Novel pleiotropic loci controlling panicle architecture across environments in japonica rice (Oryza sativa L.)

JOURNAL OF

GENETICS AND GENOMICS J. Genet. Genomics 37 (2010) 533544

www.jgenetgenomics.org

Novel pleiotropic loci controlling panicle architecture across environments in japonica rice (Oryza sativa L.) Yuan Guo a, b, Delin Hong a, * a

State Key Laboratory of Crop Genetics and Germplasm Enhancement, Nanjing Agricultural University, Nanjing 210095, China b Institute of Bast Fiber Crops, Chinese Academy of Agricultural Sciences, Changsha 410205, China Received for publication 21 November 2009; revised 11 March 2010; accepted 23 April 2010

Abstract To identify quantitative trait loci (QTLs) controlling panicle architecture in japonica rice, a genetic map was constructed based on simple sequence repeat (SSR) markers and 254 recombinant inbred lines (RILs) derived from a cross between cultivars Xiushui 79 and C Bao. Seven panicle traits were investigated under three environments. Single marker analysis indicated that a total of 27 SSR markers were highly associated with panicle traits in all the three environments. Percentage of phenotypic variation explained by single locus varied from 2% to 35%. Based on the mixed linear model, a total of 40 additive QTLs for seven panicle traits were detected by composite interval mapping, explaining 1.2%35% of phenotypic variation. Among the 9 QTLs with more than 10% of explained phenotypic variation, two QTLs were for the number of primary branches per panicle (NPB), two for panicle length (PL), two for spikelet density (SD), one for the number of secondary branches per panicle (NSB), one for secondary branch distribution density (SBD), and one for the number of spikelets per panicle (NS), respectively. qPLSD-9-1 and qPLSD-9-2 were novel pleiotropic loci, showing effects on PL and SD simultaneously. qPLSD-9-1 explained 34.7% of the phenotypic variation for PL and 25.4% of the phenotypic variation for SD, respectively. qPLSD-9-2 explained 34.9% and 24.4% of the phenotypic variation for PL and SD, respectively. The C Bao alleles at the both QTLs showed positive effects on PL, and the Xiushui 79 alleles at the both QTLs showed positive effects on SD. Genetic variation of panicle traits are mainly attributed to additive effects. QTL × environment interactions were not significant for additive QTLs and additive × additive QTL pairs. Keywords: japonica rice; simple sequence repeat (SSR); panicle traits; quantitative trait locus (QTL); pleiotropic locus

Introduction The panicle of rice is an organ of photosynthesis as well as storage of photosynthetic product (Hirota et al., 1990). Abbreviations: NPB, number of primary branches per panicle; NSB, number of secondary branches per panicle; SBD, secondary branch distribution density; PL, panicle length; NS, number of spikelets per panicle; NFG, number of filled grains per panicle; SD, spikelet density. * Corresponding author. Tel & Fax: +86-25-8439 6626. E-mail address: [email protected] DOI: 10.1016/S1673-8527(09)60073-4

Its architecture is relevant to not only yield and quality but also diseases resistance. Generally, cultivars with long and curve panicle showed elite quality but low yield in japonica rice, such as Koshihikari and Nipponbare, introduced from Japan, and series of Yanjing cultivars bred in Yan-cheng, Jiangsu Province. While cultivars with short and erect panicle usually showed high yield but poor quality. Zhendao88, bred in Zhenjiang, Jiangsu Province, Liaojing 5 and Shennong 611, bred in Liaoning Province were examples of this type panicle architecture (Lü et al., 1997; Yang et al., 1999; Jiang et al., 2007). Cultivars with high

534

Yuan Guo et al. / Journal of Genetics and Genomics 37 (2010) 533544

spikelet density are generally more resistant to lodging, but more susceptible to fasle smut diseases in panicle than cultivars with low spikelet density in japonica rice (Liu et al., 2009). Up to now, quantitative trait loci (QTLs) contributing to rice panicle architecture have been intensively explored and mapped to their residing chromosomal regions using various mapping populations (Lin et al., 1996; Xiao et al., 1996; Zhuang et al., 1997; Xiong et al., 1999; Yamagishi et al., 2002; Li et al., 2006). Three QTLs controlling the number of spikelets per panicle, gpa7, SPP1 and qSPP7, have been fine mapped (Tian et al., 2006; Xing et al., 2008; Liu et al., 2009). Three genes controlling the number of spikelets per panicle, Gn1a, Ghd7 and DEP1, have been cloned by a map-based strategy (Ashikari et al., 2005; Xue et al., 2008; Huang et al., 2009). Gn1a encodes cytokinin oxidase/dehydrogenase (OsCKX2), an enzyme that degrades the phytohormone cytokinin (Ashikari et al., 2005). Ghd7 encodes a CCT domain protein (Xue et al., 2008). DEP1 encodes a unknown PEBP like domain protein (Huang et al., 2009). Xiushui 79 is a japonica cultivar with erect and short panicle. C Bao is a japonica restorer line with curve and long panicle. The panicles of C Bao look like the paws of chicken after spikelets are filled with photosynthetic products. To understand the genetic background of panicle characteristics in japonica rice, we developed a population of recombinant inbred lines (RILs) from a cross between these two cultivars. Our previous studies demonstrated that the number of spikelet per panicle, panicle angle, the number of primary branches per panicle and the number of secondary branches per panicle were controlled by major genes plus polygenes by using the method of genetic segregation analysis (Liu et al., 2005; Chen et al., 2006; Guo et al., 2008). In the present study, we constructed a linkage map of simple sequence repeat markers, and analyzed phenotypic data of 7 panicle traits in the RIL population under three environments, to identify QTLs controlling the traits and quantify their genetic effects.

combinant inbred lines (RILs) derived from a cross between two japonica rice cultivars Xiushui 79 and C Bao by single-seed descend. The female parent, Xiushui 79, bred by Jiaxing Institute of Agricultural Sciences, Zhejiang Province, China, has 95 cm plant height, erect and compact spike. The male parent C Bao, bred by Anhui Academy of Agricultural Sciences, Anhui Province, China, is a japonica restorer line with 100 cm plant height, curve and large spike (Fig. 1). Both Xiushui 79 and C Bao headed at middle-August when sowed in middle-May in Nanjing, China. In 2006, The RILs (F8:9) and the two parents were grown in paddy field at Jiangpu Experiment Station, Nanjing Agricultural University, China (E1 environment). In the rice growing season of 2008, the RILs (F10:11) and the two parents were grown in paddy field at Jiangpu Experiment Station, Nanjing Agricultural University (E2 environment) and Foundation Seed Production Farm in Sihong

Materials and methods Experimental population, field experiments and phenotypic measurements The population used in this study consisted of 254 re-

Fig. 1. Plant type and panicle type of Xiushui 79 and C Bao in japonica rice. A: Xiushui 79. B: C Bao.

Yuan Guo et al. / Journal of Genetics and Genomics 37 (2010) 533544

County in Jiangsu Province (E3 environment), respectively. In each environment, parents and RILs were planted in two replications and each line was planted in two rows, each having 8 single-seedling hills, in a density of 17 cm × 20 cm. Regular field management was carried out. At maturity stage, the number of primary branches per panicle (NPB), the number of secondary branches per panicle (NSB), panicle length (PL), the number of spikelets per panicle (NS), and the number of filled grains per panicle (NFG) on the main stems were determined on five different plants for each RIL and parent. Secondary branch distribution density (SBD) and spikelet density (SD) were calculated by using the following formulas: SBD = NSB/NPB; SD = NS/PL. Mean values over two replications in each environment were used for analysis.

Polymorphism detection and single marker analysis On the basis of our initial SSR linkage map (Guo et al., 2009), 97 pairs of new SSR primer (915 pairs of SSR primer in all) were screened for polymorphism between the two parents using total DNA as template. The 254 RIL plants were genotyped using a selected set of polymorphic SSR markers. Single marker analysis was conducted to detect associations between traits and markers using Windows QTL Cartographer Ver. 2.5 (http://statgen.ncsu.edu/ qtlcart/WQTLCart.htm) (Zeng, 1994).

Genetic linkage map construction and QTL analysis Genomic DNA of two parents and 254 RILs used for linkage map constructing were extracted from plants in E1 environment. The genetic linkage map was constructed using the software MAPMAKER/EXP Ver. 3.0 (Lander, et al., 1987) and mapping was done using MapDraw (Liu and Meng, 2003). Additive QTLs, additive × additive QTL pairs and QTL × environment (Q × E) interactions were detected using mixed linear model approaches and with computer program QTLNetwork2.0 (Yang, 2007, 2008). A threshold probability of P = 0.005 was used. QTL nomenclature followed the propositions of the Rice Genetics Cooperative (McCouch et al., 1997).

535

Results Phenotypic difference between the parents and variations in the RIL population for the 7 panicle traits Phenotypic values of all the 7 traits investigated in C Bao were consistently larger than those in Xiushui 79 in each of the three environments. Both parents had higher values in E3 than in E1 and E2, except for NSB and SBD of Xiushui 79. By t-test, significant difference between the two parents was found in traits of NSB, SBD, PL, NS and NFG in all the three environments, and in traits of SD in E1 and E2. Phenotypic values of all the 7 traits investigated in RIL population showed a continuous distribution, and transgressive segregation in either direction was observed (Fig. 2).

Polymorphism between two parental lines and association between markers and panicle traits Of 915 SSR primers used in this study, 105 pairs of SSR primers showed polymorphism between two parental lines. Polymorphic ratio was 11.5%. Among the 12 chromosomes, chromosome 7 showed the largest polymorphic ratio (16.4%), and chromosome 1 the least (7.8%) (Table 1). Twenty-seven SSR marker loci distributed on chromosomes 1, 4, 6, 7, 8 and 9 were detected as putatively linked to panicle traits in all the three environments with a probability of more than 99.9%. Percentages of phenotypic variation explained by the markers are present in Table 2. As shown in Table 2, markers associated with NSB, SBD, NS and NFG were mainly distributed on chromosome 1, markers associated with NPB were mainly distributed on chromosome 8, and markers associated with PL and SD were mainly distributed on chromosome 9. Among the 27 markers detected, some of them associated with more than one trait. RM84, RM462 and RM3453 on chromosome 1 showed association with five traits, i.e., NSB, SBD, NS, NFG and SD. RM80 on chromosome 8 showed association with four traits, i.e., NPB, NSB, NS and NFG. RM6570, RM5652, OSR28, RM5786 and RM201 on chromosome 9 showed association with PL and SD (Table 2). The phenomena that the same makers associated with different traits implied that the molecular mechanisms of these traits are same, and the markers with pleiotropic effects would facilitate the improvement of several traits simultaneously by molecular assisted selection.

536

Yuan Guo et al. / Journal of Genetics and Genomics 37 (2010) 533544

Fig. 2. Frequency distribution of panicle traits in RIL population derived from Xiushui 79/C Bao in japonica rice. X1 and C1, Xiushui 79 and C Bao under E1 enviroment, respectively; X2 and C2, Xiushui 79 and C Bao under E2 enviroment, respectively; X3 and C3, Xiushui 79 and C Bao under E3 enviroment, respectively.

Yuan Guo et al. / Journal of Genetics and Genomics 37 (2010) 533544

537

Table 1 SSR polymorphism between Xiushui 79 and C Bao, number of markers entered to linkage map and number of distorted markers Polymorphism (%)

Chromosome length (cM)

No. of distorted markers

No. of surveyed SSRs

1

116

9

7.8

7

50.2

6

2

113

14

12.4

13

182.4

4

3

99

10

10.1

9

89.4

7

4

54

6

11.1

4

40.0

4

5

78

9

11.5

7

69.9

1

6

83

8

9.6

6

76.7

1

7

61

10

16.4

9

115.7

7

8

93

12

12.9

12

151.7

7

9

63

10

15.9

9

57.9

3

10

43

4

9.3

4

32.2

0

11

61

9

9.8

6

52.2

1

12

51

7

13.7

5

50.9

2

915

105

11.5

91

969.2

43

Total

No. of markers

No. of SSRs for map construction

Chromosome No.

Table 2 SSR markers highly significantly related to panicle traits in all three environments and percent of the phenotypic variation explained (R2) Trait a

Environment

Marker RM486

RM564

AP22

RM80

RM281

RM264

RM6948

RM433

E1

11.0

12.7

14.5

12.0

12.4

10.9

17.4

07.7

E2

10.1

13.0

10.8

10.1

06.9

05.9

10.6

06.7

E3

11.8

07.6

07.0

08.6

09.2

07.6

11.4

04.5

RM84

RM462

RM3453

RM490

RM562

RM8263

RM80

NPB

NSB E1

20.8

23.2

20.9

07.6

03.8

03.9

07.1

E2

19.1

17.8

17.6

04.7

02.3

02.2

04.3

E3

16.6

17.6

14.4

08.0

03.8

03.0

02.3

RM84

RM462

RM3453

RM346

RM410

SBD E1

22.5

23.7

24.0

03.1

03.4

E2

16.4

15.3

17.1

03.6

01.7

E3

11.4

11.8

11.2

01.7

02.8

RM551

RM6570

RM5652

RM410

RM257

OSR28

RM5786

RM201

E1

05.6

19.6

23.7

21.3

26.0

14.2

13.7

11.4

E2

03.3

27.4

26.3

27.0

32.4

18.3

16.4

13.2

E3

06.7

28.5

26.9

28.5

34.9

20.3

18.9

16.6

RM84

RM462

RM3453

RM490

RM562

RM8263

RM80

E1

09.9

13.2

09.5

05.7

06.7

03.0

07.8

E2

15.0

13.8

12.7

03.8

03.4

04.3

06.8

E3

14.7

14.5

11.2

07.2

03.4

02.8

04.4

RM84

RM462

RM3453

RM490

RM3288

RM8239

RM80

E1

05.4

08.3

06.5

06.4

03.2

04.6

07.6

E2

07.3

06.6

05.9

05.4

02.4

07.1

08.1

E3

07.2

06.5

05.6

06.0

03.2

03.4

04.1

RM84

RM462

RM3453

RM6570

RM5652

OSR28

RM5786

RM201

PL

NS

NFG

SD

a

E1

07.4

08.8

04.6

21.7

19.8

12.7

14.1

11.0

E2

12.2

11.2

08.0

24.5

19.6

19.2

18.1

14.9

E3

10.6

10.0

06.0

18.0

20.3

15.0

15.4

12.7

NPB, number of primary branches per panicle; NSB, number of secondary branches per panicle; SBD, secondary branch distribution density; PL, panicle length; NS, number of spikelets per panicle; NFG, number of filled grains per panicle; SD, spikelet density.

538

Yuan Guo et al. / Journal of Genetics and Genomics 37 (2010) 533544

Construction of SSR linkage map based on the RIL population A genetic map consisting of 91 SSR markers on 19 linkage groups was constructed by using Mapmaker software version 3.0. The remaining 14 SSR markers could not be mapped to any linkage group. The map was 969 cM long, with an average distance of 10.6 cM between adjacent markers (Fig. 3). By F2 test, 43 (47.3%)

markers showed distortion segregation from the expected allelic frequency of 1:1 (Xiushui 79 allele : C Bao allele) in the RIL population (P < 0.05). The number of distortion segregation markers varied among the chromosomes, ranging from 1 marker on chromosome 5, 6 and 11 to 7 markers on chromosome 8 (Table 1). The total frequency of Xiushui 79 allele was 0.537, and C Bao allele was 0.463 in the RIL population, fitting to the expected ratio of 1:1.

Fig. 3. SSR linkage map based on RILs derived from Xiushui79/C Bao in japonica rice and chromosomal location of QTLs for panicle traits. NPB, number of primary branches per panicle; NSB, number secondary branches number per panicle; SBD, secondary branch distribution density; PL, panicle length; NS, number of spikelets per panicle; NFG, number of filled grains per panicle; SD, spikelet density.

Yuan Guo et al. / Journal of Genetics and Genomics 37 (2010) 533544

Additive QTLs, additive × additive QTL pairs and QTL × environment interactions identification for panicle traits A total of 40 additive QTLs for 7 traits were identified, and located on all of the chromosomes except chromosome 2, 5 and 12. Four QTLs were for NPB, 6 QTLs for NSB, 5 QTLs for SBD, 5 QTLs for PL, 7 QTLs for NS, 7 QTLs

539

for NFG and 6 QTLs for SD, respectively. QTL × environment interactions were not significant for additive QTLs (Table 3). Twenty-one pairs of additive×additive QTL interaction were detected for the 7 traits totally, including 6 interactions between main effect loci and 15 interactions between two loci having no main effect at the single-locus level. QTL × environment interactions were not significant for additive × additive QTL pairs (Table 4).

Table 3 Estimated position, flanking markers, additive effects and percentage of the phenotypic variation explained (R2) of the QTLs for the panicle traits

a

Traits a

QTL

Marker interval b

NPB

qNPB-1 qNPB-7 qNPB-8 qNPB-10

RM486-RM265 RM11-RM346 RM80-RM281 RM5629-RM171

NSB

qNSB-1 qNSB-4 qNSB-6 qNSB-7 qNSB-8 qNSB-10

SBD

Distance(cM) c

LOD

A effect

P-value

R2 (%)

1.0 2.0 4.6 0.3

12.2 6.1 17.2 9.4

0.66 0.37 0.49 0.34

< 0.0001 < 0.0001 < 0.0001 < 0.0001

10.1 2.4 10.0 3.0

RM84-RM462 RM3288-RM349 RM162-RM5753 RM180 RM1235-RM331 RM5629-RM171

0.8 9.0 21.8 0.0 9.0 2.0

63.4 10.6 6.6 9.2 7.5 6.8

4.43 1.68 3.25 1.72 2.27 1.75

< 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001

17.8 2.2 2.5 3.4 2.0 2.6

qSBD-1 qSBD-3 qSBD-4 qSBD-7 qSBD-11

RM84-RM462 RM545-RM3766 RM3288-RM349 RM11-RM346 RM5349-RM206

0.8 3.0 9.0 1.0 3.0

63.8 7.1 7.6 6.9 9.6

0.27 0.09 0.10 0.15 0.12

< 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001

16.0 1.8 1.3 3.4 2.6

PL

qPL-4-1 qPL-4-2 qPL-6 qPL-9-1 qPL-9-2

RM551 RM3288-RM349 RM5314-RM454 RM6570-RM5652 RM5652-RM410

0.0 7.0 0.9 8.0 7.0

12.8 9.1 22.1 112.6 115.8

0.56 0.59 0.81 1.21

< 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001

5.6 3.6 1.2 34.7 34.9

NS

qNS-1-1 qNS-1-2 qNS-4 qNS-6 qNS-8-1 qNS-8-2 qNS-10

RM84-RM462 RM486-RM265 RM3288-RM349 RM8239-RM5314 RM152-RM1235 RM80-RM281 RM5629-RM171

0.8 0.4 5.0 1.0 1.1 2.0 0.3

37.7 8.3 14.2 6.9 8.1 10.6 9.9

15.06 7.33 12.02 5.86 7.92 6.71 7.41

< 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001

12.3 2.5 2.4 3.7 2.7 4.6 3.4

NFG

qNFG-1-1 qNFG-1-2 qNFG-4 qNFG-6 qNFG-7 qNFG-8-1 qNFG-8-2

RM84-RM462 RM486-RM265 RM3288-RM349 RM8239 RM180 RM152-RM1235 RM80

0.8 0.4 10.0 0.0 0.0 3.0 0.0

16.7 7.0 11.0 9.7 10.6 8.5 10.2

8.02 7.47 14.25 6.80 6.44 7.48 7.20

< 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001

5.9 3.4 3.8 3.9 3.1 3.3 5.5

SD

qSD-1 qSD-6 qSD-7 qSD-8 qSD-9-1

RM84-RM462 RM5753 RM418-RM11 RM1235-RM331 RM6570-RM5652

0.8 0.0 2.1 15.5 9.0

46.1 7.9 7.1 16.3 62.3

0.87 0.46 0.28 0.67 0.95

< 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001

9.3 4.3 1.4 1.5 25.4

qSD-9-2

RM5652-RM410

8.0

65.5

0.61

< 0.0001

24.6

1.35

See the footnote a of Table 2 for the full names of the traits. b Bold letters indicate the nearest marker. c Distance from the nearest marker to putative QTL.

540

Yuan Guo et al. / Journal of Genetics and Genomics 37 (2010) 533544

Table 4 Estimated position, flanking markers, additive × additive effects and percentage of the phenotypic variation explained of the additive × additive interaction QTL pairs for panicle traits Traits a QTL_i

Marker interval b

Distance(cM) c QTL_j

Marker interval

NPB

5-5

RM161-RM7473

12.0

8-11

RM6948-RM433

0.1

15.4

5-6

RM7473-RM480

2.9

6-4

RM162-RM5753

20.0

9.1

NSB

SBD

P-value

H2(AA)(%)

0.43

< 0.0001

4.3

0.73

< 0.0001

1.5

Epistasis (AA)

5-6

RM7473-RM480

2.9

8-11

RM6948-RM433

0.1

12.0

0.26

< 0.0001

4.2

8-4

RM3383-RM72

1.2

9-5

RM410-RM257

0.8

12.4

0.39

< 0.0001

2.5

qNSB-7

RM180

0.0

qNSB-10

RM5629-RM171

2.0

1.00

0 0.0005

1.3

qNSB-8

RM1235-RM331

9.0

qNSB-10

RM5629-RM171

2.0

1.66

< 0.0001

1.0

1-1

RM486-RM265

0.0

5-5

RM161-RM7473

11.0

9.1

2.13

< 0.0001

2.2

2-10

RM2127-RM48

0.0

11-2

RM7120-RM287

0.1

6.3

0.92

0 0.0009

1.2

3-4

RM3766-RM5639

0.8

8-8

RM80-RM281

0.0

6.4

1.13

< 0.0001

1.4

2-5

RM1313-RM327

8.0

11-2

RM7120-RM287

0.0

10.5

0.10

< 0.0001

2.2

3-6

RM218-RM232

1.0

8-11

RM6948-RM433

0.0

7.5

0.08

< 0.0001

2.0

2-9

RM525-RM2127

2.9

7-3

RM180-RM214

0.0

9.1

0.50

< 0.0001

1.5

5-5

RM161-RM7473

0.0

10-3

RM5629-RM171

0.0

11.2

0.64

< 0.0001

2.8

NS

qNS-1-1

RM84-RM462

0.8

qNS-4

RM3288-RM349

5.0

5.80

< 0.0001

0.8

NFG

qNFG-1-1 RM84-RM462

0.8

qNFG-4

RM3288-RM349

10.0

7.61

< 0.0001

1.6

PL

SD

a

Distance(cM) LOD

2-4

RM145-RM1313

7-1

RM82-RM125

qSD-1

RM84-RM462

qSD-7

RM418-RM11

0.0

4-1

RM551-RM335

0.3

9.7

5.82

< 0.0001

2.2

13.6

8-9

RM281-RM264

0.7

9.1

9.55

< 0.0001

2.3

0.8

qSD-9-1

RM6570-RM5652

9.0

0.27

0 0.0005

1.0

2.1

qSD-8

RM1235-RM331

15.5

0.29

0 0.0060

0.6

2-2

RM5356-RM1358

1.6

8-7

RM22957-RM80

20.2

7.3

0.63

< 0.0001

1.5

3-3

RM545-RM3766

0.0

3-8

RM7097-RM448

6.7

7.0

0.44

< 0.0001

2.1

See the footnote a of Table 2 for the full names of the traits. b Bold letters indicate the nearest marker. c Distance from the nearest marker to putative QTL.

Among the 4 additive QTLs detected for NPB, qNPB-1 and qNPB-8 explained 10.1% and 10.0% of the phenotypic variation, respectively. The alleles from Xiushui 79 at these two loci showed positive effects, and values of additive effects were 0.7 in qNPB-1 and 0.5 in qNPB-8. Four pairs of additive×additive QTL interaction between two loci having no main effect were detected for NPB, accounting for 1.5%4.3% of the phenotypic variation in the population. All of 4 interactions had positive effects on NPB. Among the 6 additive QTLs detected for NSB, qNSB-1 explained 17.8% of the phenotypic variation. The alleles from C Bao at this locus showed positive effects, and value of additive effect was 4. Two pairs of additive×additive QTL interaction between main effect loci and 3 interactions between two loci having no main effect were detected for NSB, accounting for 1.0%2.2% of the phenotypic variation in the population. All of 4 interactions had negative effects on NSB.

Among the 5 additive QTLs detected for SBD, qSBD-1 explained 16.0% of the phenotypic variation. The effects of allele from C Bao at this locus were positive, and value of additive effect was 0.3. Two pairs of additive×additive QTL interaction between two loci having no main effect were detected for SBD, accounting for 2.0%2.2% of the phenotypic variation in the population. One interaction (2-5/11-2) had positive effects on SBD, while, another interaction (3-6/8-11) had negative effects on SBD. Among the 5 additive QTLs detected for PL, qPL-9-1 and qPL-9-2 explained 34.7% and 34.9% of the phenotypic variation, respectively. The effects of allele from C Bao at these two loci were positive, and values of additive effects were 1.2 cm in qPL-9-1 and 1.4 cm in qPL-9-2. Two pairs of additive × additive QTL interaction between two loci having no main effect were detected for PL, accounting for 1.5%2.8% of the phenotypic variation in the population. Both two interactions had negative effects on PL.

Yuan Guo et al. / Journal of Genetics and Genomics 37 (2010) 533544

Among the 7 additive QTLs detected for NS, qNS-1-1 explained 12.3% of the phenotypic variation. The positive allele came from C Bao at this locus, and value of additive effect was 15. One pair of additive×additive QTL interaction between main effect loci were detected for NS, accounting for 0.8% of the phenotypic variation and had negative effects on NS. Among the 7 additive QTLs detected for NFG, qNFG-1-1 explained 5.9% of the phenotypic variation. The positive allele came from C Bao at this locus, and value of additive effect was 8. One pair of additive × additive QTL interaction between main effect loci and 2 interactions between two loci having no main effect were detected for NFG, accounting for 1.6%2.3% of the phenotypic variation in the population. Two interactions (qNFG-1-1/qNFG-4 and 2-6/4-1) had positive effects on NFG, while, one interaction (7-1/8-9) had negative effects on NFG. Among the 6 additive QTLs detected for spikelet density, qSD-9-1 and qSD-9-2 explained 25.4% and 24.6% of the phenotypic variation, respectively. The positive allele came from C Bao at these two loci, and values of additive effects were 1.0 in qSD-9-1 and 0.6 in qSD-9-2. Totally four pairs of additive × additive QTL interaction were detected for SD, two interactions were between main effect loci and other two interactions were between loci having

541

no main effect, accounting for 0.6%2.1% of the phenotypic variation. Among the 4 interactions, one interaction (2-2/8-7) had positive effects on SD, while, the remaining three interactions (qSD-1/qSD-9-1, qSD-7/qSD-8 and 3-3/3-8) had negative effects on SD.

Pleiotropic loci and correlation among traits It is found that QTLs which explained more than 10% of the phenotypic variation were mainly clustered on three intervals of rice chromosomes. In interval RM84-RM462 on chromosome 1, QTLs conditioning NSB, SBD and NS simultaneously were detected. In interval RM6570-RM5652 and RM5652-RM410 on chromosome 9, QTLs conditioning simultaneously PL and SD were detected. From Table 5, it can be seen that traits controlled by cluster loci showed highly significant correlations between them. These results indicated that trait correlation could be attributed to the effect of pleiotropy. One pleiotropic locus detected in the interval of RM6570-RM5652 on chromosome 9 explained 34.7% and 25.4% of the phenotypic variation for PL and SD, respectively, and was named as qPLSD-9-1. Another pleiotropic locus detected in the interval of RM5652-RM410 on chromosome 9 explained 34.9% and 24.6% of the phenotypic variation for PL and SD, respectively, and was named as qPLSD-9-2.

Table 5 Coefficients of correlation between panicle traits in 254 RILs derived from Xiushui 79/C Bao in japonica rice Environment

Trait a

NSB

SBD

PL

NS

NFG

E1

NPB

0.51**

0.003

0.06

0.55**

0.42**

E2

0.52**

0.1

E3

0.49**

0.06

0.46**

0.27**

0.64**

0.68**

0.41**

0.03

0.58**

0.51**

0.45**

0.85**

0.21**

0.75**

0.63**

0.53**

E2

0.89**

0.23**

0.92**

0.79**

0.68**

E3

0.83**

0.12

0.89**

0.72**

0.64**

E1

E1

NSB

0.20**

0.54**

0.48**

0.36**

E2

0.13*

0.75**

0.57**

0.58**

E3

0.13*

0.65**

0.5**

0.44**

SBD

0.25**

0.33**

0.37**

E2

0.3**

0.42**

0.38**

E3

0.17*

0.34**

0.56**

E1

E1

PL

0.83**

0.7**

E2

0.87**

0.68**

E3

0.85**

0.68**

E1

a

SD

NS

NFG

0.55**

E2

0.49**

E3

0.42**

See the footnote a of Table 2 for the full names of the traits. * and ** represent significant at 5% and 1% probability level, respectively.

542

Yuan Guo et al. / Journal of Genetics and Genomics 37 (2010) 533544

Discussion SSR marker loci detected by both single marker analysis and composite interval mapping In the present study, markers associated with panicle traits were detected by both single marker analysis and composite interval mapping. Comparing the results of these two methods, we found nine SSR markers detected by single marker analysis were identical with markers linked tightly to QTLs detected by composite interval mapping. Among the 9 marker loci, two marker loci, RM486 and RM281, were associated with NPB, One marker locus, RM462, was associated with NSB and SBD, two marker loci, RM551 and RM5652, were associated with PL, two marker loci, RM462 and RM80, were associated with NS, Four marker loci, RM462, RM3288, RM3289 and RM80, were associated with NFG, and three marker loci, RM462, RM5753 and RM5652, were associated with SD. Percentage of phenotypic variation explained by QTLs linked to the marker loci mentioned above was higher than that of other QTLs. And their Q × E interactions were not significant whatever in different years of the same place or in different places of the same year. These facts suggested that they are reliable marker loci for marker-assisted selection in different environments.

The relationship between the QTLs detected in this paper and reported genes controlling panicle traits Among the 40 additive QTLs detected in this study, nine QTLs, i.e., qNPB-1, qNPB-8, qNSB-1, qSBD-1, qPL-9-1, qPL-9-2, qNS-1-1, qSD-9-1 and qSD-9-2, each explained more than 10% of phenotypic variation. In interval RM84-RM462 of chromosome 1, three QTLs (qNSB-1, qSBD-1 and qNS-1-1) were detected. The interval RM84-RM462 was located near the Gn1a (Ashikari et al., 2005), a gene controlling the number of spikelets per panicle, suggesting that qNS-1-1 might be the same locus as Gn1a. On chromosome 9, two QTLs, qPLSD-9-1 and qPLSD-9-2 were detected, and located upstream and downstream of RM5652, respectively. Up to now, one major QTL, qPE9-1 (Yan et al., 2007), and two genes, EP2 (Zhu et al., 2010) and EP3 (Piao et al., 2009), which are responsible for erect panicle in rice have been found, and located on chromosome 9, 7 and 2, respectively. qPE9-1

was located downstream the marker RM5652, and coincided with DEP1 (Huang et al., 2009), a gene responsible for dense panicle, high grain number per panicle and erect panicle, indicated that qPLSD-9-2, qPE9-1 and DEP1 may be the same locus.

Novel pleiotropic loci controlling PL and SD simultaneously Two pleiotropic loci, qPLSD-9-1 and qPLSD-9-2, controlling PL and SD simultaneously were identified in this study. qPLSD-9-1 explained 34.7% and 25.4% of the phenotypic variation for PL and SD, respectively. qPLSD-9-2 explained 34.9% and 24.4% of the phenotypic variation for PL and SD, respectively. Up to now, four major QTLs for panicle length were reported. Among these, two QTLs located on chromosome 1 (Xiong et al., 1999) and chromosome 7 (Li et al., 2006) were detected in populations of inter-specific crosses. The other two major QTLs, located on chromosome 4 (Teng et al., 2002) and chromosome 6 (Mei et al., 2005), respectively, were detected in populations of inter-subspecific crosses. In addition, three major QTLs for spikelet density have been reported currently. Among the three QTLs, one was detected in a population derived from an indica-japonica cross and located on chromosome 4 (Xiao et al., 1996). The other two major QTLs on chromosome 7 (Lin et al., 1996) and chromosome 8 (Xing et al., 2001) were detected in populations of indica-indica crosses. Thus, we deduced that qPLSD-9-1 and qPLSD-9-2 were two novel pleiotropic loci.

Panicle morphogenetic processes being controlled by different genes Rice panicle size is determined by such underlying morphogenetic processes as: 1) primary branch formation on the panicle axis; 2) floret formation on the primary branches (mainly determined by the secondary branch formation); and 3) pre-flowering abortion of florets in the panicle (Yamagishi et al., 2004). In the present study, QTLs accounting for more than 10% of phenotypic variation for the number of primary branches per panicle were detected in interval RM486-RM265 on chromosome 1 and interval RM80-RM281 on chromosome 8, respectively. QTL accounting for more than 10% of phenotypic variation for the number of secondary branches per panicle, secondary branch distribution density and the number of

Yuan Guo et al. / Journal of Genetics and Genomics 37 (2010) 533544

spikelets per panicle were detected in interval RM84RM462 on chromosome 1. The result indicated that QTLs affecting the first underlying morphogenetic process were different from QTLs affecting the latter two processes. Yamagishi et al. (2002) reported that a QTL controlling the number of primary branches was located on chromosome 11, while QTLs controlling the number of secondary branches and average number of spikelets on the secondary branch were located on chromosome 6 by using DH lines derived from a cross between two japonica rice cultivars, Akihikari and Koshihikari. Similarly, it was also reported that panicle morphogenetic processes were controlled by different genes (Nagata et al., 2002; Jing et al., 2004; Yamagishi et al., 2004).

Genetic variation of panicle traits being mainly attributed to additive effect In the present study, additive × additive QTL interaction pairs were detected for all seven panicle traits. However, phenotypic variation explained by these interactions was small, only ranged from 0.6% to 4.3%. It indicated that additive × additive QTL interactions were the factors affecting panicle traits but not the major component in genetic variation. The genetic variations of panicle traits were mainly attributed to additive effects. The same conclusion was deduced by Shen et al., (2009) by using RILs derived from Xieqingzao B and R9308. For indica-indica cross population, the contributions of QTL main effects to the phenotypic variation of rice yield related traits were higher than those of epistatic effects (Yu et al, 1997; Xing et al, 2002; Zhuang et al., 2002). Similar results have also been observed in other mapping studies for yield traits in rice (Li et al., 1997; Liao et al., 2001).

Acknowledgements This work was supported by the Program for Changjiang Scholars and Innovative Research Team of Nanjing Agriculture University (No. IRT0432). We thank Dr. Guo-Liang Jiang of South Dakota University, USA, for his helpful comments to the manuscript.

References Ashikari, M., Sakakibara, H., Lin, S.Y., Yamamoto, T., Takashi, T.,

543

Nishimura, A., Angeles, E. A., Qian, Q., Kitano, H., and Matsuoka, H. (2005). Cytokinin oxidase regulates rice grain production. Science 309: 741745. Chen, X.G., Liu, J.B., and Hong, D.L. (2006). Genetic analysis on panicle angle and number of spikelets per panicle by using six generations of three crosses derived from erect×curve panicles in japonica rice (Oryza sativa L.). Acta. Agron. Sin. 32: 11431150 (in Chinese with an English abstract). Guo, Y., Wan, Z.B., Chen, X.G., and Hong, D.L. (2008). Genetic analysis on number of primary and secondary branches per panicle in japonica rice (Oryza sativa L.). J. Nanjing. Agri. Univ. (Natural Science) 31: 812 (in Chinese with an English abstract). Guo, Y., Cheng, B.S., and Hong, D.L. (2009). Construction of SSR linkage map and analysis of QTL for rolled leaf of restorer line in japonica rice (Oryza sativa L.). Chin. J. Rice. Sci. 23: 245251 (in Chinese with an English abstract). Jiang, J.H., Guo, Y., Chen, X.G., Xu, H.M., Hou, Y.J., and Hong, D.L. (2007). The correlation between panicle angle and rice quality and genetic analysis on rice quality in japonica rice (Oryza sativa L.). Hereditas (Beijing) 29: 714724 (in Chinese with an English abstract). Jing, Y.H., Fu, Y.C., Sun, C.Q., Zhang, P.J., Xu, Z.J., Chen, W.F., and Wang, X.K. (2004). Mapping QTLs for vascular bundle in peduncle and yield components of rice (Oryza sativa L.). J. Chin. Agri. Univ. (Natural Science) 9: 1621 (in Chinese with an English abstract). Hirota, O., Oka, M., and Takeda, T. (1990). Sink activity estimation by sink size and dry matter increase during the ripening stage of barley (Hordeum vulgare) and rice (Oryza sativa). Ann. Bot. 65: 349353. Huang, X.Z., Qian, Q., Liu, Z.B., Sun, H.Y., He, S.Y., Luo, D., Xia, G.M., Chu, C.C., Li, J.Y., and Fu, X.D. (2009). Natural variation at the DEP1 locus enhances grain yield in rice. Nat. Genet. 41: 494497. Lander, E.S., Green, P., Abrahamson, J., Barlow, A., Daly, M.J., Lincoin, S.E., and Newburg, L. (1987). Mapmaker: an interactive computer package for constructing primary genetic linkage maps of experimental and natural populations. Genomics 1: 174181. Li, Z.K., Pinson, S.R.M., Park, W.D., Paterson, A.H., and Stansel, J.W. (1997). Epistasis for three grain yield components in rice (Oryza sativa L.). Genetics 145: 453465. Li, C.B., Zhou, A.L., and Sang, T. (2006) Genetic analysis of rice domestication syndrome with the wild annual species, Oryza nivara. New Phytol. 170: 185194. Liao, C.Y., Wu, P., Hu, B., and Yi, K.K. (2001). Effects of genetic background and environment on QTLs and epistasis for rice (Oryza sativa L.) panicle number. Theor. Appl. Genet. 103: 104111. Lin, H.X., Qian, H.R., Zhuang J.Y., Lu, J., Min, S.K., Xiong, Z.M., Huang, N., and Zheng, K.L. (1996). RFLP mapping of QTLs for yield and related characters in rice (Oryza sativa L.). Theor. Appl. Genet. 92: 920927. Liu, R.H., and Meng, J.L. (2003). MapDraw: a microsoft excel macro for drawing genetic linkage maps based on given genetic linkage data. Hereditas (Beijing) 25: 317321 (in Chinese with an English abstract). Liu, J.B., and Hong, D.L. (2005). Genetic analysis on panicle angle and number of spikelets per panicle in japonica rice (Oryza sativa). Chin. J. Rice. Sci. 19: 223230 (in Chinese with an English abstract). Liu, T.M., Mao, D.H., Zhang, S.P., Xu, C.G., and Xing, Y.Z. (2009).

544

Yuan Guo et al. / Journal of Genetics and Genomics 37 (2010) 533544

Fine mapping SPP1, a QTL controlling the number of spikelets per panicle, to a BAC clone in rice (Oryza sativa). Theor. Appl. Genet. 118: 15091517. Liu, B.L., Chen, S.Q., Dong, D., Qian, L., and Xu, Z.J. (2009). Study on the relationship between rice plant type and false smut of rice. HuBei. Agri. Sci. 48: 4246 (in Chinese with an English abstract). Lü, W.Y., Cao, P., Shao, G.J., and Cao, B.C. (1997). Study on quality character of main japonica rice varieties in Liaoning province. Liaoning. Agri. Sci. 5: 711 (in Chinese). McCouch, S.R., Cho, Y.G., and Yano, M. (1997). Report on QTL nomenclature. Rice Genet. Newsl. 14: 1114. Mei, H.W., Li, Z,K., Shu, Q.Y., Guo, L.B., Wang, Y.P., Xu, X.Q., Ying, C.S., and Luo, L.J. (2005). Gene actions of QTLs affecting several agronomic traits resolved in a recombinant inbred rice population and two backcross populations. Theor. Appl. Genet. 110: 649659. Nagata, K., Fukuta, Y., Shimizu, H., Yagi, T., and Terao, T. (2002). Quantitative trait loci for sink size and ripening traits in rice (Oryza sativa L.). Breed. Sci. 52: 259273. Piao, R.H., Jiang, W Z., Ham, T.H., Choi, M.S., Qiao, Y.L., Chu, S.H., Park, J.H., Woo, M.O., Jin, Z.X., An, G., Lee, J., and Koh, H.J. (2009). Map-based cloning of the ERECT PANICLE 3 gene in rice. Theor. Appl. Genet. 119: 14971506. Shen, X.H., Cao, L.Y., Chen, S.G., Zhan, X.D., Wu, W.M., and Cheng, S.H. (2009) Dissection of QTLs for panicle traits in recombinant inbred lines derived from super hybrid rice, Xieyou 9308. Chin. J. Rice. Sci. 23: 354362 (in Chinese with an English abstract). Teng, S., Qian, Q., Zeng, D.L., Kunihiro, Y., Huang, D.N., and Zhu, L.H. (2002). QTL analysis of rice peduncle vascular bundle system and panicle traits. Acta. Bot. Sin. 44: 301306. Tian, F., Zhu, Z.F., Zhang, B.S., Tan, L.B., Fu, Y.C., Wang, X.K., and Sun, C.Q. (2006). Fine mapping of a quantitative trait locus for grain number per panicle from wild rice (Oryza rufipogon Griff). Theor. Appl. Genet. 113: 619629. Xiao, J., Li, J., Yuan, L., and Tanksley, S.D. (1996). Identification of QTLs affecting traits of agronomic importance in a recombinant inbred population derived from a subspecific rice cross. Theor. Appl. Genet. 92: 230244. Xing, Y.Z., Xu, C.G., Hua, J.P., and Tan, Y.F. (2001). Analysis of QTL×environment interaction for rice panicle characteristics. Acta. Genet. Sin. 28: 439446 (in Chinese with an English abstract). Xing, Y.Z., Tan, Y.F., Hua, J.P., Sun, X.L., Xu, C.G., and Zhang, Q.F. (2002). Characterization of the main effects, epistatic effects and their environmental interactions of QTLs on the genetic basis of yield traits in rice. Theor. Appl. Genet. 105: 248257. Xing, Y.Z., Tang, W.J., Xue, W.Y., Xu, C.G., and Zhang, Q.F. (2008). Fine mapping of a major quantitative trait loci, qSSP7, controlling the number of spikelets per panicle as a single Mendelian factor in rice. Theor. Appl. Genet. 116: 789796.

Xiong, L.Z., Liu, K.D., Dai, X.K., Xu, C.G., and Zhang, Q.F. (1999). Identification of genetic factors controlling domestication-related traits of rice using an F2 population of a cross between Oryza sativa and O. rufipogon. Theor. Appl. Genet. 98: 243251. Xue, W.Y., Xing, Y.Z., Weng, X.U., Zhao, Y., Tang, W.J., Wang, L., Zhou, H.J., Yu, S.B., Xu, C.G., Li, X.H., and Zhang, Q.F. (2008). Natural variation in Ghd7 is an important regulator of heading date and yield potential in rice. Nat. Genet. 40: 761767. Yamagishi, M., Takeuchi, Y., Kono, I., and Yano, M. (2002). QTL analysis for panicle characteristics in temperate japonica rice. Euphytica 128: 219224. Yamagishi, J., Miyamoto, N., Hirotsu, S., Laza, R.C., and Nemoto, K. (2004). QTLs for branching, floret formation and pre-flowering floret abortion of rice panicle in a temperate japonica×tropical japonica cross. Theor. Appl. Genet. 109: 15551561. Yan, C.J., Zhou, J.H., Yan, S., Chen, F., Yeboah, M., Tang, S.Z., Liang, G.H., and Gu, M.H. (2007) Identification and characterization of a major QTL responsible for erect panicle trait in japonica rice (Oryza sativa L.). Theor. Appl. Genet. 115: 10931100. Yang, J., Zhong, W.G., Zhang, Z.L., Song, X.L., Chen, Z.D., and Ji, J.A. (1999). Current situation of and improving strategy for grain quality of japonica rice in Jiangsu province. Jiangsu. Agri. Sci. 3: 68 (in Chinese). Yang, J., Zhu, J., and Williams, R.W. (2007). Mapping the genetic architecture of complex traits in experimental populations. Bioinformatics 23: 15271536. Yang, J., Hu, C.C., Hu, H., Yu, R.D., Xia, Z., Ye, X.Z., and Zhu, J. (2008). QTLNetwork: mapping and visualizing genetic architecture of complex traits in experimental populations. Bioinformatics 24: 721723. Yu, S.B., Li, J.X., Tan, Y.F., Gao, Y.J., Li, X.H., Zhang, Q.F., and Saghai Maroof, M.A. (1997). Importance of epistasis as the genetic basis of heterosis in an elite rice hybrid. Proc. Natl. Acad. Sci. USA 94: 92269231. Zeng, Z.B. (1994). Precision mapping of quantitative trait loci. Genetics 136: 14571468. Zhu, K.M., Tang, D., Yan, C.J., Chi, Z.C., Yu, H.X., Chen, J.M., Liang, J.S., Gu, M.H., and Cheng, Z.K. (2010). ERECT PANICLE2 encodes a novel protein that regulates panicle erectness in indica rice. Genetics 184: 343350. Zhuang, J.Y., Lin, H.X., Lu, J., Qian, H.R., Hittalmani, S., Huang, N., and Zheng, K.L. (1997). Analysis of QTL×environment interaction for yield components and plant height in rice. Theor. Appl. Genet. 95: 799808. Zhuang, J.Y., Fan, Y.Y., Rao, Z.M., Wu, J.L., Xia, Y.W., and Zheng, K.L. (2002). Analysis on additive effects and additive-by-additive epistatic effects of QTLs for yield traits in a recombinant inbred line population of rice. Theor. Appl. Genet. 105: 11371145.