Crop Science Grow Your Career with CSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in Crop Science
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (24)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kamoshita, A.
Right arrow Articles by Wade, L. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kamoshita, A.
Right arrow Articles by Wade, L. J.
Agricola
Right arrow Articles by Kamoshita, A.
Right arrow Articles by Wade, L. J.
Related Collections
Right arrow Rice
Right arrow Cell Biology & Molecular Genetics
Right arrow Root Development
Crop Science 42:255-265 (2002)
© 2002 Crop Science Society of America

CELL BIOLOGY & MOLECULAR GENETICS

Effects of Phenotyping Environment on Identification of Quantitative Trait Loci for Rice Root Morphology under Anaerobic Conditions

A. Kamoshitaa, Jingxian Zhangc, J. Siopongcoa, S. Sarkarungb, H. T. Nguyenc and L. J. Wade*,a

a Crop, Soil and Water Sciences Division, International Rice Research Institute, DAPO Box 7777, Metro Manila, The Philippines
b Plant Breeding, Genetics and Biochemistry Division, International Rice Research Institute, DAPO Box 7777, Metro Manila, The Philippines
c Plant Molecular Genetics Laboratory, Dep. of Plant and Soil Science, Texas Tech Univ., Lubbock, TX 79409-2122

* Corresponding author (l.wade{at}cgiar.org)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
In the rainfed lowlands, rice (Oryza sativa L.) develops roots under anaerobic soil conditions with ponded water, prior to exposure to aerobic soil conditions and water stress. Constitutive root system development in anaerobic soil conditions has been reported to have a positive effect on subsequent expression of adaptive root traits and water extraction during water stress. We examined effects of phenotyping environment on identification of quantitative trait loci (QTLs) for constitutive root morphology traits using 220 doubled-haploid lines (DHLs) from the cross of ‘CT9993-5-10-1-M’ (CT9993; japonica, upland adapted) x ‘IR62266-42-6-2’ (IR62266; indica, lowland adapted) in four greenhouse experiments. Broad sense heritability (h2) was 75, 60, and 64% on average for shoot biomass, deep root morphology, and root thickness traits, respectively. Quantitative trait loci analysis identified 18 genomic regions associated with deep root morphology traits, but only three were identified consistently across experiments. Three out of a total of eight QTLs for root thickness traits were found in more than one experiment. The maximum genetic effects caused by a single QTL were increments of 0.05 g of deep root mass below a 30-cm soil depth, 0.9% of deep root ratio, 1.6 cm of rooting depth, and 0.09 cm of root thickness, with phenotypic variation explained by a single QTL ranging from 6.8 to 51.8%. The results demonstrate the importance of phenotyping environment and suggest prospects for selection of QTLs for deep root morphology, root thickness, and vigorous seedling growth under anaerobic conditions to improve the constitutive root system of rainfed lowland rice. There was some consistency in QTL regions identified, despite the presence of QTL x environment interactions.

Abbreviations: CT9993, ‘CT9993-5-10-1-M’ • DAS, days after sowing • DH1 to DH4, Experiments 1 to 4 • DHL, doubled-haploid lines • G, genotypic variation • G x E, genotype x environment • h2, broad-sense heritability • IR6226, ‘IR62266-42-6-2’ • IRRI, International Rice Research Institute • LOD, log of the odds • PVC, polyvinyl chloride • QTL, quantitative trait locus


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
RAINFED LOWLAND RICE is grown in bunded fields, with soil conditions ranging from anaerobic to aerobic during crop growth (Wade et al., 1998). Rice plants have to develop their root system under anaerobic flooded conditions, which results in different expressions of root anatomy and gross root morphology from those under an upland aerobic environment. Champoux et al. (1995) compared selected lines in both aerobic and anaerobic conditions and found significant interactions between rooting depth and water regime. In rainfed lowlands, rice generally encounters water deficit late in the growing season. We will define constitutive traits as those which are expressed under anaerobic, non-waterstressed conditions, do not require water stress for their expression, and may demonstrate variation that is subsequently modified by adaptive traits. Adaptive traits will be defined as those, such as root penetration index or osmotic adjustment (Zhang et al., 2001), which are expressed in response to water deficit or soil physical/chemical barriers. Less research attention has been given to constitutive traits than to adaptive traits.

A deep and thick root system has been thought advantageous for improved drought tolerance in the rainfed lowland ecosystem, based on extrapolation from experience with upland rice (O'Toole, 1982; and Fukai and Cooper, 1995). Under anaerobic well-watered conditions, root system development had a positive effect on subsequent plant growth during progressive water stress (Azhiri-Sigari et al., 2000; Kamoshita et al., 2000; and Hoque and Kobata, 1998). Azhiri-Sigari et al. (2000) and Kamoshita et al. (2000) demonstrated genotypic variation in constitutive root traits, and subsequent responses of adaptive root traits, especially in deeper soil layers. Greater root elongation to depth resulted in improved water extraction. Improved seedling vigor was also valuable to growth afterward (Mitchell et al., 1998). In the field, roots are generally shallow in rainfed lowlands (Pantuwan et al., 1997), but genotypes differ in root growth in deeper layers (Samson and Wade, 1998). Despite having fewer roots in deeper layers, rainfed lowland rice can extract water from below a 15-cm soil depth (Wade et al., 1999).

Genetic improvement in the root system of rainfed lowland rice has been slow, partly because of the lack of a reliable screening system for phenotyping. Several groups are currently identifying QTLs associated with rooting characteristics (Ray et al., 1996; Champoux et al., 1995; Yadav et al., 1997; Price and Tomos, 1997; Nguyen et al., 1997). Maximizing h2 is critical, and measurement accuracy depends on the appropriateness of the phenotyping environment to the rainfed lowland target. Previous work on QTL for root system morphology was conducted under hydroponic (Price and Tomos, 1997) or upland conditions (Ray et al., 1996; Champoux et al., 1995; Yadav et al., 1997; Shashidhar et al., 1999; Zhang et al., 2000). There were no QTL reports for anaerobic lowland conditions, nor for rainfed lowland water stress in the field. Trait manifestation is also affected by plant size and phenology, and both interact with the water stress imposed, and with air temperature and solar radiation even in the absence of water stress. Epistatic effects should be estimated (Wang et al., 1999a). For example, Yadav et al. (1997) indicated that putative QTL for root traits detected as main effects also had substantial interactions with other genes. Genotype x environment interactions (G x E) will certainly complicate QTL identification, with some QTLs specific to the screening conditions imposed or encountered (Jansen et al., 1995). Assessment of QTL x environment interactions is needed.

This study examines the phenotypic variation and QTL mapping for gross root morphology under anaerobic lowland conditions. This is the first report of QTL for root morphology in rice under anaerobic conditions. The aim is to quantify the effects of phenotyping environment defined as different planting dates on the identification of QTL for constitutive root morphology traits. The potential for genetic improvement of the root system of rainfed lowland rice is discussed.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Plant Population
A population of 220 anther culture derived DHLs from the cross CT9993 x IR62266 was developed at the International Rice Research Institute (IRRI), Los Baños, The Philippines (14°11' N, 121°15' E; 23 m altitude). The gross root morphology of the two parental lines was characterized under both stress and nonstress conditions in the greenhouse (Azhiri-Sigari et al., 2000) and in the field (Sarkarung et al., 1997; and Samson and Wade, 1998). Azhiri-Sigari et al. (2000) showed that CT9993 had a slower initial growth rate under anaerobic flooded conditions, but had thicker roots and, in the later vegetative stage, deeper roots than IR62266.

Setup of Pot Experiments
Root morphology was evaluated in four pot experiments with different sowing dates at the IRRI greenhouse (Table 1). Experiments 1 and 2 (DH1, DH2) were conducted with only 144 DHLs, using a 12 x 17 alpha design with three replicates. Experiments 3 and 4 (DH3, DH4) were conducted with all 220 DHLs, using a 15 x 15 alpha design with three replicates. Alpha design is a new class of incomplete block designs, and is suited to dealing with large numbers of treatments (Paterson and Williams, 1976). An evaporative gradient from the center to the side of the greenhouse was observed after DH2, so replicates were arranged perpendicular to this gradient in DH3 and DH4.


View this table:
[in this window]
[in a new window]
  		Table 1. Characteristics of the four screening environments (DH1–DH4).

 
Four or five presoaked seeds of each DHL were sown on the wet soil and thinned to one healthy seedling per pot by {approx}10 d after sowing (DAS). The sowing dates were 30 July 1996, 25 November 1996, 1 November 1997, and 16 May 1998 in the DH1, DH2, DH3, and DH4 experiments, respectively. Most of the lines were pregerminated before sowing, but several lines failed to germinate and were missing in each experiment (Table 1).

A cylindrical pot made of polyvinyl chloride (PVC), of 20-cm internal diameter and 55-cm depth with a plastic bag insert, was filled with 20 kg of air-dried Maahas clay soil (28% clay, 44% silt, and 28% sand; pH 5.2) (Wopereis, 1993). At first, 17 kg of soil was carefully put into the plastic bag inside the pot to eliminate any gaps between the plastic bag and the inner wall of the pot, and {approx}6 kg of water was added. The whole soil layer was then mixed with a wooden stick until standing water remained. When the soil surface shrank, another 3 kg of air-dried soil was put into the pot, more water was added, and the soil was puddled again. The exterior of the pot was covered with aluminum foil to minimize any rise in soil temperature in pots in the greenhouse, so that high-temperature effects on root growth were minimized (Nagai and Matsushita, 1963).

Sufficient levels of N fertilizer (1.26 g pot of N-1, as urea 46-0-0) were supplied based on the results of Regmi (1995). Solo-phos (0-18-0) equivalent of 0.33 g pot of P-1 and muriate potash (0-0-60) equivalent of 0.62 g pot of K-1 were also applied at puddling and mixed thoroughly into the puddled soil. The level of standing water was maintained at {approx}2 to 4 cm by watering daily. No disease or insect damage occurred.

Measurements
Minimum and maximum air temperature and soil temperature at a 5-cm depth from the soil surface at 0800 and 1500 h were recorded daily in the greenhouse. Solar radiation data came from the IRRI wetland meteorological station {approx}500 m from the greenhouse. The heatsum with a base temperature of 9 °C and average daily solar radiation during the experimental period were calculated (Table 1).

The plants were sampled at 30 DAS in DH1 as a preliminary trial to assess optimal sampling time for the root system to be developed below a 30-cm soil depth. On the basis of results of DH1, the plants were sampled later in the other experiments; at 45 DAS in DH2, 42 to 44 DAS in DH3, and 48 to 50 DAS in DH4 (Table 1). Tiller number and plant height were measured 1 d before the sampling dates. Plants were cut at the soil surface. The soil mass inside the plastic sleeve was slowly pulled out of the PVC pots and the soil divided into layers of 0 to 10, 10 to 20, 20 to 25, 25 to 30, 30 to 35, 35 to 40, 40 to 45, and 45 to 50 cm from the soil surface. Roots were carefully separated from the soil on the 1-mm sieve screen (Schuurman and Goedewaagen, 1965). The dry weight of each plant component was measured after drying at 70 °C for 4 to 5 d. Shoot biomass was determined as the sum of aboveground biomass and stem base below the soil surface. Total root mass and deep root mass below the 30-cm soil depth were obtained, and the deep root ratio, the proportion of the latter to the former, was calculated. Deep root per tiller was calculated by dividing deep root mass by the total number of tillers. Maximum rooting depth was calculated from the deepest soil layer where roots were present and the longest root measured in the layer. Root thickness was measured by microcaliper at a 0- to 10-cm and 20- to 25-cm soil depth for 7 to 10 randomly chosen primary roots. A total of seven traits were analyzed. These included shoot biomass, deep root mass, deep root ratio, deep root per tiller, rooting depth, and root thickness at either a 0- to 10-cm or 20- to 25-cm soil depth.

Statistical Analysis
Analysis of variance and calculation of means were conducted for the seven traits between the parents using Systat 7.0 (Statistical Package for the Social Sciences, Inc., 1996, 1997) and among the DHLs using the SAS software (SAS Institute, 1990). The DH1 and DH4 were analyzed as an alpha design to account for variation across blocks, and DH2 and DH3, where more lines were missing, were analyzed as a randomized block design. Broad-sense heritability was calculated from the estimates of genetic and residual variances derived from the expected mean squares of the analysis of variance:

[1]
where k was the number of replications. Those lines used for QTL analysis were tested in combined analysis of variance using the data sets of four experiments to compare the mean square of G x E interaction (defined as different planting dates) with that of genotypic variation. Pearson correlation was calculated between shoot and root traits for each experiment using Systat 7.0.

Map Construction and Quantitative Trait Loci Analysis
One hundred fifty-four DHL were used to construct the map with 315 (145 restricted fragment length polymorphism, 153 amplified fragment length polymorphism, and 17 microsatellite) markers at Texas Tech University, Lubbock, TX (Zhang et al., 2001). Total map distance is 1788 centimorgans. The map length for individual chromosomes or whole genome was comparable with those of Cho et al. (1998) and Harushima et al. (1998). The linkage map was constructed with MapMaker Macintosh Version 2.0 (E.I. DuPont De Nemours and Co., Wilmington, DE). Putative QTLs (main effect QTLs assuming no epistasis) for the traits were identified in both separate analysis for each experiment and combined analysis of 4 experiments by employing composite interval mapping based on QTLMapper (version 1.0) (Wang et al., 1999a,b). QTLMapper is a user-friendly computer software for mapping putative QTLs and epistatic QTLs, and for quantifying additive effects, epistatic effects, and QTL x environment interactions. Its unique feature allows us to control background genetic variation, which can be defined as noise arising from nonrandom sampling (because of small population size) of main effect and epistatic QTLs other than the one under investigation. In our analysis, background genetic variation was controlled for both main and interaction markers. The probability was set as 0.005 to select background main effect markers, and as 0.001 to select background interaction markers (Sripongpangkul et al., 2000). The lower probability was chosen for background interaction markers to minimize the probability of false positive epistasis. Selected background interaction markers with partial determination coefficients smaller than 0.05 were manually deleted. Walking speed was set as 0.05 Morgans.

Only a subset of lines (115, 100, 127, and 154 in DH1, DH2, DH3, and DH4 experiments, respectively) (Table 1) was selected for QTL analysis of individual environments (planting dates), because not all DHLs were genotyped. Combined analysis with 154 lines was conducted, with the matrix filled 81%, to estimate epistasis and to calculate relative and general contributions of additive effect, epistasis, and QTL x environment interaction (additive x environment and epistasis x environment interactions). Main effect QTLs and epistasis QTLs were declared significant at the thresholds of 0.005 and 0.001, respectively, with log of the odds (LOD) set higher than 1.9 and 4.0, respectively (Sripongpangkul et al., 2000). A relative contribution was calculated as the proportion of variance caused by a specific genetic source in the total phenotypic variance, taken as a heritability contributed by that genetic source. The general contribution for each genetic source was calculated from the relative contributions of all the putative QTLs involved.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Temperature and Solar Radiation
The average daily minimum and maximum temperatures in DH1, DH2, DH3, and DH4 were 29 to 39, 29 to 36, 25 to 33, and 27 to 36 °C, respectively. The average daily solar radiation was listed in Table 1. Heatsum in DH1 was smallest because of the earlier harvest date. Temperature regime in DH3 and solar radiation regime in DH2 were lower than in the other experiments (Table 1). Multiple regression analysis showed that 98% of the variation in average shoot biomass of DHLs across the four planting dates was explained by heatsum and cumulative solar radiation:

[2]
where SB, H, and R refer to shoot biomass, heatsum, and cumulative solar radiation, respectively.

Phenotypic Variation
Plant size in DH1 was only one-third of that in the other three experiments, due to earlier sampling (Table 2). Among the other three experiments, DH2 and DH3 had a smaller shoot biomass than DH4 because of lower solar radiation and air temperature. The root system in DH2 and DH3 was shallower than in DH4, whereas root thickness was similar.


View this table:
[in this window]
[in a new window]
  		Table 2. Mean values for CT9993 and IR62266, ranges in doubled haploid lines (DHLs), and broad sense heritability (h2) of the seven traits in (a) DH1, (b) DH2, (c) DH3 and (d) DH4 experiments, respectively. A dash (–) indicates data not measured.

 
CT9993 had a smaller shoot biomass and thicker roots than IR62266 in all the experiments. For the four deep root morphology traits, however, CT9993 had a greater deep root mass, deep root ratio, and deep root per tiller than IR62266 only in DH4, and CT9993 had smaller or comparable values in DH2 and DH3, showing crossover interaction. Small and generally not statistically significant differences occurred in rooting depth between the parents in each of the experiments.

Significant differences were observed among DHL progeny for all traits in all the experiments except for root thickness in the 20- to 25-cm soil depth in DH3, even though the parents did not differ in deep root morphology traits in DH3. Consequently, transgressive variation was large for deep root morphology traits in each experiment. For example, in DH4, the DHLs with the deepest root morphology had 0.82 g of deep root mass and 12.8% of deep root ratio, while those of CT9993 were 0.28 g and 6.4% (Table 2d).

The level of h2 was highest for shoot biomass (averaging 0.80), followed by root thickness (averaging 0.63), and smallest for deep root morphology traits (averaging 0.48) in DH1, DH2, and DH3 (Table 2a,b,c). In DH4, h2 for deep root morphology traits was comparable with those of shoot biomass and root thickness traits, all averaging 0.72 (Table 2d). Heritability for tiller number was 0.82, 0.86, 0.85, and 0.84 in the four experiments.

In combined analysis across all four experiments (Table 3), G x E interaction was statistically significant for all the traits among progeny, but its mean square was smaller than that of genotypic variation (G). The ratio of G x E:G mean squares was <0.5 for shoot biomass and root thickness, but >0.5 for deep root mass, deep root ratio, and deep root per tiller.


View this table:
[in this window]
[in a new window]
  		Table 3. Mean square of genotype by environment interaction (G x E) compared with that of genotypic variation (G) in combined analysis of the four experiments among 139 selected doubled haploid lines from CT9993/IR62266.

 
In almost all cases, phenotypic correlations between shoot and root traits were significant, but with different coefficients of determination (Table 4). There was a strong phenotypic correlation between shoot biomass and total root mass, with r > 0.60, except in DH1. Deep root mass was correlated with shoot biomass, with a r of 0.2 to 0.3, also with the exception of DH1. Correlations between shoot biomass and root thickness in 0- to 10-cm and 20- to 25-cm depths were smaller.


View this table:
[in this window]
[in a new window]
  		Table 4. Phenotypic correlation (r2) between some of the root and shoot traits across 4 experiments (DH1–DH4) in CT9993/IR62266.

 
Quantitative Trait Loci Analysis
Table 5 presents the results of putative QTL from the separate analysis of each experiment. The lowest LOD score was 1.9, but most of the QTL had a LOD score higher than 4.0. For shoot biomass, seven QTLs were identified, with the phenotypic variation explained by a single QTL ranging from 7.7 to 56.8%. Only one QTL, G257-RM21 in Chromosome 11, was found in DH1 and DH2, explaining 27.8 and 32.2% of the phenotypic variation, respectively. In combined analysis, five of the seven QTLs were significant. For the four deep root traits, 18 QTLs were identified, with the phenotypic variation explained by a single QTL ranging from 4.7 to 51.8%. Only three QTLs, C813-RG957 in Chromosome 1, RG437-ME10_18 in Chromosome 2, and C119-C859 in Chromosome 5, were found in more than one experiment. RG437-ME10_18 was identified in three of the four deep root traits, and explained a larger percentage of the phenotypic variation than C813-RG957 and C119-C859. In combined analysis, 11 of the 18 QTLs were significant. For root thickness, eight QTLs were identified, with the phenotypic variation explained by a single QTL ranging from 6.8 to 36.4%. Three QTLs in regions EM13_3-RG158 and EM18_13-ME9_7 in Chromosome 2, and RG476-RG214 in Chromosome 4, were found in more than one experiment. In combined analysis, five of the eight QTLs were significant.


View this table:
[in this window]
[in a new window]
  		Table 5. Chromosome (Chr.) and marker intervals that were likely to contain QTLs (P < 0.005), approximate positions (Pos.) of QTLs, log of the odds (LOD) score, effect (A) and relative contributions (r2) of QTLs for the seven traits evaluated in CT9993/IR62266 across DH1, DH2, DH3, and DH4 experiments.

 
The approximate positions of putative QTLs that were significant in both separate and combined analysis are shown in Fig. 1 . A QTL for shoot biomass in DH4 and QTL for all of the four deep root traits and root thickness at the 20- to 25-cm depth in DH3 and/or DH4 were located near ME10_18 in Chromosome 2. A QTL for shoot biomass in DH3 and QTLs for rooting depth in DH3 and root thickness at the 0- to 10-cm depth in DH3 and DH4 were located near TGMSP2 in Chromosome 2. A QTL for shoot biomass in DH1 and DH2 and QTL for deep root mass and deep root ratio in DH2 were located near G257 in Chromosome 11.




View larger version (91K):
[in this window]
[in a new window]
  		Fig. 1. Approximate position of putative quantitative trait loci (QTLs) for shoot biomass, deep root mass below 30-cm soil depth, deep root ratio, deep root per tiller, maximum rooting depth, root thickness at 0- to 10-cm soil depth, and root thickness at 20- to 25-cm soil depth in CT9993/IR62266. From the QTLs identified in any of the four experiments in Table 5, only those confirmed in combined analysis with log of the odds > 2.0 were presented. Quantitative trait loci repeatedly identified in at least two experiments were marked with the letter R above.

 
All of the QTLs associated with shoot biomass, deep root traits, or root thickness in Fig. 1 interacted with the other QTLs or marker intervals that were not linked with the traits under study, showing significant epistasis effects (P = 0.001), with LOD scores >4.0. In total, 9, 35, and 16 pairs of epistasis were found for shoot biomass, deep root morphology, and root thickness traits, respectively. From them, 7, 21, and 11 pairs involved putative QTLs identified in Fig. 1, respectively. All of the putative QTLs for each trait were involved in epistasis either for the same traits or for other traits. However, most of those epistatic interactions were of such small phenotypic impact as to be relatively unimportant. The number and contribution of epistasis consisting of pairs of QTLs that were not involved in QTL main effects (non-main effect QTLs) were also generally small (Table 6). Compared with shoot biomass and deep root morphology traits, epistasis consisting of non-main effect QTL had greater importance for root thickness at the 0- to 10-cm soil depth (0.033, Table 6), but the phenotypic impact was again relatively unimportant.


View this table:
[in this window]
[in a new window]
  		Table 6. Number and general contribution (Gc) of epistasis quantitative trait loci (QTLs) for each trait identified in combined analysis. Both those epistasis involving main effect QTLs and those consisting of only non-main effect QTLs are presented. Marker loci of epistasis of non-main effect QTLs are listed.

 
Combined analysis of all four experiments showed the general contribution of main additive effect, epistasis, additive x environment interaction, and epistasis x environment interaction (Table 7). For shoot biomass, additive effect and epistasis were comparable and QTL x environment interactions were small. For deep root mass and deep root per tiller, additive effect tended to be larger than epistasis, but comparable with additive x environment interaction. For deep root ratio, additive effect, epistasis, additive x environment interaction, and epistasis x environment interaction were comparable in size. For rooting depth and root thickness, additive effect was larger than the other components.


View this table:
[in this window]
[in a new window]
  		Table 7. General contributions of additive effect, epistasis, additive x environment (E) interaction and epistasis x E interaction for seedling vigor, deep root, and thick root traits in combined analysis of four experiments.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Interaction with Environment
Though CT9993 has been qualitatively characterized as having deep roots in breeding nurseries (Sarkarung et al., 1997), its expression was inconsistent here. In these experiments, CT9993 always had thicker roots and smaller shoot biomass than IR62266. At lower temperature and solar radiation in DH2 and DH3, CT9993 did not express its deep root mass by the time of sampling (42–45 DAS). In contrast in DH4, CT9993 distributed a greater proportion of biomass to deeper roots and had a larger deep root mass than IR62266. Wade et al. (2000) and Azhiri-Sigari et al. (2000) reported that CT9993, an upland adapted line, was slower in shoot and root growth compared with rainfed lowland genotypes under anaerobic lowland conditions. Consequently, CT9993, with its suboptimal adaptation to anaerobic lowland conditions, was less vigorous, which may have limited translocation of assimilate for the development of deeper roots in DH2 and DH3. In DH4, with the favorable weather conditions that enabled higher levels of biomass production and assimilate supply, the greater capacity of CT9993 for deep root development was expressed.

The crossover interaction for deep root traits between the two parents was associated with relatively large G x E interaction for deep root traits among their progeny, and a comparable size of putative QTL and the QTL x environment interactions. Most of the QTLs for deep root traits were found in the DH4 experiment, where CT9993 was able to express its deep root character, and the favorable alleles were coming from both parents. On the other hand, QTLs in TGMSP2-ME9_7 in Chromosome 2 and in G257-RM21 in Chromosome 11, which were found in the DH2 and DH3 experiments, were the same as the QTLs for shoot biomass, and the favorable alleles were coming only from the IR62266 parent. Consequently, phenotypic expression of root traits was influenced by environment and plant age, which had important implications for the capacity to detect QTLs associated directly or indirectly with the trait of interest. Unless appropriate conditions are chosen for phenotyping and results are interpreted carefully, QTLs may be misassigned. Care must be taken to ensure those QTLs associated with a related trait are not assigned a role in directly controlling a target trait. For example, shoot biomass affected gross root morphology in DH2 and DH3. Further, improved interpretation of expression of favorable alleles for deep root traits from CT9993 was partly dependent on understanding generated from related physiological studies (Wade et al., 2000; Azhiri-Sigari et al., 2000).

The significance of phenotyping conditions for the identification of QTL for deep root traits, even in the absence of water stress, emphasizes the importance of choosing conditions comparable with the critical characteristics of the target environment in rainfed lowland fields (Wade et al., 1998). QTLs identified during the rainy season may be more important because this is when rainfed lowland rice is usually cultivated. Further, the extent of adaptation of an upland adapted parental line with a deep root system to anaerobic lowland conditions needs to be considered when choosing conditions for phenotyping of deep root traits using populations derived from a cross between an upland rice and a lowland rice under anaerobic lowland conditions. From the evidence presented, such measurement should be delayed until shoot dry weight was at least 30 g plant-1, to ensure adequate dry matter was available for expression of deep root traits. The expression of deep root morphology under anaerobic lowland conditions needs to be studied further, in relation to assimilate translocation from shoot to root and response to limited availability of oxygen for root growth.

Identification of Quantitative Trait Loci
In spite of large environmental effects, even in anaerobic well-watered conditions, three QTLs each for deep and thick root traits were identified across experiments. These QTLs could potentially be used for introgression in a marker-assisted selection program. These data suggest that QTLs near R2147 in Chromosome 1 and ME10_18 in Chromosome 2 could improve deep root ratio by 0.67 and 0.93%, respectively. A QTL near C119 in Chromosome 5 could increase rooting depth by 1.2 cm. Quantitative trait loci near EM133 and TGMSP2 in Chromosome 2 and near RG476 in Chromosome 4 could increase root thickness by 0.04 to 0.09 cm. In comparison with the values reported by Yadav et al. (1997) (e.g., effects on rooting depth ranging from 3.5–4.9 cm and those on root thickness ranging from 0.02–0.03 cm), the effects of QTLs on root thickness in this study were larger due to slightly longer growth duration, but those on deep root morphology were smaller due to limited oxygen availability in anaerobic saturated soils which may have prevented full expression of genes for deeper root morphology (Armstrong et al., 1991). The advantages of deeper root growth during progressive stress for water extraction from soil and for maintenance of plant growth were demonstrated in a pot system that simulated rainfed lowland conditions (Kamoshita et al., 2000). The extent of benefit derived from a deeper constitutive root system must also be quantified in rainfed lowland field conditions.

ME10_18 and TGMSP2 regions in Chromosome 2 contained QTLs for both root and shoot traits. These regions may contain clusters of QTL (Paterson et al., 1990; and Yadav et al., 1997), which can be resolved by fine mapping, or alternatively, the segment may affect both traits pleiotropically.

All putative QTLs interacted with the other QTLs either for the same traits or for other traits. For example, the QTLs for shoot biomass, such as those near ME10_18, TGMSP2, RZ682, and RZ536, interacted with QTLs for deep root morphology and root thickness traits. The contribution of epistasis consisting of pairs of QTLs that were not linked was small. Therefore, selection of the QTLs in a scheme of marker-assisted selection may be focused on those putative QTL for constitutive root traits and for shoot biomass that may affect expression of the root system through epistasis.

The relationship between QTLs for the root system and major genes previously reported has not been well studied. Plant hormones such as cytokinin or gibberellic acids are known to affect not only shoot growth or plant height but also root morphology (Takagi, 1990; Tanimoto, 1990; and Klepper, 1991). Several major dwarfing or semi-dwarfing genes already mapped in rice (Huang et al., 1996) were in close proximity to QTL for deep root traits in our study. For example, the dwarf Genes d-30 and d-31 were placed near RG437-RG171 in Chromosome 2 and near RG214 in Chromosome 4, respectively. Robertson (1985) and Huang et al. (1996) provided evidence that QTL and major genes were in the same loci but were different alleles. Physical mapping of the putative QTLs for deep root morphology traits would help to elucidate how rooting depth and deep root mass are genetically controlled at the molecular level.

Transgressive variation for deep root traits was large. With IR62266 possessing genes for better adaptation to anaerobic lowland conditions, and with CT9993 possessing genes for constituent deep and thick root traits, there is potential to be able to identify progenies with good initial shoot growth under anaerobic conditions that can also develop a deeper root system with thicker roots. Such progenies are important because lowland rice plants most likely establish and develop their root system in anaerobic conditions without water stress. Rice plants with an extensive root system could be advantageous for continued water extraction and growth. Together with the genetic dissection for adaptive root and shoot traits during a stress period, constitutive root morphology must be fully considered for developing drought-tolerant rice. Finally, it should be recognized that a capacity to develop deep roots in anaerobic lowland conditions may not always result in their expression in target field environments. In some situations, additional traits, such as a capacity to penetrate a hardpan, may also be needed (Samson and Wade, 1998).


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
This study identified, for the first time, QTLs for constitutive root morphology traits of rice under anaerobic conditions. Effects of environment (planting date) on identification of QTLs for root morphology traits were large and complex, as QTL x environment interactions were significant, even in the absence of water stress. This result emphasizes the importance of defining conditions for phenotyping which relate closely to the target environment where the traits are to be expressed, and in reproducing those conditions consistently in repeated screening. Epistasis was comparable with or smaller than putative main effect QTLs, but those consisting of pairs of QTLs not linked with traits were generally small. In spite of large environmental effects, several QTLs for root morphology were consistently detected across experiments: QTLs for deep root morphology traits near R2147 in Chromosome 1, near ME10_18 in Chromosome 2, and near C119 in Chromosome 5, and QTLs for root thickness near EM13_3 and near TGMSP2 in Chromosome 2 and near RG476 in Chromosome 4. Thus, there was some consistency in QTL regions identified despite the presence of QTL x environment interaction. These consistent QTLs could be used for introgression into elite rainfed lowland rice in a marker-assisted selection program.


   ACKNOWLEDGMENTS
 
This research was supported by Rockefeller Foundation Grant RF960001#435, which provided salary for Dr. Akihiko Kamoshita. DHL seeds were provided by Dr. Surapong Sarkarung. Experimental operations in the greenhouse and laboratory were assisted by Mr. Rene M. Panopio, Mr. Donato V. Lanwang, and Mr. Ramon B. Masajo.

Received for publication May 18, 2000.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 


Related articles in Crop Science:

This issue in Crop science

Crop Science 2002 42: 1-2. [Full Text]  



This article has been cited by other articles:


Home page
 Crop Sci.Home page
R. C. Babu, B. D. Nguyen, V. Chamarerk, P. Shanmugasundaram, P. Chezhian, P. Jeyaprakash, S. K. Ganesh, A. Palchamy, S. Sadasivam, S. Sarkarung, et al.
Genetic Analysis of Drought Resistance in Rice by Molecular Markers: Association between Secondary Traits and Field Performance
Crop Sci., July 1, 2003; 43(4): 1457 - 1469.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in Crop Science
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (24)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kamoshita, A.
Right arrow Articles by Wade, L. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kamoshita, A.
Right arrow Articles by Wade, L. J.
Agricola
Right arrow Articles by Kamoshita, A.
Right arrow Articles by Wade, L. J.
Related Collections
Right arrow Rice
Right arrow Cell Biology & Molecular Genetics
Right arrow Root Development

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Vadose Zone Journal
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome