Previous PageTable Of Contents

Increasing grain size and reducing screenings in wheat using a tiller inhibition gene – investigating grain morphology by image analysis

JH Mitchell 1,3, Scott Chapman 1, Greg Rebetzke 2 and Shu Fukai 3

1 CSIRO Plant Industry, QBP, 306 Carmody Rd, St Lucia 4067 Brisbane, Qld, Australia. Email jaquie.mitchell@csiro.au.
2
CSIRO Plant Industry, G.P.O. Box 1600 Canberra, ACT, 2601, Australia.
3
School of Land, Crop and Food Sciences, The University of Queensland, Brisbane 4072, Qld, Australia.

Abstract

Soil water deficit coupled with rising air temperatures during grain-filling can often lead to the development of small kernels (‘screenings’) as some proportion of the harvested wheat crop. Growers want to minimize screenings, as the value of the crop is reduced when the percentage of screenings is high. Wheat breeding lines containing the tiller inhibition (tin) gene produce fewer tillers, and when water-stressed, tend to have larger mean kernel weights and reductions in grain screenings. In 2006, a series of dryland field experiments were sown to investigate the distribution of individual kernel size parameters among backcross-derived, Silverstar lines varying for presence of the tin gene. Digital imaging was used to test the hypothesis that mean kernel weight advantage of tin lines was associated with a higher frequency of larger-sized kernels. There was a strong, positive linear relationship between kernel weight and kernel width. Lines containing the tin gene had a higher mean kernel weight and width, and a larger distribution of kernel width than free-tillering lines. The wider distribution reflected a high frequency of large kernel widths resulting in tin lines producing significantly less screenings. The physiological and morphological bases for these differences in screenings are discussed.

Keywords

drought, kernel weight, tin gene, screenings

Introduction

Grain yield improvement in wheat, as with other cereals, has been largely achieved through selection for lines with high kernel number per unit area (Fischer, 2008; Sayre et al., 1997). In some instances, this has been achieved indirectly through selection of lines with a high capacity to tiller, and tends to be associated with smaller kernels (Acreche and Slafer, 2006) even under non-limiting conditions. Water stress, in particular post-anthesis water stress, occurs frequently in the northern region of the Australian wheatbelt. The reliance of crops on stored soil moisture coupled with rising temperatures and vapour pressure deficits throughout grainfilling, commonly results in reduced photosynthesis, premature ripening and subsequently small kernels (Wardlaw, 2002).

Selection for the reduced-tillering tin gene has been shown to result in larger stems, higher harvest index and increased kernel weight (Duggan et al., 2005; Richards, 1988). To reduce the incidence of small kernels and consequently screenings, van Herwaarden et al.(1998) suggested the incorporation of the tin gene into adapted wheat germplasm may be of value. While reduced grain screenings is an important selection criterion in wheat breeding it can be challenging to measure on large numbers of lines, particularly under favourable growing conditions where considerable selection work can be based. Average kernel weight is negatively associated with percent screenings (Sharma and Anderson, 2004). However, average kernel weight provides no indication of the distribution of kernel size within a seedlot. To improve our understanding of screenings and kernel size, a measure of kernel size distribution is required. There are two methods available to establish this: weighing of individual kernels within a sample; or use of image analysis to give kernel size parameters. As ‘screenings’ are measured by industry using a 2 mm-slotted sieve it was considered that image analysis that provides a two-dimensional estimate of length and width of individual kernels within a sample, would be a more direct and therefore useful measure.

The timing and pattern of water stress in relation to crop development is critical with the period anthesis ± 10 days being identified as important in wheat (Woodruff and Tonks, 1983). It was hypothesised that tin lines with fewer tillers would have relatively synchronous flowering compared to free-tillering lines that can produce late tillers, and because of this tin lines may have a smaller kernel size distribution than free-tillering lines in a terminal water stress. Consequently, we tested the hypothesis that the mean kernel weight advantage of tin lines was associated with a greater frequency and narrower distribution of larger (wider) sized kernels. The alternate hypothesis was that tin lines were associated with a higher frequency of larger kernels representing a wider kernel size distribution.

Methods

Near-isogenic lines varying in tiller number were developed by twice back-crossing the tin gene into the free-tillering, high-screenings variety, Silverstar (Rebetzke pers.comm.). Lines containing the tin1 allele, as identified by molecular marker Xgwm136 (Spielmeyer and Richards, 2004), were classed as restricted and semi-restricted (T in Figure 2) in tillering, whereas those with the alternative wild-type allele were free-tillering (W). Two dryland field experiments were conducted in 2006 to evaluate the performance of 10 Silverstar tin (T) and 10 W lines at Kingsthorpe (Lat -27.47, long 151.83) and Gatton (-27.55, 152.33) QLD. Plots were 6 m long and 7-rows wide with a row spacing of 0.25 m. Seed amounts were adjusted for each line to target a plant density of 150 plants/m2. Experiments were randomised complete block designs with two replications. Sown into dry soil on June 9th (Kingsthorpe) and 15th (Gatton), both experiments were irrigated (41 and 63 mm respectively) to ensure good plant establishment. Initial soil moisture differed with Gatton having a relatively full profile (221 mm plant available water; PAW), while Kingsthorpe only had 74 mm PAW. For each line, kernel weight was determined for a random sample of 300 grain. The 300 grain sample was scanned on a flat-bed scanner at 300 dpi resolution. Images were analysed using the open source program ImageJ (National Institute of Health, USA, rsb.info.nih.gov/ij) and individual kernel width, length, perimeter and area estimated. Frequency data on kernel width was utilised to compare with mean kernel weight.

To determine the range in anthesis date among spikes within a plant, spikes from two plants within a plot, for four representative tin (SsrT02, SsrT16, SsrT17, SsrT65) and wild (SsrW35, SsrW43, SsrW47, Silverstar) lines were tagged as anthesis (Z65) occurred. Anthesis data was collected from another 2006 Gatton dryland trial and is used to represent the differences in range of anthesis dates between main stem spike and last tiller spike, occurring in the tin versus W lines.

Results

Environments

Due to the low initial soil moisture at Kingsthorpe, it was determined that the trial should be supplied with approximately 41 mm of irrigation, 75 d after sowing and 10 d after commencement of stem elongation. This relieved water stress around anthesis but allowed for development of post-anthesis water stress (data not shown). Pre-and post-anthesis water stress led to an average experiment yield of 2.3 t/ha with relatively high kernel screenings (18%).

In contrast, the Gatton dryland trial, sown into a relatively full soil profile, developed some symptoms of pre-anthesis water stress. However, 65 mm rain fell around anthesis, and this in combination with a further 44 mm during grain fill ensured good grain set, a relatively high yield (4.9 t/ha) and low screenings (5.8%)

Stem number

In the Kingsthorpe stress environment, the free-tillering lines produced a stem number of 4.8 stems/plant at maturity; this was 60% greater than tin lines (3 stems/plant). Within the tin lines, restricted tin lines produced an average 2.5 stems/plant while the semi-restricted tin lines produced 3.5 stems/plant.

Table 1: Average kernel weight, spike number, kernel number, kernels per spike, screenings and kernel width of Silverstar restricted (tin) and free-tillering (W) lines in two 2006 environments. Kingsthorpe pre- and post-anthesis and Gatton mild pre-anthesis water stress.

Parameter

Kingsthorpe

Gatton

 

tin

W

tin

W

Kernel weight (mg)

28.3 ± 1.1

20.7 ± 1.2

32.0 ± 1.0.

29.5 ± 1.0

Spike number (m-2)

225 ± 51

464 ± 51

307 ± 29

430 ± 29

Kernel number (m-2)

7472 ± 813

12748 ± 815

14455 ± 910

17651 ± 909

Kernels per spike

35±1.8

30±1.1

39±1.2

35±1.2

Screenings (%)

11.8 ± 3.6

25.1 ± 3.6

5.0 ± 0.7

6.7 ± 0.7

Kernel width (mm)

2.8 ± 0.005

2.5 ± 0.004

3.0 ± 0.004

2.9 ± 0.004

Range in anthesis

As hypothesised, within a plant, the tin lines had relatively synchronous (3.1 d ± 0.74) flowering compared to free-tillering lines (6.5 d ± 0.94).

Kernel weight, width and percent screenings

In both environments, kernel weight was greater (+31% and +8%) in tin than free-tillering lines, the difference largest in the more severely water stressed Kingsthorpe environment (Table 1). Maintenance of large kernel size by tin lines resulted in a 50% reduction in screenings relative to free-tillering lines at Kingsthorpe. Across all lines grown in both environments, there was a highly significant (p<0.01) negative association between percent screenings and kernel weight (Figure 1a) while there were few screenings when average kernel weight exceeded 30 mg. A highly significant (p<0.01) positive association was observed between average kernel weight and width, with tin lines maintaining a higher mean kernel width than free-tillering lines (Figure 1b).

The frequency distribution of kernel width (Figure 2) indicated that tin lines with a greater mean kernel width had a wider kernel width distribution than free-tillering lines. However, tin lines only had 1.6% of kernels with widths less than 2 mm (the industry screen slotted sieve width) compared with 5.3% in free-tillering lines. These values are considerably less than percent screenings based on weight because image analysis did not give a measure of the third dimension, kernel depth, which if thin would allow kernels to fall through slotted sieve. As expected, the kernel width was greater in the reduced water-stress Gatton environment and its distribution (data not shown) was similar between tin (0.9% below 2 mm) and free-tillering lines (1.8% below 2 mm).

Figure 1: Association between a) screenings (%) and kernel weight (mg) and b) average kernel weight (mg) and width (mm) for plot data of Silverstar tin (T) and wild (W) lines from 2006 Kingsthorpe and Gatton trials.

Figure 2: Frequency distribution of kernel width for Silverstar tin (solid) and free-tillering (dashed) lines. Data is for the water-stressed Kingthorpe site.

A strong positive association (data not shown) between stem number per plant at maturity and screenings (r =0.75, p<0.01) indicated, for every extra stem, screenings would increase by 6% in the Kingsthorpe environment. This increase in screenings reflected a reduction in kernel width by 0.14 mm for every extra stem (r = -0.84, p<0.01).

Conclusion

Silverstar-derived backcross lines containing the tin gene had a larger mean kernel weight and width, and a wider distribution of kernel width than free-tillering sister lines. The wider distribution reflected a high frequency of large kernel widths and this resulted in tin lines which produced less screenings. Despite the reduced range in anthesis date for tillers within tin lines, it is speculated that anthesis date of individual florets within spikes of tin lines may be asynchronous as there are more florets set per spike (37 kernels) in tin than free-tillering (33 kernels per spike) lines. This potential asynchrony in flowering may account for the wider distribution in kernel widths observed for tin lines and is under further investigation. These results highlight the importance of tiller and stem production on crop water dynamics, and subsequent impact on final kernel weight and size.

References

Acreche, M.M., and G.A. Slafer. 2006. Grain weight response to increases in number of grains in wheat in a Mediterranean area. Field Crops Research 98:52-59.

Duggan, B.L., R.A. Richards, A.F. van Herwaarden, and N.A. Fettell. 2005. Agronomic evaluation of a tiller inhibition gene (tin) in wheat. I. Effect on yield, yield components, and grain protein. Australian Journal of Agricultural Research 56:169-178.

Fischer, R.A. 2008. The importance of grain or kernel number in wheat: A reply to Sinclair and Jamieson. Field Crops Research 105:15-21.

Richards, R.A. 1988. A Tiller Inhibitor Gene in Wheat and Its Effect on Plant-Growth. Australian Journal of Agricultural Research 39:749-757.

Sayre, K.D., S. Rajaram, and R.A. Fischer. 1997. Yield potential progress in short bread wheats in northwest Mexico. Crop Science 37:36-42.

Sharma, D.L., and W.K. Anderson. 2004. Small grain screenings in wheat: interactions of cultivars with season, site, and management practices. Australian Journal of Agricultural Research 55:797-809.

Spielmeyer, W., and R.A. Richards. 2004. Comparative mapping of wheat chromosome 1AS which contains the tiller inhibition gene (tin) with rice chromosome 5S. Theoretical and Applied Genetics 109:1303-1310.

van Herwaarden, A.F., G.D. Farquhar, J.F. Angus, R.A. Richards, and G.N. Howe. 1998. 'Haying-off', the negative grain yield response of dryland wheat to nitrogen fertiliser - I. Biomass, grain yield, and water use. Australian Journal of Agricultural Research 49:1067-1081.

Wardlaw, I.F. 2002. Interaction between drought and chronic high temperature during kernel filling in wheat in a controlled environment. Annals of Botany 90:469-476.

Woodruff, D.R., and J. Tonks. 1983. Relationship between Time of Anthesis and Grain-Yield of Wheat Genotypes with Differing Developmental Patterns. Australian Journal of Agricultural Research 34:1-11.

Previous PageTop Of Page