PlantCell Reports

Plant Cell Reports (1996) 15:317- 321

9 Springer-Verlag1996

Transgenic fertile Scoparia dulcis L., a folk medicinal plant, conferred with a herbicide-resistant trait using an Ri binary vector Mami Yamazaki 1, Lin Son 1, Toshimitsu Hayashi 2, Naokata Morita 2, Tetsuya Asamizu 3, lsamu Mourakoshi t, and Kazuki Saito x 1 Faculty of Pharmaceutical Sciences, Laboratory of Molecular Biology and Biotechnology in Research Center of Medicinal Resources, Chiba University, lnage-ku, Chiba, 263 Japan 2 Faculty of Pharmaceutical Sciences, Toyama Medical and Pharmaceutical University, Sugitani, Toyama, 930-01 Japan 3 Toyama Prefectural Institute for Pharmaceutical Research, Yarimizu, Toyama, 939-03 Japan Received 10 January 1995/Revised version received 23 June 1995 - Communicated by K. Shimamoto

Summary Transgenic herbicide-resistant Scoparia dulcis plants were obtained by using an Ri binary vector system. The chimeric bar gene encoding phosphinothricin acetyltransferase flanked by the promoter for cauliflower mosaic virus 35S RNA and the terminal sequence for nopaline synthase was introduced in the plant genome by Agrobacterium-mediated transformation by means of scratching young plants. Hairy roots resistant to bialaphos were selected and plantlets (R0) were regenerated. Progenies (S1) were obtained by self-fertilization. The transgenic state was confirmed by DNA-blot hybridization and assaying of neomycin phosphotransferase II. Expression of the bar gene in the transgenic R0 and SI progenies was indicated by the activity of phosphinothricin acetyltransferase. Transgenic plants accumulated scopadulcic acid B, a specific secondary metabolite of S. dulcis, in amounts of 15-60% compared with that in normal plants. The transgenic plants and progenies showed resistant trait towards bialaphos and phosphinothricin. These results suggest that an Ri binary system is one of the useful tools for the transformation of medicinal plants for which a regeneration protocol has not been established.

Abbreviations: CaMV: cauliflower mosaic virus, NPT-II: neomycin pbosphotransferase, PAT: phosphinothricin acetyltransferase, PPT: phosphinothricin Introduction The limited number of medicinal plants have been manipulated

by genetic engineering and very few pathways of secondary metabolism in them have been understood on molecular level. Thus, it is required to establish the systems for efficient transformation and regeneration of each species of medicinal plants (Saito et al. 1992a). Scoparia dulcis L. (Scrophulariaceae) is a perennial herb distributed in tropical and subtropical regions, and has been used for crude folk drugs in Paraguay, Taiwan and India. From this plant, diterpenoid acids named scopadulcic acids A and B were isolated together with scoparic acid A and scopadulin (Hayashi et al., 1987; Kawasaki et al., 1987). Among these compounds, scopadulcic acid B exhibits potent inhibitory activity as to replication of herpes simplex virus type 1 (HSV-1) (Hayashi et al., 1988) and H +, K +ATPase of the gastric mucosa (Asano et al., 1990), as well as antitumor activity (Hayashi et al., 1992). It is known that there are at least two chemotypes in this plant species as to the major component of diterpenoids, i.e., the scopadulcic acid B (SDB) and scoparic acid A (SA) types (Hayashi et al., 1991). SA is assumed to be a precursor of SDB in the biosynthetic pathway. For further study to reveal a molecular basis for characterizing these biogenetic chemotypes, it is necessary to establish a method for the genetic transformation of this plant. This paper is the first report on the transformation of S. dulcis. Correspondence to: K. Saito

The binary vector system based on an Agrobacterium-Ri plasmid has been used to produce transgenic hairy roots containing the T-DNAs of helper Ri plasmids and second binary vectors (Simpson et al., 1986). This technique depends on the fact that the T-DNA derived from a Ti plasmid can be mobilized in trans by vir gene products of the Ri plasmid. In some cases mature plants can be regenerated from hairy roots. We have been exploring the Ri binary vector system as to the genetic transformation of pharmaceutically important plants (Saito et al., 1991a, b, 1992a, b). The bar gene encodes phosphinothricin (PPT) acetyltransferase (PAT) in Streptomyces hygroscopicus. PAT inactivates the synthetic herbicide, PPT (Basta | Hrchst), which inhibits glutamine synthase, thus causing the death of plant cells. The bar gene was transferred into plants to obtain transgenic herbicideresistant crops (De Block et al., 1987, 1989, Spencer et al., 1990) and medicinal plants (Saito et al., 1992b). In this study, we transferred bar into Paraguayan S. dulcis, and obtained transgenic herbicide-resistant mature plants and progenies by using an Ri binary vector system.

Material and Methods Plasmids, bacteria and plants. In a binary vector, pARK5, the bar gene was placed under the transcriptional control of the promoter for CaMV 35S RNA, and flanked by the terminator of the nos gene for nopaline synthase (Saito et al., 1992b). The chimeric NPT-II gene was also present in the T-DNA as a reporter gene for transformation, pARK5 was introduced into Agrobacterium harboring pRi15834, a wild agropine-type Ri plasmid, as described previously (Saito et al., 1992b). Sterile Paraguayan S. dulcis plants (SDB type) were cultured on A1 medium (half-strength Murashige and Skoog medium (1962), 1% sucrose and 0.8% agar, pH5.7). Plant transformation and regeneration. Agrobacterium harboring pRi15834 and pARK5 was cultured in YEB medium (5 g/l beef extract, I g/1 yeast extract, 5 g/1 peptone, 5 g/l sucrose, 2 mM MgSO4) supplemented with 50 mg/l rifampicin and 25 mg/l kanamycin for 2 days at 25"C. Precultured Agrobacterium was inoculated into the stems of 3-week-old sterile S. dulcis plants by scratching with a needle. The hairy roots excised from the infected sites were transferred to B5 medium (Gamborg et al., 1968) containing 500 mg/l Claforan| to remove Agrobacterium and 5 mgfll bialaphos to select transformed hairy roots. Bialaphos (Herbiace| Meiji Seika Kaisha Ltd.) is a tripeptide containing PPT, the PPT being released in plant cells. Adventitious shoots that regenerated from bialaphos-resistant hairy roots were transferred to A1 medium for rooting. The rooted plants were transferred to pots containing vermiculite supplemented with 10% (v/v) A1 medium without agar and then to culture soil. The hairy roots, regenerated shoots and plants on agar medium were cultured at 25 "C in the light (16 hr/day, 2000 lux).

318 DNA blot hybridization. Plant DNA was extracted as described (Dellaporta et al., 1983). The DNA was digested with HindlIl, and then electrophoresed on a 0.8% agarose gel, transferred to a Hybond N+ filter (Amersham, Bucks, UK) and then hybridized with random prime labelled 32p-probes (Takara, Kyoto, Japan) by the protocols recommended by the suppliers. The purified 0.7 kb BamHI-EcoRl fragment of pARK22, which is a pUC19 derivative containing the same T-DNA region as that of pARK5, was used as the probe for the bar gene. The HindlIl-digested fragments of pLJ 1 and pLJ85 were used as probes for TL- and TR-DNA of pRi 15834 (Jouanin, 1984; Slightom et al., 1986). The filters were finally washed with 0.1xSSC, 0.1% SDS at 65"C. PAT and NPT-H. PAT activity was determined after acetylation of PPT with 14C-acetyl CoA by thin layer chromatography (De Block et al., 1987). NPT-II activity was assayed as described previously (Reiss et al., 1984). Herbicide application. The plant leaves were suffused with 100/al of 0.5% aqueous solution of the formulated Herbiace | (bialaphos sodium salt content, 20%) as recommended by the suppliers. Extraction and determination of scopadulcic acid B. Five developed leaves from the top of regenerated plants, which were growing on the AI agar medium and of 8-10 cm height, were lyophilized. Twenty mg of freeze-dried leaves was extracted with 5 ml of CHC13 by sonication for 20 min. The extract was concentrated and then applied to a Bond Elut Cartridge (Varian, SI, 3CC). After washing with 5 ml CHCI3, the column was eluted with 3 ml of MeOH. The eluate was evaporated to dryness and the resultant residue was dissolved in 0.3 or 0.5 ml MeOH containing 0.25 mg/ml of p-toluenesulphoneacid as an internal standard. Two I.tl of the sample solution was subjected to HPLC analysis. HPLC was carried out on a column packed with Cosmosil 5CI8-AR (150 mm x 4.6 mm i.d., Nacalai Tesque), using MeOH-0.02M H3PO4 (3:1 v/v) as the mobile phase at a flow rate of 10 ml/min at 30~ with monitoring at 230 nm. Analysis of $1 progenies. Progenies were obtained by selffertilization. Surface-sterilized seeds of SI progenies from clone C and untransformed plants were sown on A 1 agar medium containing 5 mg/1 of bialaphos. After two weeks, the number of resistant seedlings was determined. PAT activity in bialaphos-resistant seedlings was also assayed.

Results

Plant transformation, regeneration and fertilization. The optimum concentration of bialaphos for the selection of transformants was determined using hairy roots of S. dulcis obtained by transformation only with the Ri plasmid. The concentration of 5 mg/l bialaphos in B5 agar medium was sufficient to inhibit the growth of excised hairy roots. We decided to use this concentration of bialaphos for the selection of resistant hairy roots. Within 2 weeks after inoculation, hairy roots had appeared at the scratch sites on stems. The hairy roots of a length about 5 mm were excised and transferred onto B5 agar medium containing Claforan | and bialaphos. Five out of 20 clones from independently infected sites grew on the selection medium (Fig. 1). After a month, adventitious shoots were regenerated from the green parts of bialaphos-resistant hairy roots. The excised shoots were transferred onto A 1 medium for rooting. The rooted plantlets were grown for three weeks on vermiculite containing 10% (v/v) A1 medium in culture pots with lids, and then transferred to planting soil. Finally, 4 clones of transformed plants, clones B, C, E and G2, were established. Among these regenerated plants, clone G2 showed the marked characteristic features, such as wrinkled leaves and short internodes, of a regenerated plant from a hairy root caused by the expression of the T-DNA genes of an Ri plasmick

Confirmation of the transgenic state of the transformants The integration of the T-DNAs of pRi15834 and pARK5 was analyzed by DNA-blot hybridization (Fig. 2). Plant DNA was digested with HindllI and then blotted on to a filter. The filter was hybridized with a BamHI-EcoRI fragment of pARK5 covering the bar gene and the nos terminator as a probe. The results indicated that clones B, C and G2 contained one to several copies of the TDNA of pARK5 (Fig. 2a). However the hybridized band in clone B was shorter than the standard. It is speculated that the bar gene in clone B was not inserted as intact form but was rearranged. This speculation is supported by PAT assays exhibiting no activity in clone B (Fig.3). Clone E did not contain the bar gone. The presence of the TL-DNA sequence of pRi15834 was indicated in all transformants examined using HindlIl-digested fragments of pLJl as probes (Fig. 2b). The signal intensities were varied among clones. All these transformants produced agropine (data not shown), indicating the integration and expression of the TR-DNA of pRi 15834 in the plant genome (De Paolis et al., 1985).

Miscellaneous techniques. Protein was determined with a kit supplied by Bio-Rad Laboratory (Richmond, Cal., USA). Agropine and mannopine were detected as reported previously (Saito et al., 1990a)

Fig. 2. DNA-blothybridizationof transformants. PlantDNA (20 ~g) was digested with HindlII,and then electrophoresedon a 1.0%agarose gel, transferredto a nylon

filterand thenhybridizedwith the 32p-probes. (a) Hybridizatioriwith the BamHI-EcoRI fragmentof pARK22, lanes: 1, 10pg of llindllI-EcoRl fragmentof pARK22; 2, 3, 4, transformedclones B, C, G2; 3', a S1 progenyof clone C: 5, non-transformedcontrol. (b) Hybridizationwith HindlIl fragmentsof plJ 1. Lanes: 1,500 pg of HindllI fragmentsof pLJl; 2, 3.3', 4, ~me transformedclonesas in (a); 5, non-Uanslormedcontrol.

319

I~if. 1. Regeneration of transgeni,- ~coparia dulcir plants from bialaphosresistant hairy roots. (a) Selection Oil B5 agaf medium supplemented with 5 rag/1 of bialaphos and the formation of adventitious shoots; (b) enlarged picture of shoot-forming hairy roots; (c) regenerated plant under acclimatization; (d) regenerated plants in a pot of culture soil.

Fig. 3. Expression of PAT activity in transformants. The reaction products of 14C-acetyl-CoA and PPT in the presence of protein extracts of plant tissue were analyzed by thin layer chromatography as described previously (De Block et al., 1987). Lanes: 1, [l-14C]-acetyl-CoA (2 GBq/mmole); 2, non-transformed plant as a negative control; 3, a negative control plus the purified PAT from St. hygroscopicus; 4, 5, 6, 7, transformed clones B, C, E, G2, transformants; 8, control hairy root transformed only with pRi15834.

Fig. 4. Resistance of a transgenic plant expzessing the bar gene to biala0ho,: A 0.5% aqueous solution of the commercial formula, Herbiace | (bialaphos sodium salt contem. 20%), was applied to one leaf ot a plant a.s rex~mmended by the supplier. (a) Control plant derived from a hairy root transformed only with pRi 15834. Seven days after application, the l~'aftreated with Herbiace| (indicated by an arrowhead) had completely died. (b) Clone C 7 days after application. No visible change was observed in the leaf treated with Herbiace| (indicated by an arrowhead).

Fig. 5. Germination of the SI seeds on the medium supplemented with bialaphos. Surface-sterilized seeds of a non-transformed plant and clone C were germinated on AI agar medium containing 5 mg/I of bialaphos for 14 days. (a) All the non-transformed .seeds (total 50 seeds) could not germinate. (b)The S1 progenies of clone C (48 / 67, 72%) grew.

Fig. 6. Expression of PAT activity in progenies. The assay was carried out in the same way as in Fig. 3. Lanes: 1-3, sam~ as lanes 1-3 in Fig. 3, 415, independent S 1 progenies of the parent plant of clone C.

320 Expression of ctu'meric genes and resistance towards a herbicide

Discussion

The expression of the chimeric bar gene was analyzed by assaying of PAT (Fig. 3). Positive activity was shown by clones C and G2, but not by clones B and E. It is speculated that the fragmented copy of bar gene in clone B (Fig. 2a) is not functional. While the reason is not clear, integrated bar gene might have been rearranged in clone B during regeneration after selection with the herbicide. Clone E probably had escaped from bialaphos selection. Regenerated transformants were examined as to their resistance to a commercial herbicide formulation (Fig. 4). An aqueous solution of Herbiace| was applied on the leaves of a regenerated plant of clone C and an untransformed control plant. The control leaves died 7 days after application. On the contrary, the leaves of clone C showed resistance to the herbicide.

Modification by means of genetic engineering has been performed for many kinds of crop plants. The critical steps in the genetic engineering of plants are the selection of transformed cells and plant regeneration from these cells. The Ri binary vector system provides a powerful tool in plant species that can be transformed with Agrobacterium containing an Ri plasmid and that can be frequently regenerated from hairy roots. This system has been applied to several plant species (Shahin et al., 1986; Simpson et al., 1986; Hamil et al., 1987; Stougaard et aL, 1987; Saito et al., 1990a, b, 1991a, 1992a, b). The Ri plasmid vector has some characteristic features compared with a completely disarmed Ti vector, as follows: (1) one can easily distinguish transformed cells as hairy roots integrated with any desirable foreign genes on a second binary vector in high frequency; in the present study, 10% (2/20) of unselected hairy roots were doubly transformed; (2) in some plant species, mature plants can be regenerated from hairy roots and offspring are also obtained (Tepfer, 1984; Tepfer et al., 1989; Costantino et al., 1984; Sukhapinda et al., 1987; Saito et al., 1992b). These features could be advantageous in some cases of genetic engineering. However, plants transformed with an Ri vector also show an unfavorable "hairy root syndrome" for application to transgenic plants. The contents of scopadulcic acid B were decreased in the regenerated plants derived from hairy roots of S. dulcis. In particular, clone G2, which showed the syndrome markedly, accumulated the lowest level of scopadulcic acid B. Such reduced production of secondary metabolites was reported in regenerated plants from hairy roots of Atropa belladonna (Jung and Tepfer, 1987; Saito et al. 1992b). However, in some cases, the regenerated plants of heavy Ri syndrome even produced higher amounts of secondary metabolites (Ko et al., 1988; Saito et al., 1991c). Thus, it would be necessary to investigate in further detail concerning to the relation of Ri syndrome and produciblity of secondary metabolites. In the case of S. dulcis, fertile regenerated plants were obtained spontaneously without the addition of any phytohormone. The S1 progenies of 72% (48/67) of showed PPT resistance. This is the normal Mendelian segregation ratio and thus it was suggested that the integrated foreign gene was stably transmitted to the second generation. The T-DNAs carrying bar and the TL- and TR-DNAs of Ri plasmid were segregated in offspring. Thus, we could obtain S 1 progenies carrying only bar but not TL- and TR-DNAs causing Ri syndrome. These progenies are desirable for molecular breeding conferring useful genetic traits by Ri binary vectors. The strategy of genetic complementation of a mutant chemotype by a DNA fragment determining such a mutation is currently being.used for anthocyanin pigment biosynthesis (Lloyd et al., 1992). However, it is necessary to establish a method of genetic transformation in order to identify the DNA sequences responsible for genetically determined chemotypes such as the SA and SDB types of S. dulcis. The protocol we presented in this study constitutes a reproducible method for the genetic transformation of this plant. It is applicable for further investigation of reverse genetics concerning DNA fragments determining biosynthetic chemotypes.

Production o f a specific secondary product, SDB

The accumulation of SDB in leaves of transgenic clones was analyzed by HPLC (Table 1). The leaves of transformants accumulated SDB in amounts of 15-60% compared with that in untransformed plants. In particular, clone G2, which showed marked phenotypic changes of the Ri syndrome, accumulated the lowest amount of SDB. Thus, it is suggested that some physiological change caused by expression of the genes in the Ri plasmid might affect the production of diterpenoid acids.

Table 1. Contents of scopadulcic acid B in regenerated plants.

clone B C E G2 HRC UTd

scopadulcic acid B (%) 0.44 a 0.44 0.32 0.11 0.45 0.74

(0.08) b (0.06) (0.10) (0.01) (0.05) (0.12)

a% dry weight, b% fresh weight, CTransformed with only pRi15834, auntransfa,mexlcon~l. The diterpenoid fraction was extracted from freeze-dried leaves of ta'ansformantsgrowingin pots containingAI agar medium. Scopadulcicacid B was determined by HPLC using p-toluenesulfonic acid as an internal standard as described under Material and Methods. Data are the means of duplicatedeterminations.

Inheritance of a transgenic trait by progenies

S 1 progenies of clone C were obtained and analyzed for a transgenic trait. Seeds of the offspring and control untransformed plants were germinated on A1 agar medium containing bialaphos. Within two weeks, all the control seeds had died, but 48 out of 67 seedlings (72%) of the S1 progenies had been able to develop cotyledons (Fig. 5). All the bialaphos-resistant seedlings showed PAT activity (Fig. 6) and same hybridization pattern with bar probe as that of their parent plant (Fig. 2a, lane 3'). The segregation ratio of resistant and sensitive progenies was statistically significant being 3 : 1 (X2=0.4, P>0.05). These results indicated that integrated bar genes in the parent clone C were linked and transmitted to the offspring as an apparent single dominant allele according to the Mendelian rule. DNA blot hybrydization of the progenies indicated that ten out of 19 bialaphos-resistant progenies (53%) and 16 out of 19 resistant progenies (84%) contained TL- and TR-DNA, respectively (data not shown). The clones carrying TL-DNA showed the phenotypes typical of Ri syndrome such as shorter internodes and less apical dominance compared with normal plants. In contrast, three clones carryng no TL- and TR-DNAs showed no such Ri syndrome.

References

Asano, S., Mizutani, M., Hayashi, T., Morita, N., and Takeguchi, N. (1990) J. Biol. Chem., 265, 22167-22173 Costantino, P., Spano, L., Pomponi, M., Benvenuto, E., and Ancora, G. (1984) J. Mol. Appl. Genet., 2, 465-470 De Block, M., Botterman, J., Vandewiele, M., Dockx, J., Thoen, C., Gossele, V., Rao Movva, N., Thompson, C., Van Montagu, M., and Leemans, J. (1987) EMBO J., 6, 2513-2518 De Block, M., De Brouwer, D., and Tenning, P. (1989) Plant Physiol., 91, 694-701 Dellaporta, S. L., Wood, J., and Hicks, J. B. (1983) Plant Mol. Biol. Reporter, 1, 19-21 De Paolis, A., Mauro, M. L., Pomponi, M., Cardarelli, M., Spano, L., and Costantino, P. (1985) Plasmid, 13, 1-7 Gamborg, O. L., Miller, R. A., and Ojima, K. (1968) Exp. Cell Res., 50, 151-158 Hamil, J. D., Prescott, A., and Martin, C. (1987) Plant Mol. Biol., 9 573-584

321 Hayashi, T., Kishi, M., Kawasaki, M., Arisawa, M., Shimizu, M., Suzuki, S., Yoshizaki, M., Morita, N., Tezuka, Y., Kikuchi, T., Berganza, L.H., Ferro, E., and Basualdo, I. (1987) Tetrahedron Lett., 28, 3693-3696 Hayashi, K., Niwayama, S., Hayashi, T., Nago, R., Ochiai, H., and Morita, N. (1988) Antiviral Res., 9, 345-354 Hayashi, T., Okamura, K., Kawasaki, M., and Morita, N. (1991) Phytochem., 30, 3617-3620 Hayashi, K., Hayashi, T., and Morita, N. (1992) Phytother. Res., 6, 6-9 Jouanin, L. (1984) Plasmid, 12, 91-102 Jung, G., and Tepfer, D. (1987) Plant Science, 50, 145-151 Kawasaki, M., Hayashi, T., Arisawa, M., Shimizu, M., Hiroe, S., Ueno H., Syogawa, H., Suzuki, S., Yoshizaki, M., Morita, N., Tezuka, Y., Kikuchi, T., Berganza, LH., Ferro, E., and Basualdo, I. (1987) Chem. Pharm. Bull., 35, 3963-3966 Ko, K. S., Ebizuka, Y., Noguchi, H., Sankawa, U. (1988) Chem. Pharm. Bull., 36, 4217-4220 Lloyd, A. M., Walbot, V., and Davis, R. W. (1992) Science, 258, 1773-1775 Murashige, T., and Skoog, F. (1962) Physiol. Plant, 15, 473-479 Reiss, B., Sprengel, R., Will, H., and Schaller, H. (1984) Gene, 30, 211-218 Saito, K., Kaneko, H., Yamazaki, M., Yoshida, M., and Murakoshi, I. (1990a) Plant Cell Rep., 8, 718-721 Saito, K., Yamazaki, M., Shimomura, K., Yoshimatsu, K., and Murakoshi, I. (1990b) Plant Cell Rep., 9, 121-124

Saito, K., Yamazaki, M., Kaneko, H., Murakoshi, I., Fukuda, Y., and Van Montagu, M. (1991a) Planta, 184, 40-46 Saito, K., Noji, M., Ohmori, S., Imai, Y., and Murakoshi, I. (1991b) Proc. Natl. Acad. Sci. USA, 88, 7041-7045 Saito, K., Yamazaki, M., Kawaguchi, A., and Murakoshi, I. (1991c) Tetrahedron., 47, 5955-5968 Saito, K., Yamazaki, M., and Murakoshi, I. (1992a) J. Nat. Prod., 55, 149-161 Saito, K., Yamazaki, M., Anzai, H., Yoneyama, K., and Murakoshi, I. (1992b) Plant Cell Rep., 11,219-224 Shahin, E. A., Sukhapinda, K., Simpson, R. B., and Spivey, R. (1986) Theor. Appl. Genet., 72, 770-777 Simpson, R. B., Spielmann, A., Margossian, L., and McKnight, T. D. (1986) Plant Mol. Biol., 6, 403-415 Slightom, J. L., Durand-Tardif, M., Jouanin, L., and Tepfer, D. (1986) J. Biol. Chem., 261, 108-121 Spencer, T. M., Gordon-Kamm, W. J., Daines, R. J., Start, W. G., and Lemaux, P. G. (1990) Theor. Appl. Genet., 79, 625631 Stougaard, J., Abildsten, D., and Marcker, K. A. (1987) Mol. Gen. Genet., 207, 251-255 Sukhapinda, K., Spivey, R., Simpson, R. B., and Shahin, E. A. (1987) Mol. Gen. Genet., 206, 491-497 Tepfer, D. (1984) Cell, 37, 959-967 Tepfer, D., Metzger, L., and Prost, R. (1989) Plant Mol. Biol., 13, 295-302