36:735-742 (1996). certain environments (Bozzini, 1988). On the other hand, durum wheat are better adapted to Mediterranean dryland than bread wheat.
Published May, 1996
Genetic Diversity in Durum Wheat Based on RFLPs, Morphophysiological Traits, and Coefficient of Parentage Enrique Autrique,* Miloudi M. Nachit, Philippe Monneveux, Steven D. Tanksley, and Mark E. Sorrells ABSTRACT Estimation of genetic diversity present in gene pools is an important determination for breeding programs. This study was conducted to determine the level of variation in durum wheat cultivars (Triticum turgidum L. var durum) based on different measures. Genetic diversity in durum wheat was measured in a collection of 113 improved cultivars and landraces of diverse ecogeographical origin by means of restriction fragment length polymorphism (RFLP), morphophysiological traits, and coefficient of parentage (COP). Thirty-nine clones and a single restriction enzyme were used to measure the RFLP-based genetic distance. Average taxonomic distances were calculated for the morphophysiological traits evaluated in four location-years. Lower genetic distances were observed for both RFLP and morphophysiological traits for the improved cultivars and for some landraces from Morocco and Jordan, while genetic distances were larger for the rest of the landraces. Patterns of variation for morphophysiological traits were associated with days to heading, plant height, and harvest index. Landraces contained 99% of the total fragments observed in the pool of improved varieties and showed 13% unique fragments. Coefficient of parentage analysis revealed 15 ancestral lines that contributed 72% of the genetic make up of improved cultivars. Thirty-nine different ancestors contributed to the remaining 28%. Correlation of distances based on different measures was higher for average taxonomic distance and Nei's genetic distance (r = 0.47) while COP relationship with the other two measures was lower. Narrower genetic diversity in breeding lines based on the measures used, suggests the utilization of other sources of variation.
URUM WHEAT is one of the oldest cultivated plants in the world and is grown mainly in the Middle and Near East regions and North Africa, which are considered the centers of origin and diversification of this crop (Vavilov, 1951). It represents about 10% of the total world wheat production area. The adaptation of durum wheat largely overlaps that of bread wheat (Triticum aestivum L.) but is less widely grown because (i) it is not suited for breadmaking, (ii) it is less winterhardy, and (iii) generally, its adaptation is restricted to E. Autrique, Homero 1337, Col. Polanco, 11550 Mexico, D. F. Mexico; S.D. Tanksley and M.E. Sorrells, Dep. of Plant Breeding and Biometry, 252 Emerson Hall, Cornell Univ., Ithaca, NY 14853; M.M. Nachit, Cereal Program, The Int. Center for Agric. Res. in the Dry Areas (ICARDA), P.O. Box: 5466, Aleppo, Syria; P. Monneveux, ENSA-INRA, Genetique et Amelioration des plantes, 2 Place Viala, 34060 Montpellier CedexOl, France. Received 2 Jan. 1995. "Corresponding author (autrique @ mail. mindvox. ciateq. mx). Published in Crop Sci. 36:735-742 (1996).
certain environments (Bozzini, 1988). On the other hand, durum wheat are better adapted to Mediterranean dryland than bread wheat. Breeder's germplasm collections are mainly high yielding or disease resistant varieties and advanced lines that are being intercrossed in the search for new genes or gene combinations. The positive results from this approach have contributed to improved durum cultivars (Porceddu et al., 1988). Yield increases have been accompanied by increases in total biomass production, harvest index, and changes in yield components such as number of spikes per unit area and spike fertility (CIMMYT, 1988). The level of genetic variation present in gene pools of cereal crops has been analyzed by studying the pedigree relationship between cultivars released over a period of time. Coefficient of parentage estimation of cultivars of oats (Avena sativa L.) (Sou/a and Sorrells, 1989), soybean [Glycine max (L.) Merr.] (Cox et al., 1985a), winter wheat (Cox et al., 1985b), rice (Oryza sativa L.) (Dilday, 1990), and barley (Hordeum vulgare L.) (Martin et al., 1991) has shown that a restricted number of ancestral genotypes account for a large proportion of the variation present in released cultivars. In barley, the hybridization of different gene pools was largely restricted to two-row and six row barley cultivars (Martin et al., 1991). Crosses between winter and spring wheat gene pools are far more common and offer a new source of diversity (CIMMYT, 1987). Breeder's use of landraces and wild species is usually restricted to cases where no variation is found in the improved materials. Landraces are genetically diverse for many phenological, morphological, physiological, and quality traits as well as resistance to biotic stress (Porceddu et al., 1988). Landraces have been collected from the centers of diversity by several institutions and plant explorers (Lawrence, 1984) and are being used for improving adaptability of high yielding cultivars grown under dry environments. Mediterranean durum landraces have been found to possess desirable traits lacking in improved cultivars (Nachit, 1992). Durum wheat world Abbreviations: COP, coefficient of parentage; PIC, polymorphic information content; QTL, quantitative trait loci; RFLP, restriction fragment length polymorphism; DSZ, standard taxonomic distance based on morphophysiological traits using z transformations; NGD, Nei's genetic distance based on RFLP.
CROPSCIENCE, VOL. 36, MAY-JUNE1996
Table 1. Landraces and improved genotype names or crosses, Parentage (COP) or morphophysiological data.
origin and cluster
numbr based on Nei’s genetic distance,
Cluster Entry? no.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 ¯ 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 41 42 43 45 47 48 49 50 51 52 53 54 55 56 57 58 59 61 62 63 65 67 68 69 70 71 72 73 74 75 76 (continued)
Genotype name Hedba 3 Qued Zenati 368 MohamedBen Bachir Oued. Zenati¶ Zenati Bouteille~:¶ Guemgoun rekhem 1181 (ARI 76-30) 1293 (ARI 76-142)
T. Dur. Ethiop.IC 8373 ST, t64~ Romanan 2 Mavragani-Iraklion Moundrous-2 Atsiki-3 Local Iraklion Khapli~§¶ Tripolino Scorsonera Sicflia Lutri Cannizzara Senatore Cappeili¶ Jordan Coll. 86 NO21 (Jord86-21) Jordan Coll. 86 NO42 (Jord86-42) Jordan Coll. 86 NO44 (Jord86-44) Jordan Coll. 86 NO53 (Jord86-53) Jordan Coll. 86 NO80 (Jord86-80) Jordan Coil. 86 NO174 (Jord86-174) M 13 M 21 M3 M 20 M 10 M 1084 M1086 M 1090 M 11 M 1150 M 15 Tremez Molle~¶ Rasp. de Aguilas~:¶ Haurani Nawani¶ Haurani 27 Normal Haurani Haurani Hamari Ahmar BCH Akbash Kishk Baladia Hamra Gezira 17 Baladia Hamra Nabel Jamal Normal Hamari Shihani Jennah Rhetffah~:¶ Durum #2~¶ Mahmoudi981 Mindum-~:¶ VernalEmmer~:§¶ H.O-FAO25918 D-7/5/Bye/Tc60//ZB/WIs/3/Cp/St464 (DLRI0) Ruff//Jo/Cr/3/GdoVz578 Haucan Tafna Gd 75/3/Stk//Ch67/Jo Akrachel Lahn-SH Syrica2 Stork Scar/GdoVz579/3/GdoVz471/Br//Pg (DLRS) Frig/Ren//Ruff/Gta/3/Ren/4/USAO640Fg//Fg*2/Ruff (DLR6) OmtelI Omrabi5 (Mrb5)
Origin Algeria Algeria Algeria Algeria Algeria Algeria Cyprus Cyprus Egypt Ethiopia Ethiopia Greece Greece Greece Greece Greece India Italy Italy Italy Italy Italy Jordan Jordan Jordan Jordan Jordan Jordan Morroceo Morrocco Morrocco Morrocco Morrocco Morrocco Morrocco Morrocco Morrocco Morrocco Morrocco Portugal Spare Syria Syria
RFLP B A B B B B B A A C B B B B B
Syria Syria Syria
B C B C A A A C A C A A A A A A A A A A A C C C C C C C B
Syrm Syria Syria Tunisia Tunisia Tumsia USA USA Turkey
Syna Syna Syna
E E E A A A A A A A A A A A A A A A
Phenotypic data 4 5
5 5 5 1 1 1 1 1
AUTRIQUEET AL.: GENETIC DIVERSITY IN DURUMWHEAT Table I. cont’d. Cluster Entry~" no. 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
Genotype name Omrabi 17 (MrblT) Tensift 1 Guerou 1 Gedifla Chahba 88 Sabil 1 Loukus 1 Bicre Furat 1 Nile Zud 1 Khabur 1 Po Hazar Deraa OmRabi 14 (Mrb14) Kabir 1 Oronte Daki = Ceyhan Karasu Gr/Boy Sajur Ain Arous Sebou Quadalete Jordan Korifla Belikh 2 Cham 1 Siliana Awali Heider = Marjawi Brachoua Aric31708.70/3/Bo//C.de C/Br/4/Cit/Gta (IC78) North. Dakota 86-10 (N.D.86-10) Wakooma DT 369
USA USA Canada Canada
A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A B A A C C A
1 3 3 3 1 4 3 3 2 2 2 2 3 2 2 1 3 1 2 2 2 2 2 2 3 2 2 2 3 2 2 3 4 4 4 4
Phenotypic data 3 2 3 3 2 2 4 4 3 4 3 2 4 3 3 3 3 2 2 3 2 3 3 3 4 2 3
Entries 1 to 61 are Landraces,all the other entries are improvedlines from the CIMMYT/ICARDA breeding programor otherwise noted. Germplasmrequested of the National Small Grains Collection, Aberdeen, ID. T. turg/dum spp. dicoccumaccessions. Ancestral landraces based on pedigree informationof all improvedcultivars.
collections from the USDA and Italy were evaluated for morphologicaltraits (Jain et al., 1975; Porceddu, 1976), flag leaf characteristics (Spagnoletti Zeuli and Qualset, 1990), and physiological and morphological traits (Yang et al., 1991). Basedon a diversity index, greater variation was found for some of the countries that are believed to be the centers of genetic diversity. Total variation in material as detected by the Shanon-Weiner diversity index showed significant differences amongthe characters but not amongthe countries (Porceddu, 1976) and sometraits showeda regional pattern of variation (Jain et al., 1975). Diversity in Ethiopia, considered a secondary center of diversity of tetraploid wheat, was reported to be low by Yanget al. (1991) while another study of 1800 landraces collected in Ethiopia showedgreat diversity for manytraits (Srivastava et al., 1988). Morphological characters, limited in number, often do not reliably portray genetic relationships because of environmentalinteractions, epistatic interactions and the largely unknowngenetic control of the traits (Smith and Smith, 1989). Genetic markers such as restriction fragment length polymorphism(RFLP) represent genetic variation at the DNAlevel, allowing an estimation of the degree of relatedness between individuals without the influence of environmentalvariation (Miller and Tan-
ksley, 1990; Beer et al, 1993; Dudley, 1994). Quantification of genetic variation between individuals could enhance the level of variation in breeding populations. The objective of this research was to study the level of diversity present in a collection of durumwheatcultivars and landraces adapted and collected in different ecogeographical areas. Genetic diversity was measured by RFLPsand compared with diversity based on morphophysiological traits and coefficient of parentage. MATERIALS AND METHODS Germplasm Atotal of 111durumlandraces,advancedlines andcultivars, and two T. turgidumvat. dicoccumaccessionsof diverse origin were used for this study (Table 1). Fifty-one landraces were selected from the collection of the International Center for Agricultural Research in the Dry Areas (ICARDA) germplasm bankand eight additional landraces wereprovidedby the U.S. National SmallGrains Collection (Aberdeen,ID). Overall, the landraces representedcollections from11 different countries from the MiddleEast, North Africa, Europe, and the North America. Forty-eight advancedlines developed at CIMMYT/ ICARDA joint breeding program included long term check cultivars, newlyreleasedcultivars and selected advancedlines. Thiscollection includedmaterialadaptedto the mandated target
CROP SCIENCE, VOL.36, MAY-JUNE 1996
environments of the CIMMYT durum program at ICARDA. Four additional improvedgenotypesrepresent cultivars, advanced lines, or both from Canadaand USA. DNAIsolation
were grouped based on COPto ancestral cultivars by the UnweightedPair-Group MeanAverage (UPGMA) clustering method and the SAHN subroutine from NTSYS-pc(Rohlf, 1990).
and Clone Selection
Ten grams of fresh leaf tissue bulked from eight to 10 17-d-old plants were frozen in liquid nitrogen, ground, and used for DNAextraction. Extraction buffer and procedure used were based on Tai and Tanksley (1990); one to one volumeof chloroform/isoamyl (24:1) was used instead potassiumacetate and the secondDNA precipitation was omitted. DNAwas cut with EcoRI(AdvancedAmericanBiotechnology, Fullerton CA)with the appropriate buffer and conditions specified by the manufacturer. Approximately20 to 25 ttg of digested DNAwere loaded into a 9 g L-~ agarose gel, electrophoresed and transferred to a HybondN + membrane (Amersham International PLC,Little Chalfont, Bucks, UK.). Filters were prehybridized with a hybridization buffer, according to Churchand Gilbert (1984). Probes were labeled with [32p]dCTPby random priming method (Feinberg and Vogelstein,1983). Filters were washedas described in Anderson et al. (1992) and filters were exposedover X-rayfilm for 5to7d. Thirty-nine probes from oat and barley cDNA libraries and a wheat genomiclibrary, described in Heunet al. (1991), were used. Probes were selected based on their chromosomal location (Andersonet al., 1992). Theseclones showedhybridization to fragmentsof the A and B genomes,based on Chinese Springaneuploids,and weredistributed across the sevenchromosome groups. Someof the clones used were selected based on their polymorphicinformation content (Andersonet al., 1993). Morphological, Phenological, and Physiological Data A subset of the germplasmscreenedfor RFLPwas evaluated for 16 morphological,phenologicaland physiological traits. Thesetraits included days to heading, days to flower, days to maturity, plant height, total numberof tillers, numberof spikes per squaremeter, length of awns,last internodelength, pedunclelength abovethe flag leave, spike length, grains per spike, total biomass, yield, harvest index, thousandkernel weight,and droughtsensitivity index. Thesubset of accessions included 40 landraces and 36 advancedlines. Data for this study was collected in a replicated trial under twodifferent growingconditions, irrigated and non-irrigated, for 2 yr at Montpellier (INRA),France. The results from this study and details of the experimentwere published elsewhere(Ali-Dib, 1992). Coefficient of Parentage The coefficient of parentage(COP)betweentwoindividuals is definedas the probabilitythat a randomallele at a locus in one individual is identical by descent to a randomallele at the samelocus in the other individual. COPvalues (r) were estimated as described by Coxet al. (1985a). Ancestral lines from 51 improvedlines, cultivars, or both were traced back with several sources of information about genealogies and pedigrees (Brajcich et al., 1986; Zeven and Zeven, 1976; Zevenand Reiner, 1991). Additional informationwas provided by Dr. Elias Elias (North DakotaState University), Dr. Victor Vallega (Istituto Sperimentaleper la Cerealicoltora, Rome), and by Dr. OsmanAbdalla (CIMMYT,Mexico). The COP values were calculated with a Fortran programdevelopedat KansasState University (Cox and Murphy,1990). Cultivars
Data Analysis Autoradiographs were scored based on the presence or absence of bands, generatinga matrix of 1 and 0. Informative bandswere used to generate a genetic distance matrix with the SIMGEND routine based on Nei’s formula from the NTSYS-pc statistical package(Rohlf, 1990). SubroutineSAHN was used to cluster the genotypesbasedon the genetic distance matrix with the UPGMA clustering method. Polymorphicinformation content (PIC) wasused to measure the relative value of each clone with respect to the amountof polymorphism it exhibits. The gene diversity measureof each clone was calculated based on (Weir, 1990, p. 124-130): n
PIC = 1 - ZP~ j=l whereP,~ is the frequencyof the jth RFLPpattern, for clone i and n indicates the total numberof patterns revealed by the clone. Meansacross locationsand years for eachone of the quantitative variables measuredwere used to computea similarity distance matrix. The data was transformed with the STAND procedure from NTSYS-pc (Rohlf, 1990). The standardization procedurereducedthe effect of different scales of measurement of the different characters. In this transformation, the mean is subtractedfromthe original value and dividedby the standard deviation. The standardized values were used in the SIMINT subroutine of NTSYS-pc (Rohlf, 1990) to computea matrix of similarities amongall pair of genotypeswith the average taxonomicdistance. Standard taxonomic distance was used because previous results indicated a higher correlation with measuresof similarity based on RFLPand isozymedata (Beer et al., 1993). Clustering of genotypes, by UPGMA method, based on the similarity matrix or based on the first eight principal componentaxes (PCA)that accounted for 91% the total variation producedsimilar grouping of genotypes except for one improvedline. Matrix Comparisons Nei’s genetic distance(NGD) matrix, coefficient of parentage matrix and the matrix based on morphophysiological data (DSZ) were compared by the MXCOMP routine of NTSYS-pc (Rohlf, 1990)that uses the normalizedMantelZ statistics. The statistical considerationsfor these analyseswerediscussedby Beeret al. (1993). RESULTS Restriction Fragment Length Polymorphism Analysis Thirty-nine clones detected a total of 232 fragments resulting in a mean of 5.7 fragments per clone and 4.3 polymorphic bands. The average number of unique genotypes revealed by each clone was eight. Thirty-six probes were polymorphicon at least one of the accessions surveyed, and distinguished each one of the genotypes surveyed. From the total number of fragments detected, 165 were polymorphic and were used to estimate the genetic relationships betweenall possible pairs of genotypes. The polymorphic information content (PIC) ranged from 0 to 0.96. PIC values were similar to those of
AUTRIQUEET AL.: GENETIC DIVERSITY IN DURUMWHEAT
29 %. From this commongroup of polymorphic markers, 16%of the alleles were fixed in the population of improved varieties and a large proportion of them were either in high (>0.85) or low (