Introduction

Rice as a moderately salt sensitive plant is one of the most valuable crop in the world and showed suffer symptoms when the concentration of soluble salt in soil reached at 0.3 %, and lead to yield declines (Akbar et al. 1972; Korbe and Abdel 1974; Maas and Hoffman 1977; Grover and Pental 2003).One-third of the global population use rice as their staple food, and cultivated area of rice in China is 30 million hm2 and its yield is more than 40 percent of the national production of alimentary crop. Salt-alkaline is one of the main abiotic stresses for agricultural production in the world. About 400–950 million hectares covered by salt-alkaline soil of the earth surface (Chandan et al. 2006). As abiotic stresses salt-alkaline soil is an important restriction factor for the reduction of rice production and growing in Asia and Africa. The trend of salt-alkaline soil was becoming worse with the global warming, improper irrigation and improper fertilization. In order to reduce the effect of salt-alkaline land to rice production, people improve soil condition with a high cost and not a through method; on the other hand they breed high salt-alkaline tolerance rice cultivars through traditional approach, which is a useful but spend time. In view of these, to understand the genetic mechanism of salt-alkaline tolerance and discover the major genes or QTL in rice and then improving superior salt-alkaline tolerance cultivar to increase the soil production would be an economic and effective method to deal with this agricultural problem.

The genetic variation of salt-alkaline tolerance of rice was very plenty, and showed different salt-alkaline tolerance at the different physiological stage (Ikehashi and Ponnamperuma 1978; Gregorio et al. 2002). According to the preceding reports, salt-alkaline tolerance of rice was a quantitative trait controlled by multiple genes, except in some mutants or transgenic strains which were controlled by single major gene (Lin et al. 2004; Qi et al. 2008; Kim et al. 2009; Tanveer et al. 2008). So, identification of QTL associated with salt-alkaline tolerance and then polymerize them to improve salt-alkaline tolerance of rice would be an effective method in the program of breeding salt-alkaline tolerance cultivars (Yeo and Flowers 1986; Flowers et al. 2000). Causse et al. (1994), Cho et al.(1998) and Mccouch et al. (1988) reported that use of molecular marker technique in mapping of rice greatly promoted studies of quantitative trait in rice. Yano et al. (1997) detected QTLs associated with heading stage and yield in rice. Kim et al. (2009) used BC3F5 which derived from the cross between Ilpumbyeo and Moroberekan identified 8 QTLs under controlled and salinity conditions at seedling stage, and of them two for reduction rate of dry weight, three for reduction rate of fresh weigh, two for reduction of leaf area, which were located on chromosomes 1, 6, 7, respectively. Islam et al. (2011) used F2 population which derived from the cross between BRRIdhan40 and IR61920-3B-22-2-1 identified 3 major QTL associated with salt damage scale in rice at seedling stage, and these 3 QTL located on chromosomes 1, 8, and 10 respectively. Hossein and Atefeh (2008) used F2:3 population which derived from the cross between Tarommahalli and Khazar identified four QTLs associated root length, two QTLs for dry root weigh and three QTLs for ion exchange under salt stress. Lin et al. (2004) used F2:3 population derived from the cross between Nona Bokra and Koshihikari detected three QTLs associated seedling survival days of rice under salt stress with the total phenotypic variation from 13.9 to 18.0 %. Jafar and Fotokian (2011) used the BILs population which derived from the cross between Tarome-Molaei and Tiqing indentified 14 QTL, and of them the QKr1.2 for K+ content was detected on chromosome 1 which with 7.8 of LOD and 30 % of the total phenotypic variation was a major QTL. All studies above show that most QTLs which were detected recently are of little phenotypic variation, and indicated that complex physiological trait as salt-alkaline tolerance may be regulated by several main genes and multiple minor genes. Otherwise, according to the classic genetic research, salt-alkaline tolerance of rice was an integrated performance of several physiological reaction which controlled by many quantitative trait loci, and also to be present epistatic effects (Flower 2004; Moeljopawiro and Ikehashi 1981; Jones 1985). The identification of major QTL will help to understand the mechanism of quantitative trait loci which only based on the single gene model, but according to the knowledge of Biochemistry and Developmental Genetic, there were would be lots of interaction among gene product (Wright, 1980). In the past few years people paid more attention to the research of epistatic loci and there were proof to show that epistatic effects would be an important genetic mechanism for species adjusting to environment(Allard 1996; Rieseberg et al. 1996). Otherwise, some researchers indicated that QTL in plant showed spatiotemporal expression (Wu et al. 2010; Cui et al. 2011). So, to analyze spatiotemporal expression of QTL associated with salt or alkaline tolerance will give us more information of the genetic mechanism in plants than to analyze QTL at a single stage (Cui et al. 2011; Wang et al. 2010; Wu et al. 2010), and it will be very importance for understanding the mechanism of genetic basis in plants. Most of preceding studies focus on the salt tolerance of rice. There were only a few of preceding studies on the alkaline tolerance of rice (Cheng et al. 2008; Qi et al. 2009;2008). But there were no reports of using the same population to indentify QTL and epistatic loci associated with salt or alkaline tolerance and analyze the difference between them at the same time, and also few reports of spatiotemporal expression of salt or alkaline tolerance at seedling stage in rice. In view of this, RILs population derived from the cross between Yiai1 and Lishuinuo are used in this study, and traits of DLR and DSR at seedling stage would be measured at 10 and 20 days under salt stress, and trait of DLR at seedling stage would be measured at 10 to 60 days, but the trait of DSR would be measured at seedling stage at 55 days after transplanted in alkaline stress. Meanwhile, with the purpose of further understand genetic mechanism of salt or alkaline tolerance in rice, SSR Markers which different between parent strain were used for construction of genetic linkage map, and based on this map QTL and epistatic loci associated with DLR and DSR at seedling stage under salt or alkaline stress were also analyzed.

Materials and methods

Mapping population

An RILs population F8 containing 200 lines derived from the cross between Yiai 1 (YA) and Lishuinuo (LSN), two rice varieties, and F1 population was made in Beijing in 2001, and F2 seeds acquired in 2002 through adding generation, and F8 generation acquired during 2002–2005, in Beijing and Hainan through adding generation and SSD method. YA was a salt-alkaline tolerant parent and LSN was a salt-alkaline sensitive parent.

Evaluation of the salt tolerance for the mapping population and the two parents

The salt tolerance of the two parents and RILs were evaluated in the greenhouse of the Institute of Crop Science, Chinese Academy of Agricultural Sciences in 2006. The seeds of parents and RILs were soaked in water for one day at the room temperature, and transported to calorstat for accelerating germination. The budding seeds near the same derived from each line were planted in vinyl chamber, each line a row and each row for 20 strains, and the distance between rows and between hills was 5 cm × 1 cm. The salt stress with the NaCl concentration of 0 and 1.5 % was carried out at the stage of 3 leaves of the seedling, respectively, and each treatment was repeated twice. Dead leaf rate (DLR) and dead seedling rate (DSR) of 10 strains of each line was evaluated at 10 and 20 days after transplanted under salt stress, then average DLR and average DSR was accounted, and the statistic unit was the average of that 10 strains. The means were used for data analysis with the following method:

DLR (%) = (the total number of dead leaves of treated strains/the total number of leaves of treated trains) × 100

DSR (%) = (the total number of dead seedlings of treated strains/the total number of seedlings of treated trains) × 100

Evaluation of the alkaline tolerance for the mapping population and the two parents

The DLR and DSR of alkaline tolerance of the two parents and RILs were evaluated in the alkaline tolerance evaluated pool in Rice Research Institute, Jilin Academy of Agricultural Sciences. The soil in the evaluated pool is from heavy salt-alkaline area of Baicheng, Jilin Province, and its pH value was above 9.5. After running water irrigation, the pH value of the soil and water in the pool was adjusted to between 8.7 and 8.9, at 25 °C. The depth of water was kept at 3-5 cm above soil at the seedling stage of rice, 2–3 days checked the water depth level and then irrigated water into the pools so as to kept the water at the lever scale, the water in the pools were also kept still. In order to maintain the stabilization of the water pH value, the open canopy was placed 2 meter above the pools to keep off rain fall. The rice seeds were sown on April 10th, 2006 with the method of seedlings grown on the flat nursery with dry soil in a greenhouse. These seeds were transplanted into the evaluated pools at the stage of 4 leaves. 20 seeds of each line were planted in a single row 20 cm × 10 cm apart. Each treatment was repeated twice. The DSR of 15 plants was investigated at 55 days after transplanted into the pools, and the DLR was also investigated at 10, 15, 20, 25, 30, 35, 40, 45, 50, 55 and 60 days. The means were used for data analysis with the following method:

DLR (%) = (the total number of dead leaves of treated strains/the total number of leaves of treated trains) × 100

DSR (%) = (the total number of dead seedlings of treated strains/the total number of seedlings of treated trains) × 100

Extraction of DNA and SSR markers selection

The total genomic DNA of both parents and RILs was extracted from 4–5 leaves of the plant at the tillering stage and the concentration of the DNA was analyzed in ultraviolet spectrophotometer. Two thousand pairs of SSR primers were selected from the Website of http://www.gramene.org/microsat/microsats.txt . The SSR primers were synthesized by the Saibaisheng Company in Beijing. Among these, 155 pairs of SSR primers selected were polymorphic between the two parents. PCR reaction system as follows: 1 µl (20 ng/µl) DNA; 0.25 µl(5U/µl) Taq polymerase; 1.0 µl (2 µM)SSRprimer; 0.75 µl(2.5 mM) dNTPs; 1 µl 10 × buffer(including Mg2+); 6.0 µl ddH2O. PCR program as follows: 94 °C for 5 min for intial; 94 °C for 30 s; 55 °C for 30 s; 72 °C for 1 min; for 35 cycles. The PCR products were separated in 8 % polyacrylamide gel electrophoresis and visualized by the method of silver staining.

Date analysis

The genetic linkage map and QTL analysis were done by using the software of Mapermaker 3.0, and QTL detected at the critical value of 3.0.The nomenclature of QTL was done as described previously (MaCouch et al. 1997). The genetic linkage map was made as described previously (Liu and Meng 2003). The mean, range of variation, standard deviation and coefficient of variation of the date was analyzed by using the SAS 9.1 (p < 0.05).

Results

Phenotypic variation of DLR and DSR of both parents and population of RILs at the seedling stage under salt stress

The results of the evaluation of DLR and DSR of both parents and RILs at the seedling stage under salt stress were shown in Table 1. Under salt stress for 10 days after transplanted, the DLR of Yiai 1 was 32.90 % and the DLR of Lishuinuo was 68.19 %, but the DSR of Yiai 1 was 6.67 %, the DSR of Lishuinuo was 18.33 %. At 20 days after transplanted, the DLR of Yiai 1 was 55.93 %, and DLR of Lishuinuo was 97.87 %, but the DSR of Yiai 1 was 26.67 %, and DSR of Lishuinuo was 87.50 %, and it showed that salt tolerance of Yiai 1 was obviously stronger than Lishuinuo. The range of variation of DLR of RILs under salt stress for 10 days and 20 days were 23.4–66.5 % and 53.0–100 %, respectively; The range of variation of DSR of RILs under salt stress for 10 and 20 days were 0–34.6 % and 12.1–100 %, respectively. When the time of salt stress was increasing, the degree of salt damage was also increased, and the mean of DLR for the RILs at 10 and 20 days were 42.94 and 81.79 %, respectively, and the mean of DSR for the RILs at 10 and 20 days were 7.10 and 63.75 %, respectively.

Table 1 Variation of dead leaf rate (DLR) and dead seedling rate (DSR) at the seedling stage for RILs under salt stress
Full size table

The distribution of DLR and DSR under salt stress for RILs showed in Figs. 1 and 2, respectively. At 10 and 20 days after transplanted in the salt stress, the DLR of the RILs showed a continuous normal or near normal distribution, it is quantitative trait controlled by multiple genes. The DSR of the RILs showed a continuous distribution with 2–3 peaks and skewed to lower value under salt stress for 10, and 20 days, it is quantitative trait controlled by a few major genes and many other minor genes together.

Fig. 1
figure 1

Distribution of dead leaf rate under salt stress for mapping population (YA Yiai 1, LSN Lishuinuo)

Full size image
Fig. 2
figure 2

Distribution of dead seedling rate under salt stress for mapping population (YA Yiai 1; LSN: Lishuinuo)

Full size image

Phenotypic variation of DLR of both parents and RILs population at the seedling stage under alkaline stress

In general, the results of the evaluation of DLR for both parents and population of RILs at the seedling stage under alkaline stress showed an increasing tendency with the time of alkaline stress increasing (Table 2). At 10–60 days after transplanted in alkaline stress, the range of variation of DLR for Lishuinuo was 18.34–58.24 %, but the range of variation of DLR for Yiai1 was 10.75–33.88 % and this indicated that the alkaline tolerance of Yiai1 was obviously stranger than Lishuinuo.

Table 2 Variation of dead leaf rate (DLR) for parents and RILs population under alkaline stress condition
Full size table

The tendency of the mean values (35.29–38.78 %) of DLR for RILs was increasing during the days of 10–20, but it was decreasing at the early time and increasing at the end time during the days of 25–40, and then it was increasing during the days of 45–60, after transplanted in the alkaline stress. The range of variance of the mean values for the DLR of RILs populaion was 22.76–38.78 % at the whole alkaline stress time, but the range of variance of the mean values for the DLR of RILs population was 8.87–12.4 %. According to the distribution of DLR for RILs population (Fig. 3), we known that the DLR of RILs showed a near normal continuous distribution during the days from 10 to 60 under alkaline stress, it was considered that the DLR is a quantitative trait controlled by multiple genes.

Fig. 3
figure 3

Distribution of dead leaf rate for RILs population during the days of 10–60 after transplanted into alkaline stress

Full size image

Phenotypic variation of DSR of both parents and RILs population at the seedling stage under alkaline stress

At 55 days after transplanted in alkaline stress, the DSR of Yiai 1 and Lishuinuo was 0 and 35.0 %, respectively, it showed that the alkaline tolerance of Yiai 1 was obviously stronger than Lishuinuo. The DSR of the RILs showed a large phenotypic variation with a range of variance from 0 to 88.9 %, and its means was 22.1 %, and its coefficient of variation was 110.8 %. According to the distribution of the DSR when treated 55 days under alkaline stress for RILs population (Fig. 4), it showed a continuous distribution which was secund a low value of phenotypic variation with 2–3 peaks, it indicated that the DSR was a quantitative trait controlled by a few major genes and many other minor genes together.

Fig. 4
figure 4

Distribution of dead seedling rate when treated 55 days under alkaline for RILs population (YA Yiai 1, LSN Lishuinuo)

Full size image

QTL analysis

Construction of genetic linkage map

In this study 155 SSR markers which were polymorphism between parents were used for the construction of genetic linkage map which with the average distance of 9.95 cM between each two markers was overlay the whole rice genome of 1541.5 cM (Fig. 5)

Fig. 5
figure 5

Likelihood intervals for additive and AA epistatic QTLs of dead leaf rate and dead seedling rate under salt or alkaline stress for mapped RILs derived from the cross between rice variety Yiai 1 and Lishuinuo. Epistetic QTL areas that each two figures with the same shading located in represent interaction interval of Epistetic QTL

Full size image

The DLR and DSR of rice under salt stress

As shown in Table 3, the one QTL, qDLRs4 - 1 which associated with the DLR at 10 days after translated in salt stress was detected in RM3471-RM261 on chromosome 4, with the LOD of 3.26, which explained 8.65 % of the observed phenotypic variance. Two QTLs, qDLRs4 - 1 and qDLRs8 - 1, which associated with the DLR at 20 days after translated in salt stress were detected in RM307-RM471 on chromosome 4 and in RM22741-RM404 on chromosome 8, Respectively, which explained 11.01 % and 18.28 % of the observed phenotypic variance, respectively, which are major QTLs.

Table 3 Additive QTL detection for DLR and DSR under salt stress and alkaline stress respectively in rice
Full size table

For the DSR, three QTLs, qDSRs4 - 2 and qDSRs8-1 which associated with the DSR at 10 days and qDSRs5 - 1 associated with the DSR at 20 days after translated in salt stress were detected. These QTLs located in RM307-RM471 on chromosome4, RM22741-RM404 on chromosome 8, and RM163-RM5401 on chromosome 5, respectively. Which the observed phenotypic variance were 27.2, 15.96 and 9.99 %, respectively, the qDSRs4-2 and qDSRs8-1 are major QTLs. Additive alleles of qDLRs4 - 1, qDLRs8 - 1, qDSRs4 - 2, and qDSRs8 - 1 originated from Lishuinuo (LSN), but additive alleles of qDSRs5 - 1 originated from Yiai1 (YA).

The DLR and DSR of rice under alkaline stress

Three QTLs (Table 3), qDLRa1 - 1, qDLRa5 - 2 and qDLRa5 - 3 associated with the DLR at 20 days, 45 days and 50 days after transplanted in alkaline stress were detected located in RM243-RM583 on chromosome 1, RM289-RM413 on chromosome 5, RM7118-RM169 on chromosome 1. The range of variation for LOD value of these three QTLs was 3.51 to 4.41, and the observed phenotypic variance of these 3 QTLs were 8.29, 33.25 and 8.32 %, respectively. The qDLRa5 - 2 is major QTL. Additive alleles of qDLRa1 - 1 originated from Yiai 1, but additive alleles of qDLRa5 - 2 and qDLRa5 - 3 originated from Lishuinuo. And there were no QTLs associated with DSR for alkaline stress were detected in this study.

Epistatic QTL of AA interaction for DLR and DSR under salt or alkaline stress

In the study of epistatic QTL of AA interaction for DLR and DSR under salt or alkaline stress, three pairs of epistatic QTLs were detected (Table 4). Of them, two pairs for DLR were detected at 30 and 35 days after transplanted in alkaline stress, respectively. One pairs for DSR was detected at 55 days after transplanted in alkaline stress. All these QTL located on chromosome 1, 5, 6. The AA values of epistatic QTL for DLR at 30, 35 days were 5.27 and -6.67, respectively; LOD values were 5.5 and 5.0, respectively; and the observed phenotypic variance were 13.08 % and 23.91 %, respectively. The AA value of epistatic QTL for DSR at 55 days was 13.14, LOD value of it was 6.2, and the observed phenotypic variance was 28.84 %.

Table 4 Epistatic QTL of AA interaction for DLR and DSR under salt or alkaline stress in rice
Full size table

qDLRa1 - 1i / qDLRa6 - 2j was with the negative value which implies that the two-locus genotypes from the same parent Yiai 1 or Lishuinuo take the negative effects, while the two-locus genotypes of recombination from parent Yiai 1 and Lishuinuo take the positive effects, and the case of plus AA values such as the other two pairs of QTL were just the opposite. Otherwise, there was no Epistatic QTL of AA interaction for DLR and DSR under salt stress.

Discussion

Genetic control of salt or alkaline tolerance in rice

Salt and alkaline tolerance in rice is not only a complicated physiological trait, but also a complicated genetic trait (Qi et al. 2008). The common understanding was that rice increases its salt tolerance through discharge of Na+ and absorption of K+ to maintain a lower rate of Na+/K+. (Gregorio and Senadhira 1993) considered that additive and dominant genes play a role in maintaining the lower rate of Na+/K+. The salt tolerance materials can be induced by the incomplete dominance genetic mutation (Zhang et al. 1995). The salt tolerance at the seedling stage in rice was controlled by a few genes (Jones 1985) and genetic variation of under salt stress showed an additive and interactional character (Moeljopawiro and Ikehashi 1981).

In the present study, the DLR of RILs population under salt or alkaline stress showed a continuous distribution, it was controlled by multiple genes, which is the same with the results in the previous studies. The results showed that rice could adapt the salt stress condition at the early stage, but with the time of salt stress increasing, the cellular tissue of rice was continue to be destructed and then the limit of the previous salt tolerance was to be broken, so at 20 days after transplanted in salt stress showed a higher DLR distribution. Under alkaline stress, the RILs population showed an increasing tendency of DLR during the days of 10–20, then showed a decrease of DLR at early stage in a little extend, but a high decrease of DLR at late stage in a little extend during the days of 25–40, and then showed an increasing tendency of DLR again during the days of 45–60. This may be because of that the rice was sensitive to alkaline stress at the early seedling stage and showed an increasing tendency of DLR during the days of 10–20; the rice adjusted itself to adapt the alkaline stress condition during the days of 25–40 and showed a decrease of DLR at early stage in a little extend, but a high decrease of DLR at late stage in a little extent; the rice was sensitive to alkaline stress again during the days of 45–60 and showed an increasing tendency of DLR again. The result was similar to the results of (Flowers and Yeo 1981) and Lutts et al. (1995). In this study, DLR is a quantitative trait controlled by multiple genes, but DSR is a quantitative trait which controlled by a few major genes and many other minor genes together. The result was similar to the results of Flowers and Flowers (2005) and (Jones 1985) indicated in their studies on barley, rice, citrus, and tomato.

Detection of QTLs for salt tolerance in rice

Zhang et al. (1999) showed that OSA3 gene on rice chromosome 12 expresses PM H+-ATPase; based on the relative abundance of PM H+ -ATPase gene transcripts in M-20 shoots, they indicated that the active role of this gene in the strict control of Na+ and Cl+ uptake into root symplast and apoplast and further translocation into the shoots and leads to the reduced gene expression in M-20 shoots under salt stress condition to adjust its salt tolerance. Michael et al. (2010) indicated that the Salto gene controlled the balance of Na+/K+ in rice cells to increase its salt tolerance. Zhang et al. (2006) indicated that the Mangrin increased the alkaline tolerance in transgenic rice. Eight QTLs associated with dry weight, fresh weight, leaf area, seedling length, and survival rate of seedlings at the seedling stage in rice were detected on chromosomes 1, 6 and 7, respectively (Kim et al. 2009). Lin et al. (2004) used F2:3 population to detect two QTLs for survival days of seedlings under salt stress. Of these QTLs, the two major QTLs with the higher effect, qSNC - 7 for shoot Na+ concentration and qSKC - 1 for shoot K+ concentration, explained 48.5 and 40.1 % of the total phenotypic variance, respectively, and were major genes. Takehisa et al. (2004) detected 27 QTLs associated with saline tolerance use the BILs population. According to all the studies above, we know that using different mapping populations always got the different results when identified QTLs associated with salt tolerance under salt stress in rice.

Contrasting the result in this study with the preceding researches, qDSRs5 - 1 was located in the same marker flanking as QSst5a (Qian et al. 2009); qDLRs4 - 1 for 10 days, and qDLRs4 - 1 for 20 days in salt stress were located in the same marker interval as qSH4 (Chen et al.2008), QSst4, QKc4b(Yang et al.2009); qDSRs4 - 2 was located in the same marker interval as QSw4,QGw4a (Chen et al.2008).These marker intervals of QTLs associated with different salt tolerance trait were detected in different mapping populations show that these marker intervals were important areas of rice genome, and would be was closely related with salt tolerance in rice, and worth to be used in the molecular marker-assisted breeding program in rice to get salt tolerance cultivars. In chromosome 8, qDLRs8 - 1 for DLR of rice, and the qDSRs8 - 1 for DSR of rice, were detected on the same marker interval, and the distance of this maker interval is 6.8 cM, which was likely to be the same QTL, or one gene with multifunction or two QTLs with compact linkage.This QTL was a new detected one, so can be used in the breeding program in rice to get salt tolerance rice cultivars in the future.

Detection of QTLs for alkaline tolerance in rice

Qi et al.(2008) used F2:3 population to detect 13 QTLs for DLR of rice at the seedling stage of the days of during 20–62 under alkaline tolerance; of them, the alleles of qDLR9 - 2 and qDLR4 originated from alkaline sensitive parent “Gaochang 106”, but the alleles of qDLR7 - 1 and qDLR6 - 2 originated from alkaline tolerance parent “Changbai 9”. Cheng et al. (2008) used two DH populations (DH1: ZYQ8 × JX17and DH2: CJ06 × TN1) identified 10 additive QTLs in H1, 14 additive QTLs in DH2 at the germination and early seedling stage under alkaline stress. At present, there were only a few of studies for alkaline tolerance in rice. In this study three QTLs associated with DLR in rice under alkaline stress were detected, among them, the observed phenotypic variance of qDLRa1 - 1 for 20 days, qDLRa5 - 2 for 45 days and qDLRa5 - 3 for 50 days, with the observed phenotypic variance of 8.29, 33.25 and 8.32 %, respectively.

Contrasting the result in this study with the preceding researches for salt or alkaline tolerance, qDLRa1-1 was located in the same marker interval as qSKc1 for shoot K+ concentration (Lin et al.2004), Std for major salt-tolerance gene (Gong et al.1998), TS1 of cDNA clone for salt response (Qian et al. 2003), and QSt1a (Qian et al. 2009). qDLRa5 - 2 was located in the same marker interval as QDss5, QSf5, QGW5 (Chen et al.2008). The observed phenotypic variance of qDLRa5 - 2 was 33.25 %. The reason of it may because of the distance between two SSR markers were very large, which of 25.6 cM. The same marker interval could be detected QTLs for salt an alkaline tolerance by using the different mapping population, and indicates that people could cultivate new rice cultivars associated with several resistant traits by pyramiding useful QTLs. Otherwise, qDLRa5 - 3 was a new detected QTL located in chromosome 5, which with the observed phenotypic variance of 8.32 %, the distance between two markers of 13.8 cM.

Detection of epistatic QTLs of AA interaction for salt or alkaline tolerance in rice

Quantitative trait loci were controlled by multiple genes, epictatic effects plays an important role in quantitative trait (Phillips 1998). Epictatic effects were the no additive genetic effects between genelocis. Cheng et al. (2008) used two DH populations (DH1: ZYQ8 × JX17and DH2: CJ06 × TN1) identified 15 epictatic QTLs in H1, 15 epictatic QTLs in DH2 at the germination and early seedling stage under alkaline stress. Zhao et al. (2011) used 16 single segment substitution lines (SSSL) and 15 double segment substitution lines (DSSL) to detected seven pairs of epictatic interactions for the grain shape. Lei and Che (2010) used a population of 284 recombinant inbred lines (RIL) from the indica-japonica rice cross Lemont × Teqing to detected nine pairs of epistasis for heading date. Jiang et al. (2009) used a RIL population derived from the—rice cross Minghui86 Jiafuzhan identified 15 pairs of epistatic QTLs with significant additive-by-additive (AA) interaction effects for three traits. Identification for epistatic QTLs associated with salt or alkaline tolerance was scarce at present.

In this study, 3 pairs of AA epistatic QTLs (Table 4) associated with alkaline tolerance. The observed phenotypic variances of them were larger than additive QTLs under alkaline stress,and showed that AA epistatic QTLs was closely related with the resistance to alkaline injured in rice, and these results were similar as Lei et al. (2008) which indicated that epistatic effects was closely related with the inheritance about grain shape and chalkiness degree in rice. But the observed phenotypic variances of AA epistatic QTLs were smaller than additive QTLs under salt stress in rice. The results above showed that the mechanism for salt tolerance was difference to alkaline tolerance in rice. How to use these results in the molecular marker-assisted breeding program in rice to get salt or alkaline tolerance cultivars should be further researched on the base of the preceding studies.

The origin of additive alleles for QTLs of salt or alkaline tolerance in rice

Three QTLs associated with DLR of rice under salt or alkaline stress were detected, of them, the additive alleles for two QTLs originated from the salt-alkaline tolerance parent YA, and the additive alleles for the other seven QTLs originated from the salt-alkaline sensitive parent LSN. The results showed that salt-alkaline tolerance alleles were not only contained in salt-alkaline tolerance variety, but also were contained in salt-alkaline sensitive variety. The phenomenon was similar to some early reports, such as the alleles of over half of the numbers of QTLs for cold tolerance at seedling stage and stooling stage originated from cold sensitive parent “Miyang23” (Han et al. 2005), and such as the alleles of six QTLs for alkaline tolerance at seedling stage in rice originated from alkaline sensitive parent “Gaochan106” (Qi et al. 2008), and such as the alleles of two QTLs associated with high yields originated from lower yield wild rice (Li et al. 2002). Based on the results in this study and the early reports above, we know that excellent gene resource could be detected from any one crop germplasm and should never be eliminated because of poor phynotype.