Journal of Experimental Botany, Vol. 53, No. 371, pp. 989–1004, May 2002
REVIEW ARTICLE
Linking drought-resistance mechanisms to drought
avoidance in upland rice using a QTL approach:
progress and new opportunities to integrate stomatal and
mesophyll responses
Adam H. Price1,5, Jill E. Cairns1, Peter Horton2, Hamlyn G. Jones3 and Howard Griffiths4
1
Department of Plant and Soil Sciences, University of Aberdeen, Aberdeen AB24 3UU, UK
Department of Molecular Biology and Biotechnology, University of Sheffield, Sheffield S10 2TN, UK
3
School of Life Sciences, University of Dundee, Dundee DD1 5EH, UK
4
Department of Plant Sciences, University of Cambridge, Cambridge CB2 3EA, UK
2
Received 20 August 2001; Accepted 10 December 2001
Abstract
The advent of saturated molecular maps promised
rapid progress towards the improvement of crops
for genetically complex traits like drought resistance
via analysis of quantitative trait loci (QTL). Progress
with the identification of QTLs for drought resistancerelated traits in rice is summarized here with the
emphasis on a mapping population of a cross
between drought-resistant varieties Azucena and
Bala. Data which have used root morphological traits
and indicators of drought avoidance in field-grown
plants are reviewed, highlighting problems and
uncertainties with the QTL approach. The contribution of root-growth QTLs to drought avoidance
appears small in the experiments so far conducted,
and the limitations of screening methodologies and
the involvement of shoot-related mechanisms of
drought resistance are studied. When compared to
Azucena, Bala has been observed to have highly
sensitive stomata, does not roll its leaves readily, has
a greater ability to adjust osmotically, slows growth
more rapidly when droughted and has a lower wateruse efficiency. It is also a semi-dwarf variety and
hence has a different canopy structure. There is a
need to clarify the contribution of the shoot to
drought resistance from the level of the biochemistry
of photosynthesis through stomatal behaviour and
leaf anatomy to canopy architecture. Recent
advances in studying the physical and biochemical
5
processes related to water use and drought stress
offer the opportunity to advance a more holistic
understanding of drought resistance. These include
the potential use of infrared thermal imaging to study
energy balance, integrated and online stable isotope analysis to dissect processes involved in carbon
dioxide fixation and water evaporation, and leaf
fluorescence to monitor photosynthesis and photochemical quenching. Justification and a strategy
for this integrated approach is described, which has
relevance to the study of drought resistance in
most crops.
Key words: Infrared thermography, Oryza sativa, photosynthesis, root growth, stable isotopes, water-use efficiency.
Introduction
Since the development of molecular markers first allowed
the construction of saturated linkage maps (Botstein
et al., 1980), it has been clear that the technology of
quantitative trait locus (QTL) analysis could be usefully
employed to analyse the genetics of complex traits. By
producing mapping populations based on crosses of
parental varieties contrasting for the trait of interest, it
should prove possible to identify which parts of the
genome improve the trait. It should also prove possible
to identify the genomic regions that influence component
To whom correspondence should be addressed. Fax: [ 44 (0)1224 272703. E-mail: [email protected]
ß Society for Experimental Biology 2002
990
Price et al.
traits theoretically linked to the main trait and approximately quantify the contribution of these component
traits. Once achieved, the targeting of genomic regions
for varietal improvement would be possible through
marker-assisted selection (Stuber et al., 1999). Combining
QTL mapping with other biotechnological advances
such as physical mapping and whole genome sequencing
opens the opportunity for the identification of the
responsible gene(s) by map-based landing (Tanksley
et al., 1995) or a candidate gene approach (Pfieger
et al., 2001).
The application of the QTL approach to the genetic
dissection of drought resistance in rice (Oryza sativa L.)
is certain to be difficult. The challenge is worthwhile
considering the very substantial impact of drought on
rice production (Evenson et al., 1996) and the relatively
limited progress previously made in improving the
drought resistance of high yielding varieties (Price and
Courtois, 1999; Fukai and Cooper, 1995; Nguyen et al.,
1997). There are two main reasons for the slow progress
in this area. Firstly, the phenomenon of drought is complex in itself. To a plant, drought is an interaction
between precipitation, evapotranspiration, irradiation,
soil physical properties (including the soil hydrological
and strength properties), soil nutrient availability, and
biological interaction with pests, pathogens and neighbouring riceuweed plants). This makes it difficult to define
with any precision a ‘typical drought’ for screening purposes which is likely to represent the conditions of the
target environment. Secondly, rice germplasm displays a
diverse range of sometimes genetically complex mechanisms of drought resistance, including mechanisms of
drought escape (short duration), drought avoidance
(e.g. deep rooting) and drought tolerance (e.g. osmotic
adjustment). Because the drought phenomenon and the
mechanisms of drought resistance interact (i.e. the same
mechanism of drought resistance will not work for all
drought environments) it will naturally prove difficult
to identify regions of the rice genome which contribute
to resistance to drought in a broad range of drought
environments. The ultimate goal of breeders is to identify
QTLs that increase yield under drought or at least
increase yield stability under drought. Since yield in the
absence of drought is itself a complex trait influenced by
many component traits, and since under drought, these
traits will interact with the drought environment and the
resistance mechanisms, the true complexity of breeding
for yield under drought is clear. In other cereal species,
good progress has been made where a single drought
resistance mechanism has been shown to be of great
importance. For example, there is great promise in work
on QTL mapping anthesis to silking interval in maize
(Ribaut et al., 1997; Frova et al., 1999) and the stay green
trait in sorghum (Crasta et al., 1999). Rice has proved
more difficult.
Mechanisms of drought resistance in rice
In this paper, drought adaptation is defined following
Levitt, which distinguishes drought resistance from
drought escape (flowering to complete life cycle before
drought), divides drought resistance into drought (stress)
avoidance (maintenance of tissue water potential) and
drought (stress) tolerance and further divides drought
tolerance into dehydration avoidance and dehydration
tolerance (Levitt, 1980). Most of the traits that contribute
potentially to the drought resistance of rice have been
thoroughly reviewed (Fukai and Cooper, 1995; Nguyen
et al., 1997, Price and Courtois, 1999). A brief summary
is given below, with greater detail given to the study of
roots, given their demonstrated importance in drought
resistance and their contribution to other potential constraints such as nutrient deficiency. One important point
to note here is that many of the observed responses to
drought in plants obey Le Chatelier’s Principle (when
subject to a perturbation a system tends to respond in
such a way as to minimize the effect of the perturbation).
For example, drought responses such as stomatal closure,
leaf rolling, enhanced root growth, enhanced ABA production, and so on act to minimize water deficits. It follows that in any particular case one of these responses
could indicate particular sensitivity to drought or may
indicate a highly responsive drought tolerance mechanism. It is often difficult or impossible to distinguish
these two and this must be remembered when choosing
indicators of drought stress.
The root system
Since a plant obtains its water and mineral requirements
from its roots and the availability of these resources
often imposes a limit to plant productivity, it is difficult to
overstate the importance of roots to plant productivity.
Root development is fundamentally involved in the
response to many plant stresses, in particular, drought
and mineral deficiency. The possession of a deep and
thick root system which allows access to water deep in the
soil profile is considered crucially important in determining drought resistance in upland rice and substantial
genetic variation exists for this (Ekanayake et al., 1985b;
Fukai and Cooper, 1995; O’Toole, 1982; Yoshida and
Hasegawa, 1982). This trait may be less important in
rainfed lowland rice, where hardpans may severely restrict
root growth. Here, the ability to penetrate a hard layer is
considered important and genetic variation in the ability
to penetrate a layer of hard wax has been demonstrated
(Yu et al., 1995). This trait may also be useful in upland
rice where high penetration resistance may limit rooting depth and where soils will harden as they dry. It is
important to note two points here. First, the penetration
of roots through uniform hard layers like wax is probably
achieved through the possession of large root diameter
Rice drought resistance
which resist buckling (Cook et al., 1997), while when the
impedance is due to a coarse textured sandy or stony
horizon, it may be that thin roots would penetrate more
easily. Second, the investment of carbon in a deep root
system may have a yield implication because of lost
carbon allocation to the shoot.
It is vitally important to appreciate that root growth is
profoundly influenced by the environment. Adverse conditions (chemical or physical) directly inhibit root growth
(e.g. low water potential or highulow temperature). Biological factors in the root environment can also have
a major influence on root distribution, but they are
generally poorly understood. Particularly important is
the presence of root feeding organisms such as nematodes, termites, mites, and aphids that can severely reduce
root proliferation or rooting depth and thereby affect
drought resistance (Audebert et al., 2000). The shoot
environment can also indirectly influence root growth
either via carbon supply or signalling processes (e.g. light
interception, water status, nutrient status). It has been
suggested that plants respond to shifts in resource supply
by allocating carbon to the organ involved in capturing the limited resource (Thornley, 1972; Dewar, 1993).
When light is limiting, plants invest in shoot biomass.
When nitrogen is limiting, they invest in root production.
At the mechanistic level, theories implicating sucrose
supply (Farrar, 1992), hormonal action (Jackson, 1993)
or a combination of both (Van der Werf and Nagel, 1996)
have been advanced to explain this phenomenon. It seems
clear that the root tip is an important component in
the sensing and signalling of environmental cues to the
whole plant (Aiken and Smucker, 1996). Responses to
drought or temperature are probably complex due to a
multiplicity of physical or biochemical processes directly
affected. At the genetic level, the response of roots to the
environment is poorly understood because roots are
intrinsically difficult to study, particularly in the natural
environment.
Shoot-related traits of drought resistance
Several mechanisms of drought resistance are associated
with the shoots of rice. These are listed in Table 1, which
includes references for the evidence of genetic variation
in each trait. These are not discussed in any further
detail because most are adequately dealt with in the
reviews cited at the beginning of this section, whilst
some are discussed in the later section describing difference between Azucena and Bala. Three recent developments need expansion. Firstly, a report (Tripathy et al.,
2000) highlights the potential value of membrane stability as a component of drought tolerance in rice. These
authors indicate that maintaining ion balance under
tissue water deficit is important in drought resistance
and show that genetic variation exists in rice. Secondly,
991
Table 1. Shoot-related traits that may be important in the
drought resistance of rice, indicating whether there is evidence
of significant genetic variation in the rice germplasm and the
reference for that evidence (see Price and Courtois, 1999, for
more details)
Potential drought
resistance-related
characteristic
Significant genetic
variation observed?
Reference
Rapid leaf rolling
Yes
Rapid stomatal
closure
Yes
High water use
efficiency
Thick epicuticular
wax
Osmotic adjustment
Dehydration
tolerance
Stay green ability
Membrane stability
Photochemical
quenching
Photoinbition
resistance
Yes
Dingkuhn et al., 1989
Price et al., 1997b
Turner et al., 1986a
Turner et al., 1986b
Dingkuhn et al., 1989
Dingkuhn et al., 1991a
Price et al., 1997b
Turner et al., 1986a
Turner et al., 1986b
Dingkuhn et al., 1991b
Yes
O’Toole and Cruz, 1983
Yes
Yes
Lilley and Ludlow, 1996
Lilley and Ludlow, 1996
Not tested
Yes
Not tested
Tripathy et al., 2000
Yes
Jiao and Ji, 2001
the most recent development in the use of stable isotope
measurements in the assessment of water use efficiency
(WUE) is to link carbon isotope discrimination to oxygen
isotope discrimination, which can distinguish between
carbon isotope discrimination due to stomatal resistance
anduor photosynthetic activity (Barbour and Farquhar,
2000). This new development may improve the costeffective determination of water use efficiency in crops
including rice. It has been pointed out, however, that a
drought-resistant plant is one that maximizes the use of
available water, rather than maximizing WUE (Jones,
1981, 1993). Thirdly, the value of improving the use of
absorbed light, resistance to photoinhibition and capacity
for non-photochemical quenching to improve drought
resistance of rice has been described (Horton, 2000).
In addition, a genetic basis for difference in resistance
to photoinhibition in rice has been demonstrated (Jiao
and Ji, 2001). Each of the traits highlighted in Table 1 are
physiologically, biochemically and genetically complex in
themselves and interact with each other. Unfortunately,
rice scientists are still some distance away from being
able to identify the ideal trait or traits to use in selection
for drought resistance or to use in the identification of
genomic regions contributing to it.
A strategy for QTL mapping
This paper does not intend to describe the detail of doing
QTL analysis. The reader is directed to the review by
992
Price et al.
Kearsey for the general principles of QTL analysis
(Kearsey, 1998) and to Price and Courtois for the application to drought resistance in rice (Price and Courtois,
1999). In outline, however, the QTL approach is as
follows; (a) choose a trait with value or potential value for
breeding; (b) identify parental lines displaying extreme
phenotypes for this trait (or parents that will produce a
segregating population through transgression, despite
their similarity); (c) cross these lines to produce progenies
which segregate for the trait (e.g. F2, back-cross, recombinant inbred or doubled haploid population) of 70 –300
plants or lines (100 lines is generally considered the
lower limit); (d) phenotype the population for the trait;
(e) screen the parents of the population for genetic polymorphism using restriction fragment length polymorphisms (RFLPs) or PCR based markers (commonly 70 –200
markers); (f ) determine the genotype of all the progenies
for the selected markers; (g) construct a genetic map from
marker data using computer programs such as MapMaker (Lander et al., 1987) or JoinMap (Stam, 1990); (h)
identify markers associated with the trait using analysis
of variance or, more powerfully, use interval mapping
techniques such as employed by the program MapMakeru
QTL, QTLCartographer (CJ Basten, BS Weir, and ZB
Zeng, Department of Statistics, North Carolina State
University) or PLABQTL (HF Utz and AE Melchinger,
Institute of Plant Breeding and Genetics, University of
Hohenhein, Germany).
The likelihood of success in this endeavour is affected
by many factors, but it is increased by choosing as divergent a set of parents as possible (at least for the trait of
interest), by producing and screening as large a population as possible, by reducing environmental error in trait
measurement to as low as possible and by achieving a
genetic distance between markers of about 20 cM or less
(Churchill and Doerge, 1994; Gallais and Rives, 1993).
Given the difficulties associated with drought resistance highlighted in the introduction, the logical way to use
a QTL approach in the exploitation of drought resistancerelated genes was to produce several mapping populations using parental varieties of contrasting drought
resistance, and conduct relevant phenotyping to identify
the regions of the rice genome that contribute to drought
resistance and its component traits (Price and Courtois,
1999). By combining field evaluations of plant performance under drought with physiological experiments on the
underlying mechanisms of drought resistance, it should
prove possible to quantify the value of different drought
resistance mechanisms. Studies to date have therefore
followed two main approaches. The first is to screen
populations for indicators of drought resistance in field
trials. In these, plants are grown in the dry season
and drought is imposed by withholding water. The second
approach is to conduct laboratory or greenhouse experiments to analyse component traits of drought resistance,
thereby reducing the variability in the environment which
hampers their measurement in the field. The individual
competent traits that have received the most attention to
date are root morphology (Champoux et al., 1995; Yadav
et al., 1997; Price and Tomos, 1997), root penetration
ability (Ray et al., 1996; Ali et al., 2000; Price et al., 2000;
Zheng et al., 2000) and osmotic adjustment (Lilley et al.,
1996; Zhang et al., 1999). Having received less attention
is leaf rolling ability and stomatal sensitivity (Price et al.,
1997b) and membrane stability (Tripathy et al., 2000).
There is certainly an opportunity to study carbon isotope discrimination as an indicator of water use efficiency (Dingkuhn et al., 1991b), non-stomatal resistance
(O’Toole and Cruz, 1983) and possibly both stay-green
ability and photosynthetic response to high light under
stress (Horton, 2000).
The objective in studying component traits is to be able
to identify genomic regions contributing to a trait which
theoretically will improve drought resistance. In order to be
of convincing value the QTLs must be shown to contribute in a range of environmental conditions, at least
those comparable to the target environment. Therefore,
there has been and will continue to be an emphasis on
conducting experiments in a range of conditions in order
to assess QTL stability across environment. Once stable
QTLs for component traits of drought resistance are
identified, the next step will be to introduce these QTLs
into a near isogenic background in order to conduct
more sophisticated experiments to give a understanding
of the underlying physiological and molecular nature of
the QTL and to evaluate the contribution to yield in the
target environment.
There are five populations that have been used to study
either drought resistance or traits related to drought
resistance for which information is published. The first
used was a population of recombinant inbred lines from
a cross between CO39 and Moroberekan which has been
used to study drought avoidance in the field and root
growth in soil-filled tubes (Champoux et al., 1995), root
penetration through wax (Ray et al., 1996) and osmotic
adjustment (Lilley et al., 1996). A double haploid population based on a cross between IR64 and Azucena has
been used to study field drought avoidance (Courtois
et al., 2000), root growth in tubes (Yadav et al., 1997),
both of these together (Hemamalini et al., 2000) and root
penetration (Zheng et al., 2000). Zhang et al. report the
use of a double haploid population based on a cross
between varieties CT9993-5-10-1-M and IR62266-42-6-2
(Zhang et al., 1999). For that population a range of
traits have been measured but few details are available
as yet. The identification of root penetration QTLs in a
population from the cross IR58821-23-B-1-2-1 Z IR5261UBN-1-1-2 has been reported (Ali et al., 2000). This
paper concentrates on just one other population, that
being a set of recombinant inbred lines based on a cross
Rice drought resistance
between two drought-resistant upland varieties, Bala and
Azucena. Comparative mapping between all these populations which have only used a few of the same RFLP
markers to produce their maps, is possible using the
extensive maps of other authors (Causse et al., 1995;
Kurata et al., 1994). Comparisons between RFLPs from
the two common sources (Cornell University and the
Rice Genome Project, Japan) is more difficult, but can
be achieved either by reference to data on comparative mapping presented at the Third Rice Genetics
symposium at IRRI in 1995 (S McCouch personal
communication), data on the Oryzabase web site
(http:uushigen.lab.nig.ac.jpuriceuoryzabaseu) or to maps
which use combinations of both sets of probes (Lu et al.,
1997; Price et al., 2000).
Progress in QTL mapping drought
resistance and related traits in the
BalaZAzucena population
A cross was made between the varieties Bala and Azucena
in 1992 and an F2 population was used to identify QTLs
for simple root traits of plants grown in hydroponics
(Price and Tomos, 1997; Price et al., 1997a). That preliminary experiment identified a number of genomic
regions affecting root thickness and maximum root
length. Using a leaf excision test, the rate of leaf rolling
and stomatal closure was also mapped in that F2 population (Price et al., 1997b). Subsequently, the population
was advanced by single seed descent to an F6 of 205
recombinant inbred lines (Price et al., 2000). A new map
has been produced that now has 102 RFLP, 34 AFLP
and 6 microsatellite markers (total 142 markers) giving a
genome length of 1732 cM and an average space between
markers of 12.2 cM.
Progress in QTL mapping root traits in the
Bala Z Azucena population
This population has been used for extensive root growth
studies in glasshouse or controlled environment rooms in
the UK. A total of 170 lines have been tested in soil-filled
boxes for 4 weeks under well watered or drought (nonwatered) conditions (Price et al., 1999). In that screen, a
box 2 Z 1 Z 1 m (length Z depth Z width) was constructed
from plywood. The roots of each plant were held within a
porous nylon fabric bag (0.9 m deepZ 0.08 m diameter)
containing sandy loam soil. After 4 weeks, the soil was
washed from each plant and maximum root length and
the thickness of the thickest root taken from the base
of each plant were measured. Using thin glass-sided
chambers (1 Z 0.3 Z 0.015 m) filled with a different sandy
loam soil and again well-watered or non-watered treatments, 140 lines were evaluated over two summers for
root length during growth and for root thickness,
rootushoot ratio and deep root weight (below 0.9 m)
993
after 56 d (Price et al., 1999, 2002a, b). The whole population has been screened in hydroponics (AH Price,
unpublished data) while for 110 lines the ability of the
roots to penetrate a simulated hard layer of wax petroleum has been conducted (Price et al., 2000). These results
indicate three important points. Firstly, under each
experiment there are a large number of relatively small
QTLs detected, distributed over the genome. Secondly,
the pattern of QTLs varies between the type of rooting
environments and between different treatments (and
even between replicate runs to some extent). Thirdly, some
consistent pattern does emerge from studying combined data on this population and other data obtained
in other populations. Figure 1 shows a summary of the
regions identified as controlling some root traits in the
Bala Z Azucena population. Based on these findings and
those of the other populations, five regions stand out, on
chromosomes 2, 5, 7, 9, and 11.
At a region of chromosome 2, at about 125 cM from
the top of the chromosome (at marker C601 in Bala Z
Azucena population), QTLs for root penetration ability
have been revealed in all four of the mapping populations
in which root penetration has been tested (Price et al.,
2000; Ray et al., 1996; Zhang et al., 1999; Zheng et al.,
2000). In the Bala Z Azucena population, there were also
QTLs for maximum root length in both well-watered
and water-limited conditions in the thin chamber experiments (Price et al., 2002b) and for root length in hydroponics measured in both the F2 (Price and Tomos, 1997)
and F6 (AH Price, unpublished data). Yadav et al. (1997)
also found a QTL for maximum root length and root
thickness in this region.
On chromosome 5, at about 85 cM (at marker C43
in this population) is a region with QTLs in the Bala Z
Azucena population for root penetration ability (Price
et al., 2000), root length and thickness in hydroponics of
both the F2 (Price and Tomos, 1997) and F6 (AH Price,
unpublished data). In each case it is the Bala alleles which
increase root growth, even though Azucena has a
better root system than Bala. In the same region, Yadav
et al. (1997) showed that Azucena alleles increase root
thickness.
On chromosome 7 at about 80 cM is a region which
has effects on root morphology most strongly in other
populations. Thus in the Bala Z Azucena population, a
QTL at marker RG650 for maximum root length was
detected in the thin chamber experiment (Price et al.,
2002b). Yadav et al. (1997) reported QTLs for maximum
root length and deep root weight, Zheng et al. (2000)
reported a QTL for root penetration ability and Zhang
et al. (1999) reported a QTL for rooting depth in the
same place.
One of the most striking observations gained from
comparing QTL mapping of root traits is that chromosome 9 has two strong clusters of QTLs detected in all
994
Price et al.
Fig. 1. Molecular linkage map with RFLP and microsatellite markers indicated (AFLP marker names omitted) of the Bala Z Azucena population
showing quantitative trait loci (QTLs) for root length, thickness or penetration ability. Indicated are those QTLs that were over the threshold value for
the analysis conducted. In the case of the glass chamber experiments (indicated GC), only QTLs which were apparent when data from well-watered and
non-watered plants were averaged are presented. QTLs on the left of the chromosome indicate that Azucena alleles increased the trait value.
populations. The lower one is near 80 cM and is associated with marker G1085 in the Bala Z Azucena. QTLs
for rootushoot ratio, root thickness, deep root weight, and
maximum root length at marker G1085 under both wellwatered and water-limited conditions have been revealed
in this population (Price et al., 2002b). In this region,
Yadav et al. (1997) have reported QTLs for maximum
root length and deep root weight, Hemamalini et al.
(2000) have reported QTLs for root thickness, Zheng
et al. (2000) have reported a QTL for root thickness, and
Champoux et al. (1995) reported QTLs for root thickness
and rootushoot ratio.
Above this region, at about 60 cM, is another region
affecting root morphology. At marker G385 in the
Bala Z Azucena population, QTLs were detected in the
thin chamber screen for root thickness and rootushoot
ratio (the latter only in water-limited treatment). In
the same place Yadav et al. (1997) reported a QTL for
deep root weight while Champoux et al. (1995) reported
QTLs for root thickness and rootushoot ratio.
At about 85 cM on chromosome 11 is another
region that appears to affect root growth. At marker
C189 in the Bala Z Azucena population, QTLs for
maximum root length in hydroponics were identified
in the F2 (Price and Tomos, 1997) and for root
penetration ability in the F6 (Price et al., 2000).
Champoux et al. (1995) identified QTLs for maximum
root length, root thickness and rootushoot ratio in the
Rice drought resistance
same place. Also here, Ray et al. (1996) identified a QTL
for root penetration.
Progress in QTL mapping drought avoidance traits
Between 110 and 176 F6 lines have been screened for
indicators of drought avoidance in sites in the International Rice Research Institute (IRRI), Philippines and the
West Africa Rice Development Association (WARDA),
Côte d’Ivoire over two dry seasons. QTLs for the visual
scores leaf rolling and leaf drying and for relative water
content have been located (Price et al., 2002c).
As with root traits, drought avoidance traits give many
QTLs which differ between sites and between years at
the same site. Again, by studying these results with those
obtained on the other two populations that have been
tested in a similar manner (Champoux et al., 1995;
Courtois et al., 2000) a pattern does emerge from the
complexity indicating some strong evidence of consistent
drought avoidance QTLs. What is particularly striking,
however, is poor alignment of drought avoidance QTLs
with the regions where there is good evidence for an effect
on root thickness or depth (i.e. on chromosomes 2, 5, 7,
9, and 11 described above). There is some evidence that
the regions of chromosomes 5, 7, 9, and 11 that affect
root growth also affect performance under drought.
On chromosome 5 QTLs in which the Azucena allele
increases relative water content and reduces leaf rolling
at WARDA were detected, but in the same place a QTL
in which the Azucena allele increased leaf rolling at IRRI
was revealed (Price et al., 2002c). (Note that the Azucena
allele here decreases root thickness, length and penetration ability in this population.) Courtois et al. (2000)
identified QTLs in this region for leaf rolling in three
screens and both leaf drying and RWC in one screen,
in each case the Azucena allele increasing drought
avoidance, in agreement with the increase in root
thickness associated with the Azucena alleles in their
population (Yadav et al., 1997). Champoux et al. (1995)
identified QTLs for leaf rolling just above RG351
on chromosome 7 in two of the growth stages, while
Courtois et al. (2000) reported QTLs for leaf drying
in one screen and relative growth rate under drought
stress in another near RG351. Courtois et al. (2000)
reported QTLs for leaf rolling in all screens at RZ12
(maps to the same place as G1085) on chromosome 9,
while Champoux et al. (1995) reported leaf rolling QTLs
in all growth stages in the same location. Only Champoux
et al. report evidence for QTLs for drought avoidance
at the root growth QTL on chromosome 11, with leaf
rolling QTLs in three growth stages (Champoux et al.,
1995). There are two important points to note. Firstly,
there is no convincing co-location of drought avoidance
QTLs with the root growth QTL on chromosome 2 in any
population. Secondly, only on chromosome 5 is there
995
any co-location between drought avoidance and root
morphology QTLs in the Bala Z Azucena population (and
even here, only in the IRRI drought screen does the effect
on drought avoidance agree with the effect on root
morphology).
Possible explanations for lack of co-location
of root and drought avoidance traits
The poor co-location of drought avoidance and root
morphological QTLs in populations in general, and in the
Bala Z Azucena population specifically, casts doubt on
the notion that improving the roots of rice will improve
drought resistance. Specifically, it is possible to offer
several theories as to why there are no drought avoidance
QTLs at root growth QTLs in the Bala Z Azucena population, each of which have profound implications for
attempts to improve drought resistance in crops.
Roots do not contribute to drought resistance
There seems to be ample evidence that rice varieties or
even progenies of segregating populations which have
better root systems assessed in controlled environment
experiments do perform better in the field under drought
in terms of visual drought avoidance (Ekanayake et al.,
1985b; Champoux et al., 1995; Price et al., 1997). Fieldbased studies of root distribution confirm that varieties
identified as being deep rooting in the laboratoryu
greenhouse are indeed better at depth in the field
(Puckridge and O’Toole, 1981; Yoshida and Hasegawa,
1982). It has been shown that deeper roots increase the
ability to extract soil water from depth (Mambani and
Lal, 1983a, b, c; Yoshida and Hasegawa, 1982), improve
shoot water status (Yoshida and Hasegawa, 1982)
and may increase yield under drought (Mambani and
Lal, 1983a). Root pulling force, a supposed integrator
of rooting capacity, has also been shown to correlate
with shoot water status under drought stress (Ekanayake
et al., 1985a). It seems unlikely to be true that roots
are unimportant, since the evidence that deeper and
thicker roots should contribute to drought resistance, at
least in some environments, seems convincing.
Root growth QTLs identified in controlled
environments are not expressed in the field
Although a primary aim of root screening has been to get
the best compromise between practicality (ease of experiment) and relevance to the field situation, it must be
accepted that both the root and shoot environment are
likely to be different to that experienced in the real
drought situation. The plant root growth is known to
be influenced by rooting environment and probably the
shoot environment. This means that artefactual results
996
Price et al.
from root confinement, soil chemical (especially nutrient
distribution) or physical properties (water distribution,
penetration resistance, temperature fluctuation) or shoot
environment will probably occur. The fact that the ranking of varieties for root morphologial characteristics
between different experiments using different methods
generally agrees (Price et al., 1997) suggests that the
genetic control of root thickness and depth is under quite
environmentally-robust genetic control. The success of
Champoux et al. at linking root thickness in particular
(measured in soil-filled tubes) to drought avoidance
strongly suggests that root morphology QTLs measured
in a screenhouse are important in conferring drought
resistance in the field (Champoux, 1995).
The environment also affects individual QTLs and the
QTLs with small effects are the more likely be influenced
by variation across replicates, sites or repeated runs.
QTLs which have R2 values around or below 10% may
be difficult to pick-up repeatedly for this reason or due to
the high statistical threshold used in QTL analysis, unless
the population size is high (100). However, several of
the root growth QTLs reported do have R2 values around
20% or above when measured in the greenhouse (e.g. an
R2 of 29% for hydroponic root length on chromosome 11
in Price and Tomos, 1997; 18% for root penetration ratio
on chromosome 2 in Price et al., 2000; 18% for root
thickness of soil grown plants on chromosome 9 in
Price et al., 2002b). These are relatively large QTLs and
one might reasonably expected them to influence root
morphology in at least some field situations.
Limitations of field trials
A meaning for drought resistance?: Drought resistance
can have different meanings for different workers. Some,
for example, consider drought resistance solely in terms
of yield under drought conditions, while others are more
concerned with stability of yield over a range of environmental conditions. Yet others consider drought resistance in terms of survival of drought conditions, rather
than productivity. It follows from these observations and
contrasting requirements that rankings of genotypes for
drought resistance depends on the method of assessment
used. It is likely that characters that favour drought
escape (such as early floweringuharvest) will be found
in different genotypes than characters that favour the
avoidance of internal water deficits in dry conditions
(such as deep roots, rapid stomatal closure or high cuticular resistance) or characters associated with biochemical
tolerance of internal water deficits. Furthermore, other
varieties may be apparently drought resistant through an
enhanced ability to recover after a dry period. It seems
likely, therefore, that no single characteristic can be used
as a screen for drought tolerance in genetically diverse
collections.
A particular problem with drought resistance screening, especially under field conditions, is that the ranking
of genotypes is very dependent on the specific environmental conditions of the trial (i.e. there is a high genotype
by environment interaction in drought screens). This is
well illustrated by the assessment of the Bala Z Azucena
population in different sites (Price et al., 2002c). Only leaf
drying scores correlated between IRRI and WARDA
(r\ 0.373, P < 0.001), while leaf rolling scores and relative
water contents did not. Indeed, there were very poor
correlations even between years at WARDA (r \0.068
for leaf rolling, 0.124 for leaf drying and 0.206 (P \ 0.05)
for RWC) although there was much better consistency at
IRRI. Reasonably good correlations were found between
leaf rolling scores in IRRI and an Indian field site in the
AzucenaZ IR64 population (r\ 0.470 and 0.66, P < 0.001)
(Courtois et al., 2000).
Drought screens in dry season: Because of the absolute
need to have a drought in each screen (project funding
is rarely able to accommodate loss of whole seasons
of data due to uncertainty in climate), drought trials
are generally conducted in the dry season when low water
availability can be guaranteed, and is combined with
supplementary irrigation. Unfortunately, these conditions
are generally very harsh for the plants and probably do
not reflect well the conditions of natural droughts which
more commonly occur as low water availability in a normally reasonably wet period. Particularly important here is
whether the irrigation allows water to soak the whole soil
profile fully (i.e. does it keep the deep soil wet) and the very
bright (often cloudless), hot and dry climatic conditions
which may cause substantial physiological stress in the
shoots even if there is adequate access to water.
Soil physical properties: The ability of roots to access
water at depth is dependent on their growth through the
soil medium. If the soil has a high strength (has a high
resistance to penetration) or if its strength increases
greatly as it dries out (as all but the sandiest soils do to a
greater or lesser extent) then the roots may fail to contribute to drought avoidance. It has been discovered that
the site where field trials at WARDA were conducted was
very hard. This is illustrated in Fig. 2. It shows the depth
to a penetration resistance of 3.0 MPa (a resistance which
should very substantially inhibit root growth) in the field
at WARDA where drought screens were conducted and
indicates that, for much of the area of the drought trial
the roots could not have been expected to grow below
20–30 cm despite a demonstrated ability to grow well
over 1 m in glasshouse experiments. The relationship
between soil hardening, root growth and drought in this
and other WARDA sites is currently being investigated.
Soil biological properties: Field screens are prone to
damage by pests and pathogens. Dry season screens
Rice drought resistance
997
Fig. 2. Results of a wet season soil penetration resistance survey of field used for drought screening at WARDA. Penetration was assessed in a grid
of 5 m separation using a cone penetrometer (Remik, Australia). Presented here is a map based on depth to a resistance of 3 MPa. Droughted plots
were sown in the leftmost 25 m of the field.
appear to be particularly prone to root-damaging pests
such as mites, termites and nematodes. This is partly due
to the greater abundance of these aerobic organisms
in drier soils (hence not a major problem in flooded rice
in any season). Recent experiments at WARDA have
encountered mite and termite damage and nematodeinduced root damage at IRRI has also been a problem
(B Courtois, personal communication).
Shoot-related mechanisms which differ between
the two parents
The variety Bala was chosen as a parent in the crossing
programme specifically because it displays a high degree
of drought resistance using standard measures of leaf
rolling and leaf drying, yet it does not have a good root
system. It must, therefore, have shoot-related mechanisms of drought resistance which contribute to drought
resistance. Experiments using Bala indicate several mechanisms of drought resistance that might be expected to
differ with Azucena. However, few of these mechanisms
have been effectively evaluated for their contribution
to yield under stress in rice, and most have not been
analysed in the mapping population. It may be that in the
drought screens so far conducted that it is these traits
that have dominated drought avoidance, either because
they are very effective in conferring drought avoidance
or because the root system does notuwas not able to
contribute. The authors believe that there is considerable
scope for investigating these mechanisms in more detail.
Leaf rolling and stomatal sensitivity: There is variation in
the degree to which the leaves of rice roll in response to
low water potential (Dingkuhn et al., 1989; Turner et al.,
1986a). The rate of leaf rolling upon leaf excision also
varies in rice varieties (Price et al., 1997b) and the variety
Bala was shown to roll unusually slowly (compared to 11
other varieties including Azucena). A leaf rolling QTL has
been identified on chromosome 1 in the F2 (Price et al.,
1997b) that coincides with a large effect QTL for leaf
rolling in the field (Price et al., 2002c). There is also
evidence that the stomata of Bala are more sensitive than
those of Azucena (Price et al., 1997b).
Shoot morphology and mass, canopy structure and response
to drought: Bala is a semi-dwarf variety while Azucena
is tall. Bala tillers more than Azucena and produces
narrower and shorter leaves. The thickness of the leaves
appears to depend on the environment since there was a
significant interaction between genotype and treatment
(irrigated or droughted) at IRRI in 1996 for specific leaf
area (P < 0.01). Bala had the thinner leaves under irrigation, but the thicker under drought when compared to
Azucena. Increased leaf thickness should increase water
use efficiency and therefore seems a useful response to
drought. What seems certain is that the canopy structure
of the two varieties will be markedly different in terms
of light and heat interception, heat and water loss and
probably gas diffusion inside the leaf, all parameters
which may be expected to affect water use efficiency and
drought-induced biouphotochemical damage to the leaf.
Under well-watered conditions, Bala has a smaller
shoot biomass than Azucena (about 25% smaller in UK
greenhouse experiments and 45% smaller in the field at
IRRI) although this is not apparent in droughted plants
(indeed, if anything, Bala are bigger under prolonged
drought in the greenhouse or field). There is a contradiction here with (unpublished) data which indicates that,
998
Price et al.
under drought the increase in plant height slows markedly
in Bala when droughted in comparison to Azucena.
coincide with one at WARDA. Further investigation into
this discrepancy could be useful.
Water-use efficiency as measured by D13C: There is
genetic variation for water use efficiency which is reflected
in difference in D13C between upland rice genotypes
(Dingkuhn et al., 1991). These data suggest that Azucena
has a high water use efficiency although unfortunately
Bala was not tested. One of the two varieties that
were used as parents to produce Bala at the Central
Rice Research Institute, Cuttack, India, namely N22
(Chaudary and Rao, 1982), was tested, however, and was
much less water use efficient than Azucena. Previously
unpublished data collected on irrigated control plants
growing in the dry season at IRRI and WARDA confirm that there is a difference in D13C between Bala and
Azucena of about 1.2 ml Y 1 (Table 2). The large standard
deviation of the segregating population compared to the
parents (IRRI only) in these data clearly indicate that
there is a genetic component to the carbon isotope discrimination. QTL mapping, however, shows (Fig. 3) that
although genomic regions controlling these values can be
identified, in no case does the region detected at IRRI
Osmotic adjustment: It has been shown that, in general,
the tropical japonica rice varieties have a poor ability to
adjust osmotically under drought, something that indica
rice varieties can do relatively well (Lilley and Ludlow,
1996). It is believed that this represents an adaptation to
drought in rainfed lowlands where roots cannot easily
contribute to drought avoidance. Azucena was shown to
be the 12th worst variety of the 61 tested while Bala
was the 10th best. An osmotic adjustment QTL has been
mapped to chromosome 8 in the CO39 Z Moroberekan
population (Lilley et al., 1996) and the same location
does appear to contribute to drought avoidance in the
Bala Z Azucena population (Price et al., 2002c).
Table 2. D13C values for youngest fully expanded leaves from
five plants of varieties Azucena, Bala and the F6 progeny grown
under irrigation in the dry season of IRRI in 1996 and WARDA
in 1997 (] standard deviation)
Variety
IRRI
WARDA
Azucena
Bala
F6
Y 27.15 ] 0.13
Y 28.27 ] 0.02
Y 27.71 ] 0.58
Y 27.26 (no replication)
Y 28.12 (no replication)
Y 27.14 ] 0.51
Summary of shoot mechanisms: Several pieces of
information suggest that the shoots of Bala have an
ability to respond to drought rather more than Azucena.
Thus stomata are more sensitive, osmotic adjustment
is more pronounced, shoot elongation growth is more
sensitive, leaf thickness increases (rather than decreases in
Azucena). On the other hand, Azucena rolls its leaves
more rapidly and D13C data indicate a more efficient
water use efficiency under low to moderate water stress.
There appears to be many differences in the way in which
the shoots of Bala and Azucena may be expected to
interact with the drought. These may well interfere with
the detection of co-located drought avoidance and root
growth QTLs, but they probably reflect the expression
of traits that would be valuable for improving drought
Fig. 3. Molecular linkage map (with marker names omitted) of the Bala Z Azucena population showing QTLs for D13C measured on five youngest
fully expanded leaves of plants grown under irrigated conditions during dry season screens at IRRI in 1996 and WARDA in 1997. The values
above each box give the LOD value obtained with QTLCartographer using composite interval mapping, where a LOD of 3.1 represents the 0.05%
genome wide threshold value. The boxes represent the Y 1 and [ 1 LOD interval and the number above or below the box indicates the LOD score
for the QTL. A positive LOD value indicates that Azucena alleles increase the trait value (reduced water use efficiency), a negative LOD indicates that
Azucena alleles reduced the trait value.
Rice drought resistance
999
resistance. The techniques and approaches that will allow
the physiological processes underlying these contrasting drought responses to be investigated in detail will
be outlined next. Using advances in the understanding
of photosynthetic physiology and biochemistry of rice
(Horton and Murchie, 2000) together with additional
methods to differentiate stomatal from mesophyll limitations, will resolve the contributions of shoots and roots
in conferring drought resistance and yield stability under
drought. This, it is hoped, will open the way to the
identification of new strategies for improving the drought
resistance of rice.
New approaches to study the physics,
biochemistry and physiology of drought
Infrared thermography to analyse energy balance
and stomatal function
One potential approach to the screening of large numbers
of lines of a crop such as rice for stomatal characteristics
uses an advance on the infrared thermometry (as used by
Garrity and O’Toole, 1995), which is the use of thermal
imaging (Jones, 1999a). The basic principle of this
approach is that the energy balance of plant leaves is
very strongly dependent on the rate of evaporation, which
in turn is strongly dependent on stomatal conductance
(Jones, 1999b). Indeed, screening for stomatal behaviour
using infrared imaging techniques has previously been
used in laboratory screens for ABA-insensitive mutants
in barley (Raskin and Ladyman, 1988) and for studying
the stomatal functionality of Arabidopsis mutants (Gray
et al., 2000). The technique can also be used as an indicator of plant stress, the precision of which can be
improved by the use of wet and dry reference surfaces
for calibration (Jones, 1999a, b). This allows a linear, 1 : 1,
relationship between stomatal conductance measured by
porometry and that calculated from thermograms to be
demonstrated in the laboratory (Fig. 4). Thermal imaging
could usefully be applied to Bala and Azucena to characterize their different relationships between plant water
status (and importantly soil water status) and stomatal
function.
A problem with the application of thermal imaging
techniques in the field is that the relationship between leaf
or canopy temperature and stomatal conductance is very
dependent on the other environmental conditions (humidity, radiation, windspeed, etc.). Preliminary field experiments, however, suggest that such an approach could also
be applicable in the field. This is because in a comparative
screening trial with repeated controls, absolute calibration of the images is not necessary, one is primarily concerned with differential responses. A major advantage of
thermal imaging as opposed to the use of infrared thermometry or porometry to measure stomatal conductance
Fig. 4. The 1 : 1 linear relationship obtained between leaf conductance
of Phaseolus vulgaris estimated from porometry measurements and that
calculated from thermograms immediately prior to the use of the
porometer. The calculations were based on calibration values obtained
for wet benchcoteTM (Whatman Ltd), dry benchcote and benchcote
covered with surgical dressing ($ and n) or calculated from
measurements made on another day under the same conditions using
wet and dry benchote only (s); (from Jones, 1999a).
in the field is the shorter time required to obtain large
numbers of data. Relative canopy temperatures of different plants in a mapping population could be measured
in a few minutes, thereby avoiding the confounding effects
of rapid changes of stomatal conductance in response to
changes in microclimate such as windspeed or radiation
which can occur when using slower techniques on large
numbers of lines.
Stable isotopes and the interaction between
water use and mesophyll conductance
Given the problems described above in relating root
growth characteristics to QTL markers, one method
allowing the source of water to be distinguished is the use
of 18O (or 2H, D) as a tracer of water abstraction within
the soil profile (Ehleringer and Dawson, 1992, see also
reviews in Griffiths, 1998). However, this is dependent on
a discernible signal between groundwater and surface,
meteoric water, and that at any given field trial, groundwater is available and accessible (see problems at
WARDA, above). The 18O signal is directly affected by
evaporation, whether at the soil surface or at the leaf,
and how this leaf signal may be used to augment carbon
isotope discrimination measurements is discussed.
Whilst providing an excellent marker for water use
under semi-arid conditions, the carbon isotope signal in
leaf organic material reflects a combination of stomatal
and mesophyll processes (Table 2; Griffiths et al., 1999).
In simplest terms, as long as the proportional drawdown of CO2 from ci (substomatal CO2 concentration)
1000
Price et al.
to cc (CO2 concentration at Rubisco) is constant, then ci
represents a good proxy for comparative water use measured by carbon isotope discrimination (D), and that
calculated from stomatal conductance (Di). If, however,
some of the morphological and biochemical constraints
vary between the parent lines (such as leaf thicknessu
rolling, or N contentuRubisco activity; Table 3), the
mesophyll conductance will alter the additional drawdown to Rubisco and hence affect carbon isotope discrimination. The greater is the drawdown from ci to cc,
the lower the discrimination expressed in organic material,
or measured in real-time, when CO2 can be collected
during on-line discrimination (Dobs). Under high temperatures, the impact of respiration and photorespiration
become quantitatively greater at low rates of net CO2
uptake (Griffiths et al., 1999). A number of studies have
suggested that leaf N is a good proxy for Rubisco activity,
and the importance of evaluating Rubisco allocation and
light use for rice in the field is discussed.
One means of distinguishing the interplay between
stomatally-led evaporative demands and internal mesophyll conductance is to analyse the 18O signal in leaf
organic material or leaf water. The problem of analysing
organic material is that the signal reflects primarily the
environmental conditions during leaf expansion, rather
than more immediately during an episodic drought, and
so analyses of leaf water 18O content have been undertaken for the two parent lines under field conditions in
WARDA. Despite the limited replication, there is evidence of a consistent difference between the two varieties
(Fig. 5), with greater enrichment generally found in Bala
rather than Azucena. Under laboratory conditions, a
detailed study of the variation between on-line, instantaneous carbon and oxygen isotope discrimination
would help to evaluate the different drought stress
characteristics of parent lines at the leaf level (Harwood
et al., 1998), while a survey of leaf water, anduor leaf
organic material of the entire mapping population in a
field trial, would help to distinguish the interplay between
stomatal and mesophyll processes.
Photosynthetic capacity and light utilization
Under field conditions, irrigated rice shows a marked
midday depression of photosynthesis, with leaf temperatures reaching 35–37 C for much of the day (Murchie
et al., 1999). While there are implications for both losses
of net C gain and increased photorespiration at this time,
there was no evidence of long-term photoinhibition in
this irrigated crop. However, it was speculated that in
more marginal habitats, such as those where upland rice
is grown, the loss in overall yield attributable to photoinhibition could be considerable (Horton and Murchie,
2000). Thus screening of Photosytem II fluorescence
Fig. 5. 18O discrimination measured in water extracted from the
youngest fully expanded leaves from Azucena (open bar) and Bala
(hatched bar) plants grown under irrigation on three sites at WARDA
in the dry season of 2001 (Bar \ standard error, n \ 3).
Table 3. The link between stable isotope measurements 13C and 18O in organic material and real-time instantaneous discrimination
during gas exchange and coupling to stomatal and mesophyll conductance, as well as environmental conditions and leaf ultrastructure
as determinants of isotope composition and link to Rubisco activity (from Griffiths et al., 1999)
Control of D
CO2 supply and demand
Long term
N
Augs
Q
----
g
h
organic
d13C Q
(d18O,VPD-led)
13
Environmental conditions
Instantaneous
Di
WUE
Physiologicalumorphological correlates
C,18O
in CO2 Q
ci
gi
O
cc
O
Dobs
8
>
>
>
>
>
>
>
<
>
>
>
>
>
>
>
:
diffusive exchange
cc
RUBISCO
CAPACITY
(Photo)respiration
refixation and exchange
8
>
>
>
>
>
>
>
>
>
<
>
>
>
>
>
>
>
>
>
:
Life formustrategy
SLM
chlorophyll
Leaf N
Ash content
8
>
>
>
>
>
>
>
>
>
>
>
<
>
>
>
>
>
>
>
>
>
>
>
:
VPD
PPFD
TEMPERATURE
RAINFALL
SOIL WATER DEFICIT
SOIL
STRENGTHu
NUTRIENTS
Rice drought resistance
1001
Conclusion
Fig. 6. Pmax and Rubisco protein content determined for a new plant
type (NPT) variety (IR65598–112–2) and IR72 for four leaves of
different ages (from Horton and Murchie, 2000).
characteristics, and the relationship between maintenance
of electron transport and capacity for non-photochemical
quenching of the mapping population under waterlimited conditions may provide another important
marker of leaf-level drought tolerance characteristics.
Photosynthetic capacity was also implicated as a key
limitation to overall productivity in irrigated rice
(Murchie et al., 1999; Horton and Murchie, 2000), and
as such would contribute to the overall mesophyll
conductance (Table 3). Despite this, during grain-filling
a significant portion of chlorophyll and leaf N was lost
from leaves, without a loss of photosynthetic capacity
(Murchie et al., 1999), supporting the notion of leaf
Rubisco as a store of mobilizable N. Rubisco content of
a new plant type (NPT) rice variety was much higher than
in IR72 (Fig. 6) apparently in excess of that needed to
sustain photosynthesis, since NPT Pmax was actually
lower than that of IR72. Varying the content of Rubisco
may be an important strategy for ensuring adequate N
reserves for grain fill in certain rice varieties (Horton
and Murchie, 2000).
In summary, the efficiency of light use will in part
be determined by the canopy architecture (which differ
markedly between Bala and Azucena), with leaf angle and
orientation both important to reduce self shading and
optimizing the use of available Rubisco, but perhaps at
the expense of photoinhibitory losses under droughtstressed conditions. In determining the trade-off between
maintaining yield and tolerance to drought, it would
be important to include additional measurements of
chlorophyll fluorescence, photosynthetic capacity and
N allocation to evaluate the mapping population of
upland rice.
New developments in plant biology open the opportunity
to investigate the mechanisms of drought resistance in
a more holistic way than in the past. If advances in the
use of stable isotopes and infrared thermometry can be
combined with advances in understanding of the fundamental biochemical processes underlying photosynthetic
activity there are opportunities for great advancement.
If this research is conducted in the context of genetic
mapping on the model monocotyledon species, anduor
combined with gene array technology, the outcomes may
be even greater. The Bala Z Azucena population offers
exciting challenges to both plant physiologists and geneticists to collaborate in elucidating the importance of
different mechanisms of drought resistance in the context
of contrasting droughting environments and to make real
progress in identifying genomic regions (for breeders) or
even candidate genes involved in these crucial processes.
Progress made using this approach for rice, will lead to
equivalent opportunities in other crops.
Acknowledgements
Most of the work on Bala and Azucena presented here has been
conducted under projects funded by the UK Department for
International Development (DFID), Plant Sciences Programme.
This paper uses some previously unpublished data for which
A Price would like to acknowledge the following; Brigitte
Courtois (IRRI), Alain Audebert and Monty Jones (WARDA)
for the collaboration on the field evaluation of the population;
John Townend (Aberdeen), Katherine Steele, John Gorham,
Julian Bridges (Bangor, UK), and Lawrence Clarke (IACR
Rothamsted, UK) for the collaboration on root screening;
Alain Audebert and Chris Mullins (Aberdeen) for the ongoing
work on soil physical properties and root growth. We are
grateful to Gary Lanigan and Nick Betson for analysing the
18O data.
References
Aiken RM, Smuker AJM. 1996. Root system regulation of
whole plant growth. Annual Review of Phytopathology
34, 325–346.
Ali ML, Pathan MS, Zhang J, Bai G, Sarkarung S, Nguyen HT.
2000. Mapping QTL for root traits in a recombinant inbred
population from two indica ecotypes in rice. Theoretical and
Applied Genetics 101, 756–766.
Audebert A, Coyne DL, Dingkuhn M, Plowright RA. 2000. The
influence of cyst nematodes (Heterodera sacchari) and
drought on water relations and growth of upland rice in
Côte d’Ivoire. Plant and Soil 220, 235–242.
Barbour MM, Farquhar GD. 2000. Relative humidity- and
ABA-induced variation in carbon and oxygen isotope ratios
of cotton leaves. Plant, Cell and Environment 23, 473–485.
Botstein D, White RL, Skolnick M, Davis RW. 1980. Construction of a genetic map using restriction fragment length
polymorphisms. American Journal of Human Genetics 32,
314–331.
1002
Price et al.
Champoux MC, Wang G, Sarkarung S, Mackill DJ, O’Toole JC,
Huang N, McCouch SR. 1995. Locating genes associated
with root morphology and drought avoidance in rice via
linkage to molecular markers. Theoretical and Applied
Genetics 90, 969–981.
Causse M, Fulton TM, Cho YG, Ahn SN, Chunwongse J, Wu K,
Xiao J, Yu Z, Ronald PC, Harrington SB, Second GA,
McCouch SR, Tanksley SD. 1995. Saturated molecular map
of the rice genome based on an interspecific backcross
population. Genetics 138, 1251–1274.
Chaudary D, Rao MJBK. 1982. Breeding rice varieties for
dryland and drought-prone areas of India. In: Drought
resistance in crops with the emphasis on rice. Manila: IRRI,
265–272.
Churchill GA, Doerge RW. 1994. Empirical threshold values
for quantitative trait mapping. Genetics 138, 963–971.
Cook A, Marriott CA, Seel W, Mullins CE. 1997. Does the
uniform packing of sand in a cylinder provide a uniform
penetration resistance? A method for screening plants for
responses to soil mechanical impedance. Plant and Soil
190, 279–287.
Courtois B, McLaren G, Sinha PK, Prasad K, Yadav R, Shen L.
2000. Mapping QTLs associated with drought avoidance in
upland rice. Molecular Breeding 6, 55–66.
Crasta OR, Xu WW, Rosenow DT, Mullet J, Nguyen HT. 1999.
Mapping of post-flowering drought resistance traits in grain
sorghum: association between QTLs influencing premature
senescence and maturity. Molecular and General Genetics
226, 579–588.
Dewar RC. 1993. A root–shoot partitioning model based on
carbon–nitrogen water interactions and Munch phloem flow.
Functional Ecology 7, 356–368.
Dingkuhn M, Cruz RT, O’Toole JC, Dorffling K. 1989. Net
photosynthesis, water use efficiency, leaf water potential
and leaf rolling as affected by water deficit in tropical
upland rice. Australian Journal of Agricultural Research
40, 1171–1181.
Dingkuhn M, Cruz RT, O’Toole JC, Turner NC, Dorffling K.
1991a. Responses of seven diverse rice cultivars to water
deficits. III. Accumulation of abscisic acid and proline in
relation to leaf water-potential and osmotic adjustment. Field
Crops Research 27, 103–117.
Dingkuhn M, Farquhar GD, De Datta SK, O’Toole JC. 1991b.
Discrimination of 13C among upland rices having different
water use efficiencies. Australian Journal of Agricultural
Research 42, 1123–1131.
Ehleringer JR, Dawson TE. 1992. Water uptake by plants:
perspective from stable isotopes. Plant, Cell and Environment
15, 1073–1082.
Ekanayake IJ, Garrity DP, Masajo TM, O’Toole JC.
1985a. Root pulling resistance in rice: Inheritance and
association with drought tolerance. Euphytica 34,
905–913.
Ekanayake IJ, O’Toole JC, Garrity DP, Masajo TM. 1985b.
Inheritance of root characters and their relations to drought
resistance in rice. Crop Science 25, 927–933.
Evenson RE, Herdt RW, Hossain M. 1996. Rice research
in Asia: progress and priorities. Wallingford, UK: CAB
International.
Frova C, Krajewski P, di Fonzo N, Villa M, Sari-Gorla M. 1999.
Genetic analysis of drought tolerance in maize by molecular
markers. I. Yield components. Theoretical and Applied
Genetics 99, 280–288.
Farrar JF. 1992. The whole plant : carbon partitioning during
development. In: Pollock CJ, Farrar JF, Gordon AJ, eds.
Carbon partitioning within and between organs. Bios Scientific
Publishers, 163–180.
Fukai S, Cooper M. 1995. Development of drought-resistant
cultivars using physio-morphological traits in rice. Field Crops
Research 40, 67–86.
Gallais A, Rives M. 1993. Detection, number and effects of
QTLs for a complex character. Agronomie 13, 723–738.
Garrity DP, O’Toole JC. 1995. Selection for reproductive stage
drought avoidance in rice, using infrared thermometry.
Agronomy Journal 87, 773–779.
Gray JE, Holroyd GH, van der Lee FM, Bahrami AR,
Sijmons PC, Woodward FI, Schuch W, Heterington AM.
2000. HIC signalling pathway links CO2 perception to
stomatal development. Nature 408, 713–716.
Griffiths H. 1998. Stable isotopes: integration of biological,
ecological and geochemical processes. Oxford: Bios Scientific
Publishers.
Griffiths H, Borland AM, Gillon JS, Harwood KG, Maxwell K,
Wilson JM. 1999. Stable isotopes reveal exchanges between
soil, plants and the atmosphere. In: Press MC, Scholes JD,
Barker MG, eds. Physiological plant ecology. Blackwell
Science, 415–441.
Harwood KG, Gillon JS, Griffiths H, Broadmeadow MSJ. 1998.
Diurnal variation of Delta(CO2)-C-13, Delta(COO)-O-18O-16 and evaporative site enrichment of Delta(H2O)-O-18 in
Piper aduncum under field conditions in Trinidad. Plant, Cell
and Environment 21, 269–283.
Hemamalini GS, Shashidar HE, Hittalmani S. 2000. Molecular
marker assisted tagging of morphological and physiological
traits under two contrasting moisture regimes at peak
vegetative stage in rice (Oryza sativa L.). Euphytica 112,
69–78.
Horton P. 2000. Prospects for crop improvement through the
genetic manipulation of photosynthesis: morphological and
biochemical aspects of light capture. Journal of Experimental
Botany 51, 475–485.
Horton P, Murchie EH. 2000. C4 photosynthesis in rice: some
lessons from studies of C3 photosynthesis in field-grown rice.
In: Sheehy JE, Mitchell PL, Hardy B, eds. Redesigning
photosynthesis to increase yield. Elsevier Science bv, 127–144.
Jackson MB. 1993. Are plant hormones involved in root to shoot
communication? Advances in Botanical Research 19, 104–187.
Jiao D, Ji B. 2001. Photoinhibition in indica and japonica
subspecies of rice (Oryza sativa) and their reciprocal F1
hybrids. Australian Journal of Plant Physiology 28, 299–306.
Jones HG. 1981. The use of stochastic modelling to study the
influence of stomatal behaviour on yield–climate relationships. In: Rose DA, Charles-Edwards DA, eds. Mathematics
and plant physiology. London: Academic Press, 231–244.
Jones HG. 1993. Drought tolerance and water-use efficiency. In:
Smith JAC, Griffiths H, eds. Water deficits: plant responses
from cell to community. Oxford: Bios, 193–203.
Jones HG. 1999a. Use of thermography for quantitative studies
of spatial and temporal variation of stomatal conductance
over leaf surfaces. Plant, Cell and Environment 22, 1043–1055.
Jones HG. 1999b. Use of infrared thermometry for estimation
of stomatal conductance in irrigation scheduling. Agricultural
and Forest Meteorology 95, 139–149.
Kearsey MJ. 1998. The principles of QTL analysis (a minimal
mathemetaics approach). Journal of Experimental Botany
49, 1619–1623.
Kurata N, Nagamura Y, Yamamoto K, et al. 1994. A 300
kilobase interval genetic map of rice including 883 expressed
sequences. Nature Genetics 8, 365–372.
Rice drought resistance
Lander ES, Green P, Abrahamson J, Barlow A, Daly MJ,
Lincoln SE, Newburg L. 1987. Mapmaker: an interactive
computer package for constructing primary genetic linkage
maps of experimental and natural populations. Genomics
1, 174–181.
Levitt J. 1980. Responses of plants to environmental stresses,
Vol. 2. Water, radiation, salt and other stresses. New York:
Academic Press, 93–128.
Lilley JM, Ludlow MM. 1996. Expression of osmotic adjustment and dehydration tolerance in diverse rice lines. Field
Crops Research 48, 185–197.
Lilley JM, Ludlow MM, McCouch SR, O’Toole JC. 1996.
Locating QTL for osmotic adjustment and dehydration
tolerance in rice. Journal of Experimental Botany 47,
1427–1436.
Lu C, Shen L, Tan Z, Xu Y, He P, Chen Y, Zhu L. 1997.
Comparative mapping of QTLs for agronomic traits of rice
across environment by using a double-haploid population.
Theoretical and Applied Genetics 94, 145–150.
Mambani B, Lal R. 1983a. Response of upland rice varieties
to drought stress. I. Relationship between root system
development and leaf water potential. Plant and Soil 73,
59–72.
Mambani B, Lal R. 1983b. Response of upland rice varieties
to drought stress. II. Screening rice varieties by means of
variable moisture regimes along a toposequence. Plant and
Soil 73, 73–94.
Mambani B, Lal R. 1983c. Response of upland rice varieties to
drought stress. III. Estimating root system configuration
from soil moisture data. Plant and Soil 73, 95–104.
Murchie EH, Chen Y-Z, Hubbart S, Peng S, Horton P. 1999.
Interactions between senescence and leaf orientation determine in situ patterns of photosynthesis and photoinhibition
in field-grown rice. Plant Physiology 119, 553–563.
Nguyen HT, Babu RC, Blum A. 1997. Breeding for drought
resistance in rice: physiological and molecular genetics
considerations. Crop Science 37, 1426–1434.
O’Toole JC. 1982. Adaptation of rice to drought-prone environments. In: Drought resistance in crops with the emphasis on
rice. Manila: IRRI, 195–213.
O’Toole JC, Cruz RT. 1983. Genotypic variation in epicuticular
wax of rice. Crop Science 23, 393–394.
Pflieger S, Lefebvre V, Causse M. 2001. The candidate gene
approach in plant genetics: a review. Molecular Breeding
7, 275–291.
Price AH, Courtois B. 1999. Mapping QTLs associated with
drought resistance in rice: progress, problems and prospects.
Plant Growth Regulation 29, 123–133.
Price AH, Steele KA, Gorham J, Bridges JM, Moore BJ,
Evans JL, Richardson P, Jones RGW. 2002a. Upland rice
grown in soil-filled chambers and exposed to contrasting
water-deficit regimes. I. Root distribution, water use and
plant water status. Field Crops Research (in press).
Price AH, Steele KA, Moore BJ, Barraclough PB, Clark LJ.
2000. A combined RFLP and AFLP linkage map of upland
rice (Oryza sativa L.) used to identify QTLs for rootpenetration ability. Theoretical and Applied Genetics 100,
49–56.
Price AH, Steele KA, Moore BJ, Jones RGW. 2002b. Upland
rice grown in soil-filled chambers and exposed to contrasting
water-deficit regimes: II. Mapping QTL for root morphology
and distribution. Field Crops Research (in press).
Price A, Steele K, Townend J, Gorham G, Audebert A, Jones M,
Courtois B. 1999. Mapping root and shoot traits in rice:
experience in UK, IRRI and WARDA. In: Ito O, O’Toole J,
1003
Hardy B, eds. Genetic improvement of rice for water-limited
environments. Manila, Philippines: International Rice
Research Institute, 257–273.
Price AH, Tomos AD. 1997. Genetic dissection of root growth
in rice (Oryza sativa L.). II. Mapping quantitative trait loci
using molecular markers. Theoretical and Applied Genetics
95, 143–152.
Price AH, Townend J, Jones MP, Audebert A, Courtois B. 2002c.
Mapping QTLs associated with drought avoidance in upland
rice grown in the Philippines and West Africa. Plant
Molecular Biology (in press).
Price AH, Virk DS, Tomos AD. 1997a. Genetic dissection of
root growth in rice (Oryza sativa L.). I. A hydroponic screen.
Theoretical and Applied Genetics 95, 132–142.
Price AH, Young EM, Tomos AD. 1997b. Quantitative trait
loci associated with stomatal conductance, leaf rolling and
heading date mapped in upland rice (Oryza sativa). New
Phytologist 137, 83–91.
Puckridge DW, O’Toole JC. 1981. Dry matter and grain
production of rice, using a line source sprinkler in drought
studies. Field Crops Research 3, 303–319.
Ray JD, Yu L, McCouch SR, Champoux MC, Wang G,
Nguyen H. 1996. Mapping quantitative trait loci associated
with root penetration ability in rice (Oryza sativa L.).
Theoretical and Applied Genetics 92, 627–636.
Raskin I, Ladyman AR. 1988. Isolation and characterization of
a barley mutant with abscisic-acid-insensitve stomata. Planta
173, 73–78.
Ribaut JM, Jiang C, Gonzalez deLeon, Edmeades GO,
Hoisington DA. 1997. Identification of quantitative trait
loci under drought conditions in tropical maize. 2.
Yield components and marker assisted selection strategies.
Theoretical and Applied Genetics 94, 887–896.
Stam P. 1993. Construction of integrated genetic linkage maps
by means of a new computer package: Joinmap. The Plant
Journal 5, 739–744.
Stuber CW, Polacco M, Senior ML. 1999. Synergy of empirical
breeding, marker-assisted selection, and genomics to increase
crop yield potential. Crop Science 39, 1571–1583.
Tanksley SD, Ganal MW, Martin GB. 1995. Chromosome
landing—a paradigm for map-based gene cloning in plants
with large genomes. Trends in Genetics 11, 63–68.
Thornley JHM. 1972. A balanced quantitative model for
root : shoot ratios in vegetative plants. Annals of Botany
36, 431–441.
Tripathy JN, Zhang J, Nguyen TT, Nguyen HT. 2000. QTLs for
cell-membrane stability mapped in rice (Oryza sativa L.)
under drought stress. Theoretical and Applied Genetics
100, 1197–1202.
Turner NC, O’Toole JC, Cruz RT, Namuco OS, Ahmad S.
1986a. Responses of seven diverse rice cultivars to water
deficits. I. Stress development, canopy temperature, leaf
rolling, and growth. Field Crops Research 13, 257–271.
Turner NC, O’Toole JC, Cruz RT, Yamboa EB, Ahmad S,
Namuco OS, Dingkuhn M. 1986b. Responses of seven diverse
rice cultivars to water deficits. II. Osmotic adjustment, leaf
elasticity, leaf extension, leaf death, stomatal conductance,
and photosynthesis. Field Crops Research 13, 273–286.
Van der Werf A, Nagel O. 1996. Carbon allocation to shoots and
roots in relation nitrogen supply is mediated by cytokinins
and sucrose: opinion. Plant and Soil 185, 21–32.
Yadav R, Courtois B, Huang N, McLaren G. 1997. Mapping
genes controlling root morphology and root distribution on
a double-haploid population of rice. Theoretical and Applied
Genetics 94, 619–632.
1004
Price et al.
Yoshida S, Hasegawa S. 1982. The rice root system, its
development and function. In: Drought resistance in crops
with the emphasis on rice. Manila: IRRI, 83–96.
Yu L, Ray JD, O’Toole JC, Nguyen HT. 1995. Use of waxpetrolatum layers to simulate compacted soil for screening
rice (Oryza sativa L.) root penetration ability. Crop Science
35, 684–687.
Zhang J, Babu RC, Pantuwan G, Kamoshita A, Blum A,
Sarkarung S, O’Toole JC, Nguyen HT. 1999. Molecular
dissection of drought tolerance in rice: from physiomorphological traits to field performance. In: Ito O,
O’Toole J, Hardy B, eds. Genetic improvement of rice for
water-limited environments. Manila: IRRI, 331–343.
Zheng H, Babu R, Pathan M, Ali L, Huang N, Courtois B,
Nguyen HT. 2000. Quantitative trait loci for root-penetration
ability and root thickness in rice: comparison of genetic
backgrounds. Genome 43, 53–61.