efficient agrobacterium tumefaciens-mediated

In Vitro Cell. Dev. Biol.—Plant 40:359–365, July/August 2004 q 2004 Society for In Vitro Biology 1054-5476/04 $18.00+0.00

DOI: 10.1079/IVP2004539

EFFICIENT AGROBACTERIUM TUMEFACIENS-MEDIATED TRANSFORMATION OF SWEET POTATO (IPOMOEA BATATAS (L.) LAM.) FROM STEM EXPLANTS USING A TWO-STEP KANAMYCIN – HYGROMYCIN SELECTION METHOD GUO-QING SONG†, HIDEO HONDA,

AND

KEN-ICHI YAMAGUCHI*

Functional Chemicals Laboratory, Mitsui Chemicals, Inc., 1144, Togo, Mobara-shi, Chiba, 297-0017, Japan (Received 30 October 2003; accepted 3 February 2004; editor J. M. Widholm)

Summary To achieve reliable stable transformation of sweet potato, we first developed efficient shoot regeneration for stem explants, leaf disks, and petioles of sweet potato (Ipomoea batatas (L.) Lam.) cultivar Beniazuma. The shoot regeneration protocol enabled reproducible stable transformation mediated by Agrobacterium tumefaciens strain EHA105. The binary vector pIG121Hm contains the npt II (pnos) gene for kanamycin (Km) resistance, the hpt (p35S) gene for hygromycin (Hyg) resistance, and the gusA (p35S) reporter gene for b-glucuronidase (GUS). After 3 d co-cultivation, selection of calluses from the three explant types began first with culture on 50 mg l21 of Km for 6 wk and then transfer to 30 mg l21 of Hyg for 6 –16 wk in Linsmaier and Skoog (1965) medium (LS) also containing 6.49 mM 4-fluorophenoxyacetic acid and 250 mg l21 cefotaxime in the dark. The selected friable calluses regenerated shoots in 4 wk on LS containing 15.13 mM abscisic acid and 2.89 mM gibberellic acid under a 16 h photoperiod of 30 mmol m22 s21. The two-step selection method led to successful recovery of transgenic shoots from stem explants at 30.8%, leaf discs 11.2%, and petioles 10.7% stable transformation efficiencies. PCR analyses of 122 GUS-positive lines revealed the expected fragment for hpt. Southern hybridization of genomic DNA from 18 independent transgenic lines detected the presence of the gusA gene. The number of integrated T-DNA copies varied from one to four. Key words: sweet potato; Ipomoea batatas; GUS; Agrobacterium tumefaciens; transgenic plants. Lawton et al., 2000), was achieved. Transient and stable expression of GFP was also reported after electroporation of protoplasts (Dhir et al., 1998; Lawton et al., 2000). However, while these direct gene transfer methods led to successful delivery of genes into sweet potato, no stable transgenic plants were regenerated and confirmed. Otani et al. (1993), using Agrobacterium rhizogenes, obtained morphologically aberrant shoots, without a selection system, from hairy roots induced on leaf explants of five cultivars among 14 tested on LS medium (Linsmaier and Skoog, 1965), but only one transgenic line was confirmed to be stable by Southern analysis. This research group later succeeded in obtaining transgenic sweet potato plants mediated by A. tumefaciens from embryogenic calluses of cv. Kokei 14 after selection with 30 mg l21 hygromycin (Hyg) following a 3-d co-cultivation (Otani et al., 1998). An average of 10 Hyg-resistant clusters, of which 53.1% could be regenerated on LS þ 15.13 mM abscisic acid (ABA) þ 2.89 mM gibberellic acid (GA3), were obtained from 1 g of embryogenic calluses (Otani et al., 1998; Kimura et al., 2001). Transformation mediated by A. tumefaciens with selection by 50 mg l21 Km yielded regenerated transgenic plants of cv. Jewel (Newell et al., 1995; Mora´n et al., 1998) and cv. White Star (Gama et al., 1996). Transgenic plants of cv. Jewel were regenerated from 2.5– 10% of freshly harvested storage roots after a series of plant growth regulator changes over 6 mo. (Newell et al., 1995). Mora´n et al. (1998) obtained 32 individual shoots from 45 leaf explants of cv. Jewel in 6 wk on MS with a combination of 6-benzylaminopurine (BA) and naphthaleneacetic

Introduction Sweet potato [Ipomoea batatas L. (Lam.)] is an attractive target for ‘plant molecular farming’ due to its clonal propagation, ease of cultivation, and high productivity of storage roots and foliage. Likewise, male sterility, incompatibility, and a hexaploid genome make it difficult to improve by conventional breeding (Sihachakr and Ducreux, 1987). The development of an efficient and reproducible transformation system is needed for genetic manipulation of sweet potato to either improve the crop or establish it as a novel ‘transgenic plant bioreactor’ (Giddings et al., 2000). However, sweet potato is not yet among the crop plants that can be routinely transformed. A great deal of effort has been made to establish an efficient transformation system for sweet potato. Using particle bombardment, transient and stable expression of marker genes in sweet potato cultivars, b-glucuronidase (GUS) in cv. Jewel and TIS-70357 after selection by 50 mg l21 kanamycin (Km) (Prakash and Varadarajan, 1992), and green fluorescent protein (GFP) gene in cv. Beauregard without antibiotic selection (Dhir et al., 1998; *Author to whom correspondence should be addressed at (present address): Lab for Sustainable Horticulture, Minami Kyushu University, Hibarigaoka, Takanabe-cho, Miyazaki 884-0003, Japan. Email keny@ nankyudai.ac.jp †Present address: Plant Transformation Center, Department of Horticulture, Michigan State University, East Lansing, MI 48824.

359

360

SONG ET AL.

FIG . 1. Schematic representation of the plant expression vector pIG121Hm.

acid (NAA). Half of the 32 Km-resistant shoots that also rooted on 100 mg l21 Km were considered stable transformants, giving a transformation frequency of 33%. However, only one transgenic line was used in Southern blot analysis, and untransformed escapes among Km-resistant shoots could not be completely excluded. Transgenic plants were also obtained at a frequency of 12.3% from embryogenic calluses of sweet potato cv. White Star within 7 wk (Gama et al., 1996). The methods reviewed here have not been widely adapted due to low transformation efficiencies. In addition, both freshly harvested storage roots and embryogenic calluses are not readily available target tissues. Considerable effort is still needed to develop an efficient sweet potato transformation and regeneration system. In the present study, stem, petiole, and leaf disk explants of sweet potato cv. Beniazuma, the major sweet potato cultivar in Japan, were used for regeneration and transformation studies. The method of transformation, including the selection method reported here, is the first to enable efficient production of transgenic sweet potato plants, especially from stem explants, and is anticipated to have potential application for other sweet potato cultivars.

Materials and Methods Plant materials. Storage roots of the sweet potato (Ipomoea batatas (L.) Lam.) cv. Beniazuma were potted in sterilized planting medium and placed in a growth chamber at 298C under a 12 h photoperiod of 30 mmol m22 s21 photosynthetic photon flux (PPF) using cool white fluorescent tubes. After 4 wk, many shoots, ,10 –30 cm in length, were produced. The stock plants were watered every 2 wk and could be maintained for several months to provide explants. LS medium was used throughout this study with 3% sucrose, solidified with 0.32% gellan gum, and the pH adjusted to 5.6 before autoclaving at 1218C for 15 min at 105 kPa. To obtain explants, shoots 10 cm in length, including the shoot tip, were excised from greenhouse stock plants, the leaves were removed, and shoot tips, ,2– 3 cm in length, were sterilized for 15 min in 5% NaOCl with 0.01% Tween 20. The shoot tips were rinsed three times in sterile water. To establish in vitro stock cultures, shorter shoot tips, ,5 –7 mm in length, were excised and 10 shoot tips were placed upright in each 90 £ 20 mm plastic Petri dish. Each dish contained 30 ml of LS þ 8.88 mM BA þ 1.14 mM indole-3-acetic acid (IAA). The dishes were placed at 268C under a 16 h photoperiod of 30 mmol m22 s21. After 3 wk, elongating shoot tips were cultured singly in Magenta boxes; each contained 50 ml of LS þ 2.85 mM IAA (herein GM). These shoot tips elongated and also rooted, and were maintained by cultivating node sections on a monthly basis. These stock cultures were maintained in the same environmental conditions as for the starting dishes. Internode stem sections from in vitro stock plants, or those taken from shoots that were sterilized as previously mentioned, were cut transversely into sections 6–10 mm in length. The sections were then cut in half along the axis, and used as explants for regeneration and transformation experiments. Leaf disks, 5–7 mm, and petiole explants, 8–10 mm, were also prepared from in vitro stock plants. For callus induction, LS þ 6.49 mM 4-fluorophenoxyacetic acid (4-FA) (herein CIM), and for shoot regeneration, LS þ 15.13 mM ABA þ 2.89 mM GA3 (herein RM), both media from Otani and Shimada (1996) originally devised for shoot regeneration via embryogenesis from shoot tips, were used. Stem, petiole explants, and leaf disks were placed on 30 ml of CIM in a Petri dish (90 £ 20 mm), 10 explants

per dish, and cultured at 268C in the dark for 4 wk for induction of calluses, and subcultured monthly. Four to 8 wk after induction, explants with primary calluses were transferred to Petri dishes containing 30 ml of RM, and maintained at 268C under a 16 h photoperiod of 30 mmol m22 s21. Shoots regenerated in another 4 wk, and individual regenerated shoots, 0.5 cm or longer, were inserted for rooting, singly on 50 ml of GM in a Magenta box under the same light and environmental conditions as previously stated. Ten shoot tips from in vitro plants, 5–7 mm in length, were also cultured on CIM for obtaining calluses. Agrobacterium tumefaciens strain and expression plasmid. The binary vector pIG121Hm containing the chimeric neomycin phosphotransferase-II gene (npt II) under the NOS promoter, the hygromycin phosphotransferase gene (hpt) under the CaMV 35S promoter, and an intron-GUS expression cassette also under the CaMV 35S promoter were used (Fig. 1). This binary vector was introduced into A. tumefaciens strain EHA105 by electroporation (Ohta et al., 1990). Sensitivity of explants to antibiotics. To determine the optimum antibiotic concentration for use in the selection of transformed cells, Km (50 and 100 mg l21) and Hyg (10 and 30 mg l21) were tested separately based on previous reports for sweet potato (Prakash and Varadarajan, 1992; Otani et al., 1993; Newell et al., 1995; Gama et al., 1996; Dhir et al., 1998; Otani et al., 1998; Mora´n et al., 1998). Later, 50 mg l21 Km followed by 30 mg l21 Hyg was also tested. Stem, petiole, and leaf disk explants were prepared, and placed in each Petri dish on 30 ml of CIM þ antibiotics and sealed with plastic wrap. Five Petri dishes, each with six explants, were used as replications for each treatment. The explants were incubated for 2 wk in the dark, and were then placed under a 16 h photoperiod of 30 mmol m22 s21 at 268C. The number of explants producing calluses was recorded after 6 wk in the light. Plant transformation. A single colony of A. tumefaciens strain EHA105: pIG121Hm was cultured in 2 ml of LB containing 50 mg l21 Km and 30 mg l21 rifampicin at 308C for 48 h. Twenty ml of the culture were then inoculated into 20 ml of the same medium and grown to an OD600 of 0.8–1.0. The culture was centrifuged at 2500 £ g for 2 min. The bacterial pellet was resuspended in an equal volume of CIM þ 100 mM acetosyringone (AS) and incubated for 1 h at 308C. Freshly prepared stem, petiole, leaf disk explants, and calluses were incubated in the bacterial suspension with rotary shaking at 100 rpm for 20 min at 308C, and then blotted dry on sterile filter paper. These explants were then placed on sterilized filter paper overlaid on 30 ml of solid co-culture medium containing 100 mM AS in each Petri dish. Three co-culture media, CIM, LS, and LS þ 4.52 mM 2, 4-dichlorophenoxyacetic acid (2,4-D), were tested for transient transformation. Co-cultivation was carried out for 3 d at 238C in the dark. After co-cultivation, the explants were washed in CIM þ 500 mg l21 cefotaxime (Cx) for 10 min, rinsed three times in sterile water, and then blotted dry on sterile filter paper. To develop a selection system, antibiotics at concentrations previously mentioned were used with stem explants. Three dishes each with 10 stem explants were used as replicates for each treatment. The experiment was repeated twice. Later, the optimum Km–Hyg selection method was used to select transformed cells from the other explant types. The initial stage of selection was carried out on CIM containing 50 mg l21 Km þ 250 mg l21 Cx at 3-wk intervals for 6 wk at 238C. In the next stage of selection, explants were transferred to CIM þ 30 mg l21 Hyg þ 250 mg l21 Cx at 268C. The entire selection procedure was carried out in the dark. Km- or Km and Hyg-resistant selected calluses were transferred to RM with or without 100 mg l21 Cx and cultured at 268C under a 16 h photoperiod of 30 mmol m22 s21. Four weeks later, the regenerating shoots were excised and placed on GM for rooting. The plantlets with fully developed roots were transferred to planting medium in plastic pots (20 £ 30 cm) and grown in a growth chamber at 268C under a 16 h photoperiod of 50 mmol m22 s21. Shoot regeneration efficiencies for the three explant types and stable transformation frequencies under the four selection conditions for stem

TRANSFORMATION OF SWEET POTATO

FIG . 2. Shoot regeneration from stem explants of sweet potato cv. Beniazuma. A, Formation of bright yellow, friable calluses on CIM after 4 wk. B, Shoot regeneration from friable calluses on RM in 4 wk. explants were analyzed for significance by ANOVA with mean separation by Duncan’s test (P # 0.05). PROCGENMOD (SAS Institute Inc.) was used. Histochemical GUS assay. GUS expression in transformed and nontransformed calluses, stem explants, shoots, plantlets, and mature leaves was determined histochemically according to Jefferson et al. (1987). Prior to fixation, the surface of the mature leaves was gently scored with a surgical blade to enable more staining solution to enter the cells. After fixation, all samples were incubated overnight at 378C in assay buffer containing 1 mM 5-bromo-4-chloro-3-indolyl-D -glucuronide, cyclohexylammonium salt (X-Gluc). Polymerase chain reaction (PCR) and Southern analysis. Total genomic DNA was isolated from the leaves of in vitro-cultured plants following the CTAB method (Rogers and Bendich, 1985). About 200–500 ng of template DNA was used for one 20 ml PCR. The primers corresponding to a 987-bp fragment of the hpt gene were as follows: hpt-F: 50 -GCC TGA ACT CAC CGC GAC-30 ; hpt-R: 50 -CGT CGG TTT CCA CTA TCG G -30 . The reaction conditions were 948C for 2 min, 35 cycles of 948C for 10 s, 558C for 1.5 min and 728C for 2 min, with a final 10-min extension at 728C. For Southern analysis, 20 mg of DNA was digested with HindIII or HindIII and EcoRI, electrophoresed in 0.8% agarose gel and subsequently transferred to a nylon membrane. A 1.9-kb SmaI/SacI fragment containing the GUS coding region isolated from the pIG121Hm plasmid was used as the probe. Labeling, hybridization, and detection were performed using the Alkphos Direct Kit (Amersham Biosciences Co.) according to the manufacturer’s instructions.

Results Regeneration potential of explants. Over 90% (55/60) of stem explants produced bright yellow calluses at the cut edges within 4 wk of culture on CIM (Fig. 2A). On RM, an average of 10 shoots were obtained from those stem explants with bright yellow calluses after 4 wk (Fig. 2B). The shoots rooted on GM in 3 wk. Stem explants from field plants and in vitro stock cultures both had the same capacity for producing regenerable calluses. Similarly, bright yellow calluses, which regenerated into shoots on RM, were induced from 90% of leaf disks and 93% of petiole explants after 8 wk on CIM. There was no difference in the mean number of shoots produced by stem explants, petiole explants, and leaf disks, 10, seven, and eight, respectively. The bright yellow calluses usually retained their shoot regeneration capacity on CIM for 2 mo. However, over this period, a loss of the bright yellow color and a decrease in the frequency of shoot regeneration were observed in some lines. Effect of selective antibiotics on shoot regeneration. All stem explants turned black and died after 4 wk on CIM containing 30 mg l21 Hyg (Fig. 3A-a). The same response occurred when Hyg was decreased to 10 mg l21. With 50 or 100 mg l21 Km, the stem explants remained alive and expanded somewhat in 4 wk, and callus was formed from 15 –20% of the explants (Fig. 3A-b). In comparison,

361

the stem explants remained green and produced bright yellow calluses on CIM without antibiotics (Fig. 3A-d). Similar responses to the antibiotics were observed for leaf disk and petiole explants and calluses induced from shoot-tip explants. Establishing optimal parameters for transformation. Among the co-cultivation media, CIM, LS, and LS þ 4.52 mM 2,4-D, CIM was optimal for transient GUS expression in stem explants mediated by strain EHA105: pIG121Hm (data not shown). The frequency of transient GUS expression, measured as the percentage of stem explants with at least one blue spot, was over 80% after 3 d of cocultivation in the dark. Therefore, the same co-cultivation environment was used for the other explant types in the following experiments. Based on the responses of stem explants separately to Km or Hyg and the results of a preliminary experiment on a two-step Km –Hyg method, we evaluated three selection methods after co-cultivation: (1) CIM with Hyg (30 or 10 mg l21); (2) CIM with 50 mg l21 Km; and (3) CIM with 50 mg l21 Km followed by CIM with 30 mg l21 Hyg. When Hyg (10 or 30 mg l21) was tested, all EHA105-inoculated stem explants turned black and died in 2 wk. With 50 mg l21 Km, resistant calluses were produced from EHA105-inoculated stem explants after 10 wk of culture (Fig. 3A-c). Resistant regenerable yellow calluses were induced from 53.3% of the inoculated stem explants (Fig. 3B). After transfer to RM without selective antibiotics, shoot regeneration from these calluses began within 4 wk (Fig. 3C). However, only 12.7% (15/126) of the regenerated shoots were GUS-positive. Compared with the two-step selection method, the transformation frequency using 50 mg l21 Km was much lower (Table 1). In the two-step selection system, stem explants were first placed on CIM containing 50 mg l21 Km þ 250 mg l21 Cx for 6 wk after cocultivation, and then transferred to CIM þ 30 mg l21 Hyg þ 250 mg l21 Cx. Six to 10 wk later, yellow, friable resistant calluses occurred on these explants (Fig. 3D, 3E-a). No difference in transformation frequency was found between stem explants from in vitro stock cultures and those from field-grown plants (Table 2). Using the same two-step selection method, Hyg-resistant calluses were produced from inoculated leaf disks and petiole explants after 18 wk (Fig. 3E-b, 3E-c). From the inoculated calluses, Hyg-resistant yellow, friable callus clusters were induced after at least 10 mo. induction (Fig. 3E-d). After transfer to RM without selective antibiotics, the selected calluses from each explant type developed a large number of red shoot buds after 2 wk (Fig. 3F). Subsequently 5 – 10 shoots per cluster regenerated from over 90% of these calluses within 4 wk (Fig. 3G). All regenerated shoots tested were found to be GUS-positive. Consequently, the transformation frequencies measured as the percentage of explants that produced GUS-positive shoots were 30.8% for stem explants, 11.2% for leaf disks, 10.7% for petiole explants, and 3.2% for calluses (Table 2). Over 85% of the shoots recovered from stem explants rooted on LS þ 2.85 mM IAA in 4 wk and developed into morphologically normal phenotypes (Fig. 3H). After transfer to planting medium, these transgenic plants likewise had no phenotypic abnormalities (Fig. 3I). In contrast, other transformants from stem explants had reduced apical dominance (Fig. 4A) or had an altered leaf and/or a reduced internode length (Fig. 4B, C), and failed to produce roots and recover a normal phenotype even after 12 wk. Expression of GUS. Expression of GUS was observed in both phenotypic-normal (Fig. 4D-b) and phenotypic-abnormal transformants (Fig. 4A-C). In mature leaves from the phenotypically normal

362

SONG ET AL.

FIG . 3. Transformation, selection, and regeneration of transgenic plants from different sweet potato explant types. A, Sensitivity of stem explants to antibiotics after 6 wk of culture: a, nontransformed stem explants on CIM containing 30 mg l21 Hyg; b, nontransformed stem explants on CIM containing 50 mg l21 Km; c, transformed stem explants on CIM containing 50 mg l21 Km; d, nontransformed stem explants on antibiotic-free CIM. B, Formation of Km-resistant calluses after 8 wk of selection on CIM. C, Young shoots from Km-resistant calluses in 4 wk on RM. D, Formation of Km- and Hyg-resistant calluses from stem explants after 12 wk of selection following two-step selection. E, Induction of Hyg-resistant calluses from different explant types: a, stem explants; b, leaf disk; c, petiole explants; d, calluses. F, Regeneration of Hyg-resistant calluses on RM after 2 wk. G, Regeneration of shoots from Hyg-resistant calluses on RM in 4 wk. H, Growth and rooting of GUS-positive transformant plants on LS þ 2.85 mM IAA. I, Growth of GUS-positive transformant plants in plastic pots (20 £ 30 cm).

transgenic plants grown in planting medium, strong GUS expression was observed (Fig. 4E-a). Blue areas, as expected, were not observed in the nontransformed tissues (Fig. 4D-a, E-b). PCR and Southern blot analyses. Among independent GUSpositive transgenic lines recovered from different explant types, one regenerated shoot from each transformed explant type was selected for PCR. The expected band (987 bp) from the coding region of hpt was found in all these transformants. It was absent in the nontransformed controls (data not shown). All of the 122

GUS-positive lines tested, 99 from stem explants, 10 from leaf disks, eight from petiole explants, and five from calluses, were found to be PCR-positive. Southern hybridization of genomic DNA samples double-digested with HindIII and EcoRI was first performed on 11 GUS- and PCRpositive lines, eight from stem explants and one each from leaf disks, petiole explants, and calluses. These lines showed the expected 2.9-kb band, indicating the presence of gusA (data not shown). The number of inserts in transgenic plants was then

363

TRANSFORMATION OF SWEET POTATO TABLE 1 EFFECT OF SELECTION METHODS ON TRANSFORMATION EFFICIENCY OF STEM EXPLANTS OF SWEET POTATO CV. BENIAZUMA Antibiotic Hyg (10 mg l21) Hyg (30 mg l21) Km (50 mg l21) Km–Hygz

No. of explants

No. of explants producing resistant calluses

No. of explants producing GUS-positive shoots

No. of GUS-positive shoots (total no.)

60 60 60 60

0 0 32 16

0 0 9 16

0 0 15 (126) 30 (30)

Transformation frequency (%)y 0.0 0.0 11.2 26.7

a a b c

z

Selection first for 6 wk on CIM with 50 mg l21 Km, and then on CIM with 30 mg l21 Hyg for 10 wk. Transformation frequency (%) ¼ Number of explants producing GUS- and PCR-positive/Total number of explants inoculated £ 100. Values with different letters are significantly different at P # 0.05. y

TABLE 2 TRANSFORMATION EFFICIENCY OF FOUR EXPLANT TYPES OF SWEET POTATO (CV. BENIAZUMA) USING THE TWO-STEP SELECTION METHOD Explant Stem explants (in vitro stocks) Experiment 1 Experiment 2 Stem explants (field plants) Experiment 1 Experiment 2 Leaf disks Petiole explants Calluses (4 wk) Calluses (8 wk) z y

No. of explants

No. of explants with GUSand PCR-positive shoots

Transformation frequency (%)z

Time period for transformationy

60 60

17 20

28.3 33.3

12–16 wk 12–16 wk

60 109 89 75 126 162

15 29 10 8 4 1

25.0 26.6 11.2 10.7 3.2 0.6

12–16 wk 12–16 wk 18–26 wk 18–26 wk Over 10 mo. Over 10 mo.

Transformation frequency (%) ¼ Number of explants producing GUS- and PCR-positive plants/Total number of explants inoculated £ 100. Period from inoculation to regeneration of transgenic shoots.

determined by hybridizing the same probe with HindIII-digested DNA of another seven independent transgenic lines, four from stem explants and one each from leaf disks, petiole explants, and calluses. HindIII has a unique cleavage site near the right border in the T-DNA region in the vector (Fig. 1). Four of the seven lines had a single band with four different patterns; the other three lines had two, three, and four bands with different patterns, respectively. All the bands observed were over 3.0 kb. The pattern of Southern hybridization signals indicated one to four copies present per transgenic genome of independent GUS-positive plant lines (Fig. 5). In contrast, no signal was detected in the genomic DNA from the nontransformed control. Discussion Regeneration of sweet potato occurs at low frequencies for various explants (Liu and Cantliffe, 1984; Otani et al., 1987; Perera and Ozias-Akins, 1991; Cavalcante et al., 1994; Desamero et al., 1994; Murata et al., 1994; Dhir et al., 1998). Otani and Shimada (1996) established an efficient method to regenerate shoots (97%) for the calluses induced from shoot tips. We demonstrated that stem explants, leaf disks, and petioles of sweet potato, similar to shoot tips, produce calluses that regenerate shoots at high frequencies (, 90 – 93%). An efficient selection method is required for the production of transgenic plants. In previous reports, either npt II or hpt alone was used as the selectable marker gene for transforming sweet potato

cells (Newell et al., 1995; Gama et al., 1996; Mora´n et al., 1998; Otani et al., 1998). Prakash and Varadarajan (1992) found that some GUS-negative calluses proliferated on the Km-containing medium (50 mg l21), and suggested that npt II under control of the NOS promoter might not be an effective selectable marker. Our results indicated that sweet potato cells are more sensitive to Hyg than to Km (Table 1). Hyg (30 mg l21) appeared effective in inhibiting escape tissues and cells, but could not be used immediately after co-cultivation even when decreased to 10 mg l21. Under the two-step selection method, a relatively lower selection pressure produced by 50 mg l21 Km at the first stage of selection likely allowed the transformed cells and perhaps some untransformed cells to divide. Then, when 30 mg l21 Hyg was used in the second selection stage, strong selection pressure effectively eliminated untransformed cells, but allowed the transformed cells to continue growth. Allowing cell division and callus formation on Km-containing medium prior to Hyg selection resulted in successful regeneration of transgenic plants from stem, petiole, leaf, and callus explants (Table 2). Under the Km – Hyg selection method and using the appropriate explant type (stem explants) of genotype cv. Beniazuma, the efficiency of transformation was considerably improved over that previously published (Otani et al., 1993; Newell et al., 1995; Gama et al., 1996; Otani et al., 1998). The most successful transformation of sweet potato mediated by A. tumefaciens in previous reports was achieved on cv. Jewel at frequencies of 2.5 –10% for freshly harvested storage roots (Newell et al., 1995) and 33% for leaf

364

SONG ET AL.

FIG . 4. GUS analyses in transgenic sweet potatoes regenerated from stem explants. A, Transformants (8 wk) without apical dominance. B, Transformants (8 wk) with reduced apical dominance and altered leaves. C, Transformants (8 wk) with short internodes and small leaves. D, Expression of GUS in a transformant with a normal appearance: a, nontransformed control; b, transformant (8 wk). E, Mature leaves from a normal transformant grown in planting medium: a, transgenic leaf gently scored; b, nontransformed control gently scored.

Newell et al., 1995; Gama et al., 1996; Otani et al., 1998). In our experiments, 30.8% of inoculated stem segments of cv. Beniazuma produced GUS- and PCR-positive transformants. On the other hand, stem explants were shown to be ideal explants for the transformation of sweet potato due to their ready availability as an explant and the simple transformation process. The high efficiency in the transformation of stem explants in the popular variety Beniazuma suggests that this novel approach warrants testing for routine stable transformation of diverse varieties of sweet potato. Acknowledgments The authors would like to thank Professor T. Shimada and Dr. M. Otani, Ishikawa Agricultural College for their helpful discussions. We would also like to thank Ms. M. Fujishige and Ms. Y. Suzuki for their expert technical assistance. This work was supported by the New Energy and Industrial Technology Development Organization (NEDO), Japan.

References FIG . 5. Southern blot analyses of gusA in transformants. Total DNAs digested with HindIII and probed with GUS. Lane 5, nontransformed control; lanes 1–4, independent transgenic lines recovered from stem explants; lanes 6, 7, 8, independent transgenic plants recovered from leaf disks, petiole explants, and calluses, respectively. M, lDNA 1-kb marker.

explants (Mora´n et al., 1998). However, the 33% frequency for leaf explants by Mora´n et al. (1998) was determined as the number of Km-resistant shoots per the total number of explants, and the data are thus not comparable with other reports (Otani et al., 1993;

Cavalcante, A. J. M.; Sihachakr, D.; Allot, M.; Tizroutine, S.; Mussio, I.; Servaes, A.; Ducreux, G. Isozyme modifications and plant regeneration through somatic embryogenesis in sweet potato [Ipomoea batatas (L.) Lam.]. Plant Cell Rep. 13:437–441; 1994. Desamero, N. V.; Rhodes, B. B.; Decoteau, D. R.; Bridges, W. C. Picolinic acid-induced direct somatic embryogenesis in sweet potato. Plant Cell Tiss. Organ Cult. 37:103–111; 1994. Dhir, S. K.; Oglesby, J.; Bhagsari, A. S. Plant regeneration via somatic embryogenesis, and transient gene expression in sweet potato protoplasts. Plant Cell Rep. 17:665–669; 1998.

TRANSFORMATION OF SWEET POTATO Gama, M. I. C.; Leite, J. R. P.; Cordeiro, A. R.; Cantliffe, D. J. Transgenic sweet potato plants obtained by Agrobacterium tumefaciens-mediated transformation. Plant Cell Tiss. Organ Cult. 46:237–244; 1996. Giddings, G.; Allison, G.; Brooks, D.; Carter, A. Transgenic plants as factories for biopharmaceuticals. Nat. Biotechnol. 18:1151–1155; 2000. Jefferson, R. A.; Kavanagh, T. A.; Bevan, M. W. GUS fusions: b-glucuronidase as a sensitive and versatile gene fusion marker in higher plants. EMBO J. 6:3901–3907; 1987. Kimura, T.; Otani, M.; Noda, T.; Ideta, O.; Shimada, T.; Saito, A. Absence of amylose in sweet potato [Ipomoea batatas (L.) Lam.] following the introduction of granule-bound starch synthase I cDNA. Plant Cell Rep. 20:663–666; 2001. Lawton, R.; Winfield, S.; Daniell, H.; Bhagsari, A. S.; Dhir, S. K. Expression of green-fluorescent protein gene in sweet potato tissues. Plant Mol. Biol. Rep. 18:139a–139i; 2000. Linsmaier, E. M.; Skoog, F. Organic growth factor requirement of tobacco tissue cultures. Physiol. Plant. 18:100–127; 1965. Liu, J. R.; Cantliffe, D. J. Organogenesis and plant regeneration in tissue cultures of sweet potato (Ipomoea batatas Poir). Plant Cell Rep. 3:112–115; 1984. Mora´n, R.; Garcı´a, R.; Lo´pez, A.; Zaldu´a, Z.; Mena, J.; Garcı´a, M.; Armas, R.; Somonte, D.; Rodrı´guez, J.; Go´mez, M.; Pimentel, E. Transgenic sweet potato plants carrying the delta-endotoxin gene from Bacillus thuringiensis var. tenebrionis. Plant Sci. 139:175–184; 1998. Murata, T.; Fukuoka, H.; Kishimoto, M. Plant regeneration from mesophyll and cell suspension protoplasts of sweet potato, Ipomoea batatas (L.) Lam. Breed. Sci. 44:35–40; 1994. Newell, C. A.; Lowe, J. M.; Merryweather, A.; Booke, L. M.; Hamilton, W. D. O. Transformation of sweet potato (Ipomoea batatas (L.) Lam.) with

365

Agrobacterium tumefaciens and regeneration of plants expressing cowpea trypsin inhibitor and snowdrop lectin. Plant Sci. 107:215–227; 1995. Ohta, S.; Mita, S.; Hattori, T.; Nakamura, K. Construction and expression in tobacco of a b-glucuronidase (GUS) reporter gene containing an intron within the coding sequence. Plant Cell Physiol. 31:805– 814; 1990. Otani, M.; Mii, M.; Handa, T.; Kamada, H.; Shimada, T. Transformation of sweet potato (Ipomoea batatas (L.) Lam.) plants by Agrobacterium rhizogenes. Plant Sci. 94:151–159; 1993. Otani, M.; Shimada, T. Efficient embryogenic callus formation in sweet potato (Ipomoea batatas (L.) Lam.). Breed. Sci. 46:257–260; 1996. Otani, M.; Shimada, T.; Kimura, T.; Saito, A. Transgenic plant production from embryogenic callus of sweet potato (Ipomoea batatas (L.) Lam.) using Agrobacterium tumefaciens. Plant Biotechnol. 15:11– 16; 1998. Otani, M.; Shimada, T.; Niizeki, H. Mesophyll protoplast culture of sweet potato (Ipomoea batatas L.). Plant Sci. 53:157–160; 1987. Perera, S. C.; Ozias-Akins, P. Regeneration from sweetpotato protoplasts and assessment of growth conditions for flow-sorting of fusion mixtures. J Am. Soc. Hort. Sci. 116:917–922; 1991. Prakash, C. S.; Varadarajan, U. Genetic transformation of sweet potato by particle bombardment. Plant Cell Rep. 11:53–57; 1992. Rogers, S. O.; Bendich, A. J. Extraction of DNA from milligram amounts of fresh, herbarium and mummified plant tissues. Plant Mol. Biol. 5:69– 72; 1985. Sihachakr, D.; Ducreux, G. Plant regeneration from protoplast culture of sweet potato (Ipomoea batatas (L.) Lam.). Plant Cell Rep. 6:326– 328; 1987.