Mutagenesis and Biotechnology Techniques as Tools for Selecting New Stable Diploid and Tetraploid Olive Genotypes and Their Dwarfing Agronomical Characterization

InOlea europaeaL. subsp. europaea, tetraploid genotypes do not exist in nature. Herein, we report the first example of selection of tetraploid olive plants, obtained by combining physical mutagenesis and biotechnology techniques. Stable tetraploid (4n) and diploid (2n) genotypes were isolated in vitro through shoot-tip fragmentation of two mixoploidmutants derived from the gamma irradiation of self-incompatible Leccino and self-compatible Frantoio cultivars. In this study, the stable mutants FRM5-4n, FRM5-2n, LM3-4n, and LM3-2n were characterized in the field for vegetative and reproductive behavior with the aim to use them as varieties or dwarfing rootstocks. The stable 4n genotype of Leccino acquired self-fertility whereas the 4n Frantoio maintained it. A high and constant yield was showed by LM3-2n during 9 years of observation, maintaining the same oil quality as the Leccino wild type (wt). Moreover, the LM3-2n acquired the capacity to be intercompatible with the diploid mutant Leccino dwarf (LD) and with the Leccino wt. This acquired property would allow for a reduction of heterozygosity in the offspring, if crossed with each other for some generations and with the Leccino wt, because it is a sort of self-fertilization. When used as rootstocks, both 4n and 2n Leccino mutants proved to be very effective for reducing the scion size of the high-vigor Canino cultivar, which is well known for its excellent extravirgin oil. Finally, it was demonstrated that the self-grafting of vigorous cultivar caused a reduction in plant size, thus suggesting that it is possible to produce semidwarf plants from vigorous genotypes to consider them in high-density olive orchards. In recent years, in many olive oil producing countries, the research has been focused on the innovation of orchard management with the aims to reduce production costs and to increase yield and oil quality. For these purposes, attempts have been made to introduce intensive or superintensive orchards, or adopt new training systems (Connors et al., 2014; Freixa et al., 2011). In several countries, there has been a rapid increase of highdensity olive cultivation regarding a limited number of mainly traditional cultivars, whereas possible new varieties with suitable characteristics are still under trial in some research institutions (Rallo et al., 2008; Rugini et al., 2016; Vivaldi et al., 2015).Moreover, the cultivars used, such as Arbequina, Koroneiki, and Arbosana, are not considered entirely acceptable by some countries, because the growers prefer to cultivate local varieties yielding high oil quality, although very few of them are suitable for intensive or superintensive plantations, mainly due to the excessive vigor of the plants. The establishment of suitable new genotypes by means of traditional crossbreeding requires a great deal of time. On the contrary, the planting of vigorous local cultivars, grafted on dwarfing rootstocks, could be a feasible strategy to achieve rapid results. However, the availability of dwarfing rootstocks is very limited and their effectiveness is genotype dependent (Fontanazza et al., 1998). It might be possible to select rootstocks from offspring of programmed crosses or in alternative to use varieties, since both do not require waiting a long time to reach the phenotypic stability of the adult phase as the seedling. Until now, little was known about the mechanisms involved in the dwarfing ability of rootstocks. Several hypotheses about the role of physiological stimuli and/or anatomical characters have been made (scion/rootstock disaffinity, water influence, hormonal factors, competition for carbohydrates or nutrients), each supported by experimental data. It was found that no single mechanism influences plant physiology. Rather, the vigor reduction induced by rootstock appears to be the result of a more complex interaction of various factors (Atkinson et al., 2003; Basile et al., 2003; Cohen and Naor, 2002; Gasc o et al., 2007; Nardini et al., 2006; Solari et al., 2006; Soumelidou et al., 1994; Trifil o et al., 2007). Mutagenesis induction is a useful technique for accelerating the genetic improvement of both varieties and rootstocks, but the difficulties observed in isolating stablemutants for vegetative propagation represent a strong deterrent. The first mutants in olive, showing different vegetative habits, were obtained 40 years ago, following the gamma ray irradiation of plants from Ascolana Tenera and Moraiolo cultivars. These mutants showed low agronomic value (Donini and Roselli, 1972) and cytogenetic instability. Subsequently, some mutants with compact phenotype from cultivars Leccino and Frantoio irradiated cuttings were able to reduce plant size when used as rootstock (Pannelli et al., 1990, 1992). Cytogenetic analyses showed that the mutants, named Frantoio Compact (FC) and Leccino Compact (LC), were mixoploids, whereas a mutant showing a dwarf vegetative habit, obtained from cv. Leccino and therefore namedLD, proved to be the only stable diploid mutant (Rugini et al., 1996). The LD mutant produced fruits similar to those of thewild type genotypes, but bloomed at least 1 week later, while both FC and LC produced normal and large-sized fruits, suggesting that they originated from diploid or tetraploid cells, respectively, as validated by Rugini et al. (1996). Recently, Caporali et al. (2014) observed that most of the flowers produced by the mixoploid LC plants are tetraploids, characterized by floral structures larger than in the corresponding diploid plants, due to the increase of the cell size, which also occurred in the fruits. However, polyploidy had little effect on the LC fruit size, because these showed a less elongated shape than in cv. Leccino. Diploid and tetraploid olive plants were isolated from themixoploid LC and FCmutants bymeans of the in vitro shoot-tip fragmentation technique (Rugini et al., 1996). The tetraploid shoots were easily distinguished from the diploid or mixoploid ones, both in in vitro culture and in the greenhouse, at the early growing stages, due to their wider and thicker leaves. It is well known that there are no tetraploid genotypes in the cultivated olive (O. europaea subsp. europaea); in fact, this is diploid with basic chromosome number n = 23, and a nuclear DNA content ranging from 2.3 pg/1C Received for publication 15 Feb. 2016. Accepted for publication 28 Apr. 2016. Most of the funding was provided by the FILAS project ‘‘MIGLIORA’’ of Latium Region. Corresponding author. E-mail: rugini@unitus.it. HORTSCIENCE VOL. 51(7) JULY 2016 799 DNA in cultivars Leccino and Frantoio (Rugini et al., 1996), to 3.90 pg/2C in cv. Dolce Agogia and 4.66 pg/2C in cv. Pendolino (Bitonti et al., 1999). Loureiro et al. (2007) observed a nuclear DNA content ranging from 2.90 to 3.20 pg/2C in wild olive. On the contrary, tetraploid and hexaploid individuals were detected respectively in the subspecies cerasiformis and maroccana of the olive complex O. europaea (Besnard et al., 2008; Rallo et al., 2003). In particular, it was hypothesized that tetraploid cerasiformis could be derived from the hybridization between the ancestors of the subspecies guanchica and europaea (Besnard et al., 2008; Rugini et al., 2011). Polyploidy is a very common condition in the plant kingdom, changing the organization and function of the genome at both genetic and epigenetic level (Comai, 2005). It is well known that polyploidization usually increases the cell dimensions, the fruit size, and sometimes the plants show a better adaptability than their diploid parents. This study focuses on the vegetative and reproductive characterization of the stable tetraploid and diploid mutants obtained from the shoot apex fragmentation of mixoploid genotypes, grown under field conditions, and on the evaluation of their effect on the vigorous cultivar Canino when used as rootstocks. Materials and Methods Plant material. Stable tetraploid mutants Frantoio and Leccino, and diploid Frantoio and Leccino plants, were evaluated in field trials. These genotypes were obtained from the mixoploid genotypes FC and LC after in vitro shoot apex fragmentation, which produced several lines (Rugini et al., 1996). In this study, only plants belonging to the lines FRM5 for Frantoio and LM3 for Leccino were analyzed and compared with Frantoio and Leccino wt. Frantoio and Leccino were used as controls (hereinafter named Frantoio wt and Leccino wt). The tetraploid mutants were named FRM5-4n and LM3-4n, whereas the diploid plants were termed FRM5-2n and LM3-2n. LD, a further diploid mutant characterized by Rugini et al. (1996), was used in some experiments. Semi-hardwood cuttings were collected from diploid and tetraploid adult micropropagated plants FRM5 and LM3 to produce new plants to be used for the experiments. The ploidy level of the twigs used as cuttings was assessed as described below. The new plantlets were transplanted into pots and bred in greenhouse conditions for 2 years, then transferred in the field at the experimental farm of Tuscia University (Viterbo, Central Italy). Furthermore, with the aim to use the mutants as dwarfing rootstocks, both diploid and tetraploid LM3 cuttings were grafted with the cv. Canino, and compared with Canino grafted on Leccino wt, and both ungrafted and selfgrafted Canino. In addition, the self-grafting of LM3-2n and LM3-4n was carried out. Bark graftingwasperformed in the spring, at a height of 20 cm from ground level in 1-year-old potted plants. Similar to the grafted plants, the ungrafted ones were cut at the same height from the ground. The field trial was carried out in the following spring. All the plants transferred in the field, both mutant genotypes and grafting combinations, were cultivated in free shape form with a central axis spaced 2.5 · 4 m and drip irrigated whenever necessary during the summer. At the end of the winter, a slight pruning was performed every year by removing only suckers and v