Plant Cell Reports (1992) 11: 314- 317
Plant Cell Reports 9 Springer-Verlag1992
Somatic embryogenesis in leaves and leaf-derived protoplasts of Actinidia deliciosa var. deliciosa cv. Hayward (kiwifruit) M. Margarida Oliveira and M. Salom~ S. Pais Departamento de Biologia Vegetal, Faculdade de Ci~ncias de Lisboa, Bloco C2, Piso 1, Campo Grande, 1700 Lisboa, Portugal Received February 12, 1992/Revised version received April 2, 1992 - Communicated by I. Potrykus
Summary. The obtention of embryogenic competence in Actinidia deliciosa var. deliciosa cv. Hayward is reported. Axillary buds from shoots submitted to cold (4~ and starvation for 1.5 months, developed leaves with embryogenic competence. These leaves, cultured in darkness for 1.5 months on a medium containing zeatin as a sole growth regulator, originated compact structures from which embryos developed. The plating orientation and sectioning of leaves strongly affected the expression of the embryogenic potential. A selected fraction of the protoplasts isolated from these leaves was able to develop in an embryogenic way. The germination of the embryos is still only occasional. A b b r e v i a t i o n s : 2,4-D - 2,4-dichlorophenoxyacetic acid; 2-iP - 6-dimethylallyl aminopurine; IAA - indole-3acetic acid; NOA - naphthoxyacetic acid; SEM - Scanning Electron Microscopy.
Introduction Actinidia deliciosa is a dioecious woody species original from China and introduced in New Zealand in the beginning of this century. Nowadays kiwifruit is an important crop plant widespread around the world, New Zealand being the larger world producer of this fruit. The increasing interest in kiwifruit has been followed by increased research on this plant. Very few reports exist on the induction of somatic embryogenesis in A. deliciosa. Embryoids were obtained from stem and root segments by Harada (1975) and Bini (1979). BrossardChriqui and Tripathi (1984) using stamen-derived roots and calluses could obtain somatic embryos, as well as shoots. The embryos, however, were unable to develop behyond the torpedo stage. From endosperm tissue of A. chinensis var. chinensis, Zenguang et al. (1983) could develop embryogenic calli and regenerate complete plantlets. Fraser and Harvey (1986) could also regenerate plantlets from embryos, but in this case from callus derived from the anther wall of A. chinensis var. chinensis and A. deliciosa var. deliciosa. Based on the highly Correspondence to: M.M. Oliveira
different embryogenic response obtained with the different plants assayed, these authors suggested a strong influence of the genotype on the regulation of the embryogenic response. In this paper we report the results on the induction of competence for embryogenesis using leaves developed from stressed shoots of A. deliciosa var. deliciosa cv. Hayward.
Material and Methods Plant material. Plantlets of Actinidia deliciosa var. deliciosa cv. Hayward (female cultivar) were propagated in vitro in hormone-free medium (H2) (Pedroso et al. 1992), and used as source material for the experiments. Induction of competent tissue. After removing roots and leaves, the shoots were submitted to one of the following treatments, during 1.5 months: 1) Cold and starvation - Shoots were maintained in sterile Petri dishes, in which a piece of wet filter paper was placed far from the shoots. The Petri dishes were carefully sealed with parafilm and kept inside plastic bags at 4~ 2) Cold stress - Shoots were inoculated in fresh H2 medium and kept at 4~ 3) Nutritional stress - After removing roots and leaves the shoots were sub-cultured on the same old medium. After the stress, the axillary buds were isolated and plated on half strength Murashige and Skoog (1962) medium supplemented with vitamins, 40 g.1-1 glucose and 2 mg.1-1 zeatin (El Glu40). After 1-2 months of culture at 22~ + 2 and a photoperiod of 16 h (26 I.tmol photons. m-2.s-1), the leaves of the regenerated shoots were used for induction of embryogenesis. Embryogenesis To identify the leaves we considered the first leaf as the first one after the apical bud. The leaves were detached and inoculated on E1 medium, containing 20 g.1-1 of glucose or sucrose (El Glu20 or E1 Suc20). Entire or sectioned leaves were plated, with the adaxial or abaxial epidermis facing the medium, and cultured in darkness for
315 1.5 months. The embryos originated were transferred to H2 medium at torpedo or cotyledonar stage, excised or not from the surrounding tissue.
Scanning Electron Microscopy Leaves with different stages of embryogenesis were prepared as previously described (Oliveira and Pals 1991).
Protoplast isolation and culture Leaves originated from buds submitted to cold and starvation were cultured in darkness for 3 days on E1 Suc20 medium and then used for protoplast isolation following the procedure described by Oliveira and Pais (1991). Due to the high amounts of mucilages and raphides that kiwifruit leaves contain, the purification step was adjusted for this particular case. Briefly after filtering through a 100~m sieve, protoplasts were mixed with half volume of W5 salt solution (Menczel et al. 1981), layered on top of a 0.6 M sucrose solution and centrifuged at 165xg for 7 minutes. The pellet, consisting of mesophyll protoplasts, mucilages, raphides and cell debris, was discharged. Fraction 1, from the interface, contained mostly protoplasts with few chloroplasts and numerous cytoplasmic strands, and also some mesophyU protoplasts. This fraction was washed twice in W5 solution, centrifuged at 120xg for 5 minutes and recovered for culture. Protoplasts of fraction 2 were spread in the mixture of enzymes and W5 solution. To this fraction, an equal volume of W5 solution was added and the protoplasts were recovered by centrifugation (120xg, 5 minutes), washed twice in W5 solution and recovered for culture. Protoplasts of both fractions were suspended in PD1 medium and cultured in liquid medium on top of agarose-solidified medium as previously described (Oliveira and Pals 1991). Results
Induction of competent tissue and embryogenesis Cold and nutritional stresses applied independently did not affect the capacity of the stressed buds to develop new shoots in E1 Glu40 medium. However, when shoots were submitted to cold and starvation for 1.5 months, the number of responding buds decreased to 60%, and this value still came down to 20% when the stress was extended to 4 months. The medium used to regenerate plantlets from the stressed buds (El Glu40) was found inadequate for nonstressed buds. In such cases only callus developed and the shoot was unable to grow. In this medium, the buds submitted to a higher stress were those that could develop better. Those submitted both to cold and starvation, could develop a fourth node after 1.5 months of culture. In a first experiment we verified that the leaves developed from buds submitted to cold and starvation showed embryogenic potential. Entire or half sectioned leaves developed embryos on white compact structures located on the veins at the adaxial surface inside the culture medium (Fig. 2). One and a half months after culture initiation, different stages of embryo development could be observed (Fig. 3). Sometimes shoots were found to develop together with embryos (Fig. 3). For the leaves that curled up away from the medium, no embryos nor even the white compact embryogenic structures,
could be found on the leaf portions that were not in contact with the medium. Leaves plated with the abaxial surface towards the culture medium, usually curled or hypertrophied and did not develop those compact structures. Such structures did also not develop when the leaves were cut in small pieces but in this case calluses largely proliferated at the cut surfaces. The results on the embryogenic competence of leaves developed from differently stressed buds are summarized in Fig. 1. ~1 _ % of leaves developing embryogenic structures I I - % of leaves reaching embryo production
100 80 60
4o
,iili!i
Stress
~"
1st
Leaves
Fig. 1- Embryogenic competence of first (lst) second (2rid) and third (3rd) leaves of shoots grown from buds stressed during 1.5 months: C. + S. - Cold (4~ and starvation (absence of culture medium); Cold - 4~ Nutr.Nutritional stress (no renewal of the culture medium). Twelve leaves were used in each assay, and data are mean values from 3 independent assays. Leaves from buds submitted only to cold stress did not produce embryos, although several embryogenic structures could be formed (Fig. 1). The best embryo production was obtained for leaves developed from buds submitted to cold and starvation. Twenty percent of the fourth leaves that were still produced by these shoots, were found to develop embryos. It was also found that the second and the third leaves were more embryogenic then first ones (Fig. 1). Light was a limiting factor for induction of embryogenesis. Leaves cultured under photoperiod could only develop, at the cut surfaces, deep green calluses with spots of cells accumulating anthocyanins. From these structures abnormal shoots developed. The connection of the embryos to the basal tissue was strong. The inoculation of the embryos and the basal tissue on the same fresh medium or on hormone-free medium, often resulted in callus production around the embryo (Fig. 4). Excised embryos (Fig. 5) tended to dedifferentiate upon sub-culture. Embryo germination was observed only occasionally in H2 medium (Fig. 6).
Scanning Electron Microscopy The SEM observations showed that the embryogenic structures seem to originate from cells below the epidermis (Fig. 7). Calluses with a friable appearance were often observed near the compact structures (Fig. 8). Sin-
316
Fig. 2: Embryos developing inside the culture medium on the adaxial leaf surface. Fig. 3: Together with a shoot (arrow), three stages of embryo development are visible a) globular, b) heart, c) torpedo. Fig. 4: Embryo that failed to germinate. Fig. 5: Embryo after excision. Fig. 6: Germination of an embryo which was not excised. Figs. 7-10: Scanning Electron micrographs. Fig. 7: Emergence of embryogenic structures (arrows), from the ruptured adaxial leaf surface. Fig. 8" The compact embryogenic structure (asterisk) often develops together with callus tissue (arrow). Fig. 9" Embryos (arrows) protruding from the embryogenic structure. Fig. 10: Shoot (arrow) and embryos (double arrow) arising from a compact structure. Figs. 11-14: Culture of leaf protoplasts (fraction 2). Fig. 11" First division of a cell with very few chloroplasts. Fig. 12: Colony of c.a. 10 cells with dense cytoplasm. Figs. 13 and 14: Embryo shapes of the colonies regenerated after 2.5 months. Scale bars: lmm for 2-4 and 6-10; 5001.tm for 5, 13 and 14; 501.tm for 11 and 12. ce early stages, embryos and shoots can be clearly distinguished. Embryos have a globular appearance (Fig. 9) different from the conical shape of shoot primordia. Moreover, shoots develop stomata and covering hairs which are never visible on the embryos. At early stages of embryo development (globular to heart) a network of fibrillar strands appears covering the cells. The strong connection of the embryo with the tissue beneath is clearly visible in Fig. 10. Protoplast Culture Two months after culture initiation the protoplasts from fraction 1 had originated colonies with a greenish colour and with variable degrees of friability. Those from fraction 2 started dividing after the first week (Fig. 11)
and progressively increased in density. Successive cell divisions proceeded asymmetrically and without a significant increase in the volume of the colony (Fig. 12). Fifteen days after culture initiation 64% of the protoplasts had died, and from the remaining ones, 43% had divided, these corresponding to 27% of first divisions, 13% of second divisions, 2.3% of third divisions and 0.9% of more then 3 divisions. In these cultures, highly dense whitish structures developed which, 2.5 months after culture initiation, showed an embryo appearance (Figs. 13 and 14). Discussion The state of competence may depend on the level of molecular activity of the genes responsible for embryo-
317 genesis (Carman, 1990). According to Tran Thanh Van (1989, personal communication) to make a plant cell, tissue or organ in culture develop in a certain way, it is necessary to "switch off" the existing physiological programme and to "switch on" the new one. Stress conditions are sometimes found to trigger the embryogenic process. Certain less common culture conditions and certain chemical compounds may affect the mechanisms that regulate the switch of a somatic cell to an embryogenic one. The system we describe here is still far from being optimized. However, it presents some interesting features such as the existence of an intermediate developmental stage (shoot development) between the induction shock (stress applied) and the expression of the embryogenic response (in leaves). This effect shows that there is some information, which is not revealed during shoot development but only expresses on the newly regenerated leaves. Moreover, it seems that there are different factors of induction (the different stresses) leading to different levels of expression. Whatever the information is (genetic, hormonal or other), we found that different leaves show a different sensitivity to the transmitted signal (differences between first, second and third leaves). When kiwifruit leaves coming or not from stressed buds, are inoculated on hormone-free medium (H2) in darkness with the adaxial surface down, they develop a great number of roots, particularly if they have cuttings (Oliveira, unpublished results). The use of zeatin in the culture medium was important for the induction of the embryos, and in this case root development was highly reduced or totally inhibited. It seems probable that kiwifruit leaves have high levels of endogenous auxin and need the exogenous administration of cytokinin to promote embryo development. Although most plants require an exogenous auxin for induction of somatic embryogenesis (Carman 1990), the need of cytokinins has been reported in some cases. In Coffea canephora embryo formation was always inhibited by the presence of auxin, while cytokinin had a beneficial effect in this process (Hatanaka et al. 1991). For A. chinensis var. chinensis endosperm culture, Zenguang et al. (1983) also report best embryo production using zeatin. Harada (1975) using 2,4-D or NOA (0.1-10 mg.1-1) in root and stem cultures, could only obtain globular embryoids which either did not develop further or dedifferentiated. Fraser and Harvey (1986), using 2-iP or zeatin (5 mg.1-1) and IAA (0.3-1.0 mg.1-1) in cultures of anther derived callus of Actinidia, could obtain embryos and shoots. These authors found that IAA was essential for embryogenesis in this material. Moreover, Fraser and Harvey (1986), reported a higher frequency of embryogenesis in one male plant of A. chinensis, and only a male clone of A. deliciosa (cv. Matua) developed a type of callus which could be induced to develop embryos and regenerate plantlets. From the female plant (cv. Hayward), the same authors reported that only organogenic calluses could be obtained. In our experiments, the need to plate the leaves with the adaxial side down, and the total absence of embryoge-
nic development in the regions not touching the medium may indicate that aeration has a negative effect on embryo induction. An identical situation was found for embryo induction in another woody species, Camellia japonica (Pedroso, personal communication). The need to keep the leaves intact or only half sectioned may be related to the development of endogenous gradients of hormones or to the absorption and transport of nutrients and growth regulators. On the other hand, the increased proliferation of callus at the cut surfaces, when leaves are cut in small pieces, was inversely correlated with embryo production, suggesting an inhibitory effect of the callus produced on the cells with embryogenic potential. The isolation of a fraction of leaf protoplasts having an embryogenic behaviour may overcome the inhibitory effect, increasing embryo production and, at the same time, providing fast dividing protoplasts for genetic manipulation experiments. Further studies on the isolation and culture of protoplasm from the different tissues of the leaf may provide precise information on which cell type is competent for embryogenesis. Embryo germination is often difficult to achieve and growth is arrested at the cotyledonary stage (Trtmouilloux-Guiller and Chtnieux 1991). From recent studies on zygotic embryogenesis in dicots, it has been stated that embryogenesis and germination are two independent events, only linked by a switch mechanism (Galau et al. 1991). The second step, germination, is usually only possible when, after the abscision from the mother plant, the embryo goes through a desiccation stage. In our case, a strong connection of the embryos to the subjacent embryogenic or leaf tissue was found, probably accounting for the inhibition of further development.
Acknowledgements: This resarch was partially supported by J.N.I.C.T. (contract 906/BIO/90) and by I.N.I.C. References Bini G (1979) In: Tecniche di colture in vitro, per la propagazione su vasta scalla delle specie ortoflorofrutticole. Pistoia 6 ottobre:211-218. Brossard-Chriqui D, Tripathi BK (1984) Can J Bot 62:1940-1946. Carman JG (1990) In vitro Cell Dev Biol 26:746-753. Fraser LG, Harvey CF (1986) Scientia Hort 29:335-346. Galau GA, Jakobsen KS, Hughes DW (1991) Physiol Plant 81: 280-288. Harada H (1975) J Hort Sci 50:81-83. Hatanaka T, Arakawa O, Yasuda T, Uchida N, YamaguchiT (1991) Plant Cell Rep 10:179-182. Menczel L, Nagy F, Kiss Z, Maliga P (1981) Theor Appl Genet 59:191-195. Murashige T, Skoog F (1962) Physiol Plant 15:473-497. Oliveira MM, Pais MS (1991) Plant Cell Rep 9:643-646. Pedroso MC, Oliveira MM, Pais MS (1992) HortSci 27 (in press). Trtmouilloux-Guiller J, Chdnieux J-C (1991) Plant Cell Rep 10:102-105. Zenguang H, Youli H, Lyin X (1983) Kexue Tongbao 28:112-117.
Plant Cell Reports 9 Springer-Verlag1992
Somatic embryogenesis in leaves and leaf-derived protoplasts of Actinidia deliciosa var. deliciosa cv. Hayward (kiwifruit) M. Margarida Oliveira and M. Salom~ S. Pais Departamento de Biologia Vegetal, Faculdade de Ci~ncias de Lisboa, Bloco C2, Piso 1, Campo Grande, 1700 Lisboa, Portugal Received February 12, 1992/Revised version received April 2, 1992 - Communicated by I. Potrykus
Summary. The obtention of embryogenic competence in Actinidia deliciosa var. deliciosa cv. Hayward is reported. Axillary buds from shoots submitted to cold (4~ and starvation for 1.5 months, developed leaves with embryogenic competence. These leaves, cultured in darkness for 1.5 months on a medium containing zeatin as a sole growth regulator, originated compact structures from which embryos developed. The plating orientation and sectioning of leaves strongly affected the expression of the embryogenic potential. A selected fraction of the protoplasts isolated from these leaves was able to develop in an embryogenic way. The germination of the embryos is still only occasional. A b b r e v i a t i o n s : 2,4-D - 2,4-dichlorophenoxyacetic acid; 2-iP - 6-dimethylallyl aminopurine; IAA - indole-3acetic acid; NOA - naphthoxyacetic acid; SEM - Scanning Electron Microscopy.
Introduction Actinidia deliciosa is a dioecious woody species original from China and introduced in New Zealand in the beginning of this century. Nowadays kiwifruit is an important crop plant widespread around the world, New Zealand being the larger world producer of this fruit. The increasing interest in kiwifruit has been followed by increased research on this plant. Very few reports exist on the induction of somatic embryogenesis in A. deliciosa. Embryoids were obtained from stem and root segments by Harada (1975) and Bini (1979). BrossardChriqui and Tripathi (1984) using stamen-derived roots and calluses could obtain somatic embryos, as well as shoots. The embryos, however, were unable to develop behyond the torpedo stage. From endosperm tissue of A. chinensis var. chinensis, Zenguang et al. (1983) could develop embryogenic calli and regenerate complete plantlets. Fraser and Harvey (1986) could also regenerate plantlets from embryos, but in this case from callus derived from the anther wall of A. chinensis var. chinensis and A. deliciosa var. deliciosa. Based on the highly Correspondence to: M.M. Oliveira
different embryogenic response obtained with the different plants assayed, these authors suggested a strong influence of the genotype on the regulation of the embryogenic response. In this paper we report the results on the induction of competence for embryogenesis using leaves developed from stressed shoots of A. deliciosa var. deliciosa cv. Hayward.
Material and Methods Plant material. Plantlets of Actinidia deliciosa var. deliciosa cv. Hayward (female cultivar) were propagated in vitro in hormone-free medium (H2) (Pedroso et al. 1992), and used as source material for the experiments. Induction of competent tissue. After removing roots and leaves, the shoots were submitted to one of the following treatments, during 1.5 months: 1) Cold and starvation - Shoots were maintained in sterile Petri dishes, in which a piece of wet filter paper was placed far from the shoots. The Petri dishes were carefully sealed with parafilm and kept inside plastic bags at 4~ 2) Cold stress - Shoots were inoculated in fresh H2 medium and kept at 4~ 3) Nutritional stress - After removing roots and leaves the shoots were sub-cultured on the same old medium. After the stress, the axillary buds were isolated and plated on half strength Murashige and Skoog (1962) medium supplemented with vitamins, 40 g.1-1 glucose and 2 mg.1-1 zeatin (El Glu40). After 1-2 months of culture at 22~ + 2 and a photoperiod of 16 h (26 I.tmol photons. m-2.s-1), the leaves of the regenerated shoots were used for induction of embryogenesis. Embryogenesis To identify the leaves we considered the first leaf as the first one after the apical bud. The leaves were detached and inoculated on E1 medium, containing 20 g.1-1 of glucose or sucrose (El Glu20 or E1 Suc20). Entire or sectioned leaves were plated, with the adaxial or abaxial epidermis facing the medium, and cultured in darkness for
315 1.5 months. The embryos originated were transferred to H2 medium at torpedo or cotyledonar stage, excised or not from the surrounding tissue.
Scanning Electron Microscopy Leaves with different stages of embryogenesis were prepared as previously described (Oliveira and Pals 1991).
Protoplast isolation and culture Leaves originated from buds submitted to cold and starvation were cultured in darkness for 3 days on E1 Suc20 medium and then used for protoplast isolation following the procedure described by Oliveira and Pais (1991). Due to the high amounts of mucilages and raphides that kiwifruit leaves contain, the purification step was adjusted for this particular case. Briefly after filtering through a 100~m sieve, protoplasts were mixed with half volume of W5 salt solution (Menczel et al. 1981), layered on top of a 0.6 M sucrose solution and centrifuged at 165xg for 7 minutes. The pellet, consisting of mesophyll protoplasts, mucilages, raphides and cell debris, was discharged. Fraction 1, from the interface, contained mostly protoplasts with few chloroplasts and numerous cytoplasmic strands, and also some mesophyU protoplasts. This fraction was washed twice in W5 solution, centrifuged at 120xg for 5 minutes and recovered for culture. Protoplasts of fraction 2 were spread in the mixture of enzymes and W5 solution. To this fraction, an equal volume of W5 solution was added and the protoplasts were recovered by centrifugation (120xg, 5 minutes), washed twice in W5 solution and recovered for culture. Protoplasts of both fractions were suspended in PD1 medium and cultured in liquid medium on top of agarose-solidified medium as previously described (Oliveira and Pals 1991). Results
Induction of competent tissue and embryogenesis Cold and nutritional stresses applied independently did not affect the capacity of the stressed buds to develop new shoots in E1 Glu40 medium. However, when shoots were submitted to cold and starvation for 1.5 months, the number of responding buds decreased to 60%, and this value still came down to 20% when the stress was extended to 4 months. The medium used to regenerate plantlets from the stressed buds (El Glu40) was found inadequate for nonstressed buds. In such cases only callus developed and the shoot was unable to grow. In this medium, the buds submitted to a higher stress were those that could develop better. Those submitted both to cold and starvation, could develop a fourth node after 1.5 months of culture. In a first experiment we verified that the leaves developed from buds submitted to cold and starvation showed embryogenic potential. Entire or half sectioned leaves developed embryos on white compact structures located on the veins at the adaxial surface inside the culture medium (Fig. 2). One and a half months after culture initiation, different stages of embryo development could be observed (Fig. 3). Sometimes shoots were found to develop together with embryos (Fig. 3). For the leaves that curled up away from the medium, no embryos nor even the white compact embryogenic structures,
could be found on the leaf portions that were not in contact with the medium. Leaves plated with the abaxial surface towards the culture medium, usually curled or hypertrophied and did not develop those compact structures. Such structures did also not develop when the leaves were cut in small pieces but in this case calluses largely proliferated at the cut surfaces. The results on the embryogenic competence of leaves developed from differently stressed buds are summarized in Fig. 1. ~1 _ % of leaves developing embryogenic structures I I - % of leaves reaching embryo production
100 80 60
4o
,iili!i
Stress
~"
1st
Leaves
Fig. 1- Embryogenic competence of first (lst) second (2rid) and third (3rd) leaves of shoots grown from buds stressed during 1.5 months: C. + S. - Cold (4~ and starvation (absence of culture medium); Cold - 4~ Nutr.Nutritional stress (no renewal of the culture medium). Twelve leaves were used in each assay, and data are mean values from 3 independent assays. Leaves from buds submitted only to cold stress did not produce embryos, although several embryogenic structures could be formed (Fig. 1). The best embryo production was obtained for leaves developed from buds submitted to cold and starvation. Twenty percent of the fourth leaves that were still produced by these shoots, were found to develop embryos. It was also found that the second and the third leaves were more embryogenic then first ones (Fig. 1). Light was a limiting factor for induction of embryogenesis. Leaves cultured under photoperiod could only develop, at the cut surfaces, deep green calluses with spots of cells accumulating anthocyanins. From these structures abnormal shoots developed. The connection of the embryos to the basal tissue was strong. The inoculation of the embryos and the basal tissue on the same fresh medium or on hormone-free medium, often resulted in callus production around the embryo (Fig. 4). Excised embryos (Fig. 5) tended to dedifferentiate upon sub-culture. Embryo germination was observed only occasionally in H2 medium (Fig. 6).
Scanning Electron Microscopy The SEM observations showed that the embryogenic structures seem to originate from cells below the epidermis (Fig. 7). Calluses with a friable appearance were often observed near the compact structures (Fig. 8). Sin-
316
Fig. 2: Embryos developing inside the culture medium on the adaxial leaf surface. Fig. 3: Together with a shoot (arrow), three stages of embryo development are visible a) globular, b) heart, c) torpedo. Fig. 4: Embryo that failed to germinate. Fig. 5: Embryo after excision. Fig. 6: Germination of an embryo which was not excised. Figs. 7-10: Scanning Electron micrographs. Fig. 7: Emergence of embryogenic structures (arrows), from the ruptured adaxial leaf surface. Fig. 8" The compact embryogenic structure (asterisk) often develops together with callus tissue (arrow). Fig. 9" Embryos (arrows) protruding from the embryogenic structure. Fig. 10: Shoot (arrow) and embryos (double arrow) arising from a compact structure. Figs. 11-14: Culture of leaf protoplasts (fraction 2). Fig. 11" First division of a cell with very few chloroplasts. Fig. 12: Colony of c.a. 10 cells with dense cytoplasm. Figs. 13 and 14: Embryo shapes of the colonies regenerated after 2.5 months. Scale bars: lmm for 2-4 and 6-10; 5001.tm for 5, 13 and 14; 501.tm for 11 and 12. ce early stages, embryos and shoots can be clearly distinguished. Embryos have a globular appearance (Fig. 9) different from the conical shape of shoot primordia. Moreover, shoots develop stomata and covering hairs which are never visible on the embryos. At early stages of embryo development (globular to heart) a network of fibrillar strands appears covering the cells. The strong connection of the embryo with the tissue beneath is clearly visible in Fig. 10. Protoplast Culture Two months after culture initiation the protoplasts from fraction 1 had originated colonies with a greenish colour and with variable degrees of friability. Those from fraction 2 started dividing after the first week (Fig. 11)
and progressively increased in density. Successive cell divisions proceeded asymmetrically and without a significant increase in the volume of the colony (Fig. 12). Fifteen days after culture initiation 64% of the protoplasts had died, and from the remaining ones, 43% had divided, these corresponding to 27% of first divisions, 13% of second divisions, 2.3% of third divisions and 0.9% of more then 3 divisions. In these cultures, highly dense whitish structures developed which, 2.5 months after culture initiation, showed an embryo appearance (Figs. 13 and 14). Discussion The state of competence may depend on the level of molecular activity of the genes responsible for embryo-
317 genesis (Carman, 1990). According to Tran Thanh Van (1989, personal communication) to make a plant cell, tissue or organ in culture develop in a certain way, it is necessary to "switch off" the existing physiological programme and to "switch on" the new one. Stress conditions are sometimes found to trigger the embryogenic process. Certain less common culture conditions and certain chemical compounds may affect the mechanisms that regulate the switch of a somatic cell to an embryogenic one. The system we describe here is still far from being optimized. However, it presents some interesting features such as the existence of an intermediate developmental stage (shoot development) between the induction shock (stress applied) and the expression of the embryogenic response (in leaves). This effect shows that there is some information, which is not revealed during shoot development but only expresses on the newly regenerated leaves. Moreover, it seems that there are different factors of induction (the different stresses) leading to different levels of expression. Whatever the information is (genetic, hormonal or other), we found that different leaves show a different sensitivity to the transmitted signal (differences between first, second and third leaves). When kiwifruit leaves coming or not from stressed buds, are inoculated on hormone-free medium (H2) in darkness with the adaxial surface down, they develop a great number of roots, particularly if they have cuttings (Oliveira, unpublished results). The use of zeatin in the culture medium was important for the induction of the embryos, and in this case root development was highly reduced or totally inhibited. It seems probable that kiwifruit leaves have high levels of endogenous auxin and need the exogenous administration of cytokinin to promote embryo development. Although most plants require an exogenous auxin for induction of somatic embryogenesis (Carman 1990), the need of cytokinins has been reported in some cases. In Coffea canephora embryo formation was always inhibited by the presence of auxin, while cytokinin had a beneficial effect in this process (Hatanaka et al. 1991). For A. chinensis var. chinensis endosperm culture, Zenguang et al. (1983) also report best embryo production using zeatin. Harada (1975) using 2,4-D or NOA (0.1-10 mg.1-1) in root and stem cultures, could only obtain globular embryoids which either did not develop further or dedifferentiated. Fraser and Harvey (1986), using 2-iP or zeatin (5 mg.1-1) and IAA (0.3-1.0 mg.1-1) in cultures of anther derived callus of Actinidia, could obtain embryos and shoots. These authors found that IAA was essential for embryogenesis in this material. Moreover, Fraser and Harvey (1986), reported a higher frequency of embryogenesis in one male plant of A. chinensis, and only a male clone of A. deliciosa (cv. Matua) developed a type of callus which could be induced to develop embryos and regenerate plantlets. From the female plant (cv. Hayward), the same authors reported that only organogenic calluses could be obtained. In our experiments, the need to plate the leaves with the adaxial side down, and the total absence of embryoge-
nic development in the regions not touching the medium may indicate that aeration has a negative effect on embryo induction. An identical situation was found for embryo induction in another woody species, Camellia japonica (Pedroso, personal communication). The need to keep the leaves intact or only half sectioned may be related to the development of endogenous gradients of hormones or to the absorption and transport of nutrients and growth regulators. On the other hand, the increased proliferation of callus at the cut surfaces, when leaves are cut in small pieces, was inversely correlated with embryo production, suggesting an inhibitory effect of the callus produced on the cells with embryogenic potential. The isolation of a fraction of leaf protoplasts having an embryogenic behaviour may overcome the inhibitory effect, increasing embryo production and, at the same time, providing fast dividing protoplasts for genetic manipulation experiments. Further studies on the isolation and culture of protoplasm from the different tissues of the leaf may provide precise information on which cell type is competent for embryogenesis. Embryo germination is often difficult to achieve and growth is arrested at the cotyledonary stage (Trtmouilloux-Guiller and Chtnieux 1991). From recent studies on zygotic embryogenesis in dicots, it has been stated that embryogenesis and germination are two independent events, only linked by a switch mechanism (Galau et al. 1991). The second step, germination, is usually only possible when, after the abscision from the mother plant, the embryo goes through a desiccation stage. In our case, a strong connection of the embryos to the subjacent embryogenic or leaf tissue was found, probably accounting for the inhibition of further development.
Acknowledgements: This resarch was partially supported by J.N.I.C.T. (contract 906/BIO/90) and by I.N.I.C. References Bini G (1979) In: Tecniche di colture in vitro, per la propagazione su vasta scalla delle specie ortoflorofrutticole. Pistoia 6 ottobre:211-218. Brossard-Chriqui D, Tripathi BK (1984) Can J Bot 62:1940-1946. Carman JG (1990) In vitro Cell Dev Biol 26:746-753. Fraser LG, Harvey CF (1986) Scientia Hort 29:335-346. Galau GA, Jakobsen KS, Hughes DW (1991) Physiol Plant 81: 280-288. Harada H (1975) J Hort Sci 50:81-83. Hatanaka T, Arakawa O, Yasuda T, Uchida N, YamaguchiT (1991) Plant Cell Rep 10:179-182. Menczel L, Nagy F, Kiss Z, Maliga P (1981) Theor Appl Genet 59:191-195. Murashige T, Skoog F (1962) Physiol Plant 15:473-497. Oliveira MM, Pais MS (1991) Plant Cell Rep 9:643-646. Pedroso MC, Oliveira MM, Pais MS (1992) HortSci 27 (in press). Trtmouilloux-Guiller J, Chdnieux J-C (1991) Plant Cell Rep 10:102-105. Zenguang H, Youli H, Lyin X (1983) Kexue Tongbao 28:112-117.
