Published July 19, 2018

RESEARCH

Genetic Diversity of a Germplasm Collection of Confectionery Sunflower Landraces from

B. Pérez-Vich, M. R. Aguirre, B. Guta, J. M. Fernández-Martínez, and L. Velasco*

Instituto de Agricultura Sostenible (IAS-CSIC), Alameda del Obsipo ABSTRACT s/n, 14004 Córdoba, Spain. Received 14 Feb. 2018. Accepted 1 June Native to North America, non-oilseed, confec- 2018. *Corresponding author ([email protected]). Assigned to tionery sunflower (Helianthus annuus L.) has Associate Editor Stella Kantartzi. been traditionally cultivated in Spain since its Abbreviations: AMOVA, analysis of molecular variance; He, expected introduction from the New World in the 16th heterozygosity; IAS-CSIC, Instituto de Agricultura Sostenible–Consejo century. This created great genetic diversity Superior de Investigaciones Científicas; INIA, National Institute for in the form of local landraces, whose charac- Agricultural Research; PCoA, principal coordinates analysis; PCR, terization and conservation is of paramount polymerase chain reaction; PIC, polymorphic information content; importance. In this research, several seed and SSR, simple sequence repeat. plant traits, as well as flowering time, were evaluated in a collection of 192 landraces of confectionery sunflower from Spain. Evalu- unflower (Helianthus annuus L.) is currently a major crop ation was conducted in Córdoba, Spain, in Sat the world scale, with an annual production over 40 Tg of 2011, 2012, and 2013. The greatest variability grains (FAOSTAT, 2017). Around 90% of sunflower production is was observed for hundred-seed weight (4.21– dedicated to oil extraction, while the major part of the remaining 19.68 g), plant height (65.00–361.67 cm), head 10% corresponds to non-oilseed or confectionery sunflower. diameter (9.00–31.00 cm), and days to flow- The latter is mainly used for confection seeds, snack food, and ering (64.31–103.00 d). Genetic diversity in the collection was also evaluated with a set pet food (Fernández-Cuesta et al., 2012). Current oilseed and of 52 simple sequence repeat (SSR) markers, confectionery cultivars widely differ for achene characteristics. which produced 167 alleles, with an average Oilseed cultivars have small black seeds with low hull proportion of 3.2 alleles per locus (from 2 to 5). The SSR and high oil content, typically ?50%. Conversely, confectionery markers disclosed moderate variability in the cultivars have large seeds that are typically black or grey with germplasm collection, with average Nei’s white stripes, high hull proportion, and low oil content, typically expected heterozygosity of 0.29 (from 0.02 to <30% (Fernández-Martínez et al., 2009). 0.50). The analysis of the structure of the germ- Sunflower is native to North America, where it was grown by plasm collection revealed the existence of two Native Americans as far back as 3000 BC, mainly in the territory separated genetic pools, one of them widely corresponding to present-day Arizona and New Mexico (Putt, distributed throughout the country and another 1997). There is some discrepancy about the date when sunflower one tracing back to a reduced area in the north reached Europe and its origin. Putt (1997) stated that Spanish of Córdoba Province, where accessions with a high level of membership in this group are explorers introduced sunflower to Spain for the first time in 1510 still relatively common. Genetic diversity of from New Mexico. An alternative hypothesis is that Spanish this germplasm can be of use for widening the explorers found the sunflower plant in Peru in 1532, where it genetic base of cultivated sunflower. Published in Crop Sci. 58:1972–1981 (2018). doi: 10.2135/cropsci2018.02.0108

© Crop Science Society of America | 5585 Guilford Rd., Madison, WI 53711 USA This is an open access article distributed under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1972 www.crops.org crop science, vol. 58, september–october 2018 was worshipped as a sacred image of the sun god, and Table 1. Collection details of the 192 confectionery sunflower Spanish landraces used in the genetic diversity analysis. it was taken to Spain (Lehner and Lehner, 1960). The latter hypothesis is followed by the Dictionary of Spanish Accession Region Province Location Language of the Royal Spanish Academy, where sunflower NC020027 Castile–La Mancha Ciudad Real Daimiel is described as “native to Peru” (RAE, 2014). NC020055 Castile–La Mancha Ciudad Real Calzada de Calatrava NC020085 Castile–La Mancha Ciudad Real Valdepeñas After its introduction in the 16th century, sunflower NC020110 Castile–La Mancha Ciudad Real Villanueva de los Infantes has been traditionally grown in Spain as an ornamental NC020143 Castile–La Mancha Ciudad Real Terrinches plant in gardens and for the use of seeds as a snack. It NC020154 Castile–La Mancha Ciudad Real Alhambra has also been used as a border plant in small vegetable NC020165 Castile–La Mancha Ciudad Real Alhambra gardens (Pardo-Pascual, 1942). This has resulted in the NC020355 Castile–La Mancha Ciudad Real Almodovar del Campo development of numerous local landraces whose conser- NC020418 Castile–La Mancha Ciudad Real Membrilla vation is of paramount importance. However, the NC020569 Castile–La Mancha Guadalajara Siguenza introduction of improved cultivars in the second half of NC024377 Castile–La Mancha Albacete Balazote the 20th century, together with a gradual replacement of NC024407 Castile–La Mancha Albacete Povedilla sunflower by maize Zea( mays L.) as a border plant in small NC024429 Castile–La Mancha Albacete Villapalacios vegetable gardens, has led to a drastic genetic erosion of NC024454 Castile–La Mancha Albacete Bienservida confectionery sunflower germplasm in Spain. In the mid- NC024619 Castile–La Mancha Cuenca Zafra de Zancara NC024640 Castile–La Mancha Cuenca Pozorrubio 1980s, the then-called National Institute for Agricultural NC026188 Castile–La Mancha Toledo Ocaña Research (INIA) started the collection of local land- NC026190 Castile–La Mancha Toledo Ocaña races of confectionery sunflower, with the collaboration NC052903 Castile and León León La Pola de Gordon of J.M. Fernández-Martínez (Instituto de Agricultura NC056389 Murcia Murcia Lorca Sostenible–Consejo Superior de Investigaciones Cientí- NC069019 Castile and León Zamora Fuentesauco ficas [IAS-CSIC], Córdoba, Spain). Collection efforts NC069151 Castile and León Salamanca Cantalpino were interrupted until the 2000s, when new collection NC069457 Castile and León Zamora Justel expeditions were conducted by L. Velasco (IAS-CSIC, NC074507 Castile and León Soria Burgo de Osma-Ciudad Córdoba, Spain). This resulted in a germplasm collection of de Osma 196 accessions of confectionery sunflower that is currently NC074589 Castile and León Burgos Melgar de Fernamental NC076152 Castile and León Palencia Calzada de los Molinos maintained at the Centre of Plant Genetic Resources of NC077705 Extremadura Badajoz Fregenal de la Sierra INIA. Part of the collection was previously evaluated for NC079052 Huelva Encinasola seed quality traits, which revealed large genetic variation NC083389 Andalusia Jaén Castillo de Locubin for some of the traits (Velasco et al., 2014). The objec- NC083633 Andalusia Cordoba Villanueva de Cordoba tive of the present research was to evaluate the germplasm NC086191 Andalusia Cadiz Villamartin collection for several seed and plant traits and to analyze NC087931 Andalusia Jaén Pozo Alcon genetic diversity using microsatellite (simple sequence NC087939 Andalusia Jaén Santiago-Pontones repeat [SSR]) molecular markers. NC094593 Andalusia Córdoba Belmez NC094594 Andalusia Córdoba MATERIALS AND METHODS NC094595 Andalusia Córdoba Encinas Reales NC094597 Andalusia Córdoba Germplasm Collection NC094598 Andalusia Córdoba The study was based on 187 accessions of the INIA germplasm NC094599 Andalusia Córdoba Espiel collection of confectionary sunflower (http://wwwx.inia.es/ NC094600 Andalusia Córdoba Espiel coleccionescrf/PasaporteCRF.asp) plus five additional accessions NC094601 Andalusia Córdoba collected by the authors. The accessions were collected in the NC094602 Andalusia Córdoba Córdoba following Spanish regions: Andalusia (129 accessions), Castile– NC094603 Andalusia Córdoba Benameji La Mancha (40), Extremadura (15), Castile and León (7), and NC094604 Andalusia Córdoba Hinojosa del Duque Murcia (1). The complete list of accessions used in this research NC094605 Andalusia Córdoba Espiel and their collection locations is given in Table 1. An overview of NC094606 Andalusia Córdoba Benameji the distribution of collection sites is shown in Fig. 1. NC094607 Andalusia Córdoba Hinojosa del Duque NC094608 Andalusia Córdoba Plant Cultivation and Tissue Collection NC094609 Andalusia Córdoba Hinojosa del Duque The accessions were grown in the field at the experimental NC094611 Andalusia Córdoba farm of the Institute for Sustainable Agriculture in 2011, 2012, NC094612 Andalusia Córdoba Espiel and 2013. Most of the accessions were characterized by a high NC094613 Andalusia Córdoba El Viso level of self-incompatibility and high plant stature, >4 m in NC094614 Andalusia Córdoba Encinas Reales some cases. Both characteristics hindered to a large extent seed NC094615 Andalusia Córdoba Villanueva de Córdoba multiplication. All the plants were bagged and sib-mated using NC094616 Andalusia Córdoba crop science, vol. 58, september–october 2018 www.crops.org 1973 Table 1. Continued. Table 1. Continued. Accession Region Province Location Accession Region Province Location NC094617 Andalusia Córdoba Espiel NC094679 Andalusia Granada Algarinejo NC094618 Andalusia Córdoba NC094680 Andalusia Granada Algarinejo NC094620 Andalusia Córdoba Fuente Obejuna NC094681 Andalusia Córdoba NC094621 Andalusia Córdoba Fuente Obejuna NC094682 Andalusia Córdoba NC094623 Andalusia Córdoba Encinas Reales NC094683 Andalusia Jaén NC094624 Andalusia Córdoba La Carlota NC094684 Andalusia Jaén Alcalá la Real NC094625 Andalusia Córdoba La Carlota NC094685 Andalusia Jaén Castillo de Locubín NC094626 Andalusia Córdoba NC094686 Andalusia Jaén Valdepeñas de Jaén NC094627 Andalusia Córdoba NC094687 Andalusia Jaén Fuensanta de NC094628 Andalusia Córdoba Hinojosa del Duque NC094688 Andalusia Jaén Martos NC094629 Andalusia Córdoba Córdoba NC094690 Andalusia Jaén NC094630 Andalusia Córdoba Añora NC094691 Andalusia Jaén Fuensanta de Martos NC094631 Andalusia Córdoba Alcaracejos NC094693 Andalusia Jaén Rus NC094694 Andalusia Jaén NC094632 Andalusia Córdoba Alcaracejos NC094695 Andalusia Jaén Castellar NC094633 Andalusia Córdoba Villanueva del Duque NC094696 Andalusia Jaén Castellar NC094634 Andalusia Córdoba Hinojosa del Duque NC094697 Andalusia Jaén NC094636 Andalusia Jaén Martos NC094698 Andalusia Jaén Santisteban del Puerto NC094637 Andalusia Jaén Martos NC094699 Andalusia Jaén Santisteban del Puerto NC094639 Andalusia Córdoba Belmez NC094700 Andalusia Jaén Santisteban del Puerto NC094640 Andalusia Córdoba Belmez NC094701 Andalusia Jaén NC094641 Andalusia Córdoba Belmez NC094702 Andalusia Jaén Villanueva del Arzobispo NC094642 Andalusia Córdoba Villanueva de Córdoba NC094703 Andalusia Jaén Villanueva del Arzobispo NC094643 Andalusia Córdoba Villanueva del Rey NC094704 Andalusia Jaén Villanueva del Arzobispo NC094644 Andalusia Córdoba Posadilla (Fuente Obejuna) NC094705 Andalusia Jaén Villanueva del Arzobispo NC094706 Andalusia Jaén Puente de Génave NC094645 Andalusia Córdoba Villaviciosa de Córdoba NC094707 Andalusia Jaén Puente de Génave NC094646 Andalusia Córdoba Villaviciosa de Córdoba NC094708 Andalusia Jaén Puente de Génave NC094647 Andalusia Córdoba Córdoba NC094709 Andalusia Jaén Puente de Génave NC094648 Andalusia Sevilla Villanueva del Río NC094710 Andalusia Jaén NC094649 Andalusia Córdoba Hinojosa del Duque NC094711 Andalusia Jaén NC094650 Andalusia Córdoba NC094712 Andalusia Jaén Segura de la Sierra NC094651 Andalusia Córdoba Hinojosa del Duque NC094713 Andalusia Granada Benalúa de las Villas NC094652 Andalusia Córdoba Benameji NC094714 Andalusia Granada Benalúa de las Villas NC094653 Andalusia Córdoba Alcaracejos NC094715 Andalusia Granada Benalúa de las Villas NC094654 Andalusia Córdoba Espiel NC094716 Andalusia Granada Benalúa de las Villas NC094655 Andalusia Córdoba Montoro NC094717 Andalusia Granada Pinos Puente NC094718 Andalusia Granada Oguijares NC094656 Andalusia Córdoba Villanueva del Duque NC094719 Andalusia Granada Oguijares NC094658 Andalusia Córdoba Villaviciosa de Córdoba NC094720 Andalusia Málaga Villanueva del Trabuco NC094659 Andalusia Córdoba Posadas NC094721 Andalusia Badajoz Higuera la Real NC094660 Andalusia Córdoba El Viso NC094722 Andalusia Córdoba Encinas Reales NC094661 Andalusia Córdoba Benameji NC094723 Andalusia Córdoba Hinojosa del Duque NC094662 Andalusia Córdoba Alcaracejos NC094724 Andalusia Córdoba Alcaracejos NC094663 Andalusia Córdoba NC094725 Extremadura Badajoz Santa Marta NC094664 Andalusia Córdoba Espiel NC094726 Extremadura Badajoz Zalamea de la Serena NC094665 Andalusia Córdoba Encinas Reales NC094727 Extremadura Badajoz Quintana de la Serena NC094667 Andalusia Córdoba Córdoba NC094728 Extremadura Badajoz Cabeza de Buey NC094668 Andalusia Córdoba NC094729 Extremadura Badajoz La Coronada NC094730 Extremadura Badajoz Quintana de la Serena NC094669 Andalusia Córdoba Baena NC094731 Extremadura Badajoz Quintana de la Serena NC094670 Andalusia Córdoba Luque NC094732 Extremadura Badajoz Valencia de las Torres NC094671 Andalusia Córdoba Moriles NC094733 Extremadura Badajoz Cabeza la Vaca NC094672 Andalusia Córdoba Moriles NC094734 Extremadura Badajoz Segura de León NC094673 Andalusia Córdoba Moriles NC094735 Extremadura Badajoz Segura de León NC094674 Andalusia Córdoba NC094736 Extremadura Badajoz Higuera la Real NC094675 Andalusia Córdoba Monturque NC094737 Extremadura Badajoz Segura de León NC094676 Andalusia Córdoba Monturque NC094738 Extremadura Badajoz Segura de León NC094677 Andalusia Córdoba Las Navas del Selpillar NC094739 Castilla la Mancha Ciudad Real Fuencaliente NC094678 Andalusia Córdoba Villafranca de Córdoba NC094740 Castilla la Mancha Ciudad Real Brazatortas NC094741 Castilla la Mancha Ciudad Real Almodovar del Campo

1974 www.crops.org crop science, vol. 58, september–october 2018 Table 1. Continued. Such modifications were: (i) addition of 0.1% (w/v) ascorbic Accession Region Province Location acid, 0.1% (w/v) diethyldithiocarbamic acid sodium salt, and NC094742 Castilla la Mancha Ciudad Real Torralba de Calatrava 0.2% (v/v) 2-mercaptoethanol to the cetyltrimethylammonium NC094743 Castilla la Mancha Ciudad Real Daimiel bromide (CTAB) extraction buffer just before use; and (ii) use of NC094744 Castilla la Mancha Cuenca chloroform instead of chloroform:isoamyl alcohol (24:1). Since NC094745 Castilla la Mancha Cuenca for all the landraces used in this study the plants were sib-mated, NC094747 Castile–La Mancha Toledo Madridejos within-accession diversity was not evaluated and equal amounts NC094748 Castile–La Mancha Toledo Madridejos of DNA of 15 plants per accession were pooled and used as a NC094749 Castile–La Mancha Toledo Corral de Almaguer template for polymerase chain reaction (PCR) amplification. NC094750 Castile–La Mancha Cuenca Almodovar del Pinar A total of 95 SSR markers distributed across the 17 sunflower NC094751 Castile–La Mancha Cuenca Almodovar del Pinar NC094752 Castile–La Mancha Cuenca Almodovar del Pinar genetic linkage groups (Tang et al., 2003) were screened initially NC094753 Castile–La Mancha Cuenca using 16 diverse accessions with different geographical distri- NC094754 Castile–La Mancha Toledo Camuñas butions and traits. This screening gave rise to 52 markers with NC094755 Castile–La Mancha Toledo Madridejos clearly scorable and polymorphic bands, providing good coverage NC094756 Castile–La Mancha Toledo Madridejos across the sunflower genome (Table 2). These 52 SSR markers NC094757 Castile–La Mancha Ciudad Real Manzanares were used for subsequent genotyping of the complete set of 192 NC094758 Andalusia Córdoba El Viso accessions. The PCR reactions were performed using 30 mL of NC094759 Castile–La Mancha Ciudad Real Villarta de San Juan reaction mixture containing 1´ PCR buffer, 1.5 mM MgCl2, NC094760 Castile–La Mancha Albacete Tobarra 0.2 mM of deoxynucleotides, 0.3 mM of 3¢- and 5¢-end primers, NC094761 Castile–La Mancha Albacete Pozohondo 0.7 units of Taq DNA polymerase (Biotaq DNA Polymerase, NC094762 Castile–La Mancha Albacete Tobarra Bioline), and 50 ng of genomic DNA. Conditions for PCR were: a pool of pollen from all the plants of the accession. This fact, denaturation at 94°C for 2.5 min, annealing temperature of together with severe occurrence of some diseases, particularly +10°C for 30 s, 72°C for 30 s, nine cycles in which the annealing Rizhopus, determined that not all the accessions produced seed temperature was decreased by 1°C every other cycle, 32 cycles at in the 3 yr. Phenotypic data for each accession were based on 94°C for 30 s, annealing temperature for 30 s, and 72°C for 30 s information collected in at least two environments, although with a final extension of 20 min at 72°C. Annealing tempera- no information on all the accessions could be obtained in a tures varied from 52 to 60°C. The amplification products were single environment. In all cases, 40 seeds per accession were separated on 3% Metaphor agarose (BMA) gels in 1´ TBE buffer germinated in moistened filter paper and sown in small pots, stained with SaveView Nucleic Acid Stain (NBS Biologicals). and the plants were transplanted in the field after 3 wk in a The gels were visualized under ultraviolet light. A 100-bp DNA growth chamber. Plants of each accession were grown in a ladder (Solis BioDyne) was used as a molecular weight marker single row 8 m long with 20 cm of plant spacing and 2-m sepa- to get an approximate size of DNA fragments. The presence or ration between accessions. The plants were furrow irrigated as absence of bands was scored as 1 or 0, respectively. The Quantity needed. Two fully expanded leaves were cut from 15 randomly One 1-D Analysis Software (Bio-Rad Laboratories, 2006) was selected plants per accession before flowering. used for the allele designation by comparing allele bands to the 100-bp DNA ladder. Phenotypic Evaluation Days to flowering, measured as the average value per acces- Data Analysis sion of the number of days between planting and flowering The bands of amplified products were scored as present (1) or time for each individual plant of the accession, was recorded. absent (0) and a binary matrix was constructed, in which the Plant height and head diameter were measured for each plant bands were treated as individual alleles. The reason for this was at the end of flowering. After harvesting, hundred-achene that more than two alleles per sample were observed in some weight, hundred-kernel weight, and the proportion of hull in cases, as the templates used in SSR analysis were pooled DNA the achene were determined. Achene length and width were from 15 plants from each accession. The frequency of each allele measured on 25 randomly selected achenes per accession. Seed in the germplasm collection was calculated, and the polymor- color and the presence and position of stripes were not included phic information content (PIC) for each locus was estimated as 2 because most of the accessions were very variable for these traits PIC = 1 − S(pi ), where pi is the proportion of samples carrying (i.e., they contained seeds of different colors and/or showed the ith allele (Singh and Singh, 2015). Nei’s expected hetero- differences for presence and position of stripes). zygosity (He) was computed for all loci as a measure of genetic Oil content was measured on previously desiccated achenes diversity. All parameters were calculated using GenAlEx 6.501 (17 h at 103°C) by nuclear magnetic resonance (Granlund and (Peakall and Smouse, 2012). Zimmerman, 1975) on an Oxford 4000 analyzer (Oxford Analyt- The binary matrix was transformed into a genetic simi- ical Instruments). For all the traits, data from 2 yr were averaged. larity matrix using Jaccard coefficient and principal coordinates analysis (PCoA) was conducted using NTSYS-pc 2.2 software Molecular Characterization (Rohlf, 2010). Identification of genetically homogeneous groups in the Leaves were frozen at −80°C, lyophilized, and ground to a fine germplasm collection was conducted using STRUCTURE powder in a laboratory mill. DNA was isolated using the protocol version 2.3.4 (Pritchard et al., 2000). The software iteratively described by Rogers and Bendich (1985) with small modifications. clusters the samples on the basis of a user-supplied K number crop science, vol. 58, september–october 2018 www.crops.org 1975 Fig. 1. Map of Spain showing the collection sites of confectionery sunflower landraces. lat, latitude; lon, longitude. of groups, with samples being assigned probabilistically to one Finally, analyses of molecular variance (AMOVA) were or several groups. An admixture model following the Hardy– conducted based on groups defined by specific areas of origin Weinberg equilibrium was used. The analysis was repeated 10 of the accessions, at the level of region, province, and even at times for each value of K (from 1 to 20) using a burn-in period the intraprovince level in the case of Córdoba, which showed of 100,000 Markov chain Monte Carlo (MCMC) iterations and a particular pattern of variation. Geographical groups were a run length of 100,000. The number of groups that best fit the only considered in those cases in which a sufficient number of data was determined using Structure Harvester (Earl and von accessions per area were available. The AMOVA analyses were Holdt, 2012) with the Evanno correction (Evanno et al., 2005). performed with GenAlEx 6.501. Permutations of the output of STRUCTURE analysis were calculated using the FullSearch algorithm provided in CLUMPP version 1.1.2b (Jakobsson and Rosenberg, 2007), and the output RESULTS was used to produce bar graphs of the population structure using Phenotypic data of the germplasm collection for seed and Origin Pro 9.1 software (OriginLab Corporation, 2017). plant traits, as well as for days to flowering, are shown

1976 www.crops.org crop science, vol. 58, september–october 2018 Table 2. Details of simple sequence repeat (SSR) markers used for evaluating genetic diversity in 192 confectionery sunflower Spanish landraces. No. of No. of SSR marker alleles PIC† GenBank‡ LG§ SSR marker alleles PIC GenBank LG ORS 7 4 0.48 BV012511 15 ORS 621 5 0.70 BV006097 11 ORS 70 3 0.52 BV006893 8 ORS 656 2 0.52 BV006715 16 ORS 166 3 0.29 BV012434 8 ORS 665 3 0.55 BV006123 3 ORS 202 4 0.38 BV012456 3 ORS 687 3 0.35 BV006138 15 ORS 229 4 0.56 BV012471 2 ORS 691 2 0.45 BV006140 10 ORS 230 4 0.45 BV012472 6 ORS 694 3 0.48 BV006721 14 ORS 297 3 0.40 BV006634 17 ORS 716 4 0.66 BV006728 1 ORS 307 4 0.32 BV005910 14 ORS 733 3 0.36 BV006733 11 ORS 309 2 0.18 BV005912 4 ORS 750 3 0.62 BV006177 16 ORS 316 2 0.50 BV005917 13 ORS 774 2 0.39 BV006192 5 ORS 342 2 0.27 BV005933 2 ORS 785 4 0.52 BV006200 4 ORS 366 2 0.23 BV005949 4 ORS 844 2 0.23 BV006252 9 ORS 420 3 0.48 BV005977 15 ORS 852 3 0.36 BV006258 5 ORS 428 3 0.50 BV005982 9 ORS 887 4 0.62 BV006290 9 ORS 437 3 0.66 BV006848 10 ORS 899 4 0.53 BV006302 16 ORS 442 2 0.51 BV006670 9 ORS 938 4 0.53 BV006337 9 ORS 456 2 0.44 BV005996 17 ORS 963 3 0.40 BV006359 4 ORS 457 3 0.60 BV005997 11 ORS 966 3 0.37 BV006361 7 ORS 505 4 0.36 BV006027 5 ORS 993 4 0.62 BV006383 16 ORS 533 4 0.72 BV006044 5 ORS 1024 4 0.65 BV006408 5 ORS 534 3 0.62 BV006045 13 ORS 1041 3 0.48 BV006423 7 ORS 543 4 0.77 BV006051 1 ORS 1065 2 0.29 BV006445 2 ORS 561 3 0.57 BV006691 17 ORS 1079 4 0.39 BV006457 14 ORS 565 5 0.57 BV006065 17 ORS 1114 2 0.42 BV006484 3 ORS 595 5 0.62 BV006853 10 ORS 1146 2 0.24 BV006511 11 ORS 613 3 0.30 BV006091 10 ORS 1161 4 0.65 BV006524 8 † PIC, polymorphic information content.

‡ GenBank locus identification.

§ LG, linkage group. in Table 3. The greatest variability was observed for germplasm collection, with an average value of Nei’s He hundred-seed weight from 4.21 to 19.68 g, for plant height of 0.29 (from 0.02 to 0.50). from 65.00 to 361.67 cm, for head diameter from 9.00 to Principal coordinate analysis showed genetic diversity 31.00 cm, and for days to flowering from 64.31 to 163.00. in the germplasm collection. Figure 2 shows the two first Data for individual accessions are given in Supplemental coordinates, which explained 4.6 and 3.0% of the varia- Table S1. Hundred-seed weight was positively correlated tion, respectively. A group of 14 accessions from Andalusia with both seed length (r = 0.78, P < 0.01) and seed width clearly differentiated from the rest of the collection. (r = 0.75, P < 0.01), which in turn were also positively Thirteen of them were collected in six villages from two correlated (r = 0.62, P < 0.01). Strong positive correla- Table 3. Mean, SD, minimum, and maximum average values tion was also observed between plant height and days to for seed length, seed width, ratio seed length/width, hundred- flowering r( = 0.78, P < 0.01) and between oil content and seed weight, percentage of kernels, seed oil content, plant percentage of kernels in the achene (r = 0.61, P < 0.01). height, head diameter, and days to flowering in a collection of Conversely, the main negative correlation (r = −0.43, P < 192 landraces of confectionery sunflower from Spain. 0.01) involved the percentage of kernels and seed width Trait Mean SD Min. Max. (Table 4). Seed length (cm) 1.36 0.16 0.91 1.75 Evaluation of the 192 landraces with the selected 52 Seed width (cm) 0.77 0.12 0.50 1.15 SSR markers produced 167 alleles, with an average of 3.2 Seed length/width 1.79 0.22 1.32 2.39 alleles per locus and minimum and maximum values of Hundred-seed weight (g) 10.29 2.62 4.21 19.68 two and five, respectively (Table 2). The allele frequency Kernels (%)\ 54.44 5.53 39.32 68.10 in the collection ranged from 2.1 to 90.1%. The average Seed oil content (%) 22.15 2.79 16.00 32.20 Plant height (cm) 173.10 57.81 65.00 361.67 PIC number for SSRs was 0.47 (from 0.18 to 0.77). Head diameter (cm) 14.88 2.83 9.00 31.00 The SSR markers disclosed moderate variability in the Days to flowering 92.72 19.87 64.31 163.00 crop science, vol. 58, september–october 2018 www.crops.org 1977 Table 4. Correlation coefficients between phenotypic traits measured in 192 landraces of confectionery sunflower collected in Spain. Seed 100-seed Plant Head Days to Oil Trait width weight % Kernels height diameter flowering content Seed length 0.62** 0.78** −0.18* −0.27** 0.12 −0.27** 0.03 Seed width 0.72** −0.43** −0.24** 0.27** −0.23** −0.11 100-seed weight −0.17* −0.29** 0.19** −0.27** −0.01 %Kernels −0.02 −0.17* −0.05 0.61** Plant height 0.25** 0.78** −0.03 Head diameter 0.06 −0.15* Days to flowering −0.17* *, ** Significant at the 0.05 and 0.01 probability levels, respectively.

well-defined geographical areas in the northern area of 0.53, P < 0.01) and negatively correlated with seed length Córdoba Province: El Viso, Hinojosa del Duque, Villan- (r = −0.32, P < 0.01), seed width (r = −0.39, P < 0.01), and ueva del Duque, and Villaralto in the Pedroches Valley, and hundred-kernel weight (r = −0.32, P < 0.01). Espiel and Villanueva del Rey in the Guadiato Valley. The The AMOVA considering the geographic groups used maximum distance between any of these villages is 36 km. for structure analysis indicated that 97% of the variation For the other accession from this group, the only informa- was found within groups and only 3% was attributable tion available is that it was collected in Córdoba Province. to differences between groups. When the AMOVA was The analysis of the structure of the germplasm collec- computed considering only those accessions with high tion revealed the existence of two major genetic groups level of membership (>85%) in the two groups defined in (Fig. 3). They were named Group 1 and Group 2. The the STRUCTURE analysis, variation between the two percentage of membership of each accession to each genetic groups was 21%, with the remaining 79% denoting varia- group is shown in Fig. 4. Group 2 was mainly represented tion within groups. in the accessions from northern Córdoba, although all the geographical areas contained some accessions with a high DISCUSSION percentage of membership in this group. The 14 distinct The results of this research indicate that the germplasm accessions identified in PCoA (Fig. 2) had, in general, a collection contains large phenotypic variation. The largest high percentage of membership in Group 2, averaging 85% variation was found for head diameter, hundred-seed (from 58 to 97%). From 51 landraces collected in the two weight, plant height, and days to flowering, with a marked main geographic areas in the north of Córdoba, Guadiato association of taller plants with late flowering. Although Valley (n = 22) and Pedroches Valley (n = 29), the highest no phenotypic information from all the accessions could percentage of membership in Group 2 was observed in be obtained in a single environment, differences were large Pedroches Valley, which contained 10 accessions with enough to be considered meaningful, taking into account >90% membership in this group (Fig. 5). The percentage that, in all cases, data from at least two environments of membership in Group 2 was positively correlated with plant height (r = 0.55, P < 0.01) and days to flowering r( =

Fig. 2. Principal coordinate (Coord.) analysis using 52 simple Fig. 3. Delta K for values of K in admixture analysis of genetic sequence repeat markers in a germplasm collection of 192 structure in a germplasm collection of 192 landraces of landraces of confectionery sunflower collected in Spain. confectionery sunflower.

1978 www.crops.org crop science, vol. 58, september–october 2018 Fig. 4. Percentage of membership in the two genetic groups Fig. 5. Percentage of membership in the two genetic groups identified in genetic structure analysis of a germplasm collection identified in genetic structure analysis of confectionery sunflower of 192 landraces of confectionery sunflower. landraces collected in two geographic areas in the north of Córdoba Province: Guadiato Valley and Pedroches Valley. were included. The large variation for flowering time in part of this germplasm collection was previously reported taken into account that the present research focused on by Velasco et al. (2014), who also reported the existence a very specific group of sunflower germplasm, made up of variation for seed quality traits such as squalene and exclusively of confectionery landraces collected in Spain. tocopherol contents and sterol profile. The existence of Analysis of population structure in sunflower has a positive correlation between flowering time and plant focused mainly on extensive collections including a wide height has been previously reported in sunflower germ- range of germplasm, but there are few studies focusing plasm (Bert et al., 2003). specifically on confectionery sunflower germplasm. Dong Previous studies on genetic diversity of sunflower et al. (2007), Kholghi et al. (2012), and Jannatdoust et germplasm have focused on different types of sunflower al. (2016) studied population structure of 70, 15, and 50 populations, including wild populations, landraces, old confectionery cultivars, respectively, from China (Dong et cultivars, and parents of modern elite hybrids, both main- al., 2007) and Iran (Kholghi et al., 2012; Jannatdoust et tainer and restorer inbred lines. A gradual narrowing of al., 2016) using amplified fragment length polymorphism allelic diversity from one end of the germplasm spectrum (AFLP), SSR, and/or retrotransposon-based markers. In (wild populations) to the other (elite oilseed inbred lines) those studies, the authors did not find a clear structure has been documented (Tang and Knapp, 2003). Mandel et defined on the basis of geographical origin. In our study, al. (2011) analyzed 433 cultivated accessions from North two separated genetic pools, one of them widely distributed America and Europe, including mainly inbred lines, throughout the country and another one tracing back to a Native American landraces, and open-pollinated vari- reduced area in the north of Córdoba Province, were iden- eties, with 34 expressed sequence tag SSRs. The authors tified. Most of the accessions of the latter area, particularly reported Nei’s He across cultivars of 0.47 and an average those from the Pedroches Valley, had a high percentage of number of alleles per locus of 6.8. Moreno et al. (2013) membership in one of the genetic groups (Group 2). The studied 14 Argentinean open-pollinated and composite high relative proportion of accessions in this geographic area populations with 16 SSRs and reported an average with a high percentage of membership in Group 2 suggests number of alleles per locus of 6.25 and Nei’s He index that this particular genetic group might have evolved in this of 0.56. Using 42 SSR markers on 137 sunflower inbred area. The presence in the Pedroches Valley of other acces- lines, 13 open-pollinated populations, and 20 composite sions with a high percentage of membership in Group 1, populations from Argentina, Filippi et al. (2015) reported as well as accessions with intermediate membership values, a Nei’s He index of 0.51, with an average number of alleles suggests the introduction of foreign germplasm. Most of per locus of 4.95. In the present research, Nei’s He was the accessions of Group 2 are very tall and late flowering. 0.29, with 3.2 alleles per locus. Although comparison of They were formerly often used for shadowing small vege- different studies is difficult because of wide differences in table gardens, and their dried stems were used as beams for plant material and methodological approach, the results drying sausage inside houses (Moreno-Valero, 1994). Such from the present study revealed a moderate level of genetic uses are not so common nowadays, which might have led diversity for confectionery landraces in Spain, compared to the introduction of shorter and earlier flowering forms with the abovementioned studies. However, it has to be from other areas.

crop science, vol. 58, september–october 2018 www.crops.org 1979 Confectionary sunflower has been a traditional Dong, G.J., G.S. Liu, and K.F. Li. 2007. Studying genetic diversity in the core germplasm of confectionary sunflower Helianthus( crop widely cultivated at a small scale in Spain (Pardo- annuus L.) in China based on AFLP and morphological analysis. Pascual, 1942). This has resulted in the development Russ. J. Genet. 43:627–635. doi:10.1134/S1022795407060051 of landraces adapted to specific geographic areas and Earl, D.A., and B.M. von Holdt. 2012. STRUCTURE HAR- specific cultural uses (Moreno-Valero, 1994). Spain’s VESTER: A website and program for visualizing STRUC- industrial development in the 1960s led to a drastic TURE output and implementing the Evanno method. reduction of active agrarian population (from 35.7% in Conserv. Genet. Resour. 4:359–361. doi:10.1007/s12686-011- 1964 to 16.3% in 1984) that moved to other sectors in the 9548-7 cities (Lieberman, 1995). This process probably contrib- Evanno, G., S. Regnaut, and J. Goudet. 2005. Detecting the num- ber of clusters of individuals using the software STRUC- uted to dispersal and mixture of germplasm, including TURE: A simulation study. Mol. Ecol. 14:2611–2620. the introduction of foreign forms after the loss of local doi:10.1111/j.1365-294X.2005.02553.x landraces. Because of the particular characteristics of FAOSTAT. 2017. FAOSTAT database. FAO, Rome. ht t p://f ao - the plants, associated with well-documented traditional stat3.fao.org (accessed 13 Feb. 2018). uses (Pardo-Pascual, 1942; Moreno-Valero, 1994), and Fernández-Cuesta, A., A. Nabloussi, J.M. Fernández-Martínez, the existence of a common genetic identity in a very and L. Velasco. 2012. Tocopherols and phytosterols in sun- restricted geographic area, our hypothesis is that the flower seeds for the human food market. Grasas Aceites Group 2 genetic pool represents ancient forms of confec- 63:321–327. doi:10.3989/gya.010112 Fernández-Martínez, J.M., B. Pérez-Vich, and L. Velasco. 2009. tionery sunflower germplasm. Conversely, accessions Sunflower. In: J. Vollmann and I. Rajcan, editors, Oil crops. with a high percentage of membership in Group 1 cannot Springer, New York. p. 155–232. doi:10.1007/978-0-387- be associated to a specific geographic area, and they are 77594-4_6 probably more recent forms that became popular due to Filippi, C.V., N. Aguirre, J.G. Rivas, J. Zubrzycki, A. Puebla, D. their shorter stature and earlier flowering. Cordes, et al. 2015. Population structure and genetic diversity In conclusion, the present research showed the characterization of a sunflower association mapping popula- existence of an overall moderate genetic diversity of tion using SSR and SNP markers. BMC Plant Biol. 15:52. doi:10.1186/s12870-014-0360-x confectionery sunflower landraces in Spain, compared Granlund, M., and D.C. Zimmerman. 1975. Oil content of sun- with other studies based on broader sets of sunflower flower seeds as determined by wide-line nuclear magnetic res- germplasm. The study revealed the existence of two onance (NMR). Proc. North Dakota Acad. Sci. 27:128–133. separated genetic pools, one of them widely distributed Jakobsson, M., and N.A. Rosenberg. 2007. CLUMPP: A cluster throughout the country and another one tracing back matching and permutation program for dealing with label to a reduced area in the north of Córdoba Province, switching and multimodality in analysis of population struc- where accessions with a high level of membership in this ture. Bioinformatics 23:1801–1806. doi:10.1093/bioinformat- group are still relatively common. Genetic diversity in ics/btm233 Jannatdoust, M., R. Darvishzadeh, R. Ziaeifard, M.A. Ebrahimi, this germplasm collection can be of use for widening the H.H. Maleki, E. Gholinezhad, and A.A. Hatamnia. 2016. genetic base of cultivated sunflower. Analysis of genetic diversity and population structure of con- fectionery sunflower Helianthus( annuus L.) native to Iran. J. Conflict of Interest Crop Sci. Biotechnol. 19:37–44. doi:10.1007/s12892-015- The authors declare that there is no conflict of interest. 0052-6 Kholghi, M., R. Darvishzadeh, I. Bernousi, A. Pirzad, and H. Supplemental Material Available Laurentin. 2012. Assessment of genomic diversity among and Supplemental material for this article is available online. within Iranian confectionery sunflower Helianthus( annuus L.) populations by using simple sequence repeat markers. Acta Agric. Scand., Sect. B 62:488–498. doi:10.1080/09064710.2 Acknowledgments 012.662243 The research was funded by National Institute for Agricultural Lehner, E., and J. Lehner. 1960. Folklore and symbolism of flow- and Food Research and Technology (INIA) Research Project ers, plants and trees. Tudor Publ. Co., New York. RF2011-00021-00-00 Lieberman, S. 1995. Growth and crisis in the Spanish economy (1940–93). Routledge, London. References Mandel, J.R., J.M. Dechaine, L.F. Marek, and J.M. Burke. 2011. Genetic diversity and population structure in cultivated sun- Bert, P.F., I. Jouan, D. Tourvieille de Labrouhe, F. Serre, J. Philip- flower and a comparison to its wild progenitor,Helianthus pon, P. Nicolas, and F. Vear. 2003. Comparative genetic annuus L. Theor. Appl. Genet. 123:693–704. doi:10.1007/ analysis of quantitative traits in sunflower Helianthus( annuus s00122-011-1619-3 L.). 2. Characterisation of QTL involved in developmen- Moreno, M.V., V. Nishinakamasu, M.A. Loray, D. Alvarez, J. tal and agronomic traits. Theor. Appl. Genet. 107:181–189. Gieco, A. Vicario, et al. 2013. Genetic characterization of sun- doi:10.1007/s00122-003-1237-9 flower breeding resources from Argentina: Assessing diver- Bio-Rad Laboratories. 2006. Quantity One user guide for version sity in key open-pollinated and composite populations. Plant 4.6.3. Bio-Rad Lab., Hercules, CA. Genet. Resour. 11:238–249. doi:10.1017/S1479262113000075

1980 www.crops.org crop science, vol. 58, september–october 2018 Moreno-Valero, M. 1994. La matanza en . Rev. Fol- Rogers, S.O., and A.J. Bendich. 1985. Extraction of DNA from clore 161:154–166. milligram amounts of fresh, herbarium and mummified plant OriginLab Corporation. 2017. Tutorials for Origin. OriginLab tissues. Plant Mol. Biol. 5:69–76. doi:10.1007/BF00020088 Corp., Northampton, MA. Rohlf, F.J. 2010. NTSYSpc. Numerical Taxonomy and Multivariate Peakall, R., and P.E. Smouse. 2012. GenAlEx 6.5: Genetic Analysis System, version 2.21h. Exeter Softw., Setauket, NY. analysis in Excel. Population genetic software for teach- Singh, B.D., and A.K. Singh. 2015. Marker-assisted plant ing and research-an update. Bioinformatics 28:2537–2539. breeding: Principles and practices. Springer, New York. doi:10.1093/bioinformatics/bts460 doi:10.1007/978-81-322-2316-0 Pardo-Pascual, M. 1942. Cultivo del girasol. Agricultura (Mari- Tang, S., V.K. Kishore, and S.J. Knapp. 2003. PCR-multiplexes for bor) 11:127–128. a genome-wide framework of simple sequence repeat marker Pritchard, J.K., M. Stephens, and P. Donnelly. 2000. Inference of loci in cultivated sunflower. Theor. Appl. Genet. 107:6–19. population structure using multilocus genotype data. Genet- doi:10.1007/s00122-003-1233-0 ics 155:945–959. Tang, S., and S.J. Knapp. 2003. Microsatellites uncover extraor- Putt, E.D. 1997. Early history of sunflower. In: A.A. Schneiter, edi- dinary diversity in Native American land races and wild tor, Sunflower production and technology. Agron. Monogr. populations of cultivated sunflower. Theor. Appl. Genet. 35. ASA, CSSA, and SSSA, Madison, WI. p. 1–21. doi:10.2134/ 106:990–1003. doi:10.1007/s00122-002-1127-6 agronmonogr35.c1 Velasco, L., A. Fernández-Cuesta, and J.M. Fernández-Martínez. RAE. 2014. Dictionary of Spanish language. 23rd ed. Real Acad. 2014. Variability of seed quality traits in a collection of Span- Española, Madrid. ish landraces of confectionery sunflower. Crop Pasture Sci. 65:242–249. doi:10.1071/CP13390

crop science, vol. 58, september–october 2018 www.crops.org 1981