Ho Sun, Lihong Zhi, Feng Teng,Zhihong Li, Zuxin Zhng,

aNational Key Laboratory of Crop Genetic Improvement,Huazhong Agricultural University,Wuhan 430070,Hubei,China

bMedical College,Hubei University of Arts and Science,Xiangyang 441053,Hubei,China

c Hubei Tenglong Seed Co., Ltd,Xiangyang 441100,Hubei, China

Keywords:

ABSTRACT Gray leaf spot (GLS) caused by Cercospora zeae-maydis and C.zeina is an extremely devastating leaf disease that limits maize production annually.The use of GLS-resistant maize hybrids is the most cost-effective approach for reducing losses.Resistance to GLS is quantitatively inherited in maize (Zea mays L.) and further sources of resistance remain to be analyzed.Here,we detected qRgls1.06,a major quantitative trait locus for GLS resistance in bin 1.06 that explained approximately 55%of the phenotype variance.Fine mapping over 2 consecutive years localized qRgls1.06 to a 2.38-Mb region.Homozygous qRgls1.06WGR/WGR plants in DZ01 background displayed higher GLS resistance and 100-grain weight than DZ01 plants.The GLS responses of several susceptible elite inbred lines were improved by the introduction of qRgls1.06 by marker-assisted backcrossing.Our findings extend the understanding of the genetic basis of resistance to GLS and provide a set of resistant germplasm for genetic improvement of resistance to GLS in maize.

1.Introduction

Gray leaf spot(GLS),a serious foliar disease in maize(Zea mays L.) caused by infection with Cercospora(Cercospora zeae-maydis and C.zeina),was first reported in Illinois in 1925[1].Since the 1990s, GLS has rapidly spread throughout maize-growing areas worldwide and currently poses a huge threat to maize production[2,3].The most cost-effective and environmentally friendly approach to control this disease is the use of resistant hybrids [4].The introduction of resistance alleles into germplasm resources in breeding programs is feasible using marker-assisted selection (MAS) [5].A key prerequisite for MAS, however, is an understanding of the genetic basis of resistance in resistant germplasm resources.

Identification of quantitative trait loci (QTL) for GLS resistance has been of wide interest since the 1990s.Early studies revealed that GLS resistance in maize is a quantitative trait that is inherited in an additive, partial dominant, or dominant manner[6].To date,a number resistance QTL have been mapped on all 10 maize chromosomes using different mapping populations [4,7-15].For example, Juliatti et al.identified four GLS-resistance QTL on chromosomes 1, 3 and 5 in a population derived from a cross of resistant line BS04 and susceptible line BS03[16].Benson et al.identified 16 QTL on chromosomes 1 to 9 using a nested association mapping population [17].In addition, Berger et al.analyzed previous research of GLS resistance QTL and identified five GLS QTL hotspots in bins 1.05-1.06, 2.05-2.06, 4.08, and 5.03 and 5.05[15].In a recent study, a resistance QTL, qGLS8, was finemapped to a ～130-kb region in which a teosinte genome segment had been introgressed into inbred line B73 [18].A recent study conducted in Southwest China where the geographical and climatic conditions are suitable for high incidence of GLS identified a major QLT, qGLS_YZ2-1, in a 2.4 Mb region of chromosome 2[19].These identified QTL laid the foundation for further fine mapping and cloning of resistance genes.

MAS can help improve traits controlled by identified genes or QTL [5,20],and in recent years,significant progress in crop disease resistance breeding has been achieved by MAS.For example,four bacterial blight(BB)resistance genes in rice,Xa-4, Xa-5, Xa-13 and Xa21, were pyramided into a breeding line by MAS, and pyramided lines generally exhibited a higher level of resistance and/or a wider spectrum of resistance to the BB pathogen than predicted from the parental behaviors[21].qMrdd1,a major QTL for resistance to maize rough dwarf disease (MRDD) on chromosome 8, was used in MAS to improve elite maize hybrids, with consequent reductions in grain yield losses due to the disease[22].Although numerous QTL for GLS resistance have been identified few have been used in maize breeding programs.One likely reason for this was that most of the QTL were minor effect QTL that explained only low amounts of phenotypic variation.MAS for a major effect locus would obviously be more effective in genetic improvement.Identification of major resistance loci is therefore of high significance for genetic improvement of GLS resistance.

In the present study, qRgls1.06, a major QTL for GLS resistance, was identified and fine mapped to a ～2.38-Mb region (187.06-189.44 Mb; B73 RefGen-v4.0) on chromosome 1.qRgls1.06 was located in a “hotspot” for GLS resistance identified in previous reports.Furthermore,qRgls1.06 was introduced into several elite inbred lines by MAS to examine its effects on GLS response, flowering time and yield-related traits.We found that qRgls1.06 has excellent potential for improvement of GLS resistance in maize.

2.Materials and methods

2.1.Phenotyping and disease scoring

Phenotypic evaluation of the QTL mapping population,recombinant family lines and backcross progenies for marker-assisted selection were carried out at Yesanguan(31.02°N,110.20°E,altitude:1108 m),and hybrids were tested at Yesanguan and Jianshi (30.36°N, 109.38°E, altitude:1390 m), both in Hubei province.Both locations are mountainous with high humidity and a 7.8-19.4 °C annual average temperature that is suitable for GLS development under natural infection and spread.The pathogen in this area is C.zeina [19,23].Individual plants were scored at the grain dough stage according to the extent of disease symptoms on the three ear-leaves as follows: 1 (highly resistant, lesions accounting 0-5% of the leaf area), 3(resistant, 6%-10%), 5 (slightly resistant, 11%-30%), 7 (susceptible,31%-70%)and 9(highly susceptible,71%-100%)(Fig.1A) [11,19].

2.2.QTL primary mapping

WGR, bred by introgression of chromosomal segments from tropical germplasm into the temperate line L1288, has been widely used as a GLS-resistant inbred line in the maize zone of southwestern China.WGR was crossed to the GLSsusceptible line DZ01, the female parent of Chinese elite hybrid Tenglong No.1.The WGR×DZ01 F1was backcrossed with DZ01 to develop a BC1F1mapping population.A population of 1054 plants, including BC1F1individuals and 60 plants of each parent, was grown at Yesanguan at an altitude of 1108 m in the spring of 2017.The GLS response of each individual was recorded.Separate DNA pools for bulked-segregant analysis sequencing (BSA-Seq) were prepared from 30 individuals with extreme resistant and susceptible GLS responses.The BC1F1population was also used to construct a genetic linkage map and to verify loci identified by BSA-seq.

The DNA quality of the two bulks was evaluated on 1%agarose gels.After quantifying DNA concentrations, genomic DNA (100 ng) was digested with the restriction endonuclease ApeKI.The resulting restriction fragments were ligated to sequencing adaptors for PCR amplification.PCR products were separated in 2% agarose gels.Fragments in the size range of 430-530 bp (corresponding to restriction fragments ranging from 300 to 400 bp) were purified using a Qiagen gel purification kit(Qiagen,Valencia,CA,USA).Finally,both DNA libraries were sequenced on a HiSeq4000 platform with 150-bp paired-end reads by GenoSeq(Wuhan,China).After removing adaptor reads and low-quality reads from the raw data, the resulting clean reads were compared with B73 RefGen V4(http://plants.ensembl.org/Zea_mays/Info/Index) using version 0.7.5a-r405 of the Burrows-Wheeler alignment tool [24].The reads were sorted with SortSam in Picard v1.91, and variant calling was conducted using GATK v3.7 in HaplotypeCaller mode [25].A set of high-quality SNPs and insertions/deletions(InDels)with minor allele frequency ≥0.2,relative heterozygosity ≤0.2, and missing genotype data ≤0.2 was obtained.Finally,the mean values of Δ2(SNP index)were calculated in a sliding window(2-Mb window size,20-kb step length).After SNP and InDel calling, we calculated the SNPindex and Δ (SNP-index) to identify candidate GLS-related QTL.The SNP index refers to the proportion of reads harboring a GLS-resistant parent genotype in the total reads.The average SNP-index of SNPs in a specific genomic interval was calculated in a sliding window (2-Mb window size, 20-kb step length).Δ (SNP-index) was the difference between average SNP-index of the resistant bulk and that of the susceptible bulk.Statistical confidence intervals of Δ (SNPindex) were calculated under the null hypothesis of no QTL(P＜0.01)[26],and the corresponding Δ(SNP-index)2graph was plotted.

Fig.1–Phenotypic evaluation of gray leaf spot(GLS).(A)GLS symptoms on a scale of 1 to 9.(B and C)GLS severity on a WGR plant(B)and leaf(C)at Yesanguan in spring of 2017.(D and E)GLS severity on an F1 plant(D)and leaf(E).(F and G)GLS severity on a DZ01 plant(F) and leaf(G).(H)Comparison of leaf disease scores among WGR, DZ01 and their hybrid.Significant differences were determined by one-way ANOVA.Sample size,60.***,P＜0.001.

2.3.Fine mapping of QTL qRgls1.06

A total of 780 BC1F1individuals was planted at Sanya (18°N,109°E)in Hainan province in the winter of 2017.Eighty BC1F1individuals heterozygous at the qRgls1.06 locus selected by genotyping with flanking markers (Table S1) were crossed with DZ01 to develop a BC2F1population.Genomic DNA extracted from the endosperms of 11,000 BC2F1seeds using 0.1 mol L?1NaOH was amplificated by PCR using linked markers M3 and M5 and PCR products were separated by a Fragment Analyzer (Advanced Analytical Technologies, Inc.,USA).In the spring of 2018 the seeds of 95 recombinants were sown at Yesanguan for phenotyping and further fine mapping.All 95 recombinants were genotyped (Table S1)and self-pollinated for progeny testing.In addition, the genetic backgrounds of 30 BC2F1individuals heterozygous for qRgls1.06 were genotyped using 187 widely dispersed polymorphic markers.The individual with the highest background recovery rate (92.25%) was then crossed with DZ01 to create a BC3F1population (Table S2).BC3F1plants heterozygous for the qRgls1.06 interval were self-pollinated at Sanya in the winter of 2018 to form a BC3F2population from which qRgls1.06 homozygotes were selected by genotyping.These improved DZ01 genotypes are referred to as DZ01RR.

2.4.Marker-assisted improvement of inbred lines and hybrids

DZ01RRindividuals (the improved female parent of Tenglong No.1)were crossed with SL06(the male parent of Tenglong No.1) to produce an improved hybrid Tenglong No.1.In the spring of 2019, the original and improved Tenglong No.1 hybrids were grown at Yesanguan and Jianshi (in Hubei province, 30°N, 109°E, altitude: 1390 m) to evaluate GLS response, flowering time and grain yieldrelated traits.For assessment of GLS response and agronomically important traits,358 BC3F2individuals and selfed progenies of the 95 recombinants were grown at Yesanguan.Inbred lines, WGR, DZ01, SL06 and DZ01RRand the original and improved Tenglong No.1 hybrids were planted in 5-row plots with 0.67 m spacing among rows and 0.3 m spacing among plants.Each selfed family was planted as a 2-row plot.

MAS for improved GLS resistance was conducted using WGR as the donor of the desirable qRgls1.06 allele,with seven elite inbred lines, namely, Zheng 58, Chang 7-2, Sheng 62,Sheng 68, N6427, N6429 and N6439, serving as receptors(recurrent parents).In the spring of 2018, WGR was crossed with each of the seven elite inbred lines at Wuhan (30°N,114°E).In the winter of 2018, 40 individuals of each F1population were planted at Sanya, and crossed with the corresponding recurrent parent.In the spring of 2019, 150 individuals of each BC1F1population were phenotypically evaluated at Yesanguan.Markers linked with qRgls1.06 that were polymorphic between WGR and each recurrent parent(Table S1) were developed in order to genotype all BC1F1individuals, and the GLS responses of all individuals were recorded.

2.5.Genetic analysis

Each individual in the BC1F1population was classified as resistant or susceptible based on pooled GLS scores (1, 3, and 5)were assigned to the R group,and 7 or 9 were categorized as the S group.

2.6.DNA preparation and marker development

Genomic DNA was extracted from leaves of plants at the 7-8-leaf stage using a modified CTAB method, with one additional purification step involving chloroform isoamylalcohol applied to obtain high-quality DNA [27].For BSASeq,the quality of DNA samples was tested by 1%agarose gel electrophoresis.An alkaline lysis method was used to extract endosperm DNA from maize seeds.InDel markers were developed from the BSA-Seq data, and primers were designed with PRIMER3 Plus software (http://122.205.95.20/cgibin/primer3plus/primer3plus.cgi) using the genomic sequence of B73 RefGen V4 (http://ensembl.gramene.org/Zea_mays/Info/Index) as a reference.One thousand simple sequence repeat(SSR)markers from the IBM 2008 Neighbors map as well as the newly developed markers were used for parental polymorphism screening in the Fragment Analyzer and genetic distance estimation in MAPMAKER v3.0 [28].Recombination fractions were converted to map distances in centimorgans (cM) using the Kosambi mapping function[29].

3.Results

3.1.A partially dominant locus for GLS resistance in WGR

Phenotypic evaluation of WGR,DZ01 and the F1population at the early milk stage in 2017 revealed that leaves of WGR developed a few fleck-type lesions (Fig.1C), a mean score of 1.33 ± 0.74 (n = 60) (Fig.1B, C, H), showing that GWR was strongly resistant to GLS.In contrast,leaves of DZ01 displayed necrotic spots, gray to tan in color and distinctly rectangular in shape and covering almost the entire leaves(Fig.1F,G,H),a mean score 8.60±0.80(n=60)at the late dough stage(Fig.1H).The disease score of F1(heterozygous) individuals was 2.76 ±1.10 (n = 60), intermediate between the parents but skewed towards the resistant parent (Fig.1D, E, H).Thus the GLS resistance of WGR was partially dominant.

Among 1054 BC1F1plants, 521 had disease scores of 1 to 5 and were accordingly assigned to the R group, whereas 533 plants with disease scores of 7 or 9 were placed in the S group.These figures obviously fitted a 1:1 ratio indicative of variation at a single locus.We concluded that WGR has a partially dominant allele conferring resistance to GLS.

3.2.Mapping of qRgls1.06 to a 2.38-Mb genomic region

After quality control, 292,326,247 (BR) and 238,872,074 (BS)clean reads covering 86.12% (BR) and 84.85% (BS) of the genome, respectively (Table S3), were obtained from the two libraries.The SNP frequencies (SNP index) of the two pools were calculated across the whole genome.The 99% quantile threshold was obtained by sorting the Δ2(SNP index) from small to large.According to the null hypothesis, we chose peak regions above the threshold value(0.17)as the candidate QTL interval.One QTL for GLS resistance, designated as qRgls1.06, was identified within an 18.42-Mb(176.56-194.98 Mb) region in bin 1.06 of chromosome 1 (Fig.2A).In addition, 1000 SSR markers from the IBM 2008 Neighbors map were screened in the parental lines.Among them, 19 markers on chromosome 1 were polymorphic between the two parental lines, and nine (Table S1) were selected to reconstruct a genetic linkage map.In addition,we developed InDel marker M4 (Table S1).These 10 polymorphic markers within the qRgls1.06 interval were subsequently used to genotype 1054 BC1F1plants and to reconstruct the genetic linkage map using MAPMAKER software.Composite interval mapping using winQTLcart2.5 software uncovered a QTL flanked by umc1972 and umc2560 that co-localized with qRgls1.06.This QTL had a high LOD score (26.51), exerted a large additive effect on the disease score(?2.70)and explained 55.77% of the phenotypic variance (Fig.2B), thus suggesting that the genomic interval flanked by umc1972 and umc2560 was a high confidence region for resistance to GLS.All BC1F1individuals were classified using three molecular markers(umc1972, M4, and umc2560) within qRgls1.06; 498 plants genotyped qRgls1.06WGR/DZ01had an average GLS score of 4.73±1.85,which was significantly different(P=1.01×10?116)from 7.71 ± 1.69, the mean score exhibited by the 495 plants with qRgls1.06DZ01/DZ01.

We designed an additional seven InDel markers within the umc1972-umc2560 interval based on variants called from the BSA-seq data(Table S3).Using flanking markers umc1972 and umc2560 and all eight InDel markers,including M4(Table S1),we identified 61 recombinants in the umc1972-umc2560 interval in the BC1F1population.These recombinants resulted from 10 crossover events,referred to as RL-1 to RL-10 (Fig.3A and B).The disease scores of RL-1 to RL-6 were significantly lower than that of DZ01 (P＜0.001) and were also obviously lower than those of RL-7 to RL-10 (P＜0.05) (Fig.3C).Three informative recombinant chromosomes, RL-4, RL-6 and RL-9,delineated qRgls1.06 within the M3-M5 interval of approximately 6.02 Mb (184.16-190.18 Mb) in B73 RefGen-v4.0.Using markers M3 and M5, we genotyped 11,000 BC2F1seeds and found 95 recombinants.In addition, we developed six InDels(designated M9-M14)within the M3-M5 region(Table S1).The 95 recombinants were genotyped using nine markers,namely,M3,M4,M5 and M9-M14,and were classified according to nine crossover events referred to as RL-11 to RL-19 (Fig.3D).Progeny testing, performed to evaluate the disease scores of these recombinant chromosomes, uncovered a significant difference in GLS response between qRgls1.06WGR/WGRand qRgls1.06DZ01/DZ01genotypes from the progeny families of RL-11 to RL-14 (P＜0.001) (Fig.3D and E); however, there was no significant difference in GLS response between qRgls1.06WGR/WGRand qRgls1.06DZ01/DZ01genotypes in the progeny families of RL-15 to RL-19 (P ＞0.05) (Fig.3E).Consequently, qRgls1.06 was narrowed down to a 2.38-Mb (187.06-189.44 Mb, B73 RefGen-v4.0) genomic region flanked by markers M13 and M4 close to the centromere in chromosome 1.

Fig.2– QTL primary mapping by bulked-segregant analysis sequencing and linkage mapping.(A)QTL primary mapping by bulked-segregant analysis sequencing.The value of Δ2(SNP index)was calculated using a window size of 2 Mb and a step length of 20 kb.The x-axis shows the physical distance of the 10 maize chromosomes,whereas the y-axis indicates the Δ2(SNP index)value.The dotted line represents the screening threshold(0.17),and physical position corresponding to the value of Δ2(SNP index)exceeding the threshold is the candidate interval.(B)QTL identification by linkage mapping.The x-axis represents the genetic distance(cM)among markers,and the y-axis corresponds to the likelihood of linkage(LOD)score.The confidence interval was flanked by umc1972 and umc2560.

3.3.Strong effect of qRgls1.06 on 100-grain weight in line families and the hybrid in the presence of disease

Flowering time is an agronomically important trait in crop plants.Earlier studies suggested that GLS response in maize was associated with flowering time [14,17,30].To explore the effect of qRgls1.06 on flowering time,days to tasseling,days to silking and days to anthesis, both qRgls1.06WGR/WGRand qRgls1.06DZ01/DZ01genotypes were examined.No significant difference was observed in the three flowering-time traits between the two genotypes, indicating that qRgls1.06 had no obvious effect on flowering time(Fig.S1A).To test the effects of qRgls1.06 on other important traits, namely, plant height,ear height,tassel length,tassel branch number,ear length,ear diameter, kernel row number, and 100-grain weight, both qRgls1.06WGR/WGRand qRgls1.06DZ01/DZ01were examined in a BC3F2population.A significant difference was detected in GLS response and 100-grain weight between qRgls1.06WGR/WGRand qRgls1.06DZ01/DZ01(Fig.4A and B), but no differences between the two genotypes were observed for the other traits(Fig.S1AJ).These results suggest that qRgls1.06 has a strong effect on 100-grain weight,a critical grain-yield component trait.

To examine the effect of the desirable qRgls1.06 allele in hybrids, we selected the DZ01RRfamily line harboring qRgls1.06WGRin DZ01 background and crossed it with inbred line SL06, which has a susceptible allele at qRgls1.06, to form the DZ01RR/SL06 hybrid.Compared with the original hybrid DZ01/SL06, the hybrid DZ01RR/SL06 exhibited significantly higher GLS resistance (with disease scores of 3.06 ± 1.12 at Yesanguan and 3.66 ± 1.09 at Jianshi) and higher 100-grain weight (39.26 ± 1.51 g at Yesanguan and 37.56 ± 1.06 g at Jianshi) across two environments (Fig.4C and D), but with no significant differences in flowering time and grain yieldrelated traits (Fig.S2A-D).The grain yield of DZ01RR/SL06 was 12,949.56 kg ha?1at Yesanguan and 12,718.27 kg ha?1at Jianshi,approximately 22.0%higher than that of DZ01/SL06 at the two locations (10,577.85 kg ha?1at Yesanguan and 10,427.31 kg ha?1at Jianshi) (Fig.4E).These results demonstrated that qRgls1.06 improved resistance to GLS and thereby increased the grain yield of the hybrid.

3.4.Improvement GLS resistance in elite inbred lines using the favorable allele of qRgls1.06

To evaluate effects on GLS response in different genetic backgrounds we introduced the qRgls1.06 allele of WGR(qRgls1.06RR) into seven susceptible lines (qRgls1.06rr) by marker-assisted backcrossing (Table S1).Heterozygous qRgls1.06Rrand homozygous qRgls1.06rrgenotypes were identified in each BC1F1population, and the disease grades of these individuals were scored.There were significant differences in GLS scores between qRgls1.06Rrand qRgls1.06rrindividuals in each BC1F1population (Figs.5; S3A-J), suggesting that the qRgls1.06Rallele was effective across a range of genetic backgrounds.

4.Discussion

Identification and characterization of genes for resistance to GLS has been an ongoing goal of maize researchers.Dozens of QTL for GLS resistance have been identified through linkage mapping and association studies.Several of these QTL were co-localized in bins 1.05-1.06.For example, Lehmensiek et al.identified a major GLS resistance QTL in bins 1.05-1.06 [4],whereas Balint-Kurti et al.and Pozar et al.separately found this QTL in IBM and NIL-derived populations [10,31].Zwonitzer et al.[11]suggested that the QTL in bin 1.06 pleiotropically conferred resistance to GLS, northern leaf blight and southern leaf blight.Using a nested association mapping population,Benson et al.identified 16 GLS resistance QTL [17], including one co-localized with QTL reported in bin 1.06 by Saghai-Maroof et al.[8]and Clements et al.[32].In a genome-wide association study, Shi et al.similarly identified 31 GLS-resistance loci in a set of inbred lines, two of which(qGLS1.05-1 and qGLS1.05-2)were found in bin 1.05[13].Metaanalysis can be used to integrate loci reported from multiple studies to highlight the GLS-resistance-associated bin 1.05-1.06 region harboring a “real QTL”[15,33].By integrating the QTL in this study with those mentioned above, we found that the major QTL qRgls1.06 was not only well co-localized with loci from previous studies but was also mapped into a narrower interval (Fig.S4).Importantly, qRgls1.06 accounted for 55.77% of the phenotypic variation and exhibited partial dominance; this mode of inheritance was rarely observed for GLS-resistance loci in previous studies, suggesting that qRgls1.06 in line WGR is a more desirable allele that provides a larger phenotypic contribution to GLS resistance.Even so,approximately 45%of the phenotypic variance in the mapping population was not explained by qRgls1.06, implicating additional minor QTL that were not identified by BSA-seq in this study.Our findings therefore lay a foundation for gene cloning and use of the WGR allele in breeding.

Fig.3–Fine mapping of qRgls1.06.(A,B, and D)Genotypes of recombinant individuals.(C)Phenotypes of recombinant individuals.(E)Phenotypes of two offspring with homozygous genotypes.The genetic structure of each genotype is depicted by gray and white rectangles,corresponding to heterozygous qRgls1.06WGR/DZ01 and homozygous qRgls1.06DZ01/DZ01 genotypes,respectively.Histograms represents the average disease scores.The average disease score of recombined individuals is shown in(C),and the black and white bars in (E)indicate the average disease scores of qRgls1.06WGR/WGR and qRgls1.06DZ01/DZ01 genotypes, respectively.Data are shown as means±s.d.N,number of measured individuals.Significant differences were assessed by Student’s t-tests.***,P＜0.001;NS,not significant.

Fig.4–Comparison of disease scores and 100-grain weight in different populations.(A and B)Disease score(A)and 100-grain weight(B)of homozygous qRgls1.06WGR/WGR and homozygous qRgls1.06DZ01/DZ01 plants in a BC3F2 population.(C,D, and E).Disease score(C),100-grain weight(D)and grain yield(E)of hybrids DZ01RR×SL06 and DZ01×SL06 at Yesanguan and Jianshi.Phenotypic data are shown as means±s.d.N,number of measured individuals.Significant differences were assessed by oneway ANOVA.***, P＜0.001.

Fig.5– Effect of the qRgls1.06 allele from WGR on GLS response in diverse genetic backgrounds.Seven elite inbred lines,namely Zheng 58,Chang 7-2,Sheng 62,Sheng 68, N6427,N6429,and N6439,were subjected to marker-assisted selection.Phenotypic data from qRgls1.06Rr and qRgls1.06rr genotypes in different BC1F1 populations are shown as means±s.d.n,numbers of individuals assessed.Significant differences were determined by Student’s t-tests.***,P＜0.001.

Although qRgls1.06 was mapped to a 2.38-Mb genomic region, further isolation of the gene underlying qRgls1.06 will be difficult.This QTL interval is proximal to the centromere region where crossover formation is normally inhibited.Furthermore, WGR is a temperate inbred line with partially tropical maize kinship, inferring a large genetic difference from B73.The 2.38-Mb genomic region corresponding to B73 RefGen V4 which is not fully equivalent to WGR in genome size and gene content.In B73 RefGen V4,the 2.38-Mb qRgls1.06 region contains 44 annotated genes, including five wallassociated kinase(WAK)genes.WAKs are receptor-like kinase(RLK)subfamily proteins with documented roles in biotic and abiotic stress responses in plants [34,35].In Arabidopsis thaliana, over-expression of WAK1 enhances plant resistance to Botrytis cinerea [36].In Oryza sativa L., expression of the OsWAK1 gene is induced by Magnaporthe grisea,and transgenic lines overexpressing OsWAK1 show improved resistance to rice blast[37].In addition,resistance to rice blast is positively regulated by OsWAK14, OsWAK91, and OsWAK92 but is negatively regulated by OsWAK112d [38].In maize, ZmWAK and ZmWAK-RLK1 are related to resistance to head smut and southern leaf blight, respectively [39,40].In summary, the WAK gene family greatly contributes to plant disease resistance.Bioinformatics analysis has revealed that the five WAK genes underlying qRgls1.06 could encode proteins with both galacturonan-binding and serine/threonine kinase domains,that are required for perception and transduction of extracellular signals[39,40].Considering the important role of the WAK gene family in plant disease resistance, we suggest these WAK genes are potential candidates for further research,although other candidate genes are not excluded.

Resistant germplasm resources are the basis for breeding resistant maize hybrids.Cercospora zeae-maydis and C.zeina are well known as causal fungi of GLS in maize [1,41]but the two species have different geographical distributions.While C.zeae-maydis is the predominant pathogen in much of the USA, C.zeina occurs in the eastern corn belt of that country[41].In China,C.zeae-maydis appears to be the causal agent in the northeastern corn planting area, whereas C.zeina predominates in the southwest [26].Unfortunately, germplasm resources with high resistance to GLS are rare.In the present study,we found that inbred line WGR had high GLS resistance largely attributable to a single, partially dominant resistance allele at qRgls1.06, which explained more than 50% of the phenotypic variance.The level of GLS resistance conferred by qRgls1.06WGRallele was stable across seven inbred lines,demonstrating its potential for breeding.Line WGR should serve as a useful donor for improvement of GLS resistance in inbred lines and hybrids.

CRediT authorship contribution statement

Zuxin Zhang conceived and designed the experiments.Hao Sun and Zhihong Li performed the experiments.Hao Sun,Feng Teng,and Lihong Zhai analyzed the data.Feng Teng and Hao Sun contributed reagents/materials/analysis tools.Hao Sun and Lihong Zhai wrote the paper.

Declaration of competing interest

Authors declare that there are no conflicts of interest.

Acknowledgments

This work was supported by the National Key Laboratory of Crop Genetic Improvement Self-Research Program (ZW18B0101), the China Scholarship Council (201908420122), the Teachers’ Scientific Ability Cultivation Foundation of Hubei University of Arts and Science (PYSB20201001), and the Xiangyang Youth Science and Technology Talent Development Plan.

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2020.08.001.