NH787 EMS mutant of rice variety Nagina22 exhibits higher phosphate use efficiency

Rice (Oryza sativa L.), a major dietary source, is often cultivated in soils poor in available inorganic orthophosphate (Pi), which is a key nutrient for growth and development. Poor soils are amended by phosphorus (P) fertilizer, which is derived from the non-renewable rock phosphate reserves. Therefore, there is a need for developing rice varieties with high productivity under low P conditions. At the ICAR-IIRR, ethyl methanesulfonate (EMS) mutagenized rice genotype Nagina22 (N22) were screened for high grain yield in Pi-deprived soil, which led to the identification of ~ 10 gain-of-function mutants including NH787. Here, detailed comparative morphophysiological, biochemical, and molecular analyses of N22 and NH787 were carried out in hydroponics and potting soil under different Pi regimes. Under Pi-deprived condition, compared with N22, NH787 exhibited higher root and vegetative biomass, the number of tillers, and grain yield. The augmented agronomic traits of NH787 were corroborated with significantly higher photosynthetic rate, pollen fertility, stigma receptivity, and the activities of antioxidant enzymes superoxide dismutase (SOD) and catalase (CAT). Further, several genes involved in the maintenance of Pi homeostasis (GPH) were differentially regulated. The study thus revealed a wide-spectrum influence of the mutation in NH787 that contributed towards its higher Pi use efficiency (PUE).


Scientific Reports
| (2021) 11:9156 | https://doi.org/10.1038/s41598-021-88419-w www.nature.com/scientificreports/ also been an attractive approach for functional genomics 22 . A programmable CRISPR/Cas9 system emerged as a promising molecular tool for genome editing 23 and Jennifer Doudna and Emmanuelle Charpentier were awarded the 2020 Nobel prize in Chemistry for developing this versatile technology. CRISPR/Cas9 system is now a favored technology for generating transgene-free rice plants [24][25][26] . A gain-of-function mutagenesis is an alternative approach based on the ectopic overexpression of transgenes under the control of a strong constitutive CaMV35S or ubiquitin promoter 27 . Functional characterization of several GPH by reverse and/or forward genetic approach has thus led to the identification of several key positive and negative regulators of sensing and signaling cascades governing the maintenance of Pi homeostasis 3,6,28 (Table 2). However, plants generated by these forward and reverse genetics approaches are often deemed as a potential transgene and are regulated by stringent countryspecific ethical legislations, and often fail to comply with the biosafety regulations [29][30][31] . Although CRISPR-edited rice was considered to comply with the regulatory approval for commercial applications 24 , recently Court of Justice of the European Union has clubbed them with GM plants [32][33][34] . One of the classical controversial cases is the Golden rice, which was engineered to produce seeds enriched with β-carotene to mitigate vitamin A deficiency in the millions of poor people 35 but has been embroiled in polarized debate over its ethicality 36 . On the contrary, mutation breeding by exposure to mutagens such as EMS or irradiation by X-rays is environmentally benign, has good safety records, and is not regulated worldwide 29 . EMS-induced mutagenesis is an attractive strategy for inducing genetic variations in the genome 37,38 and has facilitated the development of a rich repository of rice mutants that exhibit tolerance to different biotic and/or abiotic stresses 39 . N22 is an upland and short aus genotype and is tolerant to heat and drought 40,41 . An initiative was launched by the Department of Biotechnology, Govt. of India, for generating EMS-mutagenized M2 populations (~ 85,000) in the background of N22 41 . At www.nature.com/scientificreports/ the ICAR-IIRR, efforts are underway for more than a decade to screen N22 EMS mutants that exhibited altered PUE under field condition, which led to the identification of several loss-of-function and gain-of-function mutants [42][43][44][45][46][47] . Among these mutants, detailed morphophysiological and molecular analyses were carried out for the loss-of-function mutant NH101, which revealed several traits that were affected contributing towards its lower PUE compared with N22 47 . However, gain-of-function N22 mutant that shows significantly higher PUE has not been characterized as yet. Therefore, in this study, detailed comparative morphophysiological, biochemical, and molecular analyses of N22 and NH787 mutant were carried out in the hydroponics and potting soil under different Pi regimes. The analysis revealed several traits that contributed towards the higher PUE of the gain-of-function NH787 mutant.

Materials and methods
Plant materials and experimental conditions. Rice (Oryza sativa L. ssp indica) genotype N22 were mutagenized with EMS and several gain-of-function mutants i.e., NH363, NH514, NH686, NH719, NH776, and NH787 were identified, which exhibited high grain yield in Pi-deprived soil under field condition compared with N22 45 . From these gain-of-function mutants, NH787 was selected for detailed morphophysiological, biochemical, and molecular analyses. About 15 seeds each of N22 and NH787 were placed equidistant on a Pteri plate lined with germination paper soaked with deionized distilled water and wrapped in aluminum foil and kept for germination in a growth chamber (28-30 °C) for 4 days. N22 and NH787 seedlings showed significant variation in their radicle length. Therefore, the seedlings were scanned and their radicle was measured by employing the ImageJ program 48 and only those in the range of 2.0-3.0 cm were selected and transferred to the hydroponic system as described 49 containing P+ (0.3 mM NaH 2 PO 4 ) and P− (0 mM NaH 2 PO 4 ) media as described 50 for 7 days. For the potting soil experiment, N22 and NH787 were grown initially under normal soil condition for 14 days. Subsequently, the seedlings were transplanted in earthen pots which were filled with 8 kg of normal soil (P+) and low P soil (P−) with the Olsen P values of 24 kg/ha and 1.8 kg/ha for P+ and P− soil, respectively. P+ and P− soils were fertilized as described 47 . All plant experiments were carried out in accordance with the guidelines and permission was obtained to collect the rice seeds.
Quantitative analysis of the root traits. Seedlings grown in the hydroponic system were removed along with the mesh after 7 days treatment under P+ and P− conditions and placed in an inverted position in a Petri plate containing a pool of water. Under the stereomicroscope, roots were separated at the shoot: hypocotyl junction and transferred to a Petri plate containing 1% (w/v) agar. Adventitious, seminal, and lateral roots were spread gently with a camel hair brush to reveal the root system architecture (RSA). Spread-out roots were then scanned at 1000 dpi using a desktop scanner. Scanned images were then used for documenting the number and length of different root traits by using the ImageJ program 48 . determined by a 50% decrease in the absorbance at 560 nm due to rapid inhibition of O 2 − with NBT. Peroxidase (POD) activity was assayed as described 56 . The reaction mixture comprised 1.0 ml phosphate buffer (100 mM, pH 6.1), 0.5 ml each of guaiacol (96 mM), H 2 O 2 (12 mM), and 0.1 ml of the enzyme extract. The absorbance was taken at 470 nm at different time intervals (0, 1, 2, and 3 min). Catalase (CAT) was assayed as described 57 . The reaction mixture comprised 1.5 ml phosphate buffer (100 mM, pH 7.0), 0.5 ml H 2 O 2 (75 mM), and 0.05 ml of the enzyme extract. A temporal disappearance of H 2 O 2 was recorded at an interval of 30 s for 2 min at 240 nm. Ascorbate peroxidase (APX) activity was assayed as described 58 . The root and shoot tissues were ground in a solution containing 1.5 ml phosphate buffer (100 mM, pH 7.0) containing ascorbic acid (1 mM), and EDTA (0.5 mM). The solution was centrifuged at 12,000 rpm at 4 °C for 20 min and the supernatant was collected for the assay. The reaction mixture comprised 1.5 ml phosphate buffer (100 mM, pH 7.0), 0.1 ml each of EDTA (3.0 mM), H 2 O 2 (3.0 mM), 0.5 ml ascorbic acid (3 mM), and 0.05 ml of the enzyme extract. The APX activity was measured by monitoring the gradual decrease in the absorbance value at an interval of 30 s for 2 min at 290 nm.
Quantification of H 2 O 2 content. H 2 O 2 content was estimated as described 59 . Freshly harvested root and shoot tissues (~ 0.5 g) were ground in 10 ml of trichloroacetic acid, centrifuged at 12,000 rpm at 4 °C for 15 min, and the supernatant was collected for the assay. The reaction mixture comprised 0.5 ml of phosphate buffer (10 mM, pH 7.0), 2 ml of KI (1 M), and 0.5 ml of the supernatant. The reaction mixture was vortexed for 1 min, incubated in dark for 30 min, and H 2 O 2 content was quantified at 390 nm.
Assay for pollen viability. The anthers from the spikelets were collected just before anthesis, crushed in Lugol's (I 2 -KI) solution, and observed under a light stereomicroscope as described 60 . Sterile and fertile pollens were unstained and stained, respectively and their images were captured using a compound microscope (10X).

Quality traits.
Harvested grains were threshed, cleaned, and dried at 45 °C for 3 days to achieve identical moisture content. Grains (~ 25 g) were dehulled using a sheller (Satake Co. Ltd. Japan). The hulling rate was computed as described 61 . Brown rice was milled by employing Pearlest grain polisher (Kett, USA) and the milling rate was calculated as described 61 . The head rice recovery was calculated by weighing polished rice and separating head rice (≥ ¾ length of the brown rice) manually from the broken fractions. Gel consistency (GC) was computed as described 62 . Gelatinization temperature was calculated based on the alkali spread score of the milled rice as described 61 . Amylose content was estimated from the ground rice flour calorimetrically as described 63 . Length, width, and area of grains were measured by using the ImageJ program 48 . qRT-PCR analysis. Total RNA (~ 2 μg) was isolated from the ground tissue using Trizol reagent and treated with RNase-free DNase. First-strand cDNA was synthesized by using oligo (dT)-18 primer and Superscript II Reverse Transcriptase (Invitrogen). OsActin (LOC_Os03g50885) was used as an internal control. The qRT-PCR analysis was performed in triplicate using SYBR Premix Ex TaqII (TaKaRa) in a StepOnePlus Real-time PCR system (Applied Biosystems). Relative expression levels of the genes were computed by the 2 -ΔΔC T method of relative quantification 64 . Gene-specific primers are listed in Supplementary Table S1.
Statistical analysis. Two-way analysis of variance (ANOVA) was performed using open-source software R 65 with agricolae package. Statistical significance of the parameter means was determined by performing Fisher's LSD test.
Ethical approval. The authors declare that the experiments comply with the current laws of the country in which they were performed and in compliance with ethical standards.

Results and discussion
Selection of the uniformly grown seedlings for treatment under different Pi regimes in a hydroponic system. Easy-to-assemble, element-contamination-free, and the aseptic hydroponic system is suitable for documenting the developmental responses of different root traits of the rice seedlings grown under different Pi regimes 45,47,49 . The seed area of N22 and NH787 was documented by employing the ImageJ program 48 . There was no significant variation in the seed area of N22 and NH787 (Fig. S1a). Relatively, the seed area was marginally higher (~ 5%) in N22 EMS mutant NH101 47 . This suggested a variable effect of EMS mutagenesis on the seed area of N22 EMS mutants. Seeds (~ 20) of N22 and NH787 were placed equidistant on a Petri plate lined with a germination paper soaked with deionized distilled water, wrapped in aluminum foil, and maintained in a growth chamber (28-30 °C) for 4 days. Rice seed with radicle length > 0.5 cm was considered germinated 47 . The images of germinated seedlings (~ 200 each of N22 and NH787) spread over 10 Petri plates were captured by using a desktop scanner (Fig. S1b) www.nature.com/scientificreports/ of the germinated seedlings of both N22 and NH787. Earlier studies had suggested selecting only those rice seedlings whose radicle length falls within a fairly comparable size range (~ 2 to 3 cm) for subsequent transfer to a hydroponic system under different Pi regime to circumvent any erroneous interpretations 47,49 . In the model plant Arabidopsis thaliana also, the selection of uniformly grown seedlings with primary root length in the range of ~ 1.5 to 2.5 cm was recommended to minimize the effect of intrinsic variability on the subsequent treatments under different Pi regimes [66][67][68] . Therefore, the radicle length of the germinated N22 and NH787 seedlings was measured by using the ImageJ program 48 and categorized into different groups based on their radicle length (Fig. S1c). The size distribution pattern of N22 and NH787 radicle length is represented by the red (≤ 0.50 cm), black (0.51-2.0 cm), green (2.01-3.00 cm), and yellow (3.01-5.50) histograms, which exhibited a typical Gaussian curve and a noticeable variation between the genotypes. The number of seedlings with radicle length in the size range of 2.01-3.00 cm was significantly higher (49.38%) in NH787 compared with N22 (21.88%). These seedlings were eventually selected for transfer to the hydroponic system containing P+ and P− media and the rest of the seedlings (< 2.01 cm and > 3.00 cm) were discarded (Fig. S1d).
Responses of ontogenetically distinct root traits under different Pi regimes in a hydroponic system. The root system of rice comprises ontogenetically distinct embryonically developed primary and seminal roots that play a key role during the seedling stage and post-embryonically developed adventitious roots constitute the bulk of the functional root system in a mature plant 69,70 . N22 and NH787 seedlings (4-dayold) with radicle length in the size range of 2.01-3.00 cm were transferred to the hydroponic system containing P+ and P− media and grown for 7 d. After the treatment, roots of N22 and NH787 were separated at the shoot: hypocotyl junction and spread gently to reveal the architectural details of the embryonically and post-embryonically developed traits under P+ and P− conditions. Images of the spread-out roots were captured by using a desktop scanner and the ImageJ program 48 (Fig. 2a,b). The number of lateral roots (NLR) was significantly reduced (47.31%) in NH787 compared with N22 under P+ condition (Fig. S2A,B) but was comparable under P− condition (data not shown). Pi deficiency triggered a significant reduction (72.01%) in the total length of the lateral roots (TLLR) on primary, seminal, and adventitious roots of N22 (data not shown) and agreed with earlier studies on N22 47,49 . Relatively, Pi deficiency-mediated reduction of TLLR in NH787 was 48.64%, which was significantly lower compared with N22 (data not shown). This suggested that the effect of Pi deficiency on TLLR was more aggravated on N22 than NH787. Although TLLR of N22 and NH787 was comparable under P+ condition (Fig. S2A), it was significantly higher (43.42%) in the latter compared with the former under P− condition (Fig. 2a,c). In rice, elongation of the seminal root plays a key role in the acquisition of nutrients such as Pi and nitrogen (N) 72 . Therefore, the effect of Pi deficiency was investigated on the number of seminal roots (NSR) and the total length of seminal roots (TLSR) of N22 and NH787. The effect of Pi deficiency was evident on the developmental response of the seminal roots of N22, which revealed significant reductions by 90.03% and 84.68% in their NSR and TLSR, respectively compared with P+ condition (data not shown) and was congruent with studies on N22 47,49 . The effects of Pi deficiency on both NSR and TLSR of NH787 were relatively less aggravated and resulted in reductions by 69.88% and 55.99%, respectively (data not shown). NSR and TLSR of N22 and NH787 were comparable under P+ condition (Fig. S2a). However, under P− condition the NSR and TLSR of NH787 were 2.6-fold and 2.9-fold higher, respectively compared with N22 ( Fig. 2a,d,e). Pi deficiency has also been shown to exert an attenuating influence on the seminal root length of rice varieties O. rufipogon (IRGC 105491) and Curinga 72 . The total length of adventitious roots (TLAR) increased significantly (25.52%) in N22 during Pi deficiency (data not shown) and the result was in agreement with earlier studies on N22 47,49 . Relatively, the increase in TLAR was only 12.29% in Pi-deprived NH787 (data not shown). TLAR of N22 and NH787 was comparable under P+ condition (Fig. 2a) but was significantly higher (65.53%) in the latter compared with the former under P− condition (Fig. 2a,f) 10 . Comparatively, the effect of Pi deprivation was less aggravated on TRL of NH787 and exhibited a 36.74% reduction compared with P+ condition (data not shown). Although TRL of N22 and NH787 was comparable under P+ condition (Fig. S2A), it was significantly higher (65.68%) in the latter compared with the former under P− condition (Fig. 2a-g). Together, the detailed analyses of different root traits revealed that the effects of Pi deficiency were more aggravated on N22 than NH787. Further, the Pearson correlation was presented as a correlogram to determine the relationship across the developmental responses of ontogenetically distinct root traits of N22 and NH787 under different Pi regimes (Fig. S2C, Fig. 2h). Under P+ condition, TLLR was positively and significantly correlated with TRL and NSR in N22, whereas a significant positive correlation was observed between NLR, TLLR, and TRL and NAR and TLAR in NH787 (Fig. S2C). Under P− condition, NSR and TLSR in N22, and TLSR, NAR, TLAR, and TRL in NH787 exhibited a significant positive correlation (Fig. 2h). The analysis revealed that NSR, TLSR, and TLLR were positively and significantly correlated with TRL in both Pi-deprived N22 and NH787. Pearson correlation has also been used in earlier studies on various morpho-biochemical traits at various developmental stages of N22 and its EMS mutants under different Pi regimes 43

Effects of different Pi regimes on various morpho-agronomic traits of N22 and NH787 grown to maturity in potting soil.
Growth performance and morpho-agronomic traits of N22 and NH787 plants grown to maturity (50% flowering) in Pi-replete (P+) and low Pi (P−) potting soil were determined (Fig. S3,  Fig. 3 ). Whereas, during Pi deficiency, the unfilled spikelets/panicle was significantly higher in N22 (52.74%) but was comparable in NH787 with P+ condition (data not shown). It was evident from the analysis that the effects of Pi deprivation were relatively more aggravated in N22 than NH787. Earlier studies also reported the inhibitory effects of Pi deficiency on various morpho-agronomic traits of N22 and its EMS mutants [42][43][44]46,47,71 . Under P+ condition, there was no apparent difference in the phenotype of N22 and NH787 (Fig. S3A). However, the phenotype of the root, panicles, and grain was more robust in NH787 compared with N22 ( Fig. S3B-D). This was reflected in significantly higher root biomass (41.86%), vegetative biomass (19.70%), filled spikelets/panicle (89.79%), 100-grain weight (15.00%), and yield (36.96%) of NH787 compared with N22 ( Fig. S3E-G,I,J). On the contrary, unfilled spikelets/panicle was significantly higher (61.64%) in N22 compared with NH787 (Fig. S3H). However, under P+ condition the root length and number of tillers in N22 and NH787 were comparable (data not shown). Further, Pearson analysis revealed a positive and significant correlation of yield with root and vegetative   (Fig. S3K). Under P− condition, the phenotype of the plant, root, panicles, and grain were more robust in NH787 than N22 (Fig. 3a-d). The phenotypic observation was substantiated with significantly higher root biomass (9.2 folds), root length (12.21%), vegetative biomass (3.9 folds), number of tillers (2.37 folds), filled spikelets/panicle (3.15 www.nature.com/scientificreports/ folds), 100-grain weight (40.00%), and yield (6.42 folds) of NH787 compared with N22 ( Fig. 3e-i,k,i). However, unfilled spikelets/panicle was significantly higher (73.99%) in N22 than NH787 (Fig. 3j). Similar to P+ condition, under P− condition also Pearson analysis showed a positive and significant correlation of yield with root and vegetative biomass, number of tillers, and filled spikelets/panicle in both N22 and NH787 (Fig. 3h).
Photosynthetic and chlorophyll fluorescence traits of N22 and NH787 grown to maturity in potting soil. Pi deficiency adversely affects photosynthetic and chlorophyll fluorescence traits in rice 73,74 .
Pi, Apase and the enzymes involved in ROS scavenging of N22 and NH787 grown to maturity in potting soil. Pi deficiency exerts an attenuating effect on the concentration of Pi, while its effect is augmenting on the activities of Apase and ROS scavenging pathway (APX, CAT, H 2 O 2 , POD, and SOD) in the root and shoot of rice 47 . Therefore, the concentration of Pi and the activities of Apase and ROS scavenging enzymes were assayed in N22 and NH787 grown to maturity under different Pi regimes (Fig. S5, Fig. 5 (Fig. S5A, Fig. 5a (Fig. S5b, Fig. 5b  ) in NH787 compared with N22 ( Fig. 5c-g). The results highlighted differential effects on ROS-mediated redox signaling and oxidative stress in NH787 compared with N22 under different Pi regimes. Earlier studies also showed the Pi-dependent differential effects on ROS homeostasis in the EMS mutants of N22 43,47 .
Reproductive traits of N22 and NH787 grown to maturity in potting soil. In earlier studies, significant inhibitory effects of Pi deficiency were observed on the yield potential of N22 and its EMS mutants [42][43][44][45][46][47] . Therefore, the effects of Pi deprivation were investigated on the male reproductive traits of N22 and NH787 grown under different Pi regimes in a potting soil up to maturity (Fig. 6). I 2 -KI staining was used for determining the viability of pollen collected after anthesis from N22 and NH787 grown under P+ and P− conditions (Fig. 6ad). Pollen viability was significantly higher (P+ [14.73%], P− [23.76%]) in NH787 compared with N22, which suggested that Pi deficiency-mediated effect on pollen viability was more aggravated in the latter than the former (Fig. 6a-E) (Fig. 6f-h). On the contrary, the activity of CAT in NH787 was significantly lower and higher under P+ (19.05%) and P− (47.05%) conditions, respectively compared with N22 (Fig. 6I). The analysis revealed differential effects on antioxidant enzyme activities of N22 and NH787 under different Pi regimes. Under Pi-deprived condition, NH787 exhibited significantly higher CAT activity in the anthers compared with N22, which triggered the efficient conversion of H 2 O 2 into water and oxygen possibly favoring higher pollen fertility and yield (Fig. S6). Further, the effects of Pi deficiency were determined on various grain parameters of N22 and NH787 grown in P+ and P− potting soil up to maturity (Fig. 7). Single grain weight of N22 and NH787 collected from P+ and P− plants were categorized into a week (6-10 mg and 10-15 mg) and robust (16-20 mg, www.nature.com/scientificreports/ 21-25 mg, and 26-30 mg) categories and the frequency of weight distribution pattern in these categories was computed, which revealed a typical Gaussian curve (Fig. 7a). Under P+ condition, the frequency of single grain weight under different categories was comparable between N22 and NH787 (Fig. 7a). However, significant variation in the frequency of single grain weight was observed under P− condition for N22 compared with NH787 ranging from higher (6-10 mg and 10-15 mg), lower (16-20 mg and 21-25 mg), and comparable (26-30 mg) values (Fig. 7a). The analysis revealed that NH787 seeds were comparatively more robust than N22 when grown under Pi-deprived condition. Under both P+ and P− conditions, several grain quality parameters of NH787 were significantly higher than N22 comprising hulling (P+  (Fig. 7h). Different parameters such as alkali spread value, gelatinization temperature, and gel consistency are conventionally used for determining the gain quality traits of rice [75][76][77] . Therefore, alkali spread value, gelatinization temperature, and gel consistency were assayed for P+ and P− seeds of N22 and NH787, which revealed that these traits were superior in the latter compared with the former (Table S2). Although α-amylase activity in the spikelets of N22 and NH787 was comparable under P+ condition, it was significantly lower in the latter compared with the former under P− condition (Table S3). Low α-amylase activity in the rice spikelets has been correlated with the grain weight and yield 78,79 . Pearson correlation analysis was carried out to determine the relationship across agronomical and quality traits (yield, TN, GA, GW, GL, AC, HRR, milling, and hulling), pollen fertility (PF), and the activities of antioxidant enzymes (CAT, APX, POD, and SOD) and α-amylase (AA) in the anther and spikelets of N22 and NH787 under different Pi regimes (Fig. 8). Under P+ condition, a positive and significant correlation in N22 was observed with yield, TN, AA, and AC and that of NH787 with yield, TN, PF, AC, and HRR. Whereas, under P− condition, a positive and significant correlation in N22 was detected with yield, TN, APX, AA, AC, and hulling and that of NH787 with yield, TN, PF, POD, APX, GA, GW, AC, and milling. The analysis revealed that TN, AC, AA, POD, APX, and PF were correlated positively and significantly with the yield of N22 and NH787 under different Pi regimes. www.nature.com/scientificreports/ Relative expression levels of GPH in N22 and NH787 grown to maturity in potting soil. The qRT-PCR assay was employed to decipher Pi deficiency-mediated effects on the relative expression levels of functionally diverse GPH in the roots of N22 and NH787 grown to maturity in potting soil under P+ and P− conditions (Fig. 9). For this experiment, only those GPH were selected, which had been functionally characterized either by overexpression under the constitutive promoter or by mutation (T-DNA, Tos17, RNAi, or CRISPR-cas9) and implicated in their tissue-specific key roles in the sensing and signaling cascades governing the maintenance of Pi homeostasis under Pi regimes (  (Fig. 9a). Higher relative expression levels of these GPHs in NH787 could be assumed to play a pivotal role in contributing towards its higher yield potential (Fig. 3). On the contrary, the relative expression levels of several GPH were significantly reduced in Pi-deprived roots of NH787 compared with N22 that are involved in the uptake of Pi by high-affinity Pi transporter (OsPht1;10 87 ), utilization of extracellular organic P (OsPAP10a 90 ), regulation of the growth during Pi deficiency via a negative feedback loop and by interacting with OsPHR2 in a Pi-dependent manner (OsSPX1 83,91 ), regulation of Pi starvation signal transduction (OsmiRNA399a 92 ), growth, development, and maintenance of Pi homeostasis (OsLPR5 93 ), regulation of Pi starvation responses (OsPHO2 94 ), and post-translational SUMOylation of proteins (OsSIZ1 95 ) (Fig. 9b). In Arabidopsis thaliana, electrophoretic mobility shift assay revealed the binding of the transcription factor PHR1 as a dimer to an imperfect palindromic 8-bp sequence (5′-GNATATNC-3′) named as PHR1 binding sequence (P1BS) found in the promoter (2 kb Figure 8. Correlogram of the agronomical and quality traits, pollen fertility, and the activities of antioxidant enzymes and α-amylase in the anther and spikelets, respectively in N22 and NH787 grown in a potting soil up to 50% flowering under P+ and P− conditions. The scale represents Pearson correlation values with brownish and bluish shades indicate positive and negative correlation, respectively. TN Tiller number, PF pollen fertility, GA grain area, GW grain width, GL grain length, AA α-amylase, AC amylose content, HRR head rice recovery. www.nature.com/scientificreports/ upstream of ATG start codon) of several genes involved in Pi deficiency-mediated responses 96,97 . Therefore, the P1BS (GNATATNC) motif was analyzed in the promoter (3 kb upstream of ATG initiation site) of the 17 GPH revealing its presence in 14 of them, which suggested their potential regulation by OsPHR2 (Table 2). In this context, significantly higher relative expression of OsPHR2 in Pi-deprived roots of NH787 compared with N22 ( Fig. 9a) suggested its potential regulatory influence on the expression of several GPH that play a key role in the maintenance of Pi homeostasis under different Pi regimes.

Conclusions
The results provided empirical evidence towards the differential effects of the EMS mutagenesis on various morphophysiological, biochemical, and molecular traits of NH787 that conferred higher PUE under low Pi soil condition compared with N22. A schematic diagram is presented highlighting the efficacy of the EMS mutants for screening the rice variety NH787 with higher PUE in an environment-friendly manner for sustainable production (Fig. 10). NH787 is now used as a donor in breeding programs for developing low P tolerant varieties with superior grain quality and is also being evaluated in larger plots at multiple locations with variable agroclimatic conditions. Efforts are also underway to identify the candidate genes in NH787 responsible for higher PUE by employing quantitative trait loci (QTL) mapping and MutMap approach in the F2 populations revealing a discernible phenotype 98 . MutMap approach has been used in an earlier study for identifying the candidate genes conferring salt tolerance in the F2 populations of EMS mutant hitomebore salt tolerant 1 (hst1) of rice variety Hitomebore 99 .

Data availability
All data generated or analyzed during this study are included in this published article (and its Supplementary Information files). The sequence data is available on request.
Received: 12 February 2021; Accepted: 7 April 2021 Figure 10. A schematic representation highlighting the efficacy of the EMS mutants for screening the rice variety NH787 with higher PUE in an environment-friendly manner for sustainable production.