The northern corn leaf blight resistance gene Ht1 encodes an nucleotide‐binding, leucine‐rich repeat immune receptor

Abstract

Northern corn leaf blight, caused by the fungal pathogen Exserohilum turcicum, is a major disease of maize. The first major locus conferring resistance to E. turcicum race 0, Ht1, was identified over 50 years ago, but the underlying gene has remained unknown. We employed a map‐based cloning strategy to identify the Ht1 causal gene, which was found to be a coiled‐coil nucleotide‐binding, leucine‐rich repeat (NLR) gene, which we named PH4GP‐Ht1. Transgenic testing confirmed that introducing the native PH4GP‐Ht1 sequence to a susceptible maize variety resulted in resistance to E. turcicum race 0. A survey of the maize nested association mapping genomes revealed that susceptible Ht1 alleles had very low to no expression of the gene. Overexpression of the susceptible B73 allele, however, did not result in resistant plants, indicating that sequence variations may underlie the difference between resistant and susceptible phenotypes. Modelling of the PH4GP‐Ht1 protein indicated that it has structural homology to the Arabidopsis NLR resistance gene ZAR1, and probably forms a similar homopentamer structure following activation. RNA sequencing data from an infection time course revealed that 1 week after inoculation there was a threefold reduction in fungal biomass in the PH4GP‐Ht1 transgenic plants compared to wild‐type plants. Furthermore, PH4GP‐Ht1 transgenics had significantly more inoculation‐responsive differentially expressed genes than wild‐type plants, with enrichment seen in genes associated with both defence and photosynthesis. These results demonstrate that the NLR PH4GP‐Ht1 is the causal gene underlying Ht1, which represents a different mode of action compared to the previously reported wall‐associated kinase northern corn leaf blight resistance gene Htn1/Ht2/Ht3.

Maize northern corn leaf blight resistance gene Ht1 was cloned by a map‐based strategy. It encodes an NLR protein with structural homology to ZAR1 and gene classification indicates it has a different mode of action to Htn1/Ht2/Ht3.

1. INTRODUCTION

Northern corn leaf blight (NLB), caused by the fungal pathogen Exserohilum turcicum (teleomorph Setosphaeria turcica), is an important disease in maize affecting both field corn and sweet corn. It occurs worldwide in areas that are temperate and humid, which includes all major maize‐growing regions in the United States (Galiano‐Carneiro & Miedaner, 2017; Munkvold & White, 2016; Pataky & Ledencan, 2006; Perkins & Pedersen, 1987; Welz & Geiger, 2000). Infected plants exhibit cigar‐shaped lesions that expand in size as the disease progresses, resulting in completely blighted plants in very severe cases (Galiano‐Carneiro & Miedaner, 2017; Hooker, 1963; Munkvold & White, 2016; Welz & Geiger, 2000). Depending on when infection first occurs, overall yield losses may be up to 30%–50% (Galiano‐Carneiro & Miedaner, 2017; Munkvold & White, 2016; Perkins & Pedersen, 1987), with highest losses occurring when infection begins before silking (Munkvold & White, 2016, Perkins & Pedersen, 1987). Although fungicides may be employed to manage NLB, the most common means of controlling infection is by deploying resistant maize varieties (Galiano‐Carneiro & Miedaner, 2017; Munkvold & White, 2016; Pataky & Ledencan, 2006; Welz & Geiger, 2000). The first major NLB resistance locus identified in maize was Ht1 (Hooker, 1963).

Ht1 was originally identified from two separate sources, the inbred line GE440 and a popcorn accession known as Ladyfinger (Hooker, 1963; Pataky & Ledencan, 2006; Welz & Geiger, 2000), and has been widely used in maize breeding programmes since the early 1960s. Maize plants containing the Ht1 locus produce small chlorotic lesions, as compared to large brown lesions on plants lacking Ht1 (Munkvold & White, 2016; Perkins & Pedersen, 1987). These chlorotic lesions delay and suppress fungal sporulation during the growing season, which can prevent the spread of inoculum (Hooker, 1963; Munkvold & White, 2016). When introgressed into a susceptible background, the presence of Ht1 reduces the severity of NLB by approximately one‐third when plants are challenged with race 0 of the pathogen (Hooker & Kim, 1973; Lipps et al., 1997; Pataky & Ledencan, 2006).

Since 1965, it has been known that Ht1 is located on the long arm of maize chromosome 2 (Patterson et al., 1965). Subsequent analyses confirmed this result and delimited Ht1 between two restriction fragment length polymorphism markers, UMC22 and UMC122 (Bentolila et al., 1991; Coe et al., 1995; Li et al., 1998; Patterson et al., 1965; Welz & Geiger, 2000). Over the past several decades, multiple genetic loci from different maize sources have been identified that confer resistance to NLB (Chen et al., 2015; Poland et al., 2011; Rashid et al., 2020; Welz & Geiger, 2000). For instance, the gene underlying the NLB resistance locus Htn1/Ht2/Ht3 has been identified to be a wall‐associated receptor‐like kinase (Hurni et al., 2015; Yang et al., 2021). However, despite knowledge of its genomic location and the importance of Ht1 in maize breeding programmes, the underlying gene has not been described.

Here we report that the gene underlying the Ht1 locus is a nucleotide‐binding, leucine‐rich repeat protein (NLR). A survey of the maize nested association mapping (NAM) genomes showed that susceptible alleles of Ht1 have very low expression, but overexpression of the B73 allele does not result in resistance, indicating that sequence variation underlies the difference between resistant and susceptible phenotypes. Structural modelling of Ht1 suggests that it has homology to ZAR1 from Arabidopsis. RNA sequencing (RNA‐Seq) analysis revealed networks of differentially expressed genes in Ht1 transgenic plants as compared to wild‐type plants that appear to result in reduced fungal biomass in infected plants.

2. RESULTS

2.1. An NLR gene identified as candidate gene for Ht1

To identify the gene responsible for Ht1 resistance, a BC₅ mapping population of 2450 individuals was generated with the Ht1‐containing resistant inbred PH4GP and the susceptible inbred PH5W4. PH5W4 was used as the recurrent parent for population development. The BC₅ plants were inoculated at the V5 stage with race 0 inoculum and scored for NLB resistance or susceptibility at the R3 stage in Windfall, Indiana. Single‐nucleotide polymorphism (SNP) markers located within the previously identified chromosome 2 region were used to confirm and define the Ht1 interval. Ht1 was initially delimited to a c.260 kb interval (B73 v4 genome), flanked by the cleaved amplified polymorphic sequence (CAPS) markers PHM505 and b199.a6 (Figure 1). Additional markers were developed to genotype the recombinants within the region. The Ht1 locus was finally narrowed to a c.19 kb interval (B73 v4) flanked by the CAPS markers M12 and M20. Based on the B73 v4 genome, this interval contains two annotated proteins: a 60S ribosomal protein and a disease resistance protein (CC‐NBS‐LRR [NLR] class) (Figure 1). From its annotation, the disease resistance protein, Zm00001d007056, was the apparent candidate gene, but this gene had very low to no expression in B73 (Table S1).

FIGURE 1.

Map‐based cloning of Ht1. The top panel shows the location of Ht1 on chromosome 2. The middle panel shows the region in B73 v4 with relevant markers annotated. The bottom panel shows a comparison of the interval in B73 and the resistant donor line PH4GP

To determine the nature of the resistant allele, a PH4GP bacterial artificial chromosome (BAC) library was constructed and a BAC clone spanning the fine‐mapped interval was identified and sequenced. The Ht1 interval in PH4GP was c.9.5 kb in size, with only a single annotated NLR gene fully within the region (Figure 1). This NLR gene was the apparent candidate gene underlying the Ht1 locus.

2.2. Ht1 candidate gene validated in transgenic testing

To confirm that the NLR gene within the fine‐mapped region was the gene conferring NLB resistance, a construct was made in which the genomic sequence of the NLR gene was expressed under the control of its native promoter and terminator. This construct was transformed into PHR03, a maize line that is susceptible to all races of E. turcicum. Transgenic positive and null segregants were selected for a greenhouse‐based NLB assay.

V3‐stage plants were challenged with race 0, race 23N, or race 1 of E. turcicum to determine the efficacy and race‐specific resistance by the Ht1 candidate gene. While null plants were susceptible to all pathogen races, consistent with the expected spectrum of Ht1 resistance, the transgenic plants were resistant to race 0 and race 23 N, but susceptible to race 1 (Figure 2). This confirms that, unlike Htn1/Ht2/Ht3, which are different alleles of a wall‐associated kinase (Hurni et al., 2015; Yang et al., 2021), Ht1 encodes an NLR and represents a new mechanism of resistance to E. turcicum. This gene will hereafter be referred to as PH4GP‐Ht1.

FIGURE 2.

Transgenic validation of Ht1 candidate gene. (a) Expression levels (relative to maize eIF4G) of the Ht1 candidate gene in PH4GP (the resistant donor line), PHR03 (the transformation background), transgenic plants, and null siblings. The expression level is the average of six individuals and the error bars represent the standard error. Transgenic plants expressing PH4GP‐Ht1 under the control of the native promoter and terminator have a similar expression level as compared to the native allele. PHR03 and nulls have no PH4GP‐Ht1 expression. (b) Transgenic plants are resistant while null siblings are susceptible to race 0 and race 23N of Exserohilum turcicum while both are susceptible to race 1

2.3. Overexpression of B73 Ht1 allele does not result in resistance

During the map‐based cloning process, a putative disease resistance protein gene was identified in B73 between the closest flanking markers (Figure 1), but this gene had little to no expression (Table S1). After the validation of PH4GP‐Ht1, the B73 allele was re‐examined. A protein sequence alignment was performed in Geneious Prime (Biomatters Ltd) using the genomic sequence of PH4GP‐Ht1 and the B73 allele (Figure S1a). The sequences were found to be 98.5% identical between PH4GP‐Ht1 and the B73 allele.

Given the high level of identity between PH4GP‐Ht1 and the B73 allele, together with the lack of expression in B73, it was hypothesized that the susceptibility of B73 to race 0 of E. turcicum may be due to the expression level of the gene as opposed to sequence variation. To test this hypothesis, constructs were made in which the genomic sequence of PH4GP‐Ht1 or the B73 allele from the the start codon ATG to the stop codon were expressed under the control of the maize histone H2B promoter. Transgenic plants were generated by transforming the constructs into PHR03. The presence of the transgene was assessed by quantitative PCR (qPCR) and the expression level of the gene was assessed by reverse transcription (RT)‐qPCR. For both PH4GP‐Ht1 and the B73 allele, transgenic plants exhibited a similar level of expression, while the nulls for both constructs showed no expression (Figure 3a). These plants were then challenged with race 0 of E. turcicum. For PH4GP‐Ht1, all transgenic plants were resistant while the nulls were susceptible. In contrast, for the B73 allele, all the plants, both transgenics and nulls, were susceptible (Figure 3b). This result indicates that, despite the low native expression level of the B73 allele, it is the sequence variations of the gene that are responsible for the race 0 susceptibility exhibited by B73.

FIGURE 3.

Overexpression of the B73 allele does not confer resistance to race 0. (a) PH4GP‐Ht1 and the B73 allele (B73‐Ht1) under the control of the H2B promoter are expressed to a similar level in transgenic plants while null plants have no expression. The expression level is the average of six individuals and the error bars represent the standard error. (b) The overexpression of PH4GP‐Ht1 confers resistance to race 0 while the overexpression of B73‐Ht1 does not

2.4. Structural modelling of Ht1

To better understand how minor sequence variations could render the B73 allele nonfunctional, we used AlphaFold2 to predict its structure and compared it to the recently published AtZAR1 NLR (Figure 4) (Jumper et al., 2021; Wang et al., 2019). Although the whole protein identity of PH4GP‐Ht1 and AtZAR1 was low, their NB‐ARC domains were 64% similar, with conserved nucleotide‐binding and interdomain interacting motifs, indicating that Ht1 probably operates via a similar conformational switch mechanism during activation. The predicted structures of the Ht1 inactive monomer, active protomer, and homopentamer were all similar to AtZAR1. The nucleotide‐binding site is located at the domain interface of nucleotide‐binding domain (NBD), helix domain (HD1), and winged helix domain (WHD), with ADP binding to the inactive state and ATP binding to the active state. During activation, structural modelling indicated that the NBD‐HD1 and WHD‐LRR (leucine‐rich repeat) are capable of undergoing c.180° rigid body rotation, with the nucleotide polyphosphate moiety interacting with different sites of the WHD. Modelling also showed that in the inactive state ADP β‐phosphate interacts with WHD's H506 residue while R314, due to the lack of γ‐phosphate, forms a salt‐bridge with D507, with both H506 and D507 in the conserved Met‐His‐Asp (MHD) motif. In the active state, the bound nucleotide moves 1.5 nm away from the MHD motif to the vicinity of α1 of WHD. The ATP γ‐phosphate forms a salt‐bridge with R314, which in turn forms hydrogen bonds to α1's S418 carbonyl. The pore formed by the Ht1 pentamer consists of hydrophobic residues facing the membrane bilayer and hydrophilic residues facing the lumen, similar to that in AtZAR1, which facilitate its role as a cation channel. Taken together, these data indicate that Ht1 may act via a similar mechanism to AtZAR1. We next examined amino acid differences in the theoretical B73 peptide constructed via a MUSCLE alignment of PH4GP's coding sequence against the B73 genome, which was 98.5% identical to Ht1 (Figure S1a). The monomer structure was found to be highly similar between the two, although some differences were present in conserved domains. There were two substitutions in the coiled‐coil region, three substitutions in the NB‐ARC domain, and six in the LRR region, but none were predicted to significantly alter protein structure. While there were no differences that would significantly disrupt the monomer structure, the two substitutions in the coiled‐coil domain (V34A and V123A) were predicted to significantly weaken hydrophobic packing in the pentamer pore structure and therefore could reduce the stability of this conformation.

FIGURE 4.

PH4GP‐Ht1 structural modelling and comparison to B73‐Ht1. (a) Ht1 domain organization. (b) Ht1 inactive monomer. (c) Ht1 active protomer. (d) Ht1 homopentamer. (e) Active PH4GP‐Ht1 protomer model depicting the location of amino acid differences relative to B73‐Ht1

2.5. Conservation and expression of Ht1 alleles in NAM lines

The recently sequenced NAM genomes provide an excellent resource to assess the level of conservation of Ht1 across diverse maize germplasm (Hufford et al., 2021). We first phenotyped all 26 NAM lines for resistance to E. turcicum race 0 and race 1. None of the lines tested was found to be resistant to race 0, but they were susceptible to race 1, implying that none of them has a functional Ht1 allele. To assess the sequence diversity of Ht1 in the NAM lines, we performed a BLAST search of the Ht1 protein against the predicted proteomes of all NAM lines. All lines except for Ki11 contained an annotated protein similar to Ht1 arising from the same chromosome 2 region. However, none of these alleles produced an identical protein, with identities ranging from 76.7% to 99.4% and coverage ranging from 37.0% to 85.45% (Table S2). The predicted B73 protein from the NAM lines was found to only cover 48.8% of the PH4GP‐Ht1, although an earlier model from the B73 v4 reference genome was found to be more similar to the PH4GP protein (100% coverage at 95.6% identity).

Although the NAM lines were well‐annotated using RNA‐Seq data from a variety of different tissues, sequence differences could still conceivably also arise from differing annotation methods. To account for this, we performed BLAST searches against NAM line genomes using the genomic sequence of Ht1 and identified homologous regions in all lines. We then used MUSCLE to align the coding sequence of PH4GP‐Ht1 against these regions to determine whether identical proteins could theoretically be produced. No NAM lines were found to possess the ability to produce a protein identical to PH4GP‐Ht1, with theoretical identities ranging from 87.6% (CML103) to 99.5% (CML333). This analysis indicated that B73 was theoretically capable of producing a protein with 98.5% identity to PH4GP, although no model existed to support this transcript in either the NAM genome or B73 v4 genome. Overall, although all NAM lines contain an allele of Ht1, none have a genomic sequence matching the PH4GP allele at 100% identity and all annotated proteins have significant differences from PH4GP‐Ht1 due to alternate splicing as well as nonsynonymous SNPs.

Expression data of all NAM lines is available from 10 different tissue types (Hufford et al., 2021), enabling examination of the level of expression of different Ht1 alleles. The vast majority of Ht1 alleles had very little to no expression across all tissues in all NAM lines (Table S1). Notable exceptions to this included CML322, CML247, CML277, and M37W, all of which had modest expression levels of alternate Ht1 alleles that ranged from an average of 1.15 to 1.94 fragments per kilobase of transcript per million mapped reads (FPKM). The nonfunctional B73 Ht1 allele had very low expression, averaging only 0.04 FPKM across different tissue types.

2.6. Differential gene expression of NLB‐infected wild‐type and Ht1 transgenic maize

To assess gene expression changes that result from the presence of Ht1 during NLB infection, we performed an NLB infection time series experiment using wild‐type and Ht1 transgenic plants. Plants were sampled in triplicate at 0.5, 12, 24, 36, 72, and 168 h after infection (HAI) with NLB inoculum or mock inoculum. In this paper, we refer to the wild‐type and transgenic plants inoculated with either NLB or mock inoculum as Wi (wild‐type plant + inoculated with NLB), Ti (transgenic plant + inoculated with NLB), Wm (wild‐type plant + inoculated with mock), and Tm (transgenic plant + inoculated with mock). RNA‐Seq libraries were constructed from each time point and then sequenced on an Illumina NovaSeq 6000. A total of 625 million paired‐end reads were obtained after trimming adaptors and removing low‐quality reads (accession GSE206951). Salmon (v. 1.1.0) was used to quantify the high‐quality reads using the maize B73 v4 transcriptome (Patro et al., 2017). DEseq2 was then used to determine the number of differentially expressed genes (DEGs) for each pairwise comparison at each time point: Wm vs. Wi, Tm vs. Ti, and Wi vs. Ti (Love et al., 2014). Genes were considered significantly differentially expressed at p‐adjusted ≤0.01 and fold change ≥2. All pairwise comparisons had the most DEGs at 168 HAI (Figures 5 and S3). Interestingly, the comparison between mock and inoculated Ht1 transgenic plants at 168 HAI resulted in more than three times as many DEGs compared to the same comparison in wild‐type plants (4946 vs. 1583). We constructed Venn diagrams to determine the overlap between DEGs identified when comparing mock to inoculated plants in both wild‐type and Ht1 transgenics. At 168 HAI, the vast majority of the DEGs found in the wild‐type plants were also found in the Ht1 transgenic comparison, while the Ht1 transgenic comparison had thousands of unique DEGs (Figure 5c). Similar results were found when comparing a merged set of DEGs from mock versus inoculated across all time points (Figure 5d).

FIGURE 5.

Differential expression of mock and inoculated wild‐type and PH4GP‐Ht1 transgenic plants. (a) Pairwise comparison of mock versus inoculated PH4GP‐Ht1 transgenic plants. (b) Pairwise comparison of mock versus inoculated wild‐type plants. (c) Overlap of differentially expressed genes (DEGs) from the mock versus inoculated PH4GP‐Ht1 transgenic and wild‐type plant comparisons at 168 h after inoculation (HAI). (d) Overlap of DEGs from the mock versus inoculated PH4GP‐Ht1 transgenic and wild‐type plant comparisons at all time points merged together

To determine the functional enrichment of DEGs associated with response to NLB infection, we used the Panther GO (Gene Ontology) enrichment analysis tool (Figure 6 and Table S4) (Thomas et al., 2003). Several defence‐related GO terms were found to be significantly enriched when comparing mock versus inoculated wild‐type and Ht1 transgenics at both 48 and 168 HAI, including “programmed cell death induced by symbiote”, “inositol biosynthesis”, and “NAD salvage” (Laha et al., 2015; Pétriacq et al., 2016). Surprisingly, although the Tm vs. Ti comparison yielded more total DEGs, the Wi vs. Ti comparison was found to be more enriched for defence‐related GO terms. Nearly all of the most enriched DEG functions found in the Tm vs. Ti at 168 HAI were associated with photosynthesis, with the notable exception of terms associated with oxylipins, which have well‐established roles in plant defence. The Wi vs. Ti comparison also yielded significantly enriched defence terms, including “defence response to oomycetes”, “auxin influx”, and “auxin transport”, as well as many metabolic pathways that indirectly impact plant defences such as “oxylipin metabolic process” and “inositol metabolic process” (Blée, 2002).

FIGURE 6.

Rich factor plot for top 40 enriched terms at 168 h after inoculation for the wild‐type inoculated versus PH4GP‐Ht1 transgenic inoculated (Wi vs. Ti) comparison

Finally, we used the RNA‐Seq data as a proxy for E. turcicum biomass to determine if there was a biomass difference in wild‐type and Ht1 transgenic plants. For this analysis, we merged the B73 transcriptome with the public E. turcicum transcriptome and quantified the expression of the merged set via Salmon (Ohm et al., 2012). A small number of E. turcicum genes had expression in the mock treatments and were removed as spurious. The total gene expression of all E. turcicum genes relative to the total expression of all maize and E. turcicum genes was then calculated as a proxy for biomass. Early time points had no appreciable biomass, with the percentage of E. turcicum gene expression never rising above 0.06%. However, at 168 HAI E. turcicum gene expression accounted for 1.40% of all gene expression in the wild‐type inoculated plants, while it only accounted for 0.42% of the total expression in Ht1 transgenic inoculated plants. This biomass difference is in keeping with the phenotypic differences present at 168 HAI and indicates that Ht1 begins reducing E. turcicum biomass relatively early during infection.

3. DISCUSSION

Three well‐characterized NLB resistance quantitative trait loci (QTLs) in maize have been described, two of them (Htn1 and Ht2/3) having recently been cloned and found to encode WAK proteins (Hurni et al., 2015; Yang et al., 2021). Ht1 was the only remaining major NLB QTL for which the underlying gene had not yet been identified. We used a map‐based cloning approach to narrow the Ht1 interval to a 9.5 kb region in PH4GP containing a single NLR resistance gene, which we termed PH4GP‐Ht1. This result was confirmed through the transgenic insertion of PH4GP‐Ht1 to the susceptible maize variety PHR03, which resulted in resistance to E. turcicum races 0 and 23 N, but not to race 1, consistent with the expected resistance spectrum of Ht1. Interestingly, all NAM lines except for Ki11 were found to contain an allele of the Ht1 gene, but most of these alleles had little to no gene expression, making it difficult to determine the correct transcript models and make accurate protein‐level comparisons with PH4GP‐Ht1. To supplement the annotated models, we analysed theoretical maximum identity comparisons and found that most lines could, in theory, produce transcripts encoding proteins that were highly similar to PH4GP‐Ht1. To determine whether expression or minor sequence variations were more likely to be the cause of resistance phenotype differences, we overexpressed the B73 Ht1 allele with the maize H2B promoter. We determined that overexpressing the B73 allele was unable to confer resistance to race 0 of E. turcicum, indicating that sequence variation may be the more important factor in determining resistance phenotype.

To better understand the mechanism underlying PH4GP‐Ht1 resistance and determine how minor sequence variations could result in significantly different resistance phenotypes, we structurally modelled PH4GP‐Ht1. We then compared the resulting structure to AtZAR1, which has a resolved structure at the level of both inactive monomer and active homopentamer (Wang et al., 2019). Despite very low whole‐protein similarity, PH4GP‐Ht1 was predicted to have a similar structure to AtZAR1 and was also predicted to form a homopentamer on activation. Interestingly, examination of the amino acid changes in the susceptible B73 protein revealed changes in two interacting valine residues that were substituted with alanine residues of the coiled‐coil domain, which have the potential to reduce the stability of coiled‐coil interactions in the active homopentamer structure by lessening hydrophobic packing (Zhu et al., 1993). Such minor variations may contribute to the lack of resistance seen when overexpressing the susceptible B73 allele and provide further evidence that expression is not the only factor underlying differing resistance phenotypes of Ht1 alleles. In the future, systematic testing of all nonsynonymous SNPs in PH4GP and B73 Ht1 alleles could elucidate the causal variation.

We further examined the impact of PH4GP‐Ht1 on E. turcicum infection through an RNA‐Seq time series. E. turcicum RNA‐Seq reads were below the detection limit at all time points except 168 HAI, at which point a clear difference could be seen in the fungal reads of PH4GP‐Ht1 transgenic plants compared to the wild type, which had more than a threefold increase in E. turcicum gene expression, indicating that PH4GP‐Ht1 probably acts earlier on in the infection to cause a reduction in pathogen growth. When comparing mock to inoculated plants, PH4GP‐Ht1 transgenic plants had nearly three times as many inoculation‐responsive DEGs. Interestingly, the DEGs of the wild‐type plants were mostly contained within the DEG list of the transgenic plants, while transgenic plants had more than 3000 unique inoculation‐responsive DEGs in which up‐regulated DEGs were associated with defence (primarily oxylipin biosynthetic processes) and down‐regulated DEGs were primarily associated with photosynthesis. Similar changes have been found in other plant defence responses to fungal pathogens, including rice (Azizi et al., 2015) and wheat (Neugebauer et al., 2018). These results indicate that PH4GP‐Ht1 probably acts relatively early during E. turcicum infection to reduce photosynthesis and redirect resources towards defence responses.

Taken together, these results clearly demonstrate that the PH4GP allele of the Ht1 NLR underlies its NLB resistance phenotype. Although resistant Ht1 alleles are not present in the NAM founder lines, they have been widely deployed in commercial maize germplasm (Campaña & Pataky, 2005). Stacking of multiple disease resistance genes has been used to confer durable resistance to rapidly evolving pathogens in a diverse array of crop species, such as rice (Kumari et al., 2017), wheat (Zhang et al., 2021), and potato (Zhu et al., 2012). The PH4GP‐Ht1 gene represents a different mode of action compared to the other known major NLB resistance gene, a wall‐associated kinase with two known functional alleles (Htn1 and Ht2/Ht3). Ht1 and Htn1/Ht2/Ht3 not only represent different gene classes but are also known to confer resistance to different races of the E. turcicum. Stacking of these two resistance loci through either conventional breeding or genome editing therefore has the potential to confer durable broad‐spectrum NLB resistance to maize germplasm.

4. EXPERIMENTAL PROCEDURES

4.1. Plant material

A BC₅ mapping population of 2450 individuals was generated with the Ht1‐containing resistant inbred PH4GP and the susceptible inbred PH5W4. PH5W4 was used as the recurrent parent for population development.

4.2. BAC library construction, screening, and sequencing

BAC libraries were constructed, screened, and sequenced as previously described (Leonard et al., 2014). Briefly, leaves from approximately 200 etiolated seedlings were harvested, fast‐frozen, and used for high molecular weight DNA isolation. High molecular weight DNA was partially digested with HindIII and the BAC library was constructed in the pCC1BAC vector. Clones were screened and identified by hybridization using immunological detection of PCR digoxigenin‐labelled probes. Positive BACs were retrieved and confirmed by direct colony PCR. Probes specific for the Ht1 region were labelled from PCR fragments amplified using the M12 and M20 primers (see section 4.4).

BAC DNA from positive clones was purified using the Qiagen Large Construct system. Sequencing random shear sublibraries were constructed on pBluescript SK+. Plasmids were amplified from single‐colony cultures using Templiphi (GE Healthcare Life Sciences) and sequenced in an ABI 3730xl DNA sequencer using ABI PRISM BigDye (Applied BioSystems). Base calling, quality assessment, assembly, and validation were performed using phred, phrap, and exgap software.

4.3. Generation of transgenic plants

Transgenic plants were generated as previously described (Leonard et al., 2014). Briefly, constructs were introduced into Agrobacterium tumefaciens LBA4404 and used to transform maize embryos of PH184C or PHR03. Independent transgenic events were generated, and the copy number of transgene insertion was determined.

4.4. CAPS marker development

CAPS markers were developed by designing primers to PCR‐amplify low‐copy regions within the fine‐mapped interval and sequencing the PCR product in the resistant and susceptible parents to look for SNPs. Identified SNPs were then assessed for their ability to cause a difference in a restriction enzyme digestion site in the parents of the mapping population. The primers were then used for PCR on the individuals from the mapping population and the PCR product was digested with the restriction enzyme to determine the genotype of each individual. The primers and restriction enzyme used for each marker are as follows:

$$\begin{array}{c}\mathrm{PHM}505:\text{forward}{5}^{\prime }-\text{TGATCTACGAAGAGGATGAAT}-3^{\prime },\hfill \\ \text{reverse}{5}^{\prime }-\text{CGAGACCTGAACGTGACAAC}-3^{\prime },\mathit{Ase}\mathrm{I}\hfill \end{array}$$

$$\begin{array}{c}\mathrm{b}199\text{a6}:\text{forward}{5}^{\prime }-\text{AAATAAGCCGCTCAATGTGG}-3^{\prime },\hfill \\ \text{reverse}{5}^{\prime }-\text{CCAGAAAGCAAACAGAAGTGC}-3^{\prime },\mathit{Ase}\mathrm{I}\hfill \end{array}$$

$$\begin{array}{c}\mathrm{M}12:\text{forward}{5}^{\prime }-\text{TCTTGCCATGCTACACAAGC}-3^{\prime },\hfill \\ \text{reverse}{5}^{\prime }\text{GATACTGGAGGCAGCTGAGG}-3^{\prime },\mathit{Eco}\mathrm{RI}\hfill \end{array}$$

$$\begin{array}{c}\mathrm{M}20:\text{forward}{5}^{\prime }-\text{AGGTCTCGACAGAGCAGAGC}-3^{\prime },\hfill \\ \text{reverse}{5}^{\prime }-\text{CAACTTCCCTCCTGGTTCC}-3^{\prime },\mathit{Sac}\mathrm{II}\hfill \end{array}$$

4.5. Greenhouse assay

Two‐week‐old plants were inoculated with race 0 of E. turcicum. Plants were uniformly sprayed with a suspension of 25,000 conidia/ml; negative control plants were sprayed with water. Immediately after inoculation, plants were transferred into a dew chamber and kept for 16 h. Then, plants were removed from the dew chamber and transferred to a greenhouse maintained under 16 h of daylight.

4.6. Gene expression and enrichment analysis

RNA‐Seq reads from the NLB infection time series experiment using wild‐type and Ht1 transgenic plants were quantified against the maize B73 v4 transcriptome. Differential expression between each comparison at every time point was then performed using DESeq2. Genes under the FDR threshold (<0.05) were deemed as differentially expressed. Functional enrichment for these genes was performed on PANTHER geneset enrichment tool on Gene Ontology. Rich factor plots for significant (FDR < 0.05) terms were plotted using R.

4.7. RT‐qPCR for Ht1 expression level

One leaf punch was collected from vegetative leaf tissue in a 96‐well deep‐well plate with bullet tubes. Total RNA was extracted with TRK extraction buffers (PR021, PR030, PR031 from Omega Bio‐tek) and was treated with DNase I. RNA was reverse transcribed using the High‐capacity cDNA Reverse Transcription Kit (Thermo Fisher) into cDNA according to the manufacturer's procedures. Primers were designed for the Ht1 gene:

$$\mathrm{HT}1-\text{F3}:5^{\prime }-\text{GTGATCTGAGGATCGACGAA}-3^{\prime }$$

$$\mathrm{HT}1-\text{R3}:5^{\prime }-\text{TTTGAGCCGGTAGCACGACA}-3^{\prime }$$

RT‐qPCR was performed on the QuantStudio 7 real‐time PCR (384‐well Block) (Thermo Fisher) using SYBR Green (Thermo‐Fisher) with the gene‐specific primers listed above. Expression of Ht1 gene was quantified as relative to the expression of a reference gene, eIF4‐γ.

CONFLICT OF INTEREST

Corteva has filed patent applications related to the use of Ht1 (WO2017066597A1, WO2018071362A1).

Supporting information

Figure S1 Alignments protein (a) and coding sequence (b) from B73 and PH4GP Ht1 alleles

Figure S2 Differentially expressed genes with greater than twofold change between wild‐type inoculated versus transgenic inoculated

Figure S3 Top 40 enriched terms in pairwise comparisons at 168 h after incoculation for Ht1 transgenic mock versus inoculated (a) and wild‐type mock versus inoculated (b)

Table S1 Expression of NAM Ht1 alleles across 10 different tissues of the NAM founder lines

Table S2 Similarity of the PH4GP Ht1 allele to the Ht1 alleles in NAM founder lines

Table S3 Differentially expressed genes from comparisons of wild‐type and PH4GP‐Ht1 transgenic plants during an Exserohilum turcicum infection time series

Table S4 Significantly enriched gene ontology terms in the differentially expressed genes identified from comparisons of wild‐type and PH4GP‐Ht1 transgenic plants during an Exserohilum turcicum infection time series

Thatcher, S. , Leonard, A. , Lauer, M. , Panangipalli, G. , Norman, B. & Hou, Z. et al. (2023) The northern corn leaf blight resistance gene Ht1 encodes an nucleotide‐binding, leucine‐rich repeat immune receptor. Molecular Plant Pathology, 24, 758–767. Available from: 10.1111/mpp.13267

DATA AVAILABILITY STATEMENT

All relevant data can be found within the manuscript and its supporting information. The PH4GP‐Ht1 genomic and coding sequence are available in GenBank at https://www.ncbi.nlm.nih.gov/genbank/ under accession ON685201 and ON685202, respectively. RNA‐seq data are available at Gene Expression Omnibus (GEO) at https://www.ncbi.nlm.nih.gov/geo/ under accession GSE206951.