Plant Cell Reports (1994) 13:344-348 Plant Cell Reports �9 Springer-Verlag 1994

Genetic transformation and plant regeneration of watermelon using Agrobacterium tumefaciens

Pil S. Choi 1, 2, Wong Y. Soh 2, Youn S. Kim 1, 3, Ook J. Yoo 3, and Jang R. Liu 1

1 Plant Cell Biology Laboratory, Genetic Engineering Research Institute, KIST, Taejon, Korea 2 Department of Biology, Chonbuk National University, Chonju, Korea

Department of Life Science Biology, KAIST, Taejon, Korea

Received 20 September 1993/Revised version received 16 December 1993 - Communicated by G. C. Phillips

Abstract . Adventitious shoots formed on the proximal cut edges of different cotyledonary explants of watermelon [Citrullus lcmatus (Thunb.) Matsum. & Nakai; cvs. Sweet Gem and Gold Medal] cultured on Murashige and Skoog's (MS) medium with 1 mg1-1 6-benzyladenine (BA). Light (16-h photoperiod, about 7 Wm -2 cool-white fluorescent lamps) was essential for shoot formation. To obtain transformed plants, cotyledonary explants of 'Sweet Gem' were cocultured with Agrobacteri~'n tumefaciens LBA4404, a disarmed strain harboring a binary vector pBI121 carrying the CaMV 3KS promoter- fl-glucuronidase (GUS) gene fusion used as a reporter gene and NOS promoter-neomycin phosphotransferase gene as a positive selection marker, for 48 h on MS medium with i mg1-1 BA and 200 ~M fl-hydroxyacetosyringone. After 48 h of culture, explants were transferred to medium with 1 mg1-1 B/K 250 rng1-1 carbenicillin, and 100 rngl -i kanamycin and cultured in the light. Adventitious shoots formed on the explants after 4 weeks of culture. When subjected to GUS histochemical assay, young leaves obtained from the shoots showed a positive response at a frequency of up to 16%. Preculturing cotyledonary explants on MS medium with 1 rng1-1 BA for 5 d enhanced the competence of the cells to be transformed by Agrobacterita'n. Southern blot analysis confirmed that the GUS gene was incorporated into the genomic DNA of the GUS-posifive regenerants. The transformed plants were grown to maturity.

Introduct ion

Watermelon is one of the most important vegetable crops cultivated in many parts of the

Correspondence to." J. R. Liu

tropic and subtropic regions. The cultivated area for watermelon is second behind tomato and world production ranks third in metric tons among vegetables. Plant regeneration of watermelon via organogenesis in cotyledonary tissue cultures (Anghel and Rosu 19~5; Compton and Gray 1993a; Dong et al. 1991; Srivastava et al. 19~9) and v ia somatic embryogenesis in immature zygotic embryos (Compton and Gray 1993b) have been reported.

Recent advances in Agrobacteriurn-mediated transformation have made it possible to introduce foreign genes into various plant species to improve their productivity and quality beyond the limit of conventional breeding. However, stable transformation of watermelon has not yet been reported. This paper describes a transformation system for watermelon by cocultttring cotyledonary explants with Agrobacteri~n harboring the pBI121 binary vector carrying GUS gene of E. coli as a reporter gene.

Materials and methods

Plant materials and oalture conditions. Zygotic embryos of F1 hybrid watermelon (eva. Sweet Gem and Gold Medal) were dissected out of the mabare seeds and surface-disinfected with 70% ethanol for 1 rain and 1% sodium hypochlorite for 10 man. They were rinsed three t imes with sterile deionized-disffiled water.

The basal medium used throughout the experiments consisted of Murashige and Skoog 's (1962) inorganic salts, 100 rag1 -z myo-inositol, 0.4 mg1-1 thiamine'HC1, 3% sucrose, and 0.4% Gelrite (MS medium). The pH of all media w a s

adjusted to 5.8 before autoclaving. Twenty- f ive ml of medium was dispensed into 8 7 x 15-ram plastic Petri dishes. Nine zygotic embryos were placed in each Petri dish conf in ing the basal medium and incubated at 25"C in the dark. After 5 d incubation, seedlings 2 to 3 cm long germinated, and their cotyledons were excised and transversely cut into proximal and distal halves (cotyledonary explants).

345

Induction of adventitious shoots. To induce adventitious shoots, cotyledonary explants of 'Sweet Gem' and 'Gold Medal' were placed onto medium supplemented with either 0.02, 0.2, 1, 2, 4 or 9 mg1-1 BA in Petri dishes. Two to 4 Petri dishes were cultured per treatment with 4 explants per disk Cultures were maintained at 25"C in the light (16-h photoperiod, about 7 Wm -z cool-white fluorescent lamps) or in the dark. After 4 weeks of cutlure, the number of explants with adventitious shoots and the number of shoots formed per explant were counted under a dissecting microscope.

Transformation of AgrobacteriawT. pBI121, a binary vector carrying the CaMV 35S promoter-GUS gene-NOS terminator fusion and NOS promoter-neomycin phosphotransferase gene-NOS terminator fusion (Fig. 1) was transformed into Agrobacterium LBA 4404, a disarmed strah% by a heat shock treatment (Jefferson et aL 1987) and used for coculture with cotyledonary explants.

Transformation of cotyledonary explants. Cotyledonary explants of 'Sweet Gem' were gently punctured 10 to 15 times throughout the adaxial surface with the tip of a scalpel, and preculb.tred on medium with 1 mgl -I BA for 0, 1, 2, 3, 4, 5 or 6 d. Agrobac~er/um was grown in YEP liquid medium. Explants were cocultured with Agrobacterium at the log phase of growth in liquid MS medium with i mg1-1 BA and 200 ~M B-hydroxyacetosyringone (106 bacterial cells per ml) for 48 k (~-Hydroxyacetosyringone was f i l t e r - s t e ~ and added to medium after autoclaving.) Explants were briefly blotted with sterile Whatman filter paper and placed onto MS medium with 1 mg1-1 BA, 250 mg1-1 carbenicillin, and 100 mg1-1 kanamycin, and cuiix~red in the light as described above. (The antibiotics were added to medium after autoclaving.) Five Petri dishes were cultured per treatment (explant preculture for 0 to 6 d) with 9 explants per disk After 4 weeks of culture, young leaves were excised from 20 to 30 adventitious shoots per treatment and incubated with 5-bromo-4-chloro-3-indolyl- ~ -D-glucuronic acid (X-gluc) at 37"C for 24 h according to the Clone Tech manual Data recorded at 4 weeks included the percentage of explants kanamycin-resistant shoots and the frequency of histochemical GUS-positive shoots. Statistical analysis was conducted using the GLM procedure.

Southern blot analysis GUS-positive and GUS-negative adventitious shoots were separately transferred onto MS basal medium to be rooted. The regenerants were transplanted to potting soil and maintained in the phytotron (RH 60-70%, 27"C day/22"C night, 16-h photoperiod, about 80 Win-2). The genomic DNA of GUS-positive and GUS-negative regenerants was extracted according to the method of Deblaere et aL (1987), digested with EcoRI, HindIII, BarnI-II or E c o R I ] ~ for 4 h and subjected to elec~'ophoresis (0.8% agarose gel). The DNA bands were transferred to positive-charged Nylon membrane, and a 2.1 kb GUS NOS poly(A) probe labeled with digoxigenine (DIG; Boehringer Mannheim) was used for Southern hybridization.

Results and Discussion

Induca'on of adventitious shoots

After one week of culture, etiolated cotyledonary explants turned green and enlarged 2 to 3 times. After 2 weeks of culture, callus formed on the cut edges of the explants (Fig. 2A). After 3 weeks of culture, callus on the proximal cut edges of both of the proximal and distal halves

RB ATG TGA L B

�9

Fig. 1. A binary vector harboring E. coli ~-glucuronidase gene. LB, T-DNA left border; RB, T-DNA right border; NOS-pro, nopaline synthase promoter; NPT II, neomycin phosphotransferase gene; NOS-ter nopaline snythase terminator; GUS, ~-glucuronidase gene. The dotted line indicates the 2.1-kb EcoRI/B~rnHI DNA fragment used as a Southern hybridization probe.

of the cotyledonary explants gave rise to leaf primordia (Fig. 2B), indicating that the cotyledon has a polarity for competence of adventitious shoot formation. After 4 weeks of culture adventitious shoots were grown to about i cm in height. The number of adventitious shoots explant and the percentage of explants with shoots were greatest when the explants were cultured on medium with 1 rng1-1 BA for both cultivars (Fig. 3). The percentage of 'Sweet Gem' explants with shoots was higher than tha t of 'Gold Medal' explants on the whole. Likewise, 'Sweet Gem' produced about three times more shoots than 'Gold Medal' at the same concentration of BA and was more competent to form shoots over a wider range of BA concentrations. This finding is similar to that of Compton and Gray (1993a) and Srivastava et al. (1989) who reported optimal shoot formation on medium with 5 ~M BA. The distal half cotyledonary explants produced more shoots than the proximal half (in 'Sweet Gem', it was about 2 times higher; in 'Gold Medal', about 1.25 times higher), which is in contrast to the results of Compton and Gray (1993a). These researchers obtained adventitious shoots only on the proximal region of the cotyledons, and suggested that the competence for adventitious shoot formation in watermelon was restricted to the proximal region of cotyledons. However, our results indicate that cells competent for shoot formation are not localized at one site of the cotyledon in these cultivars. Light was essential for adventitious shoot formation, because no cotyledonary explants of either cultivar formed shoots in the dark (data not shown).

3 4 6

Fig . 2. Adventitious shoot formation on cultured cotyledonary explanta of watermelon and transformed plant regeneration following coculture with Agrobacterium. A: Green callus formed on the distal half cotyledon; B: Numerous shoots formed on the explant after 3 to 4 weeks of cul0Jre; C: Transformed plants Wansplanted to potting soil; D: A transformed plant bearing flower buds.

100 7-0

tn

Q.

I,~ 50- 1,5C~

0 0

B

0.2 1.0 2.0 4D 9.0 ~02 0.2 1,0 2,0 4.0 9.0

BA Conc. (rag I-~)

0.02

Fig. 3. Effect of BA concentration on adventitious shoot formation on the distal (rT) and proximal (m> ~ e ~ of 'Sweet Gem' (A) and 'Gold Medal ' (B) cotyledonary explants. Explants were cultured on MS medium with various concentrations of BA in the light and data were collected after 4 weeks of culture. Vertical bars represent standard error of the mean.

60

~ 50

40

"~ 3o

"'~ 2o

io

0

8

~.0 0

X 0 0

/ § o . ~

/ �9 ~ = 0 .7~*

o B

y = 1.5 - 5 . 2 x + 3 . 2 1 x =

=-:2: \

�9 9 2 4 ;

Days

347

Transformation of cotyleclonary explants

The percentage of explants forming kanamycin-resistant shoots after coculture with Agrobacteri~'n was increased by preculture of explants on medium with 1 rng1-1 BA for up to 5 (Fig. 4). The proportion of GUS-pesitive shoots relative to total adventitious shoots subjected to GUS histochemical assay was greatest (16%) for the 5 d preculture treatment. Thus, preculturing explants for 5 d made cells more competent for Agrobacter/um-mediated transformatiorr After 2 weeks of culture, all GUS-positive shoots rooted on MS basal mediurrL Seven regenerants were transplanted to petting soil and grown to maturity in the phytotron (Fig. PC, D).

Fig. 4. Effect of explant Ia'ecultare m the kanamycin-resistant shoot formation (O) and transfccmafion efficient (O) of 'Sweet Gem' cotyledonary explants followir~ coculture with AgroSac~r/z~n. The e.xplauts were preculmred on MS rr~dium with 1 roll -z BA before cocul~tre with Agrobacter/z~n harborir~ pBI121 binary vector. Data wese collected after 4 weeks of culixu~. The tmnsformatkm efficieucT is obtained from the number of hist~cal GUS-lx~itive shoots divided by the number of slx~ts subjected to X-gluc assay.

* Significant at P = 0.05.

Fig. 5. Southern blot analysis of GUS-positive and GUS-negative regenerants of watermelon. The 2.2 kb GUS NOS poly(A) probe labeled with DIG was hybridized with genomic DNA of regeneraats. Each lane was loaded with the following DNA. A. Lane 1:pBI121 digested with EcoRI/BomHI; Lane 2: a GUS-negative regenerant; Lane 3: GUS-positive regenerant I digested with EcoRI; Lane 4: GUS-positive re~enerant digested with EooRI/BarnHI; Lane 5: GUS-positive regenerant IT digested with EcoRI; Lane 6: GUS-posifive regeaerant II digested with EcoRI/BamHI ; Lane 7 : GUS-positive regenerant 1]I digested with EooRI; Lane 8: GUS-positive regenerant III digested with EcoRUBarnHI. B:Lanes 1 and 2 are the same to those of "A". GUS-positive regenerant ]~ digested with EcoRI (Lane 3), HindI~ (Lane 4), and BarnHI (Lane 5), respectively.

348

When the genomic DNA of three randomly selected GUS-posifive regenerants was digested with EcoRI or BamHI and subjected to Southern blot analysis using DIG-labeled GUS NOS poly(A) probe, only one band of about 23 kb greater in length than the intact pBI121 was obtained, indicating that the GUS gene was incorporated into the genomic DNA of the regenerants (pBI121 has one EcoRI site and the total size of the vector is 12.7 kb) (Fig. 5A; Lanes 3, 5 and 7). One band of 2.2 kb obtained by EcoRI or BamHI digestion as expected showed that the incorporated DNA fragment included the intact GUS gene (Fig. 5A; Lanes 4, 6 and 8). The genomic DNA of regenerant III digested with HindIII yielded one major band of 9.2 kb and several minor bands of smaller sizes (Fig. 513; Lane 4). The major band confirmed that the GUS gene was incorporated into the genomic DNA of the regenerant. The minor bands were probably due to non-specific binding of the probe with the genomic DNA.

At presen~ using the protocol described above, we are attempting to introduce a partial cDNA clone (0.7 kb) of sweet potato ADP-glucose pyrophosphorylase we cloned (unpublished) into watermelon in a reverse (antisense) orientation,

~ which we expect to produce high sugar content fruits as recently demonstrated for potato (Mtiller-RSer et al., 1992). We are also trying to

introduce chitinase gene of bean to produce fungal disease-resistant watermelon. Such an approach has been tested in tobacco (Broglie et al., 1991).

Acknowledgements. This work was supported by a grant (G70580) from the Korea Ministry of Science and Technology to J.R.L We thank Drs. Daniel J. Cantliffe, Kyung H. Paek and Sang S. Kwak for their critical reading of this manuscript, and Dr. Kyung S. Ko for his gift of ~- hydroxyacetosyringone. We are also grateful to Miss Chang S. Kim for her assistance in preparation of this manuscript.

References

Anghel I, Rosu A (1985) Rev Roum Biol-Biol V~g~t 30:43-55 Broglie K, Chet I, Hollyday M, Cressman R, Biddle P,

Knowltion S, Mauvais CJ, Brgolie R (1991) Sdence 254: I194-i197

Compton ME, Gray DJ (1993a) J Amer Soc Hort Sci 118:151-157

Compton ME, Gray DJ (1993b) Plant Cell Rep 12:61-65 Deblaere R, Reynaerts A, Hofte H, Hemalsteens JP, Leemans

J, Van Montagu M (1987) Methods in Enzymol 153:277-292

Dong J-Z, Jia S-R (1991) Plant Cell Rep 9:559-562 Jefferson RA, Kavanaugh TA, Bevan MW (1987) EMBO J

6:3901-3907 Mtiller-RSber B, Sonnewald U, Willmitzer L (1992) EMBO J

11:1229-1238 Murashige T, Skoog F (1962) Physiol Plant 15:156-163 Srivastava DR, Andrianov VM, Piruzian ES (1989) Plant Cell

Rep 8:300-302