Field identif ication of morphological and physiological traits in two special mutants with strong tolerance and high sensitivity to drought stress in upland rice (Oryza sativa L.)

HUANG Min, XU Yu-hui, WANG Hua-qi

College of Agronomy and Biotechnology, China Agricultural University, Beijing 100193, P.R.China

Abstract The two mutants idr1-1 and 297-28, which were obtained from the radiation mutation of HD297 and IAPAR9, were used as experimental materials in this study for a 2-year (2012 and 2013) experiment about f ield drought resistance identif ication in Beijing, China. Key agronomic traits and water-related physiological indexes were observed and measured, including the leaf anti-dead level (LADL), days to heading, plant height, setting percentage, aboveground biomass, leaf water potential (LWP), net photosynthetic rate (P n) and transpiration rate. The results showed that the mutant idr1-1 that was under drought stress (DS) conditions for 2 years had the highest LADL grades (1.3 and 2.0) among all the materials, and they were 2-3 grades stronger than the wild-type IAPAR9 with an average that was 21.4% higher for the setting percentage than the wild type. Compared with the IAPAR9 for the 2-year average delay in the days to heading and the reduction rates in the plant height, setting percentage, and aboveground biomass under DS compared with the well-watered (WW) treatment, idr1-1 showed 3.2% less delay and 19.1, 16.4, and 6.1% less reduction, respectively. The idr1-1 in the LWP always exhibited the highest performance among all the materials. The P n of idr1-1 under severe and mild DS comparing with that under WW was slightly decreased and even slightly increased, respectively, leading to an average reduction rate of only 0.92%, which was 26.93% less than that of IAPAR9. Under the severe DS, idr1-1 still showed the highest value of 16.88 μmol CO2 m-2 s-1 among all the materials and was signif icantly higher than that of IAPAR9 (11.66 μmol CO2 m-2 s-1). Furthermore, only idr1-1 had the increased and the highest transpiration rate values (7.6 and 6.04 mmol H2O m-2 s-1) under both mild and severe DS compared with the values under WW, when the transpiration rate of all the other materials signif icantly decreased. By contrast, the 297-28 in terms of the LADL grade under DS was the lowest (7.0), and it was four grades weaker than its wildtype HD297 and even one grade weaker than the drought-sensitive paddy rice SN265. For the 2-year average reduction rates in aboveground biomass and plant heights under DS compared with those under the WW, 297-28 was 31.6 and 31.8% higher than HD297, respectively. Meanwhile, 297-28 showed the worst performance for the LWP, P n, and transpiration rate. These results suggest that idr1-1 might be a superior drought tolerant mutant of upland rice found in China. It has a strong ability to maintain and even enhance leaf transpiration while maintaining a high plant water potential under DS, thus supporting a high P n and alleviating the delay in agronomic trait development and yield loss effectively. 297-28 is a much more highly drought-sensitive mutant that is even more sensitive than paddy rice varieties.The two mutants could be used as drought tolerance controls for rice germplasm identif ication and the drought resistant mechanism studies in the future. idr1-1 is also suitable for breeding drought-tolerant and lodging-resistant high-yield rice varieties.

Keywords: rice mutants, morphological and physiological traits, drought resistance, field identif ication

1. lntroduction

Rice is one of the most important grain crops in Asia, and it provides approximately one-third of the caloric intake in the region. According to the FAO, 140.4 million hectares of harvested area produced 667.9 million tons of rice in 2016 (FAO 2017). However, a water shortage has been an important restriction for rice production in many rainfed areas of the world (Pandey et al. 2007). In China, the agriculture industry consumes approximately 70.4% of the water resources, and the proportion of that water used by rice production is nearly 70% (Zhang 2007). Most rice production in China depends on irrigation; however, a less-than-abundant water supply still presents a long-term threat. Water shortages will probably remain a challenge to the sustainable production of rice based on recent climate predictions (Wassmann et al. 2009), which suggest that the intensity and frequency of drought will be aggravated.

The breeding of drought-tolerant cultivars is an effective and economical way to address this problem, and some of these cultivars have been released, including Morobérékan, IRAT109, B96-3, and IAPAR9. However, the conventional breeding progress towards improved drought tolerance in rice cultivars has been relatively slow (Cooper et al. 1999; Wade et al. 1999; Bernier et al. 2008). One of the constraints is the narrow genetic basis because of the germplasm shortage. The use of radiation-induced mutation is an effective way to enhance the genetic variability in many species (Ahloowalia and Maluszynski 2001; Mostafa et al. 2015; Mondal et al. 2017). In rice, many novel germplasm resources were generated by gamma irradiation. In a functional genomics study, Kong et al. (2014) reported the morphological and physicochemical properties of two starch rice mutants, thereby supporting new genotypes for food processing, and they identif ied defective genes. Aromatic rice with high iron was also obtained by gamma radiation, and this germplasm resource was used to develop a new cultivar (Tran and Ho 2017). In addition, the salt resistance of gamma-induced rice was enhanced using a proper radiation dosage (Shereen et al. 2009). Droughtresistance properties have been demonstrated in sorghum (Sihono 2010; Human et al. 2012), soybeans (Moussa 2011), and sugar beets (Sen and Alikamanoglu 2012), but few studies have focused on rice. In China, many donor parents responsible for the drought tolerance of the released cultivars were derived from IRAT109, IAPAR9, or some landraces. Additionally, material def iciencies have limited gene cloning for drought resistance in rice. For example, the donor parents used to construct populations by positional cloning were mostly derived from CT9993, Swarna, Azucena, Apo, and some other landraces (Lanceras et al. 2004; Venuprasad et al. 2009).

In the present study, we selected two pure lines of mutants derived from gamma-induced IAPAR9 and HD297 to identify their drought responses. Morphological traits, such as the leaf anti-dead level (LADL), days to heading, plant height, setting percentage and aboveground biomass, and physiological traits, such as the leaf water potential (LWP), net photosynthetic rate (Pn), and transpiration rate, were used to evaluate the drought tolerance under wellwatered (WW) and drought stress (DS) conditions. One mutant called idr1-1 conferred higher drought tolerance characteristics than IAPAR9, the wild type, whereas 297-28 showed greater susceptibility to drought than the HD297 wild type. These two mutants may provide superior germplasm for use in developing drought-resistant varieties and for studying their molecular mechanisms.

2. Materials and methods

2.1. Materials for the study

Six genotypes were used in this study: HD297 (Oryza sativa L. ssp. japonica), 297-28, IAPAR9 (O. sativa L. ssp. indicajaponica intermediate type), idr1-1, SN265 (O. sativa L. ssp. japonica), and IRAT109 (O. sativa L. ssp. japonica). HD297 is an upland rice variety that is planted widely in North China. As a standard control for the national upland rice regional trials (NURRT), HD297 exhibits a good drought tolerance phenotype. The 297-28 is a mutant induced from HD297 by gamma-ray radiation. This mutant shows thinner stems, leaves, and roots and more tillers than HD297 under WW conditions, whereas under drought stress, the leaves wilt easily, leading to sharp decreases in the yield. IAPAR9 is a drought-tolerant upland rice introduced by the Brazilian Agricultural Research Corporation (EMBRAPA) in 1992. The idr1-1 is a semi-dwarf mutant derived from IAPAR9 by γ-ray radiation. It shows stronger drought resistance than the wild type. SN265 is a super high-yield rice variety with low drought resistance in North China, and IRAT109 is a famous international upland rice variety with strong drought resistance from the Cote d'Ivoire; they were the other two controls used in this study.

2.2. Methods

Planting methodsThe f ield experiments for the above six genotypes were conducted in the summer and autumn of 2012 and 2013 at Shangzhuang Experimental Station of China Agricultural University, Beijing, China (40°08′13.4′′N, 116°11′06.6′′E). The study f ield is sandy soil. Two treatments were applied, namely, DS and WW, with three replicates for each treatment. A randomized complete block design (RCBD) was applied to all the experimental plots. The commonly used local sowing method, direct sowing, was used to sow the rice; the applied base fertilizer consisted of 48 kg N ha-1, 120 kg P2O5ha-1, 100 kg K2O ha-1, 22.5 kg ZnSO4ha-1, and 30 kg FeSO4ha-1, and a top dressing of 45 kg N ha-1urea was applied at the tillering stage. In 2012, the plot areas of the WW and DS treatments were 4.5 m2(six rows per plot, plot size 2.5 m×1.8 m) and 3.0 m2(four rows per plot, plot size 2.5 m×1.2 m), respectively. In 2013, the plot areas of the WW and DS treatments were 10.92 m2(seven rows per plot, plot size 5.2 m×2.1 m) and 15.75 m2(seven rows per plot, plot size 7.5 m×2.1 m), respectively. Chemical and manual methods were used for weed control, and local and conventional management practices for rice cultivation were followed.

Drought treatmentFor the WW treatment, irrigation was applied for the entire growth period to maintain a shallow layer of water on the surface and a wet state in the experimental f ield. For the DS treatment, natural drought was applied, in which time full irrigation was applied after sowing to ensure that all the tested materials had normal germinations, and then, from seedling emergence until the end of the experiment, the plant growth was completely dependent on natural precipitation without any irrigation. At the study site in Beijing, the periods from May to June (from the seedling to tillering stages) and from mid-August to early October (from the heading to maturing stages) occur mostly in the dry season when they receive scarce precipitation, and drought may occur between July and early August (from jointing to booting) during some years. During this time without precipitation, the soil moisture sensor TRIME-PICO TDR (IMKO, Ettlingen, Germany) was used to measure the soil water content (%, v/v) at the 0-20 cm layers of the soil surface every 3-5 days. According to our observations over many years, when the topsoil (depth 20 cm) is saturated, the soil volumetric water content is generally above 40 or 35%. When it is lower than 20%, the ground begins to crack, and some of the top rice leaves begin to die. When the soil volumetric water content is lower than 15%, the ground cracks become obvious, and most leaves show complete or partial death. When this content is less than 10%, the whole plant may die. Therefore, in general, volumetric water content ranges of 30-25%, 25-20%, 20-15%, 15-10%, and ＜10% in the topsoil (depth 20 cm) indicate soil drought stress as follows: “beginning (slight)”, “mild”, “moderate”, “severe”, and “extreme severe”, respectively. When the soil volumetric water content drops below approximately 25%, “mild” drought has begun in the soil. After that, the rice plants were subjected to 2 weeks or more of drought treatment, and their related traits were investigated.

Trait investigationThe LWP and leaf photosynthesis parameters (LPPs) under diurnal change and the key time points were investigated. The LPPs, which include two major indexes, the Pnand the transpiration rate, were measured after 2 weeks of continuous drought during the heading and grain-f illing stages (～80-100 days after sowing, or DAS). The LPPs were measured using an LI-6400XT Portable Photosynthesis System (Li-Cor Environmental, Lincoln, NE, USA) with the following parameters: a reference cell CO2(Ca) concentration of 380-400 μmol CO2mol-1, a f low of 400 μmol s-1and an instant photosynthetic photon f lux density (PPFD) under natural light. The LWP was determined with a pressure chamber instrument, the Pressure Chamber PMS600 (PMS instruments, Albany, OR, USA). The sampling method was as follows: A plot showing representative growth of rice materials was selected for measurement, and plants with robust growth and the same morphological conditions were selected as candidate plants before the measurement. Candidate plants were marked for further measurements. For each rice material, a total of 3-5 plants were measured, and the measurements were conducted on the uppermost f irst fully expanded leaf for each plant. The daily change monitoring for the LPPs and LWP was performed at four time slots on the same sunny day, from 07:00-08:00 h, 09:00-10:00 h, 14:00-15:00 h, and 17:00-18:00 h. The key time-point measurement was only conducted on sunny days at 09:00-11:00 h (for LPPs) and at 14:00-16:00 h (for LWP); measurements for the WW and DS treatments were conducted on two adjacent days (e.g., measurement for WW was only on the 1st day and for DS, it was only on the 2nd day), and they were repeated three times (taking 6 days).

The rice leaves become wilted or even died under drought stress. The leaf death performance among different genotypes ref lects the strength of the drought resistance. The LADL was evaluated by taking the ratio of the green leaf length to the total leaf length of each tested leaf by measurement (Zu et al. 2017) or visualization (Xu et al. 2018). The grading standards for the LADL in our lab was on a 1-9 resistance rating scale (1=highly resistant, 3=moderately resistant, 5=moderately sensitive, 7=sensitive, and 9=highly sensitive) (Xu et al. 2018). In this study, when the LPPs and LWP (～100 DAS) were measured, the LADL score was investigated by visual estimation for the DS treatment.

The days to heading parameter was also recorded. Before each harvest, sampling and seed investigations were performed. In each plot, one row of plants with uniform growth was selected, and a 1-m-long sample segment was accurately measured for sampling. The plants in this sample segment were sampled by removal, along with the roots, and 10 plants were randomly selected to measure the plant height, setting percentage, aboveground biomass, and other indexes. The measurement standards are shown in Table 1.

Statistical analysis of the study dataEach trait index was presented as the mean±SD, and they were statistically analyzed using SAS 9.1.3 (SAS Ins., Cary, NC, USA). The SAS general linear model (GLM) procedure was used to analyze the signif icance of the variations between the genotypes, between the treatments, and between the replicates (＊, P＜0.05;＊＊, P＜0.01). The DUNCAN method was used for multiple comparisons between varieties. The SAS TTEST procedure and Student's t-test (＊, P＜0.05;＊＊, P＜0.01) were used for comparisons between the WW and DS treatments.

3. Results

3.1. Droughtstress degrees and effects

The drought occurrence degree in the tillage layer of the soil at the heading and grain-f illing stages is shown in Fig. 1. In 2012, the soil had a high water content due to abundant rainfall, and yet a rain-free period occurred from 100- 115 days after sowing (DAS). During this period, the volumetric water content in the soil surface layer ranged between 15 and 25%, and a low value of 15% was maintained for approximately 10 days, thereby causing a mild to moderate drought on the whole. In 2013, no effective precipitation occurred in 80-120 DAS, and the volumetric water content in the soil surface layer was always below 15% (indicating moderate drought) and mostly f luctuated around 10% (indicating severe drought), thereby indicating a generally severe drought. Those 2 years provided an ideal natural drought environment for this study. Herein, 2012 is also referred to as a mild drought and 2013 as a severe drought.

3.2. Analyzing the agronomic traits of the mutants

The results for the joint analysis of variance for the effects under the DS and WW treatments are shown in Table 2. The signif icant difference tests between the DS and WW treatments showed that in 2013, the differences in all the traits between the DS treatment and the WW treatment reached the extremely signif icant level or the signif icant level. Specif ically, under the DS treatment, the aboveground biomass, plant height, setting percentage, and LADL decreased extremely signif icant, and the days to heading were signif icantly delayed. In 2012, except for the setting percentage, the other traits under DS treatment were also signif icantly much weaker or signif icantly weaker than those under the WW treatment, with the same trend being observed in 2013. Only the setting percentage did not decrease, but instead it showed an increase, which occurred primarily because during the heading stage under the WW treatment in 2012, 297-28 had lodging problems, whereas the IRAT109 experienced selective damage from rice planthoppers. These problems led to abnormal (decreased) setting percentages in both genotypes under the WW treatment, with setting percentage values that were signif icantly even lower than those under the DS treatment. Compared with the observations in 2012, the degree of difference in each of the traits between the DS and WW treatments was signif icantly higher in 2013. The results for the 2 years showed that compared with the traits under the WW treatment, of all the traits, the aboveground biomass and LADL showed the most sensitivity in responding to drought under the DS treatment, with the reduction rates for both traits being the greatest. Specif ically, the aboveground biomass showed reduction rates of 33.2 and 44.0% in 2012 and 2013, respectively; the LADL showed reduction rates of 34.4% (3.1/9) and 38.9% (3.5/9) in 2012 and 2013, respectively. The days to heading were relatively stable, and they were delayed by 4.3 and 6.4% in 2012 and 2013, respectively. This result can be explained by the fact that during the 2-year experiment (especially in 2013), the f ield drought primarily occurred during the middle and late stages of plant growth and development, which led to damaged functional leaves in the middle and upper parts and directly reduced the biomass and grain yield, whereas the drought before the heading stage was relatively mild and thus had a small impact on the days to heading.

Table 1 The list of measurement methods for the agronomic traits

Fig. 1 Soil water moisture content in this study. DS, drought stress. LPPs, leaf photosynthesis parameters; LWP, leaf water potential.

Table 2 shows that a signif icant difference is present in each of the various traits between the genotypes under the WW and DS treatments during the 2 years. Hence, a comparative analysis of the pros and cons among the genotypes for each trait is possible (Table 3).

Leaf anti-d ead level (LADL, Table 3)Under the DS treatment, the idr1-1 had the lowest 2-year average LADL value (1.7), and 297-28 had the highest value (7.0). The LADL value of each genotype remained basically the same over the 2 years, and the differences between the genotypes reached a signif icant level. The LADL value of idr1-1 (1.3 and 2.0) was signif icantly lower than that of IAPAR9 (4.0 and 5.0), with a decrease by two to three levels. By contrast, 297-28 exhibited an LADL value of 7.0 in both years, which was signif icantly higher than that of HD297 (2.3 and 3.0) with an increase of four levels; the LADL value of 297-28 was even one level higher than that of the rice variety SN265, exhibiting the characteristics of higher drought sensitivity than common rice varieties.

Days to heading (Table 3)Although drought primarily occurred during the grain-f illing and maturing stages, the days to heading for each of the genotype was nevertheless delayed. Compared with that under WW treatment, the days to heading under the DS treatment showed a delay. The degree of delay in days to heading in 2013 was signif icantly greater than it was in 2012 for all the genotypes (except 297-28). For the 2-year average delay degree, the result for the idr1-1 (4.0%) was signif icantly less than that of IAPAR9 (7.2%), with a decrease of 3.2%; it was also less than the delay for IRAT109 (6.2%). The HD297 genotype showed the lowest delay in the average days to heading, suggesting a high drought resistance ability, but the days to heading delay of 297-28 was quite similar to that of the wild type because it has a longer vegetative period and luxuriant vegetative growth under WW treatment. SN265 (8.8%) had a signif icantly higher number of delayed average days to heading than the other genotypes. The results showed that after drought stress treatment, the delay degree in the growth and development for drought-sensitive genotypes was generally greater than that for drought-tolerant genotypes.

Plant height (Table 3)Under the WW environment, the plant height of idr1-1 (57.7 cm) was signif icantly lower than that of IAPAR9 (104.0 cm), indicating that it was a dwarf mutant. Compared with that under the WW treatment, in 2013, the plant height of idr1-1 under the DS treatment decreased signif icantly, but the reduction rate was signif icantly lower than that of IAPAR9. By contrast, the plant height reduction rate for 297-28 was much higher than that of HD297 and it was close to that of SN265, which had the highest reduction rate. In 2012, except for IRAT109, idr1-1 and HD297, which did not have signif icantly decreased plant heights, all the genotypes had decreased plant heights at the extremely signif icant level or at the signif icant level. For the 2-year average of plant height decreases, the reduction rate in the plant height for idr1-1 was 19.1% less than that of IAPAR9, and for 297-28, it was 31.8% more than that of HD297.

Table 2 Comparison of agronomic traits between the well-watered (WW) and drought stress (DS) treatments using a joint analysis of variance

Table 3 Analysis of the agronomic traits of the genotypes tested in this study1)

Setting percentage (Table 3)In 2013, except for 297-28, which had a lower setting percentage under the WW treatment than under the DS treatment, all the genotypes had a signif icantly lower setting percentage under the DS treatment than under the WW treatment. The reduction rate of the setting percentage for idr1-1 reached an extremely signif icant level but was less than half of the reduction rate for IAPAR9. SN265 showed the greatest reduction in the setting percentage; its reduction rate was 60.0% more than that of IRAT109. In 2012, except for 297-28 and IRAT109, all the genotypes experienced a reduction in the setting percentage to varying degrees under DS compared to WW, but the reduction was not signif icant; idr1-1 even had a slightly increased setting percentage, which indicated that the mild drought stress did not signif icantly affect the setting percentage. Based on the average from 2-year data, idr1-1 showed a reduced setting percentage, which was 16.4% less than that of IAPAR9. During both years, as the drought stress intensity increased, the setting percentages for the genotypes decreased by various degrees. Under the DS treatment, the setting percentage of SN265 decreased from 88.5% in 2012 to 33.1% in 2013; it was the largest reduction observed (up to 62.6%) among all the genotypes. IAPAR9 also showed a signif icant reduction (21.0%). By contrast, idr1-1 showed only a marginal decrease (by 2.5%), which was close to that observed for IRAT109 (slight decrease of 1.3%). The setting percentage of 297-28 decreased slightly, but that of HD297 increased instead. These results indicate that there are signif icant differences between the genotypes in the setting percentage traits in terms of the ability to adapt to drought.

Aboveground biomass (Table 3)During both years of this experiment, all the genotypes showed decreased aboveground biomasses to varying degrees under the DS treatment compared with those under the WW treatments, and these decreases were extremely signif icant in 2013. When the rate of reduction was averaged for the 2 years, HD297 had the smallest average reduction rate (18.0%), and SN265 showed the largest (52.2%). For idr1-1, its 2-year average reduction rate was 19.9%, which was 6.1% lower than that of IAPAR9 (26.0%) and 25.0% lower than that of IRAT109. For 297-28, the 2-year average reduction rate was 49.6%, which was 31.6% higher than that of HD297 and close to that of SN265. These results suggest that the drought resistance capacity of idr1-1 was slightly stronger than that of IAPAR9 and signif icantly stronger than that of IRAT109, whereas the capacity of 297-28 was signif icantly weaker than that of HD297 and comparable to that of SN265.

3.3. Leaf water potential and photosynthetic physiology of mutants

Daily changes in the leaf water potential and P n(1) Daily change in the leaf water potential (LWP). The level of LWP (MPa) directly ref lects the water content status of the plant body. The lower the value, the less free water there is in the plant body, which implies more serious drought stress. As shown in Figs. 2 and 3, under the DS and WW treatments, the highest LWP for each genotype occurred before sunrise, during which the differences among the genotypes were the smallest. After sunrise, with the increased light intensity and temperature, the LWPs rapidly decreased from 07:00-08:00 h to 09:00-10:00 h, during which the differences among the genotypes began to occur. The LWPs reached the lowest for all the genotypes at the 14:00-15:00 h, during which the difference among the genotypes was the greatest. Later, the LWPs gradually rose back from the 17:00-18:00 h to sunset, during which the differences among the genotypes experienced a decrease. At each time point, the LWP under the DS treatment was signif icantly lower than that under the WW treatment for the same genotype. The difference in LWPs among different genotypes under the DS treatment was signif icantly higher than that under the WW treatment; the difference reached the largest at the 14:00-15:00 h, and there were still signif icant differences among the genotypes at the 17:00-18:00 h under the DS treatment.

Fig. 2 Daily change in light intensity and temperature. PPFD, photosynthetic photon f lux density.

Fig. 3 Daily changes in the leaf water potential (LWP). The measurement day for the well-watered (WW) treatment was 08-22-2013, and for the drought stress (DS) treatment, it was 08-23-2013. The thick solid line represents a genotype under the DS treatment, whereas the thin dashed line represents a genotype under the WW treatment.

For idr1-1, the LWPs for the entire day under both DS and WW treatments were higher than those of IAPAR9. Unlike the results under the WW treatment, the differences between the two genotypes under the DS treatment were obviously widened starting from early in the morning, reaching a maximum at the 09:00-10:00 h, and it was still present at the 17:00-18:00 h. By contrast, the difference between the two genotypes under the WW treatment only begin to appear after the 09:00-10:00 h, reaching its maximum until 14:00-15:00 h, and it converged at 17:00-18:00 h. Hence, the difference between the two genotypes under the DS treatment appeared signif icantly earlier and lasted longer than under the WW treatment. For 297-28, its LWPs over the entire day under both DS and WW treatments were lower than those of HD297. Signif icantly different from the WW treatment, the gap between 297-28 and HD297 under the DS treatment was signif icantly widened during most of the day from the 09:00-10:00 h to the 17:00-18:00 h; it reached its maximum at the 14:00-15:00 h, and it still existed at 17:00-18:00 h. These f indings indicated that the LWP of the 297-28 mutant was signif icantly weaker than that of HD297 and weaker than that of the reference variety SN265. SN265 is a typical rice variety that had lower LWP values over the entire day relative to the drought-tolerant reference variety IRAT109.

(2) Daily change in the Pn. Photosynthesis is undoubtedly one of the most sensitive physiological and ecological traits in the plants' transient response to f ield drought stress, and it is the most important physiological basis for the above agronomic traits, especially in terms of the aboveground biomass and grain yield formation. The changes in Pn(Fig. 4) indicate that a signif icantly different trend occurred in Pnunder the DS treatment compared to that under the WW treatment. Under the DS treatment, the Pnvalues of 297-28 and SN265 reached their maximum at the 07:00-08:00 h, and then they rapidly decreased at 09:00-10:00 h with the increase in the light intensity and temperature; the values showed a slowed-down, decreasing rate from the 09:00-10:00 h to the 14:00-15:00 h , and then during the time from the 14:00-15:00 h to the 17:00-18:00 h, photosynthesis again rapidly became weak and stopped. The Pnvalues of other genotypes were at a very high start point at the 07:00-08:00 h, they were slowly increased (IAPAR9, idr1-1) or slowly decreased (HD297, IRAT109) at the 09:00-10:00 h, and then they fell into a decline or continued to decline during the time from the 09:00-10:00 h to 14:00-15:00 h and were rapidly decreased to 0 during the time from the 14:00-15:00 h to 17:00-18:00 h. Under the WW treatment, the Pnvalues were enhanced with the increase in light intensity, and they reached their peak at the 09:00-10:00 h; then after 10:00 h, with the increased temperature and the continuous increase in the light intensity, the Pnvalues decreased signif icantly; at the 17:00-18:00 h, photosynthesis stopped as the light intensity became weak.

For idr1-1, under the DS treatment, the Pns over the entire day were signif icantly higher than those of IAPAR9 and reached a maximum at the 09:00-10:00 h, the optimal time for dry matter accumulation. Compared with the results under the DS treatment, differences in the Pns between idr1-1 and IAPAR9 under the WW treatment were signif icantly smaller, with crossed overlapping between the two genotypes at different time slots during the day. For 297-28, under the DS treatment, the Pnwas signif icantly lower than that of HD297 even as early as 07:00-08:00 h; with the increased light intensity and temperature, the difference between the two genotypes reached their maximum at the 09:00-10:00 h; later, with the decline in photosynthesis in both genotypes, the difference became small. The Pns over the whole day for 297-28 were lower than those of HD297. Under the WW treatment, the Pncurves for 297-28 and HD297 were close and parallel in the morning, and they exhibited a slight cross in the afternoon, indicating no signif icant difference between them over the entire day.

Analysis of leaf water potential and leaf photosynthesis parameters at key time points(1) Analysis of leaf water potential (LWP, 14:00-16:00 h). From the daily changes in the LWP (MPa, Fig. 3), the LWPs of the rice plants dropped to the lowest point at approximately 15:00 h (i.e., 14:00-16:00 h) and remained relatively stable. At this time slot, the difference in the LWP among the different varieties reached the maximum, and the LWP values measured at this time slot were suitable for comparisons among the species.

Fig. 4 Daily change in the net photosynthesis rate (P n). The measurement day for the well-watered (WW) treatment was 08-22-2013, and for the drought stress (DS) treatment, it was 08-23-2013. The thick solid line represents a genotype under the DS treatment, whereas the thin dashed line represents a genotype under the WW treatment.

As shown in Table 4, except for HD297 and IRAT109 in 2013, all the genotypes had lower LWPs under the DS treatment than under the WW treatment. Specif ically, the reduction rates for 297-28 and SN265 reached an extremely signif icant level, with the 2-year average reduction rates being 28.32 and 25.63%, respectively; by contrast, the idr1-1, IAPAR9, and IRAT109 did not show a signif icant decrease. Additionally, under the DS treatment for the 2-year study period, the LWP values for 297-28 (-1.51 and -1.78) and for SN265 (-1.42 and -1.61) both showed the lowest values, which were signif icantly different from those for the other genotypes. By contrast, idr1-1 showed the highest LWP values (-1.15 and -1.39); its LWP approached that of IAPAR9, IRAT109, and HD297 in 2013 and was signif icantly higher than the results of IAPAR9, IRAT109, and HD297 in 2012. Under the WW treatment for the two study years, the LWP values for 297-28 (-1.21 and -1.35) and SN265 (-1.15 and -1.26) were slightly higher or closer to most upland genotypes (except idr1-1). These results indicate that, under the DS treatment, the responses of the paddy rice varieties were sensitive, showing low LWPs, and the reduction rates were large compared to the WW treatment. By contrast, the LWPs of the upland rice varieties were signif icantly higher than those of the paddy rice varieties; their reduction rates were small and relatively stable, with sometimes even higher levels under the DS treatment than under the WW treatment.

Table 4 also shows that because of the higher degree of drought stress in 2013 than in 2012, the LWPs for all the genotypes were decreased in 2013 compared to 2012, regardless of the treatment. Specif ically, the reduction rates between the 2 years for 297-28 and SN265 under the DS treatment (17.9 and 13.4%) were signif icantly higher than those under the WW treatment (11.6 and 9.6%). HD297 and IRAT109 showed the opposite trend, and their reduction rates (6.4 and 6.9%) were signif icantly lower than those under the WW treatment (25.4 and 20.2%). The reduction rates for idr1-1 and IAPAR9 under the DS treatment (20.9 and 11.0%) were comparable to those under the WW treatment (23.1 and 8.9%). These f indings indicated that the reduction rates in the LWPs for the drought-sensitive genotypes increased as the degree of drought stress increased over the 2 years, whereas those for the drought-tolerant genotypes showed the opposite trend.

In summary, 297-28 was the most sensitive to drought, and its sensitivity was higher than that of HD297 and even that of SN265, which was consistent with how its agronomic traits responded to drought. The drought tolerance of idr1-1 was the strongest and was better than that of IAPAR9, and it could adapt to both drought-stressed and non-stressed environments.

Table 4 Performances of the leaf photosynthesis parameters (09:00-11:00 h) and leaf water potential (14:00-16:00 h)1)

(2) Analysis of the Pnand transpiration rate (09:00-11:00 h). The parameters Pn(μmol CO2m-2s-1) and the transpiration rate (mmol H2O m-2s-1) are the two most important parameters for use in assessing the strength of photosynthesis. Based on the results of the daily change in Pn(Fig. 4), the strongest photosynthesis of rice leaves under the non-DS treatments occurred at the 09:00-11:00 h, a time during which the differences among the different varieties reached the maximum. The photosynthetic parameters measured during this time slot were suitable for comparisons among varieties.

The Pns were generally decreased for all the genotypes under the DS treatment (Table 4), but the responses were signif icantly different among these genotypes. The droughtsensitive types 297-28 and SN265 had the lowest Pnvalues among all the genotypes under both severe and mild drought (6.27 and 6.60, respectively, under severe drought; 13.03 and 14.17, respectively, under mild drought), and they showed the highest 2-year average reduction rates (33.18 and 41.13%, respectively). Additionally, HD297, IRAT109, and idr1-1 had average reduction rates of 0.15, 4.42, and 0.92%, respectively; during the mild drought year (2012), HD297 and idr1-1 had slightly higher Pns under DS than under WW. The Pns of all the genotypes under the DS and WW treatments decreased in 2013 compared with 2012, exhibiting the same trend as that of the LWP. Specif ically, idr1-1 and IRAT109 decreased slightly, and HD297, IAPAR9, and 297-28 decreased more obviously, whereas SN265 showed almost no decrease under WW, yet it experienced a sharp decrease under DS.

During both years, 297-28 had very similar Pnvalues with no signif icant difference relative to HD297 under the WW treatment, but these values were 5.1 lower than those of HD297 on average under the DS treatment, showing a signif icant difference. The 2-year average of the reduction rate in the Pnvalues (from under WW to under DS) for 297-28 was 221 times that for HD297, indicating that 297-28 was highly sensitive to drought stress. idr1-1 showed similar Pnvalues under the DS treatment relative to those under the WW treatment; that is, under the mild DS treatment, it had a slightly higher Pnvalue than it did under the WW treatment, and under the severe DS treatment, its Pnvalue was only slightly decreased (by 3.2% compared with that under the WW treatment and showing no signif icant difference), and it was the highest (16.88) among all the genotypes. Compared with the IAPAR9, idr1-1 had slightly decreased Pnvalues under the WW treatment but showed a signif icant increase by 5.22 under the severe DS treatment, leading to the 2-year average reduction rate of idr1-1, which was 26.93% less than that of IAPAR9. The results showed that idr1-1 was very insensitive to drought stress and was not only superior to IAPAR9 but even better than IRAT109, which was the opposite of the 297-28 outcome.

The transpiration rate(shown in Table 4) showed basically the same trend as that of Pnbetween the years and between the genotypes. Notably, the transpiration rate values for idr1-1 during the 2 DS treatment years were not lower, but instead, they were slightly higher than under those under the WW treatment, and the values for idr1-1 under the DS treatment during the 2 years were both signif icantly higher than those for IAPAR9 and were the highest (6.04 and 7.60) among all the genotypes.

From the above results, the Pnand transpiration rate (09:00-11:00 h) had the same performance as the LWP (14:00-16:00 h) for each genotype.

4. Discussion

4.1. Drought resistance ability of the two mutants

Two years of combined experimental data showed that under drought stress, the idr1-1 mutant had a smaller reduction rate for most agronomic traits (Table 3, e.g., plant height, setting percentage, and aboveground biomass) and physiological traits (Table 4, e.g., leaf water potential, net photosynthetic rate, and transpiration rate) and had a smaller degree of delay in days to heading (Table 3) than its wildtype IAPAR9, suggesting that it has stable performance in terms of morphological development under drought stress, and idr1-1 has a stronger drought tolerance ability than the wild type. In addition, of all the tested materials, idr1-1 has the lowest reduction in aboveground biomass and the highest LADL value in terms of morphology (Table 3), a very important and effective drought-related parameter identif ied by Zu et al. (2017). In terms of physiology, idr1-1 showed the highest leaf water potential, net photosynthetic rate, and transpiration rate values (Table 4) under drought stress conditions. These results strongly supported the interpretation that idr1-1 had the strongest drought tolerance of the six tested genotypes, which was consistent with the former study (Zu et al. 2017). Considering the study of Zu et al. (2017) on 13 representative upland rice cultivars, idr1-1 might be a superior and strongly drought-tolerant mutant in upland rice, even for the entire rice cultivar germplasm of China at present. By contrast, under drought stress, the 297-28 mutant showed a greater reduction rate in each of the agronomic traits than HD297, and its physiological capacities, such as photosynthesis and water potential, were signif icantly lower than that of HD297. The drought sensitivity of 297-28 was even higher than that of SN265, a common rice variety, indicating that 297-28 is a highly drought-sensitive mutant.

4.2. Signif icance in theoretical studies

Both idr1-1 and 297-28 were viewed as two extreme mutants that exhibit strong drought resistance and high drought sensitivity, respectively, and thus they are ideal materials for studying the mechanism of drought resistance. The idr1-1 showed decreased plant height and reduced aboveground vegetative tissues (Table 3); additionally, our unpublished data indicated that idr1-1 had developed root systems with no signif icant difference in the root length (depth), root quantity, root thickness, and root weight and had a shorter leaf length compared with its wild-type IAPAR9. By contrast, the 297-28 showed exactly the opposite changes: an increased plant height, increased vegetative tissues (Table 3), and an evolution from an upland rice root system into a paddy rice root system (e.g., the root length, the number of thick roots and the root thickness were decreased), and the leaf length was increased, and the leaf width was decreased (data not shown). A study by Lilley and Fukai (1994) showed a correlation between the plant height and the comprehensive score for drought resistance. Is the decreased plant height and developed root system of idr1-1 the cause of its increase in drought resistance, and are the increased plant height and undeveloped roots of 297-28 the cause of its decrease in drought resistance? Wider rice leaves would confer better drought resistance (Farooq et al. 2010), which was fully evident in the wide and thick leaves of idr1-1. The mesophyll cells in idr1-1 have higher turgor pressure to keep the stomata open and to maintain a high transpiration rate level and photosynthesis intensity; thus, the leaves would not curl, even under drought stress. By contrast, the 297-28 leaves are thin, narrow, and long and they lose water more readily to form curling shapes, illustrating the opposite effect. These changes in the two mutants suggest that the root system, plant height, and leaf morphology are involved in the drought resistance mechanism.

4.3. Signif icance in breeding applications

Mutants serving as the standard control varieties in drought-resistance identif ication experimentsThe two mutants in this study can be used as the standard control varieties in drought-resistance identif ication experiments. At present, the two mutants have been used as control indicators for the two extremes in drought in a standard gradient drought system in the project of “Precise identif ication of the core germplasm resources in rice: precise identif ication in drought” supported by the National Key Research and Development Program of China (2016YFD0100101). The drought-sensitive mutants 297-28 can accurately indicate the onset of a drought period, whereas the strong drought-resistant mutant idr1-1 can indicate whether the drought intensity reaches the desired level. Using these mutants as representatives of each extreme, the accuracy of the drought-resistance identif ication experiments can be controlled under different test environments.

Mutants serving as the donor parents carrying droughtresistant and dwarf ing genesA previous study indicated that plant dwarf ing and lodging resistance could help increase the yield (Sasaki et al. 2002). Kumar et al. (2008) noted that when using the hybridization of drought-resistant parents and high-yield yet drought-tolerant parents, there is a better chance for obtaining high-yielding and droughttolerant recombinants from the hybridization offspring. In the past, the hybridization of paddy rice and upland rice has been mostly used in conventional breeding for droughttolerant upland rice. Therefore, as a semi-dwarf, idr1-1 is strongly drought-resistant, and as a high-combining-ability material, it would have inherent advantages in conventional breeding. After hybridization with rice, the chance to screen for high-yield and drought-resistant conventional droughttolerant varieties would be high. Our lab has used the idr1-1 mutant in the heterosis utilization of rice by taking it as a bridge for the transfer of the drought-resistant cytoplasmic male sterility (CMS) line. Unpublished data show that idr1-1 can quickly combine its drought-tolerance performance with heterosis, and it is expected to breed high-yield, droughttolerant, and strongly adaptable hybrid rice combinations that can solve food security issues.

5. Conclusion

In this study, comprehensive and comparative observations of physiological traits were performed, including morphological development, agronomic traits, water use, and photosynthesis. The results showed that idr1-1 is a dwarf mutant of the high-stalk upland rice variety IAPAR9, and it confers signif icantly enhanced drought resistance; 297-28 is a mutant of the superior drought-resistant variety HD297 and is highly sensitive to drought, with some drought resistance traits that are even weaker than those of the typical rice SN265. These two mutants could be used as important materials in drought resistance identif ication studies on rice cultivar germplasm resources and in drought resistance mechanism studies in the future.

Acknowledgements

This work was supported by the National Key Research and Development Program of China (2016YFD0100101), the 948 Program of China (2006-G51), and the European Commission within the 6th Framework Program (ECFP6) INCO-2003-B1.2 (CEDROME-015468). We thank American Journal Experts (AJE) for their English language editing.