Analysis of aneuploid lines of bread wheat to map chromosomal locations of genes controlling root hair length

Abstract

Background and aims: Long root hairs enable the efficient uptake of poorly mobile nutrients such as phosphorus. Mapping the chromosomal locations of genes that control root hair length can help exploit the natural variation within crops to develop improved cultivars. Genetic stocks of the wheat cultivar 'Chinese Spring' were used to map genes that control root hair length.

Methods: Aneuploid stocks of 'Chinese Spring' were screened using a rapid method based on rhizosheath size and then selected lines were assayed for root hair length to identify chromosomes harbouring genes controlling root hair length. A series of lines with various fractional deletions of candidate chromosomes were then screened to map the root hair loci more accurately. A line with a deletion in chromosome 5A was analysed with a 90 000 single nucleotide polymorphism (SNP) array. The phosphorus acquisition efficiency (PAE) of one deletion line was compared with that of euploid 'Chinese Spring' by growing the seedlings in pots at low and luxury phosphorus supplies.

Key results: Chromosomes 1A, 1D and 5A were found to harbour genes controlling root hair length. The 90 000 SNP array identified two candidate genes controlling root hair length located on chromosome 5A. The line with a deletion in chromosome 5A had root hairs that were approx. 20 % shorter than euploid 'Chinese Spring', but this was insufficient to reduce its PAE.

Conclusions: A rapid screen for rhizosheath size enabled chromosomal regions controlling root hair length to be mapped in the wheat cultivar 'Chinese Spring' and subsequent analysis with an SNP array identified candidate genes controlling root hair length. The difference in root hair length between euploid 'Chinese Spring' and a deletion line identified in the rapid screen was still apparent, albeit attenuated, when the seedlings were grown on a fully fertilized soil.

Keywords: Chinese Spring; Root hairs; Triticum aestivum; aneuploidy; phosphorus acquisition efficiency; rhizosheath.

Fig. 1.

Rhizosheath size of ‘Chinese Spring’ compared with a range of unrelated cultivars. Seedlings were grown for 3 d in a soil with no added mineral nutrients other than the CaCO₃ used to adjust the pH to 6·5. Columns show mean values of six replicates, with the bars indicating standard errors. Asterisks denote ‘Chinese Spring’ and ‘Maringa’ as being significantly different (P < 0·05) from all the other lines as determined with a one-way ANOVA.

Fig. 2.

Rhizosheath sizes of the nullisomic–tetrasomic series of ‘Chinese Spring’ lines. For the nullisomic A and D series, one-way ANOVAs with the Fisher least significant difference (LSD) method was used for pairwise comparisons. Columns denote means, and bars on columns denote standard errors (n = 6). The LSD (P = 0·05) used for the pairwise comparisons is shown in each panel. For the nullisomic B series, data were not normally distributed so a one-way ANOVA on ranks with Dunn’s method for pairwise multiple comparisons was used, and this did not identify any significant differences at P < 0·05. Missing columns (e.g. lines N2AT2B and N2AT2D) indicate lines that were either not viable or had very low fertility (Devos et al., 1999). The black column in each panel denotes euploid ‘Chinese Spring’ (CS). The grey columns are lines that did not differ significantly from euploid ‘Chinese Spring’ while white columns denote pairs of nullisomic–tetrasomic lines that differed significantly (P < 0·05) from euploid ‘Chinese Spring’.

Fig. 3.

Root hair lengths of selected nullisomic–tetrasomic lines that had small rhizosheaths as shown in Fig. 2. The different panels group together lines that were assayed at the same time. The same batch of euploid ‘Chinese Spring’ (CS) grain was used as a control throughout the assays. Columns denote means, and bars on columns denote standard errors (n = 6). Stocks of the N4DT4B line were depleted and only line N4DT4A nullisomic for chromosome 4D could be assayed. Asterisks show lines that were significantly different from euploid ‘Chinese Spring’ for each assay, and the bars on each panel denote the least significant difference (LSD) (P = 0·05) as determined by a one-way ANOVA.

Fig. 4.

Rhizosheath sizes of deletion lines for chromosomes 5A (A) and 1D (B). The rhizosheath size of euploid ‘Chinese Spring’ (CS) was set at 1·0, and rhizosheath sizes of all other lines are expressed as a proportion of the euploid ‘Chinese Spring’ value. The designations given to the lines are as described (Endo and Gill, 1996), and the number at the end of this designation is the percentage of chromosome arm that remains, with the ‘S’ and ‘L’ denoting the short and long arms of the chromosome. The location of the centromere in relation to the various deletions is shown. On the figure, the lines are ordered so that the further away a line is away from the centromere, the smaller is the missing fragment of chromosome. Columns denote means, and bars on columns denote standard errors (n = 6). White columns indicate lines that are significantly different from euploid ‘Chinese Spring’ (black column), grey columns are lines that did not differ from ‘Chinese Spring’ and the bars for each panel denote the least significant difference (LSD) (P = 0·05) as determined by a one-way ANOVA.

Fig. 5.

Root hair lengths of deletion lines for chromosomes 5A (A) and 1D (B). The root hair length of euploid ‘Chinese Spring’ (CS) was set at 1·0, and root hair lengths of all other lines are expressed as a proportion of the euploid ‘Chinese Spring’ value. The lines are described as given in the legend of Fig. 4. White columns indicate lines that are significantly different (P < 0·05) from euploid ‘Chinese Spring’ (black column), grey columns are lines that did not differ from ‘Chinese Spring’. The bar in (B) denotes the least significant difference (LSD) (P = 0·05) as determined by a one-way ANOVA. For (A), the data were not normally distributed, and an analysis of ranks identified all white columns as being significantly different (P < 0·05) from euploid ‘Chinese Spring’ as denoted by the asterisk.

Fig. 6.

Root hairs of line 5AL-23 (white columns) are shorter than those of euploid ‘Chinese Spring’ (CS: black columns) when grown on a fully fertilized kandosol for a short (A: 6 d) and longer (B: 19 d) growth period. For (A), seedlings were grown in rhizoboxes with a low (P5: 5 mg P kg^–1) and high (P150: 150 mg P kg^–1) P treatment, and bars on the columns denote the standard error (n = 3). A two-way ANOVA identified a significant genotype difference (P < 0·01) and no treatment effect. For (B), seedlings were grown in pots with two low (P5 and P10: 5 and 10 mg P kg^–1) and one high (P150: 150 mg P kg^–1) P treatment. A two-way ANOVA identified significant genotype differences at P5 and P10 (P < 0·05) and a P treatment effect for euploid ‘Chinese Spring’ (P < 0·05: P150 shorter than P5).

Fig. 7.

Comparison of the PAE of deletion line 5AL-23 (white columns) with euploid ‘Chinese Spring’ (CS: black columns). Plants were grown in 2 kg pots with the various P treatments (P5 and P10: 5 and 10 mg P kg^–1) for 19 d. Since line 5AL-23 was a smaller plant than euploid ‘Chinese Spring’ at all P treatments, data for both shoot biomass (A) and root length (B) are expressed as values relative to the high P treatment (P150: 150 mg P kg^–1) of each genotype where P was not limiting growth. Columns denote means, and bars on columns denote standard errors (n = 6). Analysis of data with a two-way ANOVA did not find any significant genotypic or treatment differences at P < 0·05 for either relative shoot biomass or relative root length.