Downy mildew (DM), caused by several species of Peronosclerospora and Scleropthora, is a major maize (Zea mays L.) disease in tropical and subtropical regions. The causal pathogens, also called DMs, are oomycetes (not true fungi) and are obligate parasites of higher plants. They can overwinter in soil or on plant debris and can eventually produce visible, characteristic “downy” hyphae on the leaves (Agrios, 1988). Owing to its widespread distribution and severe yield reductions, DM is one of the most destructive of the maize pathogens (Nair et al., 2005), and in Asia, it is the most important biotic stress affecting maize (Nagabhushan et al., 2014). Serious downy mildews of maize in Asia include Peronosclerospora sorghi (sorghum downy mildew; SDM) and P. heteropogoni (Rajasthan downy mildew; RDM) in India, P. maydis (Java downy mildew) in Indonesia, P. zeae (brown stripe downy mildew) in Thailand, P. philippinensis (Philippine downy mildew) in the Philippines, and P. sacchari (sugarcane downy mildew) in several Asian countries (George et al., 2003; Prasanna et al., 2010). Despite efforts toward the development of DM-resistant cultivars or seed treatment with metalaxyl fungicide, DM still emerges in localized areas as a severe pathogen (George et al., 2003). Inoculation methods for DM in maize that have been widely used are spraying inoculum and the spreader row technique (Shekhar and Kumar, 2012). Considering that DM is an obligate parasite, the spray inoculation requires abundant, synchronous sporulation within a short time, while the spreader row technique needs consecutive inoculations of the susceptible variety and the test variety (Craig, 1980; Shekhar and Kumar, 2012).

Forward genetics is an attractive tool to improve DM resistance in maize breeding. Several quantitative trait loci (QTL) studies have identified quantitative DM-resistance genes by linkage mapping in maize (Agrama et al., 1999; George et al., 2003; Nair et al., 2005; Sabry et al., 2006; Jampatong et al., 2012). Although the QTLs and population sample sizes were diverse, the QTL analysis located DM resistance–related genomic regions on chromosomes 1, 2, 3, 6, 7, and 10 (George et al., 2003; Jampatong et al., 2012). Because the identification of QTLs so far has been limited by high cost and poor resolution, a combined protocol with positional or map-based cloning is a powerful approach to identify resistance genes against DM (Anderson et al., 2011; Phumichai et al., 2012; Franchel et al., 2013). Positional cloning is a useful tool to identify a gene for a specific phenotype in the absence of sequence information about target products. In view of this advantage, positional cloning can be applied to cross-species hybridization, mutation screening, and CpG island identification (Gallavotti and Whipple, 2015). However, positional cloning is a tedious process that requires several steps. Above all, population mapping or fine mapping of target loci is an integral part of positional cloning (Anderson et al., 2011). Additionally, positional cloning depends on the presence of appropriate genetic markers that have a traceable heredity in company with the target traits (Martin and Hine, 2015). Due to the large genome size of maize, positional cloning in maize is an inefficient approach (Gallavotti and Whipple, 2015). Simko et al. (2013) developed recombinant inbred lines (RIL) and identified DM-resistant QTLs (qDM2.1 and qDM5.1) in lettuce cultivars. Recently, Gallavotti and Whipple (2015) reported successful positional cloning by sh1 mutant screening and developed molecular markers for fine mapping. However, their methods included creating a backcross population and an F₂ population.

Fig. 1.

Molecular linkage map of 59 SSR RFLP loci and location of QTLs. Numbers to the left of the chromosomes indicate the cumulative distance of the consensus map in cM. The consensus map of the IBM2 2008 neighbors was used with MapChart. The DM-resistance QTLs were obtained from the literature (Agrama et al., 1999; George et al., 2003 ; Nair et al., 2005; Sabry et al., 2006; Jampatong et al., 2012; Phumichai et al., 2012; Nagabhushan et al., 2014).

In the present study, we developed a DM inoculation technique for use in the field. For easy inoculation, spreader rows were replaced with pots of susceptible maize. After QTL marker screening and positional cloning, without creating both a backcross and F₂ population, DM-resistant gene candidates were identified by sequence information and quantitative reverse-transcription PCR (RT-PCR).

METHODS AND RESULTS

Plant materials—Five maize nested association mapping (NAM) parent lines—B73, CML228, CML270, Ki3, and Ki11—were used to identify DM resistance–related genes. The NAM parent lines, from the U.S. Department of Agriculture (USDA), were grown at a research field near Phnom Penh, Cambodia. For positional cloning, 3-wk-old leaves were harvested and stored at −80°C until DNA extraction for the QTL analysis. For analysis of transcript accumulation in healthy or DM-infected maize, 6-wk-old leaves were harvested and immersed in RNAlater solution (ThermoFisher Scientific, Waltham, Massachusetts, USA) to stabilize the mRNA until extraction.

Downy mildew inoculation and evaluation—Maize leaves were inoculated with downy mildew from infected plants using the modified spreader row technique described by George et al. (2003). The inoculation was performed twice: in April and September. Susceptible genotypes B73 and CML270 were used in the spreader rows. The spreader rows consisted of two rows with 50 × 25 cm spacing, with 20 seeds planted after every 10th row of the test entries ( Appendix S1 (apps.1600132_s1.pdf)). Seeds of the spreader plants germinated after 7 d, then DW-infected maize was placed in plastic pots in the middle of each experimental block for 2 wk to infect the spreader plants. After removing the DM-infected maize pot, test genotypes were planted with 50 × 25 cm spacing between the spreader rows. After the emergence of the test genotype, the disease reaction was monitored and scored every 7 d for 6 wk. DM resistance was evaluated at weeks 4 and 6 during the scoring period according to the scale used by Craig (1980) and Nagabhushan et al. (2014): 0% infection (no symptom) = highly resistant (HR), 1–10% infection = resistant (R), 11–25% infection = moderately resistant (MR), 26–50% infection = moderately susceptible (MS), 51–75% infection = susceptible (S), and 76–100% infection = highly susceptible (HS).

The DM inoculations were confirmed to result in 100% infection of the susceptible cultivars, B73 and CML270. DM infection began 3–4 wk post-inoculation, then spread rapidly throughout the field. The infected maize stopped growing 6 wk after inoculation, then withered ( Appendix S2 (apps.1600132_s2.pdf)). Although a few infected maize plants grew after the DM inoculation, we observed severely suppressed stamen and pistil development. In the present study, lines CML228, Ki3, and Ki11 had low to very low DM incidence, ranging from 0% to 5% in the April test and from 22.5% to 25% in the September test (Table 1).

According to the weather records in Phnom Penh, Cambodia, in 2015, temperatures were high and humidity low in April to June, while temperatures were low and humidity high in September to October ( Appendix S3 (apps.1600132_s3.docx)). Because slightly low temperatures and high humidity are favorable for the development of DM (Shekhar and Kumar, 2012), we inferred that the lower DM incidence in the April test was related to the high temperatures and relatively low humidity compared with the more favorable conditions later. Eventually, CML228, Ki3, and Ki11 were determined to have a high DM resistance in the two independent tests. By contrast, we identified B73 and CML270 as susceptible cultivars; B73 and CML270 were 100% infected by DM at 6 wk after inoculation in both the April and September tests.

DNA isolation and SSR screening for positional cloning—Genomic DNA was extracted from 0.5 g of fresh leaves using the cetyltrimethylammonium bromide (CTAB) method described by Kim (2013), including washing steps with sodium acetate (Appendix 1). Because QTL marker screening with genomic DNA is not influenced by plant maturity during DM infection, we extracted genomic DNA from 3-wk-old leaves to shorten the experimental period. Simple sequence repeat (SSR) markers reported to be linked to DM-resistant QTLs were selected from the MaizeGDB ( http://www.maizegdb.org) based on preliminary publications (Fig. 1). A total of 48 polymorphic SSR markers or restriction fragment length polymorphism (RFLP) probes from all 10 chromosomes were used for screening. The amplification was performed using a GeneTouch thermal cycler (Bioer Technology, Hangzhou, Zhejiang Province, China) with abm Taq DNA polymerase (Applied Biological Materials, Richmond, British Columbia, Canada). The PCR program consisted of one cycle of 10 min at 94°C for the initial denaturation; 33 cycles of 1 min at 94°C, 30 s at the proper annealing temperature, and 1 min at 72°C; and final elongation for 4 min at 72°C. The PCR products were separated by electrophoresis in a 1.5% agarose gel with 1× Trisacetate-EDTA (TAE) buffer and visualized using ethidium bromide staining. The DM resistance–related region was confirmed by the presence of the PCR bands for the resistant (HR) cultivars and their absence for the susceptible (HS) cultivars. The 59 QTL-linked markers (48 loci) used to identify the DM resistance– related regions and the primer information are shown in  Appendix S4 (apps.1600132_s4.docx). The QTL-linked markers were distributed on the entire chromosome; however, 30 of 48 loci were located on chromosomes 2, 3, and 6 (Fig. 1). PCR amplification revealed 22 polymorphic QTL-linked markers in five maize cultivars ( Appendix S5 (apps.1600132_s5.pdf)). After the PCR screening with the QTL-linked markers, the polymorphism associated with bnlg1702 was found to be identical to the results in Table 1. The bnlg1702 marker amplified a PCR product of approximately 370 bp in CML228, Ki3, and Ki11, but not in B73 and CML270. MaizeGDB was used to find genetic information for SSR marker bnlg1702 on chromosome 6. Sequence information and predicted transcripts were obtained using EnsemblPlants ( http://plants.ensembl.org/index.html). According to MaizeGDB information, bnlg1702 is flanked by umc2321 and AY110873, and the genetic location is estimated to be between positions 146,503,155 and 147,912,501 on chromosome 6 based on the IBM2 2008 neighbors. QTL-linked markers umc2321 and AY110873 were also used in the PCR screening of the five NAM parent lines ( Appendix S5 (apps.1600132_s5.pdf)). Although polymorphisms of umc2321 and AY110873 did not exactly match bnlg1702, umc2321 and AY110873 yielded different PCR products between DM-susceptible and DM-resistant cultivars. An approximately 1.4-Mb fragment was identified for the DM resistance–related region. In all, 47 protein-coding gene annotations were obtained using the MAKER gene annotation pipeline (Cantarel et al., 2008) and contained 63 predicted transcripts for this fragment. Gene-specific primer sets were designed based on the 63 predicted transcripts using Primer3 (Rozen and Skaletsky, 1999) to validate the actual expression of each transcript in healthy or DM-infected maize plants ( Appendix S6 (apps.1600132_s6.docx)).

Table 1.

Disease reactions at four and six weeks after inoculation of five maize cultivars against downy mildew in Phnom Penh, Cambodia.^a,b

Quantitative RT-PCR—Total RNA was extracted from 6-wk-old leaves from healthy or infected NAM parent lines by using TRIzol (Invitrogen, Carlsbad, California, USA). Because the highest DM incidences were observed in 6-wk-old plants, transcripts of the DM resistance–related genes may be abundant in leaf samples from infected 6-wk-old plants. For RT-PCR, cDNA was synthesized from 1 µg of total RNA using a Power cDNA synthesis kit (iNtRON Biotechnology, Seongnam, Gyeonggi, Korea) according to the manufacturer's instructions. Diluted cDNAs were used as a template for quantitative RT-PCR using a gene-specific primer set ( Appendix S6 (apps.1600132_s6.docx)) and the CFX96 Touch Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, California, USA) using the EverGreen 2× qPCR MasterMix (Applied Biological Materials). The amplification was performed as follows: initial denaturation at 95°C for 10 min, followed by 44 cycles of 15 s at 95°C and 1 min at 56°C. Amplicon specificity was examined using a melt curve analysis from 65°C to 95°C. Three replicates of real-time PCR experiments were performed for each predicted transcript. The relative gene expression compared with the control plant (healthy B73) was calculated using the 2^-ΔΔct method (Livak and Schmittgen, 2001). The transcripts significantly upregulated during DM infection in resistant maize cultivars—CML228, Ki3, and Ki11—were selected as putative DM resistance– related transcripts by RT-PCR. Expression of the 63 predicted transcripts was profiled by RT-PCR; however, four predicted transcripts did not amplify in healthy or DM-infected maize plants ( Appendix S7 (apps.1600132_s7.pdf)). Nine putative DM-resistance transcripts were primarily selected, and eventually five transcripts were identified as DM-resistance candidate genes after normalization with the expression level of healthy B73 cultivars (Fig. 2). As compared with transcripts in infected B73, the average transcript accumulation ratios in the five selected candidates were calculated, and then the relative ratio was expressed as a percentage of 100% for B73, 90.5% for CML270, 388.7% for CML228, 781.9% for Ki3, and 329.6% for Ki11 (Fig. 2). According to the results in Table 1, Ki3 was determined to be the most DM-resistant maize cultivar by phenotyping with the spreader row technique. Transcript abundance in infected maize plants was much greater than the average transcript accumulation in healthy B73. The expression ratio for each infected maize cultivar as compared with healthy samples was determined to be 132.5% for B73, 92.7% for CML270, 1049.6% for CML228, 1004.6% for Ki3, and 2348.9% for Ki11.

Identification of DM-resistant gene candidates—Five putative DM-resistant candidate genes were identified (Table 2). GRMZM2G042154 and GRMZM2G098209 encoded uncharacterized proteins, but GRMZM2G042154 was classified as a transporter because it contained a transmembrane domain (NTR1/PTR family). The putative transcript GRMZM2G045236 encoded the bZIP transcription factor (bZIP34). GRMZM2G121565 was identified as a putative Brassinosteroid Insensitive 1-Associated receptor Kinase 1 (BAK1), and GRMZM2G447535 corresponded to pentatricopeptide repeat-containing protein (Ppr). In the microarray by Wei et al. (2012) to profile the expression of 125 maize bZIP genes after pathogen infection, the expression of bZIP34 in the microarray did not change after infection by various fungi. On the other hand, in the present study, GRMZM2G045236 in resistant cultivars was expressed in response to DM infection (Fig. 2). Because expression profiling by microarray is only estimated by prediction and comparison, we expected that bZIP34 in the resistant maize might be expressed following pathogen attack, especially after DM infection. BAK1, a leucine-rich repeat receptor kinase that is known as a key component in brassinosteroid signaling (Nam and Li, 2002; Shi et al., 2013), forms a complex with flagellin sensing 2 (FLS2) and acts as a positive regulator in brassinosteroid signaling and as an initiation signal of inherent immunity (Chinchilla et al., 2007). Previously, Shi et al. (2013) demonstrated that BAK1 is related to powdery mildew resistance using a bak1 mutant in Arabidopsis. Although maize Bak1 is involved in powdery mildew resistance, transcript GRMZM2G121565 might play a role in resistance against DM infection. Ppr is a plant resistance gene analog and an R gene candidate (Sekhwal et al., 2015). Considering that the grape Ppr gene is upstream of the DM resistance gene, Rpv12, which was identified by QTL analysis (Venuti et al., 2013), the Ppr gene has a high possibility for linkage with Rpv12. Laluk et al. (2011) demonstrated the Arabidopsis mitochondrial-localized pentatricopeptide repeat protein (PGN) functions in the defense response against fungi and salt stress. It should be noted that the maize Ppr gene was identified as Rf2 (accession no. AAC49371) on chromosome 9 (Cui et al., 1996; Sekhwal et al., 2015). Because, according to the QTL screening (Fig. 1, 2), Ppr gene GRMZM2G447535 is located on chromosome 6, another maize Ppr gene may be involved in DM resistance. Interestingly, indel polymorphisms were found in both susceptible B73 and resistant Ki3 and Ki11 with regard to sequence alignment ( Appendix S8 (apps.1600132_s8.pdf)). The indel region requires further study to develop a molecular marker for DM resistance.

Fig. 2.

Expression profiles of candidate genes for downy mildew resistance determined using quantitative RT-PCR. Total RNA from leaves from heathy or downy mildew–infected 6-wk-old maize plants was used to synthesize cDNA. The expression intensity was calculated using the 2^-ΔΔCt method (Livak and Schmittgen, 2001), with B73 healthy as the control. The GRMZM2G042154, GRMZM2G045236, GRMZM2G098209, GRMZM2G121565, and GRMZM2G447535 transcripts were annotated as coding for uncharacterized protein (transporter), the bZIP transcription factor (bZIP34), an uncharacterized protein, brassinosteroid insensitive 1-associated receptor kinase 1 (Bak1), and pentatricopeptide repeat-containing protein (Ppr), respectively.

In the present study, we selected DM-resistant maize cultivars using the spreader row technique. As compared with artificial inoculation (inoculum spraying), our new protocol, which improves on the spreader technique, includes an easy inoculation method and is time- and labor-efficient. Rapid evaluation of DM resistance is an outstanding advantage of this protocol because infection was achieved within 3 or 4 wk after inoculation. In a previous study to evaluate 40 maize cultivars, including five NAM parent lines, for DM resistance and incidence, B73 and CML270 were determined as susceptible, while Ki3, Ki11, and CML228 were determined as resistant (Kim et al., 2016). Congruently with our DM phenotyping, B73 and CML270 were found to be susceptible. Additionally, we suggest that the choice of a susceptible cultivar as a spreader plant is a crucial step for the rapid evaluation of DM resistance. Positional cloning with the QTL screening provides a clue to the DM resistance region that is approximately 1.4 Mb long. The predicted transcripts obtained using bioinformatics tools enabled the design of gene-specific primers and led to the identification of DM-resistant gene candidates using quantitative RT-PCR without fine mapping of the bnlg1702 locus. The candidates, namely, bZIP34, Bak1, Ppr, and genes encoding uncharacterized proteins, appear to be involved in DM resistance. However, further detailed research, including an expression study or marker screening of a segregating population and an analysis of a Mu insertion mutant will be needed to obtain definite evidence and to completely verify the DM-resistance mechanism. Even so, our protocol provides useful genetic information to improve the molecular breeding of maize.

CONCLUSIONS

Identification of DM-resistant genes is one of the most important research topics for breeding DM-resistant maize. In this study, we used a combined protocol involving the spreader row technique, QTL screening with phenotyping, and quantitative RT-PCR to identify DM-resistance gene candidates. Our new strategy included field selection using the spreader row technique without mapping or segregation population preparation, identification of the DM-resistance region using positional cloning and bioinformatics tools, and expression-level profiling using quantitative RT-PCR without fine mapping. As whole-genome information becomes available for other crops, our novel protocol should be applicable to other crops and other diseases when suitably customized.

Table 2.

List of DM-resistance candidate genes selected using positional cloning and quantitative RT-PCR.