11 Rice Genetic Improvement for Abiotic Stress Tolerance in Africa

17 Pages·2013·0.12 MB·English

11
Rice Genetic Improvement for
Abiotic Stress Tolerance in Africa
Khady N. Dramé,1* Baboucarr Manneh2 and Abdelbagi M. Ismail3
1Africa Rice Center, Dar es Salaam, Tanzania; 2Africa Rice Center, Saint-Louis,
Senegal; 3International Rice Research Institute, Los Baños, Philippines
Introduction Crop and water management can be
improved through efficient irrigation schemes,
In sub-Saharan Africa, biophysical constraints better infrastructure and hydrological control,
and high reliance on low-input and rainfed agri- and optimal use of fertilizers, which can alle-
culture impose various abiotic stresses on rice viate some of the effects of abiotic stresses.
crops (Defoer etal., 2002). Most of these stresses These measures are effective in reducing yield
are associated with water availability (drought losses, and most improved varieties are respon-
and excess water), soil problems (salinity, nutri- sive to such measures, consequently stabiliz-
ent deficiencies and toxicities) and extreme tem- ing productivity and narrowing yield gaps
peratures (heat and cold). These stresses (Ismail et al., 2008). However, the costs of
fluctuate seasonally and vary spatially. They adopting these measures are prohibitive for
adversely affect rice growth and productivity, most resource-poor African farmers and local
resulting in reduced yields. Abiotic stresses are governments. Alternatively, developing rice
limiting factors in almost all rice production germplasm tolerant to the prevalent stresses
environments – in both rainfed and irrigated involves no additional costs to farmers and
areas (Defoer et al., 2002; Saito et al., Chapter can enhance and stabilize productivity consid-
15, this volume; Table 11.1). They contribute to erably (Ismail and Tuong, 2009). Hence, vari-
significantly lower on-farm yields compared to etal improvement should be given priority as
attainable yield. Results of farm-household sur- an entry point for increasing and stabilizing
veys conducted by the Africa Rice Center rice production in sub-Saharan Africa. As
(AfricaRice) in 12 sub-Saharan African coun- much as possible, this should be combined
tries showed that, for example, when drought with improved management practices to real-
and flooding occur, they induce average yield ize the full yield potential of the new tolerant
losses of 33% (AfricaRice, 2011a). Yield losses varieties developed. This is all the more impor-
due to soil-related problems are higher in the tant given that the potential impact of
lowlands than in the uplands. In Senegal and research targeted to reduce yield loss due to
Uganda, yield losses due to the soil-related con- stresses was estimated at a global cumulative
straints salinity and iron toxicityare as high as benefit of US$ 32.9 million for seven countries
40% and 25%, respectively (AfricaRice, 2011b). over 3 years (AfricaRice, 2011b). This means
* Corresponding author: [email protected]
© CAB International 2013. Realizing Africa’s Rice Promise
144 (eds M.C.S. Wopereis et al.)
Genetic Improvement for Abiotic Stress Tolerance 145
Table 11.1. Major rice production ecosystems in Africa: actual and potential yield and limiting constraints.
(Adapted from Defoer etal., 2002; Diagne etal., Chapter 3, this volume).
Yield: Actual /
Rice Share of rice Potentiala
agro-ecosystem area (%) (t/ha) Abiotic production constraints Input use
Rainfed upland 32 1.2/2–4 P and N deficiency, acidity, Al toxicity, drought, Very low
erosion declining soil fertility, cold in
highlands
Rainfed lowland 38 1.9/3–6 Water control, N and P deficiency, Low
Fe toxicity
Irrigated Sahel/ 26 3.7/6–11 N deficiency, salinity and alkalinity, extreme High
savannah temperatures
Irrigated humid/ 1.9/5–8 N deficiency, Fe toxicity Medium
sub-humid zone
Mangrove <1/2–4 Acid sulfate, salinity, Fe toxicity, Very low
excess water
Deepwater/ 4 <1/2–4 Salinity, excess water, cold in highlands Very low
Floating
aLow end of the range refers to potential yield at current input levels; high end refers to potential yields at increased input
levels.
additional income for farmers and increased Africa and the breeding activities to develop
rice production in Africa. rice varieties tolerant of these stresses and
In the rice gene pool – both cultivated and adapted to local conditions in Africa. We also
wild species – tremendous genetic variation highlight the challenges that should be
exists for tolerance to abiotic stresses. This addressed in order to sustainably increase rice
could be exploited to develop improved varieties productivity in sub-Saharan Africa.
that are more tolerant to the major abiotic
stresses that constrain rice production in sub-
Saharan Africa than existing varieties. Progress
Major Abiotic Stresses
in developing rice varieties tolerant to abiotic
Limiting Rice Production
stresses has been slow because of the complex-
ity of the tolerance mechanisms, poor under- in Sub-Saharan Africa
standing of the inheritance of tolerance, low
heritability and lack of efficient screening tech- Sub-Saharan Africa is affected by numerous
niques (Lafitte etal., 2006). However, advances types of environmental stresses related to water
in understanding stress physiology and identifi- availability, soil problems and climatic issues. In
cation of tolerance genes/QTLs (quantitative many cases, several stresses are experienced
trait loci) (Ismail etal., 2007; Jena and Mackill, simultaneously (e.g. in mangrove-swamp rice
2008; Thomson etal., 2010b) offer promising fields where submergence, salinity and iron tox-
opportunities for rice improvement with icity occur together). Rice production in such
regards to abiotic-stress tolerance. Both con- areas is severely constrained by abiotic stresses.
ventional and biotechnological approaches are Specific data on the extent of abiotic stresses in
being used by AfricaRice to exploit the rich res- rice-growing areas of sub-Saharan Africa are
ervoir of genetic resources present in the scarce. Most of the data available are either not
African germplasm pool, particularly Oryza specific to rice-growing areas or limited to a few
sativa landraces and O. glaberrima for the countries. This issue is being addressed by
improvement of local rice varieties. In this AfricaRice. Insight on farmer perceptions of
chapter, we present the major abiotic stresses abiotic constraints is provided by Diagne et al.
that constrain rice production in sub-Saharan (Chapter 4, this volume).
146 K.N. Dramé et al.
Drought led to a rapid rise in the water table and an
increase in soil sodium/alkalinity and salinity
Drought is generally avoided in areas where (Bertrand et al., 1993). In the Office du Niger
irrigation water is available throughout the (Mali), 50% of the water table is now saline and
season, but it is a consistent feature across occasionally very saline despite low mineral con-
much of the 63.5 million hectares of rainfed tent of the irrigation water (Bertrand etal., 1993).
rice sown annually, most of which is in tropical In the Senegal River delta, marine-derived soil
Asia, Africa and Latin America (Lafitte et al., salinity is an inherent problem and sodicity is
2006). It can occur at any stage during the rice increasing in irrigated flood plains in inland areas
cropping season, but it is more devastating due to high evaporation, rising groundwater
when it occurs just prior to flowering than it is tables and poor drainage (WARDA, 1993; Matlon
during the vegetative stage, with substantial et al., 1998). Miézan and Dingkuhn (2001)
effects on grain yield (Boojung and Fukai, observed that waters of the Niger and Senegal
1996). About 70% of the rice area in sub- rivers carry substantial alkalinity, and the salt
Saharan Africa is rainfed (Diagne et al., content of water sometimes increases markedly
Chapter 3, this volume; Table 11.1). The spa- between the main rivers and the actual irrigation
tial and temporal variability of rainfall in this site. However, van Asten etal. (2003) show that
region expose rice plants to frequent drought salt accumulation in the soils of Sahelian coun-
spells. Regardless of the total rainfall and dis- tries has more to do with the underlying bedrock
tribution, the poor physical properties of than with the irrigation system. Examining soils
highly weathered, coarse-textured soils in in the irrigated areas of Foum Gleita (Mauritania)
some parts of sub-Saharan Africa induce low they found that the geographical distribution of
water-holding capacity and establish water salt was not related to the presence of irrigation
deficit as a major constraint to rainfed crop or drainage canals. Also the alkaline salts present
yields in sub-Saharan Africa (Hsiao et al., in the upper soil layers in Foum Gleita did not
1980). This is particularly true for upland rice, come from irrigation water, but from the under-
which makes up 32% of rice-growing areas in lying bedrock. Additional to the salt stress itself,
sub-Saharan Africa (Table 11.1). Analysis of the high pH resulting from the sodification/
farmer perceptions in 18 countries in sub-Saharan alkalinization reduces the availability of plant
Africa across rice environments provided evi- nutrients such as zinc and increases nitrogen
dence that in 2008 an estimated 10% of rice losses through volatilization (Bertrand et al.,
farmers experienced drought affecting 37% of 1993; Miézan and Dingkuhn, 2001). According
their rice area, causing 29% of rice yield loss to van Asten et al. (2004), using data from Sourou
(Diagne et al., Chapter 4, this volume). The Valley (Burkina Faso), this shortage of nutrients
diversity of affected production systems, vari- in the soil is a bigger problem than soil degrada-
ability of drought in terms of timing and sever- tion attributable to irrigation.
ity, and the multiple traits involved in drought High rice productivity is also constrained by
tolerance require strategic research to prioritize soil salinity in many mangrove-swamp areas.
and define environment-specific approaches Mangrove rice is grown on about 200,000 ha of
for developing drought-tolerant rice cultivars cleared mangrove swamps along the rivers and
(Mannehetal., 2007). coastal estuaries of The Gambia, Guinea, Guinea-
Bissau, Nigeria, Senegal and Sierra Leone (Matlon
et al., 1998). In these areas, rice cultivation
depends on the length of the salt-free period,
Salinity which is an interplay of in situ rainfall, the volume
of freshwater flow, and salt intrusion from the
Irrigation has the potential to ensure high rice sea. Swamps located along river banks farthest
yields and is a good strategy to offset recurrent from the sea experience a salt-free period of 6–9
droughts. Unfortunately, soils of most irrigated months during which rice can be grown. Closer to
areas in sub-Saharan Africa have continued to be the sea, rice can be grown for only 3–4 months.
degraded as a result of poor irrigation practices Rice is most sensitive to salinity stress at
and the absence of efficient drainage. These have seedling and reproductive stages (Gregorio
Genetic Improvement for Abiotic Stress Tolerance 147
et al., 2002). However, salinity tolerance at While rice is adapted to waterlogged condi-
these two stages is only weakly associated tions, complete submergence for several days
(Moradi etal., 2003). Discovering and combin- can be fatal. However, the extent of damage to
ing suitable tolerance traits for both stages is rice is affected by several factors linked to flood-
essential for developing resilient salt-tolerant water conditions, including interference in
varieties. Moreover, the salt-tolerance level of normal gas exchange and light interception.
cultivars depends on environmental condi- These factors are largely affected by the dura-
tions (Asch et al., 1997). Generally, salinity tion of the flood, its depth, temperature, and
effects on rice are more severe in arid climates the level of turbidity and turbulence of the
than in humid ones. For example, salinity lev- floodwater, and vary considerably across loca-
els at an electric conductivity (EC) of 9.5 mS/ tions and seasons (Das etal., 2009). Rice yields
cm were reported to cause a 50% yield loss in in flood-prone areas are very low, mostly aver-
the humid tropics (Flowers and Yeo, 1981), aging below 1.5 t/ha, because farmers usually
whereas under the arid conditions of the grow their low-yielding traditional varieties
Sahelian dry season an equivalent yield loss (Haefele etal., 2010); and modern high-yielding
was observed at an EC of only 3.5 mS/cm varieties tolerant of these stresses are mostly
(Dingkuhnetal., 1993). non-existent. In some locations, such as Mopti
(Mali) and the Sokoto Rima River flood plains
(Nigeria), O. glaberrima (African rice) is still cul-
tivated despite its low yield potential relative to
Excess water the more widely cultivated O. sativa (Asian rice)
(Jones et al., 1997; Diarra et al., 2004;
Excess water can be of two types: transient flash Sakagami, 2012).
floods or submergence that completely inundates
the crop for a short duration (up to 2 weeks), and
longer-term flooding where water stagnates for up Phosphorus deficiency
to a few months at different water depths (e.g.
30–50 cm: partial/stagnant, semi-deep; >100 Phosphorus (P) deficiency is one of the major
cm: deep-water; up to 3 or 4 m: very deep-water or limiting factors for crop production in highly
floating rice) (Mackill etal., 2012). In sub-Saharan weathered soils in the humid tropics. Total soil P
Africa, the deep-water area accounts for less than is generally low, with only 2–4% of the total P
4% of the whole rice-growing area (Diagne etal., available to plants. This is because of the high
Chapter3, this volume), mainly in the flood plains P-fixing capacity of fine-textured soils found in
of large rivers such as the Niger River. Excess humid and sub-humid zones (Abekoe and
water is also a common constraint throughout Sahrawat, 2001). About 5.7 billion hectares
the rainfed and irrigated rice-production areas worldwide lack sufficient plant-available P
when high rainfall and/or impeded drainage (Batjes, 1997) and 50% of potential arable land
occurs, particularly early in the season (Futakuchi, worldwide has acid soils, which are widely
2005). In the flood plains, sudden floods can tem- distributed in sub-Saharan Africa. The bioavail-
porarily submerge the rice crop in some areas, and ability of inorganic phosphate is reduced in
waterlogging can continue through most of the acidsoils.
season in others – sometimes both occur within Phosphorus deficiency can be corrected
the same season. In savannah and forest zones in through fertilizer application, but the lack of
West and Central Africa, inland valleys prevail locally available P sources and the high cost of
and rice in valley bottoms often experiences water- importing and transporting fertilizers prevent
logging for several weeks after heavy rainfall many resource-poor rice farmers from applying
(Mannehetal., 2007). Analysis of farmer percep- P. Furthermore, some rice soils can quickly fix
tions in 18 countries in sub-Saharan Africa pro- upto 90% of the added P fertilizer into less-soluble
vided evidence that in 2008 an estimated 5% of forms (Dobermann etal., 1998) that cannot be
farmers experienced flooding, affecting 37% of used by plants. This tight binding of P in the soil
their farmland, causing a production loss of 27% is frequently the primary cause of P deficiency,
(Diagne et al., Chapter 4, this volume). rather than a low total P content.
148 K.N. Dramé et al.
Insufficient plant-available soil P is a major WARDA, 1997). Excessive uptake of Fe is often
constraint for rice production. This is particu- accompanied with other nutrient deficiencies,
larly apparent in sub-Saharan Africa under especially potassium, phosphorus, calcium and
upland conditions, which are commonly char- magnesium (Benckiser etal., 1984; Prade etal.,
acterized by infertile, highly acidic, P-fixing soils, 1990). However, Sahrawat et al. (1996) report
usually in areas where little or no fertilizer is cases of Fe toxicity without apparent deficiency
applied. Under lowland conditions, where P is in other nutrients.
more available than in drier areas, P deficiency
is still a major factor limiting performance of
modern rice varieties (Defoer et al., 2002).
Phosphorus deficiency is likely to become an Extreme temperatures
increasingly important constraint, as P is
removed from soils under intensive rice produc- Low-temperature stress is common for rice
tion (De Datta etal., 1990) using high-yielding grown in temperate regions and at high eleva-
modern varieties. tions in the tropics. In Africa, cold tempera-
tures occur in the highlands of East and
Southern Africa and in some areas of the Sahel
region of West Africa during the cold, dry har-
Iron toxicity mattan season, which extends from November
to February. For example, in Madagascar, mean
Iron (Fe) toxicity is widely distributed in tropi- temperatures at 1500 m vary from 17°C in
cal lowlands and is frequently reported in October, the rice-sowing period, to 20°C during
many inland valleys (mangrove swamps, rain- the reproductive stage (Zenna et al., 2010).
fed and irrigated lowlands) of sub-Saharan Minimum temperatures can fall below 10°C
Africa (Masajo etal., 1986). It is usually asso- during early vegetative stage and below 14°C
ciated with poor drainage and the presence of during reproductive stage (Zenna etal., 2010).
iron in the parent rock or in the soils of adja- In the Sahel, there are significant temperature
cent slopes through which groundwater flows fluctuations during the year with low tempera-
into the lowland (WARDA, 1988). A wide tures during panicle initiation in the wet sea-
range of soil types can be Fe-toxic, including son and high temperatures around flowering
acid-sulfate soils, acid-clay soils, peat soils and during the dry season. In both cases, spikelet
valley-bottom soils receiving interflow water sterility may occur, leading to substantial yield
from adjacent slopes (Becker and Asch, 2005). loss (Dingkuhn, 1995). In this region, day min-
A survey conducted during 2000–2001 in imum temperatures fall below 20°C during
three West African countries (Côte d’Ivoire, November to March, while maximum tempera-
Ghana and Guinea) showed that more than tures regularly rise above 40°C from April to
50% of the lowlands studied and about 60% of June (Fig. 11.1). Planting rice between mid-
the cultivated rice plots were affected by Fe tox- September and mid-November in the Sahel is
icity, and 10% of lowland crop fields were generally associated with near total spikelet
abandoned due to high Fe-toxicity stress sterility, since the reproductive phase of the
(Chérif etal., 2009). crop coincides with periods of low night tem-
In anaerobic conditions and acidic condi- peratures (Dingkhun, 1995; Manneh et al.,
tions (pH <5), as found in waterlogged soils, the 2007). Moreover, crop duration increases with
ferrous form is stabilized and readily taken up by low temperatures, considerably limiting the
plants. Rice is particularly sensitive to Fe toxicity possibility of double-cropping in areas where
at two growth stages – soon after transplanting water control is possible (Coly, 1980; Matlon
up to tillering, and during heading/flowering etal., 1998). However, the introduction of short-
(Prade et al., 1990). Depending on the variety duration varieties can make double-cropping
and the intensity of the toxicity, yield losses aver- feasible. Dingkuhn and Miezan (1995) used a
age about 30% and can accrue up to levels that simulation model based on photothermal con-
can cause complete crop failure of susceptible stants to show that short-duration genotypes
varieties (Masajo et al., 1986; Abifarin, 1989; with high plasticity of duration perform best
Genetic Improvement for Abiotic Stress Tolerance 149
50.0
45.0
40.0
35.0
C) 30.0
°
e (
atur 25.0
er
p
m
e 20.0
T
15.0
Tmax (°C)
Tmin (°C)
10.0
5.0
0.0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct NovDec
Fig. 11.1. Example of seasonal and diurnal temperature variations during two major cropping seasons in
the Sahel (case of Fanaye, Senegal River valley, northern Senegal, 2011). Tmax, maximum temperature;
Tmin, minimum temperature.
in the dry season, while medium-duration peratures below 15°C. The extent of damage
types with high plasticity perform well in the depends on the ambient air or watertempera-
wet season, and short-duration types with ture, cropping pattern, growth stage and vari-
lowplasticity of duration perform well in both ety (Zenna et al., 2010). Crop failure can be
seasons. Rising temperatures, as a consequence observed when low temperature is manifested
of climate change, could have positive effects at different growth stages, such as germina-
on the flexibility of cropping calendars, but tion, seedling, vegetative, reproductive and
the vulnerability of rice to high temperature is grain maturity (Andaya and Mackill, 2003a,b).
also expected to increase. Global mean tem- When rice is exposed to air temperatures
peratures have already risen by around 0.6°C higher than 35°C, heat injuriesoccur. By stud-
over the last century and are projected to ying weather data at the International Rice
increase by 1.4°C to 5.8°C over the next cen- Research Institute (IRRI) farm from 1979 to
tury (IPCC, 2001). 2003, Peng et al. (2004) show that grain
Rice is sensitive to temperature fluctua- yield declined by 10% for each 1°C increase
tions, but its sensitivity depends on the devel- in minimum temperature during the dry
opmental stage. Flowering stage and 9–11 season, whereas the effect of maximum
days before heading are the most sensitive temperature on crop yield was insignificant.
stages to extremes of temperatures, with very In a different study, Ziska and Manalo
cold or very hot weather leading to high spike- (1996) found that, at a constant day tem-
let sterility (Yoshida etal., 1981; Andaya and perature of 29°C, increasing night tempera-
Mackill, 2003a; Manneh et al., 2007; Zenna ture did not significantly alter growth or
etal., 2010). Rice is particularly sensitive to tem- yield; however, increasing night temperature
150 K.N. Dramé et al.
at a day temperature of 33°C resulted in signifi- in rice, with particular focus on African rice
cant declines in grain-filling and grain yield. (O.glaberrima) and related wild species, O. barthii;
and (ii) the use of molecular markers to trans-
fer tolerance QTLs to elite African germplasm,
in addition to conventional breeding. Efforts
Strategies to Improve Rice Tolerance
are also devoted to identify new QTLs for
to Abiotic Stresses for Africa tolerance.
To ensure farmers’ access to these new
Various conventional and biotechnological technologies and their rapid adoption, AfricaRice
approaches are being used to develop rice varie- is progressively introducing participatory app-
ties tolerant of abiotic stresses. It is now possible roaches into the various breeding programmes.
to develop new tolerant rice varieties through In different countries, breeding lines and varie-
conventional breeding in combination with ties selected for tolerance to salt stress, drought,
marker-assisted selection or through direct Fe toxicity and cold were evaluated by farmers
transfer of tolerance genes into rice varieties via through field visits and on-farm tests against
genetic engineering. At AfricaRice, two main their local varieties. By 2011, some 21 promising
strategies are being followed for genetic improve- stress-tolerant lines had been identified for
ment of rice tolerance to abiotic stresses potential submission to national varietal re -
(Fig.11.2): (i) the use of genetic diversity existing lease committees. In 2010, an Africa-wide Rice
Identification of donors from
ery genetic stocks
v (GRU, collections,
o
sc introductions) Identification of
e di Hybridization/Recombination major QTLs
n Fine mapping
e
G if needed
n Development of Major successful Identification of
o
ati breeding lines QTLs reported linked markers
c
erifi
V
Preliminaryevaluation in Marker-assisted selection
unreplicated trials of elite varieties
nt
me Selections Phenotypicevaluation of elite lines
p with desired genotypes in target
o
el environments
v
e
d
e
n
Li Multi-environment testing
Farmers’ and consumers’ Farmers’ and
preferences consumers’ preferences
On-farm trials and PVS
n
o
ati Varietal release and dissemination
n
mi
e
s
s Seed production
Di
(breeder, foundation, certified)
Fig. 11.2. Framework of a breeding process for genetic improvement of stress tolerance in rice. GRU,
Genetic Resources Unit; PVS, participatory varietal selection; QTLs, quantitative trait loci.
Genetic Improvement for Abiotic Stress Tolerance 151
Breeding Task Force was revived and this net- tolerant varieties. Variability also exists in the
work is now responsible for multi-environment interspecific lowland NERICA and O. glaberrima
testing of new breeding lines developed for varieties with regards to Fe-toxicity tolerance.
lowland, irrigated, upland, high-elevation and NERICA-L 19 and NERICA-L 49 combine both
mangrove environments (see Kumashiro et al., tolerance to iron toxicity and good agronomic
Chapter 5, this volume). characters (Nwilene etal., unpublished). The tol-
erance was inherited from the O. glaberrima par-
ent, TOG5681. It has also been reported that
CG14, the O. glaberrima parent ofupland NERICA
Tapping the gene pool of African rice varieties, has remarkable tolerance to Fe-toxicity
stress (Sahrawat and Sika, 2002), while
With the success of NERICA varieties, O. glaber- TOG7235, TOG6216, TOG7442 and TOG7291
rima gained renewed interest. This species is show moderate tolerance (Mendoza etal., 2000).
known for its good adaptation to sub-Saharan About 500 accessions from the AfricaRice O. gla-
Africa environmental and soil conditions. It has berrima collection are being evaluated at Vallée du
tolerance to many abiotic stresses, including Kou (Burkina Faso) to identify new sources of tol-
drought, Fe toxicity, acidity and low-input condi- erance to Fe toxicity in this species. New research
tions (Sano etal., 1984; Jones etal., 1997) and directions are also being considered in view of the
has the ability to grow in a wide range of unfa- current and anticipated rise in temperature
vourable ecosystems such as rainfed hilly areas, resulting from climate change. New projects are
deepwater floating conditions and in coastal in the pipeline to incorporate heat tolerance into
mangrove areas (Sarla and Swamy, 2005). local African varieties. For this trait, too, O. glaber-
While AfricaRice holds a collection of about rima could be exploited to confer early morning
2500 O.glaberrima accessions, only four acces- flowering in interspecific crosses to avoid spikelet
sions were used to create the NERICA varieties sterility due to heat stress (Yoshida etal., 1981;
for both upland and lowland environments. Mannehetal., 2007).
There is still a tremendous amount of unex- One of the major limitations in the use of
ploited genetic diversity in the primary gene pool O.glaberrima for the improvement of O. sativa was
of O. glaberrima. AfricaRice has embarked on the the interspecific hybrid sterility (Ghesquiere
characterization of O. glaberrima accessions and etal., 1997). Even though AfricaRice succeeded
their evaluation for different traits of interest, in the development of interspecific varieties –
including tolerance to abiotic stresses. NERICA – the introgression of useful traits from
Several O. glaberrima were collected in Mali O. glaberrima into O. sativa is still tedious and
(RAM series) and several donors of drought time-consuming. AfricaRice and Institut de
tolerance have been identified, including RAM3, recherche pour le développement (IRD) have
RAM118 and RAM152 (Ndjiondjop etal., 2007; developed interspecific bridges between the two
Bimpongetal., 2011a). The development of back- cultivated species. These interspecific bridges com-
cross interspecific lines (O. sativa × O. glaberrima) priseO. sativa lines carrying large introgressions
showed that drought tolerance from O.glaberrima of theO. glaberrima genome that are compatible
is transferable to the progenies. Several intro- with O. sativa in subsequent crosses. Marker-
gression lines from different cross combinations assisted selection was carried out on backcross
were evaluated and lines tolerant of drought have progenies to select the homozygous lines bearing
been identified (Ndjiondjop et al., 2007, 2010; the O. glaberrima allele at the S1 locus. The S1
Bimpong et al., 2011b; Bocco et al., 2011). gene encodes a gamete eliminator which induces
Similarly, good sources of salinity tolerance both male and female gamete abortion through
have been identified within O. glaberrima (e.g. allelic interaction in interspecific O. sativa ×
RAM121, TOG6224 and TOG7230), as well as O.glaberrima crosses (Sano, 1986; Garavito etal.,
some of the interspecific lowland NERICA-L vari- 2010). The fertility restoration is monitored for
eties. These tolerant varieties do not have the tol- three generations to derive fertile backcross
erance allele of Pokkali at the Saltol locus. Further inbred lines (BILs – BC F ) with improved cross-
1 3
studies are assessing whether these alleles can ability with O. sativa. For more details see Lorieux
effectively be combined with Saltol to breed highly et al. (Chapter 10, this volume).
152 K.N. Dramé et al.
In addition to O. glaberrima, genetic varia- under drought have been identified that can
bility for tolerance to abiotic stresses is also pre- result in substantial yield improvement under
sent in O. sativa, especially in the traditional drought (Bernier etal., 2007; Venuprasad etal.,
varieties, and this is being exploited. Through 2009; Vikram etal., 2011). These QTLs offer con-
germplasm exchange, tolerant varieties identi- siderable opportunities for enhancing drought
fied elsewhere are included in breeding pro- tolerance of rice varieties for Africa for both
grammes to enhance tolerance to abiotic stresses uplands and lowlands. Drought tolerance QTLs
in the local varieties. Meanwhile, agronomic qtl12.1 and DTY3.1, respectively associated with
and quality traits preferred by consumers are yield under upland stress and lowland stress
incorporated in the new tolerant lines to meet (Bernier et al., 2007; Venuprasad et al., 2009),
the market demand. are targeted for marker-assisted selection (MAS)
at AfricaRice. Near-isogenic lines (NILs) con-
trasting for grain yield under drought, developed
Speeding up the development of tolerant at IRRI, are being tested under upland and low-
varieties through molecular breeding land conditions in Benin and Nigeria. The qtl12.1
co-localizes with the Pup1 (P uptake-1) QTL that
With a better understanding of the mechanisms confers tolerance to P deficiency (Chin et al.,
and genetics of stress tolerance, breeders are 2010). AfricaRice is aiming to transfer the Pup1
now using more precise breeding approaches – QTL into the background of popular upland vari-
notably marker-assisted backcrossing (MABC) – eties. A preliminary survey using Pup1 gene-spe-
to develop varieties with higher levels of tolerance cific markers showed that almost all upland
and acceptable grain quality. For many abiotic varieties tested (including upland NERICA varie-
stresses, major QTLs/genes have been identified ties) have Kasalath (Pup1 donor) allele at all or
in rice (Jena and Mackill, 2008; Ismail and part of the loci tested, but very few have the gene
Thomson, 2011). Some of these QTLs have been OsPupK46-2, which seems to be the major deter-
fine-mapped and a MABC system developed to minant of Pup1 effect (Gamuyao et al., 2012).
incorporate them into high-yielding varieties. This gives room for improvement of both drought
Incorporation of these QTLs into popular varie- and P deficiency.
ties demonstrates substantial impacts on rice For irrigated areas affected by salinity, the
productivity in farmers’ fields (Singh et al., tolerance QTL Saltol, derived from the salt-
2009; Thomson et al., 2010a,b; Ismail and tolerant cultivar Pokkali (Bonilla et al., 2002;
Thomson, 2011; Mackill etal., 2012). Likewise Thomson et al., 2010a), is being incorporated
at AfricaRice, MABC has been initiated for all into elite irrigated rice varieties, such as Sahel
abiotic stresses for which major QTLs have been 108, Sahel 201 and ITA 212, to improve their
identified and successfully used. A certain set of tolerance to salinity. Advanced backcross breed-
QTLs is targeted for each rice production envi- ing lines (BC –BC ) have been obtained with
2 3
ronment and research conducted at different Saltol introgression using FL478 and Pokkali as
AfricaRice stations. donor parents. In the Sahel, where rice is also
Improving drought tolerance of rice has affected by cold stress at the vegetative stage,
been hindered by the low level of genetic variabil- AfricaRice scientists are targeting the cold toler-
ity and the complex inheritance of the trait, as ance QTLs (Ctb-1, Ctb-2) identified in Silewah
well as the difficulty of accurately measuring the (Saito et al., 2004) and qCTS12a identified in
level of tolerance. Molecular approaches to M202 (Andaya and Mackill, 2003b). For sub-
improve drought tolerance have been widely mergence-prone areas, African Sub1 lines were
applied to rice, and numerous QTLs identified for developed using WITA 4, a popular variety in
secondary traits associated with drought Nigeria, as recipient (Gregorio et al., unpub-
response, including rooting traits (depth, vol- lished) and new versions using NERICA-L 19,
ume, thinness, penetration ability), leaf rolling TOX4004, Kogoni, FARO44 and BW348-1 as
and death, membrane stability, and osmotic recipients are in the pipeline. The major QTL
adjustment (Laffite etal., 2006). However, very Sub1, identified from FR13A (Xu and Mackill,
few studies have mapped QTLs related to better 1996), provides tolerance to complete submer-
yield under drought. A few major QTLs for yield gence for up to 2 weeks. Asian Sub1 varieties
Genetic Improvement for Abiotic Stress Tolerance 153
have been extensively evaluated in South and With the use of molecular markers,
South-east Asia and in Africa. In all cases, similar AfricaRice expects to increase the efficiency and
results were obtained (Table 11.2), with consist- effectiveness of its breeding programmes com-
ent yield advantages of 1 t/ha to over 3 t/ha pared to conventional breeding methods.
compared with the original varieties following
submergence for durations of about 4 to over
18days (Mackill etal., 2012). However, this QTL
Challenges and Outlook
is not effective in areas that are likely to experi-
ence prolonged waterlogging or partial stagnant
flooding of over 20 cm (Singh et al., 2011). Anticipating the effects of climate
Therefore for these areas and for deep-water changeon rice production
conditions, new materials should be developed
targeting elongation ability. Climate change is emerging as one of the most
Another powerful marker-assisted approach, important challenges of the 21st century. Africa
marker-assisted recurrent selection (MARS), is is particularly vulnerable to climate change,
being implemented for the improvement of because of its high proportion of low-input,
drought tolerance in the lowlands. Allele diver- rainfed agriculture compared with other regions
sity within bi-parental populations is exploited of the developing world (IPCC, 2001). The
by MABC to increase the frequency of beneficial changes in key climatic variables (i.e. rainfall
alleles for quantitative traits. Cyclical recombi- and temperature) will likely modify the
nation of lines bearing interesting chromosomal distribution, frequency and severity of abiotic
segments will then be conducted and favourable stresses. Prediction models in some cases provide
alleles accumulated. Six populations have been contradictory conclusions, but what is clear is
developed and more than 200 lines are under that increased flooding in low-lying areas,
evaluation for drought tolerance in Nigeria, Mali greater frequency and severity of droughts in
and Burkina Faso (M.-N. Ndjiondjop, Cotonou, arid andsemiarid areas, rising of sea water level
Benin, 2012, personal communication). (which will increase salinity intrusion in
Table 11.2. Performances of rice lines bearing the resistance allele at Sub1 locus after 21 days of
submergence at AfricaRice station, Ibadan (Nigeria), in the 2009 dry season. Survival is expressed as
the number of surviving hills under submergence as a percentage of survival without submergence.
(From Gregorio, unpublished.)
Yield under submerged Yield under normal Yield
Variety Survival (%) conditions (t/ha)a conditions (t/ha) reduction (%)
TDK1-Sub1(BCF) 95.45 3.62a 4.46 18.99
3 3
Swarna-Sub1(BCF) 90.95 3.56a 4.45 19.92
3 3
BR11-Sub1 95.34 2.60b 4.11 36.49
Samba 94.42 2.59b 3.94 34.16
Mahsuri-Sub1(BCF)
3 3
Samba 96.45 2.37b 3.62 34.47
Mahsuri-Sub1(BCF)
2 3
CR1009-Sub1 96.95 2.21b 3.61 29.78
IR64-Sub1(BCF) 96.92 2.21b 3.02 26.89
3 3
IR64 21.43 1.32c 3.19 58.46
Samba Mahsuri 17.44 0.81cd 3.83 78.73
Swarna 19.49 0.64de 4.29 85.16
FARO 35 (ITA 212) 20.41 0.41de 3.61 88.61
WITA 4 4.08 0.08e 3.77 97.99
aNumbers followed by the same letters are not significantly different from each other.