Euphytica 71: 83-90 (1993)
A comparison of the somaclonal variation level of Rosa hybrida L. cv Meirutral plants regenerated
from callus or direct induction from different vegetative and embryonic tissues

Laurence Arene, Carole Pellegrino & Serge Gudin
Selection Meilland, Domaine de Saint Andre, 83340 Le Cannet des Maures, France
Received 8 October 1992; accepted 8 September 1993


Flowering plants of Rosa hybrida L. cv Meirutral have been obtained either from direct regeneration of adventitious shoots on leaf and root fragments, or through organogenesis and somatic embryogenesis on calli derived from anther, ovule, petal, sepal, receptacle, leaf, stem internode, root and zygotic embryo tissues. The calli derived from floral parts exhibited rhizogenesis. In this case direct induction of adventitious shoots from selected roots had to be performed in order to generate plants. A histological study of the morphogenetic calli was carried out. The plants regenerated directly and those regenerated from calli of leaf, stem internode, root and zygotic embryo tissues, together with reference plants propagated by cuttings, were compared on a phenotypic basis by taking into account petal number, form and colour, and plant growth habit. From these observations, it can be concluded that directly regenerated plants are as stable as reference plants while plants regenerated from callus are unstable, especially those derived from zygotic embryo tissues.

Somaclonal variation

Upon first blooming in the greenhouse, no phenotypic variants were observed with plants produced from either cutting propagation (reference plants) or direct regeneration from leaves and roots. On the contrary, 21.7 and 70% of variation could be observed with plants regenerated from calli of vegetative parts and zygotic embryos, respectively. The recorded variations remained stable and were still observed on 2 successive flushes occuring on plants propagated by cuttings taken on the variants detected during first flush of regenerated plants (5 cuttings per variant). Among the 152 recorded variant plants, 83 significantly differed from the original cultivar by their number of petals (Fig. 5), 16 by their dwarf growth habit (Fig. 6), 4 by their round shaped petals (Fig. 7),12 by their color, 33 by both their growth habit and number of petals, 4 by both their petal number and color. The variant lot differing by the number of petals was represented by 3 plants with 71 ± 5 petals, 8 plants with 25 ± 3 petals, 36 plants with 20 ± 2 petals, 30 plants with 10 ± 2 petals, and 6 plants with 5 petals. The lot differing by the petal color was represented by 6 plants with orange petals, 3 plants with indian pink petals and 3 plants with dark red petals. The 33 plants that differed by both their dwarfness and number of petals all had 10 petals. The 4 plants that differed by both their petal number and color all had 5 orange petals.


Although plant regeneration (through a callus phase or not) had already been reported in Rosa from stem segments (Yalles & Boxus, 1978; Lloyd et al., 1988; Rout et al., 1991; Matthews et al., 1991), leaf (Lloyd et al., 1988; Leffering & Kok, 1990; de Wit et al., 1990; Rout et al., 1991), root fragments (Lloyd et al., 1988; Matthews et al., 1991), zygotic embryos (Burger et al., 1990) and stamen filaments (Noriega & Sondahl, 1991), this is to our knowledge the first time that such direct or indirect regeneration results have been reported in this genus with isolated embryos, petals, sepals, anthers, receptacles and ovules.

Fig. 5. Variants (left and right) varying from the origin cultivar (middle) by petal number.

Fig. 6. Origin cultivar (right) and dwarf variant (left).

The observed differences in frequency of callus development according to the floral parts used or to the development stages of zygotic embryos and floral parts confirm the importance, in a regeneration program, of the donor tissue (Evans, 1989; Compton & Veilleux, 1991), and its age (Burger et al., 1990). Using zygotic embryos, Burger et al. (1990) reported regeneration only through adventitious shoot induction in rose. This difference with the results presented here might be explained by the use, in the present study, of 2,4-D in the induction medium. This auxin is known to have a positive influence on embryogenic (and rhizogenic) callogenesis (Evans et al., 1981; Litz, 1984; Desai et al., 1985; Rapela, 1985). Furthermore, most authors reporting somatic embryogenesis in rose also used 2,4-D (Noriega & Sondhal,1991; Rout et al., 1988).

The histological observations reported here are comparable with those shown by Lloyd et al. (1988) in the case of organogenesis and Rout et al. (1991) in the case of somatic embryogenesis.

The occurrence of variant plants only in the case of regeneration from callus confirms the well known influence of this undifferentiated cell phase on promoting somclonal variability (De Klerk, 1990) and genetic instability (D'Amato, 1985). As far as the degree of variation is concerned, the results obtained with the zygotic embryo originated plants are difficult to compare with those presented by Burger et al. (1990) because, in contrast with us, these authors used hybrid embryos resulting from cross pollinations. Furthermore, they used embryos isolated together with remaining nucellus and endosperm, therefore allowing the possibility of some regeneration from maternal tissue to occur, even though it has recently been demonstrated that with such a material, as suspected by Burger et al. (1990), regeneration originates `from zygotic rather that maternal tissues' (Aly et al., 1992). However, the percentages of plants regenerated from the three crosses they made and that significantly vary in petal number compared to the parent cultivars correspond to 73, 35 and 80% respectively. These percentages are comparable, or even below for one of them, the 70% variation level we report with plants regenerated from true zygotic embryos resulting from self-fertilization. We consider that somaclonal variation is responsible for a part of the observed variation. In contrast with de Wit et al. (1990) and Noriega & Sondahl (1991) some of the plants we obtained from calli developed on vegetative parts were not true to the cultivar donor type (21.7%). However, the former authors mentioned that during the regeneration phase, they selected only `normal looking shoots' for further complete plant production, discarding some other 'abnormal' ones. The latter authors only observed 12 regenerated plants. We consider that it cannot be concluded from these reports that regeneration through a callus phase does not lead to any somaclonal variation in rose; our results and those obtained on pea (Natali & Cavallini, 1987), tomato (Compton & Veilleux, 1991) and sugarcane (Irvine et al., 1991) demonstrate the contrary.

Fig. 7. Origin cultivar petals (left) and round shaped variant petals (right).

Regenerating plants from calli might be attractive to the breeders as the somaclonal variation generated by this system is known to be higher than that generated by other systems such as chemically induced mutagenesis (Gavazzi et al., 1987).