Genetic diversity and clonal variation within the main Sicilian olive cultivars based on morphological traits and microsatellite markers

Scientia Horticulturae 180 (2014) 130–138

Contents lists available at ScienceDirect

Scientia Horticulturae journal homepage: www.elsevier.com/locate/scihorti

Genetic diversity and clonal variation within the main Sicilian olive cultivars based on morphological traits and microsatellite markers T. Caruso, F.P. Marra ∗ , F. Costa, G. Campisi, L. Macaluso, A. Marchese Department of Agricultural and Forest Sciences (SAF), University of Palermo, Viale delle Scienze, Edificio 4 Ingresso H, 90128 Palermo, Italy

a r t i c l e

i n f o

Article history: Received 23 July 2014 Received in revised form 8 October 2014 Accepted 11 October 2014 Keywords: Olive Olea europaea SSR Clonal selection Microsatellites

a b s t r a c t The richness of Olea europaea (L.) genetic resources in Sicily is well documented. In the last 30 years, most of the local cultivars, landraces and ecotypes have been gathered together in a large ex-situ collection, containing more than 300 genotypes. In this study, 45 putative clones of the main Sicilian olive cultivars were characterized morphologically using microsatellite markers to unambiguously identify possible superior genotypes. The microsatellites employed were polymorphic (observed heterozygosity = 0.71; polymorphic information content = 0.59), discriminated 52% of the genotypes and enabled the detection of intra-cultivar polymorphism, derived from both somatic mutations, indicating the presence of polyclonal cultivars, or from gametic origin, thus suggesting the presence of cultivar-populations. A high level of genetic variability was detected within the ‘Biancolilla’, ‘Giarraffa’ and ‘Moresca’ genotypes, whereas low variation was found within the ‘Cerasuola’ and ‘Tonda Iblea’ genotypes. The combination of UPGMA cluster analysis of data obtained from microsatellite analysis, with canonical discriminant analysis (CDA), based on 18 morphological variables, measured under the same conditions, enabled intra-cultivar diversity, attributable to genetic factors rather than to environmental ones to be identified. The goodness of fit between microsatellite profiles and the CDA analysis was significantly supported by the Mantel test (r = 0.3; p < 0.001). Genotypes and clonal variants with superior traits (larger fruit size; compact tree habit, apt for high density planting; higher oleic acid content) were identified, suitable for enlarging their area of cultivation. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Olive (Olea europaea L.) is one of the most economically important tree species in the Mediterranean basin, cultivated for its fruit and oil. Italy is the world’s second largest olive producer after Spain (Faostat, 2013). Among the Italian regions, Sicily has a significant place in the olive and olive oil production, industry and exportation, being the third largest region of production after Apulia and Calabria (ISTAT, 2013). Olives have been cultivated in Sicily since antiquity; the Phoenicians brought the olive to Sicily in the sixth century BC and the Romans continued the expansion of olive cultivation throughout the Mediterranean by developing grafting techniques (Zohary and Hopf, 1994; Besnard et al., 2001; Rugini et al., 2011). The genetic richness of Sicilian olive germplasm is well documented (Bottari

∗ Corresponding author. Tel.: +39 3289866192; fax: +390916515531. E-mail addresses: [email protected], [email protected] (F.P. Marra). http://dx.doi.org/10.1016/j.scienta.2014.10.019 0304-4238/© 2014 Elsevier B.V. All rights reserved.

and Spina, 1952; La Mantia et al., 2005; Caruso et al., 2007; Marra et al., 2013; Besnard et al., 2013) and represents a heritage patrimony of economical and scientific value, particularly for breeding programmes. Parentage studies performed by Marra et al. (2013) demonstrated that both somatic mutations and new genotypes derived through sexual crosses created the rich diversification of the Sicilian olive germplasm. The first historical investigation on indigenous Sicilian heritage olive was conducted in the second half of the XIX century by Caruso (1883). Sicilian farmers probably selected late ripening cultivars, since long, mild and wet autumns allow for olive oil accumulation and complete ripening. It is likely that vigorous and drought resistant trees, with large-sized fruits were selected by local farmers over the centuries. Since the 1980s, the Department of Agricultural and Forest Sciences (SAF) of Palermo has worked extensively on the characterization and the conservation of the main local olive cultivars and has studied their efficiency in yield and in oil quality (La Mantia et al., 2005; Caruso et al., 2007; Marra et al., 2013), following the first survey conducted in the middle of the XX century by Bottari

T. Caruso et al. / Scientia Horticulturae 180 (2014) 130–138

and Spina (1952). In 2002, a large olive germplasm collection was established in Sicily, which today contains eight well-known and extensively grown cultivars, 17 minor or neglected cultivars and 122 native genotypes. Currently in the island, olive oil production is based mainly on the cultivars ‘Biancolilla’, ‘Cerasuola’, ‘Moresca’, ‘Nocellara del Belice’, ‘Nocellara Etnea’, ‘Ogliarola Messinese’, ‘Santagatese’ and ‘Tonda Iblea’ (Caruso et al., 2007). The table olive industry is also significant (8% of total olive production) and based mainly on the cultivar ‘Nocellara del Belice’ and, to a minor extent, on ‘Nocellara Etnea’, ‘Ogliarola Messinese’ and ‘Moresca’ producing large-sized fruits of high commercial value (Caruso et al., 2007). As most varieties are often known under different names, according to the areas of cultivation, many cases of homonymy and synonymy occur, complicating cultivar identification. However, a certain amount of morphological variability has been observed within each cultivar, suggesting that the name of a single cultivar could encompass a pool of different genotypes (cultivar populations) and/or populations of clones (mixture of clonal variants). Generations of farmers have played a role in selecting, conserving and genetic improving of the Sicilian olive germplasm. Many of the morphological types of each of the main cultivars, identified thanks to reports of millers, growers and nurserymen, have been collected ex situ in a regional repository and observed over the years concerning their canopy architecture, time-course of phenological phases, tree crop efficiency, and fruit, leaf, inflorescence and endocarp traits. It is largely accepted that the olive cultivar discrimination based on morphological descriptions is not completely reliable (Belaj et al., 2001) therefore DNA molecular markers, particularly microsatellites (SSRs; simple sequence repeats), are today widely used (Bracci et al., 2011) to complement morphological analyses and to unambiguously identify the accessions held in collections. Genetic variation has been reported among naturally occurring olive clones in literature with molecular markers. Clones were identified with RAPD and ISSR (Gemas et al., 2004; Gomes et al., 2008; Martins-Lopes et al., 2009), with AFLP (Frane et al., 2010), and microsatellites (Lopes et al., 2004; Muzzalupo et al., 2010; Zaher et al., 2011; Albertini et al., 2011; Ipek et al., 2012). Although currently there is intense research to develop reliable techniques for detecting mutations in genes, clone identification is still predominantly based on the study of phenotypic traits, integrated with molecular analyses. In this study, intra-cultivar variation of the most widely distributed Sicilian olive cultivars, was identified through morphological and molecular characterization, employing a reliable set of SSRs (reviewed in Baldoni et al., 2009), to detect new clonal variants and to confirm the suspected existence of polyclonal cultivars and cultivar-populations. Novel insights on olive genetic diversity were acquired, that will be useful for conservation and breeding purposes, tree nursery genetic certification and oil traceability.

2. Material and methods 2.1. Plant material A total of 45 putative clones belonging to eight standard cultivars were studied. These comprised: 10 putative clones of ‘Biancolilla’, six of ‘Cerasuola’, three of ‘Giarraffa’, eight of ‘Moresca’, 11 of ‘Nocellara del Belice’, two of ‘Ogliarola Messinese’, four of ‘Tonda Iblea’ and two genotypes of ‘Nocellara Messinese’, a minor cultivar the fruits of which could play a role in the further development of the Sicilian table industry. The cultivar ‘Nocellara Etnea’ was included in the analysis as an internal control. All genotypes

131

were grown at the Sicilian olive regional repository established in 1991 in the ‘Azienda Carboj E.S.A.’—Castelvetrano (37.30 Lat N.; Sicily). In that collection, there were nine plants of each putative clone, divided into three replicates of three plants, placed at a distance of 5 × 7 m and trained to free vase shape. Sicilian cultivars chosen as standards for genetic and phytosanitary certification purposes (Caruso et al., 2007) were planted in the same farm; they were distributed in the field according to the same experimental scheme adopted for the putative clones, divided into three replicates of three plants, in order enable the comparison with their putative clones and to exclude the influence of environmental factors on the phenotype. In the text the standard cultivars were indicated as STD.

2.2. Morphological traits and biostatistical analysis Morphological observations were carried out on all cultivars and their putative clones in 2012–2013. Eighteen morphological variables, interval and quantitative were selected for the characterization of the putative clones among the range of characters defined as the main descriptors of the olive tree (Bottari and Spina, 1952; Barranco et al., 2000; Bartolini et al., 2005; COI, 1997; Caruso et al., 2007). The choice of these traits was guided by previous studies on Italian olive genetic resources (Marra et al., 2013); the interval classification was obtained using the classes reported by Caruso et al. (2007). Canonical discriminant analysis (CDA) based on the18 morphological variables was performed using the Systat statistical program (SYSTAT Software Inc., Chicago, IL). The power of discrimination of the canonical discriminant functions (CDFs) generated was tested by the “Jackknife misclassification method”, a procedure which omits one observation, elaborates the classification function employing all other observations (n1 + n2 − 1) and uses the classification function to classify the excluded observation. This process is reiterated for each of the observations (Osuji et al., 2013). Means of standardized canonical scores of major CDFs were plotted. Two-dimensional CDFs plots were created in order to visualize the groups of putative clones in relation to the cultivar considered as standard. Ellipses marked the 68% confidence level for the analyzed groups.

2.3. DNA extraction and microsatellite evaluation Genomic DNA was extracted from young leaves according to the protocol developed by Doyle and Doyle (1987). DNA was amplified with nine fluorescently labeled SSR primer pairs, five of which were combined in two multiplexed primer sets as follow DCA: 03, 05, 18 (Sefc et al., 2000); Gapu: 45, 71b (Carriero et al., 2002); and four used in single PCR reactions: DCA13 (Sefc et al., 2000); EMO90 and EMOL (De la Rosa et al., 2002); UDO43 (Cipriani et al., 2002). The number of alleles for each SSR locus, the expected heterozygosity (He), the observed heterozygosity (Ho) and PIC (polymorphic information content) were calculate with PowerMarker v3.25 software (Liu and Muse, 2005). Genotypes that were not distinguishable at the molecular level and/or those showing extra alleles were excluded for the calculation of SSR genetic parameters. A UPGMA dendrogram was also constructed using the PowerMarker V3.25 software employing the coefficient of similarity Nei (1973). A Mantel (1967) test was also computed with PowerMarker (Liu and Muse, 2005) to check the goodness of fit between genetic profiles and CDA analyses, based on morphological traits, of all genotypes analyzed, by using Simple Matching dissimilarity matrixes; p values were also calculated.

132

T. Caruso et al. / Scientia Horticulturae 180 (2014) 130–138

3. Results 3.1. Morphological characterization The olive genotypes under study were evaluated for 18 morphological and biometric traits following the descriptors reported by Caruso et al. (2007). The morphological variability found among ‘Biancolilla’ putative clones was high. Regarding the leaf, the predominant shape was elliptic-lanceolate (4< length-to-width ratio <6); the leaf length (mm) varied between 46.54 (clone 9) and 65.31 (clone 4), and the leaf width (mm) between 10.61 (clone 6) and 13.72 (clone 8). Fruit shape was spherical (length-to-width ratio <1.25) in clones 7 and 9, unlike the remaining clones and the STD, in which the shape of the fruit was elliptical (1.25< length-to-width ratio <1.50). The weight of the fruit (g) varied between 2.3 (clone 6) and 6.8 (clone 11), the latter had the highest fruit weight among all clones. A large proportion of the clones presented elliptical pit shape (1.9< length-to-width ratio <2.2), except clone 7, showing oval pit (1.4< length-to-width ratio <1.8). Clone 11 had the largest pit. Among the ‘Cerasuola’ group, the leaf shape was ellipticallanceolate in all clones; whilst the STD displayed an elliptical shape (length-to-width ratio <4), the length (mm) of the leaf varied between 57.81 (clone 1) and 65.49 (clone 6), the leaf width (mm) varied from 13.73 (clone 3) to 16.9 STD. The fruit shape was spherical, pit shape was oval and the weight of the fruit (g) varied between 3.36 (clone 1) and 4.41 (clone 4). Regarding the putative clones of ‘Giarraffa’, the prevalent form of the leaf was elliptical-lanceolate, the leaf length (mm) ranged from 47.80 (clone 2) to 60.88 (clone 3), the leaf width (mm) varied from 12.18 (clone 2) to 13.78 (STD). The shape of the fruit was elliptical and the highest weight of the fruit (g) was found in clone 3 (7.49) and in the STD (7.33). All analyzed putative clones and STD presented elliptical pit shape. Among the putative clones of ‘Moresca’, the predominant shape of the leaf was elliptical, except for clones 1 and 10, which displayed an elliptical-lanceolate leaf shape; the leaf length (mm) varied between 54.19 (clone 3) and 62.45 (clone 6), the smallest leaf width (mm) was found in clone 10 (11.71). The fruit shape was elliptical in all putative clones; clone 10 showed fruit weight (g) greater than 6; most of the putative clones presented an elliptical pit shape, except the clone 10, which had elongated pit shape (length-to-width ratio >2.2). In all putative clones of ‘Nocellara del Belice’, the shape of the leaf was elliptic-lanceolate, leaf length (mm) varied between 52.15 (clone 2) and 74.85 (clone 8), the leaf width (mm) ranged from 12.29 (clone 1) to 17.68 (clone 7). The fruit shape was mainly spherical, except clones 1 and 11, which had elliptical fruit shape; the highest weight of the fruit (g) was found in clone 10 (10.68 g), while the lowest was found in clone 11 (3.72 g); most of the putative clones presented oval pit shape, except clones 1 and 11, which presented elliptical pit shape. Concerning ‘Nocellara Messinese’, putative clone 1 presented elliptic-lanceolate leaf, fruit and endocarp elliptical as the standard. However, fruit and pit weight was smaller compared to STD. All putative clones of ‘Ogliarola Messinese’ and the STD presented elliptic-lanceolate leaf shape, fruit elliptical and elongated pit, except for the clone 3, having elliptical pits. ‘Tonda Iblea’ putative clones were quite similar among them. They presented leaf shape elliptical, spherical fruit shape and oval pit shape. 3.2. Biostatistical analyses of morphological traits The 18 morphological and biometric data were statistically analyzed by CDA and results were represented in Fig. 1. As reported in

Table 1, the CDF1 explained 24% of the total variance present in the studied genotypes, the CDF2 however, explained 16% of the total variance. CDF3 and CDF4 contributed an additional 12% and 10% of the total variance, respectively. The combination of the first four CDFs explained 62% of the total variance. Total canonical structure was computed for CDF1/CDF2, explaining 40.3% of the total variance, and for CDF1/CDF4, which explained 33% of the total variance. For CDF1, the parameters with higher discriminating ability were: shape, width and length of the fruit. Concerning CDF2, leaf shape, area and length were found to have the greatest discriminatory power. For CDF3: the weight of the fruit and the length of the pit (or endocarp) were the factors with the greatest discriminatory power. Finally, for CDF4: length, width and shape of the pit; weight, length, width and shape of the fruit and the flesh to pit ratio had the greatest discriminatory power. In Fig. 1a, many putative clones of ‘Biancolilla’ showed a high degree of dispersion from the genotype standard. Putative clones 4, 7, 9, 11 and 12 mainly differed along the CDF2, on which morphological parameters related to the leaf carried greater weight. Putative clones 4, 5 and 9 differed also along CDF1 and CDF4, on which fruit and pit weight loaded more; clones 6, 10, 11 and 12 placed along the CDF4. The putative clones of ‘Cerasuola’ showed a low degree of dispersion with respect to their STD along CDF1 and CDF2 (Fig. 1b). However, along the components CDF4 clones 3, 5 and 6 were discriminated. All three putative clones of ‘Giarraffa’ were scattered compared to the genotype standard (Fig. 1c). Clones 5, 6 and 10 of ‘Moresca’ were rather dispersed from the standard (Fig. 1d). In particular, clone 10 showed marked differences from the other for all three components examined. Concerning putative clones of ‘Nocellara del Belice’, 1, 5, 10 and 11 were all dispersed from the standard genotype (Fig. 1e). In detail, clones 1, 10 and 11 separated along components CDF1, CDF2 and CDF4. Clone 5 differentiated along component 4. All three putative clones of ‘Ogliarola Messinese’ were differentiated for all components (Fig. 1f), as well as the putative clone 1 of ‘Nocellara Messinese’ differing for all components from the standard (data not shown). All putative clones of ‘Tonda Iblea’ did not differ significantly from the standard (Fig. 1 g). However, clone 1 showed some differences from the standard along the component CDF2, while clone 2 differed along CDF4. Overall the correct classification matrix average was 72% (data not shown) and 65% with the Jackknife classification matrix. The genotypes showing the highest correct Jackknife classification (>70%) are presented in Table 3. On the whole 44% of the genotypes were correctly classified with the Jackknife method. The discrepancies between the classification matrix and the Jackknife classification matrix depended on the different classification functions used by the two methods. As expected, the Jackknife classification matrix presented lower values, since the Jackknife procedure provides conservative estimates of the actual probability of misclassification, especially when numerous variables are included in the analysis. 3.3. SSR genetic parameters and clustering analysis The number of alleles per locus ranged from three with Gapu45 to 11 with UDO43, with an average of 5.7. The average expected heterozygosity was 0.62. The average observed heterozygosity value was 0.71, ranging from 0.32 (EMOL) to 0.96 (DCA13 and DCA03). The PIC (polymorphic information content) varied from 0.25 with EMOL to 0.81 with DCA18, with an average value of 0.59 (Table 2). Excluding the genotype employed as a control, this set of SSR markers discriminated 52% of the genotypes. The UPGMA dendrogram (Fig. 2) showed the undistinguishable putative clones from the reference cultivars. In most of the cases these SSRs enabled the

T. Caruso et al. / Scientia Horticulturae 180 (2014) 130–138

133

Fig. 1. Plots of first and second discriminant functions (CDF1/CDF2) and of first and fourth discriminant functions (CDF1/CDF4) calculated on the basis of 18 variables measured in putative clones of: ‘Biancolilla’ (a); ‘Cerasuola’ (b); ‘Giarraffa’ (c); ‘Moresca’ (d); ‘Nocellara del Belice’ (e); ‘Ogliarola Messinese (f)’; ‘Tonda Iblea’ (g). The ellipse indicates the 68% confidence interval.

distinction of somatic mutants, differing only in one or two SSR alleles from their STD or other closest genotypes (similarity index >0.95), from the siblings, sharing at least one allele for each SSR locus. All genotypes belonging to the group ‘Biancolilla’ represented a clear case of cultivar - population (Fig. 2). Putative clones 4 and

11 showed very narrow genetic relationships and could be somatic mutants of each other, but they were determined to be an unrelated individual from ‘Biancolilla’ STD. Clones 7, 8 and 12 had a closer relationship between each other but they could not be considered siblings of the STD. Putative clones 5 and 9 can be considered genuine somatic mutants of the cultivar; while ‘Biancolilla’ 12 shared

134

T. Caruso et al. / Scientia Horticulturae 180 (2014) 130–138

Fig. 1. (Continued ).

sibling relationships with the STD. The molecular identity of the ‘Biancolilla’ clones 1, 6 and 10 with the STD was also supported by the morphological similarity depicted also by the CDA analysis (Fig. 1a). Putative clones of ‘Cerasuola’ 1, 4 and 5 were found to be somatic mutants of the STD (Fig. 2), indicating that they represent a mixture of clonal variants; this was confirmed from the morphological similarity (Fig. 1b). The remaining clones presented the same SSR profiles than the STD.

Genotypes 1 and 2 of ‘Giarraffa’ were revealed as possible siblings of the STD, while ‘Giarraffa 3’ is genetically close but not a sibling, however this evidence needs to be confirmed by further investigations. Genetic relationships emerged between the ‘Giarraffa’ genotypes with ‘Ogliarola Messinese’ STD and its clones 2 and 3; ‘Nocellara del Belice’ clone 1; and ‘Moresca 10’ (Fig. 2). Concerning the cultivar ‘Moresca’, genotypes 5 and 10 were clearly different from their STD. In particular, the SSR analysis showed that the ‘Moresca’ 5 shared a genetic relationship with

T. Caruso et al. / Scientia Horticulturae 180 (2014) 130–138 Table 1 Standardized canonical coefficients of the first four canonical discriminant functions (CDF) of 18 morphological variables and relative % of variance explained.

135

Table 3 Percentage of correct classification of genotypes (matrix of classification, obtained with Jackknife method, >70%).

Traits

CDF 1

CDF 2

CDF 3

CDF 4

Genotypes

(%) Correct classification

Leaf blade shape Leaf blade length Leaf blade width Leaf apex angle Leaf base angle Leaf max width position Leaf area Fruit shape (length/width) Fruit width Fruit length Transversal diameter of fruit Fruit weight Flesh to pit ratio Pit shape (length/width) Pit width Pit length Transversal pit diameter Pit weight Variance explained (%)

0.041 −0.147 0.077 0.148 −0.178 0.156 0.254 1.019 1.612 −1.429 0.259 0.332 0.129 −0.503 0.288 0.011 −0.266 −0.353 24.00

0.834 −1.246 −0.266 −0.211 −0.156 0.087 1.163 0.191 0.712 0.306 0.265 0.114 −0.628 0.566 0.146 −0.249 −0.107 −0.606 16.00

0.143 −0.686 −0.695 −0.361 0.113 −0.187 0.553 0.069 0.020 −0.508 −0.236 1.367 −0.134 0.609 0.191 −1.068 −0.122 −0.216 12.00

0.372 0.208 0.322 0.061 −0.482 0.030 −0.381 −0.898 −1.171 1.085 0.126 −0.872 1.037 −0.812 −1.042 1.358 0.366 0.391 10.00

‘Biancolilla’ 7 ‘Biancolilla’ 8 ‘Biancolilla’ 9 ‘Biancolilla’ 11 ‘Biancolilla’ 12 ‘Cerasuola’ 1 ‘Cerasuola’ 6 ‘Giarraffa’ 1 ‘Moresca’ 1 ‘Moresca’ 3 ‘Moresca’ 6 ‘Moresca’ 10 ‘Moresca std’ ‘Nocellara del Belice’ 10 ‘Nocellara del Belice’ 11

87 79 77 83 77 70 73 87 90 80 70 100 80 80 80

clones 4 and 5 of the cultivar ‘Cerasuola’. In the case of ‘Moresca’ 10 with ‘Giarraffa 3’ (Fig. 2) the genetic relationship is confirmed by the substantial morphological similarity (Fig. 1c and d) that may represent a case of misidentification which deserves further investigation. Putative clones 1 and 9 were found to be genuine somatic mutants of the STD, while the remaining genotypes 2, 3, 4, 6 were identical to the STD for their SSR profiles (Fig. 2), suggesting that ‘Moresca’ is a polyclonal cultivar. Regarding ‘Nocellara del Belice’, while clones 6 and 10 did not differ from the STD, all the remaining clones showed moderate polymorphism (Fig. 2). Clones 1 and 2 of ‘Nocellara del Belice’ can be considered siblings of the STD, while the other clones can be regarded as somatic mutants. Therefore, ‘Nocellara del Belice’ could be regarded as a cultivar-population in which also clonal variants are present. All ‘Tonda Iblea’ putative clones revealed identical molecular profiles (Fig. 2); this was confirmed morphologically, except for clones 1 and 2 (Fig. 1g). Thus, ‘Tonda Iblea’ can be considered a polyclonal cultivar. Concerning the minor cultivar ‘Nocellara Messinese’ genotype 1, represented a case of sibling, which separated from the STD in the UPGMA dendrogram of similarity, based on SSRs (Fig. 2). Finally, the differences found among ‘Ogliarola Messinese STD’ and the clone 3 can be attributed to somatic mutations, indicating that it is polyclonal cultivar.

Table 2 N. of alleles, observed heterozygosity (Ho), expected heterozygosity (He); polymorphic information content (PIC) in eight STD cultivar and 23 putative clones. Loci SSR

N. alleles

He

Ho

PIC

DCA03 DCA05 DCA18 Gapu45 GAPu71b EMOL EMO90 UDO43 DCA13 Average

6 4 8 3 5 4 5 11 5 5.7

0.81 0.34 0.83 0.49 0.73 0.28 0.60 0.81 0.71 0.62

0.96 0.25 0.93 0.43 0.82 0.32 0.82 0.93 0.96 0.71

0.78 0.31 0.81 0.43 0.69 0.26 0.55 0.78 0.66 0.59

Fig. 2. UPGMA dendrogram showing the genetic diversity of 45 putative clones of eight Sicilian standard cultivars (STD), based on microsatellite markers. ‘Nocellara Etnea STD’ was included in the analysis as an internal control.

136

T. Caruso et al. / Scientia Horticulturae 180 (2014) 130–138

4. Discussion In the present study morphological traits and microsatellite markers were used to detect for the first time the intra-cultivar variability of the main economically important olive cultivars grown in Sicily and to identify new promising genotypes or superior clones through clonal selection. Both clonal selection and cross-breeding programmes are in progress in many countries to select new cultivars. Today desirable traits in novel olive cultivars are quite different respect to those chosen in the recent past. These include early bearing, annual production, low vigour for intensive cultivation and canopy architecture suitable for mechanical harvesting with straddle or canopy contact machines, fruits having high olive oil content and production oil with high rate of bioactive substances beneficial for the human health (Fabbri et al., 2009). 4.1. Morphological diversity A high degree of polymorphism was observed morphologically among the putative clones, confirming the observations of local farmers, millers and growers. The CDA analysis was performed to identify patterns of variability among the putative clones and their relative standards. The combination of the first four CDFs explained 62% of the total variance and was considered statistically satisfactory in order to interpret our experimental results, since values over 50% of the total variance are considered effective for cultivar discrimination in other fruit tree species (Barone et al., 1996). The other canonical functions, from CDF5 to CDF17, explaining the remaining 38% of the total variance, were less important for the discrimination of putative clones and they were excluded for the plot construction. All resulting CDFs were significant and statistically supported by p-value <0.001 and Wilks’s Lambda test. The characters with greater discriminating ability were those relating to the fruit, to the leaf and to the pit (Table 1). Thus, these characters should be preferred and prioritized in the complex activity of morphological characterization of the accession of the olive tree. Belaj et al. (2011) highlighted the importance of the symmetry of the fruit, the shape of the pit, the flesh to pit ratio and oil content for characterization of wild olive genotypes, finding that these morphological traits discriminated 95% of the wild genotypes. D’Imperio et al. (2011) demonstrated the importance of morphological traits of the pit in the discrimination of three Italian olive cultivars from Molise and their relative somatic mutants. In both works, the morphological characteristics of the leaf were only moderately useful in identifying cultivars. In the present study, the traits that have assumed great importance as a source of variability in the first four principal components were partially different from those detected by Belaj et al. (2011) and D’Imperio et al. (2011), since parameters concerning the morphology of the leaf were found also significant in the discrimination of putative clones of key olive cultivars grown on the island. Leaf morphological traits are considered highly influenced by climatic and environmental conditions and are therefore often excluded from analysis. In the context in which our investigation was conducted, however, the environmental influence on the variability of the leaves can be excluded because all genotypes were grown at the same site and subjected to the same cultural practices. Thus the morphological differences observed could be attributable solely to genetic factors. Distributions of the genotypes in the obtained CDA plots were in agreement with morphological and biometric differences observed. 4.2. SSR genetic parameters and genetic polymorphisms The majority of the SSR markers amplified a single locus and the amplification of extra alleles was only sporadically observed with

the primer pair DCA13. The average observed heterozygosity value was higher than the expected heterozygosity value, similar to the finding of Baldoni et al. (2009). The most informative markers were DCA18 and UDO43, in accordance to Baldoni et al. (2009), while the least polymorphic markers were EMOL and DCA05. EMOL was discarded by Baldoni et al. (2009) since it presented two drawbacks; a high occurrence of null alleles and allele homoplasy. The list of the most useful primers in olive reported by Baldoni et al. (2009) should be slightly updated or modified, considering that other recommended primers amplified two loci in the Sicilian germplasm, such as GAPU101, amplifying alleles, ranging from 112–147 bp (data not shown). Our data can be compared to those published by Baldoni et al. (2009) when examining four cultivars (‘Arbequina’, ‘Carolea’, ‘Koroneiki’ and ‘Leccino’), used as control in our PCR reactions, and six microsatellites in common (DCA:3, 5, 18; UDO43; EMO90 and Gapu71b), but allelic length differences and allele drop out occurred (data not shown). Apart from the morphological traits, a high degree of polymorphism was also found at the molecular level. The chosen SSRs revealed intra-cultivar genetic diversity, caused by somatic mutations, leading to polyclonal cultivars, or by sexual reproduction, originated from natural crossing and subsequent seed dissemination leading to cultivar-populations. ‘Biancolilla’, ‘Giarraffa’, ‘Nocellara del Belice’, ‘Nocellara Messinese’ and ‘Moresca’ represented a clear case of cultivar–populations, in which also clonal variants were present, whereas ‘Cerasuola’, ‘Ogliarola Messinese’ and ‘Tonda Iblea’ resulted a mixture of clonal variants. In the literature, high level of intra-cultivar genetic diversity has been reported. Clones were identified by means of RAPD and ISSR markers in the Portuguese cultivar ‘Galeca’ (Gemas et al., 2004), in ‘VerdealTransmontana’ (Gomes et al., 2008) and in the cultivar ‘Cobranc¸osa’ (Martins-Lopes et al., 2009). Frane et al. (2010), using amplified fragment length polymorphisms, identified intra-cultivar polymorphism in a pool of genotypes of the Croatian cultivar ‘Oblica’ grown in the same agro-climatic conditions. Muzzalupo et al. (2010) found genetic variability in ‘Carolea’ using SSR markers, which the authors attributed to the occurrence of somatic mutations that had accumulated over the centuries of cultivation of this genotype. High levels of genetic polymorphism were also detected with SSR markers in ‘Picual’, ‘Conserva de Elvas’, ‘Verde Verdelho’, ‘Ascolana’, ‘Coratina’ and ‘Picholine’ (Lopes et al., 2004); within ‘Picholine Marocaine’ genotypes (Zaher et al., 2011); and among ‘Gemlik’ genotypes (Ipek et al., 2012). The goodness of fit between SSR profiles and CDA analysis in this investigation was confirmed by the Mantel test (1967) (r = 0.3; p < 0.001). We performed a test to better understand the unexpected identity in the molecular profiles of the following three cases (Fig. 2): (a) ‘Giarraffa’ 1 and 2 versus ‘Ogliarola Messinese’ STD versus ‘Ogliarola Messinese’ 2; (b) ‘Giarraffa’ 3 versus ‘Moresca’ 10; (c) ‘Giarraffa’ STD versus ‘Nocellara del Belice’ 1; and to verify whether these genetic relationships fitted the observed morphological similarity through the CDA (data not shown). High levels of morphological similarity were found in agreement with the molecular data. In general, it seems possible that some of these genotypes have been wrongly identified and named. Fruits of ‘Giarraffa’, ‘Ogliarola Messinese’ and ‘Moresca’ tend to be large, early ripening, completely black when fully ripened and in Sicily are traditionally used as table olives, generically called “passuluna” (faded fruits) in the local dialect. Therefore, it is likely that the name “Giarraffa” may have been wrongly attributed to other genotypes which fruits can be used, at black ripening stage, as table olives; ‘Giarraffa 1’, ‘Giarraffa 2’ and ‘Ogliarola Messinese STD’, showing the same SSR profiles, could be sibling of ‘Giarraffa STD’, since they shared one allele for each SSR locus. This is also confirmed by the allelic patterns of ‘Giarraffa STD’ and ‘Ogliarola Messinese STD’ found by Marra et al. (2013) for 11 SSR loci, with the exception

T. Caruso et al. / Scientia Horticulturae 180 (2014) 130–138

of the locus Gapu103, where an allele drop out possibly occurred. ‘Moresca 10’ and ‘Giarraffa 3’ could represent further cases of erroneous attribution of the name, although they are genetically close to the ‘Giarraffa’ group, while, ‘Nocellara Belice 1’ may be a sibling of ‘Nocellara Belice STD’ and ‘Giarraffa STD’, which shared morphological similarity with the latter cultivar. The integration of different statistical methods applied to the analysis of morphological and molecular data does not always permit the same conclusions to be reached about the relationships between the cultivars and clones even within the same gene pool (Belaj et al., 2003; Hagidimitriou et al., 2005). From our results, it seemed clear that the combination of the CDA with the cluster analysis of the molecular profiles is the most appropriate and reliable way to study relations among polyclonal cultivars and cultivar-populations and thus to verify intra-cultivar diversity. Clonal selection, proposed since the early 1960s, for olive cultivar improvement (e.g. Serrano et al., 1999; Grati Kamoun et al., 2000; Oueslati et al., 2009; Tous et al., 2011) was confirmed to be an effective strategy which can assist in improving the standard of a cultivar. The results presented are indeed useful for undertaking selection of superior genotypes such as ‘Nocellara del Belice’ clone 10 which showed the largest fruit size and the highest flesh to pit ratio among the ‘Nocellara del Belice’ genotypes and therefore deserves particular attention by the table olive industry. Likewise ‘Biancolilla’ clone 6 has been reported to have adaptive traits to humid environment conditions while, on the contrary, ‘Biancolilla 9’, has been described by local growers as tolerant to salty wind and drought stress, since it is widely spread and cultivated in the windy island of Pantelleria, a small volcanic island in the middle Mediterranean sea where water availability is scarce all the year around and the soil water capacity is very low. ‘Cerasuola’ clone 6 has shortened internodes resulting in a dwarfing and compact vegetative growth habit with dense tree canopy in contrast to the other ‘Cerasuola’ genotypes, so it represents an interesting genotypes that should be exploited for high-density plantings and evaluated for trunk-shaker mechanical harvesting; in addition, the olive oil of this clone has an higher content of oleic acid (79%) respect to that of the STD one (74%), with total polyphenol content of 376 ppm. These genotypes could be also employed for breed new olive varieties for facing future environmental challenges or new market/consumer needs.

5. Conclusion Our research allowed the identification of the most discriminant morphological traits such as shape, width and length of the fruit; shape, area and length of the leaf; fruit weight; length, width and shape of the endocarp, which should be prioritized in the morphological characterization. Consequently, the number of traits to be studied for olive characterization could be reduced without losing important information, saving time, thus decreasing cost of morphological analysis. The combination of traditional cluster analysis (UPGMA) based on microsatellite markers and canonical discriminant analysis (CDA) based on morphological variables, measured on the same environmental conditions, was found powerful for studying relationships among genotypes. The ex-situ collection at the regional repository can be considered the only source of propagation material of Sicilian olive genotypes and standard cultivars, from which a process of nursery certification could be initiated. Traders and experts from the oil and table olive production chain may benefit from genetic and health certification as they can enhance the activity of the selection of local genotypes, guarantee varietal identity and typical productions, improve plant quality, and permit product traceability from plant to food. Furthermore, intra-cultivar variability can be

137

exploited to extend the area of cultivation of the main cultivars in new zones, where environmental conditions are quite different with respect to the site of origin of the standard clones. Results of this work can contribute to a list of the most interesting genotypes showing novel traits which could be tested by local growers in Sicily to select genotypes adapted to the various environmental conditions of the Island. Acknowledgement This research was supported and founded by Dieta Mediterranea e Salute (DI.ME.SA)—Sicilian Cluster on Agro-Industry and Fishery, PON0200451 3361785, Palermo, Italy. We thank Dr. Daniel J. Sargent for English revision. References Albertini, E., Torricelli, R., Bitocchi, E., Raggi, L., Marconi, G., Pollastri, L., Di Minco, G., Battistini, A., Papa, R., Veronesi, F., 2011. Structure of genetic diversity in Olea europaea L. cultivars from central Italy. Mol. Breed. 27, 533–547, http://dx.doi.org/10.1007/s11032-010-9452-y. Baldoni, L., Cultrera, N.G., Mariotti, R., et al., 2009. A consensus list of microsatellite markers for olive genotyping. Mol. Breed. 24, 213–231, http://dx.doi.org/10.1007/s11032-009-9285-8. Barone, E., Di Marco, L., Marra, F.P., Sidari, M., 1996. Isozymes and canonical discriminant analysis to identify pistachio (Pistacia vera L.) germplasm. Hortic. Sci. 31, 134–138. Barranco, D., Cimato, A., Fiorino, P., Rallo, L., Touzani, A., Castaneda, C., Serafin, F., Trujillo, I., 2000. World Catalogue of Olive Varieties. Consejo Oleìcola Internacional, Madrid. Bartolini, G., Prevost, G., Messeri, C., Carignani, G., 2005. Olive Germplasm: Cultivars and World-Wide Collections. FAO, Rome, Italy. Belaj, A., León, L., Satovic, Z., De la Rosa, R., 2011. Variability of wild olives (Olea europaea subsp. europaea var. sylvestris) analyzed by agromorphological traits and SSR markers. Sci. Hortic.—Amsterdam 129, 561–569, http://dx.doi.org/10.1016/j.scienta.2011.04.025. Belaj, A., Satovic, Z., Cipriani, G., et al., 2003. Comparative study of the discriminating capacity of RAPD, AFLP and SSR markers and of their effectiveness in establishing genetic relationships in olive. Theor. Appl. Genet. 107, 736–744, http://dx.doi.org/10.1007/s00122-003-1301-5. Belaj, A., Trujillo, I., De la Rosa, R., Rallo, L., Gimenez, M.J., 2001. Polymorphism and discrimination capacity of randomly amplified polymorphic markers in an olive germplasm bank. J. Am. Soc. Hortic. Sci. 126, 64–71. Besnard, G., Baradat, P., Bervillé, A., 2001. Genetic relationships in the olive (Olea europaea L.) reflects multi-local selection of cultivars. Theor. Appl. Genet. 102, 251–258, http://dx.doi.org/10.1007/s001220051642. Besnard, G., Khadari, B., Navascués, M., et al., 2013. The complex history of the olive tree: from late quaternary diversification of Mediterranean lineages to primary domestication in the northern Levant. Proc. R. Soc. London, Ser. B: Biol. Sci., 280, http://dx.doi.org/10.1098/rspb.2012.2833. Bottari, V., Spina, P., 1952. Le varietà di olivo coltivate in Sicilia. Annali della Sperimentazione Agraria 7, 937–1004. Bracci, T., Busconi, M., Fogher, C., Sebastiani, L., 2011. Molecular studies in olive (Olea europaea L.): overview on DNA markers applications and recent advances in genome analysis. Plant Cell Rep. 30, 449–462, http://dx.doi.org/10.1007/s00299-010-0991-9. Carriero, F., Fontanazza, G., Cellini, F., Giorio, G., 2002. Identification of simple sequence repeats (SSRs) in olive (Olea europaea L.). Theor. Appl. Genet. 104, 301–307, http://dx.doi.org/10.1007/s001220100691. Caruso, G., 1883. Monografia dell’olivo. In: Cantoni, G. (Ed.), Enciclopedia Agraria Italiana. UTET, Torino, pp. 501–680. Caruso, T., Cartabellotta, D., Motisi, A., et al., 2007. Cultivar di Olivo Siciliane. Identificazione, Validazione, Caratterizzazione Morfologica e Molecolare e Qualità degli Oli. Prima Edizione, Regione Siciliana - Assessorato Agricoltura e Foreste, Palermo, Italy. Cipriani, G., Marrazzo, M.T., Marconi, R., Cimato, A., Testolin, R., 2002. Microsatellite markers isolated in olive (Olea europaea L.) are suitable for individual fingerprinting and reveal polymorphism within ancient cultivars. Theor. Appl. Genet. 104, 223–228, http://dx.doi.org/10.1007/s001220100685. COI, 1997. Methodology for primary characterization of olive varieties. In: Project RESGEN-CT (67/97). EU/COI. D’Imperio, M., Viscosi, V., Scarano, M.T., D’Andrea, M., Zullo, B.A., Pilla, F., 2011. Integration between molecular and morphological markers for the exploitation of olive germoplasm (Olea europaea L.). Sci. Hortic.—Amsterdam 130, 229–240, http://dx.doi.org/10.1016/j.scienta.2011.06.050. De la Rosa, R., James, C.M., Tobutt, K.R., 2002. Isolation and characterization of polymorphic microsatellites in olive (Olea europaea L.) and their transferability to other genera in the Oleaceae. Mol. Ecol. Notes 2, 265–267, http://dx.doi.org/10.1046/j.1471-8286.2002.00217.x. Doyle, J.J., Doyle, J.L., 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem. Bull. 19, 11–15.

138

T. Caruso et al. / Scientia Horticulturae 180 (2014) 130–138

Fabbri, A., Lambardi, M., Ozden-Tokatli, Y., 2009. Olive breeding. In: Breeding Plantation Tree Crops: Tropical Species. Springer Science+Business Media, LLC, pp. 423–465. Food and Agriculture Organization of the United Nations, FAOSTAT database, 2013, available at http://faostat3.fao.org/download/Q/QC/E Frane, S., Dunjia, B.M., Slavko, P., Zlatko, C., Zlatko, S., Branka, J., 2010. Genetic variation within the olive (Olea europaea L.) cultivar Oblica detected using amplified fragment length polymorphism (AFLP) markers. Afr. J. Biotechnol. 9, 2880–2883. Gemas, V.J.V., Almadanim, M.C., Tenreiro, R., Martins, A., Fevereiro, P., 2004. Genetic diversity in the Olive tree (Olea europaea L. subsp. europaea) cultivated in Portugal revealed by RAPD and ISSR markers. Genet. Resour. Crop Evol. 51, 501–511, http://dx.doi.org/10.1023/B:GRES. 0000024152.16021.40. Gomes, S., Martins-Lopes, P., Lima-Brito, J., Meirinhos, J., Lopes, J., Martins, A., Guedes-Pinto, H., 2008. Evidence of clonal variation in olive ‘VerdealTransmontana’ cultivar using RAPD, ISSR and SSR markers. J. Hortic. Sci. Biotechnol. 83, 395–400. Grati Kamoun, N., Ayadi, M., Khlif, M., Trigui, A., Karray, B., Rekik, H., Rekik Fakhfakh, B., Hamdi, T., Arous, N., 2000. Pomological and chemical characterization of Tunisia olive tree (Olea europaea L.). In: Proceedings of the Fourth International Olive Growing Symposium, CIHEAM-IAM, Valenzano (Bari), Italy. Hagidimitriou, M., Katsiotis, A., Menexes, G., et al., 2005. Genetic diversity of major Greek olive cultivars using molecular (AFLPs and RAPDs) markers and morphological traits. J. Am. Soc. Hortic. Sci. 130, 211–217. Ipek, A., Barut, E., Gulen Hat, Ipek, M., 2012. Assessment of inter- and intra-cultivar variations in olive using SSR markers. Sci. Agric. 69, 327–335, 10.1590. ISTAT Istituto di Statistica Italiana, 2013. Tav. C27. ISTAT Istituto di Statistica Italiana, Available from http://agri.istat.it/sag is pdwout/jsp/NewDownload.jsp. La Mantia, M., Lain, O., Caruso, T., Testolin, R., 2005. SSR-based DNA, fingerprints reveal the genetic diversity of Sicilian olive (Olea europaea L.) germplasm. J. Hortic. Sci. Biotechnol. 80, 628–632. Liu, K., Muse, S.V., 2005. PowerMarker: Integrated analysis environment for genetic marker data. Bioinformatics 21, 2128–2129, http://dx.doi.org/10.1093/ bioinformatics/bti282. Lopes, M.S., Mendonc¸a, D., Sefc, M.K., Gil, F.S., Da Camara Machado, A., 2004. Genetic evidence of intra-cultivar variability within Iberian olive cultivars. HortScience 39, 1562–1565. Mantel, N., 1967. The detection of disease clustering and a generalized regression approach. Cancer Res. 27, 209–220.

Marra, F.P., Caruso, T., Costa, F., Di Vaio, C., Mafrica, R., Marchese, A., 2013. Genetic relationships, structure and parentage simulation among the olive tree (Olea europaea L. subsp. europaea) cultivated in Southern Italy revealed by SSR markers. Tree Genet. Genomics 9, 916–973, http://dx.doi.org/10.1007/ s11295-013-0609-9. Martins-Lopes, P., Gomes, S., Lima-Brito, J., Lopes, J., Guedes-Pinto, H., 2009. Assessment of clonal genetic variability in Olea europaea L. ‘Cobranc¸osa’ by molecular markers. Sci. Hortic.—Amsterdam 123, 82–89, http://dx.doi.org/10.1016/ j.scienta.2009.08.001. Muzzalupo, I., Chiappetta, A., Benincasa, C., Perri, E., 2010. Intra-cultivar variability of three major olive cultivars grown in different areas of central-southern Italy and studied using microsatellite markers. Sci. Hortic.—Amsterdam 126, 324–329. Nei, M., 1973. Analysis of gene diversity in subdivided populations. Proc. Natl. Acad. Sci. 70, 3321–3323. Osuji, G.A., Onyeagu, S.I., Ekezie, D.D., 2013. Comparison of Jackknife and resubstitution methods in the estimation of error rates in discriminant analysis. Int. J. Math. Stat. Stud. 1, 29–37. Oueslati, I., Taamalli, W., Haddada, F.M., Manai, H., Trigui, A., Zarrouk, M., 2009. Improving the fatty acid composition of ‘Chemlali’ virgin olive oil through clonal selection. J. Hortic. Sci. Biotechnol. 84, 155–160. Rugini, E., De Pace, C., Gutierrez-Pesce, P., Muleo, R., 2011. Olea. In: Kole, C. (Ed.), Wild Crop Relatives: Genomic and Breeding Resources, Temperate Fruits. SpringerVerlag, Berlin, Heidelberg, pp. 79–117. Sefc, K.M., Lopes, M.S., Mendonica, D., Rodrigues Dos Santos, M., Laimer Da Camara Machado, M., Da Camara Machado, A., 2000. Identification of microsatellite loci in olive (Olea europaea L.) and their characterization in Italian and Iberian olive trees. Mol. Ecol. Notes 9, 1171–1173. Serrano, J.F., Leitao, F., Clara, M.I., Amaral, L., Potes, M.F., Serrano, M.C., 1999. Preliminary observations on earliness of flowering and fructification of selected clones of Olea europaea L. Acta Hortic. 474, 167–169. Tous, J., Romero, A., Hermoso, J.F., Ninot, A., 2011. Mediterranean clonal selections evaluated for modern hedgerow olive oil production in Spain. Calif. Agric. 65, 34–40. Zaher, H., Boulouha, B., Baaziz, M., Sikaoui, L., Gaboun, F., Sripada, M., 2011. Morphological and genetic diversity in olive (Olea europaea subsp. europaea L.) clones and varieties. POJ 4, 370–376. Zohary, D., Hopf, M., 1994. Domestication of Plants in the Old World, second ed. Clarendon Press, Oxford.