Exploring physiological and genetic determinants of tillering in sorghum
M.M. Alam13, G.L. Hammer1, E.J. van Oosterom1, A. Cruickshank2, and D.R. Jordan2
The University of Queensland, School of Land, Crop and Food Sciences, Brisbane, Qld 4072
2DEEDI, Hermitage Research Station, Yangan Road, Warwick, Qld 4370
3 Corresponding author:
ABSTRACT
Tillering is a plastic trait of cereals that affects grain yield and adaptability through the regulation of leaf area development and fertile tiller production. Though the trait is important for crop adaptation, its physiology and genetics are poorly understood. Recent investigations in sorghum using material with a common genetic background revealed that variation in tillering in sorghum could be explained by internal plant competition, which could be quantified by a carbohydrate supply and demand index. To validate this result in a wide range of genotypes and to further explore the genetic determinants of this trait, experiments were conducted in a low tillering (glasshouse with low light intensity) and high tillering environment (spaced planting in the field). Tillering behaviour of 90 genotypes (inbred and hybrids) with diverse origin but structured genetic background was monitored. Genotypes differed significantly in total and fertile tiller number in both environments. While internal plant competition during the critical growth period for tillering explained part of the variation in tillering, the likely involvement of other factors like hormonal regulation was clear. An analysis of combining ability revealed that the expression of tillering is primarily additive and SC62C isa potential germplasm source for high tillering in sorghum. Wide genotypic variation in tillering revealed some promising genotypes for future genetic and physiological study.
INTRODUCTION
Tillering is a key component of expression of phenotypic plasticity in cereals, and its response to changes in growing environments is important in relation to climate change. It is one of the most important agronomic traits related to crop adaptation to target environments, as high tillering is generally suitable for favorable growing conditions, whereas low tillering may be more desirable for stress conditions (Daisuke Fujita, Ebron et al. 2010). Therefore, a good understanding of regulation of tillering in cereals is required to identify genotypes with adaptation to target environments. Production of few but vigorous productive tillers can restrict plant size, which can increase post-anthesis water availability and grain yield in water limited conditions (Hammer 2006). As a complex trait, tillering is controlled by multiple genes that regulate developmental processes and interact with environmental conditions. Tiller outgrowth depends on resource availability and the ratio of carbohydrate supply and demand (S/D) has been used in rice (Dingkuhn, Luquet et al. 2006; Luquet, Dingkuhn et al. 2006) and sorghum (Kim, Luquet et al. 2010; Kim, vanOosterom et al. 2010) to relate tiller appearance to plant internal competition for resources. The S/D ratio is a complex indicator of plant carbohydrate status, in which solar radiation and leaf area determine carbohydrate supply via photosynthesis and temperature and leaf size determine carbohydrate demand associated with growth (Hammer, Carberry et al. 1993; Kim, Luquet et al. 2010; Mazzella, Bertero et al. 2000; Tardieu, Granier et al. 1999). Recent investigation of tillering in a few sorghum genotypes revealed that internal plant competition for resources could explain most of the observed genotypic variation in maximum tiller number, but that genotypes also differed in their propensity to tiller (Kim, Luquet et al. 2010), possibly indicating hormonal control of tillering. The aim of the present paper is to study the physiological and genetic control of tillering for a wide ranges of sorghum genotypes.
MATERIAL AND METHODS
Plant materials
A set of 90 sorghum genotypes (51 inbred lines and 39 hybrids including one commercial) was included in this study. The inbred lines included in this study were the genotypes of previous tillering and root experiments (Table 1). The hybrids included a subset of 30 that formed an incomplete North Carolina Design II produced from 8 males and 4 females. Females used to produce hybrids were CMS lines. Parents of mapping populations, and male and female parents for North Carolina Design II wereincluded in inbred evaluation.
Table 1Diverse germplasm from elite parent lines (inbred)used in the experiments.
Genotype / Origin / CharacteristicsAi4 / China
B35 / Ethiopia / Partially converted IS 12555, Highlystay green, female parent of NC II design
B923296 / Australia / Elite stay green parent ex QPIF breeding program, female parent of NC II design
Dorado / El Salvador
ICSV745 / India / Parent of mapping population
IS 8525 / Ethiopia / Parent of mapping population for ergot resistance
IS12611C / Ethiopia
IS17214 / Nigeria
ISCV400 / Mali
Karper 669 / USA / Diverse yellow endosperm germplasm line
KS115 / USA / Large seed
LR2490-3 / China
LR9198 / China
M35-1 / India / Drought resistant
Malisor 84-7 / Mali
MLT135 / USA / Elite moderately senescent parent line ex TAMU breeding program
MP531 / Southern Africa
QL12 / Australia / Source of stay green drought resistant. A male parent of NCII design
QL33 / Australia / Elite moderately senescent parent line ex QPIF breeding program. A female parent of NCII design
QL36 / Australia / Elite moderately senescent parent line ex QPIF breeding program. Male parent of NCII design
R890562 / Australia / Elite moderately senescent parent line ex QPIF breeding program
R9188 / USA / Partially converted derivatives of sweet sorghum Rio. A male parent of NC II design.
R931945-2-2 / Australia / Elite stay green parent ex QPIF breeding program. Parent of mapping population, a male parent of NC II design.
R9403463-2-1 / Australia / Elite moderately senescent parent line ex QPIF breeding program. Parent of mapping population for brown mid rib.
R9733 / USA / Breeding line from TexasA&MUniversity breeding program
R993396 / Australia / Elite moderately senescent parent line ex QPIF breeding program. Mmale parent of NCII design.
R999003 / Australia / Selected from an interespciifc cross between S arundinaceum(African wild type high tillering) and R931945-2-2 (low tillering)
R999017 / Australia / Selected from an interespciifc cross between S arundinaceum (African wild type high tillering) and R931945-2-2 (low tillering)
R999066 / Australia / Selected from an interespciifc cross between S arundinaceum (African wild type high tillering) and R931945-2-2 (low tillering)
R999081 / Australia / Selected from an interespciifc cross between S arundinaceum (African wild type high tillering) and R931945-2-2 (low tillering)
R999110 / Australia / Selected from an interespciifc cross between S arundinaceum (African wild type high tillering) and R931945-2-2 (low tillering)
R999197 / Australia / Selected from an interespciifc cross between S arundinaceum (African wild type high tillering) and R931945-2-2 (low tillering)
R999218 / Australia / Selected from an interespciifc cross between S arundinaceum (African wild type high tillering) and R931945-2-2 (low tillering)
Rio / USA / Sweet sorghum
RS29 / India / Drought resistant
SC103-14E / Ethiopia
SC108C / Ethiopia
SC170-6-8 / Ethiopia / Partly converted version of IS12661 a caudatum line ex Ethiopia. A high yielding genotype and parent of mapping population.
SC23 / Ethiopia / A durra genotype
SC35C / Ethiopia / Fully converted land race derived from IS 12556
SC56-14E / Sudan / Source of stay green drought resistance. A male parent of NC II design.
SC62C / Sudan / A high tillering genotype, male parent of NC II design
SC636-6 / Uganda
SC999 / Ethiopia / Partially converted IS 11080
TAM422 / USA / Early hybrid parent lacking in stay green drought resistance.
Tx2536 / USA / Early hybrid parent lacking in stay green drought resistance.
Tx2737 / USA / A high yielding genotype having yellow endosperm, widely used as parent commercially in the USA
Tx2895 / USA / Widely used commercially in the USA
Tx430 / USA / yellow endosperm, widely used as parent commercially in the USA
TX623 / USA / An elite US female pedigree BTx3197/SC170-6-4-4. Female parent of NCII design.
TX7000 / USA / Early hybrid parent lacking in stay green drought resistance. Male parent of NCII design.
Experimental sites
Experiments were conducted at two locations with contrasting temperature and radiation regimes (Table 2). Experiment 1 was sown in September 2008 in a glasshouse at the University of Queensland in St Lucia and Experiment 2 was sown in December 2008 in the field at Hermitage Research Station, Warwick, Queensland. Exp. 1 tended to have higher temperature and Exp. 2 higher daily radiation (Table 2).
Table 2 Environmental conditions for the experiments under study. Radiation and temperature parameters were calculated up to final primary tiller appearance.
environmental parameters / Exp 1 / Exp2Sowing date / 11 September 2008 / 17 December 2008
Maximum daily radiation (MJm-2 day-1) / 7.8 / 24.9
Average daily minimum temperature (oC) / 20.2 / 16.0
Average daily maximum temperature (oC) / 31.9 / 28.7
Experimental design
The experimental design in the glass house was a randomized complete block with three replications. Pots of 30 cm diameter were filled with pre-sterilized and pre-fertilized University of California soil mix (containing sand and peat). Four seeds were sown in each pot and after emergence plants were gradually thinned to one plant per pot by the four leaf stage. The whorl of each axis in each plant was sprayed daily with 0.3% Ca (NO3)2 after establishments to minimise symptoms of Ca deficiency.Three weeksafter sowing, a 2% solution of AQUASOL was addedin two consecutive weeksto provide additional N. Watering was done regularly and no drought stress occurred.
The field experiment was laid out as a randomized complete block with three replications. Plot size was one row, 4.75 m long. The site was fertilized and cultivated before planting. The experiment was machine planted with 75 cm row spacing and thinned to a maximum of 5 plants per plot at the 3-leaf stage to give widely spaced plants (approx 1 m plant-to-plant spacing) to maximise expression of tillering. The experiment was rain-fed and weeds were controlled as necessary.
Both experiments were terminated around anthesis.
Observations and calculations
Data on leaf and plant size, leaf appearance and tiller number were recorded on one plant per genotype in each replication. The number of visible and fully expanded leaves on the main shoot and the number of emerged tillers were recorded three times a week. A leaf was visible when its tip was visible above the enclosing leaf whorl and fully expanded if its ligule was visible above the ligule of the previous leaf. Total leaf number (TLN) at anthesis was the number of fully expanded leaves produced on the main shoot. Leaf size was represented by the final length (LL) and maximum width (LW)of main shoot leaves 5, 7, and 9. Leaf area was obtained by multiplying length and width by a shape coefficient of 0.69(Kim, vanOosterom et al. 2010). Internode diameter (ID) was measured using digital slide callipers on the narrowest region of the first internode above the basal root zone. Plant height (PHT) was measured from base to top of the main stem inflorescence.
Thermal time (δTT)was calculated from hourly data as, where Tbase =11°C (Hammer, Carberry et al. 1993). Carbon supply/demand (S/D) index was calculated using the formula of Kim, Luquet et al(2010).δTT was used to measure leaf expansion duration of main shoot leaf 5 (LED5) as the time between tip and ligule appearance of that leaf. Incident daily global radiation for the duration of expansion of main shoot leaf5 (RADLED5) was calculated for that period.
Statistical Analysis
Statistical analysis were conducted in ASREML-R(Butler 2007) using theREML mixed model approachallowing for spatial variation across each experiment (Gilmour 1997). For each trait, this produced variance components and predicted genotype values for each location. BLUPs (Best Linear Unbiased Predictions) for each genotype for each experiment were used in a principal component analysis.
Analysis of variance of North Carolina Design II study on a subset of the hybrids was performed using alinear model(Lee, Ahmadzadeh et al. 2005).The general and specific combining ability (GCA and SCA, respectively) represent the additive and non-additive portion of the observed genetic variance respectively (Rojas and Sprague 1952). The sum of squares ratio [(male SS + female SS)/Hybrid SS ](Pixley and Frey 1991)was used to assess the relative importance of additive (GCA) and non-additive (SCA) genetic effects on tillering. The expected variationdue to female and male parents corresponds to GCA and thatdue to the male x female interaction to SCA(Hallauer and Miranda 1988). If the combining ability effectswere significant, then the GCA estimates (gi or gj) for allparental lines and the SCA estimates (sij) for all hybrid genotypeswere calculated according to Beil and Atkins (1967):
gi = (yi. – y..)
gj = (y.j – y..)
sij = (yij - yi.- y.j + y..)
where yij is the mean of thehybrid between the ith female and the jth male parent;yi.is the mean of all hybrids involving the ith female parent; y.j is the mean of all hybrids involving the jth male parent;and y.. is the mean of all hybrids. Standard errors for gi orgj estimates and sij estimates were calculated using themethods of Cox and Frey (1984), where SEGCA= or forfemales or males, respectively. MSfland MSml are the respectivefemale x location and male x location mean squares and are multipliedby the appropriate proportion of total number of observations[males x females x reps(per location) x locations]. Standard errors forsij estimates were calculated as SESCA = where MSfml is the female x male xlocation mean square. If MSfl, MSml, and MSfml were not significant,they were replaced by the pooled error to calculate standarderrors for gi,gj, and sij estimates, respectively. Two-tailedt tests were used to test the significance of the gi,gj, andsij estimates from zero, where t = GCA/SEGCA and t = SCA/SESCA,respectively (Singh and Choudhury 1977).
RESULTS
Genetic and environmental variation in tillering and related traits in sorghum
Analysis of variance of inbred lines revealed significant genotypic and environmental differences for all traits (data not shown). Total tiller number (TTN) ranged from 0 to 8.17 in Exp1 (low tillering) and from 0.60 to 15.67 in Exp2 (high tillering). The higher TTN in Exp2 was due to the higher radiation and lower temperature (Table 2), which increased the S/D ratio in Exp2 compared to Exp1.There was also a wide range of variation for all other traits (Table 3). Leaves were longer and wider in Exp1 than in Exp2, and this was due to the changed light quantity and quality in the glasshouse, which affected leaf elongation (Smith and Whitelam 1997). Greater leaf number in Exp2 may be due to longer photoperiodin the field condition (data not shown) (Kumar, Hammer et al. 2009). Variation in plant height was reduced in Exp2.
Table 3Descriptive statistics of leaf and plant size and tiller number in sorghum inbred lines in the two experiments
Traits / EXP-1 / EXP-2Max / Min / Mean / Std / Max / Min / Mean / Std
ID / 20.30 / 10.16 / 16.15 / 1.98 / 25.58 / 13.95 / 20.24 / 1.84
LL5 / 48.73 / 28.17 / 36.05 / 4.03 / 18.10 / 11.98 / 15.17 / 2.49
LL7 / 70.01 / 47.17 / 57.15 / 4.56 / 31.55 / 22.44 / 26.05 / 3.11
LL9 / 81.75 / 62.61 / 72.31 / 3.83 / 45.03 / 36.95 / 39.57 / 2.61
LW5 / 3.08 / 1.88 / 2.42 / 0.27 / 2.09 / 1.62 / 1.87 / 0.18
LW7 / 6.02 / 3.56 / 4.76 / 0.59 / 4.34 / 3.05 / 3.69 / 0.44
LW9 / 8.14 / 4.71 / 6.71 / 0.69 / 6.64 / 4.46 / 5.64 / 0.56
TLN / 17.58 / 10.01 / 14.73 / 1.67 / 22.00 / 14.49 / 17.79 / 1.30
PHT / 301.30 / 65.45 / 123.47 / 39.00 / 232.14 / 73.30 / 102.77 / 30.92
TTN / 8.17 / 0.00 / 1.10 / 1.20 / 15.67 / 0.60 / 2.84 / 1.63
Physiological control of tillering
Physiological determinants of tillering in sorghum were examined via the extent of association of plant and leaf size with tillering. Though leaf size traits were negatively associated with tillering, only the width of older leaves (7th and 9th) had significant effects. Stem diameter showed highly significant negative genetic correlations with tiller number. Total number of leaves was also associated with the reduction of tiller number. The relationship of traits with tiller number was consistent across environments (Table 4).
Table 4Genetic correlations of plant and leaf size traits with total tiller number in each experiment and across experiments
Traits / EXP-1 / EXP-2 / Across environmentsStem diameter (ID) / -0.358*** / -0.449*** / -0.414***
5th Leaf length (LL5) / -0.143ns / -0.200* / -0.159ns
7th leaf length (LL7) / -0.152ns / -0.146ns / -0.146ns
9th leaf length (LL9) / -0.164ns / -0.189* / -0.178*
5th leaf width (LW5) / -0.095ns / -0.144ns / -0.125ns
7th leaf width (LW7) / -0.298** / -0.355*** / -0.336***
9th leaf width (LW9) / -0.508*** / -0.509*** / -0.518***
Plant height (PHT) / 0.038ns / 0.119ns / 0.063ns
Total leaf number (TLN) / -0.241* / -0.221* / -0.237*
* P<0.05; ** P<0.01;*** P<0.001; ns =not significant
Tiller number was significantly negatively correlated with the increase in leaf width from Leaf 5 to Leaf 9for both inbred lines and hybrids and around 35% of variation in tiller number of inbred lines could be attributed to the rate of increase in leaf width in both environments, whereas for hybrids, this was 34% in the low tillering environment and 22% in the high tillering environment (Figure 1 and Figure 2). The Leaf Width Increase Ratio (LWIR) is a major component of the demand term of the S/D index (Kim, Luquet et al. 2010; Kim, vanOosterom et al. 2010).
Figure 1Total tiller number versus increase in leaf width between Leaf 5 and Leaf 9 in Exp1 (low tillering environment)
Figure 2Total tiller number versus increase in leaf width between Leaf 5 and Leaf 9 in Exp1 (high tillering environment)
Genetic control of tillering
The additive and non-additive genetic effects on tillering were examined for the 30 hybrids that made up the incomplete North Carolina Design II mating design. Combined analysis of variance across the two experiments revealed highly significant variation among hybrids, which was predominantly due to the contribution of male parents (Table5). Interactions were all non significant. The SS ratio for tillering was 0.89, indicating thatdifferences in tillering were primarily due to additive gene effects.
The lack of interactions and non-additive gene effects was supported by the analyses of combining ability across experiments, which showedthat SCA estimates were generally non-significant. In contrast, the additive genetic effects (GCA) for male parents was highly significant (P<0.001, Table 6), mainly because of the high tiller number of SC62C. Male parents with the lowest GCA (and thus the lowest tiller number) were R31945-2-2, R9188, QL12, and Tx7000. The GCA for female parents was significant at P<0.05 only (Table 4), with QL33 and Tx623 having the highest values and B35 the lowest. High broad sense heritability (80.3% in Exp1 and 87.4% in Exp2) indicated the strong genetic control of the trait.