The novel avirulence effector AlAvr1 from Ascochyta lentis mediates host cultivar specificity of ascochyta blight in lentil.

Ascochyta lentis is a fungal pathogen that causes ascochyta blight in the important grain legume species lentil, but little is known about the molecular mechanism of disease or host specificity. We employed a map-based cloning approach using a biparental A. lentis population to clone the gene AlAvr1-1 that encodes avirulence towards the lentil cultivar PBA Hurricane XT. The mapping population was produced by mating A. lentis isolate P94-24, which is pathogenic on the cultivar Nipper and avirulent towards Hurricane, and the isolate AlKewell, which is pathogenic towards Hurricane but not Nipper. Using agroinfiltration, we found that AlAvr1-1 from the isolate P94-24 causes necrosis in Hurricane but not in Nipper. The homologous corresponding gene in AlKewell, AlAvr1-2, encodes a protein with amino acid variation at 23 sites and four of these sites have been positively selected in the P94-24 branch of the phylogeny. Loss of AlAvr1-1 in a gene knockout experiment produced a P94-24 mutant strain that is virulent on Hurricane. Deletion of AlAvr1-2 in AlKewell led to reduced pathogenicity on Hurricane, suggesting that the gene may contribute to disease in Hurricane. Deletion of AlAvr1-2 did not affect virulence for Nipper and AlAvr1-2 is therefore not an avirulence gene for Nipper. We conclude that the hemibiotrophic pathogen A. lentis has an avirulence effector, AlAvr1-1, that triggers a hypersensitive resistance response in Hurricane. This is the first avirulence gene to be characterized in a legume pathogen from the Pleosporales and may help progress research on other damaging Ascochyta pathogens.

INTRODUCTION

Fungal plant pathogens have evolved diverse lifestyles and molecular mechanisms to support colonization of their host plant species. Pathogen survival depends on the ability of the organism to modify the host to facilitate the acquisition of nutrients required for growth and reproduction. Broadly, necrotrophic fungal pathogens kill host tissues and extract nutrients from dead host cells, and biotrophic fungi colonize and obtain nutrients from living tissue in a parasitic relationship with the host (Dangl & Jones, 2001; Vleeshouwers & Oliver, 2014). Hemibiotrophic fungi display two phases during plant infection, with an initial biotrophic growth stage followed by a switch to necrotrophic growth (Lo Presti et al., 2015; Vleeshouwers & Oliver, 2014).

Plant–pathogen interactions are mediated primarily by secondary metabolite and protein effectors (Friesen et al., 2008; Muria‐Gonzalez et al., 2015; Rouxel & Balesdent, 2010; de Wit, 1997; de Wit et al., 2009). Two broad classes of effectors act in different ways. Avirulence effectors induce hypersensitive responses (HRs) in resistant host plants. Localized necrosis at the site of pathogen attack restricts further growth of the invading organism (de Wit, 1997; de Wit et al., 2009). This process has been described as effector‐triggered immunity (ETI) (Jones & Dangl, 2006) and fits the canonical model of gene‐for‐gene interaction in plant disease resistance described by Flor (1971). Avirulence effectors have been found in all biotrophs examined but are rare in necrotrophs. Generally, for necrotrophs, model dynamics are inverted such that the interaction of virulence factors from the pathogen and susceptibility genes in the host initiate effector‐triggered susceptibility (ETS) rather than ETI. Here, secreted fungal effector molecules diffuse through the apoplast and interact with host receptors to promote disease by inducing necrosis throughout infected tissues (Faris et al., 2010; Liu et al., 2009, 2012).

The Dothideomycetes class is a large and diverse group of fungal species and includes a wide range of plant pathogens, many of which are pathogenic on important crop species (Ohm et al., 2012). Within the Dothideomycetes, the order Pleosporales, suborder Pleosporineae includes the notable plant pathogen species Parastagonospora nodorum, Pyrenophora teres, Pyrenophora tritici‐repentis, Cochliobolus spp., Leptosphaeria maculans, and Ascochyta spp., among others (Zhang et al., 2009). These are narrow host range pathogens that invade through the epidermis and cause necrosis to kill mesophyll cells. Genome sequencing has facilitated the study of molecular mechanisms of pathogenicity for many of these species (Condon et al., 2013; Hane et al., 2007; Moolhuijzen et al., 2018; Rouxel et al., 2011; Syme et al., 2013, 2018). Ascochyta lentis (teleomorph: Didymella lentis, syn. Ascochyta fabae f. sp. lentis) is the causal organism of ascochyta blight in lentil (Lens culinaris) (Kaiser et al., 1997), which is an important foliar disease of lentil worldwide. The disease causes considerable reduction in grain quality and yield due to stem girdling, flower and pod abortion, and seed staining (Gossen & Morrall, 1983). A. lentis is from the family Didymellaceae, within the Pleosporales, and provides an additional plant disease model to research molecular mechanisms of pathogen–host interaction that contrasts with the pathogens from cereal and oilseed crops described above. A near‐complete genome assembly for A. lentis has been recently published (Lee et al., 2021), and agroinfiltration, genetic transformation, and gene knockout protocols are available for research in the lentil–ascochyta pathosystem (Debler & Henares, 2020; Debler et al., 2021; Henares et al., 2019). However, so far very limited information is available about the pathogen–host interaction at the molecular level.

Shifts in the virulence or aggressiveness of A. lentis populations have contributed to resistance breakdown of cultivars in Canada (Ahmed et al., 1996) and Australia (Davidson et al., 2016). In Australia, an increase in aggressiveness of sexually recombining A. lentis isolates towards the lentil cultivar Nipper was observed for isolates collected from 2005 to 2014. This gradual shift towards increased pathogenicity caused the breakdown of resistance in the widely grown Australian cultivar within 4 years of its release (Davidson et al., 2016).

Here we report the identification of the first effector gene from A. lentis from map‐based cloning using a biparental fungal population. The population was derived from isolates P94‐24 and AlKewell, which differ in virulence towards lentil cultivars PBA Hurricane XT (abbreviated to Hurricane herein) and Nipper. Identification of this novel effector can be used as a tool to accelerate lentil breeding and improve lentil cultivar selection for better disease management. Furthermore, knowledge about this avirulence mechanism in lentil ascochyta blight will help to inform studies of other Ascochyta species, with their respective legume hosts, including field pea, faba bean, and chickpea, being important sources of plant‐based protein.

RESULTS

Genetic mapping reveals a Hurricane pathogenicity quantitative trait locus on chromosome 3

A. lentis isolates AlKewell and P94‐24 were selected for the study, having different virulence profiles towards lentil cultivars Nipper and Hurricane, as summarized in Table 1. P94‐24 was virulent on Nipper and avirulent on Hurricane, while AlKewell had the converse virulence profile. A biparental A. lentis mapping population (N = 100) was produced to investigate the genetic mechanisms of virulence in A. lentis. Biparental progeny were screened for disease response on lentil cultivars Nipper, Hurricane, and the susceptible control cultivar ILL6002. Average Nipper and Hurricane responses were normalized to the ILL6002 response for each progeny and parental strain, and are presented in Figure 1a. Normalization of disease scores to the susceptible control cultivar corrected for variability in temperature and humidity conditions between experiments that affect disease progression and final disease assessment. For the six experiments, average scores for ILL6002 were 29%, 49%, 22%, 8.3%, 45%, and 51% with P94‐24, and with AlKewell, 20%, 36%, 22%, 7.5%, 38%, and 35%. Disease scores for Experiment 4 were low for both strains and in making the assumption that the reduced disease in this experiment would be uniform for both parental and progeny strains, normalization was considered the best strategy to correct for such variation. A. lentis parent P94‐24 had the highest level of disease on Nipper and no disease on Hurricane, and AlKewell exhibited the reciprocal host specificity. Biparental strains exhibited a wide range of normalized disease scores for both host cultivars and the distribution of mean scores is represented as a histogram (Figure 1b). Regression analysis of the progeny found that there was no linear correlation between normalized scores for Hurricane and Nipper (Figure 1c), with an R 2 value of 0.084. When setting a threshold for virulence at the normalized disease score of 0.1, for the 100 biparental strains, 41 were pathogenic on only Hurricane, 26 strains were pathogenic on only Nipper, 14 strains were pathogenic on both cultivars, and 16 strains were not pathogenic on either host. The chi‐squared test for Hurricane found that the segregation ratio for avirulence to virulence was not significantly different to 1:1 (χ2 = 1, p = 0.32) but for Nipper the segregation ratio was significantly different from 1:1 (χ2 = 4, p = 0.046) at the significance level p < 0.05. The chi‐squared tests suggest a single gene might be responsible for disease response in Hurricane but that the response in Nipper may be more complex.

TABLE 1

Virulence of Ascochyta lentis isolates on lentil cultivars Nipper and PBA Hurricane XT, and the susceptible control, cv. ILL6002

Cultivar	Year introduced	A. lentis isolate
		P94‐24	AlKewell
Nipper	2006	V	Av
PBA Hurricane XT	2015	Av	V
ILL6002	NA	V	V

Abbreviations: Av, avirulent; V, virulent; NA, not applicable as ILL6002 is a breeding line.

FIGURE 1

Disease scores for a segregating Ascochyta lentis mapping population. (a) Normalized disease scores for PBA Hurricane XT and Nipper infected with A. lentis isolates AlKewell and P94‐24 (boxed), and biparental progeny at 14 days after inoculation. Error bars show the standard error for four biological replicates. A horizontal line at 0.1 represents an arbitrary threshold below which strains were considered to be avirulent. (b) Distribution histogram of normalized disease scores for biparental progeny. (c) Linear regression analysis of biparental progeny for normalized Nipper versus Hurricane disease scores

Genotyping of the P94‐24 × AlKewell A. lentis population (N = 95) using double‐digested restriction site‐associated DNA sequencing (ddRADseq) yielded 587 high‐quality markers after filtering (Table S1), and these were mapped to the published reference genome assembly for isolate Al4 (Lee et al., 2021). A genotype marker map is provided in Figure S1. Quantitative trait locus (QTL) mapping for normalized Hurricane disease scores identified a single QTL on A. lentis chromosome 3 (logarithm of odds [LOD] score = 17, LOD threshold at p < 0.005 = 4.71; Figure 2). There was no genetic association of ILL6002 disease scores in the biparental study and the normalized Nipper disease produced only minor, nonsignificant QTL peaks (Figure 2).

FIGURE 2

Quantitative trait locus (QTL) analysis by linear mixed model for ascochyta blight disease scores for the susceptible control lentil cultivar ILL6002, and for Nipper and PBA Hurricane XT. Logarithm of odds (LOD) thresholds for Nipper (dashed line) and Hurricane (dotted line) are shown. A significant QTL with LOD score of 17.0 (LOD threshold, p < 0.005 = 4.71) was found on Ascochyta lentis chromosome 3

The region on A. lentis Al4 chromosome 3 from 2.14 Mb to the end of the chromosome at 2.317 Mb where the major QTL was identified contains 28 annotated genes (Figure 3a). Four of these genes encode proteins with a secretion signal peptide and one of these (g2688) was predicted to encode an effector protein (EffectorP 2.0 probability score 0.77). The region of chromosome 3 from 2.2 to 2.75 Mb contains repetitive AT‐rich DNA sequences as shown in Figure 3a where the GC content graph (blue) averages 35% and contains long terminal repeat and other transposable elements (Lee et al., 2021). A single gene sequence of 738 bp with 52% GC content was annotated as g2688 in the reference genome assembly at chromosome 3 position 2,243,993 to 2,243,255.

FIGURE 3

Gene features for Ascochyta lentis g2688. (a) Genomic context of the g2688 gene on A. lentis chromosome 3 (red box). Genes (green) and coding sequences (CDS) (yellow) are indicated along the region of the chromosome. The GC content graph below the chromosome shows GC content (green) and AT content (blue). (b) Alignments for AlKewell and P94‐24 forms of g2688, respectively, and corresponding protein sequences. Nucleotide identity and polymorphisms between isolates are shown above the alignments as green and white bands, respectively. cDNA and protein diagrams for each isolate are shown below as grey bars on which nucleotide and amino acid changes are shown with nucleotide and amino acid‐specific colours. Signal peptides are indicated with a green box. Cysteine residues are indicated with blue boxes and four amino acid sites under positive selection are indicated with red boxes. Asterisks at around the 450 bp position show amino acid residues potentially mutated as a consequence of repeat‐induced point mutation

In silico analysis of candidate effector g2688 reveals a small cysteine‐rich protein that is under positive selection

The gene (g2688) associated with the QTL encodes a protein that is 230 amino acid residues in length and has a single intron of 48 bp confirmed by cDNA sequencing (Figure 3b). It has a signal peptide of 25 amino acid residues and the mature polypeptide of 205 residues has a calculated molecular mass of 22.5 kDa, an isoelectric point of 5.5, and nine cysteine residues (4.4%). The g2688 ortholog from P94‐24 has 28 single‐nucleotide polymorphisms (SNPs) over the 738 bp of the predicted gene compared with AlKewell. The start and stop translation codons, and the intron and intron boundaries were unaffected by nucleotide differences between AlKewell and P94‐24. There were only two SNPs in the 500 bp upstream from the transcription start site, indicating that the promoter region is probably unchanged between isolates. Figure 3b shows the nucleotide and amino acid alignments for the putative effector gene. Additional detail about nucleotide substitutions that produce the amino acid changes is provided in Figure S2. There were 23 amino acid polymorphisms between isolates resulting from the 28 SNPs, which amounts to 89% amino acid sequence identity for the mature polypeptide. To assess positive selection for the gene, we applied pairwise comparison of calculated nonsynonymous and synonymous mutations. For P94‐24 and AlKewell full‐length g2688 cDNAs, the pairwise dN/dS ratio was 9.2.

Using tBLASTn searches of the NCBI nonredundant nucleotide sequence database and Ascochyta species genome assemblies, we identified five sequences with around 25% amino acid identity. This included homologous sequences in three Colletotrichum species, L. maculans, and a distant homolog in Fusarium oxysporum f. sp. apii (Texts S1 and S2). A homologous sequence was found in the Ascochyta viciae‐villosae (GenBank accession GCA_004335205.1). No homologous sequences were found for other Ascochyta species such as A. viciae (GCA_004335155.1), A. fabae (GCA_004335285.1) or A. rabiei (GCA_004011695.1). A neighbour‐joining tree of the protein sequences (Figure S3) shows three relatively diverged clusters of genes among small numbers of species from the Dothideomycete and Sordariomycete plant‐pathogenic fungi. The A. viciae‐villosae cDNA has 41 and 40 nucleotide polymorphisms when compared to the g2688 cDNAs from P94‐24 and AlKewell, respectively, producing 27 and 28 amino acid polymorphisms between respective encoded proteins (Figure S4). Pairwise dN/dS ratios for the A. viciae‐villosae g2688 homolog and the two forms of the gene from A. lentis were 1.3 for P94‐24 and 1.4 for AlKewell. Using the three homologous cDNA sequences from A. lentis (two) and A. viciae‐villosae (one), we investigated selection pressure using programs from the HyPhy webserver (Kosakovsky Pond et al., 2005) (Table S2) and four amino acid sites were identified as being under positive selection. These are indicated by red boxes in Figure 3b.

Analysis of the two A. lentis g2688 cDNA sequences using RIPCAL (Hane & Oliver, 2008), to assess the effects of repeat‐induced point (RIP) mutation, showed that of the 28 nucleotide polymorphisms between strains, 13 mutations are CpN → TpN transitions that could have resulted from an RIP mutation process, and three of those 13 transitions are CpA → TpA changes particularly characteristic of RIP mutation. The directional nature of RIP mutation allows us to infer which strain was potentially subject to RIP mutation. In the majority (8 of 13) of CpN → TpN transitions, the P94‐24 sequence was the site of mutation. Two of the four amino acid changes that were at positively selected sites may have resulted from RIP mutation and in both cases it was the P94‐24 sequence subjected to the C → T change. The two adjacent positively selected sites that were potentially mutated by RIP are shown in Figure 3b at nucleotide positions 448–455. Here, two alanine residues in AlKewell were changed to an arginine and an asparagine residue in P94‐24.

Functional characterization of g2688 suggests the P94‐24 isoform is an avirulence gene

The functional roles of the A. lentis g2688 effector isoforms were investigated by transiently expressing the proteins in lentil cultivars using agroinfiltration (Figure 4). The host cultivar responses shown are representative of at least three replicates and additional images for other replicates are provided in Figure S5. Agroinfiltration with a negative control, green fluorescent protein (GFP), and a positive necrosis control, necrosis‐inducing effector nep1‐like protein (NLP) (Bailey, 1995; Debler et al., 2021), produced the expected responses in the lentil cultivars tested.

FIGURE 4

Functional characterization of P94‐24 and AlKewell g2688 isoforms, labelled as g2688‐P and g2688‐K, respectively, by agroinfiltration in different lentil cultivars. Representative images of agroinfiltrated leaves from three independent experiments, each with three technical replicates, observed at 5 days postinfiltration

Agrobacterium‐mediated expression of g2688 isoforms in Hurricane and Nipper leaflets showed that Hurricane is sensitive to g2688 from P94‐24 and neither Hurricane nor Nipper are sensitive to the AlKewell isoform. Agroinfiltration demonstrated that g2688 from P94‐24 is indeed a necrosis‐inducing effector, but with Hurricane being resistant to the P94‐24 isolate, our hypothesis is that g2688 from P94‐24 is an avirulence‐type effector that triggers plant defence in the resistant cultivar by interacting with a resistance gene in Hurricane. With this evidence we have named the A. lentis effector AlAvr1 and the two forms of the gene as AlAvr1‐1 (P94‐24) and AlAvr1‐2 (AlKewell). Naming of the genes as such allows for the possibility that variant forms with cultivar‐specific avirulence reactions may be discovered in future research. The corresponding protein names are nonitalicized. We have not used the lower case avr that has been used elsewhere (Plissonneau et al., 2018) to denote alleles that mediate virulence, with the anticipation that lentil genotypes in which AlAvr1‐2 has an avirulence function may be identified at some stage. Isolates that do not have the P94‐24 isoform of AlAvr1, such as AlKewell, do not induce the resistance response and are virulent on Hurricane. The lentil cultivar PBA Bolt displayed the same sensitivity to AlAvr1‐1 (Figure 4) and possibly carries the same resistance gene as Hurricane, with PBA Bolt and Hurricane having a common parent, the ascochyta‐resistant breeding line 96‐047L*99R060 (Agriculture‐Victoria‐Services, 2013; Sudheesh et al., 2016). Disease assays with parental A. lentis isolates AlKewell and P94‐24 showed that PBA Bolt was resistant to P94‐24 but was susceptible to AlKewell, similar to Hurricane (Figure S6). Neither form of AlAvr1 induced a necrosis response in the non‐host Nicotiana benthamiana. The fact that AlAvr1‐2 from AlKewell does not induce an HR in either Hurricane or Nipper suggests that the gene is not associated with avirulence towards either of these cultivars. None of the other cultivars agroinfiltrated with AlAvr1‐1 or AlAvr1‐2 exhibited a necrosis response.

Gene expression analysis of AlAvr1 reveals that both isoforms are expressed early during infection

Reverse transcription‐quantitative PCR (RT‐qPCR) was used to determine gene expression over the early stages of infection until 7 days after inoculation. Figure 5 shows that AlAvr1‐1 and AlAvr1‐2 in isolates P94‐24 and AlKewell, respectively, were expressed at high levels on both lentil cultivars across the early stages of infection, prior to the onset of necrosis at 7 days after inoculation. Expression of AlAvr1‐2 by AlKewell (Figure 5a) was approximately 5‐fold the actin expression level within 24 h after inoculation and expression on Nipper was similar to Hurricane at 24 h and 3 days postinoculation. The expression of AlAvr1‐2 in AlKewell on Nipper peaked at around 100‐fold the actin level from 3 to 5 days and on Hurricane the peak was at 3 days. Expression then was reduced towards 7 days for both cultivars. Expression of AlAvr1‐1 in P94‐24 followed a similar expression profile across the first 7 days of the infection cycle as for AlKewell (Figure 5b), although expression on both Nipper and Hurricane was earlier and higher. The highest level of AlAvr1‐1 expression in P94‐24 was 1000‐fold the actin level at 3 days after inoculation on Nipper.

FIGURE 5

Gene expression of AlAvr1‐2 in AlKewell (a) and AlAvr1‐1 in P94‐24 (b) during infection of lentil PBA Hurricane XT and Nipper. The expression level of AlAvr1‐1 and AlAvr1‐2 in respective wild‐type (WT) isolates during the course of infection was examined using reverse transcription‐quantitative PCR. Quantitative weight measurement was based on a standard curve, and the level of gene expression was normalized to actin gene expression. Error bars represent standard error of the mean from three biological replicates. Data is plotted on a log10 scale. Results not connected by the same letter are significantly different (Student's t test, p < 0.01 (a) and p < 0.001 (b))

Gene knockout studies confirm that AlAvr1‐1 acts as an avirulence gene and AlAvr1‐2 contributes to virulence on Hurricane

The effector functions of AlAvr1‐1 and AlAvr1‐2 were further investigated using gene knockout to produce mutant strains deficient for each of the AlAvr1 effector forms in their native strain. Wild‐type (WT) and mutant strains were grown on half‐strength potato dextrose agar (PDA) to compare growth rate, morphology, and sporulation to demonstrate that knockout of respective genes did not affect growth and developmental phenotypes (Figure S7). Results from nine replicate inoculated plants for each strain × host combination are shown in Figure 6. AlAvr1‐2 gene deletion in AlKewell reduced pathogenicity on Hurricane by approximately 40% compared with WT AlKewell, whereas the deletion did not alter virulence of AlKewell towards Nipper (Figure 6a). Conclusive evidence for the avirulence function of AlAvr1‐1 was found for disease assays of P94‐24 WT and the corresponding gene‐deleted ΔAlAvr1‐1 mutant. Figure 6b shows gain of virulence for the P94‐24 mutant on Hurricane when the AlAvr1‐1 gene was deleted. This is the first evidence of an avirulence gene that correlates with different responses on lentil hosts, and we propose two different pathotypes: pathotype 1 isolates, which harbour AlAvr1‐1 that is recognized by lentil lines with the corresponding resistance gene that interacts with AlAvr1‐1, and pathotype 2, which lacks AlAvr1‐1.

FIGURE 6

Pathogenicity of AlAvr1 gene knockout strains. (a) Infection of PBA Hurricane XT and Nipper lentil with AlKewell wild‐type (WT) and the AlAvr1‐2 knockout mutant ΔAlAvr1‐2 in the AlKewell background (b) P94‐24 WT and the AlAvr1‐1 knockout mutant ΔAlAvr1‐1 in the P94‐24 background. Results are reported as the mean of % leaf area damage from nine individual plants scored 14 days postinfection and the data were square‐root‐transformed. Results not connected by the same letter are significantly different (Tukey HSD test, p < 0.0001)

DISCUSSION

The development of new crop cultivars with improved genetic resistance against pathogens invariably drives changes in pathogen populations. New adapted strains can emerge through mutation of effector genes, and subsequent selection of isolates that can overcome the genetic resistance of the new cultivar. L. maculans provides a good model for the investigation of avirulence effectors in Dothideomycete plant pathogens as described here for the lentil ascochyta blight pathogen, A. lentis.

L. maculans differs from the majority of necrotrophic fungal pathogens in that it has been shown to encode avirulence effectors, with seven avirulence effectors characterized to date (Rouxel & Balesdent, 2010). The genome of L. maculans has an unusually high content of AT‐rich and repetitive DNA regions that are populated with transposable elements (Rouxel et al., 2011). These features are similarly found in A. lentis (Lee et al., 2021). The mobile nature of transposable elements and the propensity of repetitive DNA regions in fungal genomes to undergo RIP mutations contributes to the high rates of mutation of effector genes that are frequently located in these genomic regions (Oliver, 2012; Raffaele & Kamoun, 2012). The primary modes of mutation in L. maculans avirulence genes that lead to loss of host resistance include complete gene deletion, gene inactivation due to RIP mutations that introduce stop codons, and the accumulation of other mutations that change the amino acid sequence and specificity of the avirulence protein (Gout et al., 2006; Rouxel & Balesdent, 2017; Rouxel et al., 2011).

Genetic analysis of A. lentis strains with different host preferences in the current study reveals a probable case of host resistance‐mediated gene selection. The isolate P94‐24 represents Nipper‐virulent isolates and with the characterization of a molecular mechanism for avirulence in the current study these are now described as pathotype 1. AlKewell is an isolate with virulence on Hurricane and is described as pathotype 2. QTL analysis of the P94‐24 × AlKewell mapping population identified the effector gene AlAvr1‐1 in P94‐24 and functional studies showed that the protein induces a necrosis or HR in Hurricane but not in Nipper. We speculate that the hypersensitivity to AlAvr1‐1 protein in Hurricane is the result of interaction with a presumed resistance gene in Hurricane, and that this putative resistance gene confers resistance to pathotype 1 isolates. Key evidence for the role of the AlAvr1‐1 protein in the resistance of Hurricane towards pathotype 1 isolates is that gene deletion of AlAvr1‐1 confers virulence of the mutant P94‐24 strain on Hurricane.

The lentil cultivar Nipper was introduced in Australia in 2006 as a highly ascochyta blight‐resistant cultivar and was widely grown until becoming increasingly susceptible to A. lentis isolates with progressively higher levels of aggressiveness (Davidson et al., 2016). In 2015 Hurricane was released as a moderately ascochyta‐resistant cultivar but within a year, aggressive A. lentis isolates were collected from Hurricane crops (Blake et al., 2017). The virulence profile of these isolates was the same as for the A. lentis isolate AlKewell that was collected in 2001 (Blake et al., 2019). Pathotype 2 A. lentis isolates such as AlKewell have the alternative allelic homolog of AlAvr1, AlAvr1‐2, that does not induce necrosis or hypersensitivity in agroinfiltration functional assays in the lentil cultivars tested. Gene deletion of AlAvr1‐2 did not change the host preference of AlKewell, with neither WT nor mutant strain being able to infect Nipper. However, loss of AlAvr1‐2 in the AlKewell deletion mutant did reduce pathogenicity of the strain and this result suggests that the AlAvr1‐2 effector has a function in promoting disease. The avirulence protein NIP1 from Rynchosporium secalis elicits avirulence in barley cultivars that carry the Rrs1 resistance gene and also acts as a toxin to promote disease in susceptible rrs1 barley genotypes (Hahn et al., 1993; Knogge, 1996; Rohe et al., 1995). The avirulence gene from Cladosporium fulvum, Avr5, similarly acts as an avirulence effector in resistant host cultivars and as a virulence factor in tomato cultivars that lack the Cf‐5 resistance gene (Mesarich et al., 2014). NIP1 and Avr5 represent a type of effector mechanism that acts in two different modes, and A. lentis AlAvr1 appears to function in a similar way.

AlAvr1‐1 has a clear avirulence function in Hurricane and in addition both AlAvr1‐1 and AlAvr1‐2 were shown to contribute positively to disease in minor ways, with pathogenicity of gene‐deletion mutants being lower for both ΔAlAvr1‐1 and ΔAlAvr1‐2 than respective WT strains. Deletion of AlAvr1‐1 in P94‐24 produced a slight but not significant reduction in pathogenicity on Nipper. While AlAvr1‐1 deletion eliminated avirulence on Hurricane, pathogenicity was not fully restored to the WT level observed for P94‐24 on Nipper. Deletion of AlAvr1‐2 in AlKewell significantly reduced pathogenicity of the mutant strain on Hurricane, consistent with a role as a toxin, but did not alter the disease response of Nipper.

The avirulence of AlKewell on Nipper suggests that there is an avirulence mechanism that mediates Nipper resistance. Surprisingly, although virulence and avirulence of the parental strains AlKewell and P94‐24 are reciprocal for Hurricane and Nipper, and for a substantial number of the haploid biparental progeny, only minor nonsignificant QTL peaks for the Nipper disease response were identified in this study. This could be due to multiple genes conferring these traits for Nipper in contrast to the single major gene identified for Hurricane. It is also possible that toxin functions of AlAvr1‐1 or AlAvr1‐2 play a role in the disease process in lentil and confound the genetic dissection of the apparent Nipper avirulence trait in the mapping population.

The AlAvr1 locus is located in a large AT‐rich region, analogous to the genomic context for AvrLm1 from L. maculans (Gout et al., 2006). Rather than a complete gene loss or gene inactivation process of mutation to drive this major shift in host preference of A. lentis pathotypes 1 and 2, it is evident that pathotype 2 strains were present in A. lentis populations prior to the release of the cultivar Hurricane, with AlKewell having been collected in 2001. With approximately 10% of the coding sequence of AlAvr1‐1 being polymorphic between the two pathotypes and a pairwise dN/dS value of 9.2, it is apparent that high positive selection pressure has determined the primary amino acid sequence of AlAvr1‐1. With so few AlAvr1 homolog sequences in genome databases, methods for analysis of evolutionary selection have some limitation. However, inclusion of a closely related AlAvr1 homolog from A. viciae‐villosae in the analysis of evolutionary selection has enabled the identification of sites that have been positively selected in the AlAvr1‐1 effector from pathotype 1. Four amino acid sites are predicted to be under positive selection in AlAvr1‐1. Two of these, positions 150 and 152, are alanine residues in AlAvr1‐2 in AlKewell, and in P94‐24 AlAvr1‐1 alanine residues are substituted with arginine and asparagine, respectively. These amino acid residues are larger and have both aliphatic and polar features that could contribute to functional differences between the proteins. Positive selection to confer avirulence in an effector protein is not likely because pathogen avirulence is not an evolutionary advantage (Skamnioti & Ridout, 2005). A secondary role for AlAvr1‐1, and possibly AlAvr1‐2, as a virulence factor is supported by the observed, strong positive selection of AlAvr1‐1. Furthermore, RIP mutation patterns were identified for almost half of the nucleotide polymorphisms in the cDNA sequences between AlKewell and P94‐24.

The identification of a novel avirulence effector in a dothideomycete pathogen presents us with interesting questions about pathogen evolution and host specialization. A. lentis is taxonomically classified in the Dothideomycete class, Pleosporales order of ascomycetes, which comprises several necrotrophic plant pathogens of economic importance, including L. maculans. While L. maculans produces a number of secondary metabolite toxins, such as sirodesmin PL (Gardiner et al., 2004; Sjödin & Glimelius, 1989) and phomalide (Pedras et al., 1993), the species produces multiple avirulence effectors. L. maculans is described as a hemibiotroph, having a long latent period of asymptomatic biotrophic growth in which the pathogen penetrates leaves and grows through xylem and cortex tissues of petioles and stems, followed by necrotrophic growth in main stems (Howlett et al., 2001). Transcriptome studies in L. maculans infection of canola have shown that toxin synthesis genes are down‐regulated during the biotrophic establishment phase and up‐regulated at 11 days postinoculation, after the transition from biotrophy to necrotrophy (Sonah et al., 2016; Urquhart et al., 2021). During biotrophic growth, expression of avirulence genes and salicylic acid defence signalling genes characteristic of host responses to biotrophic pathogens (Oliver & Ipcho, 2004) are more prominent (Lowe et al., 2014). AlAvr1‐1 and AlAvr1‐2 are similarly expressed during the first 7 days of asymptomatic growth in A.lentis during infection of both Nipper and Hurricane. AlAvr1‐1 expression in P94‐24 starts earlier than AlAvr1‐2 in AlKewell, and the peak expression level on Nipper at 3 days after inoculation is approximately 10‐fold higher. We have previously proposed that A. lentis is a hemibiotroph (Henares et al., 2019), with a long asymptomatic phase of around 7 days, and it is possible that there are parallels between L. maculans and A. lentis regarding the presence of avirulence effectors and the hemibiotrophic lifestyle observed for both species.

The characterization of AlAvr1‐1 and its orthologous allelic form, AlAvr1‐2, from A. lentis has provided important information about ascochyta blight disease in lentil. Knowledge about these effector genes will enable new research on the genetics of host resistance with the goal of improving the sustainability of production and management of ascochyta blight disease for this important legume crop. This first reported evidence for avirulence in an Ascochyta pathogen opens new perspectives on our understanding of this disease, with potential implications for research in other Ascochyta species and their respective crop diseases.

EXPERIMENTAL PROCEDURES

Materials

A. lentis isolates P94‐24 and AlKewell were obtained from the archive collection held at the South Australian Research and Development Institute (SARDI). P94‐24 was collected at Turretfield (South Australia) in 1994 and AlKewell was collected at Kewell (Victoria, Australia) in 2001. Lentil seed was sourced from Pulse Breeding Australia (Horsham, Australia) and PB Seeds (Horsham, Australia). Isolates were cultured on half‐strength PDA incubated at room temperature (18–22℃) under long‐wave UV light (photoperiod 12 h:12 h, light:dark). Spores were collected from mature pycnidia at 14 days after plating and filtered through sterile cotton wool.

Biparental mapping population

PCR assays were used to determine mating types of isolates of known reactions to Hurricane and Nipper (Blake et al., 2019) according to Chérif et al. (2006). Primers are listed in Table S3. A biparental A. lentis mapping population was produced from AlKewell (MAT1‐1) and P94‐24 (MAT1‐2) as described by Skiba and Pang (2003). Briefly, conidiospore suspensions (106 spores/ml) were prepared as described in Davidson et al. (2016), mixed together, and added to dried autoclaved lentil stubble pieces in a 10‐ml tube. Inoculated stubble pieces were placed onto wetted autoclaved filter paper in Petri plates and incubated at 8℃ under black light with photoperiod 12 h:12 h, light:dark. Asci and ascospores developed after 6 weeks and were confirmed by microscopy (Skiba & Pang, 2003). Haploid ascospores were discharged from mature asci as described by Skiba and Pang (2003) onto water agar, and single ascosporic cultures (n = 100) were stored as 5 mm2 agar plugs in sterile water at 5℃.

Pathotyping of a biparental A. lentis population

The biparental A. lentis population was assessed for virulence or avirulence under controlled conditions on Nipper, Hurricane, and the susceptible control ILL6002. The three differential lentils were grown in potting soil. Seedlings were inoculated with pycnidiospore suspensions (106 spores/ml) containing 0.025% Tween 20 at 14 days after sowing. Inoculated plants were misted for 5 s every 2 h for 9 days. Plants were maintained at 18℃ with photoperiod 12 h:12 h, light:dark. Disease symptoms started at 7 days and plants were scored for percentage leaf area damage at 14 days. Figure S6 shows disease symptoms for parental A. lentis strains AlKewell and P94‐24 on ILL6002, Nipper, and Hurricane. The “percentage leaf area damage” assessment is an overall measure of the proportion of necrotic tissue of whole plants, including leaves and stems. Biparental progeny were tested in a randomized complete block design with four replicates of each isolate × cultivar combination. Phenotyping was conducted across five separate experiments. To enable comparison of disease estimations across experiments, mean disease scores for Nipper and Hurricane were normalized to the mean disease score for ILL6002 within each experiment.

Isolate genotyping

DNA from 95 AlKewell × P94‐24 biparental progeny was isolated as described by Henares et al. (2019). ddRADseq genotyping‐by‐sequencing (GBS) of the biparental population was performed by Australian Genome Research Facility (Melbourne, Australia). DNA samples from isolates and progeny were digested using restriction enzymes PstI and MspI, and end‐sequencing of resulting fragments was performed on the Illumina NextSeq500 platform with 150 cycles in mid‐output mode. ddRADseq sequencing data were processed using Stacks v. 1.47 (Catchen et al., 2011, 2013) for sequence trimming and variant calling. Markers were filtered with loci removed where data were missing for more than 20% of samples or where allele frequency was less than 2%.

QTL analysis of the AlKewell × P94‐24 mapping population

QTL mapping of ILL6002 disease and normalized disease scores for Hurricane and Nipper was performed using the R‐based program R/qtl2 (Broman et al., 2019). ddRADseq marker positions were mapped to the 23‐contig A. lentis Al4 genome assembly (GenBank accession GCA_004011695.1; Lee et al., 2021) using BLASTn (Altschul et al., 1990). Unmapped loci were binned to the “unmapped” group designated “chr24”. R/qtl2 was run in mixed linear model mode with kinship calculated in R/qtl2 to account for relationships between isolates. The statistical significance of the results was tested using the permutation test in R/qtl2 with 1000 iterations.

Genome sequencing and analysis

Additional genome assemblies were produced using Oxford Nanopore sequencing (Oxford Nanopore Technologies). Isolates AlKewell and P94‐24 were grown in yeast extract dextrose liquid medium for 3 days with shaking at 180 rpm at 22℃ and DNA isolation. Nanopore sequencing and assembly were performed as described by Debler and Henares (2020). Whole‐genome alignment for comparison of A. lentis genome assemblies was performed using Nucmer v. 4.0.0 (Kurtz et al., 2004) with the “‐‐mum” argument. Percentage GC content was calculated using Bedtools v. 2.26.0 with the “bedtools nuc” command (Quinlan & Hall, 2010). RIP mutation was assessed using the program RIPCAL (Hane & Oliver, 2008). Signal peptides in annotated proteins were predicted using SignalP v. 5.0 (Almagro Armenteros et al., 2019) and effector probabilities were estimated using the program EffectorP v. 2.0 (Sperschneider et al., 2016, 2018).

Functional characterization of the g2688 effector using agroinfiltration

To test the functional properties of the AlAvr1 effector biparental population trait‐associated locus, we developed constructs based on the pEAQ‐HT vector (Sainsbury & Lomonossoff, 2008; Sainsbury et al., 2009) for transient expression in lentil using agroinfiltration as described by Debler et al. (2021). Vector diagrams are shown in Figure S8. Coding sequences corresponding to A. lentis g2688 from P94‐24 and AlKewell, with a signal peptide‐encoding sequence from the lentil PR1 protein, were designed for manufacture as gBLOCK DNA molecules from Integrated DNA Technologies. DNA sequences were inserted into the pEAQ‐HT transient expression vectors under the control of the CaMV 35S promoter. Plasmid constructs for the two AlAvr1 homologs were transformed into Agrobacterium tumefaciens GV3101 (pMP90), as described in Debler et al. (2021). Three‐week‐old lentil plants were infiltrated with Agrobacterium suspension at the adaxial leaf surface using a 1‐ml plastic syringe. The nonhost N. benthamiana was used as control. The dicot‐specific, necrosis‐inducing effector nep1‐like protein (NLP) from Peyronellaea pinodes was used as a positive necrosis control while green fluoresecent potein (GFP) was used as negative control (Debler et al., 2021). Agroinfiltrated leaves were checked for necrosis and photographed after 5 days.

Assessment of evolutionary selection models

Pairwise dN/dS values for AlAvr1 cDNA sequence from AlKewell and homologous cDNAs from P94‐24 and A. viciae‐villosae isolate ONG641 (GCA_004335205.1) were determined using the dNdS‐Calculator program (Shaw et al., 2012). More detailed methods to determine the mode of evolutionary selection were implemented using programs at the HyPhy webserver (https://www.hyphy.org/methods/selection‐methods/), including, FEL (fixed effects likelihood), FUBAR (fast unconstrained Bayesian approximation), MEME (mixed effects model of evolution), and aBSREL (adaptive branch‐site random effects likelihood). Additional homologous sequences from the NCBI nonredundant database were identified using tBLASTn (Altschul et al., 1990) and a neighbour‐joining tree for AlAvr1 and homologous protein sequences was constructed in Geneious v. 8.1.9 (Biomatters Ltd). The phylogenetic tree was drawn using the Interactive Tree of Life web server (iTOL v. 5; https://itol.embl.de/).

RT‐qPCR

The gene expression pattern of AlAvr1 during plant infection was determined using RT‐qPCR on an Applied Biosystems, QuantStudio 6 instrument (Thermo Fisher Scientific). Three biological replicates with three technical repeats of P94‐24 and AlKewell‐infected lentil cDNA were used as template in qPCRs with allele‐specific primer pairs that amplify 184 and 183 bp DNA sequences from AlAvr1‐1 and AlAvr1‐2 cDNA, respectively (Table S3 and Figure S9). Gene expression in infected leaf samples was monitored at 6 h, 12 h, 24 h, 3 days, 5 days, and 7 days postinoculation. RNA extraction was carried out using TRIzol reagent (Thermo Fisher Scientific) and cDNA was synthesized using an iScript cDNA synthesis kit (Bio‐Rad). An intron‐spanning primer pair for the A. lentis actin gene was used to check for DNA contamination of the total RNA. PCR amplification was carried out in 10‐µl reactions containing 5 µl of 2× SYBR Mastermix (Applied Biosystems), 250 nM of each primer, and 1 µl of cDNA template. PCR conditions included a touch‐down step as follows: 10 min, 95°C; 10 cycles of 15 s at 95°C, 15 s at 70°C, 40 s at 74°C; 35 cycles of 15 s at 95°C, 15 s at 60°C, and 2 min at 74°C; and final elongation of 5 min at 74°C. Gene expression was quantified based on a standard curve using genomic DNA. qPCR was performed in triplicate for each biological replicate and expression is reported relative to actin expression. To compare means of more than two data sets, analysis of variance (ANOVA) and Student's t test were implemented using the program JMP (v. 14.3.0; SAS Institute).

Gene knockout studies

Gene deletion strains of P94‐24 and AlKewell with disrupted AlAvr1‐1 and AlAvr1‐2 genes, respectively, were produced using A. tumefaciens‐mediated transformation (ATMT) with gene‐disruption constructs pTAR‐Hyg‐AlAvr1‐1 or pTAR‐Hyg‐AlAvr1‐2 (Figure S10) as described by Debler and Henares (2020). Deletion mutants were designated ΔAlAvr1‐1 and ΔAlAvr1‐2. Briefly, for gene disruption in AlKewell, 500 bp flanking regions of the AlAvr1‐2 gene were PCR‐amplified as outlined in Figure S10c. To construct the disruption mutant in P94‐24, the 1000 bp upstream flanking sequence and a 500 bp downstream sequence flanking the AlAvr1‐1 gene were similarly PCR‐amplified (Table S3 and Figure S10a). PCR used a Platinum SuperFi enzyme Master mix (Thermo Fisher Scientific). The empty vector pTAR‐0‐hyg (Debler & Henares, 2020) was digested with KpnI‐HF at 37°C for 60 min to generate vector backbone and hygromycin selectable marker cassette fragments. Vector, selectable marker, and A. lentis target gene flanking fragments were assembled using NEBuilder HiFi DNA assembly master mix (NEB). The resulting vectors (Figure S10e,f) were independently transformed into Escherichia coli OM2 competent cells plated onto Luria‐Bertani agar, and putative transformants were screened using PCR to confirm correct assembly of the fragments and verified by Sanger sequencing. Constructs were transformed into A. tumefaciens AGL1 competent cells as described by Henares et al. (2019). Successful gene disruption in the knockout strains was confirmed by PCR on total DNA of transformants using primers that amplified upstream and downstream of the expected insertion site (Figure S10g,h). Targeted gene disruption was further verified using low coverage Oxford Nanopore sequencing as described above. The pathogenicity of the WT and gene‐deletion strains was assessed as described for the seedling assay protocol above. Three biological replicates with three technical repeats were carried out for each lentil cultivar. Data were square root‐transformed to normalize residuals and ANOVA followed by the Tukey HSD test was used to determine which treatments were significantly different to each another.

FIGURE S1 Ascochyta lentis Al4‐mapped ddRADseq markers

Click here for additional data file.

FIGURE S2 Replica image of Figure 3b. Additional detail about the amino acid substitutions and the nucleotide differences that caused the changes in amino acid sequence between g2688 from Ascochyta lentis isolate P94‐24 and g2688 from isolate AlKewell. Variant nucleotides and the resulting amino acid changes are highlighted with coloured shading in the magnified sequences shown under the aligned cDNA and protein sequences

Click here for additional data file.

FIGURE S3 Phylogeny of g2688 homologs. Neighbour‐joining tree of amino acid sequences homologous to Ascochyta lentis homologs. The Ascochyta viciae‐villosae g2688 homolog was identified by tBLASTn of Ascochyta species genome assemblies. More distant homologous sequences were found for Leptosphaeria maculans, Colletotrichum spp., and Fusarium oxysporum f. sp. apii, and the tree is rooted to the L. maculans sequence

Click here for additional data file.

FIGURE S4 Ascochyta lentis and Ascochyta viceae‐villosae g2688 orthologs. cDNA and protein alignment of A. lentis AlKewell, P94‐24 and A. viceae‐villosae g2688 orthologs. Nucleotide identities and polymorphisms are indicated in the consensus sequence by green and white bars. Nucleotide and amino acid polymorphisms in specific sequences are shown in the cDNA and protein diagrams for each isolate

Click here for additional data file.

FIGURE S5 Agroinfiltration replicates. Images of all replicates of agroinfiltrated lentil and Nicotiana benthamiana leaves with positive necrosis control NLP, negative necrosis control GFP, and Ascochyta lentis g2688 protein from P94‐24 (g2688‐P) and from AlKewell (g2688‐K). Necrosis was observed for NLP infiltrations for all lentil and N. benthamiana leaves but not for GFP. g2688‐P induced necrosis only for PBA Hurricane XT and PBA Bolt and g2688‐K did not induce necrosis in any of the lentil cultivars

Click here for additional data file.

FIGURE S6 Disease response lentil cultivars with WT Ascochyta lentis strains. Plant disease assays with parental A. lentis isolates AlKewell and P94‐24 on ILL6002, Nipper, PBA Hurricane, and PBA Bolt. PBA Hurricane and PBA Bolt display a similar sensitivity to agroinfiltrated AlAvr1‐1 protein and have a common ascochyta‐resistant parent, the breeding line 96‐047L*99R060. PBA Bolt was shown to be resistant to P94‐24 and susceptible to AlKewell, similar to PBA Hurricane

Click here for additional data file.

FIGURE S7 Characterization of wild‐type and mutant strains. Growth rate and morphology, and sporulation of wild‐type (WT) Ascochyta lentis strains, P94‐24 and AlKewell, and gene knockout strains ΔAlAvr1‐2 in the AlKewell background and the ΔAlAvr1‐1 in the P94‐24 background. Strains were inoculated on half‐strength potato dextrose agar and radial growth (a), colony morphology (b), and sporulation (c) were assessed after 7 days. The mutant strains were shown not to be different with respect to these phenotypic traits, to the respective WT strains

Click here for additional data file.

FIGURE S8 Agroinfiltration constructs. Constructs for transformation of Agrobacterium tumefaciens and subsequent agroinfiltration of lentil and Nicotiana benthamiana were produced using the modified pEAQ‐HT‐DEST vector as described in Debler et al. (2021). The vectors pEAQ‐HT‐DEST1‐lenPR1‐his g2688 P and pEAQ‐HT‐DEST1‐lenPR1‐his g2688‐K carried the coding sequences for Ascochyta lentis g2688 from P94‐24 (g2688‐P) and AlKewell (g2688‐K), in frame with a N‐terminal lentil PR1 signal peptide and C‐terminal 6× histidine tag. The pEAQ‐HT‐DEST1 vector backbone contains cowpea mosaic virus 5′‐ and 3′‐UTR sequences that flank the transgene and also a gene expression cassette for production of the P19 gene silencing suppressor to enhance gene expression (Sainsbury & Lomonossoff, 2008; Sainsbury et al., 2009)

Click here for additional data file.

FIGURE S9 Reverse transcription‐quantitative PCR primer sites. qPCR primers (red/orange) were designed to amplify cDNA sequence towards the 3′ end of transcripts, using a common reverse primer corresponding to conserved nucleotide sequence and pathotype 1 and 2 gene‐specific primers corresponding to a polymorphic region of the gene. qPCR primers amplify a 184 bp product for AlAvr1‐1 from P94‐24 or a 183 bp fragment from AlAvr1‐2 from AlKewell. Primer sequences were as follows: AlAvr1‐1‐qPCR‐F, 5′‐GCGGAAGACCCAACAGC‐3′, AlAvr1‐2‐qPCR‐F, 5′‐CGGAGCATCCGCCAAA‐3′, and the common reverse primer AlAvr1‐qPCR‐R, 5′‐GACGTACTGAAGGGTGGGGT‐3′

Click here for additional data file.

FIGURE S10 Gene knockout constructs. Vector maps and validation of knockout constructs and transformed Ascochyta lentis strains. 5′ untranslated region (UTR) and 3′ UTR flanking regions for AlAvr1‐1 (a, b) and AlAvr1‐2 (c, d) were amplified from P94‐24 and AlKewell genomic DNA using primers as indicated and detailed in Table S3. Knockout constructs were assembled using Gibson assembly as described in (Debler & Henares, 2020), as shown for AlAvr1‐1 (b) and AlAvr1‐2 (d), respectively, with the hygromycin expression cassette inserted between upstream and downstream flanking DNA. The resulting construct for AlAvr1‐1 knockout, pTAR‐hyg‐AlAvr1‐1, is shown (e) with the full hygromycin expression cassette with PTrpC and TTrpC promoter and terminator, and an equivalent construct for AlAvr1‐2 (pTAR‐hyg‐AlAvr1‐2) with the AlKewell‐specific flanking DNA was similarly produced. Transformation of A. lentis strains P94‐24 and AlKewell with pTAR‐hyg‐AlAvr1‐1 and pTAR‐hyg‐AlAvr1‐2 led to production of knockout strains ΔAlAvr1‐1 and ΔAlAvr1‐2, respectively. Gene knockout was achieved by substitution of the target gene with the hyg expression cassette through the process of homologous recombination as shown in (f). Deletion of the targeted AlAvr1‐2 and AlAvr1‐1 genes in respective strains was confirmed by PCR on genomic DNA using primers flanking the insertion site (g) and on cDNA using coding sequence‐specific primers (h). On genomic DNA the knockout strains have a larger amplicon having the larger PTrpC‐hph‐TTrpC compared to the native AlAvr1‐1 or AlAvr1‐2 (g) and on cDNA templates, the gene knockout is indicated by loss of the coding sequence amplicon. Whole‐genome sequencing of wild‐type (WT) and knockout strains showed the insertion of the hyg cassette in the exact location corresponding to the flanking DNA sequences and only a single insertion for both strains. Dot plot comparisons for WT with respective knockout strains are shown for AlAvr1‐2 in AlKewell (i) and AlAvr1‐1 in P94‐24 (j)

Click here for additional data file.

TABLE S1 Mapped ddRADseq marker sequences

Click here for additional data file.

TABLE S2 Analysis of selection pressure using HyPhy webserver programs. Positive selection results from analyses using HyPhy programs for assessing selection pressure in the chromosome 3 putative effector from Ascochyta lentis. The Branch method aBSREL predicted positive selection in P94‐24 for the g2688 orthologs (p < 0.005). Site methods FEL, MEME, and FUBAR predicted positive selection at four amino acid sites that were different between AlKewell and P94 24 (p < 0.1 for FEL and MEME; posterior probability α < β > 0.9 for FUBAR)

Click here for additional data file.

TABLE S3 Primer sequences

Click here for additional data file.

TEXT S1 Gene, cDNA, and protein sequences from Ascochyta lentis P9424, AlKewell, and Ascochyta viciae‐villosae

Click here for additional data file.

TEXT S2 Protein sequences of g2688 homologs from NCBI NR database

Click here for additional data file.