Research

Stay-green alleles individually enhance grain yield in sorghum under drought by modifying canopy development and water uptake patterns Andrew K. Borrell1, Erik J. van Oosterom2, John E. Mullet3, Barbara George-Jaeggli4, David R. Jordan1, Patricia E. Klein5 and Graeme L. Hammer2 1

Queensland Alliance for Agriculture and Food Innovation (QAAFI), Hermitage Research Facility, University of Queensland, Warwick, Qld 4370, Australia; 2QAAFI, University of

Queensland, Brisbane, Qld 4072, Australia; 3Department of Biochemistry and Biophysics, Texas A&M University, College Station, TX 77843, USA; 4Department of Agriculture, Fisheries & Forestry Queensland (DAFFQ), Hermitage Research Facility, Warwick, Qld 4370, Australia; 5Department of Horticultural Sciences, Texas A&M University, College Station, TX 77843, USA

Summary Author for correspondence: Andrew K. Borrell Tel: +61 7 46603640 Email: [email protected] Received: 26 February 2014 Accepted: 19 April 2014

New Phytologist (2014) 203: 817–830 doi: 10.1111/nph.12869

Key words: canopy development, drought adaptation, grain yield, quantitative trait loci (QTLs), sorghum, stay-green, water uptake.


Stay-green is an integrated drought adaptation trait characterized by a distinct green leaf phenotype during grain filling under terminal drought. We used sorghum (Sorghum bicolor), a repository of drought adaptation mechanisms, to elucidate the physiological and genetic mechanisms underpinning stay-green.
Near-isogenic sorghum lines (cv RTx7000) were characterized in a series of field and managed-environment trials (seven experiments and 14 environments) to determine the influence of four individual stay-green (Stg1–4) quantitative trait loci (QTLs) on canopy development, water use and grain yield under post-anthesis drought.
The Stg QTL decreased tillering and the size of upper leaves, which reduced canopy size at anthesis. This reduction in transpirational leaf area conserved soil water before anthesis for use during grain filling. Increased water uptake during grain filling of Stg near-isogenic lines (NILs) relative to RTx7000 resulted in higher post-anthesis biomass production, grain number and yield. Importantly, there was no consistent yield penalty associated with the Stg QTL in the irrigated control.
These results establish a link between the role of the Stg QTL in modifying canopy development and the subsequent impact on crop water use patterns and grain yield under terminal drought.

Introduction Food security in the face of water scarcity is one of the most pressing issues confronting humanity today (UNCTAD, 2011). Global food demand is expected to increase by up to 70% by 2050 (UNWWDR4, 2012) and this requires another 1 billion tonnes of cereals and 200 million tonnes of livestock products to be produced yearly (UNCTAD, 2011). Rain-fed farming is practised on 80% of cultivated land and accounts for 60% of the world’s food production (UNCTAD, 2011). Because water supply in rain-fed cropping systems is highly variable, adaptation strategies for crop plants to dry environments become increasingly important. Understanding the physiological basis of the stay-green drought adaptation trait in sorghum is one such strategy. Sorghum (Sorghum bicolor) is an important global crop grown for food, feed, fibre and fuel, and is particularly well adapted to hot and dry conditions (Paterson et al., 2009). It is a major summer crop in tropical and subtropical rain-fed farming systems in which at least some post-flowering drought is likely to occur. In Ó 2014 The Authors New Phytologist Ó 2014 New Phytologist Trust

such terminal drought environments, in which crops run out of water during grain filling, sorghum hybrids displaying high levels of stay-green, an ability to retain green leaf area during grain filling, generally produce higher grain yield than those with intermediate or low phenotypic expression of the trait (Borrell et al., 1999, 2000a,b; Vadez et al., 2011; Jordan et al., 2012). The staygreen phenotype has also been reported to enhance grain yield under post-anthesis drought stress in other cereals, including wheat (Triticum aestivum) (Christopher et al., 2008; Adu et al., 2011; Bogard et al., 2011; Lopes & Reynolds, 2012), maize (Zea mays) (Crafts-Brandner et al., 1984; Gentinetta et al., 1986; Rajcan & Tollenaar, 1999; Zheng et al., 2009) and rice (Oryza sativa) (Mondal & Choudhuri, 1985; Wada & Wada, 1991; Fu et al., 2009; Hoang & Kobata, 2009). The stay-green drought adaptation trait has been a focus of sorghum crop improvement programmes in Australia and the USA for over 30 yr (Rosenow, 1977; Henzell et al., 1992; Jordan et al., 2012). The physiological basis of stay-green remains unclear, although a number of studies in sorghum have increased our understanding of this complex trait (Borrell et al., 2000a,b; New Phytologist (2014) 203: 817–830 817 www.newphytologist.com

New Phytologist

818 Research

Harris et al., 2007; van Oosterom et al., 2011; Vadez et al., 2011; Jordan et al., 2012; Burke et al., 2013; Choudhary et al., 2013). Earlier research focused on the nitrogen status of plants during terminal water deficit (Borrell & Hammer, 2000; Borrell et al., 2001), but later studies also evaluated the role of plant water management (Hammer, 2006; Vadez et al., 2011; Choudhary et al., 2013). When crop growth is limited by water supply, grain yield is the product of transpiration (T), transpiration efficiency (TE) and harvest index (HI) (Passioura & Angus, 2010). Within this framework, grain yield is a function of post-anthesis T (Turner, 2004), which can be manipulated via water availability at anthesis (Fig. 1). Under post-anthesis water stress, shifting even small amounts of water use from the pre- to post-anthesis period can substantially increase grain yield (Hammer, 2006; Manschadi et al., 2006). A stay-green phenotype will emerge under waterlimited conditions when the balance between the supply and demand of water for crop growth is favourable post-anthesis (Borrell et al., 2009; Jordan et al., 2012). This can be achieved via the modification of root architecture (Mace et al., 2012), canopy development (Borrell et al., 2000a), or both (Fig. 1). Here, we focus on the effects on grain yield of: (1) canopy development on plant size at anthesis, and hence crop water use before anthesis; and (2) water extraction during grain filling. Several sorghum genotypes exhibit the stay-green trait, including BTx642 (formerly B35), SC56 and E36-1 (Rosenow et al., 1983; Kebede et al., 2001; Haussmann et al., 2002). A number of researchers have mapped the quantitative trait loci (QTLs) that contribute to this trait in a range of populations, many of which were derived from crosses with BTx642, a derivative of an Ethiopian durra landrace (Tuinstra et al., 1996, 1997, 1998; Crasta et al., 1999; Boffa et al., 2000; Subudhi et al., 2000; Tao et al., 2000; Xu et al., 2000). In a recombinant inbred population derived from a cross between BTx642 and Tx7000, four major

QTLs were identified: Stg1, located on SBI-03, Stg2 (SBI-03), Stg3 (SBI-02) and Stg4 (SBI-05). These QTLs were found to explain c. 20%, 30%, 16% and 10%, respectively, of the phenotypic variability for green leaf area retention during grain filling under post-anthesis drought stress (Xu et al., 2000; Sanchez et al., 2002). Stay-green QTLs, including Stg1-4 derived from BTx642, have been found to affect water extraction, TE, green leaf area retention and grain yield in two genetic backgrounds (Vadez et al., 2011), but the near-isogenic lines (NILs) in these studies contained unknown numbers of Stg QTLs. The aim of the current research was to determine the function of Stg1–4 by evaluating their individual effects on components of canopy development (phenology, tillering, leaf size, leaf appearance rate (LAR)), crop water use (pre- and post-anthesis) and grain yield under a range of management 9 environment conditions. We show that the stay-green phenotype, to some extent, is an emergent consequence of canopy development processes. Another aim of this paper is to provide a physiological understanding that will ultimately assist in discovering the genes underpinning the causal mechanisms driving stay-green.

Materials and Methods Generation of RTx7000 NILs containing BTx642 DNA from the stay-green loci The NILs of sorghum (Sorghum bicolor (L.) Moench) genotype RTx7000 that contained one or more of the Stg loci from BTx642 were constructed from a cross of BTx642 and RTx7000, followed by backcrossing F1 plants to RTx7000 either four (BC4, 6000 NIL series) or six (BC6, 2000 NIL series) times (Harris et al., 2007). Four RTx7000 NILs, each containing only Stg1, Stg2, Stg3 or Stg4, were used in our experiments to determine the

Fig. 1 Flow chart of crop physiological processes that determine plant size and crop water use of sorghum (Sorghum bicolor) at anthesis, with flow-on consequences for water uptake during grain filling and grain yield. The effect of individual stay-green (Stg) quantitative trait loci (QTLs) on each process is indicated by arrows contained in either dotted squares (input traits; shaded grey) or dotted circles (derived traits). Upward arrow indicates increased size or number, downward arrow reduced size or number, and sideways arrow indicates no or little effect. The number of arrows represents the magnitude of the effect. The direction and number of arrows associated with each Stg QTL summarize the trait data from up to seven experiments and 14 environments relating to canopy development (Table 2), grain yield (Table 3) and crop water use (Fig. 7 and the Results section). New Phytologist (2014) 203: 817–830 www.newphytologist.com

Ó 2014 The Authors New Phytologist Ó 2014 New Phytologist Trust

New Phytologist impact of BTx642 alleles in each locus on canopy development, water uptake and grain yield: 6078-1 (Stg1) and 6085-9 (Stg4) NILs from the 6000 series, and 2219-3 (Stg2) and 2290-19 (Stg3) from the 2000 series. Field experiments Three field experiments were conducted across two locations in Australia’s north-eastern grain belt: Biloela (BIL: 24°240 S, 150°300 E; elevation, 175 m) in central Queensland, and Warwick (WAR: 28°120 S, 152°060 E; elevation, 480 m) in south-eastern Queensland (Table 1). Experiment names are designated by a combination of location (BIL, WAR), year (e.g. 05 for 2005) and how the experiments were conducted (FLD for field). Field experiments were used to assess the effects of Stg QTL on phenology, canopy development, water uptake and grain yield under field conditions. The first experiment was sown at BIL on 18 February 2002 (BIL02FLD) on a soil with a dark sandy clay loam A horizon over a brown silty clay B horizon (Uf6.42, Gn3.52, Gn3.92) (Northcote, 1979). The experiment included two irrigation treatments and 36 genotypes (RTx7000, BTx642 and 34 NILs), but only data on the two parents and four Stg NILs are considered here. The experimental design was a split plot with three replicates and irrigation treatments as main plots and genotypes as subplots. Main plots were 54 m 9 19 m, and subplots consisted of three rows of 9 m length with a row spacing of 0.9 m. Data were collected only from the centre row of each plot. The water regimes included well watered (high water, HW) and post-flowering deficit (low water, LW). The experimental block was furrow irrigated three times before sowing to fully wet the soil profile and all plots were irrigated three more times before flowering. In addition, the HW treatment was irrigated five more times. All plots were fertilized before sowing with 221 kg N ha1 applied as urea, and additional micro-nutrients as required for optimal growth. Insects and weeds were controlled as required (i.e. to maintain optimal crop growth). Two experiments were conducted at WAR (Table 1) on a cracking and weakly self-mulching brownish-black clay (Talgai shallow phase, Ug 5.14) (McKeown, 1978; Northcote, 1979). The first experiment (WAR04FLD) was sown on 11 December 2003 (HW and LW) and the second (WAR05FLD) on 7 January 2005 (HW) and 21 January 2005 (LW). The experiments were conducted under non-limiting nutrient conditions and were planted on full profiles of subsoil moisture. As the LW treatment was dependent on rainfall exclusion via rain-out shelters, water treatments could not be randomly allocated within replicates and the HW treatment was a separate block adjacent to the rain-out shelter (LW treatment). Each treatment block was split into two plant densities: high density (HD, 20 plants m2) and low density (LD, 10 plants m2). Density treatments were randomly allocated within each of four replicates, and genotypes (RTx7000, Stg1-3 NILs in WAR04FLD and Stg1-4 NILs in WAR05FLD) were randomly allocated within each density. Hence, four treatments with increasing levels of water deficit were created, ranging from HWLD (least stressed) to HWHD, LWLD and LWHD (most stressed). Additional environmental conditions (mean Ó 2014 The Authors New Phytologist Ó 2014 New Phytologist Trust

Research 819

maximum and minimum temperatures, daily radiation and relative humidity) are presented in Table 1. Plots at WAR consisted of four rows (with approximately east to west orientation) of 3 m length with 0.5 m row spacing. They were oversown and thinned to the appropriate density c. 2 wk after emergence. To eliminate border effects, only the middle two rows of each plot were used for measurements, and three guard rows were planted around the perimeter of each treatment block. All experiments were irrigated by overhead sprinklers, with irrigation in the LW treatment ceasing at 49 and 27 d before anthesis in WAR04FLD and WAR05FLD, respectively. All experiments were fertilized at or just before sowing with 240 kg N ha1 applied as urea, and additional micro-nutrients as required for optimal growth. Weeds and insects were controlled with appropriate herbicides and insecticides. Plants in all experiments were grown until maturity. Semi-controlled environment experiments Experiments in semi-controlled environments were conducted at WAR in ventilated, plastic-covered growth tunnels, orientated north to south. The front and sides of the tunnels were covered with white knitted shade cloth to allow air flow, and the top was covered with white solar weave (Gale Pacific Pty Ltd, Melbourne, Vic., Australia). The solar weave excluded rainfall and transmitted c. 70% of the incident solar radiation. Experiments were conducted by growing individual plants in lysimeters (LYS) or small pots (POT) (Table 1). LYS experiments provided a lower plant density for phenotyping than field experiments, whereas conditions in the tunnel provided an environment for phenotyping of mapping populations in POT experiments with less error variance than FLD experiments. LYS and POT experiments thus complemented the FLD experiments in assessing the roles of the various Stg QTLs. The lysimeters were made from cylindrical polyvinyl chloride (PVC) tubes of 300 mm in diameter and 750 mm in height. This volume was sufficiently large to minimize effects on plant growth (Yang et al., 2010). They were filled with a 3 : 1 : 1 mix of alluvial clay soil, loam and feedlot manure; 30 g of Osmocote Plus (16% N, 3.5% P, 10% K plus trace elements; Scotts Pty Ltd, Baulkham Hills, NSW, Australia) were added to each lysimeter at the time of filling. Two LYS experiments were conducted: WAR06LYS was sown on 25 February 2006 and WAR07LYS on 22 February 2007. Both experiments were laid out as a randomized complete block design with either 10 genotypes and seven replications (WAR06LYS) or 14 genotypes and four replications (WAR07LYS). However, only data on the four Stg NILs (6078-1, 2219-3, 2290-19, 6085-9) and the senescent recurrent parent (RTx7000) are used here. Ten seeds were sown per lysimeter and emerged seedlings were thinned to one plant 8–10 d after emergence. Plants were well watered and harvested soon after anthesis. Plant water use was measured weekly in lysimeters. Details on the growing conditions are presented in Table 1. Pot experiments were conducted during the summers of 2007/ 2008 (WAR08POT) and 2008/2009 (WAR09POT) (Table 1). WAR08POT was conducted in 7-l planter bags (Garden City New Phytologist (2014) 203: 817–830 www.newphytologist.com

New Phytologist (2014) 203: 817–830 www.newphytologist.com

11 December 2003

WAR04FLD

22 February 2007

27 December 2007

31 October 2008

WAR07LYS

WAR08POT

WAR09POT

HWLD

HWLD

HWLD

HWLD

HWLD, HWHD, LWLD, LWHD HWLD, HWHD, LWLD, LWHD

HWLD, LWLD

Treatments

RTx7000, Stg1

RTx7000, Stg1

RTx7000, Stg1, Stg2, Stg3, Stg4

RTx7000, Stg1, Stg2, Stg3, Stg4

RTx7000, RTx642, Stg1, Stg2, Stg3, Stg4 RTx7000, Stg1, Stg2, Stg3 RTx7000, Stg1, Stg2, Stg3, Stg4

Genotypes

15.6

17.7

13.3

10.2

14.1 (HW) 12.2 (LW)

16.8 (LW)

11.6

32.5

33.2

36.5

27.8

27.0 (HW) 25.7 (LW)

28.5 (LW)

27.1

Average daily maximum temperature (°C)

15.6

13.1

13.9

11.6

19.3 (HW) 18.3 (LW)

21.8 (LW)

20.9

Average daily radiation (MJ m2)

70.4

79.8

61.0

62.7

67.6 (HW) 68.6 (LW)

70.4 (LW)

77.5

Average daily relative humidity (%)

N/A

N/A

N/A

N/A

551 (HW) 411 (LW)

260 (LW)

390 (HW) 360 (LW)

Average grain yield (g m2)

GY, gsize, gnumber (LW) GLAA, FLN, LAR, TTFLA, CNPP, LA2–9, LA10+, LAmax, GY GLAA, FLN, LAR, TTFLA, CNPP, LA10+, LAmax GLAA, FLN, LAR, TTFLA, CNPP, LA10+, LAmax FLN, LAR, TT_FLA, CNPP GLAA, FLN, LAR, TT_FLA, CNPP

GLAA, CNPP

Traits

List of experiments (seven experiments and 14 managed environments), including sowing date, treatments, genotypes, meteorological data, grain yield and traits measured. BIL, Biloela; CNPP, culm number per plant; FLD, field; FLN, main shoot final leaf number; GLAA, green leaf area index at anthesis; gnumber, number of grains; gsize, grain size (measured as mass); GY, grain yield; HD, high density; HW, high water; LA2–9, cumulative area of main shoot leaves 2–9; LA10+, cumulative area of main shoot leaves 10 and up; LAmax, area of the largest main shoot leaf; LAR, leaf appearance rate; LD, low density; LW, low water; LYS, lysimeter; POT, pot; TTFLA, thermal time to flag leaf appearance; WAR, Warwick.

25 February 2006

WAR06LYS

7 January 2005 (HW) 21 January 2005 (LW)

18 February 2002

BIL02FLD

WAR05FLD

Sowing date

Experiment code

Average daily minimum temperature (°C)

Table 1 Details of seven experiments (14 managed environments) undertaken to determine the physiological and genetic basis of the stay-green (Stg) trait in sorghum (Sorghum bicolor)

820 Research

New Phytologist

Ó 2014 The Authors New Phytologist Ó 2014 New Phytologist Trust

New Phytologist Plastics Pty Ltd, Woodridge, Qld, Australia) filled with pure alluvial clay. For support, each bag was set up within a larger pot that was filled with sand. All pots were placed on a capillary watering mat (TopUpTM mat, Anova Solutions Pty Ltd, Chapel Hill, Qld, Australia) to keep the soil in the pots moist at all times. Plants were well watered at all times and 2.5 g l1 of Miracle-Gro® Max Feed (24% N, 6% P, 12% K, 4% S plus trace elements) soluble plant food (Scotts Pty Ltd) were applied weekly until plants were harvested when 11 leaves had fully expanded. A similar process was followed for WAR09POT, except that larger 19-l pots with a diameter of 330 mm were used and plants were harvested at anthesis. Five grams of Starter Z (10.5% N, 19.5% P, 2.2% S, 2.5% Zn) were added to the larger pots before planting. Each POT experiment included RTx7000, the Stg1 NIL and progeny from their cross. Only results of the parental lines are discussed here. Experiments were laid out as a randomized block design with either four (WAR08POT) or eight (WAR09POT) replications. For all pot experiments, 5–10 seeds were sown per pot, and seedlings were thinned to one plant per pot about 1 wk after emergence. Leaf number, leaf size and phenology The number of fully expanded leaves (ligule visible above that of the previous leaf) was counted weekly for all axes on four adjacent, well-bordered plants in FLD experiments and on all plants in LYS and POT experiments. LAR was calculated as the slope of the linear regression of fully expanded leaf number on cumulative thermal time. Upper leaves were excluded from LAR calculations, as the final few leaves appear nearly synchronously (Carberry et al., 1993; Borrell et al., 2000a), possibly as a consequence of the combined effects of rapid stem elongation and short length of the leaf blade. Daily thermal time was calculated from threehourly interpolations from daily minimum and maximum temperatures, using 11, 30 and 42°C as the cardinal temperatures (Hammer et al., 1993) and linear interpolations between these points. The areas of all fully expanded leaves on the main shoots and tillers were measured in WAR05FLD on two tagged plants per plot at three harvest times, corresponding to the expansion of the sixth, 12th and flag leaves. Emergence was defined as the date on which 50% of the plants in each experiment had emerged from the soil, and anthesis when, on average, 50% of the anthers had extruded from the main shoot of four tagged plants per plot in FLD studies or from each plant in LYS and POT studies. Physiological maturity was defined as the date on which basal grains in 90% of the same tagged plants in FLD studies attained a black layer on the placental area of the grain (Eastin et al., 1973), which represents the cessation of assimilate import into the grain. Tillering In FLD experiments, total and fertile culm numbers per plant were recorded based on all plants in one of the central rows of each plot, excluding plants near the end of the rows. In the LYS and POT experiments, culm number was measured for each Ó 2014 The Authors New Phytologist Ó 2014 New Phytologist Trust

Research 821

plant. The culm number per plant is defined as the number of stems per plant, including the mainstem, that is culm number = tiller number + 1. The rank of each emerged primary tiller was defined by the position of the main shoot axillary bud from which it originated, for example, Tiller 3 (T3) emerged from the sheath of main shoot Leaf 3. Biomass sampling Biomass samples were taken at anthesis and maturity in the two field experiments at WAR to determine biomass production during grain filling (Table 1). A single row of 1 m length was cut at ground level from one of the two centre rows at anthesis and two rows of 1 m length were cut from the two central rows at maturity. At least 0.5-m intervals of crop were left between sampling areas within a row and no adjacent areas were sampled. The green leaf area of the whole sample was obtained with a planimeter (Delta-T DIAS image analysis system, Cambridge, UK). Samples were dried at 80°C for at least 2 d before obtaining the dry mass. After drying, panicles were threshed. Grain yield and grain mass (100 grain weight) were measured and grain number was derived. Crop water use In WAR05FLD, soil water content was measured by neutron moderation (Model 503 DR Hydroprobe, CPN International Inc., Martinez, CA, USA). One access tube per plot was sited at the middle inter-row position. Readings were taken weekly at 20cm depth intervals to a depth of 1.8 m throughout crop growth. A calibration equation to calculate soil water content from neutron moisture meter counts was determined from field measurements taken previously at the same site during crop growth at various stages of soil water depletion. Soil water content in the 0–20-cm layer was determined gravimetrically by taking a core within a radius of 1 m of the access tube. For each time of measurement, total soil water content was determined by summing the soil water contents for each depth interval. Finally, temporal patterns of cumulative soil water content were developed for each plot. Data analyses For the WAR04FLD and WAR05FLD experiments, a factor experiment with eight levels was defined by combining location, year, water treatment and density treatment. In all experiments, a linear mixed model was fitted for each trait using the ASReml program (Butler et al., 2009) with the R software package (R Development Core Team, 2012). The fitted model contained fixed effects for experiment, genotype and their interaction. Random effects included fitting separate effects for replicates and residuals for each experiment. Predicted values for genotype 9 experiment (G 9 E) were calculated using the R function predict.asreml. Significance levels of the fixed effects were determined using a Wald chi-squared test. Associations among traits were analysed using the Excel procedure for linear regressions. New Phytologist (2014) 203: 817–830 www.newphytologist.com

New Phytologist

822 Research

Results Stg QTLs reduced green leaf area at anthesis (GLAA) and slightly increased earliness Stg QTLs reduced (P < 0.001) GLAA relative to RTx7000 across eight managed environments varying in water supply (Table 2). Averaged across environments, Stg4 showed the largest decline in GLAA relative to RTx7000 (21%). The G 9 E interaction for GLAA was small but significant (P < 0.05), indicating that the lower GLAA of individual Stg lines relative to RTx7000 was not consistent across environments. The genotype main effect on days to anthesis (DTA) was highly significant (P < 0.001). On average across four environments, DTA of Stg1, Stg3 and Stg4 was significantly shorter than for RTx7000 (Table 2), although, in absolute terms, the difference was small (c. 1 d). The genotype main effect on thermal time to flag leaf appearance (TTFLA) was highly significant (P < 0.001). On average across six environments, TTFLA of Stg3 and Stg4 was significantly shorter than for RTx7000 (Table 2), although the difference amounted to less than 2 d. For Stg1 and Stg2, the difference in TTFLA from RTx7000 was < 10°Cd (or < 1 d) and was not significant (Table 2). Decreased tillering of Stg QTLs was associated with larger leaf size The lower GLAA of Stg NILs compared with RTx7000 was associated with a reduction in tillering (Fig. 2, R2 = 0.46, n = 5,

P > 0.05). Across eight managed environments, each of the four Stg QTLs produced significantly (P < 0.01) fewer tillers than RTx7000 (Table 2). Stg2 reduced tillering the most and Stg4 the least (Fig. 2). The expression of tillering varied among experiments, resulting in a significant (P < 0.001) environment main effect. This was largely a result of low tillering under high plant density (Table 2). Stg loci also reduced (P < 0.001) tillering under both LW and HW conditions in BIL02FLD. Low tillering of Stg isolines was associated with a larger size of leaves 2–9 relative to RTx7000 (Fig. 3). Across multiple experiments, culm number was significantly negatively correlated with the total area of leaves 2–9 (R2 = 0.97, n = 5, P < 0.01) and Stg2 significantly (P < 0.05) increased the total area of leaves 2–9 relative to RTx7000 (Table 2). There was also a non-significant trend for larger lower leaves in the remaining Stg QTLs (Fig. 2). For both culm number and total area of leaves 2–9, the difference from RTx7000 was lowest for Stg4 and greatest for Stg2. However, there was no association with LAR, which did not differ significantly among genotypes (Table 2). Stg QTLs decreased total leaf number and maximum leaf size Despite the positive association between culm number and GLAA (Fig. 2), the results of Stg4 suggested that its low GLAA was determined by traits other than tillering (Fig. 2). The earlier phenology of Stg3 and Stg4 compared with RTx7000, Stg1 and Stg2 was associated with a slightly, but significantly, lower final leaf number (FLN) (Table 2). As time to flag leaf appearance is

Table 2 Components of green leaf area at anthesis for sorghum (Sorghum bicolor) inbred line RTx7000 and four stay-green (Stg) near-isogenic lines grown in field and lysimeter studies Effect Experiment (exp) BIL02FLD_HWLD BIL02FLD_LWLD WAR05FLD_HWHD WAR05FLD_HWLD WAR05FLD_LWHD WAR05FLD_LWLD WAR06LYS WAR07LYS Genotype (geno) RTx7000 Stg1 Stg2 Stg3 Stg4 exp geno exp : geno LSD

GLAA (m2 m2)

FLN

LAR

TTFLA

CNPP

LA2–9 (cm2)

LA10+ (cm2)

317 309 325 302

2301 2608 2534 2723 1464 2937

395 448 474 491 274 475

62.8 63.8 61.5 63.1

300 317 326 314 310 ns ns * 20.7

2554 2488 2395 2433 2270 *** *** *** 92.6

447 431 422 426 406 *** *** *** 18.4

63.4 62.4 63.7 62.3 62.1 *** *** ns 0.7

2.58 2.02 4.02 4.00 3.93 4.20 0.50 1.10

17.1 18.0 15.9 16.4 15.7 18.4

0.0310 0.0317 0.0248 0.0259 0.0303 0.0389

557 571 642 634 519 472

1.87 1.88 1.14 2.12 1.18 2.18 4.80 2.40

3.18 2.78 2.70 2.81 2.51 *** *** * 0.31

17.2 17.0 17.0 16.7 16.5 *** *** ** 0.3

0.0304 0.0307 0.0303 0.0305 0.0303 *** ns ns 0.0008

576 567 574 558 556 *** *** ** 12.5

2.60 2.09 1.88 2.18 2.24 *** ** *** 0.34

LAmax (cm2)

DTA

Analysis of variance for green leaf area index at anthesis (GLAA), main shoot final leaf number (FLN), leaf appearance rate (LAR), thermal time to flag leaf appearance (TTFLA), culm number per plant (CNPP), cumulative area of main shoot leaves 2–9 (LA2–9), cumulative area of main shoot leaves 10 and up (LA10+), area of the largest main shoot leaf (LAmax) and days to anthesis (DTA). BIL, Biloela; FLD, field; HD, high density; HW, high water; LD, low density; LW, low water; LYS, lysimeter; WAR, Warwick. Significant differences observed between treatments: ***, P < 0.001; **, P < 0.01; *, P < 0.05; ns, no significant difference. Note: Using the mixed model method with a Wald test for fixed effects results in a pairwise standard error of difference (SED) comparison for each pair of effects within the genotype by treatment interaction. We have presented an average least significant difference (LSD) value based on the average SED value. New Phytologist (2014) 203: 817–830 www.newphytologist.com

Ó 2014 The Authors New Phytologist Ó 2014 New Phytologist Trust

New Phytologist

Research 823 (a)

(b) Fig. 2 Green leaf area per square metre at anthesis as a function of culm number per plant for sorghum (Sorghum bicolor) inbred line RTx7000 (open square) and four stay-green near-isogenic lines (closed squares) for data averaged across four experiments and eight managed environments (Tables 1 and 2).

(c)

Fig. 3 Culm number per plant as a function of cumulative area per plant of leaves 2–9 for sorghum (Sorghum bicolor) inbred line RTx7000 (open square) and four stay-green near-isogenic lines (closed squares) for data averaged across one experiment with four managed environments (Tables 1 and 2).

determined by FLN and LAR, this result was consistent with the observation that Stg QTLs had little effect on LAR (Table 2). FLN, in turn, was positively correlated with the area of the largest leaf (LAmax) (Fig. 4a, R2 = 0.73, n = 5, P < 0.1). However, only for Stg4 did the low FLN translate into low LAmax, as LAmax of Stg3 was similar to that of Stg1 and Stg2, despite its lower FLN. Moreover, Stg2 exhibited a lower maximum leaf size than Stg1, despite equivalent FLN. The combined effect of smaller and fewer leaves of Stg NILs relative to RTx7000 resulted in a significant (P < 0.001) decline in the total area of the upper leaves (L10 onwards, LA10+) (Table 2). LA10+ was highly correlated with LAmax (Fig. 4b, R2 = 0.97, n = 5, P < 0.01). On average, relative to RTx7000, Stg4 had the greatest reduction in LA10+, whereas the reduction in Stg1 was the lowest and not significant. In a combined analysis of six managed environments, LA10+ was positively correlated (r2 = 0.85, n = 5, P < 0.05) with GLAA (Fig. 4c). The low GLAA of Stg4 (Table 2) was hence predominantly associated with its low leaf number and small leaf size (Fig. 4).

Ó 2014 The Authors New Phytologist Ó 2014 New Phytologist Trust

Fig. 4 Leaf number per culm and area per leaf are determinants of green leaf area at anthesis (GLAA) of sorghum (Sorghum bicolor). (a) Final main shoot leaf number per plant vs area of the largest leaf (LAmax), (b) area of the largest leaf vs cumulative size of leaves 10 and up (LA10+), and (c) cumulative size of leaves 10 and up vs GLAA for RTx7000 (open squares) and four stay-green near-isogenic lines (closed squares) for data averaged across three experiments and six managed environments (Tables 1 and 2).

The effect of Stg QTLs on the size of upper leaves in the main shoot was reproduced for tillers. In WAR06LYS, Stg1 reduced the size of the largest leaves in tillers T3 and T4 in a manner similar to the main shoot (Fig. 5). For the main shoot, Stg1 reduced the leaf size of the fourth leaf below the flag leaf, whereas this happened for the fifth and sixth leaves below the flag leaf for T3 and T4, respectively. In addition, the decline in maximum leaf size was greater in T3 and T4 than in the main shoot, in both absolute and relative terms. Hence, the effects of Stg1 on the size of upper leaves seemed to be more pronounced in tillers than in the main shoot.

New Phytologist (2014) 203: 817–830 www.newphytologist.com

824 Research (a)

(b)

(c)

New Phytologist from 24% (Stg3) to 31% and 32% (Stg1 and Stg2, respectively), and, in WAR05FLD, from 12–17% for Stg2, Stg3 and Stg4 to 36% for Stg1. Neither the density main effect nor any of the interactions were significant for grain yield in either of the experiments. In order to determine whether the increased grain yield of Stg QTLs under post-anthesis drought stress could be linked to preanthesis processes that reduce plant size, an association needs to exist between plant size at anthesis and grain yield that could capture the potential benefits of increased water use during grain filling on grain yield (Fig. 1). Plant size at anthesis can be represented by GLAA, whereas, under drought, when biomass accumulation is a function of water use, biomass increase during grain filling (bioGFP) represents post-anthesis water availability. Data for each genotype in each year were averaged across replications and densities, as the density effect on grain yield under drought was not significant (Table 3). In both years, the association between GLAA and bioGFP was significantly negative (R2 = 0.94, n = 4, P < 0.05 for 2004; R2 = 0.83, n = 5, P < 0.05 for 2005; Fig. 6a), indicating that a larger plant size at anthesis (greater GLAA) resulted in a smaller increase in biomass during grain filling. Biomass accumulation during grain filling was positively associated with grain number (Fig. 6b, R2 = 0.74, n = 9, P < 0.01), which, in turn, was highly positively associated with grain yield (Fig. 6c, R2 = 0.97, n = 9, P < 0.001). As these parameters were all predominantly determined by post-anthesis conditions, both relationships had a common regression across experiments. There was no consistent yield penalty associated with the Stg QTLs in the irrigated control of WAR05FLD. Although Stg2 and Stg3 yielded significantly less grain than RTx7000 in the HWLD treatment, this result was reversed in the HWHD treatment (Table 3). Averaged across the two densities, grain yields of Stg QTLs were close to those of RTx7000, with the exception of Stg1, which yielded significantly more grain under both densities under well-watered conditions and, on average, had a yield advantage of 22% under well-watered conditions. As was the case for drought stressed treatments, differences in grain yield were predominantly related to grain number rather than to individual grain mass. Stg QTLs affected crop water use before and after flowering

Fig. 5 Stay-green (Stg) quantitative trait loci affect leaf size distributions of sorghum (Sorghum bicolor) main shoot and tillers. Size of successive individual leaves on main shoot (a), tiller 3 (b) and tiller 4 (c) of plants of RTx7000 (solid line) and Stg1 (dotted line) grown in lysimeters at Warwick (WAR06LYS). Data are averaged across four replications. Error bars,  2SE of the mean.

Connecting canopy size at flowering to grain yield Under drought stress in WAR04FLD and WAR05FLD, individual Stg NILs consistently yielded more grain than RTx7000 and, in all but three of the 14 treatment 9 Stg QTL combinations, this difference was significant (Table 3). Across both densities, the average yield benefit under drought stress in WAR04FLD ranged New Phytologist (2014) 203: 817–830 www.newphytologist.com

Stg loci affected the pattern of cumulative crop water use under post-anthesis drought. Canopy size at anthesis was significantly (P < 0.05) less in all Stg NILs compared with RTx7000 in WAR06LYS under low vapour pressure deficit (VPD), and plant water use at anthesis was significantly (P < 0.05) less in Stg2, Stg3 and Stg4. Similar results were attained under field conditions with neutron moderation studies (LWHD treatment) in WAR05FLD, where Stg1 used significantly (P < 0.05) less water than RTx7000 before anthesis (139 mm vs 172 mm), but significantly (P < 0.05) more water after anthesis (197 mm vs 145 mm). This indicates that increased water use during grain filling was associated with both pre-anthesis water savings of 33 mm and increased accessibility of water, as the total water use of Stg1 was 19 mm greater than for RTx7000 (Fig. 7a). Relative to RTx7000, Stg2 and Stg3 also increased (P < 0.05) water use during grain Ó 2014 The Authors New Phytologist Ó 2014 New Phytologist Trust

New Phytologist

Research 825

Table 3 Grain yield, mass and number for sorghum (Sorghum bicolor) inbred line RTx7000 and four stay-green (Stg) near-isogenic lines grown under six environments WAR04FLD

WAR05FLD

Effects

den

gen

den : gen

wat

den

gen

wat : den

wat : gen

den : gen

wat : den : gen

Grain yield Grain number Grain mass

ns ns ***

** ns ns

ns ns ns

*** ns ***

ns ns ns

** ns *

ns ns *

ns ns *

ns ns ns

ns ns *

Predicted means

Genotype

Trait

Experiment code

RTx7000

Stg1

Stg2

Stg3

Stg4

LSD

Grain yield (g m–2)

WAR04FLD_LWHD WAR04FLD_LWLD WAR05FLD_LWHD WAR05FLD_LWLD WAR05FLD_HWHD WAR05FLD_HWLD WAR04FLD_LWHD WAR04FLD_LWLD WAR05FLD_LWHD WAR05FLD_LWLD WAR05FLD_HWHD WAR05FLD_HWLD WAR04FLD_LWHD WAR04FLD_LWLD WAR05FLD_LWHD WAR05FLD_LWLD WAR05FLD_HWHD WAR05FLD_HWLD

184 214 352 358 459 577 12 591 12 790 24 585 22 658 20 025 26 280 15.4 18.4 17.0 18.1 22.8 22.0

290 283 473 493 581 681 14 317 17 171 25 146 28 149 28 431 29 466 20.0 19.1 20.7 18.0 20.4 23.3

278 305 434 367 557 477 21 193 14 149 25 808 22 634 28 989 24 695 15.2 19.5 19.0 18.5 19.6 19.4

252 274 421 375 556 502 17 899 12 566 25 959 24 478 28 204 22 396 15.9 21.2 17.2 16.9 19.8 22.3

na na

49 49 62 62 62 62 4530 4530 3294 3294 3294 3294 2.2 2.2 2.5 2.5 2.5 2.5

Grain number (per m2)

Grain mass (mg)

409 423 542 574 na na 25 307 27 087 26 425 26 393 na na 18.0 17.2 20.5 21.8

Analysis of variance for grain yield, grain number and individual grain mass across two experiments and six environments. Significant differences observed between treatments: ***, P < 0.001; **, P < 0.01; *, P < 0.05; ns, no significant difference; na, not applicable. FLD, field; HD, high density; HW, high water; LD, low density; LW, low water; WAR, Warwick; den, density treatment; gen, genotype; wat, water treatment. Note: least significant difference (LSD) values are based on the average standard error of difference (SED) values from the mixed model method with a Wald test for fixed effects, either for main effects or interactions depending on which was significant. Trait data that are significantly higher (P < 0.05) than RTx7000 are given in bold text; trait data that are significantly lower (P < 0.05) than RTx7000 are given in italics; trait data that are not significantly different from RTx7000 are given in standard font.

filling by 26 and 47 mm, respectively, in the LWHD treatment. Crop water use during grain filling was positively correlated with grain yield in both the LWHD and LWLD treatments in WAR05FLD (Fig. 7b, R2 = 0.74, n = 4, P < 0.05), with the Stg1 NIL using more (P < 0.05) water and producing more (P < 0.05) grain than RTx7000 under both densities.

Discussion The novelty of this paper is that, for the first time, individual Stg QTLs have been shown to confer drought adaptation in a cereal by restricting pre-anthesis plant size, and thereby conserving water for post-anthesis grain growth. Understanding the impact of individual Stg QTLs on crop growth and development under drought provides an insight into the processes regulated by each of these QTLs. This is essential in understanding the G 9 E interactions that have been observed for the expression of staygreen, as it can provide insights into the merits of pyramiding multiple QTLs for specific target environments. This could make introgression of Stg QTLs in crop improvement programmes more targeted and hence more efficient. Ó 2014 The Authors New Phytologist Ó 2014 New Phytologist Trust

Reduced tillering in stay-green NILs caused by larger leaves in the tillering phase Reduced tillering was a major component of the decline in canopy size at flowering exhibited by Stg NILs, with Stg2 decreasing tillering the most and Stg4 the least. Genotypic differences in tillering have been related to the internal carbon supply–demand balance of the plant, because differences in the assimilate demand by the main shoot can affect assimilate availability for tiller production (Bos & Neuteboom, 1998; Lafarge et al., 2002). Increased early vigour of the main shoot (Fig. 1), and hence increased carbon demand, can occur either through increased LAR or increased leaf size of lower leaves that expand during tillering (Kim et al., 2010a,b; van Oosterom et al., 2011). Although previous studies found LAR to be higher in genotypes containing the BTx642 source of stay-green compared with their senescent counterparts (Borrell et al., 2000a; van Oosterom et al., 2011), it did not differ significantly among NILs and RTx7000 in the current study, suggesting that the effects of Stg QTLs on the internal carbon balance of the plant operated predominantly through leaf size. Consistent with this, the strong negative New Phytologist (2014) 203: 817–830 www.newphytologist.com

New Phytologist

826 Research (a)

(a)

(b)

(b)

(c)

Fig. 7 Stay-green 1 (Stg1) quantitative trait locus affects crop water use of sorghum (Sorghum bicolor). (a) Temporal pattern of cumulative water use for RTx7000 (open squares) and Stg1 near-isogenic line (NIL) (closed squares) grown under low water, high density (LWHD) treatment at WAR05FLD. The arrow indicates anthesis at 64 d after emergence (DAE). Error bars,  2SE of the mean. (b) Grain yield as a function of crop water use during grain filling for RTx7000 (open squares) and Stg1 NIL (closed squares) grown under LWHD and low water, low density (LWLD) treatments at WAR05FLD. Data are averaged across four replicates in Fig. 7(a,b). Fig. 6 Linking canopy size of sorghum (Sorghum bicolor) at anthesis with grain yield under drought. (a) Green leaf area at anthesis vs change in biomass during grain filling (bioGFP), (b) bioGFP vs grain number, and (c) grain number vs grain yield for RTx7000 (open squares) and four staygreen near-isogenic lines (closed squares) grown in the field under postanthesis drought stress in two seasons (WAR04FLD and WAR05FLD). Data are averaged across four replications and two plant densities.

further supported by recent evidence that sugars function as important regulators of plant development, and that limiting sugar availability to axillary buds inhibits bud outgrowth (Mason et al., 2014).

correlation between the combined area of leaves 2–9 and culm number suggests that introgression of Stg QTLs into RTx7000 reduced culm number through an increase in the size of lower leaves, with Stg2 exhibiting the largest effect and Stg4 the smallest (Fig. 3). This result concurs with the negative association between leaf size and culm number that has been reported for pearl millet (van Oosterom et al., 2001), rice (Tivet et al., 2001), sorghum (Kim et al., 2010a; Alam et al., 2014) and perennial ryegrass (Lolium perenne) (Bahmani et al., 2000). Current results thus support the hypothesis of a causal link between larger lower leaves and decreased tillering through the carbohydrate supply–demand balance of the plant modulated by Stg QTLs. Our hypothesis is

Individual Stg QTLs can also constrain GLAA by limiting the cumulative size of the upper leaves (LA10+, Fig 4c). Evidence from six environments (Table 2) showed that Stg4 constrained the cumulative size of LA10+ the most and Stg1 the least. The size of LA10+ is determined by LAmax on the main shoot, which, in turn, depends on FLN (Fig. 1). As lower FLN accelerates phenology (Carberry et al., 1993), it is possible that the lower FLN and earlier anthesis of Stg3 and Stg4 compared with RTx7000 (Table 2) was caused by maturity genes located in these QTLs, rather than by a direct effect of a ‘stay-green’ gene per se. The reduced LAmax of Stg3 and Stg4 could be a consequence of lower FLN (Fig. 4a), as observations for pearl millet (van Oosterom

New Phytologist (2014) 203: 817–830 www.newphytologist.com

Stay-green QTLs reduce upper leaf area

Ó 2014 The Authors New Phytologist Ó 2014 New Phytologist Trust

New Phytologist et al., 2001), sorghum (Carberry et al., 1993) and maize (Keating & Wafula, 1992) have shown that the two are physiologically linked. This association, however, could not explain the relatively low LAmax of Stg1 and Stg2, which was similar to that of Stg3, even though FLN was closer to that of RTx7000 (Fig. 4a, Table 2). Hence, the Stg QTLs increased the cumulative size of leaves 2–9, yet decreased the size of upper leaves (L10+), suggesting major developmental changes around the expansion of L9 and L10. The increased size of lower leaves and reduced size of upper leaves of Stg NILs compared with RTx7000 could be a consequence of the carbon balance of the plant, hormonal action or a combination of both. Modus operandi of stay-green QTLs and differences among stay-green QTLs Although all four Stg QTLs affected the same processes (tillering, size of the upper and lower leaves), they varied in their relative effects on each process and hence in the magnitude of reduction in canopy size at anthesis. It is possible that these effects are all pleiotropic, modulated by a single gene in each QTL region, or by a single regulatory gene coordinating these responses in each QTL. The most plausible explanation would be that Stg modulates the expression of genes controlling hormones that affect leaf area development and tillering either directly, or indirectly via the carbon supply–demand balance (Fig. 1). However, the decreased main shoot leaf number of Stg3 and Stg4 QTLs could also be caused by maturity genes located in these regions. A candidate gene for flowering time (Sb02g026230, indeterminate growth 1) (Colasanti et al., 2006) is located within the Stg3 QTL boundary, whereas a candidate gene from the vernalization pathway in Arabidopsis, SbVRN1 (Sb05g006090) (Levy et al., 2002), has been identified within the Stg4 QTL boundary. Although the stay-green phenotype is only apparent after anthesis, it is largely an emergent consequence of constitutive physiological processes initiated during vegetative growth. The presence of different mechanisms to reduce canopy size suggests that it could be manipulated genetically by pyramiding various combinations of Stg QTLs into a common background. Hence, Stg2, which primarily affects tillering, could be combined with Stg4, which largely affects the size of the upper leaves, to produce a particularly small canopy which may be beneficial under severe terminal drought (Fig. 1). Earlier studies on NILs containing Stg2 plus Stg4 have found this combination to be synergistic under severe terminal drought (A. K. Borrell, unpublished). Therefore, differences in leaf area development could be exploited by plant breeders to custom-make canopy size to optimize genotype 9 environment 9 management interactions to fit particular drought-prone environments. Impact of canopy development on grain yield under drought via water availability and drought adaptation The reduced canopy size resulting from Stg QTLs increased grain yield under post-anthesis drought stress (Fig. 6). Because biomass accumulation under drought stress is limited by water Ó 2014 The Authors New Phytologist Ó 2014 New Phytologist Trust

Research 827

uptake (Hammer et al., 2010), the strong negative correlations between GLAA and bioGFP in Fig. 6(a) imply that the reduction in canopy size at anthesis via the introgression of Stg QTLs increases water availability during grain filling under terminal drought, irrespective of the growth conditions before anthesis. The effects of Stg QTLs on temporal water extraction patterns (Fig. 7) were consistent with this, as Stg QTLs shifted crop water use from the pre- to post-anthesis period. Under postanthesis water stress, shifting even small amounts of water use from the pre- to post-anthesis period can substantially increase grain yield in cereals (Hammer, 2006; Manschadi et al., 2006; Kirkegaard et al., 2007) and, consequently, grain yield is a function of post-anthesis T (Turner, 2004). Within this context, the positive relationship between bioGFP and grain number (Fig. 6b) is consistent with previous studies, which have shown that grain number is determined by the crop growth rate around anthesis and, in particular, the growth rate of the reproductive organ (Vega et al., 2001; van Oosterom & Hammer, 2008). The significant correlation between grain number and grain yield (Fig. 6c) highlighted the relatively small genotypic differences in individual grain mass (Table 3), as reduced assimilate availability under drought stress resulted in an adjustment in sink size through reduced grain number. The presence of an experimental effect on the relation between bioGFP vs GLAA (Fig. 6a), but not on the two subsequent relationships (Fig. 6b, c), illustrates that the G 9 E interaction for grain yield is at least partly an emergent consequence of pre-anthesis processes, consistent with observations by van Oosterom et al. (2011). The current results thus support the hypothesis that the positive effects of Stg QTLs on grain yield under terminal drought (Fig. 6c) are likely to be an emergent consequence of their effects on canopy development (Figs 2–5) via an effect on water uptake patterns (Fig. 7), whereas environmental effects on canopy development can account for some of the observed G9E interactions on grain yield. Importantly, all four Stg QTLs increased (P < 0.05) grain yield relative to RTx7000 under well-watered conditions in the HD treatment (Table 3). This supports earlier studies conducted with the BTx642 source of stay-green (Borrell et al., 2000b; Jordan et al., 2012). There was, however, a yield penalty associated with Stg2 and Stg3 under well-watered conditions in the LD treatment. Although biomass accumulation under well-watered conditions is limited by the amount of radiation intercepted by the crop (Hammer et al., 2010), a sorghum crop with a leaf area index (LAI) of 3.5 intercepts c. 90% of the available light (Inthapan & Fukai, 1988; Lafarge & Hammer, 2002). As the lowest LAI at anthesis was c. 3.4 under well-watered conditions, the effect of reduced canopy size on radiation interception was likely to be minimal, consistent with the absence of a yield penalty under HD for all four QTLs, and only a small penalty under LD for two of four QTLs. Our results are consistent with other studies (Kassahun et al., 2010; Vadez et al., 2011) showing that Stg QTLs can affect water extraction, TE and grain yield in different genetic backgrounds. Nonetheless, Vadez et al. (2011) observed that effects of Stg QTLs can depend on the genetic context, thus highlighting the New Phytologist (2014) 203: 817–830 www.newphytologist.com

828 Research

need to evaluate Stg QTLs in multiple genetic backgrounds. However, the NILs in those studies generally contained multiple Stg QTLs, which prevented the evaluation of individual QTL effects and may have contributed to the observed context dependence. The positive impact of stay-green on yield across multiple genetic backgrounds was unequivocally reported by Jordan et al. (2012) from breeding trials that sampled 1668 unique hybrid combinations and 23 environments, with mean yields ranging from 2.3 to 10.5 t ha1. Overall, these results highlight the role of Stg QTLs in modifying canopy development, and the consequence of such modifications on crop water use and grain yield under terminal drought. The increased water use during grain filling by Stg NILs suggests a role for roots, which could enhance water uptake via better water extraction from explored soil or increased volume of soil explored by roots (Manschadi et al., 2006). Some evidence exists that particular Stg QTLs in sorghum co-locate with QTLs for nodal root angle (Mace et al., 2012), which can affect the lateral spread of roots of mature plants (Singh et al., 2012). Importantly, qRA1_5 co-located with the Stg4 QTL described in the current paper, suggesting that root architecture can be a component of increased water use by Stg NILs. Global implications The likelihood of post-anthesis drought occurring in Australia’s northern grain belt is c. 70% (Chapman et al., 2002), and the grain yield of sorghum is also likely to be affected by post-anthesis drought stress in rain-fed farming systems of India’s west central monsoon region (DeLacy et al., 2010), the southern USA (Rosenow et al., 1983; Bandaru et al., 2006) and sub-Saharan Africa (Dingkuhn et al., 2006; Kouressy et al., 2008; Abdulai et al., 2012; Bhosale et al., 2012). Therefore, the Stg QTLs can potentially increase the yield, sustainability and profitability of farmers in rain-fed environments across the globe, without major penalties during wetter years.

Acknowledgements We thank Simon Hamlet and Andrew Douglas, as well as the operational staff at Hermitage Research Facility (Queensland Government), for assistance with field experiments, and Colleen Hunt and Emma Mace (Queensland Government) for assistance with statistical analysis of data and information on candidate genes, respectively. Henry Nguyen (Texas Tech University; University of Missouri-Columbia) and the late Darrell Rosenow (Texas A&M University Agricultural Research and Extension Center) are thanked for developing and supplying seed of 34 near-isogenic RTx7000 lines containing one or more of the Stg loci from BTx642 for use in these studies.

References Abdulai AL, Parzies H, Kouressy M, Vaksmann M, Asch F, Brueck H. 2012. Yield stability of photoperiod sensitive sorghum (Sorghum bicolor L. Moench)

New Phytologist (2014) 203: 817–830 www.newphytologist.com

New Phytologist accessions under diverse climatic environments. International Journal of Agricultural Research 7: 17–32. Adu MO, Sparkes DL, Parmar A, Yawson O. 2011. ‘Stay green’ in wheat: comparative study of modern bread wheat and ancient wheat cultivars. ARPN Journal of Agricultural and Biological Science 6: 16–24. Alam MM, Hammer GL, van Oosterom EJ, Cruickshank AW, Hunt CH, Jordan DR. 2014. A physiological framework to explain genetic and environmental regulation of tillering in sorghum. New Phytologist 201: 155– 167. Bahmani I, Hazard L, Varlet-Grancher C, Betin M, Lemaire G, Matthew C, Thom E. 2000. Differences in tillering of long- and short-leaved perennial ryegrass genetic lines under full light and shade treatments. Crop Science 40: 1095–1102. Bandaru V, Stewart BA, Baumhardt RL, Ambati S, Robinson CA, Schlegel A. 2006. Growing dryland grain sorghum in clumps to reduce vegetative growth and increase yield. Agronomy Journal 98: 1109–1120. Bhosale S, Stich B, Rattunde HF, Weltzien E, Haussmann B, Hash CT, Ramu P, Cuevas H, Paterson A, Melchinger A et al. 2012. Association analysis of photoperiodic flowering time genes in west and central African sorghum [Sorghum bicolor (L.) Moench]. BMC Plant Biology 12: 32. Boffa JM, Taonda SJB, Dickey JB. 2000. Field-scale influence of karite (Vitellaria paradoxa) on sorghum production in the Sudan zone of Burkina Faso. Agroforestry Systems 49: 153–175. Bogard M, Jourdan M, Allard V, Martre P, Perretant MR, Ravel C, Heumez E, Orford S, Snape J, Griffiths S et al. 2011. Anthesis date mainly explained correlations between post-anthesis leaf senescence, grain yield, and grain protein concentration in a winter wheat population segregating for flowering time QTLs. Journal of Experimental Botany 62: 3621–3636. Borrell AK, Bidinger FR, Sunitha K. 1999. Stay-green associated with yield in recombinant inbred sorghum lines varying in rate of leaf senescence. International Sorghum and Millets Newsletter 40: 31–34. Borrell AK, Hammer GL. 2000. Nitrogen dynamics and the physiological basis of stay-green in sorghum. Crop Science 40: 1295–1307. Borrell AK, Hammer GL, Douglas ACL. 2000a. Does maintaining green leaf area in sorghum improve yield under drought? I. Leaf growth and senescence. Crop Science 40: 1026–1037. Borrell AK, Hammer GL, Henzell RG. 2000b. Does maintaining green leaf area in sorghum improve yield under drought? II. Dry matter production and yield. Crop Science 40: 1037–1048. Borrell AK, Hammer GL, van Oosterom E. 2001. Stay-green: a consequence of the balance between supply and demand for nitrogen during grain filling? Annals of Applied Biology 138: 91–95. Borrell AK, Jordan DR, George-Jaeggli B, Hammer GL, Van Oosterom E, Klein P, Mullet J 2009. Fine mapping candidate genes for “stay-green” in sorghum: are we there yet? In: 3rd International Conference on Integrated Approaches to Improve Crop Production under Drought-Prone Environments (InterDrought-III). Shanghai, China: Shanghai Academy of Agricultural Science, L 5.03. Bos HJ, Neuteboom JH. 1998. Morphological analysis of leaf and tiller number dynamics of wheat (Triticum aestivum L.): responses to temperature and light intensity. Annals of Botany 81: 131–139. Burke JJ, Chen J, Burow G, Mechref Y, Rosenow D, Payton P, Xin Z, Hayes CM. 2013. Leaf dhurrin content is a quantitative measure of the level of preand postflowering drought tolerance in Sorghum. Crop Science 53: 1056–1065. Butler DG, Cullis BR, Gilmour AR, Gogel BJ. 2009. ASReml-R reference manual release 3. Brisbane, Qld, Australia: Qld Department of Primary Industries and Fisheries. Carberry PS, Muchow RC, Hammer GL. 1993. Modeling genotypic and environmental control of leaf area dynamics in grain sorghum 2. Individual leaf level. Field Crops Research 33: 311–328. Chapman SC, Cooper M, Hammer GL. 2002. Using crop simulation to generate genotype by environment interaction effects for sorghum in water-limited environments. Australian Journal of Agricultural Research 53: 379–389. Choudhary S, Mutava RN, Shekoofa A, Sinclair TR, Prasad PVV. 2013. Is the stay-green trait in sorghum a result of transpiration sensitivity to either soil drying or vapor pressure deficit? Crop Science 53: 2129–2134.

Ó 2014 The Authors New Phytologist Ó 2014 New Phytologist Trust

New Phytologist Christopher JT, Manschadi AM, Hammer GL, Borrell AK. 2008. Developmental and physiological traits associated with high yield and stay-green phenotype in wheat. Australian Journal of Agricultural Research 59: 354–364. Colasanti J, Tremblay R, Wong A, Coneva V, Kozaki A, Mable B. 2006. The maize INDETERMINATE I flowering time regulator defines a highly conserved zinc finger protein family in higher plants. BMC Genomics 7: 158. Crafts-Brandner SJ, Below FE, Harper JE, Hageman RH. 1984. Differential senescence of maize hybrids following ear removal. I. Whole plant. Plant Physiology 7: 360–367. 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 262: 579–588. DeLacy IH, Kaul S, Rana BS, Cooper M. 2010. Genotypic variation for grain and stover yield of dryland (rabi) sorghum in India 2. A characterisation of genotype 9 environment interactions. Field Crops Research 118: 236–242. Dingkuhn M, Singh BB, Clerget B, Chantereau J, Sultan B. 2006. Past, present and future criteria to breed crops for water-limited environments in West Africa. Agricultural Water Management 80: 241–261. Eastin JD, Hultquist JH, Sullivan CY. 1973. Physiologic maturity in grain sorghum. Crop Science 13: 175–178. Fu JD, Yan YF, Lee BW. 2009. Physiological characteristics of a functional stay-green rice “SNU-SG1” during grain-filling period. Journal of Crop Science and Biotechnology 12: 47–52. Gentinetta E, Ceppi D, Lepori C, Perico G, Motto M, Salamini F. 1986. A major gene for delayed senescence in maize. Pattern of photosynthates accumulation and inheritance. Plant Breeding 97: 193–203. Hammer GL. 2006. Pathways to prosperity: breaking the yield barrier in sorghum. Agricultural Science 19: 16–22. Hammer GL, Carberry PS, Muchow RC. 1993. Modeling genotypic and environmental control of leaf area dynamics in grain sorghum 1. Whole plant level. Field Crops Research 33: 293–310. Hammer GL, van Oosterom E, McLean G, Chapman SC, Broad I, Harland P, Muchow RC. 2010. Adapting APSIM to model the physiology and genetics of complex adaptive traits in field crops. Journal of Experimental Botany 61: 2185– 2202. Harris K, Subudhi PK, Borrell A, Jordan D, Rosenow D, Nguyen H, Klein P, Klein R, Mullet J. 2007. Sorghum stay-green QTL individually reduce post-flowering drought-induced leaf senescence. Journal of Experimental Botany 58: 327–338. Haussmann BIG, Mahalakshmi V, Reddy BVS, Seetharama N, Hash CT, Geiger HH. 2002. QTL mapping of stay-green in two sorghum recombinant inbred populations. Theoretical and Applied Genetics 106: 133–142. Henzell RG, Brengman R, Fletcher D, McCosker T 1992. Relationships between yield and non-senescence (‘stay-green’) in some grain sorghum hybrids grown under terminal drought stress. In Foale MA, Henzell RG, Vance PN, eds. 2nd Australian Sorghum Conference. Gatton, Qld, Australia: Australian Institute of Agricultural Science, Melbourne, 355–359. Hoang TB, Kobata T. 2009. Stay-green in rice (Oryza sativa L.) of drought-prone areas in desiccated soils. Plant Production Science 12: 397–408. Inthapan P, Fukai S. 1988. Growth and yield of rice cultivars under sprinkler irrigation in south-eastern Queensland. 2. Comparison with maize and grain sorghum under wet and dry conditions. Australian Journal of Experimental Agriculture 28: 243–248. Jordan DR, Hunt CH, Cruickshank AW, Borrell AK, Henzell RG. 2012. The relationship between the stay-green trait and grain yield in elite sorghum hybrids grown in a range of environments. Crop Science 52: 1153–1161. Kassahun B, Bidinger F, Hash C, Kuruvinashetti M. 2010. Stay-green expression in early generation sorghum [Sorghum bicolor (L.) Moench] QTL introgression lines. Euphytica 172: 351–362. Keating BA, Wafula BM. 1992. Modeling the fully expanded area of maize leaves. Field Crops Research 29: 163–176. Kebede H, Subudhi PK, Rosenow DT, Nguyen HT. 2001. Quantitative trait loci influencing drought tolerance in grain sorghum (Sorghum bicolor L. Moench). Theoretical and Applied Genetics 103: 266–276.

Ó 2014 The Authors New Phytologist Ó 2014 New Phytologist Trust

Research 829 Kim HK, Luquet D, van Oosterom E, Dingkuhn M, Hammer G. 2010a. Regulation of tillering in sorghum: genotypic effects. Annals of Botany 106: 69– 78. Kim HK, van Oosterom E, Dingkuhn M, Luquet D, Hammer GL. 2010b. Regulation of tillering in sorghum: environmental effects. Annals of Botany 106: 57–67. Kirkegaard JA, Lilley JM, Howe GN, Graham JM. 2007. Impact of subsoil water use on wheat yield. Australian Journal of Agricultural Research 58: 303– 315. Kouressy M, Dingkuhn M, Vaksmann M, Heinemann AB. 2008. Adaptation to diverse semi-arid environments of sorghum genotypes having different plant type and sensitivity to photoperiod. Agricultural and Forest Meteorology 148: 357–371. Lafarge TA, Broad IJ, Hammer GL. 2002. Tillering in grain sorghum over a wide range of population densities: identification of a common hierarchy for tiller emergence, leaf area development and fertility. Annals of Botany 90: 87– 98. Lafarge TA, Hammer GL. 2002. Predicting plant leaf area production: shoot assimilate accumulation and partitioning, and leaf area ratio, are stable for a wide range of sorghum population densities. Field Crops Research 77: 137–151. Levy YY, Mesnage S, Mylne JS, Gendall AR, Dean C. 2002. Multiple roles of Arabidopsis VRN1 in vernalization and flowering time control. Science 297: 243–246. Lopes MS, Reynolds MP. 2012. Stay-green in spring wheat can be determined by spectral reflectance measurements (normalized difference vegetation index) independently from phenology. Journal of Experimental Botany 63: 3789–3798. Mace E, Singh V, van Oosterom E, Hammer G, Hunt C, Jordan D. 2012. QTL for nodal root angle in sorghum (Sorghum bicolor L. Moench) co-locate with QTL for traits associated with drought adaptation. Theoretical and Applied Genetics 124: 97–109. Manschadi AM, Christopher J, Devoil P, Hammer GL. 2006. The role of root architectural traits in adaptation of wheat to water-limited environments. Functional Plant Biology 33: 823–837. Mason MG, Ross JJ, Babst BA, Wienclaw BN, Beveridge CA. 2014. Sugar demand, not auxin, is the initial regulator of apical dominance. Proceedings of the National Academy of Sciences, USA 111: 6092–6097. McKeown FR. 1978. A land classification of the Hermitage Research Station. Brisbane, Qld, Australia: Division of Land Utilisation. Queensland Department of Primary Industries. Mondal WA, Choudhuri MA. 1985. Comparison of phosphorus mobilization during monocarpic senescence in rice cultivars with sequential and non-sequential leaf senescence. Plant Physiology 65: 221–227. Northcote KH. 1979. A factual key for the recognition of Australian soils. Adelaide, SA: Rellim Technical Publications. van Oosterom EJ, Borrell AK, Deifel KS, Hammer GL. 2011. Does increased leaf appearance rate enhance adaptation to postanthesis drought stress in sorghum? Crop Science 51: 2728–2740. van Oosterom EJ, Carberry PS, O’Leary GJ. 2001. Simulating growth, development, and yield of tillering pearl millet I. Leaf area profiles on main shoots and tillers. Field Crops Research 72: 51–66. van Oosterom EJ, Hammer GL. 2008. Determination of grain number in sorghum. Field Crops Research 108: 259–268. Passioura JB, Angus JF. 2010. Chapter 2 – Improving productivity of crops in water-limited environments. Advances in Agronomy 106: 37–75. Paterson AH, Bowers JE, Bruggmann R, Dubchak I, Grimwood J, Gundlach H, Haberer G, Hellsten U, Mitros T, Poliakov A et al. 2009. The Sorghum bicolor genome and the diversification of grasses. Nature 457: 551–556. R Development Core Team. 2012. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Rajcan I, Tollenaar M. 1999. Source : sink ratio and leaf senescence in maize: I. Dry matter accumulation and partitioning during grain filling. Field Crops Research 60: 245–253. Rosenow DT. 1977. Breeding for lodging resistance in sorghum. In: Loden HD, Wilkinson D, eds. Proceedings of the 32nd Annual Corn and Sorghum Research Conference. Washington, DC, USA: American Seed Trade Association, 171– 185.

New Phytologist (2014) 203: 817–830 www.newphytologist.com

New Phytologist

830 Research Rosenow DT, Quisenberry JE, Wendt CW, Clark LE. 1983. Drought tolerant sorghum and cotton germplasm. Agricultural Water Management 7: 207–222. Sanchez AC, Subudhi PK, Rosenow DT, Nguyen HT. 2002. Mapping QTLs associated with drought resistance in sorghum (Sorghum bicolor L. Moench). Plant Molecular Biology 48: 713–726. Singh V, van Oosterom EJ, Jordan DR, Hammer GL. 2012. Genetic control of nodal root angle in sorghum and its implications on water extraction. European Journal of Agronomy 42: 3–10. Subudhi PK, Rosenow DT, Nguyen HT. 2000. Quantitative trait loci for the stay green trait in sorghum (Sorghum bicolor L. Moench): consistency across genetic backgrounds and environments. Theoretical and Applied Genetics 101: 733–741. Tao YZ, Henzell RG, Jordan DR, Butler DG, Kelly AM, McIntyre CL. 2000. Identification of genomic regions associated with stay green in sorghum by testing RILs in multiple environments. Theoretical and Applied Genetics 100: 1225–1232. Tivet F, Pinheiro BD, DeRaissac M, Dingkuhn M. 2001. Leaf blade dimensions of rice (Oryza sativa L. and Oryza glaberrima Steud.). Relationships between tillers and the main stem. Annals of Botany (London) 88: 507–511. Tuinstra MR, Ejeta G, Goldsbrough P. 1998. Evaluation of near-isogenic sorghum lines contrasting for QTL markers associated with drought tolerance. Crop Science 38: 835–842. Tuinstra MR, Grote EM, Goldsbrough PB, Ejeta G. 1996. Identification of quantitative trait loci associated with pre-flowering drought tolerance in sorghum. Crop Science 36: 1337–1344. Tuinstra MR, Grote EM, Goldsbrough PB, Ejeta G. 1997. Genetic analysis of post-flowering drought tolerance and components of grain development in Sorghum bicolor (L.) Moench. Molecular Breeding 3: 439–448.

Turner NC. 2004. Agronomic options for improving rainfall-use efficiency of crops in dryland farming systems. Journal of Experimental Botany 55: 2413– 2425. UNCTAD. 2011. Water for Food. Innovative water management technologies for food security and poverty alleviation. United Nations, New York and Geneva, 1–3. UNWWDR4. 2012. Managing water under uncertainty and risk. Paris, France: The United Nations World Water Development Report 4. Vadez V, Deshpande SP, Kholova J, Hammer GL, Borrell AK, Talwar HS, Hash CT. 2011. Stay-green quantitative trait loci’s effects on water extraction, transpiration efficiency and seed yield depend on recipient parent background. Functional Plant Biology 38: 553–566. Vega CRC, Andrade FH, Sadras VO. 2001. Reproductive partitioning and seed set efficiency in soybean, sunflower and maize. Field Crops Research 72: 163– 175. Wada Y, Wada G. 1991. Varietal differences in leaf senescence during ripening period of advanced indica rice. Japan Journal of Crop Science 60: 529–536. Xu WW, Subudhi PK, Crasta OR, Rosenow DT, Mullet JE, Nguyen HT. 2000. Molecular mapping of QTLs conferring stay-green in grain sorghum (Sorghum bicolor L. Moench). Genome 43: 461–469. Yang Z, Hammer G, van Oosterom E, Rochas D, Deifel K. 2010. Effects of pot size on growth of maize and sorghum plants. In: George-Jaeggli B, Jordan DR, eds. 1st Australian Summer Grains Conference. Gold Coast, Qld, Australia: Grains Research and Development Corporation. Zheng HJ, Wu AZ, Zheng CC, Wang YF, Cai R, Shen XF, Xu RR, Liu P, Kong LJ, Dong ST. 2009. QTL mapping of maize (Zea mays) stay-green traits and their relationship to yield. Plant Breeding 128: 54–62.

New Phytologist is an electronic (online-only) journal owned by the New Phytologist Trust, a not-for-profit organization dedicated to the promotion of plant science, facilitating projects from symposia to free access for our Tansley reviews. Regular papers, Letters, Research reviews, Rapid reports and both Modelling/Theory and Methods papers are encouraged. We are committed to rapid processing, from online submission through to publication ‘as ready’ via Early View – our average time to decision is