Root formation from transgenic calli of Ginkgo biloba

Transcription

1 Tree Physiology 23, Heron Publishing Victoria, Canada Root formation from transgenic calli of Ginkgo biloba RADIA AYADI 1 AND JOCELYNE TRÉMOUILLAUX-GUILLER 1,2 1 Laboratoire de Biologie Moléculaire et de Biochimie Végétale, (EA 2106) -Université François Rabelais, Faculté des Sciences Pharmaceutiques, 31 Avenue Monge, F Tours, France 2 Author to whom correspondence should be addressed (guiller@univ-tours.fr) Received July 8, 2002; accepted January 17, 2003; published online June 2, 2003 Summary The objective of this study was to produce Ginkgo biloba L. hairy roots for future investigations into the feasibility of producing terpenoids in differentiated cell cultures. Zygotic embryos of G. biloba were inoculated with the wild agropine-type Agrobacterium rhizogenes strain A4. Three months after bacterial infection, primordia-like nodular structures formed at the root wound sites and developed into tiny calli. Calli cultivated on hormone-free Lloyd and McCown (1980) solid medium were transferred onto Murashige and Skoog (1962) solid medium, and grew rapidly for the first few months. As browning appeared and growth slowed, calli were transferred onto Ball (1959) or White (1954) solid media. These calli became nodular with several nodules from which roots, displaying characteristic features of hairy roots, developed. Transgenic calli cultivated in agitated, hormonefree liquid media led to the formation of root meristems and root tips by a cyclic development process. Stable integration of the rola, rolb and rolc genes into calli, mature roots and root meristem and root tip mixtures was confirmed by polymerase chain reaction (PCR). The absence of a 437 bp amplificate corresponding to the vird1 gene confirmed that there was no bacterial contamination of G. biloba tissues. The reverse transcription-pcr method was used to verify the expression of rola and rolc genes in mature roots. Expression of rola, rolb and rolc in root meristems and root tips was confirmed by 3 RACE-PCR analysis, which excluded amplification of possible rol gene transcripts produced by residual bacteria. This paper shows, for the first time, the feasibility of developing G. biloba roots from transformed calli. Keywords: Agrobacterium rhizogenes, hairy root, PCR, 3 RACE-PCR, rolabc genes, RT-PCR, transgenic roots, zygotic embryos. Introduction Ginkgo biloba L. is a medicinal plant species recommended for the treatment of certain aging disorders. Unique diterpenes detected in leaves and root bark are at least partially responsible for the pharmaceutical properties of this tree. However, the use of undifferentiated cell cultures for the production of secondary metabolites has been problematic because phytochemical biosynthesis is often restricted to certain organs or specialized tissues. Agrobacterium rhizogenes, a plant pathogenic bacterium, can induce the growth of adventitious roots with root hairs (so-called hairy root syndrome) (Willmitzer et al. 1982) that accumulate all the secondary metabolites detected in roots of the mother plant, at similar (Christen et al. 1992) or even at higher concentrations (Ahn et al. 1996). Agrobacterium rhizogenes strain A4 harbors an agropine-type Ri plasmid, possessing two separate regions carrying the rol genes (TL-DNA) (Bercetche et al. 1987) and aux + ags genes (TR-DNA) (Camilleri and Jouanin 1991), which can be transferred independently into the plant genome. Little information is currently available about the development of G. biloba transgenic roots. Laurain et al. (1997) used A. rhizogenes strain A4 to initiate a putatively transformed diterpene-producing cell suspension. Likewise, Balz et al. (1999) described the formation of transformed stem-derived calli of G. biloba; however, no rooting was observed. Shunan et al. (1997) reported hairy root production in G. biloba, but DNA integration was not confirmed by molecular analysis. Our main objective was to obtain organized tissue cultures, particularly transgenic roots, capable of accumulating high terpenoid contents. In this study, the feasibility of developing G. biloba roots from transformed embryo-derived calli was demonstrated by stable integration and expression of rola, rolb and rolc genes from A. rhizogenes. Materials and methods Bacterial strain Agrobacterium rhizogenes strain A4 was grown at 28 C on LPGA solid medium as described by Laurain et al. (1997). Prior to inoculation, the Agrobacteria were introduced into LPGA liquid medium and placed on a rotary shaker at 100 rpm for 24 h in the dark. Plant materials Ginkgo biloba ovules were harvested from a tree in the Botanical Garden of Tours and surface-sterilized (Laurain et al. 1997). Zygotic embryos were extracted from sterile ovules and plated for days on Murashige and Skoog (1962)

2 714 AYADI AND TRÉMOUILLAUX-GUILLER solid medium (MS) in preparation for transformation. As a negative control, roots were developed on Ball (1959) solid medium from zygotic embryos that would not be transformed. Hairy root cultures of Daucus carota L. Ann. were established from mature carrot roots according to the method described by Bercetche et al. (1987) to check the virulence of the bacterial strain and to serve as a positive control in the PCR, RT-PCR and 3 RACE-PCR screenings. Transformation of Ginkgo biloba embryos Cotyledons, hypocotyls and roots of zygotic embryos at the cotyledonary stage were wounded with a sterile needle. Four sets of experiments (I IV) were performed to test the possible effects of embryo size (6 ± 1, 7.4 ± 2, 13.5 ± 2 and 15.7 ± 2.0 mm in length) on transformation efficiency. For each experimental set, 12 embryos were immersed into a 24-h-old A. rhizogenes suspension in LPGA medium (absorbance value of at 600 nm) for 4 (set II) or 5 h (sets I, III and IV) at 28 C under dark conditions. Twelve control embryos were dipped into A. rhizogenes-free LPGA medium. Infected and noninfected zygotic embryos were washed with a 0.5 g l 1 cefotaxime solution (Claforan, Uclaf-Roussel, Paris, France) to remove bacteria and transferred onto hormone-free woody plant medium (WPM) (Lloyd and McCown 1980) (Kalys Biotechnologies, Roubaix, France). The WPM was supplemented with WPM vitamins, 20 g l 1 sucrose (Prolabo, Briare le canal, CE), 1 g l 1 activated charcoal (Kalys Biotechnologies), 7 g l 1 agar HP 696 (Kalys Biotechnologies) at ph 5.8, and after autoclaving, supplemented with (or without) 500 mg l 1 filter-sterilized cefotaxime. After culturing for 6 months, calli were transferred onto MS solid medium devoid of growth regulators, but containing vitamins and antibiotics, for 12 months. Bacteria-free subcultures derived from a fastgrowing callus were grown at 24 C in the dark or in a 12 h light/12 hdark photoperiod, alternating Ball (1959) solid medium containing Gamborg s B5 vitamins (Gamborg et al. 1976), 20 g l 1 sucrose, 250 mg l 1 antioxidant mixture (Sigma, St. Louis, MO) and 300 mg l 1 L-glutamine (Kalys Biotechnologies) with White (1954) solid medium containing Gamborg s B5 vitamins, 20 g l 1 sucrose, 350 mg l 1 casein hydrolysate (Sigma), 250 mg l 1 yeast extract (Sigma) and 5 g l 1 antivitrifying agent EM2 (Sigma). Root meristem and root tip mixtures For the first transfer from solid to liquid medium, approximately 100 to 200 mg of calli were inoculated into 100-ml Erlenmeyer flasks containing 25 ml of agitated (100 rpm) antibiotic-free liquid medium: White (W, W/2) or Ball (B/2) medium at full or half-strength containing Gamborg s B5 vitamins and 20 g l 1 sucrose, or MS medium contained mineral salts at full (MS), half (MS/2) or one-third (MS/3) strength with the addition of MS vitamins, 20 g l 1 sucrose and 300 mg l 1 L-glutamine (MS/2 only). All liquid cultures were stored in the dark at 24 ± 1 C. At least three replicates were performed for each medium. The Ball, White and MS/2 containing L-glutamine media were tested over a period of approximately 12 months. Molecular analysis by polymerase chain reaction (PCR) Plant samples were quickly frozen in liquid nitrogen and ground to a fine powder. Ginkgo biloba DNA was extracted from approximately 500 mg of transformed embryo-derived calli, transgenic roots and uninfected roots (negative control) using the DNeasy Plant Minikit (Qiagen, Hilden, Germany) according to the manufacturer s instructions. The DNA of D. carota was extracted from hairy roots (positive control). The PCR mixture consisted of 150 ng of plant DNA, 6 µl of 25 mm MgCl 2, 0.5 µl of Taq polymerase (5 U µl 1 ) (Promega, Madison, WI), 5 µl of 10 buffer, 1 µl of 10 mm dntps, 1 µl each of 10 µm rola 1 A 2 or rolb 1 B 2 or rolc 1 C 2 primers and was adjusted to a final volume of 50 µl with sterile water. The primers, synthesized by MWG-Biotech (Courtaboeuf, France), were: rola (5 -CAGAATGGAATTAGCCGGA CTA-3 ) and rola (5 -ΤΤΑΑΤCCCGTAGGTTTGTTTC G-3 ) for amplification of the 307 bp rola gene; rolb 1 61 (5 -ATGGAT CCCAAATTGCTATTCC-3 ) and rolb (5 -GTTTACTG CAGCAGGCTTCATG-3 ) for amplification of the 762 bp rolb gene; and rolc 1 18 (5 -ATGGC TGAAGACGACCTGTGT-3 ) and rolc (5 -GCCGAT TGCAAACTTGCACT C-3 ) for amplification of the 539 bp rolc gene. The rola, rolb and rolc genes were independently amplified from all samples. A control for bacterial contamination was carried out by PCR of vird1 with primers vird1s (5 -ATGTCGCAAGGCAGTAAGCCCA-3 ) and vird1as (5 -GGAGTCTTTCAGCA TGCAGCA-3 ), resulting in a 437 bp amplificate. Amplification was carried out in an i-cycler TM (Bio-Rad, Philadelphia, PA) and consisted of a DNA pre-denaturation step at 94 C for 2 min, and 40 cycles of 1 min at 94 C (denaturation), 1 min at 62 C (annealing) and 1 min at 72 C (extension), followed by 7 min at 72 C (final extension). The PCR products (25 µl) were separated on a 1.5% agarose gel (Promega) and detected by staining with 0.4 µg l 1 ethidium bromide (Promega) and exposure to UV light (300 nm). Amplification of specific mrnas by reverse transcriptionpolymerase chain reaction (RT-PCR) Ginkgo biloba rola, rolb and rolc transcripts were isolated from mature transformed roots with an RNeasy Total RNA Plant Minikit (Qiagen) according to the manufacturer s instructions. Total RNA from D. carota hairy root (positive control) were extracted using the same method. First-strand cdna synthesis was performed with the gene-specific antisense primers (rola 2, rolb 2 and rolc 2 ) synthesized by reverse transcriptase activity. Second-strand cdna was amplified by the PCR process. Reverse transcription (RT) reactions were carried out in a final volume of 50 µl containing 1 µg of total RNA, 2 µl of 25 mm MgSO 4, 0.5 µl of AMV-reverse transcriptase (2.5 U), 1 µl of RNAsine, 10 µl of AMV-RT buffer, 1 µl of 10 mm dntp (Promega) and 5 µl of 10 µm of each antisense primer. In the thermocycler, the RNA samples were heated at 48 C for 45 min (for cdna synthesis from specific mrnas) and 94 C for 2 min (for inactivation of AMV-RT and RNA/cDNA primer denaturation) and then kept at 4 C. TREE PHYSIOLOGY VOLUME 23, 2003

3 PRODUCING GINKGO BILOBA HAIRY ROOTS IN VITRO 715 Isolation and amplification of specific poly(a) mrnas by 3 RACE-PCR The 3 RACE kit (GIBCO-BRL, Carlsbad, CA) was used to isolate total RNA from transformed root meristem and root tip mixtures of G. biloba and from D. carota hairy roots. The cdnas of mrnas with a poly(a) tail were made by the nonspecific retro-transcription effect, with an adapter primer containing an oligo-dt sequence (AP) and an adapter primer in the presence of transcriptase inverse (AUAP). The PCR amplification of specific cdnas was performed with rola, rolb and rolc primers (sense and antisense), and DNA bands were visualized on an agarose gel. Results and discussion Development of transgenic calli Twelve embryos infected with wild A. rhizogenes strain A4 per experimental set (48 total) and 12 uninfected wounded embryos were cultured on hormone-free WPM solid medium. Slight cellular proliferations that led to 1 2 nodular structures resembling root primordia, similar to those observed by Phelep et al. (1991) during the transformation of Allocasuarina verticillata Lam. explants with strain A4, appeared at wound sites for about one-twelfth of the infected G. biloba embryos cultivated for 3 4 months (Table 1). Embryo size did not influence transformation efficiency. Each root primordium transferred onto hormone-free WPM solid medium evolved into a tiny callus, whereas the uninoculated embryos exhibited no cellular proliferation and ultimately died. One fast-growing callus was selected to check transformation status and root development on solid and in liquid media. alternated on Ball solid medium (established for G. biloba root development) and White solid medium (largely used to cultivate G. biloba tissues), modified by the addition of various substances that decrease browning. Calli were subcultured on Ball medium every 30 days for 3 months and then on White medium for 2 3 months. On Ball medium, roots displaying typical features of the hairy root syndrome (hormone independence, plagiotropic growth and presence of lateral branching) (Figures 1A C) developed from white nodules. Some roots attained a length of 4 to 5 cm within 2 weeks (Figures 1B and 1C). A 2 3 month subculture period requirement has been reported for the development of Picea abies (L.) Karst. (Attree et al. 1989) and Hevea brasiliensis Muell. Arg. (Montoro et al. 1993) embryos from aging calli. Initiation of the root meristem and root tip mixture Transfer of calli to hormone-free liquid media resulted in the cyclic development of root meristems and root tips (Figure 2). The MS/2 (with or without L-glutamine), Ball and White media had similar rhizogenesis efficiencies. However, short roots stopped growing when they reached a length of 20 mm (Figures 3A and 3B). We observed that G. biloba transgenic roots generally grew slowly and became brown regardless of the type of medium. Because of their high phenolic content, woody species possess low rooting capacity, a reduced hairy root phenotype and develop only a few lateral branchings or Development of transgenic roots Calli were initially cultivated on hormone-free WPM medium, then transferred onto MS, a rich saline medium conducive to generating G. biloba zygotic embryo-derived biomass (Carrier et al. 1990). The callus cultures grew rapidly for several months, then growth rate declined progressively. To initiate roots or at least palliate possible necrosis, calli were Table 1. Formation of transformed zygotic embryo-derived calli of Ginkgo biloba after 3 4 months of culture. Four sets of experiments (I IV) were performed. The bacterial infection was mediated by Agrobacterium rhizogenes strain A4. In total, 60 embryos at the early cotyledonary stage were used (48 infected embryos, i.e., 12 per experimental set and 12 uninfected control embryos). Experimental Zygotic embryo Co-culture No. of calli after sets (at cotyledonary period 3 4 months of stage) length (mm) (h) culture I 6.0 ± II 7.4 ± III 13.5 ± IV 15.7 ± Control embryos 10.7 ± Figure 1. Development of Ginkgo biloba transgenic roots from embryo-derived calli cultured on Ball solid medium after transformation by Agrobacterium rhizogenes strain A4. (A) Root emerging from a small callus. (B) First lateral branching (order 2 root) on a long root approximately 4 cm in length. (C) Elongation of the first lateral branching. TREE PHYSIOLOGY ONLINE at

4 716 AYADI AND TRÉMOUILLAUX-GUILLER Figure 2. Cycle of root development from transformed calli of Ginkgo biloba cultured in agitated hormone-free liquid medium: (1) cell aggregate formation, (2) release of single cells into the medium, (3) formation of white meristem roots (estimated at 1 2 mm in diameter), (4) initiation of root tips (2 8 mm), (5) formation of short (maximum 20 mm in length) well-developed roots and (6) the release of single cells. none at all, as reported in Larix ssp. (McAfee et al. 1993) or Pinus nigra Arnold (Mihaljevic et al. 1996). Figure 3. Well-developed transformed roots of Ginkgo biloba appearing in agitated, hormone-free liquid medium from small cell aggregates. (A, B) Root length varied from 2 8 mm (called root tips) to 20 mm (short roots). Stable genetic transformation of calli and roots by bacterial rol gene integration Stable integration of the rola, rolb and rolc genes into G. biloba zygotic embryo-derived tissues was confirmed by PCR. Bands of amplified DNA of 307 (rola), 762 (rolb) and 539 (rolc) bp from genomic DNA extracts of calli (Figures 4a, 4c and 4e), mature roots (Figures 4b, 4d and 4f) and root meristem and root tip mixtures (data not shown) were observed. Likewise, the same sized bands were amplified from D. carota hairy roots (positive control) (Figures 4a f) and were absent in DNA extracted from G. biloba uninfected roots (negative control). The lack of a 437 bp band corresponding to Figure 4. Stable integration of rola, rolb and rolc genes was determined by PCR analysis of genomic DNA extracted from transformed Ginkgo biloba cultures. Detection of rola (307 bp), rolb (762 bp) and rolc (539 bp) in transformed calli (TC) (a, c, e) and in transformed mature roots (TMR) (b, d, f). (g) PCR analysis of the rola gene (307 bp) in Daucus carota hairy root extracts (C +, positive control), G. biloba untransformed roots (C, negative control), TC and TMR. (h) PCR analysis of vird1 (437 bp) in Agrobacterium rhizogenes (A4), TC and transformed roots (TR) of G. biloba. Abbreviation: M = molecular marker (DNA molecuqlar weight standard, 100 bp ladder, Promega). TREE PHYSIOLOGY VOLUME 23, 2003

5 PRODUCING GINKGO BILOBA HAIRY ROOTS IN VITRO 717 bacterial vird1 in transformed calli and transformed roots indicates that the G. biloba transformed tissues were free of bacterial genomic contamination (Figure 4h). RolA, rolb and rolc gene expression in mature roots The absence of genomic DNA contamination (corresponding to rol gene fragments) in G. biloba RNA extracts was verified by PCR before an RT-PCR analysis was performed. The 307 and 539 bp bands were present, confirming the expression of rola and rolc genes (Figures 5a and 5c). A 762 bp band corresponding to the rolb transcript was not seen, although all three genes were expressed in extracts of D. carota hairy root (positive control) (Figures 5a c). One explanation for the absence of rolb transcripts is that transcription was blocked by the methylation of foreign genes. However, this is unlikely as rolb is considered to be the most efficient root-promoting gene in woody species. The rola and rolc genes are known to support root growth and are preferentially expressed in elongated roots (Jasik et al. 1997), whereas Altamura et al. (1994) and Di Cola et al. (1996) showed that the rolb gene acts mainly as a morphogen. Therefore, in our transformed mature roots, rolb mrnas may have declined or been destroyed while rola and rolc transcripts were still being synthesized. Rol gene expression in root meristems and root tips Amplification of specific cdnas by PCR with rola, rolb and rolc primers from transformed root meristem and root tip transcripts with a poly(a) tail, enabled us to reveal the presence of the 307, 762 and 539 bp bands (Figure 6). The expression of rola, rolb and rolc genes in G. biloba root meristems and root tips was determined by 3 RACE-PCR, which excluded the amplification of bacterial RNA. In agreement with our previous hypothesis, rola, rolb and rolc genes were expressed in the G. biloba root meristem and root tip mixtures. These results suggest that the rolb gene was involved in the beginning of root development as reported by Welander et al. (1998). Conclusions Transformed callus-derived roots of G. biloba (confirmed by stable integration and expression of rola, rolb and rolc genes) were successfully generated. Transgenic short roots may be good tools for investigating the production of pharmaceutically active terpene-derived ginkgolides. Whereas some studies have produced terpenoids from undifferentiated G. biloba cells, ginkgolides and bilobalide were undetectable or were present in low concentrations (Jeon et al. 1995). However, HPLC analysis revealed that differentiated cell cultures accumulated terpenoids at concentrations similar to those found in leaves of mature trees (Laurain et al. 1997), suggesting that detection of ginkgolides in short roots should be possible. As well, the cyclic development of root meristems and root tips opens opportunities for fundamental studies of the rol gene and promoter expression. Acknowledgments We are grateful to Prof. Dr. Marc Rideau for valuable advice, and Evelyne Danos, Claude Girard and Robert Nadreau for their technical assistance. Figure 5. Expression of rola (307 bp), rolb (762 bp) and rolc (539 bp) genes was determined by RT-PCR analysis of RNA samples extracted from Ginkgo biloba transformed mature roots (TMR) and Daucus carota hairy root (C +, positive control). Specific cdna strands were synthesized from total RNA, then amplified using specific primers. Definition: M = molecular marker (DNA molecular weight standard, 100 bp ladder, Promega). Figure 6. Expression of rola, rolb and rolc genes was determined by 3 RACE-PCR analysis of RNA samples extracted from Ginkgo biloba transformed root meristems and root tips in liquid cultures (TRT) and Daucus carota hairy roots (C +, positive control). Specific cdna strands were synthesized from total RNA with an oligo-ap (AUAP + dt) and amplified using specific rola, rolb and rolc gene primers. The DNA bands of 307 (rola), 539 (rolb) and 762 bp (rolc) were detected on an ethidium bromide-stained agarose gel. M=molecular marker (DNA molecular weight standard, 100 bp ladder, Promega). TREE PHYSIOLOGY ONLINE at

6 718 AYADI AND TRÉMOUILLAUX-GUILLER References Ahn, J.C., B. Hwang, H. Tada, K. Ishimaru, K. Sasaki and K. Shimomura Polyacetylenes in hairy roots of Platycodon grandiflorum. Phytochemistry 42: Altamura, M.M., F. Capitani, L. Gazza, I. Capone and P. Costantino The plant oncogene rolb stimulates the formation of flower and root meristemoids in tobacco thin cell layers. New Phytol. 126: Attree, S.M., D.I. Dunstan and L.C. Fowke Initiation of embryogenic callus and suspension cultures, and improved embryo regeneration from protoplasts, of white spruce (Picea glauca). Can. J. Bot. 67: Ball, E Growth of the embryo of Ginkgo biloba under experimental conditions. III. Growth rate of the root and shoot upon media absorbed through the cotyledons in vitro. Am. J. Bot. 46: Balz, J.P., D. Courtois, J. Drieu, K. Drieu, J.P. Reynoird, C. Sohier, B.P. Teng, A. Touché and V. Pétiard Production of ginkgolides and bilobalide by Ginkgo biloba plants and tissue cultures. Planta Med. 65: Bercetche, J., D. Chriqui, S. Adam and C. David Morphogenetic and cellular reorientations induced by Agrobacterium rhizogenes (strains 1855, 2659 and 8196) on carrot, pea and tobacco. Plant Sci. 52: Camilleri, C. and L. Jouanin The TR-DNA region carrying the auxin synthesis genes of the Agrobacterium rhizogenes agropinetype plasmid pri A4: nucleotide sequence analysis and introduction into tobacco plants. Mol. Plant Microbe Interact. 4: Carrier, D.J., G. Cosentino, R. Neufeld, D. Rho, M. Weber and J. Archambault Nutritional and hormonal requirements of Ginkgo biloba embryo-derived callus and suspension cell culture. Plant Cell Rep. 8: Christen, P., T. Aoki and K. Shimomura Characteristics of growth and tropane alkaloid production in Hyoscyamus albus hairy roots transformed with Agrobacterium rhizogenes A4. Plant Cell Rep. 11: Di Cola, A., P. Costantino and L. Spano Cell commitment and rolb gene expression in the induction of root differentiation. Plant Cell Tissue Organ Cult. 46: Gamborg, O.L., T. Murashige, T.A. Thorpe and I.K. Vasil Plant tissue culture media. In Vitro 12: Jasik, J., B. Boggetti, G. Caricato and S. Mantell Characterization of morphology and root formation in the model woody perennial shrub Solanum aviculare Forst. expressing rolabc genes of Agrobacterium rhizogenes. Plant Sci. 124: Jeon, M.H., S.H. Sung, H. Huth and Y.C. Kim Ginkgolide B production in cultured cells derived from Ginkgo biloba L. leaves. Plant Cell Rep. 14: Kohli, A., D. Gahakwa, P. Vain, A. Laurie and P. Christou Transgene expression in rice engineered through particle bombardment: molecular factors controlling stable expression and transgene silencing. Planta 208: Laurain, D., J. Trémouillaux-Guiller, J.C. Chénieux and T.A. Van Beek Production of ginkgolides and bilobalide in transformed and gametophyte derived cell cultures of Ginkgo biloba. Phytochemistry 46: Lloyd, G. and B. McCown Commercially-feasible micropropagation of mountain laurel, Kalmia latifolia, by use of shoot-tip culture. Proc. Int. Plant Prop. Soc. 30: McAfee, B.J., E.E. White, L.E. Pelcher and M.S. Lapp Root induction in pine (Pinus) and Larch (Larix) spp. using Agrobacterium rhizogenes. Plant Cell Tissue Organ Cult. 34: Mihaljevic, S., S. Stipkovic and S. Jelaska Increase of root induction in Pinus nigra explants using agrobacteria. Plant Cell Rep. 15: Montoro, P., H. Etienne, N. Michaux-Ferrière and M.P. Carron Callus friability and somatic embryogenesis in Hevea brasiliensis. Plant Cell Tissue Organ Cult. 33: Murashige, T. and F. Skoog A revised medium for rapid growth and bioassays with tobacco tissue cultures. Physiol. Plant. 15: Phelep, M., A. Petit, L. Martin, E. Duhoux and J. Tempé Transformation and regeneration of a nitrogen-fixing tree, Allocasuarina verticillata Lam. Bio/Technology 9: Shunan, L., S. Tianenand and L. Genbao Transformation of Ginkgo hairy root and establishment of its suspension culture clone. Wuhan Univ. J. Sci. 2: Welander, M., N. Pawlicki, A. Holefors and F. Wilson Genetic transformation of the apple Rootstock M26 with the rolb gene and its influence on rooting. J. Plant Physiol. 153: White, P.R The cultivation of animal and plant cells. The Ronald Press, New York, 239 p. Willmitzer, L., J. Sanchez-Serrano, E. Buschfeld and J. Schell DNA from Agrobacterium rhizogenes is transferred to and expressed in axenic hairy root plant tissues. Mol. Gen. Genet. 186: TREE PHYSIOLOGY VOLUME 23, 2003