GENETICS OF IMMUNITY Unraveling Genomic Complexity at a Quantitative Disease Resistance Locus in Maize

Multiple disease resistance has important implications for plant fitness, given the selection pressure that many pathogens exert directly on natural plant populations and indirectly via crop improvement programs. Evidence of a locus conditioning resistance to multiple pathogens was found in bin 1.06 of the maize genome with the allele from inbred line “Tx303” conditioning quantitative resistance to northern leaf blight (NLB) and qualitative resistance to Stewart’s wilt. To dissect the genetic basis of resistance in this region and to refine candidate gene hypotheses, we mapped resistance to the two diseases. Both resistance phenotypes were localized to overlapping regions, with the Stewart’s wilt interval refined to a 95.9-kb segment containing three genes and the NLB interval to a 3.60-Mb segment containing 117 genes. Regions of the introgression showed little to no recombination, suggesting structural differences between the inbred lines Tx303 and “B73,” the parents of the fine-mapping population. We examined copy number variation across the region using next-generation sequencing data, and found large variation in read depth in Tx303 across the region relative to the reference genome of B73. In the fine-mapping region, association mapping for NLB implicated candidate genes, including a putative zinc finger and pan1. We tested mutant alleles and found that pan1 is a susceptibility gene for NLB and Stewart’s wilt. Our data strongly suggest that structural variation plays an important role in resistance conditioned by this region, and pan1, a gene conditioning susceptibility for NLB, may underlie the QTL. THE genes and loci that influence host–pathogen interactions vary in allele effects, specificities, and linkage relationships. While disease resistance can be conditioned by single genes with large effect (Bent 1996; Jones and Dangl 2006), the emerging model of resistance for many plant diseases is complex in nature, with many genes and loci functioning in concert and each contributing a small proportion of the total phenotypic variation (Kump et al. 2011; Poland et al. 2011; Cook et al. 2012b). Each locus has a unique profile, with some loci contributing broadspectrum protection against diverse pathogen species and strains. Investigating these intricacies offers the opportunity to understand the diverse ways in which plants defend themselves against microbial assault. Correlated responses to multiple diseases have been observed in various germplasm panels, implying that there are loci and genes that condition broad-spectrum resistance (Rossi et al. 2006; Gurung et al. 2009; Wisser et al. 2011). At the chromosomal segment level, disease and insect resistance loci colocalize in a nonrandom fashion (McMullen and Simcox 1995; Williams 2003; Wisser et al. 2005) and loci have been identified that confer resistance to diverse pathogen isolates and taxa (Zwonitzer et al. 2010; Chung et al. 2011; Belcher et al. 2012). There is evidence to suggest that gene clusters can confer resistance to more than one disease. A cluster of germin-like proteins confers resistance to rice blast and sheath blight of rice (Manosalva et al. 2009). Similarly, resistance gene homologs, which are known to colocalize with broad-spectrum disease resistance loci, can cluster in the genome and contribute a diversity of specificities (Lopez et al. 2003; Ramalingam et al. 2003). Pleiotropy remains uncommon in maize, and correlated responses may be due to Copyright © 2014 by the Genetics Society of America doi: 10.1534/genetics.114.167486 Manuscript received April 3, 2014; accepted for publication June 6, 2014; published Early Online July 9, 2014. Supporting information is available online at http://www.genetics.org/lookup/suppl/ doi:10.1534/genetics.114.167486/-/DC1. Corresponding author: 303G Plant Science Bldg., 236 Tower Rd., Cornell University, Ithaca, NY 14853. E-mail: [email protected] Genetics, Vol. 198, 333–344 September 2014 333 linkage or population structure (Wallace et al. 2014), although in some cases, individual genes have been shown to condition multiple disease resistance (MDR). For example, the putative ABC transporter Lr34 of wheat provides protection against leaf rust, stripe rust, and powdery mildew (Krattinger et al. 2009). Pattern recognition receptors are able to detect molecular patterns from diverse organisms to confer disease resistance (Zipfel and Rathjen 2008). While in some cases single genes or alleles common across diverse germplasm confer disease resistance, increasingly, the role of structural variation in plants is being explored and its effects on phenotypic variation recognized (Springer et al. 2009; Chia et al. 2012; McHale et al. 2012). As quantitative trait loci (QTL) are subjected to fine mapping, some loci fractionate into many QTL, each conditioned by one or more genes (Steinmetz et al. 2002; Studer and Doebley 2011; Johnson et al. 2012). In some cases, the allele effect conditioned by each QTL is small enough that the individual locus cannot be identified in isolation (Buckler et al. 2009; Poland et al. 2011). In other cases, single resistance loci, such as Rhg1, are conditioned by multiple genes present in varying copy numbers in different lines (Cook et al. 2012a; Maron et al. 2013). Whole-genome studies have in fact suggested that structural variation is generally associated with disease resistance: structural variation in plants colocalizes with resistance nucleotide-binding proteins, receptor-like proteins, and disease resistance QTL (Lai et al. 2010; McHale et al. 2012; Xu et al. 2012). The conventional approach of genetic isolation and transgenic complementation remains the gold standard for demonstrating the function of a gene. This approach, however, is proving inadequate for dealing with the complexity underlying some loci, particularly for structural variation. Strong evidence for the importance of copy number variation in explaining trait variation (Cook et al. 2012a; Maron et al. 2013) and the emerging model of plant defense with many loci each contributing a small effect combine to challenge this paradigm (Kump et al. 2011; Poland et al. 2011; Cook et al. 2012b). There is a need for a new approach that can take advantage of whole-genome analyses, address presence/absence variation, and examine loci with small effects. This study represents such an approach and provides insights into a genetically complex locus affecting diverse traits. In maize, chromosomal bin 1.06 has been identified as a key locus for stabilizing yield under adverse conditions, including both biotic and abiotic stress (Landi et al. 2002, 2010; Tuberosa et al. 2002). In addition to plant architectural traits and yield under abiotic stress, resistance to many diseases has been localized to bin 1.06, including northern leaf blight (NLB), Stewart’s wilt, southern leaf blight (SLB), common rust, gray leaf spot (GLS), and ear and stalk rot caused by multiple fungi (Wisser et al. 2006; Chung et al. 2010b; Zwonitzer et al. 2010). In a QTL study of the recombinant inbred line (RIL) population Ki14 3 B73 evaluated for three foliar fungal diseases, NLB, GLS, and SLB, a 33-Mb region spanning bins 1.05 and 1.06 was the only locus identified that conferred resistance to all three diseases (Zwonitzer et al. 2010). A number of QTL studies for NLB resistance in maize have identified QTL at bin 1.06, ranging in physical size from 3 to 30 Mb (Freymark et al. 1993; Welz et al. 1999; Wisser et al. 2006; Chung et al. 2010b, 2011; Van Esbroeck et al. 2010; Poland et al. 2011). Additionally, bin 1.06 harbors the dominant Stewart’s wilt resistance gene Sw1 (Ming et al. 1999). Both NLB, caused by the fungus Setosphaeria turcica, and Stewart’s wilt, caused by the bacterium Pantoea stewartii, are foliar, hemibiotrophic diseases important to maize production. Both pathogens spread through the vascular tissue, causing wilted lesions by plugging xylem vessels (Jennings and Ullstrup 1957; Roper 2011). The importance of genes localized to maize bin 1.06 in resistance to both NLB and Stewart’s wilt has been described in multiple mapping populations. Using a population of Tx303 3 B73 introgression lines (Szalma et al. 2007), Chung et al. (2010b) showed that the NLB resistance QTL at 1.06 protects against fungal penetration. To explore the genomic complexity of this important region, we constructed high-resolution mapping populations at this locus and evaluated NLB and Stewart’s wilt resistance, using a set of Tx303 3 B73 near-isogenic lines (NILs) (Szalma et al. 2007; Chung et al. 2010b). Fine mapping allowed us to dissect the linkage relationship between the major-effect Stewart’s wilt QTL and the minor-effect NLB QTL and to identify candidate genes. Using association mapping, we further refined the list of candidate genes for NLB resistance and using mutants confirmed a role for the receptor-like kinase, pan1, in plant defense. Furthermore, multiple lines of evidence indicated a lack of genomic stability at the region, including reduced recombination across portions of the fine-mapping region in the NIL population and indicators of copy number variation. Materials and Methods