Study On The Water Requirement Of Paddy And Upland Rice And The Genetic Basis Of Drought Tolerance At Reproductive Stage In Rice

The shortage of water resources has become the crisis faced by the world. Using water-saved technology and developing drought tolerance rice varieties in limited water condition can not only save water in a large degree but also increase and stabilize yield, save energy and minimize environment pollution. Water-saving in crops involves three aspects: firstly, agricultural water-saving, to improve the farmland water use efficiency (WUE) by adjusting the crops layout and cultivation according to the characters of crops; secondly, physiological water-saving, to irrigate the plant at scheduled times and amounts according to the drought resistant and water requirement of crops; thirdly, biological water-saving, to find the WUE genes and molecular markers, clone relative genes, study their function, combine the transgenic technology and marker-assisted selection with traditional breeding, pyramid drought-resistant and water-saving genes, and develop high yield and WUE cultivars. In this research, the differences of water requirement between paddy and upland rice were compared and analyzed; a genetic linkage map was constructed for mapping QTLs effecting drought tolerance in an RIL population of 187 lines from a cross between an indica paddy rice and upland japonica rice. Main results of the present study are as following:1. The growth, photosynthetic characteristics and grain yield were studied using paddy rice "Shanyou63" and upland rice "Zhonghan3" with five water irrigation treatments: A (irrigated water was 25% of CK), B (irrigated water was 43.75% of CK), C (irrigated water was 62.5% of CK), D (irrigated water was 81.25% of CK) and E (conventional irrigation, irrigated water was 12000 m~3/hm~2, CK). The results showed that irrigation treatments affected obviously the characteristics above in two kinds of rices. Compared with those under conventional irrigation, losses of grain yield of Shanyou63 in treatment A and B were 80.58% and 42.98%, respectively, while the differences of grain yield were not significant among treatments C, D and E. Loss of grain yield of Zhonghan3 was 68.42% and the difference was significant in treatment A, while not significant among other treatments.2. The irrigated water was 7500m~3 and 5250m~3 per ha to produce high yields for Shanyou63 and Zhonghan3, respectively. When irrigated water was beyond the amount, grain yield of the two cultivars decreased. Compared with paddy rice, upland rice save about 30% of irrigated water to produce high yield. Yield of hybid rice yet decreased with the growth of unit irrigated water. At meeting water requirement of hybid rice, the controlled irrigation may keep paddy rice at high photosynthesis rate and chlorophyll content, have reasonable yield components, and eventually result in high yield. The drought tolerance would play important action in upland rice and result in high yield under limit water condition. In all of treatments, the yields of upland rice were lower than that of paddy rice because of less effective panicles per plant in Zhonghan3.3. A complete and well-distributed molecular linkage map with 213 SSR markers was constructed using a set of recombinant inbred lines derived from the cross of Zhenshan 97B (indica, lowland rice)/IRAT109 (japonica, upland rice). The genetic map spanned 1825.0cM and covered all of 12 rice chromosomes with an average distance of 8.6cM between adjacent markers. The proportion of alleles from Zhenshan 97B and IRAT109 in the RI population was 0.53 and 0.47 respectively. The skewness of the genetic structure of the RI population was 0.54, which approximately fitted a normal distribution.This RI population was good for map construction and QTL mapping.4. A special designed drought screening facility, which could form normal and stress water condition in the same field, was used to evaluate DT of RILs. The ANOVA results of the SWC in 2003 and 2004 showed significant difference between water treatments, soil depths and stress stages. The ranges of the water gradient between stress and control was 20.5%, 13.8%, 9.8% (in 2003) and 23.4%, 15.5%, 9% (in 2004) at 25cm, 50cm and 75cm depths, respectively. Highly significant differences were observed among the lines for yield, yield components and others traits. The effect of water treatments and treatment-by-genotype interaction were significant for yield and its components. The 2-years average yields of Zhenshan97B, IRAT109 and RI population were 13.17g, 10.13g, and 8.10g in well-watered condition. Yield loss from drought stress was 75% for Zhenshan97B, 15% for IRAT109, and 37% for RILs on average. Yield loss varied from 0% to 100% for different lines. The water treatments were properly applied in this study for the population to fully exhibit the variance of drought tolerance.5. Highly significant positive correlation between GY and BY, then SN, SF, HI, PN and TGW under well-watered condition, while under drought stress condition, HI has highest positive correlation with GY, then SF, BY, PN, TGW and SN. Path analysis partitioned the correlation coefficients into direct and indirect effects. In well-watered condition, BY had the highest direct effect on GY followed by HI. SN, SF, PN and TGW had similar but small direct effect on GY. However, SN had highest indirect effects on GY by other traits, then SF. In drought stress condition, HI had the highest direct effect on GY followed by BY. SN, SF, PN and TGW had small direct effect on GY. However, SF had highest indirect effects on GY by other traits, then PN and TGW. There was complicated relation between GY and its components. Phenotypic correlation can not show their internal genetic relation. 6. Quantitative trait loci (QTL) mapping was carried out for grain yield, yield components and other drought-tolerant relative traits in rice under normal water condition and drought stress in 2003 and 2004. Based on composite interval mapping method at the threshold LOD≥2.0, under normal water condition, total 34 main-effect QTLs were detected, 3, 3, 2, 4, 3, 3, 4, 3, 4, 3, 0 and 2 QTLs for GY, BY, PN, PL, SN, SF, TGW, HI, HD, PH, RWC and CT, respectively, in 2003; In 2004, total 23 main-effect QTLs were detected, 2, 2, 2, 3,1, 3, 4,1,1, 3,1 and 0 QTLs for the 13 traits, respectively. All of them, 13 QTLs were simultaneously or adjacently identified in both of the years, 2,1, 2,1, 1, 2,1 and 3 QTLs for BY, PN, PL, SN, SF, TGW, HD and PH. Under drought stress treatment, total 22 main-effect QTLs were detected, 1, 2, 2, 3, 3, 2, 2, 2, 0, 2, 2, 0 and 1 QTLs for GY, BY, PN, PL, SN, SF, TGW, HI, HD, PH, RWC, CT and LRol, respectively, in 2003; In 2004, total 26 main-effect QTLs were detected, 1, 3, 3, 2, 2, 2, 2, 2, 1, 3, 1, 1 and 3 QTLs for 13 traits. All of them, 9 QTLs were simultaneously or adjacently identified in both of the years, 1, 1, 1, 2, 1, 1 and 2 QTLs for BY, PN, PL, SN, SF, TGW and PH. 8 QTLs were simultaneously or adjacently identified under two contrastive water conditions across two years, 1, 1, 1, 1, 1, 1 and 2 QTLs for BY, PN, PL, SN, SF, TGW and PH, respectively. Other QTLs were detected only in one year or under one kind of water condition.7. A total of 65 QTLs were identified for 13 traits under normal condition and drought stress, based on mixed linear-model composite interval mapping method at P=0.005. These QTLs could be grouped into three major types based on their behaviors under normal water condition and drought stress treatment. TypeⅠincluded 14 QTLs that expressed under both drought stress and non-stress conditions, 2,1, 2, 2, 3, 2 and 2 QTLs for BY, PN, PL, SN, SF, TGW and PH, respectively; TypeⅡcomprised 21 QTLs that expressed under non-stress condition but not under stress, 3, 1, 1, 2, 2, 4, 2, 3, 1, 1 and 1 QTLs for GY, BY, PN, PL, SF, TGW, HI, HD, PH, RWC and CT, respectively; TypeⅢincluded 16 QTLs that were apparently induced by stress, 2, 2, 1, 1, 1, 1, 1, 2, 1, 1, 1 and 2 QTLs for GY, BY, PN, PL, SN, SF, TGW, HI, PH, RWC, CT and LRol, respectively. The phenotypic variation explained by individual QTL ranged from 1.29% to 14.76%. In this study, only three loci were detected with significant QTL×E interaction. One for SF (qSF-7b) and one for RWC (qRWC-8) showing Q×E interaction (h~2=7.36% and 2.32%, respectively) under normal condition. The other was for SN (qSN-9), which can explaine 9.17% of total variance for SN under drought stress condition.8. QTLs of phenotypically correlated traits are mapped together. Several QTL for GY overlapped or were linked with several QTLs for its components and favorable alleles came from one parent in this study. For example, the RM526-RM6 interval on chromosome 2 contained the qGY-2b,qPN-2,qSF-2a, qSF-2b, qTGW-2a and qTGW-2b; the RM132-RM231 interval on chromosome 3 contained the qGY-3, qBY-3, qPL-3, qSF-3, qHD-3, qPH-3 and qRWC-3; and the other region RM470-RM349 on chromosome 4 contained the qGY-4, qPN-4a, qPN-4b, qSN-4a and qSN-4b. IRAT109 contributed the favorable alleles for all QTLs except PN. This result showed close-linked genes or pleiotropic effect controlling yield and its components, and explained the significant positive correlation between GY and its components. These genomic regions will not only be useful for rice improvement, but also contribute to our understanding of the genetic control of GY under drought stress.9. Total 159 pairs of significant interactions for 13 traits were detected under two contrastive water conditions. These epistatic interactions could be grouped into three types based on the involved QTLs property. TypeⅠthe interaction included two main QTLs. Only one significant QTL pair was detected under normal water condition and drought stress treatment, respectively; TypeⅡthe interaction was between one main QTL and one unsignificant QTL, which was comprised 33 pairs; TypeⅡthe interaction was between two unsignificant QTLs, which was in large scale and included 124 pairs. The contribution rates of a single QTL pair varied from 0.10% to 9.20%, and the total contribution rates of all QTL pairs for each trait varied from 0.67% to 41.79%. A large number of significant interactions indicated that the interaction might have great effect on rice DT. In addition, some interaction intervals were detected affecting more than one trait. One interactive interval might have complete reverse effect on phenotypic variation in different traits. Understanding the relationship of interactions clearly is helpful to obtain the best combine of these interations and to improve the DT of rice.10. Late season drought tolerance of rice was negatively associated or co-lacated with some morphological traits like deference of canopy temperature, reduction of plant height, delay in flowering time, and drought respond index across two-year experiments. So, these traits can be used as indirect index in field screening for late season drought tolerance in rice.