molecular breeding for maize improvement: an...

Indian Journal of BiotechnologyVol 2. January 2003, pp 85-98

Molecular Breeding for Maize Improvement: An Overview

B M Prasanna'" and D Hoisington/'Division of Genetics, Indian Agricultural Research Institute, New Delhi 110 012, India

21nternational Maize and Wheat Improvement Center (CIMMYT), Apdo. Postal 6-641, 06600 Mexico, D F, Mexico

The maize genome is one of the most extensively analyzed among the plant genomes. Consequently, maize hasbeen at the forefront in development and evaluation of an array of molecular markers for varied purposes in geneticsand breeding. Besides the well-demonstrated utility of molecular markers in genotype differentiation and analysis ofgenetic diversity in maize germplasm, application of DNA-based markers is also of considerable significance totropical/sub-tropical maize production systems, such as in India, for mapping and marker-assisted selection forresistance to major biotic/abiotic stresses affecting production and productivity. Significant impetus in this directionhas been provided in recent years through the Asian Maize Biotechnology Network (AMBIONET). This articleprovides an overview of the recent efforts under AMBIONET in relation to: (i) the molecular characterization ofinbred lines developed by various public sector institutions in India; (ii) the analysis of genetic diversity in the Indianmaize germ plasm using microsatellite markers; and (Hi) the mapping of quantitative trait loci conferring resistanceto different downy mildews affecting maize in tropical Asia. Judicious integration of conventional and molecularapproaches in maize breeding programmes is vital for efficient utilization of genetic resources, and improving theproduction and post-harvest characteristics of the elite germplasm. This shall, in turn, require further strengtheningof synergistic linkages and partnerships among national and international research institutions to harness therapidly emerging information and technologies related to molecular breeding in maize.

Keywords: Zea mays, markers, fingerprinting, genome mapping, marker-assisted selection

IntroductionMaize (Zea mays Linn.) holds a unique position in

world agriculture as a food, feed and industrial croppar excellence. Although the developed countries,particularly USA, contribute predominantly to themaize production, demand for maize in developingcountries is expected to surpass the demand for bothwheat and rice by the year 2020 (Pingali & Pandey,2001). However, average productivity of maize inseveral developing countries is still considerably low.About 45 million hectares of maize is grown in thelowland tropics, where a range of climatic, biotic andabiotic constraints severely affect productivity. The

"A uthor for correspondence:Tel: 25824285; Fax: 25766420E-mail: [email protected]:AFLP: Amplified fragment length polymorphism; EST:Expressed sequence tag; GS: Genetic similarity; MAS: Marker-assisted selection; PAGE: Polyacrylamide gel electrophoresis;peR: Polymerase chain reaction; PIe: Polymorphism informationcontent; QPM: Quality protein maize; QTL: Quantitative traitloci; RAPD: Random amplified polymorphic DNA; RFLP:Restriction fragment length polymorphism; RILs: Recombinantinbred lines; SNP: Single nucleotide polymorphism; SSR: Simplesequence repeats.

challenges are diverse and complex, and there is nosingle technological solution. While significantprogress has been made in relation to maizeimprovement in India using traditional breedingstrategies (Dhillon & Prasanna, 2001), considerablescope exists to further enhance maize productivity.Modern molecular tools and techniques cancomplement conventional approaches to allowbreeders to effectively address priority research areas.

The term 'molecular breeding' is now popularlyused for the utilization of molecular (DNA-based)tools, including markers, to enhance the efficiency ofthe breeding process. DNA markers have the potentialto aid plant breeding programmes through diverseways, such as fingerprinting of elite genetic stocks,analysis of genetic diversity, and increasing theefficiency of selection for difficult traits. Among thearray of DNA-based markers available to plantscientists, the ones most commonly used are RFLPs,RAPDs, SSRs and AFLPs. Excellent reviews areavailable discussing the genetic bases of variousDNA-based markers, the means for detectingmolecular polymorphism, and the strengths andconstraints associated with different markers for

86 INDIAN J BIOTECHNOL, JANUARY 2003

Table l--Characteristics and utility of different molecular markers for applied molecular genetics in crop plants

RFLPs RAPDs SSRs AFLPs SNPs

Fingerprinting ++ -/+ ++ ++ +++Genetic diversity ++ ++ + ?

Tagging qualitative genes ++ ++ + +++ ++Mapping polygenic traits ++ -/+ + ++ ++Marker-assisted selection ++ ++ +/++ ++Comparative genome ++mapping

Principle Endonuclease DNA amplification PCR of simple Endonuclease DNArestriction; with random sequence repeats restriction; sequencing

Southern blotting primers amplification usingadapters and

specific primers

Type of polymorphism Single base changes; Single base changes; Changes in Single base changes; Single baseInsertions/Deletions Insertions/Deletions length Insertions/Deletions di fferences

of repeats

Genomic abundance High Very high Medium Very high Very high

Level of polymorphism Medium Medium High High Very high

Inheritance Codominant Dominant Codominant Dominant..? Codominant

Detection of allelic variants Yes No Yes No Yes

o. of loci detected 1-5 1-10 30-100

Need for sequence 0 0 Yes No YesinformationTechnical difficulty Medium Low Low Medium/High Medium/High

Reliability High Intermediate High Medium/High High

Quantity of DNA required 2-15~g 10-50 ng 2-15 ng 2-15 ng 2-15 ng

Use of radioisotopes YeslNo No YeslNo YeslNo No

Probes/pri mers required gDNA/cDNA Random 9- or Specific 16- Speci fic adapters SpecificlO-mer 3D-mer and primers primers

oligonucleotides primers

Start-up costs Medium Low Medium High High

Development costs Medium Low High Medium/High Medium/High

various applications (Karp et al, 1997; Liu, 2002).The most appropriate marker(s) for a particularapplication will depend on the target crop, it'sbreeding behaviour, specific objectives of theexperiment, the resolution required, and theoperational/financial constraints, if any. A comparisonof the various marker systems in relation to theircharacteristics and applicability is provided inTable I. For example, for genetic linkage mapdevelopment, any type of molecular marker may beused. However, codominant markers (e.g., RFLPs,SSRs or SNPs), provide more genetic information inF2 and backcross generations than markers detecting

predominantly presence/absence or dominantpolymorphisms (e.g., RAPDs or AFLPs). Forcomparative mapping within and across crop species,the use of RFLP as anchor loci are the best choice asthey detect evolutionarily conserved loci in a morepredictable manner than loci detected byhypervariabJe SSRs and AFLPs.

Among the different types of PeR-based DNAmarkers available for diverse applications inmaize breeding, SSR markers are often preferredfor reasons of cost, simplicity and effectiveness. SSRmarkers are robust, codominant, hypervariable,abundant, and uniformly dispersed in plant genomes

PRASANNA & HOISINGTON: MOLECULAR BREEDING IN MAIZE

(Powell et at, 1996a,b). In maize, more than 1000mapped SSR markers are available in the publicdomain (MaizeDB; http://www.agron.missouri.edu).Mogg et at (1999) showed that by sequencing theflanking regions of maize microsatellites, a SNPcould be found every 40 bp. Given that the maizegenome is estimated to be 2.5 x 109 bp in size, there isa potential for up to 62 million SNPs in maize. Withthe recent initiation of a large-scale EST sequencingprogramme in maize (http://www.zmdb.iastate.edu/zmdb/EST_project.html), a new and potentially richsource of SNPs has been uncovered (Edwards &Mogg, 2001). While both SSRs and SNPs can bereliably applied on a large scale with only smallquantities of DNA required for PCR amplification,SNPs are highly amenable for automation, and there-fore, offer significant advantages for plant breedingpurposes. SSRs, however, are the preferred choicewhen codominant, multi allelic information is required,or when the infrastructure and resources are limited.

Applications of Molecular MarkersMolecular markers are increasingly being adopted

by researchers involved in crop improvement as aneffective and appropriate tool for addressing severalbasic and applied research areas relevant toagricultural production systems (Mohan et at, 1997;Prioul et al, 1997). DNA fingerprinting and geneticdiversity analysis using molecular markers is ofsignificant utility in effective management ofgermplasm collections (Warburton & Hoisington,2001). Increasing emphasis is also being placed oncomprehensive analysis of genetic diversity inbreeding materials of major crops (Mohammadi &Prasanna, 2002). Accurate assessment of the levelsand patterns of genetic diversity using molecularmarkers is particularly helpful in maize breeding for(i) maintenance and broadening of the genetic base ofthe elite germplasm; (ii) assignment of lines toheterotic groups; (iii) selection of appropriate parentallines for hybrid combinations; and (iv) generation ofsegregating progenies with maximum geneticvariability for further selection.

DNA-based markers are also being used to discoverand exploit the evolutionary relationships betweenvarious genera within a family (e.g., the grass family,Poaceae), and various species within a genus. Geneticmapping of members of the agriculturally-importantgrasses, including rice, wheat, maize, sorghum andsugarcane, with common DNA probes has revealed

87

remarkable conservation of gene content and geneorder (Devos & Gale, 2000), reinforcing the paradigmof the "grasses as a single genetic system" (Bennetzen& Freeling, 1993; Freeling, 2001). Comparativegenomics has significant implications for theapplication of genetic information generated in onemember of the grass family (such as rice or maize orsorghum) to the potential improvement ofagronomically important traits in other members.

This article provides a brief overview of (i) theapplication of molecular markers to characterizemaize germplasm and analyze genetic diversity; and(ii) the mapping and marker-assisted selection forspecific agronomically important traits in maize. Indiscussing the above, the focus will be on the recentstudies undertaken under the AMBIONETprogramme in India, and the work being carried out atCIMMYT's Applied Biotechnology Center inMexico.

Molecular Profiling of Maize GermplasmMaize breeders in India, as in most developing

countries, have differentiated inbred lines mainly onthe basis of major morphological characters such asplant height, anthocyanin colouration of various plantparts, tassel type, tassel branching, days to flowering,ear characters, cob colouration, grain colour and grai ntype (Virk & Witcombe, 1997). Althoughmorphological descriptions are important forascertaining the agronomic utility of germplasm, suchdescriptions are not very reliable because of complex'genotype x environment' interactions that requireassessment in multiple locations/environments (Smith& Smith, 1989). Detailed studies in various cropspecies, particularly in maize, have established thatmethods that solely depend on morphological data areneither consistent nor effective in unambiguousdifferentiation of elite breeding materials (Smith &Smith, 1988, 1989; Bar-Hen et at, 1995). Geneticheterogeneity, different combinations of allelesproducing similar phenotypes, and environmentalinfluence on genotypes, result in morphologicalsimilarities or differences that may not beproportional to the underlying genetic differences.

In the past, isozyme and zein chromatographic data(Stuber & Goodman, 1983; Smith, 1988) have beenused to characterize elite inbred lines and commercialhybrids of maize (Bar-Hen et at, 1995). Isozymeanalysis is relatively simple and less costly incomparison with molecular marker analysis; however,


M M M

~ - - (!- -• • !J

c

d

~ B

inadequate genomic coverage, relatively low levels ofpolymorphism, developmental regulation andpleiotropic effects impose major constraints ineffecti vely using these markers in genotypedifferentiation and analysis of genetic diversity(Smith & Smith, 1986; Dubreuil et all 1996). Inrecent years, PCR-based SSR markers have beeneffectively used for differentiation of US andEuropean maize germplasm, as they are particularlysuited for genotype discrimination (Smith et al, 1997;Warburton et at, 2002).

In India, no systematic efforts were made toeffectively apply molecular markers for geneticfingerprinting or analysis of genetic diversity in themaize inbred lines developed by public sectorinstitutions, including those that are commonly usedfor hybrid maize breeding. Recently, however, studieshave been carried out to profile a selected set ofIndian maize genotypes, including inbred lines andsingle-cross hybrids, using both morphological andmicrosatellite markers, and to analyze the geneticdiversity in the maize inbred lines that are commonlyused in the public sector institutions (Pushpavalli etai, 2001, 2002; Mohammadi et ai, 200.2a).Observations were recorded on 20 'categoricaldescriptors' (qualitative or visually assessedquantitative characters) in 47 Indian inbred lines toascertain their utility in effective differentiation ofgenotypes. The 'categorical' descriptors, however,revealed very low polymorphism in the Indian linesanalyzed, with a total of only 55 variants, highlightingthe severe limitations of utilizing only morphologicaldata for establishing the identity or distinctness ofgenotypes. In contrast, molecular profiling of 69inbred lines, comprising 58 Indian lines, 6 CIMMYTlines developed at Mexico (used as 'reference

genotypes'), and 5 lines from the CIMMYT-AsianRegional Maize Programme (CIMMYT-ARMP),Thailand, using 58 polymorphic SSR markersrevealed high levels of polymorphism (435 alleles)(Fig. 1A). Identification of 109 unique/rare SSRalleles (found in not more than 1 or 2 of the genotypesanalyzed) facilitated effective discrimination of thegenotypes analyzed. The high level of polymorphismdisplayed by the SSR loci was also reflected by theaverage PIC value (0.70). On the basis of high PICvalues (>0.75) and distinct allelic size ranges, SSRmarkers such as dupssr17, bntg1647, and bntg198,could be effectively used in differentiating the Indianmaize inbred lines. Distinct and non-overlapping sizeranges of the amplification products of SSR loci withhigh PIC would also facilitate multiplexing forimproving the assay efficiency (Fig. IB), as suggestedby Mitchell et at (1997).

The AMBIONET study also revealed high level ofSSR heterozygosity in some of the Indian maizeinbred lines. There could be various reasons includingresidual heterozygosity due to inadequate cycles ofinbreeding, improper pollination control during seedmultiplication, seed stock contamination oraccumulation of mutations at diverse SSR loci.Amplification of similar sequences in differentgenomic regions due to duplications is anotherpossible reason for occurrence of double-bandphenotypes. High levels of heterozygosity for some ofthe inbred lines such as CMl15, CMl17, CM 123,CM124, CM205 and CM208 were revealed earlierthrough isozyme analysis (Mauri a et at, 2000). SSRprofiling of these inbreds confirmed the aboveobservation.' However, some inbreds such as CM IIIand CM1l6, which were considered as 'geneticallypure' based on isozyme analysis (Mauri a et at, 2000),

Fig. l--{A) SSR polymorphism revealed by bnlg439 (A) in Indian maize inbred lines using PAGE and silver staining technology. ¢x-174/Hinfl digest (size range of 66-726 bp) was used as molecular weight standard (M). (B) Polymorphism in selected Indian maizeinbreds revealed by multiplexed, fluorescent-labeled SSR primers, ph084 (a), ncJ 30 (b), phi308707 (c), and phi089 (d); the methodology.using semi-automated DNA sequencers, facilitates accurate sizing of SSR alleles, besides enhancing the assay efficiency.


showed high levels of heterozygosity in SSR analysis.Such incongruities in data derived from biochemicalversus DNA-based markers could result due to thelow level of isozyme polymorphisms and limitedgenomic coverage.

A sequential method combining markerinformation and agro-morphological description,proposed by Smith et al (1991) suggests: (i)comparison of two lines at marker loci and declaringthem distinct only if their genetic similarity (GS)value is below a predetermined threshold; (ii)comparison of the two lines for agro-morphologicaltraits only if their GS value is beyond this threshold.In the second step, the environmental variation for themorphological traits allows construction of statisticaltests to determine the 'minimum distance' between thetwo inbreds, which assumes considerable significancein the context of plant variety protection. It would beinteresting to ascertain the broader applicability andeffectiveness of this proposal for routine profiling ofelite breeding materials.

With the availability of high throughputtechnologies that can make use of fluorescent-labeledSSR markers through multiplexing, fingerprinting hasbeen extended to classification of genetically diversematerials such as landraces, populations, open-pollinated varieties, and germplasm accessions.Earlier studies on characterization of populations haverelied on only a few individuals per population, as thecost and time required to characterize each line tendsto be the limiting factor. The efficiency and accuracyof population fingerprinting can be enhanced by usinga bulking strategy for individuals of a specificpopulation, followed by analysis of the bulks usingmultiplexed SSR primers and semi-automated DNAsequencing technology. A set of 7 tropical maizepopulations and 57 inbred lines at CIMMYT wererecently fingerprinted using 85 multiplexed SSRprimers, leading to identification of 53 highlydiscriminatory SSR markers (Warburton et al, 2002).

Analysis of Genetic Diversity in Indian MaizeGermplasm using Molecular Markers

Pedigree information provides a broad estimate ofthe expected genetic relatedness among lines, but forallogamous crops such as maize, such information isoften unobtainable or unreliable especially wheninbred lines were derived from a broad basepopulation (Melchinger et al, 1991; Messmer et al,1993). DNA-based markers, particularly SSRs and

89

AFLPs, have provided powerful tools for analyzinggenetic diversity (Pejic et al, 1998; Vuylsteke et al,2000). AMBIONET studies on molecular profiling ofIndian maize germplasm provided valuableinformation about genetic relationships in thebreeding materials (Pushpavalli et al, 2001, 2002;Mohammadi et al, 2002a,b). Analysis of SSR allelefrequencies revealed a reasonably broad genetic basein the Indian maize germplasm. However, much ofthe SSR allelic variation in the inbred lines analyzedwas contributed by the inbred lines developed atPunjab Agricultural University (PAU), Ludhiana.Cluster analysis of the genetic dissimilarity matrix forthe genotypes under study using Ward's method,besides application of 'pattern-finding methods' suchas principal coordinate analysis, aided in determininggenetic relationships, which were broadly inagreement with the available pedigree data.

The cluster pattern based on the AMBIONET studyrevealed essentially three main groups having twosub-clusters each. The majority of the Indian inbredlines that were derived earlier from the Colombiangermplasm (CMll1, CM114, CM120, CM300) wereclustered in Group J, while some Colombian lines(CM104, CMI05 and CM115) were placed in GroupII. Almost all of the early-maturing inbred linesdeveloped at JARI, New Delhi, such as CM135,CM136, CM137 and CM138 clustered in Group II.Some of the inbreds developed at PAU, Ludhiana,and analyzed in this study (CM122, CM123, CM124,CM125, CM140) clustered in Group II. The validityof the clusters was reflected when the patterns wereanalyzed in relation to the well-known pedigreeinformation for some inbred lines, particularly thosedeveloped by PAU, Ludhiana (Dhillon et al, 1998;Saxena et al, 2000). For instance, based on pedigreedata, CM122 must be highly related to CM140, andboth these genotypes should also show close geneticrelationship with CM124 and CM125. These expectedgenetic relationships were clearly validated by theresults of various clustering procedures that formedthe basis for the consensus cluster pattern. Group IIalso included CM202 and its close derivatives,CM211 and CM208. These inbred lines also showclose genetic association primarily with CM 122 andCM140. This could be due to the fact that CM202 wasextensively utilized in the derivation of Makki SafedPool C4, which also comprises CM122 and CM140.

The study indicated the genetic distinctness ofsome of the Ludhiana inbred lines (LM5, LM6 and


CM 139), as they were placed in a distinct cluster(Group III). LMS and LM6, parental lines of thesingle-cross hybrid 'Paras', were derived fromTuxpeno Pool and Makki Safed Pool C2, respectively.Suwan-l, developed by Kasetsart University,Thailand, had also contributed to the development ofTuxpeno Pool (Dhillon et al, 1998; Saxena et al,2000). AMB 109, a line developed from Thailandgermplasm (AMATL line) has shown closerelationship with LMS, while other AMB lines (whichmainly includes lines from Thailand and Philippines)displayed closer relationship with LM6 and CM139.The semi-exotic Pools A and B have closecorrespondence with Makki Safed and TuxpenoPools, respectively. This was also reflected in theplacement of CM139 between LMS (from TuxpenoPool) and LM6 (from Makki Safed Pool). Theaccuracy of the cluster pattern derived from this studywas reflected by the grouping of some of the BIOlines, which served as 'controls'. For instance, BIOS(an advanced line from LMS) grouped closely withLMS; similarly, BI07 (an advanced line from LM6)with LM6, and BI02 (an advanced line from CM211)with CM211.

The AMBIONET study in India also indicatedclose genetic associations among some downymildew-resistant lines such as AMBI12, AMBI19,AMBllS and AMBI09 (obtained from theCIMMYT-ARMP, Thailand) and NAII16, anunreleased Indian inbred that is highly resistant tosorghum downy mildew (SDM; Peronosclerosporasorghi) and Rajasthan downy mildew (P.heteropogoni) (Nair et al, 2001). This can beattributed to the utilization of SDM-resistantgermplasm from Thailand in the development ofNAI1l6 and some AMB lines (e.g., AMB1l2). Theclose clustering of the CIMMYT-ARMP lines andtheir genetic distinctness from a majority of the Indianmaize lines highlights the possibility for furtherexpansion of the genetic base of Indian maizegermplasm through efficient use of these genotypes.In contrast to the CIMMYT-ARMP lines, the sixinbred lines from CIMMYT, Mexico, showeddispersion in various clusters. The AMBIONET studyserves as an effective foundation for further analysisof genetic relationships of the inbred lines beingdeveloped by the National Agricultural ResearchSystem (NARS).

In another recent study at IARI, a set of 23 QPMlines, including 13 inbreds developed by the national

programme as well as 10 selected tropical/sub-tropical QPM lines developed by CIMMYT wereanalyzed for their grain uality, agronomicperformance and molecular polymorphism using SSRmarkers. Polymorphic profiles for 36 SSR loci haveaided in effectively differentiating the QPM inbredlines. The study resulted in identification of SSRmarkers, such as bnlg439, phi037, bnlg125, dupssr34and bnlgI05, with high polymorphism informationcontent in the selected QPM genotypes. Analysisusing SSR markers indicated high levels ofheterozygosity in majority of the Indian QPM linesand in one CIMMYT QPM inbred, CMLl88. Clusteranalysis using SSR data, followed by canonicaldiscriminant analysis, clearly distinguished the IndianQPM inbreds from those developed at CIMMYT(Kassahun & Prasanna, 2002). The cluster patternswere largely in congruence with the availablepedigree information of the QPM inbreds studied. Thestudy demonstrated the utility of SSR markers inanalysis of genetic relationships among QPM lines,and shall aid in planned utilization of CIMMYT QPMlines in QPM breeding programmes being undertakenin India.

One of the potential applications of molecularmarker data of inbred lines is to identify parentsuseful for developing or improving single-crosshybrid performance. Although it is unlikely that themarkers such as SSRs affect the phenotypicexpression of the targeted quantitative trait(s) directly,they can serve to identify adjacent (linked) genomicsegments. In such a case, marker divergence of inbredlines can be useful to predict hybrid performance.This is of particular value in crops like maize wheresignificant effort and resources are devoted to field-testing of newly created lines in various single-crosscombinations to identify lines with superiorcombining ability. A number of studies have beencarried out in maize to ascertain the associationbetween molecular marker divergence and hybridperformance, leading to different results (Stuber et 01.1999). Majority of the studies, however, indicate thatgenotypic differences may be useful for preliminaryselection of loci/alleles for possible improvement ofhybrids (Mohammadi et al, 2002b), but probably willnot accurately reflect performance of a hybrid. Fieldevaluation of nearly 92 hybrid combinations derivedfrom 48 Indian maize inbred lines in threeseasons/environments, recently carried out by theAMBIONET-India team, in conjunction with the SSR


allele data for these inbred lines, indicated thatmolecular marker divergence is not significantlycorrelated with the hybrid performance, reinforcingthe conclusion made above. A consensus opinion isnow emerging on this key issue that: (i) the genomeshould be well-saturated with uniformly spacedmarkers and/or high level of linkage equilibrium mustexist for marker data to be reasonably successful inpredicting hybrid performance; and (ii) marker datacan be more useful for predicting hybrid performanceof lines that are related and from a narrow geneticbase than those derived from highly divergent geneticbackgrounds.

QTL Mapping in MaizeThe discovery of extensive, yet easily visualized,

variability at the DNA level, coupled with thedevelopment of statistical packages that can help inanalyzing variation in a quantitative trait incongruence with molecular marker data in asegregating population, led to mapping of QTLinfluencing an array of agronomically important traitsin diverse crop plants including maize. A QTL maybe defined as a region of the genome that isassociated with an effect on a quantitative trait.Conceptually, a QTL can be a single gene, or a clusterof tightly linked genes that affect the trait. Excellentreviews dealing with various aspects of QTL mappingin crop plants are available (Beavis, 1998; Liu, 2002;Hackett, 2002). QTL mapping and identification ofmolecular markers closely linked to QTL with majoreffects on a target trait can permit MAS in backcross,pedigree, and population improvement programmes(Young, 1999; Ribaut et al, 2002a,b). This isespecially useful for crop traits that are otherwisedifficult or impossible to select for by conventionalmeans.

Molecular markers have been used to identify andcharacterize QTL associated with diverse traits inmaize including grain yield, characters concernedwith domestication, environmental adaptation, diseaseand insect pest resistance, and drought and heat stresstolerance (Stuber, 1995; Stuber et al, 1999).Comprehensive information about such experimentscan be obtained from the MaizeDB(htrp.z/agron.missouri.edu). A case study withpotential utility in effective management of downymildew diseases in maize in tropical Asian countriesis discussed below.

91

Mapping QTL Influencing Resistance to DownyMildews in Asia-A Case Study

Downy mildews occur predominantly intropical/sub-tropical regions of China, India,Indonesia, Japan, Nepal, Pakistan, Philippines andThailand in Asia, where the disease is an importantfactor limiting maize production (Pingali & Pandey,2001). The major downy mildews that infect maize inthe region include the sorghum downy mildew[Peronosclerospora sorghi (Weston & Uppal)],Philippine downy mildew (P. philippinensis [Weston]Shaw), Java downy mildew [Po maydis (Raciborski)],sugarcane downy mildew [Po sacchari (Miyabe)Shirai & Hara] , and brown stripe downy mildew[Scleropthora rayssiae var. zeae Payak & Renfro].While P. sorghi causes downy mildew in bothsorghum and maize, the maize strain of P. sorghi inThailand, which is now reclassified as P. zeae rarelyinfects sorghum. Different types of downy mildew arereported in India, including sorghum downy mildew,brown stripe downy mildew and sugarcane downymildew. In Rajasthan (India), the downy mildewpathogen which forms oospores in the wild grass,Heteropogon contortus (speargrass) was renamed P.heteropogoni (Siradhana et al, 1980) and the disease,caused when maize is infected by the conidial stage ofthe fungus, is now referred to as Rajasthan downymildew (White, 1999). Despite the introduction ofdowny mildew resistant cultivars and the use ofmetalaxyl fungicide, severe incidence of the downymildews still occurs in localized areas (Dalmacio,2000). Cost concerns related to seed treatment withfungicide, and the emerging problem of chemicalresistance build-up in the pathogen (Raymundo,2000), point to the use of resistant varieties as a morecost-effective and environmentally-safe alternative incontrolling this disease.

Identification of molecular markers linked todowny mildew resistance genes should have a majorimpact on maize breeding across the tropical Asianregion. As a Network activity under the AMBIONETprogramme, four countries (India, Indonesia, Thailandand Philippines) undertook a QTL mapping projectaimed at identifying downy mildew resistance genes(George et al, 2002). The mapping was based onevaluation of a set of recombinant inbred lines (RILs)derived from the cross of Ki3 (resistant) by CMLl39(susceptible). The downy mildew resistant parent,Ki3, is a tropical yellow flint line with late maturityfrom Suwan-l, a cultivar developed in Thailand


against P. zeae. The susceptible parent, CML139, is asubtropical yellow-red semi-flint line withintermediate maturity. developed from CIMMYTmaterials for tropical corn borer resistance. Groh et al(1998) constructed a molecular map using 135 RILsdeveloped from the Ki3 x CML139 cross and 143RFLP markers. for QTL mapping of southwesterncorn borer (SWCB) resistance. In the AMBIONETstudy, the same 135 RIL families were evaluated fordowny mildew reaction (during 2000-2001) atMandya in southern India against sorghum downymildew (P. sorghii; at Udaipur in western Indiaagainst Rajasthan downy mildew (P. heteropogonii;at Maros in Indonesia against Java downy mildew (P.nwydis); at Farm Suwan in Thailand against sorghumdowny mildew (P. zeae); and at Southern Mindanaoin Philippines against Philippine downy mildew (P.philippinensisi. The phenotypic data, thus, compriseddowny mildew disease incidence data from individualenvironments as well as pooled data acrossenvironments. Composite interval mapping (Zeng,1994) was carried out for joint analysis of data acrossenvironments to map QTL and to estimate theirgenetic effects.

The AMBIONET study led to the identification ofQTL with significant effects on resistance to the fiveimportant downy mildew diseases affecting maizeproduction in the Asian region. The QTL that weredetected highlighted differences in the pathogenpopulations that characterize the four locations.Three QTL, two on chromosome 2 and one onchromosome 7, significantly influenced resistanceonly to particular pathogen populations. The firstQTL on chromosome 2 (position 158 cM), withresistance due to alleles from the susceptible parentCMLl39, was specific to sorghum downy mildew atMandya. The second QTL on chromosome 2 (position234 cM) and the QTL on chromosome 7, withresistance due to alleles from the resistant parent Ki3,were found to inf1uence specifically P. heteropogoniat Udaipur in India. The most important genomicregion. having the highest LR values in the analysis ofdata from individual locations as well as in the jointanalysis, and having a consistent expression againstthe different downy mildews, was found onchromosome 6. This QTL was consistently expressedacross environments despite the significant effect ofthe environments having distinct pathogenpopulations.

The proportion of the phenotypic varianceexplained by each of the five QTL (R" values) variedacross environments. Collectively; the five QTL.dentified in this study explained phenotypic variationin disease susceptibility ranging from 24% (Thailand)to 54% (Udaipur, India). Most significantly, the QTLon chromosome 6 had the largest contribution,accounting for nearly 20% and 31% of the phenotypicvariance for P. sorglzi and P. heteropogoni diseasesusceptibility at Mandya and Udaipur, respectively,and explaining more than half of the total phenotypicvariance due to the five QTL in each of the fourenvironments. Significant QTL x E interactions andlarge estimates of cr2 ge observed across the locationsindicated major inf1uence of the environment.particularly the characteristic pathogen populations onthe expression of downy mildew resistance.

Significantly, the major QTL on chromosome 6(bin 6.05) is located in a region holding clusters ofresistance genes in maize. Groh et al (1998) identifieda QTL in Ki3 at an adjacent region 011 chromosome 6conferring resistance to leaf feeding damage causedby south western corn borer. Other genes that havebeen located on chromosome 6 in bin 6.0 I includemdml which confers resistance to the maize dwarfmosaic virus (MOM V) (Simcox et al, 1995): II'sJIlI

which confers resistance to a related potyvirus, wheatstreak mosaic virus (WSMV) (McMullen & Louie.1991); rhml , which confers resistance to the fungalpathogen Cochliolobus heterostrophus (Zaitlin et 01.1993); and a QTL conferring resistance to sugarcanemosaic virus (SCMV) (Zhang & Li, personalcommunication).

Selection for QTL using genetic markers can beeffective if a significant association is found betweenthe quantitative trait and the genetic markers. TheAMBIONET study identified three SSR markersumc l l , umc23a, and umcll3 tightly linked to theQTL on chromosome 6 (George et al, 2002).indicating their possible use for MAS. Beyond theirpossible use in MAS, another potential application ofthese results would be the identification and analysisof candidate genes to deduce information about thenature and function of the detected gene(s) Indetermining resistance to downy mildews in Asia.

Chances of successful application of MAS fordowny mildew resistance are better when QTL areidentified in the germplasm used in the nationalbreeding programme. For this purpose, theAMBIONET-India team also screened nearly RO

PRASANNA & HOISINGTON: MOLECULAR BREEDING TNMAIZE 93

--------------------------~~~1

A

NI PI P2 ..•..••1------------ BCIFI -----.--------~~ ~II

B

Fig. 2-Genotyping of a panel of BCIFI mapping population (for QTL mapping of downy mildew resistance in maize), for twopolymorphic SSR loci, bnlg490 (A) and bnlg1655 (B); the mapping population was derived using CM139 (PI) and NAIll6 (P2) used asrecurrent (susceptible) and donor (resistant) parents, respectively. A lOO-bp ladder was used as the molecular size standard (M) for thegels run on a super-fine resolution 3.5% agarose.

inbred lines, including 50 Indian genotypes, against P.sorghi and P. heteropogoni at Mandya and Udaipur,respectively. The study led to the identification ofNAIl16, an excellent source of resistance againstboth the downy mildew diseases in India (Nair et at,2001). A backcross mapping population wasdeveloped using CM139 (elite susceptible inbred) andNAIl 16. Analysis of the genotypic and phenotypicdata from this mapping population (Fig. 2 A & B)confirmed the effect of the major QTL detected onchromosome 6 (bin location 6.05) by the earlierAMBIONET study of the RILs. Introgression of thismajor QTLgoverning resistance to downy mildewsinto CM139 using marker-assisted backcrossing isnow ready to be undertaken.

Marker-assisted Selection (MAS) in Maize BreedingMAS consists of identifying associations between

markers and alleles of the gene/QTL of interest, andthen using these associations to develop improvedlines or populations (Ribaut & Hoisington, 1998;Knapp, 1998). Through marker-assisted backcrossing,individuals can be backcrossed until they contain theparticular genomic segment in the genetic background

of the recipient or recurrent parent. For MAS to beeffective, recombination between the marker and thegene/QTL must be minimal. This is achieved usingclosely linked flanking markers.

The basic purpose of marker-assisted backcrossingis to speed up line conversion, and to reduce thelinkage drag of the transferred gene(s). Throughclassical backcross (BC) breeding, the transfer of asingle dominant gene would require six BCgenerations to recover 99% of the recurrent parentgenome. This procedure is time-consuming andlabour-intensive for breeding of crops such as maize,where turnover times of new lines and hybrids arefast. In a BCI generation the proportion of therecurrent parent genome would be distri butednormally around a mean of 75% (in later BCgenerations, the distribution would becomeincreasingly skewed) but given a sufficient samplesize, it would contain plants with more than 85%recurrent parent genome. These plants can beidentified with molecular markers to accelerate thebreeding process (Tanksley et at, 1989). Withoutmolecular markers flanking the target gene it is nearly


impossible to remove the linkage drag coming as"baggage" with the introgressed segment (Murray etal, 1988).

MAS is now being routinely applied in thebreeding programmes of several crops, includingmaize, for (1) tracing favourable alleles in thegenomic background of genotypes of interest; and (2)identifying individual plants in large segregatingpopulations that carry the favourable alleles. Forinstance, two of the prominent examples of utilizationof molecular markers in line conversions through aBC approach being practiced in maize at CIMMYTare: (1) introgression of the opaque2 (02) gene onchromosome 7 for the development of QPM lines, and(ii) transfer of a major QTL identified on the shortarm of chromosome 1 that is associated with maizestreak virus (MSV) resistance. The utility of MAS inQPM breeding is particularly worth discussing asQPM has considerable relevance to various maize-growing countries, particularly in the developingworld, including India.

The maize grain accounts for about 15% to 56% ofthe total daily calories in diets of people in about 25developing countries, particularly in Africa and LatinAmerica (FAO Agrostat, 1992), where animal proteinis scarce and expensive and consequently, unavailableto a vast sector of the population. Maize seed-proteinquality can be improved by selecting for thehomozygous recessive 02 allele state (Mertz et al,1964). The presence of the homozygous 02 allele stateis correlated with changes in the amino acid balancewithin the endosperm, and more specifically, afavourable increase in the proportion of lysine andtryptophan. Cloning and sequencing of the 02 gene(Schmidt et al, 1990) allowed detection of three SSRmarkers (phi057, phil12 and umcl066) within thesequence of the gene itself. CIMMYT has beenroutinely screening thousands of genotypes, usingthese three SSRs, in segregating populations toidentify genotypes that have one copy of the 02mutant allele (BC strategy) and those that have twocopies (self-pollination strategy). Selection isconducted before flowering to allow the pollination ofonly the selected plants. Integration of MAS for 02 isa relatively simple and effective strategy foraccelerating QPM development, and this strategy iscurrently being employed in various countriesincluding India (Prasanna et al, 2001).

MAS can also be of great relevance toimprovement of polygenic traits. By combining the

QTL approach (selection for favourable QTL effects)with backcrossing, useful genes that controlquantitative traits can be identified and transferred toadvanced breeding lines (Lande & Thompson, 1990;Tanksley & Nelson, 1996; Hospital & Charcosset,1997). In maize, Stuber et al (1992, 1999) mappedQTL associated with seven major traits (includinggrain yield), and were also able to generate improvedversions of inbred lines using obsolete inbreds asdonors. The efficiency of using molecular markers forimprovement of polygenic traits in maize breedingprogrammes was also demonstrated in a few otherstudies (Ribaut et al, 2002a). Despite these examples,MAS for polygenic traits in maize, as in most othercrop plants, is still in its infancy. Manipulatingquantitative traits is difficult due to the involvementof a large number of genes involved in traitexpression, often with varying effects, interactionsbetween the genes (epistasis), and QTL xenvironment interactions (Beavis & Keim, 1996).This implies that several regions/QTL must bemanipulated simultaneously to have a significantimpact, and that the effect of individual regions is noteasily identified.

CIMMYT researchers have devoted considerableefforts during the past three decades to improve pre-and post-flowering drought tolerance in maize.Although significant progress has been achieved forimproving drought tolerance in CIMMYT maizegermplasm through conventional breeding (Banzigeret al, 2000), the approach is slow and time-consuming. Use of molecular markers and QTLinformation based on carefully managed replicatedtests has the potential to alleviate the problemsassociated with inconsistent and unpredictable onsetof moisture stress or the confounding effect of otherstresses such as heat. The approach was primarilybased on breaking down the complex trait of droughttolerance into simpler components, and to manipulategenomic regions related to components like anthesis-silking interval (ASI) that are closely associated withdrought tolerance. To this end, CIMMYT conductedseveral experiments on QTL analysis and MAS fortransfer of drought tolerance to tropical maize, andobtained encouraging results. An integrated strategyof QTL mapping, MAS and functional genomics isnow being explored to further provide usefulinformation and tools to effectively complementconventional selection for drought tolerance in maize(Ribaut et al, 2002a).


Integrating MAS ill Maize BreedingDespite a wealth of published literature on QTL

mapping, particularly in recent years, a number ofconstraints have imposed severe limitations oneffective utilization of QTL information in plantbreeding through MAS. Salient among theseconstraints are: (i) identification of a limited numberof major QTLs controlling target traits; (ii)inadequacies/experimental deficiencies in QTLanalysis leading to overestimation/underestimation ofthe number and effects of QTLs; (iii) lack ofQTLlmarker associations applicable over differentsets of breeding material; (iv) strong QTL xenvironment interaction; and (v) difficulty inprecisely evaluating epistatic effects. Recently, novelstrategies have been proposed (Ribaut & Betran,1999; Ribaut et 01, 2002b), particularly using maize asa model system, to overcome some of these majorconstraints. The efficacy of such strategies inimproving the efficiency of gene introgression usingmolecular markers, and reducing the cost of MASexperiments, is being analyzed at CIMMYT.

The cost-effectiveness of using molecular markers(SSRs) in MAS experiments in maize was alsoestimated (Dreher et 01, 2000). The study revealedthat when only a few SSR markers are used and whenseveral hundred genotypes are screened, MAS is cost-effective. Using SSR markers to select for theopaque? gene during QPM development exemplifiesthe utility of MAS as an efficient substitute forphenotypic selection, considering the recessive natureof the gene, absence of obvious visual selection due tothe interaction of this gene with modifiers involved inkernel vitreousness or hardness (an essential characterin QPM), and greater cost per sample when theendosperm protein quality is analyzed throughchemical analysis. The cost-effectiveness of MASover phenotypic selection, particularly for complexpolygenic traits, is also likely to improve in the future,with the availability of more efficient and high-throughput techniques for detection of molecularpolymorphism.

Future ProspectsUnderstanding the complex web of interactions

between genes and environmental factors, andeffective application of this information forplant/animal improvement is a challenging endeavourfor biologists. To obtain relevant information, it isimperative to exploit the tools of both classical and

95

molecular genetics. The developments in the recentyears in relation to molecular marker technology andQTL analysis have allowed identification of genomicregions involved in an array of agronomicallyimportant traits in diverse crop species, particularlymaize. Such information is also providing clues tobetter understand genome organization as well asgenetic phenomena such as epistasis, pleiotropy andheterosis. However, the impact of marker-based QTLanalysis on varietal development has been less thanexpected, primarily due to two reasons: (i)experiments related to QTL discovery and varietaldevelopment have largely been independent, aspointed out by Tanksley & Nelson (1996); and (ii) fortraits such as grain yield, QTL expression is usuallydependent upon the genetic background, unlike traitssuch as disease or insect resistance which are usuallyless complex in comparison with grain yield (Stuberet 01, 1999). Development of high-throughput,reproducible molecular marker technologies, coupledwith advances in genomics research, are nowpromising to offer powerful tools to maize researchersfor more effective integration and utilization of MASfor diverse applications in breeding programmes.

Besides some highly encouraging developments inmolecular breeding, structural and functionalgenomics research in maize is progressing at a healthypace (Stuber et 01, 1999; Coe et 01, 2002). From atime when the maize genome was considered to betoo complex to consider large-scale physical mappingand sequencing, we have now reached a point whereseveral research teams in the developed world arestriving to generate contig maps for maize anddetermine the sequence and function of the severalthousands of ESTs that are already identified (Coe et01, 2002). The advances in maize functionalgenomics, including gene expression profiling andproteomics (Lee et 01, 2002), should allow us in thenear future to identify the key genes as well aspathways involved in expression of traits such asbiotic/abiotic stress tolerance (Cushman & Bohnert,2000; Seki et 01, 2001).

The NARS in India have demonstrated thecommitment and capacity to effectively apply modernbiotechnology, particularly molecular markers, forcrop improvement. In maize, we should focusprimarily on three areas: assessment of geneticdiversity, application of marker-assisted selectionusing previously identified QTL and their flankingmarkers, and development of multiple trait-targeted


mapping populations based on parents that areadapted to the Indian agricultural productionsystem(s). The data already developed under theAMBIONET programme in relation to molecularprofiling of Indian maize inbred lines, besides agro-morphological data already available from thebreeders, provide valuable information in clearlyunderstanding the genetic diversity in the inbredgenetic base, thereby permitting more efficientutilization of genetic resources, both indigenous andexotic, in breeding programmes. In addition,application of optimized molecular marker technologywill reduce the costs of genetic resource conservationand further improve their utilization. Towards thisgoal, core collections of maize in India, that includemaximum genetic diversity and best representexisting variation, must be developed.

Cost-effective application of molecular markertechnology to agriculturally important problems inIndia cannot be done in isolation. Researchers in Indiacan immensely gain by building effective linkageswith partners elsewhere to harness the synergy ofcollective efforts in molecular breeding, asexemplified by AMBIONET. Networking canfacilitate development of an integrated system forefficient application of molecular tools andtechniques, including QTL mapping and MAS, inmaize breeding programmes of the NARS. Thiswould, in turn, significantly aid in development ofimproved germplasm, including cuitivars, with greateryield potential and ability to overcome major bioticand abiotic constraints limiting maize productivity, inthe minimum possible time and with minimaloperational expenses. Collaborative research underAMBIONET is presently focused on the applicationof molecular marker technology to problems ofnational and regional importance, such as molecularcharacterization of locally important maize lines,mapping of QTLs for resistance to major diseases -downy mildews, SCMV and Banded leaf and sheathBlight - and tolerance to abiotic stresses (drought andlow nitrogen conditions), and integration of MAS inthe breeding programmes.

AcknowledgementResearch work related to molecular profiling,

genetic diversity analysis of Indian maize gerrnplasm,and genome mapping for downy mildew resistancewas facilitated by CIMMYT, Mexico, underAMBIONET, with financial support from AsianDevelopment Bank. We express our gratitude to other

AMBIONET teams (Indonesia, Thailand andPhilippines) for sharing unpublished results from thenetwork activity on QTL mapping for downy mildewresistance. Special thanks to AMBIONET resourcepersonnel from CIMMYT and India, particularlyDrs. M L C George (AMBIONET Coordinator),N N Singh, R S Rathore. T A S Setty, and the pastand present members of the AMBIONET-India Lab atIARI, New Delhi, for sharing of data and fordedicated support to AMBIONET-India researchactivities.

ReferencesBanziger M, Pixley K V, Vivek B & Zarnbezi B T, 2000.

Characterization of elite maize germplasm grown in easternand southern Africa: Results of the 1999 regional trialsconducted by CIMMYT and the Maize and WheatImprovement Research Network for SADC (MWIRNET).CIMMYT, Harare, Zimbabwe.

Bar-Hen A, Charcosset A, Bourgoin M & Cuiard J. 1995.Relationships between genetic markers and morphologicaltraits in a maize inbred lines collection. Euphytica. 84. 145-154.

Beavis W D, 1998. QTL analyses: power, precision, and accuracy.ill Molecular Dissection of Complex Traits, edited by A HPaterson. CRC Press, Boca RatonlNew York. Pp 145-162.

Beavis W D & Keim P, 1996. Identification of quantitative traitloci that are affected by environment. in Genotype-by-Environment Interaction, edited by M S Kang & H G Gauch.CRC Press, Boca Raton, Florida. Pp 123-149.

Bennetzen J L & Freeling M, 1993. Grasses as a single geneticsystem: genome composition, colinearity and compatibility.Trends Genet, 9, 259-261.

Coe E, Cone K, McMullen M, Chen S, Davis G. Gardiner J.Liscum E, Polacco M, Paterson A, Sanchez-Villeda H.Soderlund C & Wing R, 2002. Access to the maize genome:an integrated physical and genetic map. Plant Physiol, 128,9-12.

Cushman J C & Bohnert H, 2000. Genomic approaches to plantstress tolerance. Curr Opiu Plant Bioi, 3. I 17-124.

Dalmacio S, 2000. Importance of and growing concern for maizediseases in the Asian region. in Proc Seventh Asian RegionalMaize Workshop, February 23-27, 1998, edited by S KVasal, F Gonzalez Ceniceros & Fan Xing Ming. Los Banos.Philippines. Pp 267-276.

Devos K M & Gale M D, 2000. Genome relationships: the grassmodel in current research. Plant Cell, 12, 637-646.

Dhillon B S & Prasanna B M, 200 I. Maize. ill Breeding FieldCrops, edited by V L Chopra. Oxford & IBH, New Delhi. Pp147-189.

Dhillon B S, Malhi N S, Saxena V K & Grewal M S, 1998.Development and improvement of heterotic populations inmaize. Crop Improv, 25, 6-14.

Dreher K, Morris M L, Ribaut J.-M, Khairallah M, Pandey S &Srinivasan G, 2000. Is marker-assisted selection cost-effective compared to conventional plant breeding methods')The case of Quality Protein Maize. Third Annu Con] 1111

Consortium Agric Biotechnol Res, August 25-28, Ravello.Italy.


Dubreuil P, Dufour P, Krejci E, Causse M, De Vienne D, GallaisA & Charcosset A, 1996. Organization of RFLP diversityamong inbred lines of maize representing the mostsignificant heterotic groups. Crop Sci, 36, 790-799.

Edwards K J & Mogg R, 200l. Plant genotyping by analysis ofsingle nucleotide polymorphisms. ill Plant Genotyping: TheDNA Fingerprinting of Plants, edited by R J Henry. CABIPublishing, Wallingford, UK. Pp 1-13.

FAO Agrostat, 1992. Food Balance Sheets. FAO, Rome, Italy.Freeling M, 2001.· Grasses as a single genetic system.

Reassessment 2001. Plant Physiol, 125, 1191-1197.George M L, Prasanna B M, Rathore R S, Setty T A S, Singh N

N, Kasim F, Azrai M, Vasal S, Balla 0, Regalado E, VargasM, Khairallah M, Jeffers D & Hoisington D, 2002.Identi fication of QTL conferring resistance to downymildews of maize in Asia. Theor Appl Genet (Accepted).

Groh S, Gonzalez-de-Leon D, Khairallah M M, Jiang C,Bergvinson D, Bohn M, Hoisington D A & Melchinger A E,1998. QTL mapping in tropical maize III. Genomic regionsfor resistance to Diatraea spp. and associated traits in twoRIL populations. Crop Sci, 38, 1062-1072.

Hackett CA, 2002. Statistical methods for QTL mapping incereals. Plant Mol Bioi, 48, 585-599.

Hospital F & Charcosset A, 1997. Marker-assisted introgressionof quantitative trait loci. Genetics, 147, 1469-1485.

Karp A, Kresovich S, Bhat K V, Ayad W G & Hodgkin T, 1997.Molecular tools in plant genetic resources conservation: Aguide to the technologies. IPGRI Technical Bulletin No.2.International Plant Genetic Resources Institute: Rome, Italy.

Kassahun B & Prasanna B M, 2002. Simple sequence repeatpolymorphism in Quality Protein Maize (QPM) lines.Eupliytica (In Press).

Knapp S, 1998. Marker-assisted selection as a strategy forincreasing the probability of selecting superior genotypes.Crop Sci, 38, 1164- 1174.

Lande R & Thompson R, 1990. Efficiency of marker-assistedselection in the improvement of quantitative traits. Genetics,124, 743-756.

Lee J-M, Williams M E, Tingey S V & Rafalski J A, 2002. DNAarray profiling of gene expression changes during maizeembryo development. Func Integr Genotuics, 2, 13-27.

Liu B-H, 2002. Statistical Genomics: Linkage, Mapping and QTLAnalysis. 2nd Edition. CRC Press, Boca Raton, Florida.

Mauria S,. Singh N N, Mukherjee A K & Bhat K V, 2000.Isozyme characterization of Indian maize inbreds. Euphytica,112,253-259.

McMullen M D & Louie R, 199 I. Identification of a gene forresistance to wheat streak mosaic virus in maize.Phytopathology, 81, 624-627.

Melchinger A E, Messmer M M, Lee M, Woodman WL &Lamkey K R, 1991. Diversity and relationships among U.S.maize inbreds revealed by restriction fragment lengthpolymorphisms. Crop Sci, 31, 669-678.

Mertz E T, Bates L S & Nelson ° E, 1964. Mutant that changesprotein composition and increases lysine content of maizeendosperm. Science, 145,279-280.

Messmer M M, Melchinger A E, Herrmann R G & BoppenrnaierJ, 1993. Relationships among early European maize inbreds:II. Comparison of pedigree and RFLP data. Crop Sci, 33,944-950.

97

Mitchell S E, Kresovich S,Jester C A, Hernandez C J & Szewc-McFadden A K, 1997. Application of multiplex PCR andfluorescent-based, semi-automated allele sizing technologyfor genotyping plant genetic resources. Crop Sci, 37, 617-624.

Mogg R, Hanley S & Edwards K J, 1999. Generation of maizeallele-specific oligonucleotides from the flanking regions ofmicrosatellite markers. Plant and Animal Genome VIIConference Abstract Guide. Scherago International Inc, P491.

Moharnmadi S A & Prasanna B M, 2002. Analysis of geneticdiversity in crop plants - salient statistical tools andconsiderations. Crop Sci (Accepted).

Mohamrnadi S A, Sudan C, Garg A, Prasanna B M & Singh N N.2002a. Characterization of Indian maize inbred lines andanalysis of genetic diversity using microsatellite markers.Tlieor Appl Genet (Communicated).

Mohammadi S A, Prasanna B M, Sudan C & Singh N N, 2002b.A rnicrosatellite marker based study of chromosomal regionsand gene effects on yield and. yield components in maize.Cell Mol Bioi Left, 7, 599-606.

Mohan M, Nair S, Bhagwat A, Krishna T G, Yano M, Bhatia C R& Sasaki T, 1997. Genome mapping, molecular markers andmarker-assisted selection in crop plants. Mol Breed, 3, 87-103.

Murray M G, Ma Y, Romero-Severson J, West D P & Cramer JH, 1988. Restriction fragment length polyrnorphisms: whatare they and how can breeders use them? Proc 43"d AI/IIl/Com Sorghum Res COIlf, Chicago. USA. American SeedTrade Association, Washington, DC. Pp 72-87.

Nair S K, Setty T A, Rathore R S, Kumar R, Singh N N &Prasanna B M, 2001. Towards molecular marker mapping ofgenes conferring resistance to sorghum downy mi IdewiPeronosclerospora sorghi) in maize. Maize Genet CoopNewslett, 75, 47-48.

Pejic I, Ajrnone-Marsan P, Morgante M, Kozurnplick V,Castiglioni P, Tararnino G, & Motto M, 1998. Comparativeanalysis of genetic similarity among maize lines detected byRFLPs, RAPDs, SSRs and AFLPs. Theor Appl Genet, 97,1248-1255.

Pingali P L & Pandey S, 200 I. Meeting world maize needs:technological opportunities and priorities for the publicsector. ill CIMMYT 1999/2000 World Maize Facts andTrends, edited by P L Pingali. CIMMYT, Mexico, D F Pp 1-24.

Powell W, Morgante M, Andre C, Hanafey M, Vogel J, Tingey S,Rafalski A,1996a.The comparison of RELP, RAPD, ALFPand SSR(microsatellite) markers for gerrnplasm analysis.MolBreed 2,225-238.

Powell W, Machray G C, Provan J, 1996b.Polymorphism revealedby simple sequence repeats. Trends Plant Sci.l, 215-222.

Prasanna B M, Vasal S K, Kassahun B & Singh N N, 2001.Quality Protein Maize. Curr Sci, 81, 1308-1319.

Prioul J -L, Quarrie S, Causse M & de Vienne D, 1997. Dissectingcomplex physiological functions through the use of

. molecular quantitative genetics. J Exp Bot, 48, 1151-1163./Pu"shpavalli S N C V L, Sudan C, Moharnmadi S A, Nair S K.

Prasanna B M, Gadag R N & Singh N N, 2002. Analysis ofsimple sequence repeat (SSR) polymorphism in Indian maizeinbred lines. J Genet Breed, 56 (In Press).


Pushpavalli S N C V L, Sudan C, Singh N N & Prasanna B M,2001. Differentiation of elite Indian maize hybrids usingsimple sequence repeat markers. Indian J Genet, 61, 304-308.

Raymundo A 0, 2000. Downy mildew of maize in Asia: Newperspectives in resistance breeding. in Proc Seventh AsianRegional Maize Workshop, February 23-27, 1998, edited byS K Vasal, F Gonzalez Ceniceros & Fan Xing Ming. LosBanos, Philippines. Pp 277-284.

Ribaut J-M & Betran F J, 1999. Single large-scale marker-assistedselection (SLS-MAS). Mol Breed, 5,531-541.

Ribaut J-M & Hoisington 0, 1998. Marker-assisted selection:New tools and strategies. Trends Plant Sci, 3, 236-239.

Ribaut J-M, Banzinger M, Betran J, Jiang C, Edmeades G 0,Dreher K & Hoisington 0, 2002a. Use of molecular markersin plant breeding: Drought tolerance improvement in tropicalmaize. in Quantitative Genetics, Genomics, and PlantBreeding, edited by M S Kang. CABl Publishing,Wallingford, UK. Pp 85-99.

Ribaut J -M, Jiang C & Hoisington 0, 2002b. Simulationexperiments on efficiencies of gene introgression bybackcrossing. Crop Sci, 42, 557-565.

Saxena V K, Malhi N S, Singh N N, Vasal S K, 2000. Heterosis inmaize: Groupings and patterns. in Proc Seventh AsianRegional Maize Workshop, Feb 23·27, 1998, edited by S KVasal, F G Ceniceros & F Xing Ming. Los Banos,Philippines. Pp 124-133.

Schmidt R J, Burr F A, Aukerman M J & Burr B, 1990. Maizeregulatory gene opaque-2 encodes a protein with a "leucine-zipper" motif that binds to zein DNA. Proc Natl Acad SciUSA, 87,46-50.

Seki M, Narusaka M, Abe H, Kasuga M, Yamaguchi-ShinozakiK, Carninci P, Hayashizaki Y & Shinozaki K, 200!.Monitoring the full expression pattern of 1300 Arabidopsisgenes under drought and cold stresses by using a full-lengthcDNA microarray. Plant Cell, 13,61-72.

Simcox K 0, McMullen M 0 & Louie R, 1995. Cosegregation ofthe maize dwarf mosaic virus resistance gene, mmdml, withthe nucleolus organizer region in maize. Theor Appl Genet,90,341-346.

Siradhana B, Dange, S R S, Rathore R S & Singh S D, 1980. Anew downy mildew on Maize in Rajasthan, India. Curr Sci,49,316-317.

Smith J S C, 1988. The diversity of US hybrid maize germplasm:Isozymic and chromatographic evidence. Crop Sci, 28, 63-70.

Smith J S C, Chin E C L, Shu H, Smith 0 S, Wall S J, Senior ML, Mitchell S E, Kresovich S & Ziegle J, 1997. Anevaluation of the utility of SSR loci as molecular markers inmaize (Zea mays L.): Comparison with data from RFLPs andpedigree. Theor Appl Genet, 95, 163-173.

Smith J S C & Smith 0 S, 1986. Environmental effects on zeinchromatograms of maize inbred lines revealed by reversed-phase high-performance liquid chromatography. Theor ApplGenet, 71, 607-612.

Smith J S C & Smith 0 S, 1988. Associations among inbred linesof maize using electrophoretic, chromatographic, andpedigree data: II. Multivariate and cluster analysis of data

from Iowa Stiff Stalk Synthetic derived lines. Theor ApplGenet, 76, 39-44.

Smith J S C & Smith 0 S, 1989. The description and .assessmentof distance between inbred lines of maize l.The use ofmorphological traits as descriptors. Maydica, 34, 141-150.

Smith J S C, Smith 0 S, Boven S L, Tenborg R A & Wall S r,1991. The description and assessment of distances betweeninbred lines of maize. III: A revised scheme for the testing ofdistinctiveness between inbred lines utilizing DNA RFLPs.Maydica, 36, 213-226.

Stuber C W, 1995. Mapping and manipulating quantitative traitsin maize. Trends Genet, 11,477-481.

Stuber C W & Goodman M M, 1983. Allozyme genotypes forpopular and historically important inbred lines of corn (Zeamays L.). USDA Agr Res Results Ser NO.16.

Stuber C W, Lincoln S E, Wolff D W, Helentjaris T & Lander ES, 1992. Identification of genetic factors contributing toheterosis in a hybrid from two elite maize inbred lines usingmolecular markers. Genetics, 132, 823-829.

Stuber C W, Polacco M & Senior M L, 1999. Synergy ofempirical breeding, marker-assisted selection, and genomicsto increase yield potential. Crop Sci, 39, 1571-15.83.

Tanksley S 0 & Nelson J C, 1996. Advanced backcross QTLanalysis: a method for the simultaneous discovery andtransfer of valuable QTL from unadapted germplasm intoelite breeding lines. Theor Appl Genet, 92, 191-203.

Tanksley S 0, Young N D, Paterson A H & Bonierbale M W,1989. RFLP mapping in plant breeding: New tools for an oldscience. Bio/Technology, 7, 257-263.

Virk 0 S & Witcombe J R, 1997. An analysis of varietal testing.in New Seeds for Small Farmers: Challenges andOpportunities for Change in the Regulatory Framework.edited by J R Witcombe, 0 S Virk & J Farrington. Oxfordand IBH, New Delhi. Pp 1-3.

Vuylsteke M, Mank R, Brugmans B, Stam P & Kuiper M, 2000.Further characterization of AFLP data as a tool in geneticdiversity assessments among maize (Zea mays L.) inbredlines. Mol Breed, 6, 26'5-276.

Warburton, M L & Hoisington 0, 2001. Applications ofmolecular marker techniques to the use of internationalgcrmplasm collections. in Plant Genotyping: The ON AFingerprinting of Plants, edited by R 1 Henry. CABIPublishing, Wallingford, UK. Pp 83-93.

Warburton M L, Xia X, Crossa J, Franco F, Melchinger A E.Frisch M, Bohn M & Hoisington 0, 2002. Geneticcharacterization of CIMMYT inbred lines and openpollinated populations using large-scale fingerprintingmethods. Crop Sci. (Accepted).

White 0 G, 1999. Downy mildews. in Compendium of CornDiseases, Third edition, American PhytopathologicalSociety, USA. Pp 25-32.

Young N D, 1999. A cautiously optimistic vision for marker-assisted breeding. Mol Breed, S, 505-510.

Zaitlin D, DeMars S & Ma Y, 1993. Linkage of 1'11111, a recessivegene for resistance to southern corn leaf blight. to RFLPmarker loci in maize (Zea mays L.) seedlings. Gel/oil/e. 36.555-564.

Zeng Z B. 1994. Precision mapping of quantitative trait loci.Genetics, 136, 1457-1468.

molecular breeding for maize improvement: an...

Documents