Theor Appl Genet (2008) 117:1361–1377 DOI 10.1007/s00122-008-0869-1 ORIGINAL PAPER QTL analysis of pasta quality using a composite microsatellite and SNP map of durum wheat W. Zhang · S. Chao · F. Manthey · O. Chicaiza · J. C. Brevis · V. Echenique · J. Dubcovsky Received: 27 May 2008 / Accepted: 15 August 2008 / Published online: 9 September 2008 © Springer-Verlag 2008 Abstract Bright yellow color, Wrmness and low cooking environment and their interactions were analyzed using loss are important factors for the production of good-qual- factorial ANOVAs for each trait. We identiWed major QTLs ity pasta products. However, the genetic factors underlying for pasta color on chromosomes 1B, 4B, 6A, 7A and 7B. those traits are still poorly understood. To Wll this gap we The 4B QTL was linked to a polymorphic deletion in the developed a population of 93 recombinant inbred lines Lpx-B1.1 lipoxygenase locus, suggesting that it was associ- (RIL) from the cross between experimental line UC1113 ated with pigment degradation during pasta processing. The (intermediate pasta quality) with the cultivar Kofa (excel- 7B QTL for pasta color was linked to the Phytoene lent pasta quality). A total of 269 markers, including 23 synthase 1 (Psy-B1) locus suggesting diVerence in pigment SNP markers, were arranged on 14 linkage groups covering biosynthesis. QTLs aVecting pasta Wrmness and cooking a total length of 2,140 cM. Samples from each RIL from loss were detected on chromosomes 5A and 7B, and in both Wve diVerent environments were used for complete pasta cases they were overlapping with QTL for grain protein quality testing and the results from each year were used for content and wet gluten content. These last two parameters QTL analyses. The combined eVect of diVerent loci, were highly correlated with pasta Wrmness (R > 0.71) and inversely correlated to cooking loss (R < ¡0.37). The loca- tion and eVect of other QTLs aVecting grain size and weight, gluten strength, mixing properties, and ash content Communicated by P. Langridge. are also discussed. Electronic supplementary material The online version of this article (doi:10.1007/s00122-008-0869-1) contains supplementary Introduction material, which is available to authorized users. W. Zhang · O. Chicaiza · J. C. Brevis · J. Dubcovsky (&) Approximately 30 million tons of durum wheat (Triticum Department of Plant Sciences, One Shields Av., turgidum L. var. durum) is produced every year in diVerent University of California, Davis, CA 95616, USA regions of the world (http://www.fas.usda.gov/pecad/high- e-mail: [email protected] lights/2005/07/durum2005). Since durum wheat is mainly S. Chao used for pasta, the varieties that meet the requirements of USDA-ARS Biosciences Research Laboratory, high-quality pasta products receive premium prices in the 1605 Albrecht Blvd, Fargo, ND 58105, USA global market. These requirements include bright yellow color, high protein content and pasta Wrmness, and small F. Manthey Department of Plant Sciences, cooking loss (for a review see Troccoli et al. 2000). North Dakota State University, Fargo, ND 58105, USA Pasta color is determined by grain carotenoid content and carotenoid degradation by lipoxygenases during pasta V. Echenique processing (Troccoli et al. 2000). The main carotenoid pig- Departamento de Agronomía, Universidad Nacional del Sur, CERZOS (CONICET), San Andrés 800, 8000 Bahía Blanca, ment in the durum grain is lutein, which contributes to both Argentina pasta quality and nutritional value (Hentschel et al. 2002). 123 1362 Theor Appl Genet (2008) 117:1361–1377 Yellow pigment is mainly controlled by additive gene and pasta color, high protein content, and strong gluten. eVects and has high heritability (Clarke et al. 2006; Clarke UC1113 is a breeding line from the UC Davis wheat breed- et al. 2000; ElouaW et al. 2001). Equally important parame- ing program selected from CIMMYT cross CD52600 ters for pasta quality are pasta Wrmness and cooking loss, (KIFS//RSS/BD1419/3/MEXIS-CP/4/WAHAS/5/YAV79). which are associated with grain protein content (GPC) and This line has excellent agronomic performance but interme- gluten strength (Sissons et al. 2005). diate pasta quality parameters. A total of 93 recombinant The construction of detailed durum wheat genetic maps inbred lines (RIL) and the two parental lines were grown at using molecular markers has facilitated a more precise UC Davis, California in 2003, 2004, 2006 (Sacramento delimitation of the chromosome regions aVecting some of Valley, CA) and at the Desert Research and Extension Cen- these pasta quality traits. The Wrst durum maps were con- ter (Imperial Valley, CA) in 2005 and 2006. Fertilization structed using restriction fragment length polymorphism included at least 220 kg/ha of nitrogen and optimum irriga- (RFLP) markers (Blanco et al. 1998) and were then com- tion. The Weld trials were organized in a randomized com- plemented with simple sequence repeat (SSR) markers plete block design (RCBD) with three replications (plots (Korzun et al. 1999). Additional maps were published more were 1.2-m wide by 2.5–3.6 m long). Seeds from the three recently integrating diVerent types of molecular markers replications were pooled for quality analyses, which were (ElouaW and Nachit 2004; Maccaferri et al. 2008; Nachit performed at the Durum Wheat Quality Laboratory at et al. 2001; Peleg et al. 2008; Pozniak et al. 2007). Several North Dakota State University, Fargo, ND. of these maps have been used to identify quantitative trait High levels of stripe rust infection were observed at UC loci (QTL) for quality traits including grain yellow pigment Davis during the 2003 season and the RILs from the content (ElouaW et al. 2001; Patil et al. 2008; Pozniak et al. UC1113 £ Kofa population showed segregation for diVer- 2007; Zhang and Dubcovsky 2008), protein content ent levels of resistance (no completely susceptible lines (Blanco et al. 2006; Joppa et al. 1997; Olmos et al. 2003; were observed). The 2004 season at UC Davis showed low Uauy et al. 2006), test weight and kernel weight (ElouaW levels of stripe rust, and no infection was detected in the and Nachit 2004), and gluten strength (ElouaW et al. 2000). UC1113 £ Kofa RILs. In 2006 the RILs were treated with However, some of the previous QTL have been identi- fungicide to prevent damage from stripe rust. The two Wed in crosses including wild tetraploid parental lines experiments at Imperial Valley had no stripe rust or other (ElouaW and Nachit 2004; Gonzalez-Hernandez et al. 2004; diseases. In summary, only the 2003 results from the exper- Joppa et al. 1997) and therefore, have limited application to iment at UC Davis might have been aVected by stripe rust. modern durum germplasm. In addition, no information is available for more complex parameters that require full pasta evaluations, such as pasta color, pasta Wrmness or Genetic map cooking loss. Even for the traits for which there is some QTL information, additional mapping populations are Parental lines were screened using 1,235 wheat microsatel- needed to obtain a more complete picture of the diVerent lite markers previously mapped in the A and B genomes. genetic factors aVecting pasta quality traits in the modern AmpliWcation fragments were separated in 6% non-dena- durum germplasm. To Wll this gap in our understanding of turing acrylamide gel (29:1) and stained directly with ethi- the genetic factors underlying important quality parameters dium bromide (http://maswheat.ucdavis.edu/PDF/SSR_ we developed a genetic map between adapted durum wheat Protocol.pdf). The same parental lines were screened also varieties of diVerent quality and performed full pasta analy- with 275 wheat SNP markers generated by the NSF-Wheat ses for Wve diVerent environments. The QTL analyses of SNP project (http://wheat.pw.usda.gov/SNP) by template- these data provided valuable information and molecular directed dye-terminator incorporation assay with Xuores- markers that will be useful to accelerate the selection of cence polarization detection (FP-TDI) (Chao et al. 2008; durum varieties with improved pasta quality. Hsu et al. 2001). Linkage analysis was carried out using MapMaker ver- sion 3.0b (Lander et al. 1987). Map distances were com- Materials and methods puted with the Kosambi mapping function. The map was initially constructed at a LOD of 2.0. Additional markers Materials were added using the TRY command and their order was Wne-tuned using the RIPPLE command. Regions for which The durum mapping population was produced from the the marker orders are supported by LOD score values lower cross between UC1113 and Kofa. Kofa is a Desert Durum® than 2.0 were indicated by vertical lines on the left side of variety developed by Western Plant Breeders (now West- the map (Fig. 1). Centromere localizations were estimated Bred) that has excellent pasta quality with optimal semolina based on previous determinations of the arm location of 123 Theor Appl Genet (2008) 117:1361–1377 1363 Fig. 1 Linkage map of the 1A 1B 2A Kofa £ UC1113 population. 0.0 Bla 0.0 wmc382 Map positions are given in cM. 23.8 wmc95 0.0 gwm374 Red morphological marker 24.4 cfa2158, wmc329 15.4 wmc406 19.3 ksuE18, wmc193 22.8 gwm296 (Black Glume); Blue SNP mark- 39.1 wmc24 20.8 barc8 ers; Bold RFLP, STS and protein 47.5 barc148 28.2 gwm413 55.2 wmc296 73.2 wg983 34.4 gwm273 55.9 gwm122 markers, rest SSR markers. Dis- 35.1 wmc626 75.2 barc83 65.8 barc309 gwm135 40.2 wmc85 tances connected by a vertical 76.4 45.3 BE500714_237 66.6 gwm275 BM140362_603 line on the left side of the map 87.0 46.4 barc240 70.0 gwm515 89.3 wmc312 47.1 71.3 gwm249 indicate marker orders supported 91.6 cfa2129 BE443797_436 79.0 barc220, gwm71 113.3 dupw38 60.5 barc302 by LOD scores < 2.0.