Susceptibility pays off: insights into the mlo-based powdery mildew resistance

Powdery mildew (PM) is a worldwide-occurring plant disease caused by ascomycete fungi of the order Erysiphales. A conspicuous number of plant species are susceptible to this disease, the occurrence of which is increasing due to the influence of climate change. Symptoms are easy to recognize by the powdery whitish fungal structures growing on the surface of plant organs. Severe infections cause significant losses in crops, such as tomato, cucumber and wheat, as well as in ornamentals, like rose and petunia. Accordingly, breeding crops with a robust immunity to this disease is of great economic importance. A significant step in this direction was the discovery of <em>mlo</em> (mildew locus o) mutant alleles of the barley <em>HvMlo</em> gene, which are responsible for the non-race specific resistance to the barley PM pathogen, <em>Blumeria graminis</em> f.sp. <em>hordei </em>(<em>Bgh</em>). During the years, this recessively inherited resistance was observed to be durable, contrary to the short life-span of resistances conferred by dominant resistance (R-) genes used in barley breeding programs. Studies on the histological mechanisms of the <em>mlo</em>-based resistance showed that the PM pathogen was stopped during penetration of the cell wall by the formation of a papilla. This structure prevents the formation of the feeding structure of the pathogen, called a haustorium. After sequencing many plant genomes, we are discovering that <em>MLO</em> genes are not only typical of this cereal, but are ubiquitously present in higher plant species in multiple copies per species, forming a gene family. The impairment of some members of a number of ever increasing plant species lead to broad-spectrum resistance towards their adapted PM pathogens. For example, in tomato the <em>ol-2</em> gene, naturally harbored by the cherry tomato <em>Solanum lycopersicum</em> var. <em>cerasiforme</em>, represents the loss-of-function allele of the <em>SlMLO1</em> gene, conferring resistance to the PM pathogen <em>Oidium neolycopersici </em>(<em>On</em>). Consequently, the use of <em>mlo </em>mutants represents a suitable alternative to the classical use of R-genes in breeding programs. In <strong>Chapter 2</strong>, we describe the <em>in silico</em> identification of the complete tomato <em>SlMLO</em> gene family using the available information in the SOL genomic network database. In total, 16 tomato <em>SlMLO</em> members were cloned from leaf, root, flower and fruit of the susceptible tomato cv. Moneymaker to confirm the sequences retrieved from the database and to verify their actual expression in these tissues. We observed the presence of various types of splicing variants, although their possible functional meaning has not been investigated. Motif analyses of each of the translated protein sequences and phylogenetic studies highlighted, on one hand, amino acid stretches that characterize the whole MLO family, and, on the other hand, stretches conserved in MLO homologs that are phylogenetically related. Following a gene expression study upon <em>On</em> inoculation, we identified members of the <em>SlMLO</em> family that are upregulated few hours after pathogen challenge. Except <em>SlMLO1</em>, none of the three newly identified homologs in clade V, thus phylogenetically close to <em>SlMLO1</em>, are induced. Interestingly, two homologs, each found in different clades, are upregulated similarly to <em>SlMLO1</em>. Using an RNAi approach, we silenced the additional clade V-<em>SlMLO</em> homologs, namely <em>SlMLO3</em>, <em>SlMLO5</em> and <em>SlMLO8</em>, to investigate their possible role in PM resistance. We observed that none of these homologs if individually silenced, leads to PM resistance. However, if <em>SlMLO5</em> and <em>SlMLO8</em> are silenced together with <em>SlMLO1</em>, a significantly higher level of resistance is achieved compared to plants carrying the <em>ol-2</em> allele. The role of <em>SlMLO3</em> could not be verified. We, therefore, concluded that there are three <em>SlMLO</em> genes in tomato unevenly contributing to the PM disease, of which <em>SlMLO1</em> has a major role. <strong>Chapter 3</strong> focuses on the components of the tomato <em>mlo</em>-based resistance. In Arabidopsis, it is known that four members of the SNARE protein family, involved in membrane fusion, are involved in <em>mlo</em>-based resistance. In this chapter, we focused on the identification of tomato homologs of the Arabidopsis syntaxin PEN1 (AtSYP121). Among the group of syntaxins identified in tomato, two were closely related to each other and also to <em>AtPEN1</em>, denominated <em>SlPEN1a</em> and <em>SlPEN1b</em>. Another Arabidopsis syntaxin that shows a high level of homology with <em>PEN1</em>, called <em>SYP122</em>, was also found to group together with the newly identified <em>SlPEN1</em> genes. However, the role of <em>SYP122 </em>in plant immunity was not shown in literature. After obtaining individual silencing RNAi constructs, we transformed the resistant <em>ol-2</em> line, and we challenged the obtained transformants with the adapted PM <em>On</em>, and the non-adapted <em>Bgh</em>. Interestingly, we observed a significant <em>On</em> growth and an enhanced <em>Bgh</em> cell entry only in <em>SlPEN1a</em> silenced plants but not in <em>SlPEN1b</em> silenced ones. We performed a protein alignment of tomato and Arabidopsis functional and non-functional PEN sequences. The presence of three differently conserved non-synonymous amino-acid substitutions is hypothesised to be responsible for the specialization in plant immune function. In <strong>Chapter 4</strong> and <strong>Chapter 5,</strong> we build up a body of evidence pointing to the fact that the function of the <em>MLO</em> susceptibility genes is highly conserved between monocot and dicot plant species. In <strong>Chapter 4</strong> we started by identifying and functionally characterizing two new <em>MLO</em> genes of Solanaceous crops affected by the PM disease, tobacco (<em>Nicotiana tabacum</em>) and eggplant (<em>Solanum melongena</em>). We named them <em>NtMLO1 </em>and <em>SmMLO1</em> in the respective species, as they are the closest homologs to tomato <em>SlMLO1</em>. By overexpressing these genes in the resistant <em>ol-2</em> line, we obtained transgenic plants that were susceptible to the PM pathogen <em>On</em>. This finding demonstrates that both heterologous MLO proteins can rescue the function of the impaired <em>ol-2 </em>allele in tomato. In addition, we found in tobacco NtMLO1 an amino acid (Q198) of critical importance for the susceptibility function of this protein. In <strong>Chapter 5</strong>, we used the same approach adopted in Chapter 4 to show that other MLO proteins of more distant dicot species, like pea PsMLO1, can rescue the loss-of-function of the tomato <em>ol-2</em> allele. And finally, we stretched this concept also to monocot MLO proteins, using barley HvMlo. While performing these experiments, we could verify that the function of the monocot and dicot susceptibility MLO proteins does not rely on the presence of class-specific conservation. The latter can be the reason for the phylogenetic divergence, placing monocot MLO proteins in clade IV and dicot MLO proteins in clade V of the phylogenetic MLO tree. However, functional conservation might depend on crucial shared amino acids of clade IV and V MLO proteins. Therefore, we also conducted a codon-based evolutionary analysis that resulted in the identification of 130 codons under negative selection, thus strongly maintained during evolution. In <strong>Chapter 6</strong> we introduce the PM disease in cucumber caused by <em>Podosphaera</em> <em>xanthii </em>(<em>Px</em>). We cloned the candidate susceptibility gene for PM in cucumber, <em>CsaMLO8</em>, from susceptible and resistant genotypes. The latter was described as an advanced cucumber breeding line characterized by hypocotyl resistance. In this line, we found the presence of aberrant splicing variants of the <em>CsaMLO8</em> mRNA due to the insertion in its corresponding genomic region of a Class LTR retrotransposon. Heterologous expression of the wild-type cucumber allele in the tomato <em>ol-2</em> line restored its PM susceptibility, while the heterologous expression of the aberrant protein variant failed to do so. This finding confirms that the resistance of the advanced cucumber breeding line is due to the disruption of the coding region of this gene. We also showed that the expression of <em>CsaMLO8</em> in the susceptible genotype is induced by <em>Px</em> in hypocotyl tissue, but not in cotyledon or leaf. Finally, by examination of the resequencing data of a collection of 115 cucumber accessions, we found the presence of the TE-containing allele in 31 of them among which a wild cucumber accession that might have been used in breeding programs to obtain resistance to the PM disease in cucumber. In <strong>Chapter 7</strong> a novel loss-of-function allele of the <em>SlMLO1</em> gene is described, designated <em>m200</em>. This allele was found in a resistant plant (M200) from a mutagenized tomato Micro-Tom (MT) population obtained with the chemical mutagen ethyl methanesulfonate (EMS). The <em>m200</em> mutation corresponds to a nucleotide transversion (T à A) which results in a premature stop codon. The length of the predicted SlMLO1 protein in the M200 plant is only 21 amino acids, thus much shorter than the predicted protein of the previously described <em>ol-2</em> allele, consisting of 200 amino acids. Thanks to the development of a High-Resolution Melting (HRM) marker designed to detect the <em>m200</em> mutation, we observed that this allele confers recessively inherited resistance in backcross populations of the resistant M200 plant with MT and Moneymaker. Histological study showed that the resistance of the <em>m200</em> mutant is associated with papilla formation. Finally, we compared the rate of <em>On</em> penetration in epidermal cells of <em>m200</em> plants with the one of plants carrying the <em>ol-2</em> allele and the transgenic plants in which multiple <em>SlMLO</em> homologs were silenced, generated in Chapter 2. Ultimately, in <strong>Chapter 8</strong> the results of the previous chapters are discussed in the context of 1) practical applications in breeding programs aimed at introducing the <em>mlo</em>-based resistance in new crops, 2) possible research aimed at unraveling the function of the MLO protein and 3) the role of other SNARE proteins.