Hindawi Publishing Corporation

e Scientific World Journal

Volume 2014, Article ID 150432, 15 pages
http://dx.doi.org/10.1155/2014/150432

Research Article
The Genetic Structure of Wild Orobanche cumana Wallr.
(Orobanchaceae) Populations in Eastern Bulgaria Reflects
Introgressions from Weedy Populations

Rocío Pineda-Martos,1 Antonio J. Pujadas-Salvà,2 José M. Fernández-Martínez,1

Kiril Stoyanov,3 Leonardo Velasco,1 and Begoña Pérez-Vich1
1
Institute for Sustainable Agriculture (IAS-CSIC), Finca Alameda del Obispo, Avenida Menéndez-Pidal s/n, 14004 Córdoba, Spain
2
Department of Agricultural and Forestry Sciences and Resources, University of Córdoba, Campus de Rabanales,
Edificio Celestino Mutis, Carretera de Madrid Km 396, 14071 Córdoba, Spain
3
Department of Botany, Agricultural University of Plovdiv, 12 Mendeleev, 4000 Plovdiv, Bulgaria

Correspondence should be addressed to Leonardo Velasco; [email protected]

Received 10 February 2014; Revised 17 June 2014; Accepted 1 July 2014; Published 20 July 2014

Academic Editor: António Amorim

Copyright © 2014 Rocı́o Pineda-Martos et al. This is an open access article distributed under the Creative Commons Attribution
License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly
cited.

Orobanche cumana is a holoparasitic plant naturally distributed from central Asia to south-eastern Europe, where it parasitizes
wild Asteraceae species. It is also an important parasitic weed of sunflower crops. The objective of this research was to investigate
genetic diversity, population structure, and virulence on sunflower of O. cumana populations parasitizing wild plants in eastern
Bulgaria. Fresh tissue of eight O. cumana populations and mature seeds of four of them were collected in situ on wild hosts.
Genetic diversity and population structure were studied with SSR markers and compared to weedy populations. Two main gene
pools were identified in Bulgarian populations, with most of the populations having intermediate characteristics. Cross-inoculation
experiments revealed that O. cumana populations collected on wild species possessed similar ability to parasitize sunflower to those
collected on sunflower. The results were explained on the basis of an effective genetic exchange between populations parasitizing
sunflower crops and those parasitizing wild species. The occurrence of bidirectional gene flow may have an impact on wild
populations, as new physiological races continuously emerge in weedy populations. Also, genetic variability of wild populations
may favour the ability of weedy populations to overcome sunflower resistance mechanisms.

1. Background Orobanche cumana is naturally distributed from central

Asia to south-eastern Europe, where it parasitizes wild Aster-
Broomrapes (Orobanche spp. and Phelipanche spp.) are a aceae species, mainly Artemisia spp. [4]. Even though it has
group of around 170 holoparasitic plant species mainly dis- been considered by some authors as an intraspecific taxon of
tributed in the northern hemisphere. They do not have pho- Orobanche cernua L. [5], its treatment as a separate species is
tosynthetic activity and entirely depend on a host plant for nowadays widely accepted [2, 6, 7]. The Black Sea coast in
nutrition [1]. Even though most of the Orobanche spp. only eastern Bulgaria is one of the main natural distribution areas
parasitize wild plants, some of them have become noxious for O. cumana, where this species is mainly found parasitizing
weeds on a variable range of cultivated hosts [2]. This is the Artemisia maritima L. [8].
case of Orobanche cumana Wallr. (sunflower broomrape), Though domesticated in eastern North America and
which is nowadays one of the most limiting factors for sun- widely used as a staple food in the pre-Columbian period [9],
flower (Helianthus annuus L.) production in Europe and Asia the transformation of sunflower into one of the major world
[3]. oil crops started in Russia in the second half of the nineteenth
2 The Scientific World Journal

century [10]. Plants of O. cumana parasitizing sunflower were populations of Orobanche foetida Poir. collected on a wild
observed for the first time in Russia in the 1890s [11]. In host and a population growing on cultivated vetch (Vicia
Bulgaria, O. cumana parasitization on sunflower was first sativa L.) using AFLP analyses.
detected in 1935 [12]. Currently, O. cumana is present in the Studies on genetic diversity within and between O.
main sunflower-producing countries around the world, par- cumana populations are scarce and focused on weedy popula-
ticularly in Central and Eastern Europe, Spain, Turkey, Israel, tions collected on sunflower. Gagne et al. [27] studied genetic
Russia, Ukraine, Iran, Kazakhstan, and China [2]. Moreover, diversity in eight populations from several countries using
the parasite has spread to new areas in recent years [13, 14]. RAPD markers. They identified large interpopulation and low
Broomrape seed transport has been suggested as one of intrapopulation genetic variation, concluding the existence of
the main factors in the dispersion of the infestation [15]. two main gene pools, one comprising populations from
Broomrape seeds are extremely small (dust-like seeds), and Eastern Europe and another one including populations from
individual plants can produce an impressive number that Southern Spain. Pineda-Martos et al. [28] identified two main
remain viable in the soil for up to 20 years, which are easily gene pools for O. cumana in Spain, comprising populations
dispersed by water, wind, animals, humans, machinery, or from the Guadalquivir Valley (Southern Spain) and Cuenca
though attachment to sunflower seeds [15, 16]. Province (Central Spain), respectively. Both groups were
Unlike most weedy Orobanche spp., which have a broad genetically distant, but both intra- and interpopulation
range of host crops, weedy O. cumana only parasitizes sun- genetic variation were in general extremely low within each
flower [2]. The high host specificity of O. cumana is probably gene pool due probably to a founder effect. However, a
associated with the mode of inheritance of genetic resistance reduced number of populations exhibited larger genetic
in sunflower. Whereas in most host crops genetic resistance to diversity, which was attributed to the presence of individuals
Orobanche spp. is horizontal, that is, polygenic and nonrace from both gene pools and the occurrence of crosses between
specific, resistance to O. cumana in sunflower is primarily them. Even though O. cumana is considered to be primarily a
vertical, that is, monogenic, dominant, and race specific [16]. self-pollinated species [29], the occurrence of a certain rate of
The development of sunflower resistant cultivars has been cross pollination has been experimentally demonstrated [30].
paralleled by the appearance of O. cumana populations that There is no information on the population structure of
overcame sunflower genetic resistance, a recurrent process O. cumana populations parasitizing wild species and their
that has continued until today [11]. Several physiological races genetic relationship with weedy populations in areas where
of O. cumana have been reported. Vrânceanu et al. [17] iden- they coexist. There is also no information on their virulence
tified races A through E using five sunflower differential lines on sunflower. The objective of this research was to investigate
carrying the dominant resistance genes Or1 through Or5, the genetic diversity, population structure, and ability to
respectively. More virulent races named as F, G, and H were parasitize sunflower of O. cumana populations growing on
later detected in the main sunflower cultivation areas of the wild plants in the Black Sea coast of Bulgaria, as well as their
Old World [3]. In Bulgaria, races D and E were predominant relationship with weedy populations parasitizing sunflower.
till few years ago [18], but a more virulent race G has become
increasingly important in recent years [19].
There are few studies on genetic interactions between wild 2. Materials and Methods
and weedy forms of parasitic plant species. Knowledge about
such interactions is important because wild vegetation may 2.1. Orobanche cumana Populations. Two field expeditions
play a role as reservoir of genetic diversity for overcoming were conducted in July 2006 and June 2012 along the Black
genetic resistance mechanisms in the host crops [20, 21]. But Sea coast of Bulgaria, where the distribution of O. cumana in
on the other hand, evolution of virulence in weedy popu- the wild has been largely documented [8, 31–33], to collect
lations may also have an impact on the distribution of the fresh tissue and mature seeds of O. cumana populations
species in the wild [22]. Botanga et al. [23] used seeds of parasitizing wild Asteraceae species. Six populations were
eight populations of the parasitic weed Striga asiatica (L.) located in both expeditions, one of them in both years
Kuntz collected on wild hosts to conduct infestation exper- (Table 1, Figure 1). Samples from the latter population were
iments on susceptible maize and sorghum cultivars. None of managed separately in the study to evaluate potential changes
the populations parasitized on sorghum, whereas five out of between both collection dates. Voucher specimens of the
the eight populations failed to parasitize on maize. The populations are housed in the herbarium of the University
authors concluded the occurrence of local adaptation of the of Córdoba, Spain (herbarium code COA). Duplicated
parasite to a host species as well as a high degree of host spe- specimens can also be found at the herbarium SOA
cialization. Similarly, Botanga and Timko [24, 25] reported (Agricultural University of Plovdiv, Bulgaria). Populations
the stratification by host preference of Striga gesnerioides CUMBUL-1 (COA-45783 and COA-45784), CUMBUL-2
(Willd.) Vatke genotypes parasitizing cowpea [Vigna unguic- (COA-45789), CUMBUL-3 (COA-45790), CUMBUL-4
ulata (L.) Walp.] and the wild legume Indigofera hirsuta L. (COA-45785), CUMBUL-6 (COA-53262 and COA-54519),
Conversely, Olivier et al. [26], using isozyme loci, showed and CUMBUL-7 (COA-54510) were collected on A. maritima
little genetic differentiation based on host specificity among L. (Table 1). Figure 2 shows details of population CUMBUL-1.
Striga hermonthica (Del.) Benth. populations parasitizing Population CUMBUL-5 was found parasitizing Anthemis
sorghum, pearl millet, maize, and wild grasses. Similarly, Vaz arvensis L., Chamaemelum nobile (L.) All., and another
Patto et al. [20] found low genetic differentiation between species of the Asteraceae that could not be identified, though
 The Scientific World Journal

Table 1: Host species, collecting site, characteristics, and number of individuals analyzed for the studied Orobanche cumana populations.
Population Host species Collecting site Region Latitude, Longitude, Altitude Year 𝑛
O. cumana populations collected on wild hosts
CUMBUL-1 Artemisia maritima Bulgaria, Burgas, Atanasovsko Lake South-Eastern Bulgaria 42∘ 33󸀠 02.7󸀠󸀠 N; 27∘ 29󸀠 24󸀠󸀠 E; 14 m 2006 16
CUMBUL-2 Artemisia maritima Bulgaria, Burgas, Pomorie-Aheloj South-Eastern Bulgaria 42∘ 37󸀠 02.8󸀠󸀠 N; 27∘ 37󸀠 31.1󸀠󸀠 E; 17 m 2006 30
CUMBUL-3 Artemisia maritima Bulgaria, Kranevo North-Eastern Bulgaria 43∘ 20󸀠 05.6󸀠󸀠 N; 28∘ 3󸀠 41.9󸀠󸀠 E; 112 m 2006 6
CUMBUL-4 Artemisia maritima Bulgaria, Balchik North-Eastern Bulgaria 43∘ 24󸀠 36.9󸀠󸀠 N; 28∘ 9󸀠 23.5󸀠󸀠 E; 21 m 2006 29
CUMBUL-5 1 Anthemis arvensis Bulgaria, Kavarna, Gorun-Tyulenovo North-Eastern Bulgaria 43∘ 29󸀠 12.6󸀠󸀠 N; 28∘ 31󸀠 13.3󸀠󸀠 E; 44 m 2006 28
CUMBUL-5 2 Chamaemelum nobile Bulgaria, Kavarna, Gorun-Tyulenovo North-Eastern Bulgaria 43∘ 29󸀠 12.6󸀠󸀠 N; 28∘ 31󸀠 13.3󸀠󸀠 E; 44 m 2006 20
CUMBUL-6 Artemisia maritima Bulgaria, Burgas, Poda Protected Area South-Eastern Bulgaria 42∘ 26󸀠 35.91󸀠󸀠 N; 27∘ 27󸀠 58.64󸀠󸀠 E; 7 m 2012 23
CUMBUL-7 Artemisia maritima Bulgaria, Burgas, Atanasovsko Lake South-Eastern Bulgaria 42∘ 33󸀠 05.88󸀠󸀠 N; 27∘ 29󸀠 22.91󸀠󸀠 E; 8 m 2012 14
O. cumana populations collected on sunflower
CUMBUL-8 Helianthus annuus Bulgaria, Sadovo Central Bulgaria 42∘ 07󸀠 13.49󸀠󸀠 N; 24∘ 54󸀠 53.40󸀠󸀠 E; 156 m 2012 20
CUMBUL-9 Helianthus annuus Bulgaria, Plodiv Central Bulgaria 42∘ 03󸀠 35.43󸀠󸀠 N; 24∘ 49󸀠 26.28󸀠󸀠 E; 189 m 2012 20
CUMBUL-10 Helianthus annuus Bulgaria, Balgarevo North-Eastern Bulgaria 43∘ 24󸀠 58.14󸀠󸀠 N; 28∘ 26󸀠 43.83󸀠󸀠 E; 81 m 2012 18
IASCum-2 Helianthus annuus Spain, Sevilla, Écija Southern Spain 37∘ 34󸀠 24󸀠󸀠 N; 5∘ 8󸀠 45󸀠󸀠 W; 181 m 2008 12
IASCum-3 Helianthus annuus Spain, Sevilla, Osuna Southern Spain 37∘ 15󸀠 19󸀠󸀠 N; 5∘ 3󸀠 49󸀠󸀠 W; 304 m 2008 12
IASCum-4 Helianthus annuus Spain, Cuenca, Montalbo Central Spain 39∘ 51󸀠 03󸀠󸀠 N; 02∘ 39󸀠 54󸀠󸀠 W; 838 m 2008 12
𝑛, final studied sample size [including a number of plants excluded from the analysis because of lack of amplification, belonging to each of the populations CUMBUL-1 (four plants excluded), CUMBUL-4 (one
plant), CUMBUL-5 1 (two plants), CUMBUL-5 2 (two plants), CUMBUL-6 (three plants), and CUMBUL-10 (two plants)].
3
4 The Scientific World Journal

CUMBU
L
CUMBU -5 1
L-5 2
C U
CUMBU MBUL-10
NE L-4
CUMBU
L-3

SE
C

CUMBU
L-2
CUMBU
L-7
CUMBU
L-1
CUMBU
L-6

CUMBU
CUMBU L-8
L-9

Wild hosts

Sunflower

Figure 1: Geographical distribution of Orobanche cumana Bulgarian populations collected on wild and cultivated hosts (left side of the figure)
and map of mean membership probabilities per population as obtained using Bayesian clustering analysis resulting from STRUCTURE at
𝐾 = 2 (right side of the figure). Pie size is proportional to the size of each population.

Figure 2: Details of population CUMBUL-1 of Orobanche cumana parasitizing Artemisia maritima in Burgas, Bulgaria.

in the latter case only two plants were presen and they were areas of Spain in which contrasting gene pools have been
not collected. Plants collected on A. arvensis (CUMBUL-5 1; identified [28], to be used as a control (Table 1).
COA-45791) and C. nobile (CUMBUL-5 2; COA-45792) Mature seeds were collected in bulk from 5 to 30 mature
were analyzed separately to evaluate potential differences plants of populations CUMBUL-1, CUMBUL-2, CUMBUL-4,
associated with the host plant. The populations were located and CUMBUL-5 1. No mature plants were available at the
at a distance of less than 3 km from agricultural fields. time of the collection expeditions for the other populations,
Fresh tissue (young stalks) from 6 to 30 individual plants including the O. cumana populations parasitizing sunflower
(Table 1), depending on population size, was collected in in Bulgaria. Alternatively, seeds from three populations of O.
situ for each population and kept under drying conditions cumana collected in sunflower fields in Bulgaria (OC-9, OC-
in ziplock bags with silica gel for subsequent freezing 11, and OC-13) were used for virulence studies. Populations
at −80∘ C. Fresh tissue of three O. cumana populations OC-9 and OC-13 were kindly provided by Professor Rossitza
parasitizing sunflower crops in two different areas of Batchvarova, AgroBioInstitute, Sofia, Bulgaria. Population
Bulgaria (Table 1) was collected in the 2012 expedition. OC-11 was collected by one of the authors (K. Stoyanov).
Additionally, fresh tissue was also collected in situ from Spanish race F population OC-88 was also used as a control
three populations parasitizing sunflower in two different for virulence studies.
The Scientific World Journal 5

2.2. DNA Extraction and SSR Analysis. Frozen tissue was with other frequency-based estimators of population struc-
lyophilized and ground to a fine powder. DNA was extracted ture for codominant data such as Nei’s 𝐺ST , Nei’s standardized
from individual O. cumana plants using a modified version of 𝐺ST , Hedrick’s standardized 𝐺ST , Hedrick’s further standard-
the protocol described in Pérez-Vich et al. [34]. Microsatellite ized 𝐺ST for small number of populations, and Jost’s estimate
analyses were carried out as described in Pineda-Martos et of differentiation, following calculations detailed in [39]. The
al. [28], using the same set of fifteen high-quality, polymor- pairwise relationship between the genetic distance matrices
phic SSR primer pairs (Table S1 in Supplementary Mate- was tested through a Mantel’s test with 999 permutations.
Since the different statistical measures were highly correlated
rial available online at http://dx.doi.org/10.1155/2014/150432).
(𝑟 > 0.94, 𝑃 = 0.001 for all comparisons, excepting those
Amplification products were resolved by electrophoresis on
including the Jost’s estimate of differentiation in which 𝑟 >
3% Metaphor agarose (BMA, Rockland, ME, USA gels in 1x 0.90, 𝑃 = 0.001), only the results based on the genetic distance
TBE buffer at 100 V constant voltage, with SaveView Nucleic coefficient 𝐺ST with the corrections of Nei and Chesser [40]
Acid Stain (NBS Biologicals Ltd., Huntingdon, UK) incor- and Nei [41] are presented. To assess genetic relationships
porated in the gels and visualized under UV light. A 100 bp among populations, the matrix of 𝐺ST pairwise distances was
DNA ladder (Solis BioDyne, Tartu, Estonia) was used as a used as input for a principal coordinates analysis (PCoA)
standard molecular weight marker to get an approximate size using GenAlEx ver. 6.5. PCoA has the main advantage of not
of DNA fragments. Bands were scored manually with the aid requiring strong assumptions about the underlying genetic
of Quantity One 1-D Analysis Software (Bio-Rad Laboratories model [42].
Inc., Hercules, CA, USA) at least twice independently for each To identify genetically homogeneous groups (gene pools),
population. Bayesian model-based clustering algorithms implemented in
the software package STRUCTURE ver. 2.3.4 [43] were
applied. Cluster grouping in STRUCTURE is based on itera-
2.3. Molecular Data Analysis tive analysis using 𝐾 number of groups previously defined by
the user, with individuals in the sample being assigned prob-
2.3.1. Genetic Diversity Analysis. For each SSR locus, the abilistically to one or several groups. The admixture model
number of alleles (Na), observed and expected heterozygosity and the allele frequencies correlated model were used [44].
(Ho and He), and 𝐹ST were calculated using GenAlEx ver. 6.5 No prior information was used to define the clusters. For each
[35]. Additionally, each locus was tested for departure from value of 𝐾 (from 1 to 14), 10 independent runs were made that
Hardy-Weinberg equilibrium (HWE) and linkage disequi- were used to estimate the probability of the data Pr(𝑋 | 𝐾).
librium within each of the populations with Arlequin ver. For each run, 1,000,000 Monte Carlo Markov chain (MCMC)
3.5.1.3 [36]. To characterize the genetic diversity of O. cumana iterations were carried out after a burn-in period of 200,000
populations collected on wild hosts and the control popula- steps. To detect the number of genetically homogeneous
tions collected on sunflower, the percentage of polymorphic groups (𝐾) that best fits the data, the STRUCTURE HAR-
loci (𝑃), the average observed number of alleles (Na), the VESTER website [45], which implements the Evanno method
number of different alleles with a frequency ≥ 5% (Na ≥ 5%), [46], was used. The 10 runs from the most probable number
the number of effective alleles (Ne), the number of private of 𝐾 groups were averaged applying the FullSearch algorithm
alleles unique to a single population (Npa), the observed and provided in the CLUMPP ver. 1.1.2b software [47] and the
expected heterozygosity (Ho and He), the Shannon’s diversity output was entered into Distruct ver. 1.1 for display [48]. To
index (I), and the fixation index (𝐹is ) were calculated for explore the genetic structure further, the STRUCTURE anal-
all loci at each population. All calculations were carried out yses were also carried out only with the 11 Bulgarian popula-
using GenAlEx ver. 6.5. 𝐹is was used to estimate the selfing tions, as described above. We also used the program InStruct
rate (𝑆) from 𝑆 = 2𝐹is /(1 + 𝐹is ) [37]. As additional measures [49] for analyzing population structure, since this program
of intrapopulation diversity, the mean number of pairwise is an extension of STRUCTURE that does not assume
differences between individuals within each population, esti- Hardy-Weinberg equilibrium and can incorporate selfing in
mated as the mean number of differences between all pairs of the model. In addition, it can estimate the level of selfing in
SSR haplotypes in each population, and the genotypic rich- each population. Five independent chains were run for each
ness (𝑅), defined as (𝐺 − 1)/(𝑁 − 1), where 𝐺 is the number of 𝐾. Each chain was run for 1,000,000 iteration steps, with a
burn-in of 500,000, and a thinning of 10. Graphical rep-
MLGs (the observed number of multilocus genotypes) and
resentations of population assignments from InStruct were
𝑁 is the number of samples per population, were determined
produced from the program Distruct ver. 1.1 [48].
using Arlequin ver. 3.5.1.3, and GenClone 2.0 [38], respec-
Finally, an analysis of molecular variance (AMOVA) [50]
tively. within populations, among populations, and among popula-
tion groups (based on a priori grouping variables such as wild
2.3.2. Genetic Differentiation Analysis. To evaluate genetic or cultivated host or based on the gene pools determined with
differentiation between populations, initial frequency-based clustering methods) was carried out to determine the distri-
analysis was carried out by calculating pairwise genetic dis- bution of variation at different hierarchical levels. The vari-
tances between populations using the genetic distance coef- ance components were tested statistically by nonparametric
ficient 𝐺ST as implemented in GenAlEx ver. 6.5 using 1000 randomization tests using 1000 permutations. Fixation
random permutations to assess significance. Pairwise dis- indices (𝐹-statistics) were also estimated by AMOVA. All cal-
tance matrices were also calculated using GenAlEx ver. 6.5 culations were carried out with Arlequin ver. 3.5.1.3.
6 The Scientific World Journal

2.4. Parasitization Ability and Virulence on Sunflower. Mature 3. Results

seeds were collected for wild O. cumana populations
3.1. Genetic Diversity and Population Structure. All SSR
CUMBUL-1, CUMBUL-2, CUMBUL-4, and CUMBUL-5 1.
markers were polymorphic (Table S1). The total number of
However, the amount of available seed was very low, which
alleles scored was 38, ranging from 2 to 4 for each SSR locus.
restricted the number of sunflower genotypes for virulence Allelic diversity was generally low for all fifteen SSR loci when
studies as well as the number of plants per genotype. considering the whole set of 260 individual O. cumana plants
Accordingly, their parasitization ability and virulence on sun- (Table S1). All the loci exhibited an important heterozygote
flower was evaluated in two separated experiments. The first deficiency (Table S1). A significant deviation (𝑃 < 0.05) from
experiment was aimed at determining whether the popula- Hardy-Weinberg equilibrium was found for almost all loci
tions had the ability to parasitize sunflower genotypes with when all samples were considered. Linkage disequilibrium
no genetic resistance to weedy O. cumana physiological races. was significant (𝑃 < 0.05) in 238 out of 430 paired loci
Two confectionery sunflower landraces, B117 and B206, with comparisons when considering all the samples. It has been
no known resistance to any O. cumana race were used. Both established that linkage disequilibrium is predicted to
landraces were collected by L. Velasco in isolated vegetable approach zero for an ideal population, in the absence of forces
patches in Valdepeñas (Jaén Province, Spain) and Quin- such as genetic drift, population mixing, mutation, natural
tana de la Serena (Badajoz Province, Spain), respectively. selection, or inbreeding [51]. High linkage disequilibrium
Orobanche cumana population OC-88 [with known virulence observed suggested the existence of some genetic structure,
(race F)] was used as a positive control. In a second exper- apart from other factors determining the organization of
iment, the virulence of the populations was tested on a set genetic variation in the studied populations, as it will be
of sunflower lines with varying levels of genetic resistance to further discussed below.
O. cumana physiological races. Jdanovski 8281 (J8281) is a Genetic diversity within each population, measured by
line incorporating resistance gene Or2 that confers resistance the mean number of observed and effective alleles, the
to O. cumana race B [17]. AC03-1589 is a line incorporating expected heterozygosity, and Shannon’s diversity indexes,
resistance gene Or3 that confers resistance to O. cumana was in general low, and only one population (CUMBUL-4)
race C, kindly provided by Dr. Maria Păcureanu, National contained a substantial number of private alleles (Table 2).
Agricultural Research and Development Institute, Fundulea, As expected from previous studies, Spanish populations
Romania. S1358 is a line incorporating resistance gene Or4 were characterized by extremely low level of intrapopulation
that confers resistance to O. cumana race D [17]. P-1380 is genetic diversity due probably to a founder effect [28],
a line containing the resistance gene Or5, which determines with no polymorphic loci being detected in two out of the
resistance to O. cumana race E [17]. P96 is a line with three populations (Table 2). In contrast, populations from
recessive resistance to O. cumana race F [34]. B117 with no Bulgaria exhibited higher diversity values, with the exception
known resistance to any O. cumana race was used as positive of population CUMBUL-3, which showed no polymorphic
control. Populations OC-9, OC-11, OC-13, and OC-88 were loci. However, it is important to note that this was the smallest
used as controls. Because the amount of seed of popula- population, in which only six individual plants could be col-
tions CUMBUL-1, CUMBUL-4, and CUMBUL-5 1 was not lected. Amongst the other Bulgarian populations, the highest
enough for evaluating them on all sunflower lines of the genetic diversity corresponded to the populations collected
second experiment, it was decided not to test them on line on wild hosts CUMBUL-2, CUMBUL-5 2, and CUMBUL-7,
S1358. which showed He, I and pairwise difference (between indi-
Seeds of O. cumana populations were used to inoculate viduals) values over 0.25, 0.4, and 3.5, respectively (Table 2).
small pots 7 × 7 × 8 cm filled with a mixture of sand and peat The lowest genetic diversity corresponded to populations
(1 : 1 by vol). Twenty-five mg of O. cumana seeds per pot was CUMBUL-8 and CUMBUL-9, collected on sunflower, which
used. The soil mixture containing O. cumana seeds was showed He, I and pairwise difference (between individu-
carefully mixed to obtain a homogeneously infested substrate. als) values below 0.05, 0.1, and 0.5 respectively (Table 2).
Seeds of sunflower cultivars were germinated on moistened The other six populations, excluding CUMBUL-3, showed
filter paper in Petri dishes and two-day-old seedlings were intermediate diversity values, ranging from 0.10 to 0.23 for
planted in the pots inoculated with O. cumana seeds. Eight He, from 0.18 to 0.35 for I, and from 1.8 to 3.1 for pairwise
pots (replications) per combination of sunflower cultivar and differences between individuals. The fixation index (𝐹is ) and
O. cumana population were used. The plants were maintained selfing rate (𝑆) values were high for the populations studied
in a growth chamber for 21 days at 25∘ C/20∘ C (day/night) (Table 2).
with a 16 h photoperiod for incubation. After this time, the For measuring differentiation between populations, pair-
plants were transplanted to pots containing 3 L of an unin- wise 𝐺ST values were computed (Table S2). No significant or
fested sand-silt-peat (2 : 1 : 1 by vol) soil mixture and main- very low (𝐺ST ≤ 0.01) differentiation was found for popula-
tained under open air conditions. The plants were watered tions CUMBUL-8 and CUMBUL-9, collected on sunflower at
as needed and were not fertilized. The number of O. cumana close locations, CUMBUL-2 and CUMBUL-7, collected on A.
shoots per sunflower plant was counted at sunflower matu- maritima at near sites the same year, and CUMBUL-5 1 and
rity. Differences between mean numbers of O. cumana shoots CUMBUL-5 2, collected at the same location but on dif-
per plant for each O. cumana population and sunflower ferent wild hosts. Populations CUMBUL-1 and CUMBUL-7,
cultivar were analyzed through one way ANOVA and Tukey’s which were collected at the same site but with a six-year
range test using IBM SPSS Statistics version 19. difference, showed slightly higher 𝐺ST values (0.107). The
 The Scientific World Journal

Table 2: Genetic diversity parameters of Orobanche cumana populations from Bulgaria collected on wild hosts and on sunflower (prefix CUMBUL-) and from Spain collected on sunflower
(prefix IASCum-).
Genotypic richness
Population 𝑃 Na (±SE) Na ≥ 5% (±SE) Ne (±SE) Npa (±SE) Ho (±SE) He (±SE) 𝐼 (±SE) Pairwise differences 𝐹is (±SE) 𝑆
𝐺 𝑅
O. cumana populations collected on wild hosts
CUMBUL-1 66.7 1.733 (0.15) 1.667 (0.13) 1.391 (0.10) 0.000 (0.00) 0.004 (0.01) 0.229 (0.05) 0.349 (0.08) 3.111 (1.66) 7 0.40 0.984 (0.01) 0.992
CUMBUL-2 86.9 2.000 (0.14) 1.933 (0.15) 1.521 (0.10) 0.000 (0.00) 0.000 (0.00) 0.297 (0.05) 0.458 (0.07) 4.393 (2.20) 8 0.24 1.000 (0.00) 1.000
CUMBUL-3 0.0 1.000 (0.00) 1.000 (0.00) 1.000 (0.00) 0.000 (0.00) 0.000 (0.00) 0.000 (0.00) 0.000 (0.00) 0.000 (0.00) 1 0.00 — —
CUMBUL-4 53.3 1.600 (0.16) 1.333 (0.13) 1.145 (0.05) 0.067 (0.07) 0.000 (0.00) 0.105 (0.03) 0.184 (0.06) 1.595 (0.96) 6 0.18 1.000 (0.00) 1.000
CUMBUL-5 1 80.0 2.000 (0.17) 1.867 (0.16) 1.250 (0.07) 0.000 (0.00) 0.032 (0.01) 0.171 (0.04) 0.306 (0.06) 2.449 (1.34) 11 0.37 0.776 (0.04) 0.874
CUMBUL-5 2 80.0 1.933 (0.15) 1.933 (0.15) 1.418 (0.10) 0.000 (0.00) 0.014 (0.01) 0.248 (0.05) 0.400 (0.07) 3.597 (1.86) 7 0.32 0.920 (0.04) 0.958
CUMBUL-6 73.3 1.800 (0.14) 1.667 (0.16) 1.294 (0.10) 0.000 (0.00) 0.003 (0.01) 0.181 (0.05) 0.300 (0.07) 2.629 (1.43) 10 0.41 0.968 (0.03) 0.984
CUMBUL-7 73.3 1.800 (0.14) 1.800 (0.14) 1.467 (0.12) 0.000 (0.00) 0.010 (0.01) 0.258 (0.05) 0.398 (0.08) 3.947 (2.04) 7 0.46 0.975 (0.01) 0.987
Mean 64.2 1.733 (0.06) 1.650 (0.12) 1.311 (0.03) 0.008 (0.01) 0.008 (0.002) 0.186 (0.02) 0.299 (0.03) 2.715 (0.05) 7.1 0.30 0.946 0.971
O. cumana populations collected on sunflower
CUMBUL-8 40.0 1.400 (0.13) 1.133 (0.09) 1.039 (0.02) 0.000 (0.00) 0.021 (0.01) 0.034 (0.01) 0.071 (0.03) 0.446 (0.41) 3 0.11 0.194 (0.10) 0.325
CUMBUL-9 13.3 1.133 (0.09) 1.133 (0.09) 1.014 (0.01) 0.000 (0.00) 0.000 (0.00) 0.013 (0.01) 0.026 (0.02) 0.195 (0.25) 2 0.05 1.000 (0.00) 1.000
CUMBUL-10 46.7 1.467 (0.13) 1.467 (0.13) 1.175 (0.06) 0.000 (0.00) 0.015 (0.01) 0.123 (0.04) 0.201 (0.06) 1.825 (1.07) 8 0.41 0.915 (0.04) 0.956
Mean-Bulgaria 33.3 1.333 (0.10) 1.244 (0.11) 1.076 (0.05) 0.000 (0.00) 0.012 (0.01) 0.057 (0.03) 0.099 (0.05) 0.822 (0.51) 4.3 0.19 0.703 0.760
IASCum-2 0.0 1.000 (0.00) 1.000 (0.00) 1.000 (0.00) 0.000 (0.00) 0.000 (0.00) 0.000 (0.00) 0.000 (0.00) 0.000 (0.00) 1 0.00 — —
IASCum-3 0.0 1.000 (0.00) 1.000 (0.00) 1.000 (0.00) 0.000 (0.00) 0.000 (0.00) 0.000 (0.00) 0.000 (0.00) 0.000 (0.00) 1 0.00 — —
IASCum-4 6.7 1.067 (0.07) 1.067 (0.07) 1.026 (0.03) 0.000 (0.00) 0.000 (0.00) 0.019 (0.02) 0.030 (0.03) 0.290 (0.32) 2 0.09 1.000 (0.02) 1.000
Mean-Spain 2.2 1.022 (0.02) 1.022 (0.02) 1.009 (0.01) 0.000 (0.00) 0.000 (0.00) 0.006 (0.01) 0.010 (0.01) 0.097 (0.10) 0.7 0.03
𝑃: percentage of polymorphic loci, Na: average observed allele number, Na > 5%: number of different alleles with a frequency ≥ 5%, Ne: number of effective alleles, Npa: number of private alleles unique to a single
population, Ho: observed heterozygosity, He: expected heterozygosity, 𝐼: Shannon’s diversity index, Pairwise differences: mean number of pairwise differences between individuals within each population (±SD),
𝐺: number of distinct multilocus genotypes (MLGs), 𝑅: genotypic richness; 𝐹is : fixation index, and 𝑆: selfing rate.
7
8 The Scientific World Journal

Principal coordinates (1 versus 2) Table 3: Proportion of membership of each Bulgarian Orobanche

98 IASCum-2 cumana population in inferred STRUCTURE groups for 𝐾 = 2.
IASCum-3 Populations collected on wild hosts are highlighted in bold.
6 51
52
PCo 2 (26.6%)

7 Population Genetic group 1 Genetic group 2

1 2
10 IASCum-4 CUMBUL-1 0.491 0.509
4
CUMBUL-2 0.677 0.323
CUMBUL-3 0.984 0.016
3 CUMBUL-4 0.969 0.031
CUMBUL-5 1 0.138 0.862
PCo 1 (39.4%) CUMBUL-5 2 0.239 0.76
CUMBUL-6 0.104 0.896
Principal coordinates (1 versus 3)
CUMBUL-7 0.541 0.459
3 CUMBUL-8 0.023 0.977
98 4
5 15 2 2
IASCum-2 CUMBUL-9 0.022 0.978
IASCum-3
PCo 3 (17.8%)

6 7 10 CUMBUL-10 0.647 0.352

1

IASCum-4 assignments that recurred at all monitored levels of 𝐾 (Figure

S2).
PCo 1 (39.4%)
A more detailed analysis of population structure includ-
Figure 3: Principal coordinates analysis of pairwise genetic dis- ing only Bulgarian populations was carried out. STRUC-
tances among 14 Orobanche cumana populations and subpopu- TURE analyses indicated the existence of two (𝐾 = 2;
lations (260 individuals). Primary groups identified with either Figure S3) major genetic groups, mainly represented by
the 1st versus 2nd axis plot or with the 1st versus 3rd axis plot populations CUMBUL-3 and CUMBUL-4 on one hand
are highlighted with solid boxes. Populations from Spain (prefix (Gene Pool 1), and CUMBUL-8 and CUMBUL-9 on the other
IASCum) are named with their complete name, and populations hand (Gene Pool 2) (Table 3; Figure 4(a)). The remaining
from Bulgaria (prefix CUMBUL) are named with their number, seven populations were categorized in-between these two
without prefix. groups, although the average proportion of membership
was shifted towards Gene Pool 1 for populations CUMBUL-
2 and CUMBUL-10, whereas populations CUMBUL-5 1,
highest differentiation values (𝐺ST > 0.8) were found between CUMBUL-5 2, and CUMBUL-6 were clearly shifted towards
the following three groups of populations: (i) IASCum-2, Gene Pool 2 (Table 3; Figure 1). When the membership value
IASCum-3, and IASCum-4 collected on sunflower in Spain, of each individual for each population was analyzed in detail,
(ii) CUMBUL-8 and CUMBUL-9 collected on sunflower in it was shown that an important number of individuals from
Bulgaria, and (iii) CUMBUL-3 and CUMBUL-4 collected on populations CUMBUL-5 1, CUMBUL-5 2, and CUMBUL-6
wild A. maritima in Bulgaria (Table S2). In principal coor- [19 individuals out of 28 (67.9%), 10 out of 20 (50%), and 15 out
dinate analyses, the first three axes explained 39.4%, 26.6%, of 23 (65.2%), resp.] showed a high (>0.90) membership value
and 17.8%, respectively of the variation, producing five differ- for Gene Pool 2 (Figure 4(b)). Classifications of individuals
entiated groups of populations: (i) IASCum-2 and IASCum-3, at 𝐾 = 2 by the algorithms of STRUCTURE and InStruct
(ii) IASCum-4, (iii) CUMBUL-8 and CUMBUL-9, (iv) were very similar qualitatively (Figure 4(a)). Within-cluster
CUMBUL-3 and CUMBUL-4, and (v) the remaining seven selfing rates estimated from InStruct analyses were very high
populations, six of them collected on wild hosts in Bulgaria (on average, 0.947 for Gene Pool 1 and 0.951 for Gene Pool 2).
and one of them collected on sunflower in Bulgaria (Figure 3). Different AMOVA analyses were carried out within the
Bayesian-based analysis of the structure of the whole set O. cumana populations collected in Bulgaria. First, AMOVA
of populations including those from Spain and Bulgaria with analyses were conducted on populations collected on wild
STRUCTURE revealed a close relationship among popula- hosts. When no population structure was considered, 53.6%
tions whatever their geographical origin, with an optimal 𝐾 of the genetic variance was attributable to differences among
value of 2 (Figures S1 and S2). Secondary peaks were populations, while the remaining 46.4% was due to differ-
observed at 𝐾 = 4 and 7 (Figure S1), and the standard devia- ences within populations (Table 4). When populations were
tion of Pr(𝑋 | 𝐾) began to increase substantially at 𝐾 values structured according to clustering results, differences among
higher than these (Figure S1). Visualization of the cluster groups accounted for 50.4% of the total variance, while
membership for 𝐾 = 2 to 𝐾 = 7 showed a general trend differences among populations of each group only accounted
towards classification of populations IASCum-2, IASCum-3, for 17.6% (Table 4). When populations collected on sunflower
IASCum-4, CUMBUL-3, CUMBUL-4, CUMBUL-8, and were added to the model, no significant structuring according
CUMBUL-9 within uniform pools, while the rest of the to the ecological status of the populations was detected
populations were included within mixed pools (Figure S2), (Table 4). Structured analysis based on clustering groups
The Scientific World Journal 9

STRUCTURE

InStruct

CUMBUL-5 1

CUMBUL-5 2

CUMBUL-6

CUMBUL-9
CUMBUL-3

CUMBUL-8
CUMBUL-2

CUMBUL-4

CUMBUL-7

CUMBUL-10
CUMBUL-1

(a)
6
3 CUMBUL-1
0
16
12
8 CUMBUL-2
4
0
8
4 CUMBUL-3
0
30
20 CUMBUL-4
10
0
20
10 CUMBUL-5 1
0
10
Number of individuals

5 CUMBUL-5 2
0
20
15
10 CUMBUL-6
5
0
4
CUMBUL-7
2
0
20
10 CUMBUL-8
0
20
15
10 CUMBUL-9
5
0
15
10 CUMBUL-10
5
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Number of individuals within intervals of membership for Gene Pool 2
(b)

Figure 4: Results from STRUCTURE and InStruct analyses: (a) population structure obtained from STRUCTURE and InStruct analyses of
eleven Bulgarian Orobanche cumana populations, with each individual being represented by a single vertical bar divided into two shades.
Each shade represents one gene pool (𝐾) and the length of the shaded segment shows the individual’s estimated proportion of membership
in that cluster and (b) number of O. cumana individuals from each Bulgarian population within intervals of membership for Gene Pool 2 in
the STRUCTURE analyses.
10 The Scientific World Journal

Table 4: Analysis of molecular variance (AMOVA) of Orobanche cumana populations from Bulgaria.

AMOVA statistics
Hierarchical structure and source of variation 𝐹-statisticsa 𝑃 value
df Sum of squares Variance components % Variance
Bulgarian populations collected on wild hosts (8 populations; 166 individuals)
Not structured
Among populations 7 491.05 1.69 53.64 𝐹ST = 0.54 <0.001
Within populations/group 324 473.19 1.46 46.36
Structured based on gene poolsb
Among groups 1 294.36 2.29 50.37 𝐹CT = 0.50 0.032
Among populations/group 6 196.69 0.80 17.56 𝐹SC = 0.35 <0.001
Within populations/group 324 473.19 1.46 32.07 𝐹ST = 0.68 <0.001
Total of Bulgarian populations (wild and cultivated host) (11 populations; 224 individuals)
Not structured
Among populations 10 713.97 1.74 59.54 𝐹ST = 0.60 <0.001
Within populations/group 437 517.64 1.18 40.46
Structured based on ecological statusc
Among groups 1 93.81 0.14 4.55 𝐹CT = 0.05 0.234
Among populations/group 9 620.16 1.68 56.05 𝐹SC = 0.59 <0.001
Within populations/group 437 517.64 1.18 39.40 𝐹ST = 0.61 <0.001
d
Structured based on gene pools
Among groups 2 423.05 1.51 42.01 𝐹CT = 0.42 0.002
Among populations/group 8 290.91 0.90 25.05 𝐹SC = 0.43 <0.001
Within populations/group 437 517.64 1.18 32.94 𝐹ST = 0.67 <0.001
a
𝐹-statistics represents differentiation among groups (𝐹CT ), among populations within groups (𝐹SC ), and among populations within the whole population (𝐹ST ).
b
The gene pools defined with clustering analyses comprised (i) populations CUMBUL-3 and -4 and (ii) populations CUMBUL-1, -2, -5 1, -5 2, -6, and -7.
c
The structured groups based on the ecological status were (i) wild hosts (populations CUMBUL-1, -2, -3, -4, -5 1, -5 2, -6, and -7) and (ii) cultivated host
(sunflower) (populations CUMBUL-8, -9, -10).
d
The gene pools defined with clustering analyses were (i) populations CUMBUL-3, -4, (ii) populations CUMBUL-8 and -9, and (iii) populations CUMBUL-1,
-2, -5 1, -5 2, -6, -7, and -10.

produced similar results to the analysis of populations col- Table 5: Number of emerged Orobanche cumana shoots per
lected on wild hosts alone; that is, variation among groups sunflower plant (mean ± standard deviation) in the evaluation of
accounted for 42.0% of total variation, while variation among O. cumana populations CUMBUL-1, CUMBUL-2, and CUMBUL-4,
populations at each group accounted for 25.1% (Table 4). collected in Bulgaria on Artemisia maritima, CUMBUL-5 1, col-
lected in Bulgaria on Anthemis arvensis, and control population OC-
88, collected in Spain on cultivated sunflower, on two sunflower lines
3.2. Parasitization Ability and Virulence on Sunflower. A first (B117 and B206) with no genetic resistance to O. cumana, conducted
experiment demonstrated that O. cumana populations in pots in 2007a .
CUMBUL-1, CUMBUL-2, CUMBUL-4, and CUMBUL-5 1,
B117b B206a
collected on wild hosts, had the ability to parasitize sunflower
lines B117 and B206, with no resistance genes, though some CUMBUL-1 14.5 ± 9.9b 35.3 ± 13.1c
differences between populations were observed (Table 5). On CUMBUL-2 39.5 ± 7.7c 36.3 ± 11.3c
B117, populations CUMBUL-1 and CUMBUL-5 1 produced CUMBUL-4 5.2 ± 4.7a 1.7 ± 1.0a
similar number of shoots per sunflower plant to the control CUMBUL-5 1 11.5 ± 8.9ab 14.3 ± 4.5b
population OC-88, while CUMBUL-2 produced around four OC-88 10.3 ± 6.0ab 18.5 ± 7.2b
times more shoots per plant and CUMBUL-4 produced about a
Eight pots per each combination of sunflower cultivar and O. cumana
half of shoots per plant than the control. On B206, both population.
b
CUMBUL-1 and CUMBUL-2 yielded more shoots per plant Means with different letters for each sunflower cultivar differ significantly
than the control, while CUMBUL-4 produced less shoots per (𝑃 < 0.05).
plant than the control (Table 5). In a second experiment the
virulence of the populations collected on wild hosts, together
with Bulgarian populations collected on sunflower, was collected on wild hosts and those collected on sunflower
evaluated on sunflower lines with varying degrees of genetic (Table 6). The results were similar on sunflower line AC03-
resistance. On sunflower line J8281, resistant to O. cumana 1589, resistant to race C, except for a significantly higher
race B, the number of shoots per sunflower plant did not number of shoots in population CUMBUL-5 1. Similarly, the
differ significantly between Bulgarian O. cumana populations only wild population evaluated on line S1358 resistant to
The Scientific World Journal 11

Table 6: Number of emerged Orobanche cumana shoots per sunflower plant (mean ± standard deviation) in the evaluation of O. cumana
populations CUMBUL-1, CUMBUL-2, and CUMBUL-4, collected in Bulgaria on Artemisia maritima, CUMBUL-5 1, collected in Bulgaria
on Anthemis arvensis, OC-9, OC-11, and OC-13, collected in Bulgaria on cultivated sunflower, and OC-88, collected in Spain on cultivated
sunflower, on six sunflower lines with different levels of genetic resistance, conducted in pots in 2008a . The O. cumana race to which each
sunflower line is expected to be resistant (if any) is given in parenthesis.

B117b J8281 (B) AC03-1589 (C) S1358 (D) P-1380 (E) P96 (F)
CUMBUL-1 17.7 ± 6.3bc 2.6 ± 2.4ab 1.1 ± 1.2a NEc 0a 0
CUMBUL-2 20.6 ± 3.2c 1.3 ± 1.7ab 0.5 ± 1.1a 2.0 ± 1.4a 0a 0
CUMBUL-4 10.9 ± 7.1ab 0.1 ± 0.4a 0.6 ± 0.7a NE 0a 0
CUMBUL-5 1 12.6 ± 6.7ab 2.0 ± 1.1ab 5.1 ± 2.6b NE 0.4 ± 0.7a 0
OC-9 13.8 ± 7.4abc 3.0 ± 2.5ab 0.5 ± 0.8a 0.8 ± 1.0a 0a 0
OC-11 8.6 ± 5.3a 2.5 ± 2.1ab 0.2 ± 0.4a 0.7 ± 0.8a 0a 0
OC-13 14.3 ± 7.9abc 4.1 ± 1.9b 1.1 ± 1.1a 1.2 ± 0.8a 0a 0
OC-88 9.9 ± 6.9a 16.6 ± 6.5c 1.6 ± 1.3a 1.2 ± 1.2a 6.9 ± 4.3b 0
a
Eight pots per each combination of sunflower cultivar and O. cumana population.
b
Means with different letters for each sunflower cultivar differ significantly (𝑃 < 0.05).
c
NE = not evaluated.

race D (CUMBUL-2) did not differ from the Bulgarian while the rest of the genome is predominantly shaped by other
populations collected on sunflower. When the populations evolutionary sources, namely, recombination and migration
were tested on sunflower line P-1380, resistant to race E, only [29].
population CUMBUL-5 1 produced a few number of shoots Nevertheless, an important observation in this study was
per plant, whereas neither the other populations collected that the genetic structure of wild O. cumana populations
on wild species nor the Bulgarian populations collected on reflected introgressions from weedy populations parasitizing
sunflower did possess the ability to parasitize P-1380. None sunflower. This was shown not only by the analysis of
of the populations parasitized on race-F resistant line P96 population structure, but also by similar levels of virulence on
(Table 6). sunflower of weedy and wild O. cumana populations. To the
best of our knowledge, this is the first study on molecular
4. Discussion diversity and virulence on sunflower of O. cumana pop-
ulations parasitizing wild hosts. Previous studies focusing
The genetic structure of O. cumana populations analyzed in exclusively on weedy populations have shown the existence of
this study was not determined by the fact that the populations several gene pools in this species, with low genetic diversity
were collected on wild or cultivated hosts. This was an within each gene pool [27, 28, 57–59]. Gagne et al. [27]
unexpected result, since, within a number of largely self- identified two gene pools, one of them comprising popula-
pollinated parasitic plant species, host specificity has been tions from eastern Europe (Romania, Bulgaria, and Turkey)
found as a mechanism of accelerating isolation and subse- and another one including populations from southern Spain.
quently genetic divergence among populations, for example, Studies on Spanish populations identified two well-separated
in Orobanche minor Sm. [52–54], Striga asiatica [23], and S. gene pools, one of them in the south (Guadalquivir Valley)
gesnerioides [24, 25]. Conversely, Vaz Patto et al. [20] studied and another one in the central area (Cuenca Province) [28,
the genetic structure of five Moroccan O. foetida populations, 58, 59]. The study of Spanish populations [28] revealed that,
four of them parasitizing wild plants (Scorpiurus muricatus although intrapopulation genetic diversity was in general
L. and Ornithopus sativus Brot.) and another one para- extremely low, some populations showed larger diversity,
sitizing cultivated vetch. The authors found that the vetch- which was hypothesized to be produced by genetic recom-
parasitizing population was closer to the three populations bination between individuals from both gene pools. In the
parasitizing S. muricatus, while the population collected on O. present research, two contrasting gene pools were identified
sativus was the most genetically divergent. This suggested that in Bulgaria, one of them best represented by weedy popula-
parasitization of wild or cultivated hosts was not among the tions from the central area (CUMBUL-8 and CUMBUL-9),
main factors determining genetic differences between these and another one represented by wild populations from the
populations. Since host specificity in Orobanche spp. is mainly eastern coast (CUMBUL-3 and CUMBUL-4), which showed
determined by induction of seed germination by specific in all cases low intrapopulation diversity. The fact that some
chemical stimulants exuded by the host root [55], host- wild populations had higher genetic diversity values and
induced selection is expected to have an impact on very small contained individuals that exhibited membership values very
portions of the genome, probably even at a single locus by close to a weedy gene pool (>0.90) suggested the existence
modifying the binding site of the stimulant receptor [56]. of genetic flow between both gene pools, which could be
Such limited genetic modifications, despite having a huge attributed to cross fertilization and/or seed movement. It is
phenotypic impact, might not be detected with overall important to note that in the Black Sea coast of Bulgaria
genome scans such as the one carried out in this research, weedy and wild O. cumana populations coexist at short
12 The Scientific World Journal

distances. The existence of cross fertilization within this self-pollination, as reported previously for O. cumana pop-
species has been demonstrated in controlled experiments at a ulations parasitizing sunflower [27], and for other predomi-
local scale [30] as well as in the molecular evaluation of field- nantly self-pollinating broomrape species such as Phelipanche
collected weedy populations, where heterozygous individuals ramosa (L.) Pomel [68].
for unique alleles of different gene pools have been iden- Orobanche spp. differ for host specificity. Within this
tified [28]. In relation to gene flow through seed dispersal, genus, O. cumana is one of species with the narrowest range
Orobanche seeds are easily dispersed by water, wind, and of host plants. In the wild, it mainly parasitizes Artemisia spp.
animals. Individual broomrape plants produce an impressive [4], whereas sunflower is the only crop in which O. cumana
number of seeds from 50,000 to 500,000 [1] that maintain occurs as a parasitic weed [2]. Orobanche cumana belongs to
their viability in the soil for up to 20 years [11]. These seeds the native flora of Bulgaria, where it parasitizes wild species
are of near-microscopic size, from 250 to 380 𝜇m long and of the Asteraceae, mainly A. maritima [8]. Conversely, the
from 150 to 240 𝜇m wide, with a weight from 1.0 to 2.5 𝜇g and genus Helianthus is from North American origin [69]. The
are considered as “dust-seeds” [11, 15, 60, 61]. These factors are first report of O. cumana parasitization on sunflower dates
regarded as adaptations for being an obligate parasite, in back to the 1890s in Russia [11] and to 1935 in Bulgaria [12].
order to be dispersed through vegetation so as to be as close It is unknown whether O. cumana possesses natural ability to
as possible to the host plant and increasing the probability of parasitize sunflower or this ability arose in particular geno-
finding an appropriate host [61]. Additionally, at a landscape types following mutation [70]. The possibility that O. cumana
scale, Orobanche cumana seed dispersion is highly influenced possesses natural ability to parasitize sunflower cannot be dis-
by human-derived agricultural and cultivation practices, as carded, since molecules of the same nature to those involved
well as crop-seed trade and the use of contaminated sunflower in O. cumana stimulation of germination by sunflower root
seed stocks [15, 29], which might overpass spatial distances or exudates occur commonly in plant organs of Asteraceae
barriers to gene flow common in natural ecosystems. species [55]. The results of the present research did not shed
Wild and cultivated host plants represent different habi- light on this aspect, since both the population structure
tats for parasitic plants, especially when cultivated plants analysis as well as the virulence study indicated that the wild
carry qualitative resistance genes, as is the case of the sun- populations used in the study contain introgressions from
flower-O. cumana system [16]. The use of cultivars expressing weedy populations. The existence of genetic flow between
vertical resistance mechanisms has contributed to a rapid O. cumana populations parasitizing sunflower and those
development of O. cumana physiological races in most parasitizing wild species opens up an interesting field of
cultivation areas on the Old World, including Bulgaria [3, 19], research on how increasing virulence in weedy populations
which may explain why weedy populations of O. cumana observed in recent years in Bulgaria [19] may influence the
generally show low genetic diversity, since new physiological parasitization ability of O. cumana on wild species and on
races most likely evolve from single mutations events [62]. how genetic variability of wild populations may favour the
This is in general agreement with reports on plant pathogen- ability of weedy populations to overcome sunflower resis-
host interactions [22, 63, 64]. For parasitic plant-host inter- tance mechanisms.
actions, higher intrapopulation variability was reported in a
Striga gesnerioides population parasitizing the wild legume
Indigofera hirsuta L. when compared to populations growing Conflict of Interests
on cultivated cowpea [25]. Another study identified genetic The authors declare that there is no conflict of interests
diversity differences for a population of S. hermonthica grown regarding the publication of this paper.
on rice accessions of varying resistance to Striga, with the
lowest diversity corresponding to a highly resistant rice
accession [65]. This was not exactly the case of the present Acknowledgments
study, since we found genetic diversity values in wild popu-
lations similar or even lower than those reported in weedy The research was partially funded by Fundación Ramón
O. cumana recombinant populations [28]. This could be Areces, Madrid. R. Pineda-Martos was the recipient of a
explained on the basis of the existence of introgression from PhD fellowship from the Spanish National Research Council
weedy populations into wild O. cumana populations. Studies (CSIC) (JAEPre 08 00370).
in nonparasitic plant species, for example, in rice, have shown
that introgression from cultivated species can considerably References
shape genetic diversity of wild populations [66, 67].
The study of O. cumana populations in their natural [1] M. Gevezova, T. Dekalska, K. Stoyanov et al., “Recent advances
habitat provided new data about its breeding system. A clear in broomrapes research,” Journal of BioScience & Biotechnology,
heterozygote deficiency similar to that observed in popula- vol. 1, no. 2, pp. 91–105, 2012.
tions parasitizing sunflower, deviation of genotypic frequen- [2] C. Parker, “The parasitic weeds of the Orobanchaceae,” in Par-
cies at most loci from Hardy-Weinberg equilibrium, high asitic Orobanchaceae-Parasitic Mechanisms and Control Strate-
inbreeding and selfing rate values, and the relatively low levels gies, D. M. Joel, J. Gressel, and L. J. Musselman, Eds., pp. 313–
of genetic variation within populations coupled with sub- 344, Springer, Heidelberg, Germany, 2013.
stantial differences among populations, supported that [3] J. M. Fernández-Martı́nez, L. Velasco, and B. Pérez-Vich, “Pro-
wild populations of this species show a high degree of gress in research on breeding for resistance to broomrape,” in
The Scientific World Journal 13

Proceedings of the 18th International Sunflower Conference, 2012, [19] P. Shindrova and E. Penchev, “Race composition and distri-
http://www.asagir.org.ar/asagir2008/buscar congreso.asp. bution of broomrape (Orobanche cumana Wallr.) in Bulgaria
[4] I. V. Novopokrovsky and N. N. Tzvelev, “Orobanchaceae Lindl,” during 2007–2011,” Helia, vol. 35, no. 57, pp. 87–94, 2012.
in Flora SSSR, V. L. Komarov, Ed., no. 23, pp. 19–117, Botaniceskij [20] M. C. Vaz Patto, R. Dı́az-Ruiz, Z. Satovic, B. Román, A. J.
Institut Akademija Nauk SSSR, Leningrad, Russia, 1958. Pujadas-Salvà, and D. Rubiales, “Genetic diversity of Moroccan
[5] G. Beck-Mannagetta, “Orobanchaceae,” in Das Pflanzenreich, populations of Orobanche foetida: evolving from parasitising
A. Engler, Ed., vol. 4, pp. 1–348, von Wilhelm Engelmann, wild hosts to crop plants,” Weed Research, vol. 48, no. 2, pp. 179–
Leipzig, Germany, 1930. 186, 2008.
[6] A. J. Pujadas-Salvà and L. Velasco, “Comparative studies on [21] J. R. Qasem, “Parasitic flowering plants on cultivated plants in
Orobanche cernua L. and O. cumana Wallr. (Orobanchaceae) in Jordan—the present status and management,” Pakistan Journal
the Iberian Peninsula,” Botanical Journal of the Linnean Society, of Weed Science, vol. 16, no. 2, pp. 227–239, 2010.
vol. 134, no. 4, pp. 513–527, 2000. [22] J. J. Burdon and P. H. Thrall, “Pathogen evolution across the
[7] B. Román, C. Alfaro, A. M. Torres et al., “Genetic relationships agro-ecological interface: implications for disease manage-
among Orobanche species as revealed by RAPD analysis,” ment,” Evolutionary Applications, vol. 1, no. 1, pp. 57–65, 2008.
Annals of Botany, vol. 91, no. 6, pp. 637–642, 2003. [23] C. J. Botanga, J. G. Kling, D. K. Berner, and M. P. Timko, “Gen-
[8] D. Delipavlov, “Orobanchaceae,” in Flora Reipublicae Bulgari- etic variability of Striga asiatica (L.) Kuntz based on AFLP
cae, S. I. Kozuharov and B. A. Kuzmanov, Eds., vol. 10, pp. 291– analysis and host-parasite interaction,” Euphytica, vol. 128, no.
325, Marin Drinov Academic Publishing House, Sofia, Bulgaria, 3, pp. 375–388, 2002.
1995. [24] C. J. Botanga and M. P. Timko, “Genetic structure and analysis
[9] B. K. Blackman, M. Scascitelli, N. C. Kane et al., “Sunflower of host and nonhost interactions of Striga gesnerioides (Witch-
domestication alleles support single domestication center in weed) from Central Florida,” Phytopathology, vol. 95, no. 10, pp.
eastern North America,” Proceedings of the National Academy 1166–1173, 2005.
of Sciences of the United States of America, vol. 108, no. 34, pp. [25] C. J. Botanga and M. P. Timko, “Phenetic relationships among
14360–14365, 2011. different races of Striga gesnerioides (Willd.) Vatke from West
[10] E. D. Putt, “Sunflower early history,” in Sunflower Production Africa,” Genome, vol. 49, no. 11, pp. 1351–1365, 2006.
and Technology, A. A. Schneiter, Ed., Agronomy Monograph no. [26] A. Olivier, J. C. Glaszmann, C. Lanaud, and G. D. Leroux,
35, pp. 1–21, ASA-CSSA-SSSA, Madison, Wis, USA, 1997. “Population structure, genetic diversity and host specificity of
[11] D. Škorić, “The genetics of sunflower,” in Sunflower Genetics the parasitic weed Striga hermonthica (Scrophulariaceae) in
and Breeding, D. Škorić and Z. Sakač, Eds., pp. 1–163, Serbian Sahel,” Plant Systematics and Evolution, vol. 209, no. 1-2, pp. 33–
Academy of Sciences and Arts, Novi Sad, Serbia, 2012. 45, 1998.
[12] V. Encheva and P. Shindrova, “Broomrape (Orobanche cumana [27] G. Gagne, P. Roeckel-Drevet, B. Grezes-Besset et al., “Study of
Wallr.)—a hindrance to sunflower production in Bulgaria,” in the variability and evolution of Orobanche cumana populations
Proceedings of the 3rd International Workshop on Orobanche and infesting sunflower in different European countries,” Theoretical
Related Striga Research, pp. 619–622, Royal Tropical Institute, and Applied Genetics, vol. 96, no. 8, pp. 1216–1222, 1998.
Amsterdam, The Netherlands, 1994.
[28] R. Pineda-Martos, L. Velasco, J. Fernández-Escobar, J. M.
[13] P. Jouffret and V. Lecomte, “Orobanche sur tournesol. Un Fernández-Martı́nez, and B. Pérez-Vich, “Genetic diversity of
parasite en extension,” Lettre d’Informations Regionales Orobanche cumana populations from Spain assessed using SSR
CETIOM, Zone Sud, 2010, http://www.cetiom.fr/fileadmin/ markers,” Weed Research, vol. 53, no. 4, pp. 279–289, 2013.
cetiom/regions/Sud/PDF/2010/oleomail/OleoMail TO Prophy-
[29] Z. Satovic, D. M. Joel, D. Rubiales, J. I. Cubero, and B.
laxie O cumana sud 07092010.pdf.
Romn, “Population genetics in weedy species of Orobanche,”
[14] M. Amri, Z. Abbes, S. B. Youssef, M. Bouhadida, H. B. Salah, and Australasian Plant Pathology, vol. 38, no. 3, pp. 228–234, 2009.
M. Kharrat, “Detection of the parasitic plant, Orobanche
cumana on sunflower (Helianthus annuus L.) in Tunisia,” [30] M. I. Rodrı́guez-Ojeda, J. M. Fernández-Martı́nez, L. Velasco,
African Journal of Biotechnology, vol. 11, no. 18, pp. 4163–4167, and B. Pérez-Vich, “Extent of cross-fertilization in Orobanche
2012. cumana Wallr,” Biologia Plantarum, vol. 57, no. 3, pp. 559–562,
2013.
[15] M. Castejón, F. Romero-Muñoz, and L. Garcı́a-Torres, “Orob-
anche cernua seed dispersal through sunflower achenes,” Helia, [31] B. Achtarov, “Beitrag zur Kenntnis der Flora Bulgariens , nebst
vol. 14, pp. 51–54, 1991. kritischen Bemerkungen,” Bulletin of the Bulgarian Botanical
Society, vol. 9, pp. 57–73, 1943.
[16] B. Pérez-Vich, L. Velasco, P. J. Rich, and G. Ejeta, “Marker-
assisted and physiology-based breeding for resistance to [32] N. Stojanov and B. Stefanov, Flora Bulgarica, University Press,
Orobanchaceae,” in Parasitic Orobanchaceae—Parasitic Mech- Sofia, Bulgaria, 1948.
anisms and Control Strategies, D. M. Joel, J. Gressel, and L. J. [33] K. Stoyanov, “The trophic plasticity of genus Orobanche L.
Musselman, Eds., pp. 369–391, Springer, Heidelberg, Germany, (Orobanchaceae) in Bulgaria,” Agricultural Sciences, vol. 4, no.
2013. 12, pp. 67–72, 2013.
[17] A. V. Vrânceanu, V. A. Tudor, F. M. Stoenescu, and N. Pirvu, [34] B. Pérez-Vich, B. Akhtouch, S. J. Knapp et al., “Quantitative trait
“Virulence groups of Orobanche cumana Wallr. differential loci for broomrape (Orobanche cumana Wallr.) resistance in
hosts and resistance sources and genes in sunflower,” in Pro- sunflower,” Theoretical and Applied Genetics, vol. 109, no. 1, pp.
ceedings of the 9th International Sunflower Conference, pp. 74– 92–102, 2004.
80, International Sunflower Association, Paris, France, 1980. [35] R. Peakall and P. E. Smouse, “GenALEx 6.5: genetic analysis in
[18] P. Shindrova, “Broomrape (Orobanche cumana Wallr.) in Excel. Population genetic software for teaching and research-
Bulgaria—distribution and race composition,” Helia, vol. 29, no. an update,” Bioinformatics, vol. 28, no. 19, Article ID bts460, pp.
44, pp. 111–120, 2006. 2537–2539, 2012.
14 The Scientific World Journal

[36] L. Excoffier and H. E. L. Lischer, “Arlequin suite ver 3.5: a [53] C. J. Thorogood, F. J. Rumsey, S. A. Harris, and S. J. Hiscock,
new series of programs to perform population genetics analyses “Gene flow between alien and native races of the holoparasitic
under Linux and Windows,” Molecular Ecology Resources, vol. angiosperm Orobanche minor (Orobanchaceae),” Plant System-
10, no. 3, pp. 564–567, 2010. atics and Evolution, vol. 282, no. 1-2, pp. 31–42, 2009.
[37] T. de Meeûs, K. D. McCoy, F. Prugnolle et al., “Population genet- [54] C. J. Thorogood, F. J. Rumsey, and S. J. Hiscock, “Host-
ics and molecular epidemiology or how to “débusquer la bête”,” specific races in the holoparasitic angiosperm Orobanche minor:
Infection, Genetics and Evolution, vol. 7, no. 2, pp. 308–332, 2007. implications for speciation in parasitic plants,” Annals of Botany,
[38] S. Arnaud-Haond and K. Belkhir, “GENCLONE: a computer vol. 103, no. 7, pp. 1005–1014, 2009.
program to analyse genotypic data, test for clonality and [55] K. Yoneyama, C. Ruyter-Spira, and H. Bouwmeester, “Induction
describe spatial clonal organization,” Molecular Ecology Notes, of germination,” in Parasitic Orobanchaceae—Parasitic Mech-
vol. 7, no. 1, pp. 15–17, 2007. anisms and Control Strategies, D. M. Joel, J. Gressel, and L. J.
Musselman, Eds., pp. 167–194, Springer, Heidelberg, Germany,
[39] P. G. Meirmans and P. W. Hedrick, “Assessing population struc-
2013.
ture: F 𝑆𝑇 and related measures,” Molecular Ecology Resources,
vol. 11, no. 1, pp. 5–18, 2011. [56] D. M. Joel, S. K. Chaudhuri, D. Plakhine, H. Ziadna, and J.
C. Steffens, “Dehydrocostus lactone is exuded from sunflower
[40] M. Nei and R. K. Chesser, “Estimation of fixation indices and
roots and stimulates germination of the root parasite Orobanche
gene diversities,” Annals of Human Genetics, vol. 47, no. 3, pp.
cumana,” Phytochemistry, vol. 72, no. 7, pp. 624–634, 2011.
253–259, 1983.
[57] M. Castejón, M. J. Suso, F. Romero-Muñoz, and L. Garcı́a-
[41] M. Nei, Molecular Evolutionary Genetics, Columbia University Torres, “ Isoenzymatic study of broomrape (Orobanche cernua)
Press, New York, NY, USA, 1987. populations infesting sunflower (Helianthus annuus),” in Pro-
[42] T. Jombart, D. Pontier, and A.-B. Dufour, “Genetic markers in ceedings of the 5th International Symposium on Parasitic Weeds,
the playground of multivariate analysis,” Heredity, vol. 102, no. pp. 313–319, International Maize and Wheat Improvement
4, pp. 330–341, 2009. Center (CIMMYT), Nairobi, Kenya, 1991.
[43] J. K. Pritchard, M. Stephens, and P. Donnelly, “Inference of [58] R. Pineda-Martos, R. Batchvarova, J. M. Fernández-Martı́nez,
population structure using multilocus genotype data,” Genetics, L. Velasco, and B. Pérez-Vich, “Genetic diversity of Orobanche
vol. 155, no. 2, pp. 945–959, 2000. cumana populations from Spain and Eastern Europe,” in
[44] D. Falush, M. Stephens, and J. K. Pritchard, “Inference of popu- Proceedings of the 10th World Congress of Parasitic Plants, D.
lation structure using multilocus genotype data: linked loci and Rubiales, J. Westwood, and A. Uludag, Eds., p. 145, IPPS,
correlated allele frequencies,” Genetics, vol. 164, no. 4, pp. 1567– International Parasitic Plant Society, Kusadasi, Turkey, 2009.
1587, 2003. [59] L. Molinero-Ruiz, A. B. Garcı́a-Carneros, M. Collado-Romero,
[45] D. A. Earl and B. M. von Holdt, “STRUCTURE HARVESTER: a S. Raranciuc, J. Domı́nguez, and J. M. Melero-Vara, “Pathogenic
website and program for visualizing STRUCTURE output and and molecular diversity in highly virulent populations of the
implementing the Evanno method,” Conservation Genetics parasitic weed Orobanche cumana (sunflower broomrape) from
Resources, vol. 4, no. 2, pp. 359–361, 2012. Europe,” Weed Research, vol. 54, no. 1, pp. 87–96, 2014.
[46] G. Evanno, S. Regnaut, and J. Goudet, “Detecting the number of [60] D. M. Joel, “Identification of Orobanche seeds,” in Proceedings of
clusters of individuals using the software STRUCTURE: a the 4th International Symposium on Parasitic Flowering Plants,
simulation study,” Molecular Ecology, vol. 14, no. 8, pp. 2611– H. C. Weber and W. L. Forstreuter, Eds., pp. 437–444, Philipps-
2620, 2005. Universität, Marburg, Germany, 1987.
[61] L. Plaza, I. Fernández, R. Juan, J. Pastor, and A. Pujadas, “Micro-
[47] M. Jakobsson and N. A. Rosenberg, “CLUMPP: a cluster match-
morphological studies on seeds of Orobanche species from the
ing and permutation program for dealing with label switching
Iberian Peninsula and the Balearic Islands, and their systematic
and multimodality in analysis of population structure,” Bioin-
significance,” Annals of Botany, vol. 94, no. 1, pp. 167–178, 2004.
formatics, vol. 23, no. 14, pp. 1801–1806, 2007.
[62] M. I. Rodrı́guez-Ojeda, R. Pineda-Martos, L. C. Alonso et al.,
[48] N. A. Rosenberg, “DISTRUCT: a program for the graphical “A dominant avirulence gene in Orobanche cumana triggers Or5
display of population structure,” Molecular Ecology Notes, vol. 4, resistance in sunflower,” Weed Research, vol. 53, no. 5, pp. 322–
no. 1, pp. 137–138, 2004. 327, 2013.
[49] H. Gao, S. Williamson, and C. D. Bustamante, “A markov chain [63] F. Guérin, P. Gladieux, and B. L. Cam, “Origin and colonization
monte carlo approach for joint inference of population structure history of newly virulent strains of the phytopathogenic fungus
and inbreeding rates from multilocus genotype data,” Genetics, Venturia inaequalis,” Fungal Genetics and Biology, vol. 44, no. 4,
vol. 176, no. 3, pp. 1635–1651, 2007. pp. 284–292, 2007.
[50] L. Excoffier, P. E. Smouse, and J. M. Quattro, “Analysis of mole- [64] C. Xhaard, B. Fabre, A. Andrieux et al., “The genetic structure
cular variance inferred from metric distances among DNA hap- of the plant pathogenic fungus Melampsora larici-populina on
lotypes: application to human mitochondrial DNA restriction its wild host is extensively impacted by host domestication,”
data,” Genetics, vol. 131, no. 2, pp. 479–491, 1992. Molecular Ecology, vol. 20, no. 13, pp. 2739–2755, 2011.
[51] T. Ohta, “Linkage disequilibrium due to random genetic drift [65] K. Huang, R. Whitlock, M. C. Press, and J. D. Scholes, “Variation
in finite subdivided populations,” Proceedings of the National for host range within and among populations of the parasitic
Academy of Sciences of the United States of America, vol. 79, no. plant Striga hermonthica,” Heredity, vol. 108, no. 2, pp. 96–104,
6, pp. 1940–1944, 1982. 2012.
[52] C. J. Thorogood, F. J. Rumsey, S. A. Harris, and S. J. Hiscock, [66] Z. P. Song, X. Xu, B. Wang, J. K. Chen, and B.-. Lu, “Genetic
“Host-driven divergence in the parasitic plant Orobanche minor diversity in the northernmost Oryza rufipogon populations
Sm. (Orobanchaceae),” Molecular Ecology, vol. 17, no. 19, pp. estimated by SSR markers,” Theoretical and Applied Genetics,
4289–4303, 2008. vol. 107, no. 8, pp. 1492–1499, 2003.
The Scientific World Journal 15

[67] H.-W. Cai, X.-K. Wang, and H. Morishima, “Comparison of

population genetic structures of common wild rice (Oryza
rufipogon Griff.), as revealed by analyses of quantitative traits,
allozymes, and RFLPs,” Heredity, vol. 92, no. 5, pp. 409–417,
2004.
[68] M. C. Vaz Patto, M. FernÁndez-Aparicio, Z. Satovic, and D.
Rubiales, “Extent and pattern of genetic differentiation within
and between European populations of Phelipanche ramosa
revealed by amplified fragment length polymorphism analysis,”
Weed Research, vol. 49, no. 1, pp. 48–55, 2009.
[69] C. B. Heiser, D. M. Smith, S. B. Clevenger, and W. C. Martin Jr.,
“The North American sunflowers (Helianthus),” Memoirs of the
Torrey Botanical Club, vol. 22, pp. 1–218, 1969.
[70] A. Höniges, A. Ardelean, X. Xi, K. Yoneyam, and K. Wegmann,
“Towards understanding Orobanche host-specificity,” Roma-
nian Agricultural Research, vol. 29, pp. 313–322, 2012.