Towards map-based cloning of partial resistance QTLs of barley to Puccinia hordei

Partial resistance of barley to Puccinia hordei and near-nonhost resistance to non-adapted rust fungi inherit polygenically. The two types of resistance seem to share some genes and have a similar prehaustorial mechanism of resistance, but partial resistance is less strong than near-nonhost resistance of barley. Partial resistance to adapted, “host”, rust fungi seems, therefore, like a weak form of nonhost resistance to non-adapted rust fungi. If partial resistance and nonhost resistance are indeed based on the same principles, one can understand nonhost resistance by studying partial resistance and vice versa. To study partial and nonhost resistance, as well as their association, the candidate gene(s) for resistance must be cloned and characterized for their action.

Five resistance quantitative trait loci (QTLs) for partial resistance (Rphq2, Rphq3, Rphq4, Rphq11 and Rphq16) and one nonhost resistance QTL (Rnhq) were selected to pursue map-based cloning. First, the effect of the QTLs was verified in near-isogenic lines (NILs). The NILs of Rphq2, Rphq3, Rphq4 and Rnhq (QTL-NILs) were available in L94 genetic background. L94 is extremely susceptible to Puccinia hordei, and, at seedling stage, somewhat susceptible to certain non-adapted rust fungi. The experimental barley line SusPtrit is also susceptible to P. hordei but, at seedling stage, also very susceptible to at least nine species of non-adapted rust fungi. In Chapter 3, we developed NILs in SusPtrit background for Rphq2, Rphq3, Rphq11, Rphq16 and two alleles of Rnhq, viz. L94 and Vada alleles. The effect of each QTL in L94 and SusPtrit genetic backgrounds was tested not only against different isolates of P. hordei but also against different species and isolates of non-adapted rust fungi. The QTL-NILs suggested that the effects of the partial resistance genes depended on rust species and rust isolates. Some introgressions conferred resistance to a broader spectrum of rust species and isolates than others, the broadest being the Rphq11-introgression. The NILs may overestimate the spectrum of effectiveness of the partial resistance genes because some NILs contain inadvertent donor genome in the background and the introgressed QTL region may contain several linked resistance genes, each with a narrow resistance spectrum. The introgression would then confer a resistance spectrum that is the combination of the spectra of several linked resistance genes. Allowing for the possibility of linkage of narrow-spectrum resistance genes, our study suggests that some genes may be involved in partial as well as nonhost resistance. Data also suggest that genetic background may play a role in the resistance conferred by the QTL-introgression.

The NILs also allow fine-mapping of the QTL as was done for Rphq2 in a previous study. In Chapter 4, we target to fine-map another two partial resistance QTLs of our interest, viz. Rphq11 and Rphq16. We, however, did not use the NILs for fine-mapping of Rphq11 and Rphq16. Instead, after validating the effect of Rphq11 and Rphq16 using the early breeding materials for developing NILs of Rphq11 and Rphq16, we developed fixed QTL-recombinants (i.e. homozygous recombinants at the Rphq11/Rphq16 QTL alleles, homozygous susceptible at the non-targeted QTL alleles). The genomic background of fixed QTL-recombinants was still segregating, but expected not to be relevant for the resistance level. Rphq11 was fine-mapped into a 0.2 cM genetic interval and a 1.4 cM genetic interval for Rphq16, before the NILs were ready. The strongest candidate gene for Rphq11 is a phospholipid hydroperoxide glutathione peroxidase (PHGPx). This gene corresponds to the new Rphq11 peak marker – WBE129, located within the refined 0.2 cM genetic intervals and was one of the candidate genes for Rphq11 identified through e-QTL mapping on Steptoe/Morex challenged with the same rust isolate. There was no clear candidate gene identified for Rphq16.

A QTL has to be fine-mapped into a sufficiently narrow genetic window to make physical mapping feasible. Rphq2 with a genetic window of 0.1 cM is ready for physical mapping. In Chapter 5, we have constructed two non-gridded Bacterial Artificial Chromosome (BAC) libraries of barley from Vada and SusPtrit. Based on the insert sizes of the BAC clones, the estimated genome coverage of the Vada BAC library is 2.6x and of the SusPtrit BAC library 3.7x. The genome coverage of Vada is comparable to the BAC library of Morex, HVVMRXALLhB and SusPtrit to HVVMRXALLeA. The estimation of genome coverage based on microsatellite markers indicates, however, Vada and SusPtrit BAC libraries to have 5.0x and 6.8x genome coverage, respectively. Based on genome insert size, the BAC library of Vada gives at least 93% probability of identifying a clone corresponding to any sequence of Vada and for the BAC library of SusPtrit a probability of 98% is expected. Together, the two BAC libraries give more than 99% probability of recovering any specific sequence from the barley genome. A tiling path of three BAC clones was constructed for Vada, which cover the Rphq2 genetic window. The physical window of Rphq2 in Vada BAC contig is approximately 195 Kbp. For SusPtrit, the three BAC clones forming the contig did not cover the entire genetic window of Rphq2. The physical length bridged by them is approximately 226 Kbp. The TriAnnot pipeline annotated 12 genes in both the Vada and the SusPtrit contig, but only four of the annotated genes are shared between Vada and SusPtrit. The candidate genes for Rphq2 might be a resistance factor in Vada or a susceptibility factor in SusPtrit. The peroxidases and kinases are good candidates to represent Rphq2. It is possible that one of the peroxidase or kinase gene members in the physical window of Rphq2 explains the resistance phenotype observed. Another possibility is that peroxidase or kinase gene members function as a complex QTL. A member of the Seven in absentia protein family (SINA) can be a candidate as well. The gene families to which previously cloned genes for partial resistance belong were not found to be represented in the Rphq2 region.

We propose to perform functional analysis of candidate genes through Agrobacterium-mediated stable transformation of the resistance allele into a susceptible genotype, such as SusPtrit. Unfortunately, SusPtrit is, as so many barley accessions, not amenable to Agrobacterium-mediated transformation. In Chapter 2, we developed a doubled haploid (DH) mapping population (n=122) by crossing SusPtrit with Golden Promise to develop a ‘Golden SusPtrit’, i.e., a barley line combining SusPtrit’s high susceptibility to non-adapted rust fungi with the high amenability of Golden Promise for transformation. Using the DH population, we identified nine genomic regions occupied by QTLs against four non-adapted rust fungi and P. hordei isolate 1.2.1 (Ph.1.2.1). From 12 DH lines that were most susceptible to the tested non-adapted rust fungi, we selected four DHs for an Agrobacterium-mediated transformation efficiency test. We obtained a DH line (SG062N) with transformation efficiency of 11 to 17 transformants per 100 immature embryos. The level of susceptibility of SG062N to non-adapted rust fungi is either similar to or higher than the level of susceptibility of SusPtrit. Against P. hordei, the latency period conferred by SG062N at seedling stage is as short as that conferred by SusPtrit. SG062N, designated ‘Golden SusPtrit’, will be a valuable experimental line that could replace SusPtrit in future nonhost and partial resistance studies, especially for stable transformation using candidate genes that may determine the differences in resistance levels against adapted and non-adapted rust fungi.