidic peroxidases, are involved in the tolerance to oxidative stress [39]. OsP5CS (LOC_Os05g38150) encodes an important enzyme for biosynthesis of proline and OsGDSL (LOC_Os02g57110), a GDSLlike lipase gene, was increased by osmotic stress [40]. OsSKIPa (LOC_Os02g52250), OsbZIP23 (LOC_Os02g52780), SNAC1 (LOC_Os11g03300), and OsDREB1A (LOC_Os09g35030), all being transcription factor genes, were also reported to be responsible for the tolerance to oxidative stress [41][42][43][44]. The gene expression levels of the above genes were greatly higher in osipk1_1 than that in the WT under abiotic stress. These findings imply that the transcription factors above are associated with the downstream manipulation of the responses to abiotic stress in osipk1_1.
Since there is only one tested mutant of osipk1 in our research, its excellent performance may be caused by the 33-nt deletion of OsIPK1, but it may also be due to accidental occurrence during the material cultivation process, e.g., tissue culture-derived variation, transposon (Tos17) insertion mutagenesis, pleiotropy. Firstly, antibiotics, like hygromycin in our research, might also enhance variation frequency [45,46], and thus several round screenings of calli on hygromycin-containing media might significantly induce mutations in rice. Second, to generate transgenic plants by using Agrobacterium transformation spent more time than regular tissue culture, while it has been well documented that the longer the tissue culture time, the higher the somaclonal variation frequency, because the somaclonal variation occurred throughout the culture process [47]. Therefore, further studies are needed to ascertain the real causative factor(s) leading to the enhanced tolerance of osipk1_1 to salt and drought stresses. For example, progenies derived from a cross between osipk1_1 and its WT parent could be used for such purposes.
In conclusion, by targeted mutagenesis of the inositol 1,3,4,5,6-pentakisphosphate 2-kinase gene using the CRISPR/Cas9 system, we generated osipk1 mutant lines. Physiological, biochemical, and molecular data showed that the osipk1_1 mutant had a better tolerance to salt and drought stresses. While the genetic cause remains undetermined, the enhanced tolerance of osipk1_1 might be exploited in rice breeding.

Generation of Mutants Using Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) and CRISPR-Associated Protein (Cas9)
To produce osipk1 mutants, the target site was chosen for OsIPK1 (LOC_Os 04g56580/Os 04g0661200) in the third exon ( Figure 1A). The UniProt (http://www.uniprot.org/) was used for target selection to minimize the off-target effect [48] and the sgRNAs for the precise site were designed by the CRISPR-P 1.0 website (http://crispr.hzau.edu.cn/CRISPR/). DNA base pairing sequence (C1-F and C1-R, Table S1) was synthesized by Tsingke (Hangzhou, China) for the construction of the vector of CRISPR/Cas9 (pH_ipk1) carrying an element of CYP81A6-hpRNAi using the plasmid of pHUN4c12s [49], which was improved from pHUN4c12 [50]. The plasmid pH_ipk1 was created and transformed into Agrobacterium tumefaciens and introduced into rice calli induced from the mature seed of 'Xidao 1 (a japonica rice cultivar) according to Li et al. [51].

Growth and Identification of Mutants
The transgenic plantlets were regenerated from hygromycin-resistant calli and acclimatized in a growth chamber (a photoperiod of 12 h, 30 • C) for seven days before being moved into outdoor facilities. Total genomic DNA was extracted from leaves of transgenic T 0 seedlings according to a modified method of cetyltrimethylammonium bromide (CTAB) [52]. The existence of the gene of hygromycin phosphotransferase (HPT) was tested by polymerase chain reaction (PCR) amplification using the primer pairs HygR-F and HygR-R (Table S1) [53]. The site-specific mutation was identified by amplification of PCR using primers (P1-F and P1-R, Table S1) encompassing the target sites in the gene of OsIPK1 ( Figure 1A). The analysis of high resolution melt (HRM) was carried out according to Li et al. [54] to verify for the mutations. The fragments of a selected target with diverse colored lines on analysis of HRM were sequenced by Tsingke (Hangzhou, China) and the DSDecode (http://skl.scau.edu.cn/dsdecode/) program was used to decode the mutation sequences [55].

Development of Transgene-Free Mutant Lines
T 1 plants were foliar sprayed with 1.0 g·L −1 bentazon (0.1 L·m −2 ) at approximately four-leaf stage according to Lu et al. [49]. Because CYP81A6, which metabolizes and renders tolerance to bentazon was silenced in the transgenic plants, the transgene-free plants will survive the betazone spraying and the transgenic ones will die. No less than 10 survived T 1 plants from every single T 0 plants were further verified for the presence of site-specific mutations and absence of T-DNA. The T-DNA free osipk1 mutants were advanced to T 2 lines and used for further tests.

Assay of Agronomic Traits
Xidao #1 (WT) and its transgene-free osipk1 mutant were grown side by side at the facility of Zhijiang Seed Tec. Ltd., Hangzhou, Zhejiang, China. Sixty seedlings were planted for each line, and 20 randomly chosen inner plants of each plot were assessed for agronomic traits in the field, recorded before and after harvest. Different agronomic parameters, such as 1000-grain weight, seed-set, panicle length, number of tillers per plant, and plant height were evaluated.

Assay of Seed Germination
For a controlled germination test, the rice seeds were immersed in distilled water at 28 • C for two days and germinated on filter paper immersed with ultrapure water for seven days in darkness at 28 • C. The percentage of germination was noted at regular intervals and assayed. The tests were repeated thrice with 100 seeds for each repeat.

Assay of Seed Phosphorus
The contents of inorganic P (Pi) in seeds were examined both quantitatively and qualitatively by following the protocol of Chen et al. [56] with slight modifications. The qualitative assay was utilized for observation of the high inorganic P (HIP). The seeds of rice were placed into 96-well plates, mixed with the solution of 0.4 M HCl at 4 • C for 12 h. Aliquots of supernatant (10 µL) were used for the determination of Pi content by following the method of Larson et al. [9]. The colorless samples characterize contents of parent varieties, while the change of a blue color infers induced content of Pi (HIP) ( Figure 3A).
The levels of Pi in rice seeds were quantitatively measured by following the method of Wilcox et al. [57] with minor modifications in triplicate. The grains of brown rice were pestled into flour and approximately 0.5 g flour was mixed with trichloroacetic acid including 25 mM MgCl 2 , 12.5 % (w/v) by gentle shaking at 4 • C for about 12 h. The supernatants were utilized for the assay of Pi by following the method of Raboy et al. [58] after centrifugation for 15 min at 10,000× g.
The accumulation of PA-P was measured for flour of brown rice by following the protocol of McKie et al. [59] using the commercial assay kit (Megazyme, Ireland) in triplicate. Approximately 1 g grains of brown rice was mixed with 20 mL of 0.66 M hydrochloric acid at room temperature for 2 h and then centrifuged for 20 min at 12,000× g. Immediately, the supernatant (0.5 mL) was neutralized by the addition of the same volume solution of 0.75 M sodium hydroxide. The mixed sample extract was used in the procedure of enzymatic dephosphorylation reaction and the absorbance was measured at 655 nm.
The contents of total phosphorus in rice seeds were measured by following the method as previously described in triplicate [60]. Approximately 0.2 g samples of brown rice were digested with 6 mL HNO 3 for 60 min at 160 • C in a system of microwave digestion (Mars6, USA). The solution was concentrated for about 120 min at 140 • C until 1 mL digested solution was left, and then diluted into 50 mL distilled water [61]. Inductively coupled plasma mass spectrometry (ICP-MS) (PerkinElmer, USA) was used to analyze the total phosphorus in rice seeds.

Stresses Treatment
For the treatment of stresses, 14-day-old rice plants were grown in 1 2 Murashige and Skoog (MS) [62] liquid culture medium, added with 20 mM mannitol or 100 mM NaCl, and then planted for 7 days before collecting the samples. After a 21 d cultivation, rice seedlings were sampled for physiological, biochemical, and molecular assays. The expression levels of genes associated with phytic acid biosynthesis and stress response were measured.

Measurement of Inositol Triphosphate (IP 3 ) and Phytic Acid (IP 6 ) Content
Inositol triphosphate (IP 3 ) was extracted by following the method of Campion et al. [63]. Approximately 2 g sample of fresh rice leaves was ground and blended in 3 mL 1 N HCl/methanol/chloroform (1.33/1/1). The homogenate was kept for 20 min at 4 • C, and then centrifuged at 3000× g for 20 min (4 • C). The organic phase (bottom) was discarded and the aqueous phase (top) was extracted again with a methanol/chloroform (v/v, 1/1). The top phase was dried in a rotary evaporator, and then suspended in 100 µL distilled water for determination of IP 3 content using the commercial assay kit of enzyme-linked immunosorbent assay (ELISA) (Jiangsu Jingmei Biotechnology Co., Ltd., Yancheng, China) [64].
The accumulation of phytic acid (IP 6 ) in rice leaves was extracted and measured by following the protocol as previously described [65]. About 0.3 g of rice leaves were ground and mixed in 10 mL of 0.2 M HCl. The homogenate was kept for 2 h at room temperature, and then centrifuged at 5000× g for 15 min. The supernatant was added with hydrated FeCl 3 and then mixed with 6 mL of 5 M NaOH. After added with HNO 3 , the absorbance of the mixed solution was measured at 510 nm.

Hydrogen Peroxide (H 2 O 2 ) Content Measurement
H 2 O 2 accumulation was determined using the commercial assay kit for H 2 O 2 content (Solarbio, Beijing, China) according to the method described by Zhang et al. [66] and Jiang et al. [67]. Approximately 10 mg of fresh rice leaves were homogenized and mixed with 500 µL acetone followed by centrifugation at 10,000× g for 8 min (4 • C). The supernatant was mixed with an equal volume of detection reagent of H 2 O 2 by vortexing and incubated for 10 min at room temperature. The absorbance at 415 nm was determined, and the standard curve was used to assay the accumulation of H 2 O 2 .

Determination of Malondialdehyde (MDA)
The level of MDA was measured by following the method of Tang et al. [68]. About 100 mg rice leaves were homogenized and mixed with 10 mL trichloroacetic acid (v/v, 10%). The mixture was centrifuged for 15 min at 12,000× g. The reaction mix solution including 2 mL of thiobarbituric acid and 2 mL of extract was heated for 40 min at 95 • C, and then centrifuged once more for 15 min at 12,000× g after rapid cooling on ice. Finally, the absorbances were measured at 450, 532, and 600 nm.

Measurement of Free Proline (Pro)
The content of free proline was measured as previously reported [69]. About 1 g of rice tissue samples were ground with an aqueous solution of 3% (v/v) sulfosalicylic acid. After centrifugation at 37 • C at 12,000× g for 15 min, 2 mL of supernatant was added to the same volumes of glacial acetic acid and acidic ninhydrin. This mixture was heated at 95 • C for 40 min, cooled on ice, and then followed by the addition of 4 mL of toluene to extract the colored reaction product separated from the aqueous phase. The absorbance of the toluene phase at 520 nm was determined and the standard curve of proline was utilized to assay the accumulation of free proline.

Measurement of Anti-Oxidant Enzymes Activities
The activities of anti-oxidant enzymes, peroxidase (POD), catalase (CAT), and superoxide dismutase (SOD), were determined as reported [70][71][72][73] by using the commercial kits (Solarbio, Beijing, China). The anti-oxidant enzyme activities were standardized as follows: one unit of POD activity causes the absorbance to induce 0.01 at 470 nm per min; one unit of SOD activity causes the inhibition of photoreduction of 4-nitro blue tetrazolium chloride by 50%; one unit of CAT activity causes reduction of 1 nmol H 2 O 2 per min.

Analysis of Gene Expression
The RNAprep Pure Plant Kit (Tiangen, Beijing, China) was used to extract the total RNA of fresh rice tissues and HiScript III 1st Strand cDNA Synthesis Kit (Vazyme Biotech Co., Ltd., Nanjing, China) was used to generate cDNA from total RNA by reverse transcription [74]. AceQ qPCR SYBR Green Master Mix (Vazyme, Nanjing, China) was used to perform quantitative real-time PCR (qRT-PCR) and primers used in this study for qRT-PCR are shown in Table S1. The levels of relative gene expression were performed by using the method of 2 −∆∆Ct [75] and the ACTIN gene (Os03g0718100) was utilized as the internal reference.

Statistical Analysis
Three replicates were executed for molecular, physiological, and biochemical assays, and six replicates were conducted for the measurements of dry biomass, plant height, and root length. The experimental data values are shown as the means standard error (SE) based on replications. The statistical analyses were completed by following the students' t-test. The means were compared by analysis of variance (ANOVA), and the significant differences between means of groups were performed using Bonferroni post-tests.

Data Availability Statement:
The data presented in this study are available in the article or supplementary material.