Biotechnology Letters

, Volume 35, Issue 5, pp 799–810

Mapping of the quantitative trait locus (QTL) conferring partial resistance to rice leaf blast disease

Authors

  • S. Ashkani
    • Department of Crop Science, Faculty of AgricultureUniversiti Putra Malaysia
    • Department of Agronomy and Plant BreedingShahr-e-Rey Branch, Islamic Azad University
    • Department of Crop Science, Faculty of AgricultureUniversiti Putra Malaysia
    • Institute of Tropical AgricultureUniversiti Putra Malaysia
  • H. A. Rahim
    • Agrotechnology and Bioscience DivisionMalaysian Nuclear Agency
  • M. A. Latif
    • Department of Crop Science, Faculty of AgricultureUniversiti Putra Malaysia
    • Bangladesh Rice Research Institute (BRRI)
Original Research Paper

DOI: 10.1007/s10529-012-1130-1

Cite this article as:
Ashkani, S., Rafii, M.Y., Rahim, H.A. et al. Biotechnol Lett (2013) 35: 799. doi:10.1007/s10529-012-1130-1

Abstract

Malaysian rice, Pongsu Seribu 2, has wide-spectrum resistance against blast disease. Chromosomal locations conferring quantitative resistance were detected by linkage mapping with SSRs and quantitative trait locus (QTL) analysis. For the mapping population, 188 F3 families were derived from a cross between the susceptible cultivar, Mahsuri, and a resistant variety, Pongsu Seribu 2. Partial resistance to leaf blast in the mapping population was assessed. A linkage map covering ten chromosomes and consisting of 63 SSR markers was constructed. 13 QTLs, including 6 putative and 7 putative QTLs, were detected on chromosomes 1, 2, 3, 5, 6, 10, 11 and 12. The resulting phenotypic variation due to a single QTL ranged from 2 to 13 %. These QTLs accounted for approx. 80 % of the total phenotypic variation within the F3 population. Therefore, partial resistance to blast in Pongsu Seribu 2 is due to combined effects of multiple loci with major and minor effects.

Keywords

Leaf blastPartial resistanceQuantitative trait locus (QTL)Rice (Oryza sativa)SSR markers

Introduction

Rice blast disease is caused by the fungus, Magnaporthe oryzae, formerly known as Magnaporthe grisea (Couch and Kohn 2002), and is one of the most severe diseases of rice (Oryza sativa). The disease can cause losses >50 % in rice-producing areas (Talbot 2003) including in Malaysia. The most recent outbreak in Malaysia occurred in 2006. More than 60 % of 4,000 Ha cultivated with rice were devastated (Rahim 2010).

The control of the disease is a main objective of plant breeding and pathology research. M. oryzae has been studied extensively, but the diversity and the rapid evolution of the pathogen have never been permanently brought under control. To control this fungal disease, genetic engineering and breeding for resistant varieties are effective and economical methods. Gene pyramiding may provide broad spectrum and durable resistance (Tabien et al. 2002).

Partial resistance is polygenic, non-race specific and durable. This resistance is conferred by quantitative trait loci (QTLs) and reduces pathogen reproduction in compatible interactions (Koizumi 2007). Partial resistance is stable against different pathogenic races of the rice blast fungus; therefore, the use of partial resistance is one of the most promising blast control measures (Zenbayashi et al. 2002). Rice cultivars with durable resistance have been reported in several countries. The cultivar Pongsu Seribu 2 shows a stable and high level of partial resistance to leaf blast in Malaysia (Rahim 2010). This cultivar may possess many qualitative and quantitative resistance genes. The genetic basis of the durable partial resistance in the rice variety Pongsu Seribu 2 is not clear.

Molecular breeding approaches involving DNA markers, such as QTL mapping, marker-aided selection and genetic transformation, have been used to develop new resistant rice cultivars. DNA marker-technologies, such as SSRs, RFLP, RAPD, and AFLP, enable the study of genetic diversity (Latif et al. 2011), the identification and characterisation of stress resistance-related genes (Ahmed et al. 2012) and the construction of a high-resolution genetic map for isolating genes associated with important traits. SSR markers have been used extensively to identify markers linked to blast resistance genes and to locate these genes and QTLs on rice chromosomes (Fjellstrom et al. 2004; Zhu et al. 2004). The QTL detection approach has been employed to map major and minor genes involved in partial blast resistance in rice, as well as in estimating the number, the location and the effect of the genomic region of these genes (Zenbayashi et al. 2002; Tabien et al. 2002; Sallaud et al. 2003). To date, over 96 blast resistance genes or major QTLs have been identified in mapping studies (Koide et al. 2009). The chromosomal locations and linked DNA markers of the genes and regions involved have been reported in several rice cultivars. In the present study, quantitative traits loci (QTLs) related to leaf blast resistance and their chromosomal locations were identified in Pongsu Seribu 2. Identification of the QTLs associated with resistance will help advance molecular marker-assisted breeding and map-based cloning of the major genes and will open new avenues for the genetic manipulation of these candidate genes.

Materials and methods

Plant materials

Pongsu Seribu 2 is an indica rice cultivar from Malaysia with a high level of resistance to leaf blast. This cultivar was used as the donor parent of resistance genes in a cross with a susceptible but popular rice cultivar, Mahsuri. The seeds from each plant of the F2 population were used to produce F3 families under greenhouse conditions in 2010 at the Malaysian Rice Research Centre, MARDI, Pulau Pinang. A total of 188 F3 families were used for the evaluation of resistance to leaf blast and in QTL analysis.

Collection of Magnaporthe oryzae pathotype

P7.2, one of the most virulent M. oryzae pathotypes, was obtained from the Malaysian Agricultural Research and Development Institute (MARDI), Seberang Perai. The pathotype used in this experiment was selected based on its virulence towards Mahsuri (Mh), which is susceptible, and Pongsu Seribu 2 (PS2), which is resistant.

Preparation of inoculums

Potato/dextrose/agar (PDA) was used for growing the selected pathotype of M. oryzae. The single spore isolation procedure was carried out to make sure that the isolate was pure. The spores were isolated from a single colony on a Petri dish under a microscope using a sterile glass needle. Each colony was transferred to PDA slants and incubated at 28–30 °C for 7 days and then used as master culture. The spores used for inoculation were prepared as described by Chen (2001), and adjusted to 105 spores/conidia per ml. Before inoculation, 0.05 % Tween 20 was added to the suspension to increase the adhesion of the spores to the plant.

Seedling preparation and experimental design

The seeds of the F3 families and the two parental varieties/cultivars, Mahsuri and Pongsu Seribu 2, were soaked in water for 1 day and germinated on moist Whatman filter paper (no. 1) in Petri dishes for 3 days. The germinated seeds were then planted in plastic trays (36 × 23 × 10 cm) containing 3 Kg soil with NPK (5 g at 15:15:15) and 3 g ammonium sulphate per 3 Kg soil, as described by Prabhu et al. (2003) with minor modifications. The experiment was replicated three times and completely randomised. In each replication, ten seeds from each F3 line and the parental cultivars were sown in 10 cm rows totalling 21 rows per tray and 10 trays in each block. The plants were grown in a greenhouse at 25–30 °C for 2–3 weeks until they reached the four-leaf stage (Filippi and Prahbu 2001).

Pathogen inoculation

Plants (10 plants per line) at 22 days, with three or four fully expanded leaves, were inoculated by spraying the leaves with 25 ml spore suspension (105 spores/ml) until run-off. The inoculated plants were incubated in a moisture/dew chamber, and the relative humidity was maintained at 100 % for 24 h at 25–28 °C. After incubation, the plants were placed in a greenhouse (a controlled environment) from 25 to 30 °C (Filippi and Prahbu 2001). The relative humidity was maintained above 90 % by covering the plants with black netting and by using a controlled water sprinkler.

Evaluation of a partial resistance to leaf blast in F3 lines

Three components of disease resistance, blast lesion degree (BLD), blast lesion type (BLT) and percentage disease leaf area (% DLA), were estimated and were considered in QTL analysis (Table 1). The most seriously diseased leaves from each plant of each line were used to visually estimate BLD, BLT and  % DLA. The scoring of the traits  % DLA and BLT was based on the methods of Correa-Victoria and Zeigler (1993). The disease severity of 27-day-old plants was scored using a rating from 0 to 100 % for  % DLA. The BLT was scored either 0 (highly resistant: no symptoms), 1–2 (lesions 1–2 mm, no sporulation), 3 (round lesions 2–3 mm with little sporulation), or 4 (spindle-shaped lesions of more than 3 mm with heavy sporulation). The scoring of the BLD trait was carried out using the Standard Evaluation System of the International Rice Research Institute (IRRI 1996) and the protocols described by Mackill and Bonman (1992). The scores from five plants from each F3 family and the 2 parental varieties were averaged, and the mean scores of the replicates were used for QTL analysis.
Table 1

List of disease traits used for identification of QTLs in the study

Pathotypes of M. oryzae

Trait code

Trait name

Description

P7.2

BLD

Blast lesion degree

No observed lesion due to pathotype 7.2, 7 days after inoculation with blast spores in standard evaluation system.

P7.2

BLT

Blast lesion type

Type of observed lesion due to pathotype 7.2, 7 days after inoculation with blast spores.

P7.2

% DLA

Percentage of disease leaf area

Disease lesion area of a rating from 0 to 100 % due to pathotype p7.2, 7 days after inoculation with blast spores

DNA extraction and microsatellite analysis

The genomic DNA from the F3 families and two parental varieties were extracted from fresh leaf tissue by the CTAB method. The parents were checked for polymorphism with the 125 SSR primer pairs that are related to blast resistance genes from the Gramene database (www.gramene.org). Sixty-three polymorphic microsatellite markers were found between the parental cultivars and were located on 10 rice chromosomes as follows: 1, 2, 3, 4, 5, 6, 8, 10, 11 and 12. The forward primers were fluorescently labelled at the 5′-end with either 6-FAM or HEX dye from the G5 set (Applied Biosystems) and incorporated into PCR. PCR amplification was carried out in a Gene Amp System 9700 (Applied Biosystems, Foster City, Calif.). Two types of PCR reactions were carried out. Normal PCR amplification was performed using the methods described by McCouch et al. (2002), with the exception that 12.5 μl reaction mixture was used. The PCR programme was 5 min at 94 °C; 35 cycles of 30 s at 94 °C, 30 s at 55 °C and 30 s at 72 °C; and 5 min at 72 °C for final extension. For some specific cDNA-derived microsatellites, annealing at 61 and 67 °C was employed. Multiplex PCR conditions were performed as described by Pessoa-Filho et al. (2007), with some modification. PCR amplifications were carried out in 10 μl containing 10 ng DNA template, 80 μM dNTPs, 2 mM MgCl2, 0.5 U DNA polymerase, 0.2 μM each of forward and reverse primers and the volume was adjusted with PCR grade water.

Genotyping of F3 plants

Sixty-three fluorescently-labelled polymorphic SSR primer pairs, which included 13 multiplex panels distributed on ten of 12 rice chromosomes, were chosen to genotype 188 lines of the mapping population. A capillary electrophoresis instrument (3100 Genetic Analyzer, ABI) was used for data collection. The fragment size data were scored and analysed using Gene-Mapper version 4.0 and Genotyper Version 3.7 software (ABI) to determine the genotypes of progeny at the Malaysian Nuclear Agency, Agrotechnology and Biosciences Division.

Fragment size determination

The genotypic data were obtained using GeneMapper version 4.0. The software provided the allele size (fragment length) and the chromatograph of each individual plant. The allele size of each individual was compared with the allele size of the parental marker. The individual that showed a peak height and position similar to the marker in Pongsu seribu 2 was designated A; individuals similar to Mahsuri were designated B. Individuals that showed peak heights and positions from both parents were designated AB or H. The marker data generated for 188 lines using 63 SSR primers were used for map construction.

Linkage map construction

An SSR linkage map was constructed using Map Manager QTX version b19 (Manly et al. 2001). The conversion of recombination fractions into centimorgans (cM) was performed using the Kosambi mapping function (Kosambi 1994) with a search linkage criterion of 0.05. The correspondence of linkage groups and the order of the markers on chromosomes were inferred based on the existing genetic linkage map of rice (Lopez-Gerena 2006; Cho et al. 2008) and the rice physical map (www.gramene.org).

Mapping of QTLs

QTL analysis was also performed using Map Manager QTX version b19 (Manly et al. 2001). The association between individual marker loci and QTLs for blast lesion degree (BLD), blast lesion type (BLT) and percentage disease leaf area (% DLA) for pathotype P7.2 was evaluated using a single-marker analysis (SMR). The locations and the effects of the detected QTL were estimated using simple interval mapping (SIM) (Haley and Knott 1992). Likelihood Ratio Statistics (LRS) were converted to an LOD score by dividing them by 4.61. A putative QTL was reported if it was detected by single marker analysis and interval mapping at an LOD >3.0. The QTX software was used to search for epistatic interactions between QTLs. The data were reanalysed and the results were compared using QGene Software ver. 4.2 (Nelson 1997) for all 3 mapping methods: Single Marker regression (SMR), Simple Interval Mapping (SIM) and Composite Interval Mapping (CIM). For CIM, an automatic forward–backward stepwise regression was used for the selection of cofactors (forward p < 0.01, backward p < 0.01) (Wang et al. 2006). Significant thresholds for composite interval mapping were determined using 1,000 permutations for each trait. Identified QTLs were named according to the method proposed by McCouch et al. (1997) and Tabien et al. (2002).

Results

Trait evaluation

The average leaf blast severity scores in Pongsu Seribu 2 at the final assessment were 1.7 and 1.1 for BLD and BLT traits, respectively, which equalled approx. 16.7 % DLA. In Mahsuri, the scores were 4.9 and 2.7 for BLD and BLT, respectively, with 48.4 % DLA. The parental cultivars showed a significant (p < 0.01) difference in leaf blast resistance. The frequency distribution in the F3 lines (Figs. 1, 2 and 3) did not show a continuous variation with normal distribution. The Kolmogorov–Smirnov normality test showed a p > 0.01 for the blast lesion trait. The distributions of disease severity were skewed towards resistance.
https://static-content.springer.com/image/art%3A10.1007%2Fs10529-012-1130-1/MediaObjects/10529_2012_1130_Fig1_HTML.gif
Fig. 1

Trait frequency distribution of leaf blast disease reaction scale for pathotype P7.2 among two parental lines and 188 F3 families for blast lesion degree (BLD)

https://static-content.springer.com/image/art%3A10.1007%2Fs10529-012-1130-1/MediaObjects/10529_2012_1130_Fig2_HTML.gif
Fig. 2

Trait frequency distribution of leaf blast disease reaction scale for pathotype P7.2 among two parental lines and 188 F3 families for blast lesion type (BLT)

https://static-content.springer.com/image/art%3A10.1007%2Fs10529-012-1130-1/MediaObjects/10529_2012_1130_Fig3_HTML.gif
Fig. 3

Trait frequency distribution of leaf blast disease reaction scale for pathotype P7.2 among two parental lines and 188 F3 families for % disease leaf area (% DLA)

We used a logarithmic transformation of the data and a maximum likelihood method for QTL mapping analysis. By simple linear regression, it was indicated that disease severity was strongly correlated with blast lesion degree (BLD), blast lesion type (BLT) and percentage disease leaf area (% DLA) in the F3 lines (r2 > 0.95, p < 0.001). The trait variations (means of the parents and the F3 population, minimum and maximum scores) for pathotype P7.2 are shown in Table 2.
Table 2

Phenotypic values (trait variation) for leaf blast resistance of F3 families and their parents for selected pathotype, P7.2

Pathotype

Traits

Means of parents

F3 Families

P. Seribu 2

Mahsuri

Max.

Min.

Mean

Std. Deviation

Correlation coefficient

p-Values

Size

P7.2

BLD

1.7

4.9

9

0

3.71

2.02

r2 > 0.95

<0.01

188

BLT

1.1

2.7

4

0

2.26

0.96

r2 > 0.95

<0.01

188

% DLA

16.7

48.4

90

5

35.5

22.5

r2 > 0.95

<0.01

188

Molecular linkage map construction

Linkage maps (Fig. 4) of 63 polymorphic SSR markers covering ten chromosomes were constructed for 188 F3 families using Map Manager QTX. These markers were distributed on chromosomes Ch01, Ch02, Ch03, Ch04 Ch05, Ch06, Ch08, Ch10, Ch11 and Ch12. Ten linkage groups (chromosomes) contained approx. seven loci on average, with a minimum of three (e.g., Ch03, Ch10) and a maximum of ten loci (e.g., Ch11). Polymorphic markers were mainly distributed on chromosomes 2, 11, 1 and 6, leading to an expansion of these chromosomes. Genomic regions with a low density of markers appeared as large gaps on the map. In particular, chromosomes 3, 4 and 10 had the shortest distance with the lowest number of loci (3 and 4 SSR loci). Chromosomes 11 and 2 (with 10 and 9 SSR loci, respectively) had the highest number of loci and the greatest map distances. The average interval sizes between markers on the map were 75.97 cM.
https://static-content.springer.com/image/art%3A10.1007%2Fs10529-012-1130-1/MediaObjects/10529_2012_1130_Fig4_HTML.gif
Fig. 4

Linkage map with sixty three SSRs markers and QTL positions detected for blast pathotype, P7.2 for BLD, BLT and % DLA traits in F3 families of Pongsu Seribu 2 × Mahsuri

QTL mapping

A total of 13 QTLs were identified as being related to resistance to blast pathotype P7.2 by single-point marker analysis (SMA) using Map Manager QTX. To localise these QTLs, the marker regression method resulted in six putative QTLs (qRBr-1.2, qRBr-2.1, qRBr-5.1, qRBr-6.1, qRBr-11.1 and qRBr-11.2, Logarithmic of Odds (LOD) >3.0 or LRS >15) and seven suggestive QTLs (qRBr-1.1, qRBr-3.1, qRBr-6.2, qRBr-10.1, qRBr-10.2, qRBr-11.3 and qRBr-12.1 LOD <3.0 or LRS <15). The identified QTLs are presented in Table 3.
Table 3

Statistical characteristics of 13 QTLs detected for three blast lesion traits (BLD, BLT and  % DLA) with SMR (single-marker regression) and SIM (simple interval mapping) for pathotype P7.2 in F3 families

Traita

QTL

Chrb

SMA (closest marker)

IM

for putative QTLs

CIM

for putative QTLs

LRSc

LODd

PV (%)e

Af

pg

P/Nh

BLD,DLA

qRBr-1.1

1

RM462

  

4.7

 

2–3

0.4–4.4

0.0309

BLD,BLT,DLA

qRBr-1.2

1

RM428

RM428–RM1

RM428–RM1

18.6–19.4

4–4.2

9–10

0.3–8.3

0.0000

BLD,BLT,DLA

qRBr-2.1

2

RM208

RM208–RM166

 

21.9–24.1

4.7–5.2

11–12

0.4–10.7

0.0000

BLD,BLT,DLA

qRBr-3.1

3

RM168

  

7–7.8

 

4

0.2–6.1

0.0082

BLD,BLT,DLA

qRBr-5.1

5

RM413

RM413–RM13

 

19.7–21.2

4.2–4.5

10–11

0.3–9.1

0.0000

+

BLD,BLT,DLA

qRBr-6.1

6

RM8225

RM8225–RM136

 

21.9–23.5

4.7–5.0

11–12

0.4–9.9

0.0000

+

BLD,BLT,DLA

qRBr-6.2

6

RM6836

  

4.9–5.9

 

3

0.2–4.7

0.0169

+

BLD,BLT,DLA

qRBr-10.1

10

RM244

  

5.5–5.8

 

3

0.2–4.5

0.0165

BLT

qRBr-10.2

10

RM304

  

4.1

 

2

0.1

0.0437

BLD,BLT,DLA

qRBr-11.1

11

RM5961

RM 5961–RM21

RM 5961–RM21

17.8–18.5

3.8–4.0

9

0.3–8.2

0.0000

BLD,BLT,DLA

qRBr-11.2

11

RM1233

RM 1233–RM6293

RM 1233–RM6293

24.2–25.6

5.2–5.5

12–13

0.4–9.8

0.0000

BLD,BLT,DLA

qRBr-11.3

11

RM6293

  

9.6–10

 

5

0.2–6.4

0.0015

BLD,BLT,DLA

qRBr-12.1

12

RM101

  

10.5–10.8

 

5–6

0.3–6.6

0.0011

QTL nomenclature is according to McCouch et al. (1997) and Tabien et al. (2002)

a The quantitative traits are defined in Table 1, (BLD blast lesion degree, BLT blast lesion type,  % DLA percentage of disease leaf area)

bChromosomal assignment of SSRs

cLRS Likelihood ratio statistic (LRS) for the association of the traits with this locus at p = 0.05

dLogarithm of the odds ratio for putative QTLs

ePercent of the phenotypic variance (PV%) explained by the QTls at this locus

fThe additive regression coefficient for the association

gThe probability of an association this strong happening by chance

hPositive (P)(+) or negative (N)(−) effect of the QTL-associated Pongsu Seribu 2 allele

Out of the 13 QTLs, eleven QTLs (qRBr-1.2, qRBr-2.1, qRBr-3.1, qRBr-5.1, qRBr-6.1, qRBr-6.2, qRBr-10.1, qRBr-11.1, qRBr-11.2, qRBr-11.3 and qRBr-12.1) led to statistically detectable reductions in the BLD, BLT and  % DLA traits. One QTL (qRBr-1.1) was detected as being involved in two traits, BLD and % DLA. Another QTL (qRBr-10.2) was detected as being involved in BLT. The identified QTLs were significantly associated with reduced blast disease damage and accounted for 80 % of the phenotypic variation in the disease index. The Likelihood Ratio Statistics (LRS) for the association of a trait with a locus at p ≤ 0.05 ranged from 4.1 to 25.6. The additive effect of a single QTL ranged from −10.72 to 9.9. The six significant QTLs with LOD scores of >3 accounted for approx. 65 % of the phenotypic variance in mean BLD, BLT and % DLA across experiments. The map positions of the QTLs in the rice genetic linkage map were on eight of the 12 rice chromosomes, and no QTLs mapped to chromosomes 4, 7, 8, or 9 (Fig. 4). Interval mapping analysis was performed using QTX software for the three different traits and revealed six putative QTLs on chromosomes 1, 2, 5, 6 and 11. This finding confirmed the single-marker analysis result. These loci were significantly involved in resistance to blast pathotype P7.2. The likelihood ratio values for these regions were above 15. The corresponding marker intervals on these chromosomes were RM428-RM1, RM208-RM166, RM413-RM13, RM8225-RM136, RM5961-RM21, and RM1233-RM6293 (Table 3). Q-Gene detected six putative QTLs that were above the threshold logarithm of the odds of 3.0 and were associated with significant reductions in BLD, BLT and % DLA against pathotype P7.2. In a permutation analysis using the interval mapping method, the LOD scores after 1,000 iterations for the six putative QTLs detected on chromosomes 1, 2, 5, 6 and 11 were found to be above the 5 % threshold level.

All genomic regions containing QTLs for different leaf blast traits were identified using 2 mapping methods. The regions identified by Single Marker Analysis (SMA) or Simple Interval Mapping (SIM) using QTX software were also identified by Q-Gene software. Identical SMA and SIM analysis results were obtained using both software. The QTLs for three different traits in a particular genomic region (known as QTL clusters) with the same marker intervals were considered to be co-localised QTLs. The map locations of six putative QTLs that were detected by Q-gene package ver. 4.2 are shown in Supplementary Fig. 1. In a composite interval mapping analysis with Q-Gene, three QTLs, (qRBr-1.2, qRBr-11.1 and qRBr-11.2) were detected on Ch01 and Ch11. These regions contained the flanking markers RM428-RM1, RM5961-RM21 and RM1233-RM6293, which had significant LOD scores for the QTLs of 3.0, 4.5 and 4.81, respectively (Fig. 5). This result led to an increased resolution of the QTL locations on Ch01 and Ch11. The LOD peaks for the remaining QTLs that were detected by the composite interval mapping analysis were below the threshold level.
https://static-content.springer.com/image/art%3A10.1007%2Fs10529-012-1130-1/MediaObjects/10529_2012_1130_Fig5_HTML.gif
Fig. 5

Composite interval mapping(CIM) in Chromosomes 1, 2, 5, 6 and 11 for blast pathotype P7.2 for BLD, BLT and % DLA traits in F3 families of Pongsu Seribu 2 × Mahsuri using Q gene

Epistatic QTLs (E-QTLs)

Epistatsis can result in resistance to a broad spectrum of blast isolates. For the epistatic interaction analysis (EIA), the Map Manager QTX programme was used to determine the locus combinations representing interacting QTLs for two loci. A number of significant loci combinations for the three different quantitative traits evaluated were identified for two loci interactions. Eight E-QTLs were detected for BLD, 14 E-QTLs were detected for BLT and 11 E-QTLs were detected for % DLA (Table 4). The Likelihood Ratio Statistics (LRS) for a two loci interaction at p ≤ 0.01 ranged from 20.5 to 53.9. The phenotypic variation explained (PV) by two QTL combinations ranged from 0.9–23.6 % for BLD, 0.3–23.8 % for BLT and 0.8–24.1 for % DLA.
Table 4

Epistatic QTLs (E-QTLs) for blast resistance traits at two loci identified with Map Manager QTX program

Traits

Two loci interactions

 

No. of QTLs

Range of LRS

Range of PV (R2)

BLD

8

35.4–53.9

0.9–23.6

BLT

14

20.5–47.2

0.3–23.8

% DLA

11

20.5–52.1

0.8–24.1

BLD blast lesion degree, BLT blast lesion type, %DLA percentage of disease leaf area due to pathotype P7.2, LRS likelihood ratio statistic, PV phenotypic variance

Discussion

The use of molecular markers allows genetic dissection in detail and the precise localisation of QTLs on rice chromosomes. In the present study, the linkage map with 63 SSR markers created from the F3 lines derived from the cross between Pongsu Seribu 2 and Mahsuri was used to locate QTLs involved in field resistance to rice blast. These QTLs were on chromosomes 1, 2, 3, 5, 6, 10, 11 and 12. Six putative and 7 suggestive QTLs were detected as responsible for the reduction of BLD, BLT and % DLA against pathotype P7.2 by SMA and IM analysis. To ensure the reliability of QTL recognition, Map Manager QTX and Q-Gene were employed to identify main-effect QTLs (M-QTLs). The genomic locations of M-QTLs were similarly identified with both programmes; therefore, these QTLs may be considered as reliable QTLs. Based on a single-marker analysis, most of these markers were associated with one or more of the resistance traits measured here. The proportions of the total phenotypic variance explained by the significant QTLs were found to be approx. 65 % for each trait. Three M-QTLs on chromosomes 2, 6 and 11 (qRBr-2.1, qRBr-6.2 and qRBr-11.2) accounted for 12, 12 and 13 % of the total phenotypic variance, respectively. This finding indicates that these three genes play a major role in the expression of field resistance to blast in Pongsu Seribu 2 and in lines with resistance similar to or higher than Pongsu Seribu 2.

Although some M-QTLs were identified for three traits studied, the majority of the identified QTLs did not display high phenotypic variance. This suggests that the multigenic inheritance of field resistance was confirmed by the QTL analysis. The highly polygenic nature of the trait-analysed QTLs with lower phenotypic variation is expected. These small-effect QTLs should be considered when introgressing and pyramiding resistance QTLs. The QTLs contributing to a high level of trait variance or with major phenotypic effects in blast resistance identified in Pongsu Seribu 2 might be race-specific. Other QTLs with minor phenotypic effects conferring resistance to blast could be nonspecific. The additive effect of the three QTLs identified for pathotype P7.2, qRBr-5.1, qRBr-6.1 and qRBr-6.2, were positive. The additive effect of the 10 QTLs qRBr-1.1, qRBr-1.2, qRBr-2.1, qRBr-3.1, qRBr-10.1, qRBr-10.2, qRBr-11.1, qRBr-11.2, qRBr-11.3 and qRBr-12.1 was negative. Most of the QTLs from the F3 population showed a favourable effect of having trait-improving alleles from Pongsu Seribu 2. In previous studies, a different mapping population was used to identify a partial resistance QTL on the rice chromosome (Cho et al. 2008; Lopez-Gerena 2006; Zenbayashi et al. 2002; Fjellstrom et al. 2004). Some of the QTLs identified in the present study were mapped to regions previously described as containing Pi genes, and some QTLs mapped to regions with no reported Pi genes.

In this study, one of the main effect QTLs for blast resistance, qRBr2.1, was detected on chromosome 2. This QTL was flanked by markers RM208–RM166 at a distance of 1.2 cM from RM208, with an LOD score of 5.0. It accounted for 12 % of the phenotypic variation in the F3 population. In previous studies, a single dominant blast resistance gene named Pi-g(t) had been mapped to the long arm of chromosome 2 in Chinese indica rice and was flanked by similar markers (Zhou et al. 2004). In this study, one QTL, named qRBr-3.1, was located on chromosome 3 at RM168. This QTL was identified as close to a previously mapped partial resistance QTL (Sirithunya et al. 2002) and accounted for 4 % of the leaf blast severity variation in the F3 population. The two QTLs qRBr-6.1 and qRBr-6.2 were mapped to chromosome 6 and had an effect on all three resistance traits. The corresponding marker intervals for qRBr-6.1 were RM8225-RM136 and had a distance of 0.2 cM to the RM8225 marker. This significant QTL accounted for 12 % of the phenotypic variation. In contrast, qRBr-6.2 at RM6836 had a map distance of 0.4 cM to this marker and exhibited a lower percentage of observed phenotypic variance (PV values of 3 %).

In previous studies, major blast resistance genes have been named on rice chromosome 6, such as Pi-8, Pi-9(t), Pi-13(t) and Pi-22(t) (Pan et al. 1996), Pi-27 (Sallaud et al. 2003), Pi-3(t) (Mackill and Bonman 1992) and Pi-tq1 (Tabien et al. 2000). Our detected QTLs on chromosome 6, qRBr 6.1 and qRBr 6.2, were localised near some of these loci. Ashkani et al. (2011, 2012) reported that the RM8225 marker is suitable for the marker-assisted selection of a Pi gene conferring blast resistance in Pongsu Seribu 2. Several QTLs and Pi genes, such as Pik genes, have been mapped to a genetic region on chromosome 11 (Fjellstrom et al. 2004; Sharma et al. 2005; Lopez-Gerena 2006; Cho et al. 2008). The three resistance QTLs, qRBr-11.1, qRBr-11.2, qRBr-11.3 mapped to this region in our study. One of these QTLs, qRBr-11.2, accounted for 13 % of the phenotypic variance and had the largest effect on the phenotypic variation in this experiment. The QTLs qRBr-11.1, with the corresponding marker interval RM5961-RM21, and qRBr-11.3, localised at RM6293, explained less of the observed phenotypic variance, with values of 9 and 5 %, respectively. The SSR marker RM1233 was linked with the resistance QTL qRBr-11.2. This marker has already been suggested for the selection of resistance sources carrying the resistance pi gene (Ashkani et al. 2011 and 2012); these QTLs might therefore be the major genes or candidate genes of Pik. An allelic test should be conducted to clarify the allelic relationship between the gene and the QTL. The remaining QTLs (qRBr-1.1, qRBr-1.2, qRBr-5.1, qRBr-10.1, qRBr-10.2 and qRBr-12.1) that were identified in our study were novel. Among these, qRBr-1.1 was mapped to the RM462 marker on chromosome 1.

This QTL accounted for a phenotypic variance (PV%) above 3 % in two resistance traits. The QTL qRBr-1.2 was also mapped to chromosome 1, between the RM428-RM1 markers. This QTL had an effect on three resistance traits and explained 12 % of the phenotypic variance. The qRBr-5.1 locus on chromosome 5 was mapped between markers RM413-RM13. This QTL had an effect on all three resistance traits, and the PV% for this QTL was 11 %. The QTLs qRBr-10.1 at RM244 and qRBr-10.2 at RM304 had an effect on three and one of the resistance traits, respectively, with low PV% (<3 %). The QTL qRBr-12.1 at RM101 had an effect on all three resistance traits, explaining 6 % of the phenotypic variance. QTL analysis using various genetic backgrounds and populations allowed us to identify several QTLs that shared the same chromosomal locus, suggesting that common loci for differentiation among rice varieties exists and could be valuable. In the present study, an epistatic interaction analysis (EIA) undertaken with Map Manager QTX identified several epistemic QTLs. Using this software, two loci interactions were identified for three different traits: BLD, BLT and % DLA. Fourteen interactions were identified for the BLT trait, 11 interactions for % DLA and 8 interactions for BLD. QTL interactions in the F3 segregating population revealed that 8 interacting QTLs existed for BLD traits, and these traits were also identified for BLT and % DLA. The remaining interacting QTLs were diverse for the three different traits. This diversity may be attributed to the variability of traits across parents. The number of E-QTLs was greater than the number of main-effect QTLs. The likelihood ratio statistics (LRS) and the percentage of phenotypic variance attributed to these QTLs were comparatively higher than the QTLs identified by SMA and IM mapping methods using QTX and Q-GENE software.

In the present study, genotypic and phenotypic data were collected for three different traits and analysed in detail for the identification of main-effect QTLs (M-QTLs) and epistatic QTLs (E-QTLs). This was accomplished using the Map Manager QTX and Q-GENE programmes. A total of 13 QTLs, distributed on rice chromosomes 1, 2, 3, 5, 6, 10, 11 and 12, were identified using the SMA and interval mapping (IM) methods. A single resistance QTL accounted for 3–13 % of the observed phenotypic variation and accounted for approx. 80 % of the total phenotypic variation within the F3 lines. Although some QTLs detected in this study were located on the same chromosome as reported previously, further investigation of the different molecular markers being used is necessary. The results suggest that leaf blast tolerance in rice and in the Pongsu Seribu 2 cultivar is governed by the combined effects of multiple loci with major and minor effects and QTL × QTL. Identification of the putative QTLs with LOD scores >3 (qRBr-1.2, qRBr-5.1, qRBr-6.1, qRBr-11.1 and qRBr-11.2 on chromosomes 1, 2, 5 and 11) is valuable and could be investigated further for the development of blast resistance markers for resistance selection. Because resistance genes/stable QTLs on some chromosomes contributed by Pongsu Seribu 2 are highly likely to confer durable resistance, introgression of these genes into sensitive varieties through marker-assisted selection could fulfil the objective of breeding for resistance to field blast.

Acknowledgments

This research activity was funded by the Long term Research Grants Scheme (LRGS), Rice Food Security Project, Ministry of Higher Education, Malaysia. We would like to express our sincere thanks to the Malaysian Nuclear Agency for providing research facilities. We are also grateful to the Malaysian Rice Research Centre, MARDI, for supplying pathogen isolates and providing greenhouse facilities.

Supplementary material

10529_2012_1130_MOESM1_ESM.doc (70 kb)
Supplementary material 1 (DOC 70 kb)

Copyright information

© Springer Science+Business Media Dordrecht 2013