Pc

Y

a

b

c

d

e

a

ARRA

KDORY

1

gedriAmflbrpoi

0d

Field Crops Research 123 (2011) 130–138

Contents lists available at ScienceDirect

Field Crops Research

journa l homepage: www.e lsev ier .com/ locate / fc r

hysiological characterization of introgression lines derived from an indica riceultivar, IR64, adapted to drought and water-saving irrigation

oichiro Katoa,∗, Amelia Henryb, Daisuke Fujitab, Keisuke Katsurac, Nobuya Kobayashib,d, Rachid Serrajb,e

Institute for Sustainable Agro-ecosystem Services, The University of Tokyo, Tokyo 188-0002, JapanInternational Rice Research Institute, PO Box 7777, Metro Manila, PhilippinesExperimental Farm, Kyoto University, Osaka 569-0096, JapanJapan International Research Center for Agricultural Sciences, Tsukuba 305-8686, JapanInternational Centre for Agricultural Research in the Dry Areas, PO Box 5466, Aleppo, Syria

r t i c l e i n f o

rticle history:eceived 10 March 2011eceived in revised form 8 May 2011ccepted 8 May 2011

eywords:rought adaptationryza sativaoot system architectureield potential

a b s t r a c t

Water scarcity threatens sustainable rice production in many irrigated areas around the world. To copewith the scarcity, aerobic rice culture has been proposed as a promising water-saving technology. Theobjective was to elucidate the physiological attributes behind the performance of rice introgression linesin water-saving culture. We evaluated yield potential and physiological adaptation traits to water deficitof BC3-derived lines with the genetic background of an elite indica cultivar, IR64, in the field and inpot experiments. One line, YTH183, had 26% higher yield than IR64 under non-stress conditions (895vs. 712 g m−2 on average). This was attributed to enlarged sink capacity due to large grain size, whichcontributed to more efficient use of assimilates and hence a higher harvest index. YTH183 also showedbetter dehydration avoidance under intermittent soil drying, due to the adaptive response of deep rooting

to water deficiency. The grain yield of YTH183 exceeded that of IR64 by 92–102% under moderate waterdeficit caused by limited irrigation in aerobic rice culture (143 vs. 72 g m−2). Two introgressed segmentson chromosomes 5 and 6 might, at least in part, confer the higher yield potential and greater dehydrationavoidance in YTH183 simultaneously. Advanced backcross breeding combined with molecular geneticsand physiological characterization of introgressed segments would be effective for developing new ricecultivars with high yield potential and drought adaptation traits.

. Introduction

Declining water availability threatens the sustainability of irri-ated rice (Oryza sativa L.) production in many countries (Pengt al., 2009). Among the new water-saving technologies beingeveloped, aerobic rice culture can greatly reduce irrigation waterequirements (Tuong et al., 2005) with being intensively studiedn China, Brazil, India, Australia, and parts of southeast Asia andfrica (Bouman et al., 2007; Humphreys et al., 2010). It aims ataximizing crop water productivity by growing plants without

ooding or puddling, in aerobic soil (Kato et al., 2009). Aero-ic rice culture was reported to use less than 50% of the waterequired by conventional flooded rice, but with a significant yield

enalty (Peng et al., 2006). Accordingly, its success relies entirelyn the development of new suitable cultivars. However, breed-ng for this purpose confronts completely different challenges

∗ Corresponding author. Tel.: +81 42 463 1793; fax: +81 42 464 4391.E-mail address: agrkato@mail.ecc.u-tokyo.ac.jp (Y. Kato).

378-4290/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.fcr.2011.05.009

© 2011 Elsevier B.V. All rights reserved.

from the breeding of rainfed rice, in breaking the yield poten-tial while achieving acceptable adaptation to the moderate waterdeficit of aerobic soils (Atlin et al., 2006). Because rice cultivarsfor aerobic culture must perform well in both well-watered andwater-limited conditions, it is likely that multiple mechanisms willbe necessary to confer a consistent yield advantage (Guan et al.,2010).

It is well-known that there is a wide range of genetic varia-tion in yield potential and drought adaptation in rice. In particular,improvement in root system architecture towards deeper root sys-tem would be most effective for better adaptation to occasional orintermittent soil water deficit between irrigation events in aero-bic rice culture (Gowda et al., 2011). On the other hand, rice yieldpotential is often constrained by the limited sink capacity eitherdue to lower spikelet density (spikelet number per unit area) orsmaller grain size (Akita, 1999). There is an opportunity for further

increasing yield potential by enhancing sink capacity in modernrice cultivars (Ohsumi et al., 2011). By pyramiding the alleles con-ferring these characteristics into one genome proportionately, itmight be possible to develop superior rice cultivars with high yield

Resea

p2

lbvgrtaiaihpgwAniocvs

abbtgw

2

2

gepwmbrgms(otur

2

osfiLsfs2

Y. Kato et al. / Field Crops

otential and irrigation-use efficiency in water-saving culture (Luo,010; Serraj et al., 2011).

Advances in quantitative genetics such as quantitative traitocus (QTL) analysis have fundamentally changed conventional ricereeding in the last decade, and made it possible to tailor genotypesia marker-assisted selection (Ashikari et al., 2005). Molecularenetics has been already adopted in breeding for disease and pestesistance (Fukuoka et al., 2009). It could also provide a powerfulool for raising yield potential (Adachi et al., 2010) and resistance tobiotic stress (Septiningsih et al., 2009) when combined with phys-ological annotation of favourable alleles. Furthermore, Tanksleynd Nelson (1996) suggested the combination of backcross breed-ng with molecular genetics. The advanced backcross QTL analysisas enabled applications of molecular genetics in plant breedingrograms (Collins et al., 2008). Such analysis detects QTLs in eliteenetic backgrounds through the use of promising backcross lineshile maintaining genetic variability at the whole-genome scale.dvanced backcross breeding may prove efficient for developingew rice genotypes adapted to water-saving agriculture, because

t pools the valuable alleles from unadapted germplasms as a listf genotypes which are ready to be isolated. By using these lines,rop scientists can safely avoid the problems of genetic sterility andariation in phenology often encountered in molecular breedingtudies.

The objective of this study was to elucidate the physiologicalttributes underlying the performance of lines with the geneticackground of an indica rice cultivar, IR64, in advanced backcrossreeding for water-saving agriculture. We targeted aerobic rice cul-ure as a promising water-saving technology, and evaluated theenotypes with an emphasis on yield potential and adaptation toater deficit.

. Materials and methods

.1. Plant materials

Introgression rice (O. sativa L.) lines with the IR64 genetic back-round (BC3-derived lines from “New Plant Type” donors; see Fujitat al., 2009 for details) were used for this study. IR64 (recurrentarent) is a semi-dwarf, lowland-adapted indica cultivar that isidely grown in the tropics. It has good grain quality, but hasedium yield potential and is drought susceptible. The genetic

ackground of these introgression lines is uniform compared withecombinant inbred lines, and the introgression lines have beenenotyped (Fujita et al., 2010a). Thus, these introgression linesake a useful resource for both physiological and molecular genetic

tudies (Farooq et al., 2010). We selected eight introgression linesYTH146, 183, 243, 254, 272, 288, 302 and 304) on the basisf their performance in preliminary observations at the Interna-ional Rice Research Institute (IRRI), The Philippines (N. Kobayashi,npublished data). Several check cultivars were also evaluated aseferences.

.2. Experiment 1: Evaluation of yield potential

Field experiments were conducted at the experimental farmf the University of Tokyo, Japan (35◦43′N, 139◦32′E), during theummer (May–September) of 2009 and 2010; at the experimentalarm of Kyoto University, Osaka, Japan (34◦51′N, 135◦37′E), dur-ng the summer of 2010; and at the experimental station of IRRI,os Banos, The Philippines (14◦11′N, 121◦15′E), during the dry sea-

on (January–May) of 2010 (Table 1). Weather data were obtainedrom the Tokyo and Osaka meteorological stations (20 km from theites), or were recorded at meteorological stations located withinkm of the field in Los Banos. The soils are a Typic Melanudand

rch 123 (2011) 130–138 131

(clay loam) at the University of Tokyo, a Typic Fluvaquent (clayloam) at Kyoto University, and a Typic Tropaqualf (heavy clay) atIRRI’s lowland farm. Fields were puddled and kept flooded aftertransplanting to evaluate the yield potential of the rice genotypes.A non-puddled, saturated trial with direct-sowing was also set up inTokyo in 2009 (daily flush irrigation to maintain soil water potentialabove −10 kPa at 20 cm depth).

Several check cultivars and the IR64 introgression lines werearranged in a randomized complete block design in each trial, withthree (Tokyo and Osaka) or four (Los Banos) replicates per genotype(Table 2). IR68522-10-2-2 is a tropical japonica lowland-adaptedline, one of the first-generation New Plant Type series (Fujita et al.,2009). NSIC Rc158 is a second-generation New Plant Type culti-var (Peng et al., 2008) recently bred for high yield potential in thetropics. PSBRc80 is a check cultivar, well adapted to water-savingculture in lowlands (Bueno et al., 2010). IR55423-01 has one ofthe highest yield potentials among upland-adapted cultivars in thetropics (Atlin et al., 2006). Takanari is an indica lowland-adaptedcultivar with the highest yield potential in central Japan (Takai et al.,2006).

Each plot consisted of eight 4-m rows (25- or 30-cm spacing),except in Tokyo in 2009 (six 4-m rows, 20-cm spacing). One 4-leaf seedling in Tokyo and Osaka or three seedlings in Los Banoswere transplanted into each hill (20–25 hills m−2): between lateMay and early June in Osaka and Tokyo, and in mid-January in LosBanos. In the saturated trial, four or five seeds were directly sownon each hill at the same density as for transplanted rice. We applied32–43 kg ha−1 of P and 32–83 kg ha−1 of K basally; and 95 kg ha−1

of N in three splits in Los Banos, 120 kg ha−1 of N in three splitsin Osaka, and 160 kg ha−1 of N in five splits in Tokyo. Diseasesand pests were controlled by the standard practices in the studyareas.

At physiological maturity, 1.0 m2 of rice plants (avoiding borderrows) was harvested at ground level to determine the grain yieldand yield components. Grain and straw were dried in an oven at80 ◦C for 3 days and grain yield and filled grain weight (mg per grain)were calculated on a 14% moisture content basis. Sink capacity [alsodefined as ‘yield capacity’ by Yoshida (1972)] was calculated astotal spikelet number m−2 × single-grain weight according to Maeet al. (2006). Sink capacity represents the potential capacity of ves-sels that receive carbohydrates during grain filling (Yoshida, 1972).Grain filling ratio was calculated as dry weight of filled grains/sinkcapacity. At the beginning of anthesis, 8–12 plants were sampled,and the post-anthesis biomass accumulation was also calculated.The apparent translocation of dry matter from non-grain parts wascalculated as the dry weight of filled grains minus the post-anthesisbiomass accumulation. For calculating the apparent translocation,we have not taken leaf and tiller senescence into account. At fullanthesis stage, the SPAD value (a measure of chlorophyll content)and the stomatal conductance of IR64 and YTH183 were deter-mined. The SPAD values of the flag leaf of 20–30 stems per plotwere measured with a SPAD-502 chlorophyll meter (Minolta Co.,Ltd., Osaka, Japan). The stomatal conductance of the abaxial sideof the flag leaf of seven stems per plot was measured with an SC-1 leaf porometer (Decagon Devices Inc., Pullman, WA, USA) in themorning (10:00–12:00) under sunny conditions. Data in each trialwere analysed by using the generalized linear model procedure(SAS Institute, 2003), and genotype means were compared by leastsignificant difference at P = 0.05. The data of IR64 and YTH183 werealso compared by paired t-test.

2.3. Experiment 2: Yield response to diverse irrigation intensities

Field experiments were conducted at the IRRI upland farmduring the 2009 dry season (Table 1) to evaluate the agronomicperformance of the IR64 introgression lines in aerobic culture in

132 Y. Kato et al. / Field Crops Research 123 (2011) 130–138

Table 1Description of the trials in the three experiments.

Location Year Season Condition N–P–K (kg ha−1) Mean temperature (◦C) Solar radiation (MJ m−2 d−1)

Experiment 1Los Banos 2010 Dry Lowland field 95–32–32 27.5 18.7Tokyo 2009 Summer Lowland field 160–39–67 23.7 14.2Tokyo 2010 Summer Lowland field 160–39–67 25.1 17.1Osaka 2010 Summer Lowland field 120–43–83 25.6 18.2

Experiment 2Los Banos 2009 Dry Upland field 90–30–30 26.9 15.3

–100–3

tlodltivlawgtise1bS

gilgogot3c

tibu

TG

M

Experiment 3Los Banos 2009 Dry GreenhouseTokyo 2010 Summer Upland field

he tropics. The soil is an Andaqueptic Haplaquol (clay loam). Aine-source sprinkler irrigation system was set up to create vari-us irrigation application rates across contiguous plots as recentlyescribed (Centritto et al., 2009). The system consisted of a single

ine of sprinklers 4.5 m apart in the centre of the field, perpendicularo the crop rows. The sprinklers give linearly decreasing amounts ofrrigation with distance from the line. A 10-m-wide plot was subdi-ided into five zones, each 2 m across, at 2–4 m from the sprinklerine (Zone A), 4–6 m (Zone B), 6–8 m (Zone C), 8–10 m (Zone D),nd 10–12 m (Zone E). The sprinklers were operated at predawn,hen the wind speed was low. Initially, the entire field was irri-

ated 2 or 3 times a week by conventional sprinkler to maintainhe surface soil close to field capacity. The irrigation treatment wasnitiated 1 month after sowing by decreasing irrigation frequencyo that plants furthest from the line received less than optimal lev-ls of irrigation (Centritto et al., 2009). Volumetric soil moisture at0-cm intervals to a depth of 70 cm was monitored in each zoney using a frequency domain reflectometry sensor (Diviner 2000,entek Sensor Technologies, Stepney, SA, Australia).

Fourteen genotypes were grown: three indica upland-adaptedenotypes (IR55423-01, IR55419-04, IR71525-19-1-1), two trop-cal japonica upland-adapted genotypes (IRAT109, Azucena), twoowland-adapted genotypes (IR64, Takanari), and the seven intro-ression lines. Seeds were sown directly into dry soil at 80 kg ha−1

n 7 January 2009, in rows 25 cm apart and hills 10 cm apart. Eachenotype was planted in four replicates, with each plot consistingf four rows (10 m long). N fertilizer was applied in two splits at aotal rate of 90 kg ha−1. P and K fertilizers were basally applied at0 kg ha−1 each. Weeds were controlled manually, and pests wereontrolled by the standard practices in the study area.

The anthesis date was recorded in each subplot, and 2 m of the

wo central rows (1.0 m2) of each subplot was harvested at phys-ological maturity. Shoot biomass, grain yield (14% grain moistureasis), and harvest index were recorded. Data were analysed bysing the REML algorithm of the MIXED model procedure (SAS

able 2rain yield (g m−2) under non-stress conditions.

Los Banos 2010DS, flooded Osaka 2010, flooded Toky

NSIC Rc158 729 a –PSBRc80 663 ab –IR55423-01 632 ab –IR68522-10-2-2 402 c –Takanari – 979a 1248IR64 433 c 689 b 95YTH183 707 a 921 a 1070YTH254 – 668 b 93YTH304 626 ab –YTH302 585 b –YTH243 582 b –YTH146 570 b –YTH272 524 bc –YTH288 446 c –

eans followed by the same letter within columns are not significantly different (P < 0.05

– –9–67 25.1 17.1

Institute, 2003), and the Tukey–Kramer method was used for com-parison of means at P = 0.05.

2.4. Experiment 3: Root response to hydrological conditions

The rooting responses of IR64 and the five introgression linesto hydrological conditions were evaluated in a naturally lit green-house at IRRI during the 2009 dry season. Three water regimeswere established: flooded, aerobic and water-limited. Genotypeswere arranged in a split-plot design with six replicates (main-plot,water regimes). We used cylinders (18 cm diameter, 90 cm height)lined with plastic bags and filled with IRRI upland farm soil at abulk density of 0.98 g dry soil cm−3. Chemical fertilizer equivalentto N, P, K = 0.3, 0.3, 0.3 g per cylinder was incorporated into the top30 cm of the soil. Seeds were sown on 21 December 2008, and threeseedlings per cylinder were grown. In the flooded regime, a watertable of 2–3 cm above the soil surface was maintained. In the aero-bic regime, full irrigation was applied daily to keep the soil near fieldcapacity, and any excess water drained from a hole at the bottom ofthe cylinders. In the water-limited regime, irrigation was withheldfrom 16 days after sowing (DAS) except for two small irrigationsto avoiding leaf drying. The water input at 16–39 DAS was 34%less in the aerobic treatment compared with the flooded treatmentand 90% less in the water-limited treatment. Plants were harvestedat 39 DAS to determine the shoot biomass and root length. Soilsamples were separated into four sections (0–15, 15–30, 30–60,and 60–90 cm), but we combined the root lengths below 15 cmbecause only a few roots were observed below 30 cm. After thesamples were gently washed on a 1-mm sieve, the roots were con-served in 50% ethanol until scanning. An 8-bit greyscale image

of each root sample was acquired at a resolution of 400 dpi on aflatbed image scanner with a transparency unit (Epson V700, LongBeach, CA, USA). The root length was determined by WinRHIZO v.2009b software (Regent Instruments, Blain City, QC, Canada). Data

o 2010, flooded Tokyo 2009, flooded Tokyo 2009, saturated (−10 kPa)

– – –– – –– – –– – –a – –

4 c 877 c 606 bb 1045 a 732 a

4 c – –– 946 b 553 b– – –– 860 c 577 b– – –– – –– – –

).

Research 123 (2011) 130–138 133

wI

2fi

rotcr2iaTiwhadbaWtlplbp1ml

3

3(

Ih1tNt

Yeawotws(std

fTSYv

0

10

20

30

40

50

120110100908070605040Days after sowing

A

B

C

D

E

Soil moisture (v/v)

Y. Kato et al. / Field Crops

ere analysed with the generalized linear model procedure (SASnstitute, 2003).

.5. Experiment 4: Growth response to dry-down stress in aerobiceld

Subsequently we investigated the root and shoot growthesponses to soil drying in aerobic fields at the experimental farmf the University of Tokyo during the summer of 2010. Three geno-ypes (IR64, YTH183 and YTH254) were arranged in a randomizedomplete block design with four replicates. Plot size was six 2-mows. Seeds were directly sown into dry soil at 80 kg ha−1 on 12 May010 in rows 25 cm apart and hills 10 cm apart. Chemical fertil-

zer (N, P, K = 60, 39, 67 kg ha−1) was applied before sowing, andmmonium sulphate (N = 40 kg ha−1) was top-dressed on 37 DAS.he soil water potential at 20 cm depth was kept near field capac-ty by sprinkler irrigation until 66 DAS, and then two drying cycles

ere imposed during 66–90 DAS. In each cycle, irrigation was with-eld until the soil water potential at 40 cm depth reached −70 kPa,nd then full irrigation was applied. At the onset and end of soilrying, plants from 10 hills were harvested to determine the shootiomass. The fraction of radiation interception was measured twiceweek with a ceptometer (AccuPAR, Decagon Devices Inc., Pullman,A, USA) to calculate the radiation use efficiency (RUE = dry mat-

er accumulation/intercepted radiation). At 77 DAS, the pre-dawneaf water potential and stomatal conductance were measured byressure chamber (DIK-7002, Daiki Rika Kogyo, Saitama, Japan) and

eaf porometer. At 90 DAS, root samples were collected midwayetween the rows: Four soil cores (5 cm diameter, 70 cm length)er plot were taken with a soil sampler and divided into 0–15-cm,5–30-cm and 30–70-cm sections, and the root length was deter-ined as described above. Data were analysed with the generalized

inear model procedure (SAS Institute, 2003).

. Results

.1. Evaluation of yield potential in IR64 introgression linesexperiment 1)

Several IR64 introgression lines yielded significantly more thanR64 under non-stress conditions (Table 2). In particular, YTH183ad the highest yield among the introgression lines, and yielded2–63% more than IR64. The yield of YTH183 was 12% higher thanhat of IR55423-01 (upland-adapted check) and similar to that ofSIC Rc158 (high-yielding check) in the tropics, but 6–14% lower

han that of Takanari (high-yielding check) in Japan.To analyse the mechanisms underlying the higher yield of

TH183, we compared the yield components (Table 3). Under allnvironments, the harvest index of YTH183 was 10–38% (24% onverage) higher than that of IR64. The grain weight of YTH183as also 20% higher than that of IR64. The grain weights of

ther introgression lines were not significantly different fromhat of IR64 (data not shown). The panicle number of YTH183as similar to or slightly higher than that of IR64, but panicle

ize (number of spikelets) did not differ. Thus, the sink capacitytotal spikelet number m−2 × single-grain weight) of YTH183 wasignificantly larger than that of IR64, except in Osaka (Table 3). Onhe other hand, the grain filling ratio was not significantly differentespite the large difference in sink capacity.

The shoot biomass of YTH183 was not significantly differentrom that of IR64 in three out of the five environments (Table 4).

he post-anthesis dry matter accumulation showed a similar result.tomatal conductance at anthesis (gs) tended to be higher inTH183 than in IR64 (although not significantly), and the SPADalue at anthesis tended to be lower in YTH183. Although pho-

Fig. 1. Time course of soil moisture at 10 cm depth under different irrigation intensi-ties in aerobic rice culture at IRRI during the 2009 dry season. Bars indicate standarderrors (n = 4).

toassimilation indicated by shoot biomass was not improved, theapparent translocation of dry matter from non-grain parts to graintended to be higher in YTH183 than in IR64, according to differencesin shoot mass before and after grain filling (Table 4).

3.2. Yield response to diverse irrigation intensities in IR64introgression lines (experiment 2)

From irrigation Zones A–E, the total water inputs (irrigation plusrainfall) were 1029, 1006, 940, 857 and 747 mm. The gradient inthe soil moisture was sometimes interrupted by rainfall, but wateravailability at 10 cm depth corresponded fairly well to the irriga-tion intensity during most of the growth period (Fig. 1). The soilmoisture content below 20 cm remained near field capacity anddid not differ between zones (data not shown), indicating that theyield response to irrigation in aerobic culture concerned mainly thephysiological response to the cyclically drying soil in the surfacelayer (e.g., at 100 DAS).

Grain yield decreased significantly in the order of Zones A, B,C > Zone D > Zone E (Table 5). Irrigation had a greater effect onharvest index than on shoot biomass. Overall grain yield differedsignificantly among the 14 genotypes. Among the introgressionlines, only YTH183 yielded more than IR64 (by 65%). Its yield wassimilar to that of tropical japonica upland-adapted IRAT109, but29% lower than that of IR55423-01 and IR71525-19-1-1. Genotypicvariation in yield was more closely associated with harvest index(r = 0.92**) than with shoot biomass (r = 0.61*). YTH183 yieldedmore than IR64 at all irrigation intensities (Fig. 2). Interestingly,increasing severity of stress magnified the yield difference: YTH183yielded 92–102% more than IR64 in Zones D and E which corre-sponded with an 82–130% greater harvest index.

3.3. Root and shoot responses of IR64 introgression lines to dryingsoils (experiments 3 and 4)

The hydrological effect on shoot biomass was moderate, withonly 40% reduction in water-limited regime compared with floodedregime (Fig. 3). On the other hand, root system architectureresponded dramatically to water regimes, reflecting the adaptiveplasticity of root growth. Total root length in the surface soil(0–15 cm) in the water-limited regime was reduced by 78% com-pared with the flooded regime, while that in the subsurface soil(15–90 cm) was over 7 times longer than that in the flooded regime.The genotype differences in shoot dry weight and root length at

0–15 cm depth were not significant. However, the genotype differ-ences in root length at 15–90 cm depth in the flooded regime weresignificant. Although the genotype × water interaction in this traitwas not significant, YTH254 and YTH183 tended to produce more

134 Y. Kato et al. / Field Crops Research 123 (2011) 130–138

Table 3Yield components of IR64 and YTH183 under non-stress conditions.

Harvest index Singlegrain weight (mg) Panicle number (m−2) Spikelets per panicle Sink capacity (g m−2) Grain filling ratio

Los Banos, 2010DS, floodedIR64 0.36 26.6 467 – – –YTH183 0.49* 29.6** 462 ns – – –

Osaka, 2010, floodedIR64 0.36 24.4 280 139 818 0.72YTH183 0.48** 30.9** 289 ns 128 ns 984 ns 0.81 ns

Tokyo, 2010, floodedIR64 0.40 24.0 270 178 994 0.83YTH183 0.44* 29.6** 311* 177 ns 1401* 0.67 ns

Tokyo, 2009, floodedIR64 0.41 26.4 247 154 864 0.87YTH183 0.50* 30.9** 272 ns 163 ns 1179* 0.76 ns

Tokyo, 2009, saturated (−10 kPa)IR64 0.35 23.7 284 139 806 0.65YTH183 0.39** 29.1** 319 ns 154 ns 1225** 0.51 ns

* and ** indicate P < 0.05 and 0.01, respectively. Sink capacity = total spikelet number m−2 × single-grain weight. Grain filling ratio = dry weight of filled grains/sink capacity.

Table 4Dry matter growth, SPAD value, stomatal conductance and apparent translocation of IR64 and YTH183 under non-stress conditions.

Entry Shoot biomass atmaturity (g m−2)

Postanthesis biomassincrease (g m−2)

Anthesis SPADvalue

Anthesis gs

(mol m−2 s−1)Apparent translocation(g m−2)

Los Banos, 2010DS, floodedIR64 1189 421 37.6 0.68 −49YTH183 1412* 603* 39.3 ns 0.72 ns 5 ns

Osaka, 2010, floodedIR64 1668 583 42.3 – 9YTH183 1640 ns 577 ns 39.7* – 215*

Tokyo, 2010, floodedIR64 2040 830 41.6 0.61 −10YTH183 2084 ns 848 ns 41.2 ns 0.65 ns 72 ns

Tokyo, 2009, floodedIR64 1833 – 39.6 0.40 –YTH183 1785 ns – 37.8* 0.45 ns –

Tokyo, 2009, saturated (−10 kPa)IR64 1502 – 34.4 – –

32.6

* 0.86 −

r7

a

TG

Mi

YTH183 1610* –

indicate P < 0.05. gs; stomatal conductance. Apparent translocation = grain yield ×

oot growth in the subsurface soil than IR64 under water limitation:1% and 57% longer, respectively.

We then evaluated the growth responses of YTH254, YTH183nd IR64 to soil drying in aerobic fields. During the dry-down

able 5rain yield, biomass, harvest index and days to anthesis for 14 genotypes grown under d

Grain yield (g m−2) Shoot biom

GenotypeIR71525-19-1-1 313 a 820 abIR55423-01 311 a 890 aIR55419-04 309 ab 757 bIRAT109 254 b 564 cYTH183 223 b 712 bcYTH304 166 c 618 cAzucena 164 c 795 abYTH243 158 c 577 cYTH146 157 c 576 cYTH272 144 c 548 cTakanari 144 c 702 bcYTH302 136 c 580 cIR64 135 c 662 bcYTH288 21 d 622 c

Irrigation levelA 240 � 766 �B 223 � 695 �C 226 � 720 �D 165 � 619 �E 87 � 566 �

eans followed by the same letter within columns are not significantly different (P < 0.0rrigation level was shown as averaged value of genotypes.

ns –

postanthesis biomass increase.

stress cycles (66–90 DAS), the soil moisture at 40 cm depth reached−70 kPa at 77 and 89 DAS (Fig. 4). Root length density at 15–30 cmand 30–70 cm depth at the end of stress was 35% and 79% higher,respectively, in YTH183, and 79% and 215% higher in YTH254, than

ifferent irrigation intensities in aerobic rice culture (2009 dry season, Los Banos).

ass (g m−2) Harvest index Days to anthesis

0.38 b 73 e0.34 bc 80 cd0.39 b 72 e0.45 a 71 e0.30 c 78 d0.25 cd 80 cd0.20 d 80 cd0.25 cd 82 c0.25 cd 82 c0.24 cd 80 cd0.19 d 86 b0.22 d 83 c0.20 d 83 c0.03 e 101 a

0.30 � 80 �0.31 � 80 �0.31 � 80 �0.25 � 81 ��0.14 � 83 �

5). Data for each genotype was mean across all five irrigation levels. Data at each

Y. Kato et al. / Field Crops Research 123 (2011) 130–138 135

* *

**

0

100

200

300

400

EDCBAZone

(a) Grain yield (g m-2)

*

*

*

0.0

0.1

0.2

0.3

0.4

0.5

EDCBAZone

IR64YTH183

(b) Harvest index

F rent irs spect

idsdrc

4

gh2afft2nueff2twciatd

t

TG

ML

ig. 2. (a) Grain yield and (b) harvest index of IR64 and YTH183 grown under diffetandard errors (n = 4). * and ** indicate significant difference at P < 0.05 and 0.01, re

n IR64 (Table 6). Both introgression lines maintained higher pre-awn leaf water potential and stomatal conductance (the latter notignificantly) than IR64 during the dry spells. Their superiority ofehydration avoidance by deep root growth contributed to higheradiation use efficiency and biomass production than IR64 underyclical soil drying.

. Discussion

Genetic improvement for adaptation to water scarcity in irri-ated rice should simultaneously target high yield potential andigh yield stability under moderate water deficit (Serraj et al.,009). Advanced backcross breeding combined with physiologicalnd molecular genetic studies would be the most effective strategyor achieving this purpose. Introgression lines make useful materialor improving both current elite cultivars and our understanding ofhe physiological relationship between traits and yield (Guan et al.,010; Farooq et al., 2010). Several researchers have tried to con-ect plant physiology to the molecular breeding of rainfed rice bysing introgression lines, but with much emphasis on drought tol-rance (Lafitte et al., 2006; Xu et al., 2005). However, breeding riceor water-saving culture requires completely different conceptsrom drought resistance breeding (Atlin et al., 2006; Serraj et al.,011): it should focus more on yield potential, with some adap-ation to cyclically drying soils between irrigations. In this study,e have identified one introgression line (YTH183) having signifi-

antly enhanced yield potential and better dehydration avoidancen aerobic culture than IR64 (Tables 2 and 6). YTH183 had beenlways ranked among the top five high-yielding genotypes among

he 334 introgression lines in the IRRI upland breeding nurseryuring 2006–2008 (N. Kobayashi, unpublished data).

YTH183 out-yielded IR64 by 26% under non-stress condi-ions (895 vs. 712 g m−2), in both temperate and tropical regions

able 6rowth response to the dry-down stress in aerobic fields in Tokyo during summer of 201

RLD at 0–15 cm(cm cm−3)

RLD at15–30 cm(cm cm−3)

RLD at30–70 cm(cm cm−3)

Pre-dawn LWPat 77 DAS(MPa)

IR64 21 ns 3.4 b 0.39 c −0.46 bYTH183 22 4.6 ab 0.70 b −0.26 aYTH254 18 6.1 a 1.23 a −0.26 a

eans followed by the same letter within columns are not significantly different (P < 0.0WP, leaf water potential; and RLD, root length density.

rigation intensities in aerobic rice culture at IRRI in 2009 dry season. Bars indicateively.

(Table 2). Similar results (yield increases of 18%–46%) were repeat-edly obtained in irrigated lowlands at IRRI: 793 vs. 673 g m−2 in2008 (Bueno et al., 2010), and 597 vs. 410 g m−2 in 2009 (data notshown). The increased yield potential of YTH183 is attributable tothe higher harvest index (total grain weight/total biomass) in allcases. The harvest index of IR64 was much lower (0.36–0.41) thanthe current optimum value for this trait (0.50–0.55; Peng et al.,2008). The improved harvest index of YTH183 is attributable to theenlarged sink capacity due to larger grain size (Table 3). Alterna-tively, it might be derived from improved sink regulation whichwas suggested as a putative mechanism of higher yield potential inhybrid rice than inbred rice (Lafarge and Bueno, 2009). On the otherhand, the source supply for grain growth, i.e., post-anthesis biomassaccumulation by YTH183 and the concentration and amount ofnon-structural carbohydrate in the vegetative organs at anthesis(data not shown), was largely not improved (Table 4). Our resultswere in agreement with a previous claim that the sink capacitymost constrains rice yield potential even in modern cultivars (Akita,1999). Mae et al. (2006) reported that larger grain size in new ricecultivars contributed to enhancing yield potential. Likewise, Songet al. (2007) suggested that GW2, a major QTL for grain size onchromosome 2, increased yield potential of indica rice in China. Fur-thermore, Ohsumi et al. (2011) demonstrated that there is room forincreasing yield potential of japonica rice through manipulating thesink capacity by using near-isogenic lines differing in panicle size.The increase in sink capacity allowed more effective mobilizationof pre-stored carbohydrates and post-anthesis photoassimilates tograins, and hence increased the harvest index. Nevertheless, theyield potential of YTH183 is still lower than that of the current

highest-yielding cultivar in Japan, Takanari (Table 2), probably dueto lower post-anthesis biomass accumulation, lower radiation useefficiency and lower leaf N status (data not shown). This differenceclearly indicates that the next step in increasing the yield potential

0.

gs at 77 DAS(mol m−2 s−1)

RUE at 67–90DAS (g MJ−1)

CGR at 67–90DAS(g m−2 d−1)

Biomass at 90DAS (g m−2)

0.48 ns 0.92 b 16.8 b 698 b0.55 1.23 a 21.8 a 750 ab0.54 1.18 a 22.2 a 823 a

5). CGR, crop growth rate; RUE, radiation use efficiency; gs, stomatal conductance;

136 Y. Kato et al. / Field Crops Research 123 (2011) 130–138

NS

NS

NS

0

5

10

15

20

YTHIR64254

YTH183

YTH304

YTH146

YTH243

FloodedAerobicWater-limited

(a) Shootdry weight (g/pot)

bab

a a

b b

NSNS

0

10

20

30

40

50

60

70

YTHIR64254

YTH183

YTH304

YTH146

YTH243

(c) Rootlength at 15-90 cm (m/pot)

NS

NS

NS

0

100

200

300

400

YTHIR64254

YTH183

YTH304

YTH146

YTH243

(b) Root length at 0-15 cm (m/pot)

Fig. 3. Shoot and root responses to hydrological conditions in the pot experiment atIbm

ootew

gcdimtaTs

-80

-60

-40

-20

0 908070605040302010

Soi

l wat

er p

oten

tial (

kPa)

Days after sowing

20 cm40 cm

Fig. 4. Time course of soil water potential at 20 and 40 cm depth in aerobic fieldof University of Tokyo in 2010 summer. Bars indicate standard errors (n = 4). The

RRI during the 2009 dry season. Bars indicate standard errors (n = 6). Means followedy the same letter are not significantly different among genotypes (P < 0.05). NSeans genotypic difference is not statistically significant for a given environment.

f an elite indica cultivar such as IR64 should target an improvementf photosynthetic capacity so as to improve biomass accumula-ion. Towards improvement of the photosynthetic efficiency, muchffort is currently made to introduce the C4 photosynthesis path-ay into elite rice cultivars at IRRI (Hibberd et al., 2008).

In addition to superior performance under conventionally irri-ated culture, YTH183 showed good adaptation to water-savingulture: it out-yielded IR64 by 92–102% under moderate watereficit in aerobic culture (Fig. 2; 143 vs. 72 g m−2). Crop phys-

ologists suggest that rice is mainly a “drought avoider”, andechanisms for dehydration avoidance rather than dehydration

olerance would be important for plant adaptation to water short-ge in most cases (Kamoshita et al., 2008; Serraj et al., 2009).he detailed evaluation of rooting patterns in the pot experimenthowed the specific adaptive response of YTH183 and YTH254 to

tensiometer could not measure soil water potential below −80 kPa at 20 cm depth.

drying soil as root exploration into deeper soil, as well as the con-stitutively deep root growth (Fig. 3). This adaptive root growth wasalso confirmed in the field experiment (Table 6). Rice has a con-siderably compact and shallow root system compared with otherdryland crops (Angus et al., 1983). In rainfed lowland and aerobicrice culture, root length below 30 cm depth has frequently beenproposed as a putative trait of adaptation, as it contributes to morewater uptake during dry spells (Bernier et al., 2009; Henry et al.,2011; Hirayama et al., 2007). The quantity of deep roots required forimproving dehydration avoidance traits (i.e., transpiration rate andplant water status) depends on canopy size and on the intensity andfrequency of soil water deficit. At near-full canopy coverage in theaerobic field, the 79% increase in root length density at 30–70 cmdepth in YTH183 increased leaf water potential and crop growthrate by 30% relative to IR64 under 23 days’ dry-down stress (whensoil moisture at 40 cm reached −70 kPa).

We physiologically dissected the superior performance ofYTH183 in aerobic culture. Two remarkable characteristics, viz.enhanced sink capacity via large grain size and rooting responseto water deficit, were simultaneously introgressed from the donor.YTH183 and YTH254, which both showed improved root growthunder water deficiency, have one common introgressed segmenton the long arm of chromosome 6 (Fig. 5). Interestingly, QTLsfor root length under a deficiency of P (Shimizu et al., 2008)or N (Obara et al., 2010) have been fine-mapped in this region(RM1370–RM8242). As the donors of YTH183 and YTH254 are sib-lings (New Plant Type cultivars), the improved root growth of thesegenotypes might be derived from the same QTL on chromosome6. Another introgressed segment in YTH183 contains a major QTLfor grain size (qSW5) on chromosome 5 (Shomura et al., 2008).Our previous analysis using the 334 introgression lines detecteda QTL for grain weight (qGW5) in the same region as qSW5; theintrogressed segment increased the weight (Fujita et al., 2009).Furthermore, we mapped qGW5 vicinity to qSW5 in a YTH199 (asibling of YTH183) × IR64 F2 population (Fujita et al., 2010b). Theresults indicate that the introgressed segment on chromosome 5(RM3193–RMw513) confers the greater grain weight in YTH183.Although there remain nine genomic regions derived from thedonor in the genome of YTH183, we suggest that the above twointrogressed segments might, at least in part, be responsible forthe higher yield potential and greater dehydration avoidance inYTH183. To ascertain the role of each introgressed segment in root

morphology, sink capacity and hence yield potential, analysis of aset of near-isogenic lines would be prerequisite. These resources are

Y. Kato et al. / Field Crops Research 123 (2011) 130–138 137

31 1211109876542

IR64IR69093-41-2-3-2

(a) YTH183

Chromosome31 1211109876542

IR64IR69125-25-3-1-1 (donor)

(b) YTH254

Chromosome

Fuji

ce

5

yoaidpwlTmtatc

A

E(taifbAL

R

A

(donor)

Fig. 5. Graphical genotypes of (a) YTH183 and (b) YTH254 adapted from

urrently being developed from the IR64 introgression lines (Fujitat al., 2010b), and will be available in the near future.

. Conclusions

We have identified one rice introgression line with a 26% higherield potential than the elite cultivar IR64 (895 vs. 712 g m−2

n average) by advanced backcross breeding. This increase isttributable to enhanced sink capacity via enlarged grain size. Thentrogression line simultaneously showed an adaptive response torying soils by increased root growth at depth, which maintainedlant water status and dry matter accumulation under moderateater deficit. As a consequence, the grain yield of the introgression

ine was double that of IR64 under water deficit in aerobic culture.he multiple physiological mechanisms identified in these experi-ents coincide with the genomic regions identified previously for

hese traits. At least two genomic regions on chromosomes 5 and 6re suggested for further studies of yield potential of indica rice cul-ivars and of adaptation to soil moisture depletion in water-savingulture.

cknowledgements

We thank R. Torres, L. Holongbayan, N. Turingan, L. Satioquia,. Mico, N. Driz, J. Villa (IRRI), K. Ichikawa, R. Soga, and K. YatsudaUniversity of Tokyo) for their technical assistance in carrying outhese experiments. The IR64 introgression lines were developed inn IRRI–Japan collaborative research project. This work was fundedn part by a Grant-in-Aid for Scientific Research (No. 20780010)rom the Japan Society for the Promotion of Science (to YK), andy the Generation Challenge Program grant “Targeting Drought-voidance Root Traits to Enhance Rice Productivity under Water-imited Environments” (to RS).

eferences

dachi, S., Tsuru, Y., Nito, N., Murata, K., Yamamoto, T., Ebitani, T., Ookawa, T.,Hirasawa, T., 2010. Identification and characterization of genomic regions on

ta et al. (2010a). For DNA marker information, see Fujita et al. (2010a).

chromosome 4 and 8 that control the rate of photosynthesis in rice leaves. J.Exp. Bot. 62, 1927–1938.

Akita, S., 1999. Agronomic and eco-physiological approach for breaking through thestagnation of rice yield. In: Horie, T., Geng, S., Amano, T., Inamura, T., Shiraiwa, T.(Eds.), Proceedings of International Symposium “World Food Security and CropProduction Technologies for Tomorrow”. 8–9 October 1998, Kyoto, Japan, pp.77–80.

Angus, J.F., Hasegawa, S., Hsiao, T.C., Liboon, P.S., Zandstra, H.G., 1983. The waterbalance of post monsoonal dryland crops. J. Agric. Sci. 101, 699–710.

Ashikari, M., Sakakibara, H., Lin, S.Y., Yamamoto, T., Takashi, T., Nishimura, A., Ange-les, E.R., Qian, Q., Kitano, H., Matsuoka, M., 2005. Cytokinin oxidase regulatesrice grain production. Science 309, 741–745.

Atlin, G.N., Lafitte, H.R., Tao, D., Laza, M., Amante, M., Courtois, B., 2006. Developingrice cultivars for high-fertility upland systems in the Asian tropics. Field CropsRes. 97, 43–52.

Bernier, J., Serraj, R., Kumar, A., Venuprasad, R., Impa, S., Gowda, V., Oane, R., Spaner,D., Atlin, G., 2009. The large-effect drought-resistance QTL qtl12.1 increaseswater uptake in upland rice. Field Crops Res. 110, 139–146.

Bouman, B.A.M., Lampayan, R.M., Tuong, T.P., 2007. Water Management in Rice:Coping with Water Scarcity. IRRI, Los Banos, Philippines, 54 pp.

Bueno, C.S., Bucourt, M., Kobayashi, N., Inubushi, K., Lafarge, T., 2010. Water produc-tivity of contrasting rice genotypes grown under water-saving conditions in thetropics and investigation of morphological traits for adaptation. Agric. WaterManage. 98, 241–250.

Centritto, M., Lauteri, M., Monteverdi, M.C., Serraj, R., 2009. Leaf gas exchange, carbonisotope discrimination and grain yield in contrasting rice genotypes subjectedto water deficits during reproductive stage. J. Exp. Bot. 60, 2325–2339.

Collins, N.C., Tardieu, F., Tuberosa, R., 2008. Quantitative trait loci and crop perfor-mance under abiotic stress: where do we stand? Plant Physiol. 147, 469–486.

Farooq, M., Kobayashi, N., Ito, O., Wahid, A., Serraj, R., 2010. Broader leaves resultin better performance of indica rice under drought stress. J. Plant Physiol. 167,1066–1075.

Fujita, D., Santos, R.E., Ebron, L.A., Yanoria, M.J., Kato, H., Kobayashi, S., Uga, Y.,Araki, E., Takai, T., Tsunematsu, H., Imbe, T., Khush, G.S., Brar, D.S., Fukuta, Y.,Kobayashi, N., 2009. Development of introgression lines of an indica-type ricevariety, IR64, for unique agronomic traits and detection of the responsible chro-mosomal regions. Field Crops Res. 114, 244–254.

Fujita, D., Santos, R.E., Ebron, L.A., Telebanco-Yanoria, M.J., Kato, H., Kobayashi, S.,Uga, Y., Araki, E., Takai, T., Tsunematsu, H., Imbe, T., Fukuta, Y., Kobayashi, N.,2010a. Development and characterization of introgression lines of an indica-type rice variety, IR64, for unique agronomic traits. JIRCAS Working Report 66,pp. 1–95.

Fujita, D., Tagle, A.G., Koide, Y., Fukuta, Y., Kobayashi, N., 2010b. Detection ofQTLs for grain weight and development of near-isogenic lines by usingindica-type rice variety, IR64, introgression lines. Breed. Res. 12 (extra 2),

287.

Fukuoka, S., Saka, N., Koga, H., Ono, K., Shimizu, T., Ebana, K., Hayashi, N., Takahashi,A., Hirochika, H., Okuno, K., Yano, M., 2009. Loss of function of a proline-containing protein confers durable disease resistance in rice. Science 325,998–1001.

1 Resea

G

G

H

H

H

H

K

K

L

L

L

M

O

O

38 Y. Kato et al. / Field Crops

owda, V.R.P., Henry, A., Yamauchi, A., Shashidhar, H.E., Serraj, R., 2011. Root biologyand genetic improvement for drought avoidance in rice. Field Crops Res. 122,1–13.

uan, Y.S., Serraj, R., Liu, S.H., Xu, J.L., Ali, J., Wang, W.S., Venus, E., Zhu, L.H., Li,Z.K., 2010. Simultaneously improving yield under drought stress and non-stressconditions: a case study of rice (Oryza sativa L.). J. Exp. Bot. 61, 4145–4156.

enry, A., Gowda, V.R.P., Torres, R.O., McNally, K.L., Serraj, R., 2011. Variation in rootsystem architecture and drought response in rice (Oryza sativa L.): phenotypingof the OryzaSNP panel in rainfed lowland fields. Field Crops Res. 120, 205–214.

ibberd, J.M., Sheehy, J.E., Langdale, J.A., 2008. Using C4 photosynthesis to increasethe yield of rice—rationale and feasibility. Curr. Opin. Plant Biol. 11, 228–231.

irayama, M., Nemoto, H., Hirasawa, H., 2007. Relation between root system anddrought resistance in Japanese upland rice (Oryza sativa L.) varieties withmedium to late maturing under field conditions. Jpn. J. Crop Sci. 76, 245–252.

umphreys, E., Kukal, S.S., Christen, E.W., Hira, G.S., Singh, B., Yadav, S., Sharma,R.K., 2010. Halting the groundwater decline in north-west India—which croptechnologies will be winners? Adv. Agron. 109, 155–217.

amoshita, A., Babu, R.C., Boopathi, N.M., Fukai, S., 2008. Phenotypic and genotypicanalysis of drought-resistance traits for development of rice cultivars adaptedto rainfed environments. Field Crops Res. 109, 1–23.

ato, Y., Okami, M., Katsura, K., 2009. Yield potential and water use efficiency ofaerobic rice (Oryza sativa L.) in Japan. Field Crops Res. 113, 328–334.

afarge, T., Bueno, C.S., 2009. Higher crop performance of hybrid rice over eliteinbreds of same phenology under contrasted weather conditions: II. Does sinkregulation, rather than sink size, play a major role? Field Crops Res. 112,238–244.

afitte, H.R., Li, Z.K., Vijayakumar, C.H.M., Gao, Y.M., Shi, Y., Xu, J.L., Fu, B.Y., Yu, S.B.,Ali, A.J., Domingo, J., Maghirang, R., Torres, R., Mackill, D., 2006. Improvementof rice drought tolerance through backcross breeding: evaluation of donors andselection in drought nurseries. Field Crops Res. 97, 77–86.

uo, L.J., 2010. Breeding for water-saving and drought-resistance rice (WDR) inChina. J. Exp. Bot. 61, 3509–3517.

ae, T., Inaba, A., Kaneta, Y., Masaki, S., Sasaki, M., Aizawa, M., Okawa, S., Hasegawa,S., Makino, A., 2006. A large-grain rice cultivar, Akita 63, exhibits high yieldswith high physiological N-use efficiency. Field Crops Res. 97, 227–237.

bara, M., Tamura, W., Ebitani, T., Yano, M., Sato, T., Yamaya, T., 2010. Fine-mappingof qRL6.1, a major QTL for root length of rice seedlings grown under a wide

range of NH4

+ concentrations in hydroponic conditions. Theor. Appl. Genet. 121,535–547.

hsumi, A., Takai, T., Ida, M., Yamamoto, T., Arai-Sanoh, Y., Yano, M., Ando, T.,Kondo, M., 2011. Evaluation of yield performance in rice near-isogenic lineswith increased spikelet number. Field Crops Res. 120, 68–75.

rch 123 (2011) 130–138

Peng, S., Bouman, B., Visperas, R.M., Castaneda, A., Nie, L., Park, H.K., 2006. Compari-son between aerobic and flooded rice in the tropics: agronomic performance inan eight-season experiment. Field Crops Res. 96, 252–259.

Peng, S., Khush, G.S., Virk, P., Tang, Q., Zou, Y., 2008. Progress in ideotype breedingto increase rice yield potential. Field Crops Res. 108, 32–38.

Peng, S., Tang, Q., Zou, Y., 2009. Current status and challenges of rice production inChina. Plant Prod. Sci. 12, 3–8.

SAS Institute, 2003. SAS User’s Guide: Statistics. Version 9.1 2002–2003. SAS Inst.,Inc., Cary, NC, USA.

Septiningsih, E.M., Pamplona, A.M., Sanchez, D.L., Neeraja, C.N., Vergara, G.V., Heuer,S., Ismail, A.M., Mackill, D.J., 2009. Development of submergence-tolerant ricecultivars: the Sub1 locus and beyond. Ann. Bot. 103, 151–160.

Serraj, R., Kumar, A., McNally, K.L., Slamet-Loedin, I., Bruskiewich, R., Mauleon, R.,Cairns, J., Hijmans, R.J., 2009. Improvement of drought resistance in rice. Adv.Agron. 103, 41–98.

Serraj, R., McNally, K.L., Slamet-Loedin, I., Kohli, A., Haefele, S.M., Atlin, G., Kumar,A., 2011. Drought resistance improvement in rice: an integrated genetic andresource management strategy. Plant Prod. Sci. 14, 1–14.

Shimizu, A., Kato, K., Komatsu, A., Motomura, K., Ikehashi, H., 2008. Genetic analysisof root elongation induced by phosphorus deficiency in rice (Oryza sativa L.):fine QTL mapping and multivariate analysis of related traits. Theor. Appl. Genet.117, 987–996.

Shomura, A., Izawa, T., Ebana, K., Ebitani, T., Kanegae, H., Konishi, S., Yano, M., 2008.Deletion in a gene associated with grain size increased yields during rice domes-tication. Nat. Genet. 40, 1023–1028.

Song, X.J., Huang, W., Shi, M., Zhu, M.Z., Lin, H.X., 2007. A QTL for rice grain widthand weight encodes a previously unknown RING-type E3 ubiquitin ligase. Nat.Genet. 39, 623–630.

Takai, T., Matsuura, S., Nishio, T., Ohsumi, A., Shiraiwa, T., Horie, T., 2006. Rice yieldpotential is closely related to crop growth rate during late reproductive period.Field Crops Res. 96, 328–335.

Tanksley, S.D., Nelson, J.C., 1996. Advanced backcross QTL analysis: a method forthe simultaneous discovery and transfer of valuable QTLs from unadaptedgermplasm into elite breeding lines. Theor. Appl. Genet. 92, 191–203.

Tuong, T.P., Bouman, B.A.M., Mortimer, M., 2005. More rice, less water—integratedapproaches for increasing water productivity in irrigated rice-based systems inAsia. Plant Prod. Sci. 8, 231–241.

Xu, J.L., Lafitte, H.R., Gao, Y.M., Fu, B.Y., Torres, R., Li, Z.K., 2005. QTLs for droughtescape and tolerance identified in a set of random introgression lines of rice.Theor. Appl. Genet. 111, 1642–1650.

Yoshida, S., 1972. Physiological aspects of grain yield. Ann. Rev. Plant Physiol. 23,437–464.