DUAN Can-xing ,ZHAO Li-ping ,WANG Jie ,LlU Qing-kui ,YANG Zhi-huan ,WANG Xiao-ming

1 Institute of Crop Sciences,Chinese Academy of Agricultural Sciences/National Key Facility for Crop Gene Resources and Genetic Improvement,Beijing 100081,P.R.China

2 Department of Biological Center,Harbin Academy of Agricultural Sciences,Harbin 150028,P.R.China

Abstract The gray leaf spot caused by Cercospora zeina has become a serious disease in maize in China.The isolates of C.zeina from Yunnan,Sichuan,Guizhou,Hubei,Chongqing,Gansu,and Shaanxi were collected.From those,127 samples were used for genetic diversity analysis based on inter-simple sequence repeat (ISSR) and 108 samples were used for multi-gene sequence analysis based on five gene fragments.The results indicated that populations of C.zeina were differentiated with a relatively high genetic level and were classified into two major groups and seven subgroups.The intra-population genetic differentiation of C.zeina is the leading cause of population variation in China,and interpopulation genetic similarity is closely related to the colonization time and spread direction.The multi-gene sequence analysis of C.zeina isolates demonstrated that there were nine haplotypes.Genetic diversity and multi-gene sequence revealed that Yunnan population of C.zeina,the earliest colonizing in China,had the highest genetic and haplotype diversity and had experienced an expansion event.With the influence of the southwest monsoon in the Indian Ocean,C.zeina from Yunnan gradually moved to Sichuan,Guizhou,Shaanxi,Gansu,and Chongqing.Meanwhile,C.zeina was transferred directly from the Yunnan into the Hubei Province via seed and then came into Shaanxi,Henan,and Chongqing along with the wind from Hubei.

Keywords: maize,gray leaf spot,Cercospora zeina,population,disperse routes

1.lntroduction

Maize is a crop with the most production quantity and the second most-harvested area in the world (FAO 2018).China is the second-largest maize producer.In 2018,a total area of 42.13 million hectares was planted with a total output of 257.17 million tons,ranking maize as the first among all crops produced in China (National Bureau of Statistics of China 2019).Maize production,however,is threatened by many pests and diseases,with an average annual reduction of 22.5% (ranging from 19.5 to 41.1%) (Savaryet al.2019).Gray leaf spot (GLS)caused byCercosporasp.is one of the most severe leaf diseases in maize production,which causes a large number of rectangular lesions on leaves (Munkvold and White 2016) and also infects the sheaths and husks.GLS was first recorded in 1924 in Alexander County,Illinois,USA (Tehon and Daniels 1925).During the late 1970s and early 1980s,GLS spread and broke out in the several northern US states (Latterell and Rossi 1983).GLS occurred in Brazil in 1934 but it wasn’t a major disease in maize production until 2001 (Latterell and Rossi 1983;Brunelliet al.2008).In the central and southern regions of Africa,GLS was epidemic in the late 1980s and became the most harmful disease affecting African maize production (Wardet al.1997,1999).In Asia,GLS threatening maize production was first reported in Liaoning Province in northern China (Wuet al.1992).

When susceptible maize plants were infected with the GLS pathogen,there were many spots on leaves causing premature death with stalk collapse,which resulted in 60 to 80% yield loss (Latterell and Rossi 1983;Dhamiet al.2015).From 2012 to 2015,GLS was one of the top 10 most destructive maize diseases in North America.In 2015,the outbreak of GLS in the USA and Canada Ontario,resulting in the yield loss of 6 571 681 tons,accounting for 879.4 USD million (Muelleret al.2016).

Cercospora zeae-maydisandCercospora zeinaare the most important causal agents of GLS.Other species includeC.sorghivar.maydis,Cercosporasp.,C.apii,andC.beticola(Crouset al.2006;Kinyuaet al.2010;Neveset al.2015).Cercospora rodmaniiwas isolated from maize leaves with symptoms of GLS collected in Yunnan Province,China (Zhao 2016).The pathogenC.zeae-maydisis globally distributed,however,C.zeinais regional.For example,whileC.zeae-maydishas spread across America,C.zeinahas only been detected in northeastern USA (Dunkle and Levy 2000;Hsieh 2011).In Brazil,C.zeae-maydiswas found in Sao Paulo,Minas Gerais,and Bahia,whileC.zeinaoccurred in Goias,Sao Paulo,Minas Gerais,and Parana.The two pathogens only appeared simultaneously in the samples of Indianópolis in Minas Gerais (Brunelliet al.2008).In Africa,C.zeinawas the main causal agents of GLS (Dunkle and Levy 2000;Okori 2004;Meiselet al.2009;Kinyuaet al.2010).In Asia,the GLS caused byC.zeinaoccurred in China (Liu and Xu 2013) and Nepal(Dhamiet al.2015).

Gray leaf spot of maize was discovered in China only in the past 30 years.The disease caused byC.zeaemaydisoccurred mainly in the spring maize area of northeastern and northern China and endangered maize production (Luet al.1998,2008).In 2001,sporadic occurrence of GLS caused byC.zeinabegan in the western Yunnan Province of Southwest China.In 2002 and 2003,the outbreak occurred in Dehong and Tengchong,Yunnan (Sun 2005).Since 2007,GLS has spread rapidly to all maize growing regions of Yunnan Province and become a destructive and economically important disease in maize (Liet al.2008;Liuet al.2013;Jiaet al.2013;Zhaoet al.2015).Since then,the prevalence of GLS in Southwest China has become a new threat to maize production (Wanget al.2009,2010;Zhouet al.2011;Liuet al.2013;Zhanget al.2014).In 2019,the GLS caused byC.zeinabroke out in Yulin,the northern region of Shaanxi Province (2 000 km away from Dehong,Yunnan) and the susceptible maize varieties were severely affected (Zhaoet al.2015).In the future,this GLS is likely to spread to the most important corn-producing region of China,the northeast spring corn-planting area,which leads to a considerable risk.

As a disease in move since the 1980s (Latterell and Rossi 1983),GLS has moved rapidly throughout Africa,Asia,and South America over the past two decades.However,there is a lack of research on the disease spread,population characteristics and variation ofC.zeina.In this study,with samples collected from 2008 to 2015,we established the dispersal routes ofC.zeinathroughout China and clarified the factors of the pathogen variation using inter-simple sequence repeats (ISSRs)and multi-gene sequence analysis.The results provided relevant information for controlling GLS caused byC.zeinain China.

2.Materials and methods

2.1.Sample collection and DNA extraction

A total of 165C.zeinaisolates were collected from maize fields with GLS symptoms in Yunnan,Sichuan,Guizhou,Hubei,Chongqing,Shaanxi,Gansu,and Henan in China (Appendix A).Single spore cultures were obtained from diseased leaf spotsviaa singlespore isolating technique (Zhang 1983).All 165 isolates were identified asC.zeinaby morphological and cultural characteristics,histoneH3gene sequence with primer pair CyIH3F/CyIH3R and species-specific primer pairs CzeaeHIST/CyIH3R forC.zeae-maydis,CzeinaHIST/CyIH3R forC.zeina,CmaizeHIST/CyIH3R forCercosporasp.(Crouset al.2006) and CZM2F/CZM2R for detecting speciesC.sorghivar.maydiswith 1 020 bp fragment (Kinyua 2010).A total of 127C.zeinaisolates,with some geographical distance between any two isolates,were selected for population genetic diversity analysis and 108 were used for multi-gene sequence analysis (Appendix A).

The colonies ofC.zeinaon PDA media were cut into 2-mm diameter plugs and added into 100 mL of potato dextrose broth containing 10 to 15 glass beads (5-mm diameter),and were incubated for 15 d at 25°C with continuous shaking (120 r min-1).The mycelia were then filtered and frozen in liquid nitrogen and dried in a vacuum freeze dryer.The dried mycelia were crushed by a proofing machine.Genomic DNA of isolates was extracted using the Sangon Biotech Plant DNA Extraction Kit (Shanghai,China) following manufacturer’s instructions.

2.2.lSSR amplification and multi-gene sequencing

Among the 96 ISSR primers released by the University of British Columbia (Vancouver,Canada),a total of 18 primer pairs (UBC807,UBC808,UBC809,UBC810,UBC811,UBC816,UBC817,UBC823,UBC834,UBC835,UBC836,UBC864,UBC885,UBC888,UBC889,UBC890,UBC891,and UBC895) with specific amplification products and rich diversity were selected and used for genetic diversity analysis.SixC.zeaemaydisstrains from Northeast China were used as a reference in the cluster treemap.PCR were performed on a GeneAmp PCR System 9700 Thermal Cycler (ABI,Norwalk,CT,USA) in 20-μL reaction volumes containing 2.0 μL of genomic DNA,1.0 μL of forward and reverse primers,respectively,0.5 μL dNTP,0.5 μLTaq,2.0 μL buffer,and 13.0 μL dH2O.The PCR reaction was carried out according to the following procedures: 94°C for 5 min for pre-denaturation,94°C for 45 s,48°C for 30 s,72°C for 60 s of 35 cycles,and 72°C for 10 min.The products were observed on 1.5% agarose gel,followed by goldview staining.The fragment size was estimated by using a 100-bp molecular size ladder.

Five gene fragments were chosen for sequence analysis,includingITS(447 bp),β-tubulin(702 bp),Histone H3(375 bp),18sRNA(1 506 bp),andActin(749 bp).ITS1/ITS4 primers were used to amplify the ITS region.Primers for the other four genes were designed using Primer Premier 5 (Lalitha 2000) according to theCercospora β-tubulin,Histone H3,18sRNA,andActingene sequences in NCBI (http://www.broadinstitute.org/).In primers design,we covered as many intron sequences as possible to increase the probability of sequence polymorphism (Table 1).The PCR products were sequenced by Beijing GBI Technology (China).Genes with mutations were sequenced three times to confirm.

Table 1 Primers used for amplified fragments of five genes

2.3.Data analysis

Population genetic diversity was subjected to UPGMA cluster analysis (unweighted group average method)using NTSYSpc Software (version 2.11).The following parameters were calculated using Popgene 32 Software:the observed number of alleles (Na) corresponding to the amplification results of diversity primers,the effective number of alleles (Ne),Nei’s gene diversity index (H),Shannon’s information index (I) within and between groups,Nei’s genetic similarity (GSi),and genetic distance (GD).

The multi-gene sequence analysis was further subjected in ClustalX1.83 for sequence alignment and to verify whether they matched with the target gene sequences,and sequence analysis was followed by artificial correction and shearing with BioEdit.The five genes were spliced asITS-β-tubulin-HistoneH3-18sRNA-Actingenes,and the number of variable sites (S),number of haplotypes(h),haplotype diversity (Hd),nucleotide diversity (Pi),and mismatch distribution of nucleotide (Schneider and Excoffier 1999) in SNP were calculated by DnaSP 5.0(Librado and Rozas 2009).The Tajima’s D,Fu and Li’s D&F and Fu’s Fs were calculated using DnaSP 5.0.The mutations in the sequence were detected to determine whether those were consistent with neutral evolution and population expansion or reduction.Network4.5.0.2 was used to construct all haplotype network branch maps with the Median-Joining Model.The genetic differentiation of fixed parameters (Fst),the number of migrants per generation,and molecular variance analysis (AMOVA)were calculated using Arlequin 3.11.

3.Results

3.1.lntra-and inter-population genetic diversity

Genetic diversity parameters determined in eight geographical populations ofC.zeina,including 127 isolates,are listed in Table 2.At the population level,the values of Na,Ne,H,and I were 1.6759,1.4226,0.2443,and 0.3619,respectively.The results suggested that sinceC.zeinainvaded into China in 2001,the population had a relatively high genetic variance and differentiation after 15 years of proliferation and colonization.By analyzing the intra-population genetic variance from the eight regions,the results showed that Na ranged from 1.3333 to 1.8889,Ne from 1.2597 to 1.5251,H from 0.1421 to 0.3090,and I from 0.2044 to 0.4601,which indicated a relatively high genetic diversity existed among populations in China,including those in Hubei,Yunnan,and Shaanxi was high,with the H values of 0.3090,0.2963,0.2825 and I values of 0.4601,0.4415,0.4231,respectively.The high intra-population genetic diversity may be related to the time of colonization of the fungus and where it came from.The Chongqing population had the lowest genetic diversity with the H and I values of 0.1421 and 0.2044,respectively,consistent with the shortest colonization time.

Based on the GSi and GD generated from ISSR data (Table 3),a comparison between genetic identity and genetic distance showed genetic similarity among eight geographic populations,with the interpopulation GSi ranging from 0.6819 to 0.9642 (average 0.8625) and GD from 0.0364 to 0.3828 (average 0.1514).According to the inter-population genetic similarity data,population similarity ranking of Yunnan population with other regions was as following:Hubei＞Guizhou＞Shaanxi＞Henan＞Sichuan＞Gansu＞Chongqing.Genetic similarity and distance reflected the early and late stages of colonization and the diffusion direction of each geographic population ofC.zeina.The Yunnan and Hubei populations had the highest genetic similarity (0.9339),indicating the genetic backgrounds of these populations were most closely related.Meanwhile,the Yunnan population had the lowest genetic similarity with the Chongqing population (0.6819),which may be due to the last invasion into Chongqing.The Sichuan population had high similarity with Shaanxi (0.9621) and Gansu (0.9604) populations,while the highest similarity occurred between the Gansu and Shaanxi populations,demonstrating a close genetic communication between the three geographic populations.Based on the order of GLS occurrence in these three provinces,the sources of pathogens in Gansu and the west of Shaanxi provinces came from Sichuan.Henan population had high similarity with that in Shaanxi Province (0.9327),indicating thatC.zeinain Henan may have come from the Shaanxi population.It can be inferred thatC.zeinain Yunnanmay be the earliest disease diffusion source and the Hubei population became the second source for east of Shaanxi,Henan and Chongqing.

Table 2 Genetic diversity of Cercospora zeina causing maize gray leaf spot in China1)

Table 3 Genetic identity and genetic distance between eight populations of Cercospora zeina1)

3.2.Phylogenetic analysis

According to the ISSR amplification results,the 127C.zeinaisolates can be classified into two groups and seven subgroups.Group A contained 86 isolates collected from the eight regions,among which subgroup A1 had 63 samples (73%).Group B included 41 isolates from four regions,in which samples from Yunnan and Hubei provinces accounted for 63 and 29%,respectively(Table 4).The Yunnan and Shaanxi populations were distributed among the four subgroups from groups A and B,which were the places with most differentiated subgroups.Three subgroups were identified in the Hubei population.No subgroup differentiation was identified in samples from Chongqing.

Table 4 Genetic grouping of Cercospora zeina isolates based on analyses of ISSR amplification1)

The neighbour-joining tree of 127C.zeinaisolates was constructed with six isolates ofC.zeae-maydisused as reference (Fig.1).The result showed that subgroup A1 could be further divided into three small groups a,b,and c.The group a contained the earliest isolates from Yunnan in 2008 and 2009,and samples from Hubei in 2010,and also the earliest isolates from Sichuan and Guizhou.Small group c also comprised of the earliest isolates from each location.As A1 contained the early isolates from all the geographic populations,it suggested that A1 subgroup is the original population in China,A2 subgroup included only isolates of Hubei,A3 subgroup is formed from 15 isolates of Guizhou,Sichuan,Shaanxi and Gansu,and A4 and A5 subgroups contained only one isolate of Yunnan and Sichuan,respectively.

Fig.1 Dendrogram of Cercospora zeina causing maize gray leaf spot in China.YN,Yunnan;HB,Hubei;GZ,Guizhou;SC,Sichuan;HN,Henan;SX,Shaanxi;GS,Gansu;CQ,Chongqing.

In group B,the B6 subgroup consisted of 26 Yunnan isolates and B7 consisted of 12 Hubei isolates collected in2011 and 2012,and B8 subgroup contained three isolates from each of the three locations,which indicated that the differentiation of group B isolates had the characteristics of geographical populations.Group A could be grouped with group B at the correlation coefficient of 0.450,indicating that genetic variations occurred among group B isolates.

The genetic differentiation in Yunnan population was first observed among isolates collected in 2011 (B6 subgroup),and new subgroups B8 and A4 occurred among those collected in 2012 and 2014.The Hubei population differentiation appeared in 2011 (B6) and 2012(A2).Analysis of subgroup distribution of 37 isolates from Yunnan showed that the genetic differentiation occurred randomly.

3.3.Multi-gene phylogenetic analysis of C.zeina

Differentiation and geographic distribution of haplotypesThe 3 779 bp DNA fragments containing five genes (ITS-β-tubulin-Histone H3-18sRNAActin) were analyzed with 108 isolate sequences.The results indicated that theHistone H3gene had no polymorphic site.However,sequence variations were discovered in the other four genes,including one deletion site,one insertion site,and 21 SNPs,forming nine haplotypes (Table 5).There were five haplotypes in the Yunnan population,four in Sichuan and Gansu,three in Guizhou,two in Hubei and Chongqing,and one in Shaanxi.Haplotype 1 (H1) was the shared haplotype and was distributed in various geographic populations,while haplotype 2 (H2) was observed in six geographic populations (Appendix B).

Yunnan was the original place of GLS caused byC.zeinain China,and the number of haplotypes in each geographic population gradually decreased from Yunnan to northeast direction.Thus,we hypothesized thatC.zeinain other regions came from Yunnan.There were four haplotypes in Sichuan,which was associated with the fact that the GLS occurred earlier there than in other regions.In Guizhou,there were three haplotypes.Three haplotypes were shared in Gansu and Sichuan,the short spatial distance between haplotypes 1 and 7 indicated thatC.zeinain Gansu should come from Sichuan.Hubei and Chongqing had the same haplotypes,but the GLS occurred seven years earlier in Hubei.Based on genetic diversity data and wind direction,the fungus in Chongqing should come from Sichuan and Hubei.Shaanxi had only one haplotype,which may come from the diffused haplotype in the Northeast Sichuan.

Haplotype diversity of C.zeinaIn seven geographic populations,Hd varied from 0.000 to 0.733 and Guizhou had the highest value while Shaanxi had the lowest(Table 5).Pi ranged from 0.00000 to 0.00028,with the highest value in Yunnan and the lowest in Shaanxi.The mean Pi varied from 0.00000 to 1.07563,with the highest value in Yunnan,and the lowest in Shaanxi.As a whole,there are obvious differences among genetic diversity ofC.zeinahaplotype in the seven regions.Yunnan population was the highest,indicating that more genetic variations appeared in Yunnan.Yunnan Province therefore should be the origin ofC.zeinain China.

Population dispersion of C.zeinaAnalysis of the neutral evolution of each geographic population indicated that the values of Tajima’s D,Fu’s and Li’s D,and Fu’s and Li’s F significantly deviated from zero (P＜0.02) in the Yunnan population (Table 6).The Fu’s Fs of all geographic populations was significantly less than zero,indicatingC.zeinain Yunnan population had exhibited expansion.Distribution of nucleic acid mismatches(Fig.2) indicated a unimodal distribution in Yunnan populations,confirming that there have been expansion events.Yunnan was the earliest location where GLS caused byC.zeinaoccurred in China.The relatively long history of GLS occurrence,the complex maize production environment,and diverse maize varieties led to the population expansion ofC.zeinaand then formed a rich genetic diversity.Population expansion events have not been detected in other populations except the Yunnan,which is associated with the relatively short colonization time ofC.zeina.With the disease spreading to maize planting areas with different ecological environments,there may be new events of geographic population expansion in those provinces.

Fig.2 Mismatch distributions of Cercospora zeina populations of China (A) and Yunan Province,China (B).

Table 5 Haplotype diversity indices of Cercospora zeina populations of China

Table 6 Neutrality test of Cercospora zeina populations of China1)

Haplotype network of C.zeinaThe evolutionary haplotype analysis (Fig.3) showed that the inter-population shared haplotype H1 was the central and ancestral haplotype.The other eight haplotypes,however,were differentiated from H1 and were considered to be the young haplotypes.Haplotype H2,H3,H7,H8,and H9 were closely related to the ancestral haplotype.The haplotype H2 was shared by six regions except for Shaanxi,H7 was shared by Sichuan and Gansu provinces,while haplotypes H3,H8,and H9 were unique in Yunnan,Guizhou,and Gansu,respectively.Haplotype H5 from Yunnan was away from the ancestral haplotype and was distributed at the outermost end of the network.As the youngest haplotype,H5 developed 12 years afterC.zeinaentered the Yunnan Province.The results of this analysis suggested that the Yunnan Province was the first region of theC.zeinacolonization in China.

Fig.3 Network based on the haplotypes of Cercospora zeina populations of China.YN,Yunnan;SC,Sichuan;GZ,Guizhou;HB,Hubei;GS,Gansu;SX,Shaanxi;CQ,Chongqing.

Haplotype genetic differentiation and gene flow of C.zeinaHaplotype genetic differentiation and gene flow data (Table 7) indicated that the Fst values ranged from -0.09528 to 0.31407,where there were significant differences between Guizhou and Shaanxi populations and between Gansu and Shaanxi populations (P＜0.05).The Fst values in other geographic populations were not significantly different.Gene flow existed between any two geographical populations,but there was more frequent gene flow between Yunnan population and any other of the six populations.This again suggested that the Yunnan Province maybe the origin ofC.zeinain China.

Haplotype AMOVA of C.zeinaThe haplotype AMOVA(Table 8) showed that genetic variations from intrapopulations were dominant,and the inter-population genetic variations were less significant.The Fst value was not statistically significant (P＞0.1),indicating the genetic differentiation amongC.zeinapopulations was not obvious.This result was consistent with the data based on ISSR amplification.The data showed thatC.zeinaspread northward from Yunnan accompanied by genetic differentiation.TheC.zeinadifferentiation not only exhibited the continuity of population diffusion but also the correlation and particularity of regional differentiation.

Table 7 Pairwise Fst and Nm between Cercospora zeina populations in China1)

Table 8 AMOVA of Cercospora zeina populations of China

4.Discussion

4.1.The causal agents of genetic variation in population of C.zeina in China

At the time of this investigation in 2015,C.zeinahas already been causing GLS in China for 15 years.The disease has spread from the Yunnan Province in Southwest China to Guizhou,Sichuan,Chongqing,West Hubei,Shaanxi,and Gansu,consisting of maize-planting areas with substantial ecological differences.The total spread distance of this disease was about 2 000 km,with an average of more than 100 km per year (Zhaoet al.2015),which was highly consistent with the report from the USA (Latterell and Rossi 1983).The highly infected areas suffered fromC.zeina-caused GLS are primarily located in the southwest maize planting area with high altitudes,such as fields in mountains at 1 200-2 100 m in Yunnan,and at more than 1 000 m in Sichuan Province and Enshi in Hubei Province.In places with medium and low altitudes (500-1 000 m),severe GLS is associated with high rainfall and low temperatures during maize growing season.

Isolates from 2008 to 2015 were collected and subjected to ISSR amplification and multi-gene sequencing.Both results showed that there had been a certain level of variation in the genetic structure of the pathogen population.This variation has not yet reached a significant level among populations,while the genetic differentiations within population were significant,which indicated thatC.zeinapopulations in China originated from the Yunnan Province.The geographic populations were therefore formed by the diffussion ofC.zeinato different regions.AfterC.zeinacolonization occurred in different regions,the various abiotic environments had no directional effects on the population genetics.The intrapopulation variation ofC.zeinais related to other factors.

Organisms adapt to the environment and hostviagenetic mutations (Hugheset al.2008;Smithet al.2020).The factors affecting biological mutations can originate directly from the organism include gene mutation,gene migration,sexual recombination,somatic recombination,and the population size (Burdon and Thrall 2008;Khademiet al.2019).Alternatively,mutations can be generated by multiple factors from complex environments (Nevo 2001).The effects of abiotic environments acted chiefly on the genetic differentiations between populations.For example,temperature and rainfall have effects on the inter-population genetic diversity ofCaragana microphylla(Huanget al.2016).Factors such as latitude,temperature,and water availability could create a regional divergence in wild emmer wheat (Renet al.2013).The genetic structure of maize population and the formation of three geographic populations,Mexico and Southern Andes group,Mesoamerica lowland group,and Andean group,were altered due to long-term geographic isolation and adaptive evolution (Bedoyaet al.2017).In the biological environments,the genetic diversity of the host/parasite has considerable impacts on the genetic differentiation of insect and parasite populations.For example,Phytophthora parasiticaandPhytophthora infestansfrom different host plants have been shown to develop genetic structure differences within their respective populations(Shenet al.2003;Chenet al.2010),and different kinds of apples have corresponding effects on the genetic diversity ofVenturia inaequalispopulations (Fuet al.2010).Glomus intraradices,an arbuscular mycorrhizal fungus,genotypes have been shown to have significant preferences for different host plant species,Glycine max,Helianthus annuus,andAllium porrum,respectively (Crollet al.2008).

The genetic variation ofC.zeinain China mainly occurred within the geographical populations.The formation of genetic diversity ofC.zeinapopulations in China may be related to the diverse maize varieties.The southwest maize-growing region is the most complex ecology area in China.It belongs to the subtropical and temperate zones with a humid and semi-humid climate where maize is planted at an altitude between 200 and 3 000 m,and 22 maize growing sub-regions have developed (Chen 2016).Since one maize variety can only adapt to a small production area,which has led to the most abundant genetic diversity of maize varieties in Southwest China.The sampling area forC.zeinain this study ranged from 24°05′N (Mangshi,YN) to 35°07′N(Tongchuan,SX) and the sampling period was between 2008 and 2015,during which some varieties had also been replaced.The complex background of maize varieties may be an essential cause for the added genetic differentiation withinC.zeinapopulations in China.There are many examples of pathogen coevolution with their hosts,such as on wheat,stripe rust caused byPuccinia striiformisf.sp.tritici,leaf rust infected byP.triticina,stem rust byP.graminisf.sp.tritici,andpowdery mildew induced byBlumeria graminisf.sp.tritici.During the long interactions between wheat varieties and their obligate parasitisms,there is a continual variation in the virulence,aggressiveness,parasitic fitness,and genetic diversity of pathogens (Sharma-Poudyalet al.2013;Aliet al.2017;Liet al.2018;Wuet al.2019;Kolmeret al.2020).The coevolution process of the genetic diversity ofC.zeinaand host has been increasingly apparent.In Yunnan and Hubei populations along with an earlier outbreak of GLS,the B6 and B7 subgroups with a certain population size occurred,respectively,which had distinct genetic differences with initial A1 subgroup.Due to a short period of GLS occurrence in Chongqing,the genetic variation within population is low.Sichuan is also a place with a relatively early GLS occurrence,where disease began in the southern mountains around 2006.We did not collect samples from South Sichuan,and used merely the samples of the Northeast Sichuan for the ISSR analysis,where the disease appeared after 2013,which may lead to the low genetic diversity of Sichuan population.H1 was the major haplotype ofC.zeinain the western part of Sichuan,where GLS was earlier occurred than that in Northeast Sichuan.

4.2.Dispersal routes of C.zeina

The meteorological phenomena,especially the wind,are essential factors for the long-distance transmission of various organisms with aerial dispersal (Closeet al.1978;Felicísimoet al.2008;Gillespieet al.2012;Golan and Pringle 2017).The conidia,ascospores,and uredospores produced by fungi were spreadviashort-distance and long-distance wind dispersal (Brown and Hovm?ller 2002;Corredor-Moreno and Saunders 2020;Leyronaset al.2018).Fungal conidia can be transmitted over long distances through cyclonic winds.For example,the uredospores of sugarcane rust can be transferred from Cameroon in Africa to Dominica in North America,6 500 km away,with the West African Monsoon and African Easterly Jet (Brown and Hovm?ller 2002).Puccinia graminisf.sp.triticithat causes wheat stem rust could also spread between Africa and Asia by wind (Meyeret al.2017).

The climate of Southwest China is mainly affected by the southwest monsoon from the Indian Ocean (Ming 2007),with the monsoon season coinciding with maize growing season.According to the outbreak records of GLS caused byC.zeinain China,Yunnan Province should be the earliest colonized area and the southwest monsoon (occurs from May to July each year) gradually pushes the disease toward north and east (Zhaoet al.2016).Due to the north-south Hengduan Mountains and the subtropical high in the western Pacific,C.zeinaspread faster northward.The molecular analysis of the pathogens also demonstrated the diffusion process and direction ofC.zeina.In the places with early GLS occurrence,the populations exhibited a high genetic diversity while those in locations with late occurrence exhibited the opposite effect.Based on the field investigation and genetic structure analysis,diffusion routes ofC.zeinain China were inferred (Fig.4).The fungus came into Yunnan in 2001,followed by the southern regions of Sichuan in 2006,Hubei in 2007,Guizhou in 2010,Shaanxi in 2012,Gansu and Henan in 2013,and Chongqing in 2014.If the pathogen diffused onlyviathe airflow,it was assumed thatC.zeinaentered Hubei after 2012.However,GLS was discovered in maize at Lichuan of Enshi in 2007.Molecular evidence also suggested that Hubei population was the closest to the Yunnan population,suggesting thatC.zeinahas a new dispersal route.The analysis showed that the source of HubeiC.zeinashould be transmitted by maize seeds propagated at the low and hot valleys of Yunnan Province in winter,at the same time GLS occurred severely and then infected the seeds (He Yueqiu,personal communication).

Fig.4 Dispersal routes of Cercospora zeina in China.

Cercospora zeinaspreads mainlyviaairflow at short and long distances.In general,the pathogen would colonize for some years before the disease outbreaks.The GLS was first recorded in the USA,but a clear record ofC.zeinaappeared in 1998 (Wanget al.1998).In Brazil,GLS occurred in 1934 when it was the secondary disease until it broke out with the appearance ofC.zeina(Brunelliet al.2008).In South Africa,GLS occurred in 1988 and gradually spread to Uganda,Zimbabwe,Zambia,Kenya,and other East Africa countries.The predominant pathogen wasC.zeina,along with a fewC.sorghi(Brunelliet al.2008;Meiselet al.2009;Kinyuaet al.2010;Okoriet al.2015).The GLS was recorded later in Asia (Liuet al.2013;Katwalet al.2013;Dhamiet al.2015;Khaiyamet al.2017).Thus,the spread ofC.zeinacausing GLS may be from Africa (South Africa) to South America (Brazil)to North America (USA) and then to Asia (China).The genetic diversity analysis ofC.zeinapopulations in each continent indicated thatC.zeinain South Africa had high genetic diversity and the estimate of genetic differentiation within populations determined from Nei’s gene diversity was 16.64 (Mulleret al.2016).The genetic similarity inC.zeinaAfrica population was 97.6% (30 strains),in the US population was 98.8% (9 strains) or 94% (12 strains),in the Brazil population was 92% (29 strains-2001/2002) (Wanget al.1998;Dunkle and Levy 2000;Brunelliet al.2008),and in Chinese population is 86%.The genetic similarity between Africa and US populations was 97.6% (Dunkle and Levy 2000),while it was 96.2% between Brazil and USA,and 89%between Brazil and Africa (Brunelliet al.2008).

The comparative study of population genetic diversity,polygene sequence,and whole-genome sequencing have become important tools to explore the global dispersal routes of plants,animals,and microorganisms(Pamidimarri and Reddy 2014;Juhásováet al.2016;Radomskiet al.2020;Song and Cui 2017).Therefore,joint global research will be vital to reveal the process of GLS epidemic caused byC.zeinaand then control it.

Why didC.zeinacause GLS epidemics in many areas of the world? As the two primary pathogens of GLS,was there interspecific competition betweenC.zeinaandC.zeae-maydis? Many researches showed that the two species competed in the field.For example,in the 75 strains collected in Brazil from 2001 to 2002,C.zeinaaccounted for 39.4% whileC.zeaemaydisdid 57.7% (Brunelliet al.2008).However,in the 149 strains in Goiás and Paraná in 2011,the primary maize production area in Brazil,C.zeinaaccounted for 94.0% whileC.zeae-maydisonly did 4.7%,with only two strains ofC.sorghivar.maydis(Neveset al.2015).In 46 isolates collected in Yunnan between 2008 and 2011,the ratio ofC.zeinawas 87.0%,whileC.zeaemaydiswas merely 13.0% (Liuet al.2013).Since 2012,all isolates wereC.zeinaandC.zeae-maydishas not been detected.WithC.zeinabecoming the dominant species,the severity of GLS in the field increased.

To control GLS caused byC.zeae-maydisin northern China,we investigated maize germplasm for resistance to GLS and obtained some resistant accessions,such as CN165,81565,Qi319,Dan9046,and 141 (Wanget al.2014).Those accessions were planted in Yunnan with high GLS incidence to observe the disease resistance.However,those germplasm did not exhibit the high level resistance toC.zeina,indicating thatC.zeinawas more pathogenic thanC.zeae-maydis(Liuet al.2013).We speculate that the genome ofC.zeinaencodes more or stronger pathogenic effectors than that ofC.zeae-maydis,which lead to stronger virulence and aggressiveness inC.zeina.Actually,GLS outbreaks were closely linked to the appearance ofC.zeinapopulations,whether in the eastern USA,Brazil,Africa,or Asian countries.Cercospora zeae-maydishas ever been widely distributed in various maize-growing areas in China,including the southwest,but GLS has merely been a secondary disease until theC.zeinainvasion,leading to severe GLS in these areas.This information indicated thatC.zeinawas gradually replacingC.zeaemaydisin the fields and becoming the predominant fungus,with more robust environmental adaptability and pathogenicity.As a result,the risk will remarkably increase ifC.zeinaspreads further to the remaining maize producing areas in northern China.

5.Conclusion

Using the population genetic diversity and multi-gene sequence analysis,we demonstrated the genetic structure variation had occurred in theC.zeinapopulations in China after 15 years of proliferation.With the main influence of the southwest monsoon from the Indian Ocean,C.zeinaspread from Yunnan to seven other regions.Also,this fungus was spreadviathe long-distance transmission of infected seeds.

Acknowledgements

This work was supported by the China Agriculture Research System from MOAR and MOF (CARS-02)and the Agricultural Science and Technology Innovation Program of the Chinese Academy of Agricultural Sciences(CAAS-ASTIP-2017-ICS).

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on http://www.ChinaAgriSci.com/V2/En/appendix.htm