Introduction

The melon (Cucumis melo L.) has eight cultivated subspecies, namely, vars. cantalupensis, chito, conomon, dudaim, inodorus, flexuosus, reticulatus and makuwa. Of these, the oriental melon (Cucumis melo L. var. makuwa) is cultivated in large areas of Asia, especially throughout the temperate regions of China, Japan and Korea.

In melon, as in many other horticultural crops, cross-breeding is the most common approach to produce new cultivars. However, the effectiveness of conventional breeding methods is restricted in melon due to the limited number of established cultivars as well as the difficulty in obtaining hybrids between varieties. These factors have led breeders to adopt alternative approaches, including micropropagation and genetic engineering. Consequently, successful plant regeneration in melon has been reported using cotyledon (Dirks and Buggenum 1989; Niedz et al. 1989; Ntui et al. 2009) and leaf (Kathal et al. 1988; Thiruvengadam et al. 2010; Yadav et al. 1996) explants, and Agrobacterium-mediated genetic transformation, in particular, is now becoming an acceptable alternative breeding option in melon breeding programs.

Agrobacterium tumefaciens-mediated transformation and plant regeneration of several melon varieties has been achieved via cotyledon or leaf explants (Akasaka-Kennedy et al. 2004; Bordas et al. 1997; Chovelon et al. 2011; Fuchs et al. 1998; Galperin et al. 2003; Nuñez-Palenius et al. 2006; Ntui et al. 2010a, b; Ren et al. 2012; Vallés and Lasa 1994). Bordas et al. (1997) reported transformed melon plants that overexpressed the yeast HAL1 gene related to salt tolerance. Resistance to various plant diseases has also been achieved in transgenic melon through the incorporation of genes coding for the coat protein (CP) gene of various plant viruses, including cucumber mosaic virus, watermelon mosaic virus, zucchini yellow mosaic virus (ZYMV) (Fuchs et al. 1998) and watermelon silver mottle virus (Huang et al. 2011). Other researchers have reported the development of genetically engineered melon plants through transformation with the ACC oxidase antisense gene (Nora et al. 2001; Nuñez-Palenius et al. 2006). There is, however, relatively little published data on the regeneration and/or transformation of oriental melon, with most of such information originating from the studies of Moon et al. (2000) on the regeneration of whole oriental melon plants using cotyledon explants and Wu et al. (2009) on the regeneration of transgenic melon resistant to ZYMV.

To date, kanamycin (Km) has been the most widely used selection agent in plant transformation strategies, including melon (Bordas et al. 1997; Fuchs et al. 1998; Galperin et al. 2003; Nora et al. 2001; Ntui et al. 2010a, b). The stringency of Km selection for transgenic shoots is, however, highly species dependent, and in melon, the emergence of high rate of false positive and non-transgenic shoots is a major problem (Gaba et al. 2004; Ntui et al. 2010a). Wu et al. (2009) developed an improved cotyledon-cutting method and reported achieving a reduced frequency of false positive and non-transgenic shoots. Nevertheless, there is an urgent need for a selection agent that will reduce the reduction of “escapes” and thus enable the efficient selection and production of transgenic oriental melon plants.

The aim of this study was to determine both the optimal conditions for plant regeneration via organogenesis from cotyledon explants excised from seeds of the oriental melon (C. melo L. var. makuwa) and the best concentration of antibiotic(s) for the selection of transgenic plants using the neomycin phosphotransferaseII (nptII) gene. Based on the results, we have developed an efficient and successful transformation protocol for oriental melon using geneticin (Gt) as the selection agent.

Materials and methods

Plant materials and regeneration

Six cultivars of oriental melon (C. melo var. makuwa), namely, ‘Geumrang’, ‘Geumssaragi’, ‘Geumguan’, ‘Geumdong’, ‘Gohyang’ and ‘Geumnodajieunchun’, which were purchased from various Korean seed companies, were used for the shoot regeneration experiment. Seeds were surface sterilized for 20 min in a diluted commercial bleach solution [Clorox®; 0.6 % NaOCl (v/v) final concentration] containing three drops of Tween-20 per 50 mL of solution, followed by three washes with sterile distilled water. The seeds were then germinated on half-strength (1/2) Murashige and Skoog (MS) medium (Murashige and Skoog 1962) solidified with 0.8 % agar (Duchefa Biochemie, Haarlem, the Netherlands) in the dark. 5 days after germination, actively growing cotyledons were cut transversely across the midrib into segments and the segments (explants) placed on shoot induction medium [SIM; MS basic medium with 30 g L−1 sucrose and 7 g L−1 agar supplemented with 1.0 mg L−1 benzyladenine (BA) and 0.01 mg L−1 indoleacetic acid (IAA)]. The pH of the medium was adjusted to 5.8 prior to the addition of agar and then autoclaved at 121 °C for 15 min. Following an initial 3-week culture in the dark, the explants were cultured under continuous light at 24 °C. All cultures were maintained for 4 weeks without subculture to fresh medium.

The regenerated shoots were transferred onto shoot elongation medium (SEM; MS basic medium with 30 g L−1 sucrose and 7 g L−1 agar supplemented with 0.5 mg L−1 BA), and the resulting elongated shoots were then cultured onto the hormone-free MS medium to induce roots. The rooted shoots were washed with tap water and planted in pots containing autoclaved artificial soil (vermiculite:perlite, 1:1). After acclimation, they were transplanted into pots containing soil and cultivated in a greenhouse.

Determination of optimal Km and Gt concentrations

To determine the optimal concentrations of Km and Gt on shoot induction and callus formation, we prepared cotyledon explants of cv. ‘Geumnodajieunchun’ from 5-day-old seedlings grown on 1/2 MS as described above and placed the segments on SIM supplemented with different concentrations of Km (0, 1, 5, 10, 50, 100, 250, 500, 750 and 1,000 mg L−1) and Gt (0, 0.1, 0.5, 1, 2.5, 5, 10, 25, 50 and 100 mg L−1), with the abaxial surface of the explants in contact with the medium. All antibiotics were filter sterilized and added to the cooled medium (<50 °C) after autoclaving. After 3 weeks of culture in the dark, tissues were cultured in continuous light (40–50 μmol m−2 s−1 light intensity) and maintained for 4 weeks without subculture to fresh medium.

Agrobacterium-mediated transformation and plant regeneration protocols

The avirulent A. tumefaciens strain LBA4404 (Hoekema et al. 1983) harboring a binary plasmid pBI121 that contains the β-glucuronidase (GUS) gene and nptII gene was used for transformation. A glycerol stock of A. tumefaciens was thawed and then streaked onto a solid yeast extract peptone (YEP) medium containing 100 mg L−1 Km and 100 mg L−1 rifampicin. A single isolated colony was used to initiate a culture in 5 mL YEP liquid medium, and successive 50 mL cultures were each inoculated with 1 mL of this seed culture and then incubated for 24 h at 28 °C with shaking (200 rpm). Cotyledon explants of cv. ‘Geumnodajieunchun’ obtained from 5-day-old seedlings were infected with Agrobacterium (at OD600 0.6–0.7) at 25 °C and co-cultured on SIM for 2 days in the dark. After co-cultivation, the explants were washed three times with the YEP liquid medium containing 250 mg L−1 cefotaxime to remove any excess Agrobacterium and then cultured on a SIM medium containing the selection agents for 3 weeks without subculture to fresh medium. The effects of 100 mg L−1 Km and 30 mg L−1 Gt on shoot regeneration from transformed explants were then compared in terms of selection efficiency and production of transgenic plants. The regenerated shoots were transferred to SEM supplemented with 100 mg L−1 Km or 30 mg L−1 Gt for elongation, and individual elongated shoots were then re-cultured on MS medium containing 50 mg L−1 Km for rooting. The plantlets with developed roots were cultivated in the greenhouse.

Histochemical GUS assay

The histochemical analysis of GUS activity in the transformed cotyledon explants was performed according to Jefferson et al. (1987). In brief, explants were incubated overnight at 37 °C with 0.1 mM 5-bromo-4-chloro-3indolyl-β-d-glucuronide (X-Gluc) in 50 mM sodium phosphate buffer (pH 7) supplemented with 0.5 M potassium ferricyanide, 10 mM EDTA and 0.1 % triton X-100. The GUS reaction was stopped by rinsing with 70 % ethanol until pigments, such as chlorophyll, cleared completely. Photographs were taken using a camera (DSC-T1; Cannon, Tokyo, Japan).

Polymerase chain reaction

Total genomic DNA was extracted from young leaves of transgenic plants using the DNeasy Plant kit (Qiagen, Hilden, Germany). The presence of the nptII gene in transgenic plants was confirmed by PCR using the primer set 5′-GAGGCTATTCGGCTATGACTG-3′ (forward) and 5′-ATCGGGAGCGGCGATACCGTA-3′ (reverse). The amplification was carried out in a DNA thermal cycler 480 (Perkin-Elmer, Foster City, CA); the PCR program was set at 94 °C for 5 min, followed by 30 cycles of 94 °C for 30 s, 55 °C for 30 s, and 72 °C for 30 s, with a final extension at 72 °C for 5 min.

Southern blot analysis

Genomic DNAs isolated from transgenic and wild-type (WT) plants were digested with HindIII, and the digestion products of each line were separated in a 1 % agarose gel and then blotted onto Hybond N+ membrane (GE Healthcare UK Ltd, Buckinghamshire, UK). The nptII PCR product was used as a probe. Labeling, hybridization and detection were performed using the AlkPhos Direct Labeling and Detection System with CDP-Star (GE Healthcare UK Ltd, Buckinghamshire, UK) according to the manufacturer’s instructions. Hybridization was carried out at 65 °C overnight. Stringent washes were performed using a primary wash buffer at 65 °C for 20 min and a secondary wash buffer at 25 °C for 10 min. The signal was detected on an X-ray film after an exposure time of 24 h.

Experimental design and statistical analyses

Treatments were replicated in five Erlenmeyer flasks, each containing six cotyledon explants excised from independent seedlings. Both the number of explants with shoots and the number of shoots per explant were counted at 2-week intervals, but only data collected at the 8-week time point were used for data analysis. The data were statistically analyzed using analysis of variance (ANOVA) and Duncan’s multiple range test (P < 0.05) using the SAS statistical program (SAS Institute, Cary, NC).

Results and discussion

Plant regeneration via organogenesis

In the first step to establish a simple system for Agrobacterium-mediated transformation of oriental melon using cotyledon explants, we determined the optimal concentrations of plant growth regulators required to induce shoot development from cotyledon explants. Repeated experiments involving a series of different concentrations of BA and IAA revealed that the best combination was 1.0 mg L−1 BA and 0.01 mg L−1 IAA (data not shown). These concentrations differ considerably from those previously reported by Moon et al. (2000) where 0.5 mg L−1 BA and 0.1 mg L−1 IAA were optimal, possibly due to the different cultivars used in the two studies. However, they are consistent with data obtained earlier in our laboratory (Choi et al. 2001) showing that combinations involving high concentrations of cytokinin and low concentrations of auxin promote adventitious shoot formation on leaf explants in persimmon. Similar results recently reported that the highest rate of shoot induction was achieved when the explants of melon inbred lines were cultured on MS medium supplemented with 1.5 mg L−1 BA and 0.1 mg L−1 IAA (Zhang et al. 2011).

Successful shoot formation in plant regeneration systems depends on many factors, such as genotype, explant source, gelling agent and hormone concentration (Bell et al. 2011; Boszoradova et al. 2011; Choi et al. 2001; Cogbill et al. 2010; Ivanova and Van Staden 2011). Among these factors, genotype is the one essential factor influencing the efficiency of plant regeneration. The frequency of shoot regeneration could fluctuate from 50 to 90 % depending on melon genotypes tested (Ficcadenti and Rotino 1995). Galperin et al. (2003) also reported a wide range of regeneration frequency in the five melon genotypes included in their study. The oriental melon varieties evaluated in our study all showed a similar range in regeneration frequency, with no large differences among the six genotypes: ‘Geumrang’, ‘Geumssaragi’, ‘Geumguan’ and ‘Geumdong’ all achieved 100 % shoot regeneration, while ‘Gohyang’ and ‘Geumnodajieunchun’ achieved >90 % regeneration (Fig. 1). This result is consistent with that the shoot regeneration frequency reported by Moon et al. (2000) for oriental melon cultivar ‘Tongilhwang’. The numbers of shoots induced on explants of the different cultivars were >3.6 in ‘Geumrang’ and ‘Geumssaragi’, >3.3 in ‘Geumguan’, >2.9 in ‘Geumdong’ and >1.8 in ‘Gohyang’ and ‘Geumnodajieunchun’ (Fig. 1).

Fig. 1
figure 1

Effects of genotype on shoot regeneration from cotyledon explants of oriental melon. Cotyledon explants at 5 days after germination were used. Regeneration frequencies and numbers of shoots were counted after 8 weeks of cultivation on shoot induction medium. GR Geumrang; GG Geumguan; GS Geumssaragi; GD Geumdong; GH Gohyang; GE Geumnodajieunchun. Means for each of frequency of regeneration and number of shoots/explant with the same letter are not significantly different using Duncan’s multiple range test (P < 0.05)

Full size image

The cotyledon explants produced many shoot buds, and these buds elongated under suitable conditions. Buds isolated from each explant were cultured on SEM containing a low concentration of cytokinin (0.5 mg L−1 BA), resulting in elongated shoots. In oriental melon, this step is indispensable to obtaining elongated shoots from small buds, as has been described in watermelon (Dong and Jia 1991). The elongated shoots were then transferred to hormone-free MS medium for root formation. Over 90 % of shoots formed roots after 2 weeks of culture, with the average number of primary roots per shoot ranging from 5 to 7. After acclimation, the plantlets were transplanted to pots.

Effects of Km and Gt concentration

One of the most commonly used selection marker genes for screening transgenic plants is nptII, which encodes a phosphotransferase capable of phosphorylating aminoglycoside antibiotics, including Km, Gt, neomycin and paromomycin (Yoshikura 1989). To date, Km has been the most widely used of these aminoglycoside antibiotics in melon transformation protocols (Bordas et al. 1997; Fuchs et al. 1998; Galperin et al. 2003; Nora et al. 2001). In plant transformation systems, however, the balance between antibiotic concentration to achieve stringent selection and genuine transgenic shoot regeneration is highly species- and cultivar-dependent, thereby necessitating optimization of the chosen antibiotic(s) prior to transformation. The optimal concentrations of Km and Gt for shoot induction were determined by testing the explants on SIM supplemented with different concentrations of Km and Gt.

Shoot formation from cotyledon explants was not affected by 1.0–25 mg L−1 of Km. The frequency of shoot regeneration decreased rapidly with increasing Km concentration, and shoots were not formed from cotyledon explants cultured at 50 mg L−1 Km. However, explants retained their green color even at higher concentrations of Km (50–750 mg L−1 Km) (Fig. 2A). The selection of resistant shoots by Km is known to be critically sensitive to minute changes in Km concentration. In a study of genetic transformation and shoot regeneration in honeydew and western shipper cantaloupe melons, Ren et al. (2012) also reported that some tissues produced calli and shoot primordia on 100–150 mg L−1 Km but that shoot regeneration was completely inhibited on 150 and 200 mg L−1 Km. In contrast, in other studies in which Km was used for the selection of regenerated melon plants following genetic transformation, the authors report optimum concentrations that range from 50 mg L−1 [‘BU-21/3’ (Galperin et al. 2003)] to 100 mg L−1 [‘Pharoy’ and ‘Amarillo Canario’ (Bordas et al. 1997); ‘Silver light’ (Wu et al. 2009); ‘Egusi’ (Ntui et al. 2010a, b)]. Based on our results and those of these other studies, 100 mg L−1 Km was therefore chosen as the suitable concentration for the selection of oriental melon transformants.

Fig. 2
figure 2

Effects of kanamycin (Km) and geneticin (Gt) concentrations on shoot regeneration from cotyledon explants of oriental melon. Regeneration was determined after 8 weeks of cultivation on regeneration medium. The antibiotic concentrations were 0, 1, 5, 10, 50, 100, 250, 500, 750 or 1,000 mg L−1 Km (A) and 0, 0.1, 0.5, 1, 2.5, 5, 10, 25, 50 or 100 mg L−1 Gt (B). Means for each of frequency of regeneration and number of shoots/explant with the same letter are not significantly different using Duncan’s multiple range test (P < 0.05)

Full size image

Geneticin is a more active antibiotic and has been reported to be much more sensitive than Km in the regeneration and selection of nptII-transformed apple tissue (Norelli and Aldwinckle 1993). Petri et al. (2005) reported that Gt inhibited the regeneration of apricot leaves at almost all concentrations tested. In our experiments, calli, but not shoots, were formed at a low concentration (10 mg L−1) of Gt (Fig. 2B), while both callus and shoot formation were completely inhibited at 25 mg L−1 Gt. Taking these results into consideration, we therefore chose 30 mg L−1 Gt as the suitable concentration for the selection of oriental melon transformants.

Genetic transformation and plant regeneration

Following transformation, callus formation was initiated from the cut ends of the cotyledon explants within 2 weeks of culture initiation, and >90 % of the explants formed calli after 6 weeks of culture on SIM containing antibiotics (Fig. 3a, Table 1). After approximately 7–8 weeks of culture, the callus cluster covered the cotyledon explant, and adventitious shoots appeared on the callus (Fig. 3b). Adventitious shoots were elongated on SEM supplemented with antibiotics, and the small and multiple buds were sub-cultured at 3-week intervals in fresh SIM with antibiotics.

Fig. 3
figure 3

Agrobacterium-mediated transformation and plant regeneration from cotyledon explants of oriental melon. Cotyledon explants transformed with avirulent A. tumefaciens strain LBA4404 containing the binary plasmid pBI121 produced adventitious shoots on shoot induction medium containing 30 mg L−1 Gt and 500 mg L−1 cefotaxime (a, b). The adventitious shoots were elongated on MS medium containing 0.5 mg L−1 benzyladenine (BA), 30 mg L−1 geneticin, and 250 mg L−1 cefotaxime (c) and then transferred to MS medium with 50 mg L−1 Km for rooting (d). The plantlets were successfully acclimatized in the greenhouse (e). β-Glucuronidase (GUS) activity was detected histochemically as blue-stained tissue in cotyledon explants with callus (f), in adventitious shoots (g) and in plantlets (h)

Full size image
Table 1 The effects of kanamycin and geneticin antibiotics on regeneration and transformation efficiency in Agrobacterium-mediated transformation
Full size table

Shoot regeneration frequencies were 41.2 % at 100 mg L−1 Km and 15.2 % at 30 mg L−1 Gt at 8 weeks after transformation. However, large numbers of shoots cultured on medium with Km was bleached at 16 weeks after transformation (Table 1). Thus, the frequencies of elongated shoots were reduced to 6.6 % at 100 mg L−1 Km and 9.8 % at 30 mg L−1 Gt at 16 weeks after transformation. That is, more false positive shoots appeared on selection medium supplemented with Km than on that containing Gt. Our results are consistent with those of Ntui et al. (2010a) who reported that the frequency of false positive shoots produced in ‘Egusi’ melon was about 75–93 % of regenerated shoots on selection medium with 100 mg L−1 Km. The false positive shoots regenerated from SIM containing either Km or Gt were unable to elongate and form roots, and the leaves of such shoots gradually turned yellow or white and eventually died. In contrast, 16 weeks after transformation, the transformation frequencies based on the PCR assay were 2.9 % at 100 mg L−1 Km and 7.1 % at 30 mg L−1 Gt.

Individually elongated, healthy green shoots of >1 cm in length that developed on selection medium were classified as independent transgenic lines (Fig. 3c) and multiplied on SEM containing antibiotics for further analysis. These independent lines were then rooted on MS medium supplemented with 50 mg L−1 Km (Fig. 3d). The plantlets which had well-developed roots were then selected and transferred to soil in a pot (Fig. 3e).

Transformation was confirmed by the histochemical GUS activity assay. The calli of explants sampled after 4 weeks of culture on selection medium were light blue (Fig. 3f), while the roots and leaves of shoots that had survived on SIM containing Gt were dark blue (Fig. 3g, h). Integration of nptII was investigated by PCR analysis, revealing the amplification of a 700 bp fragment from GUS-positive transgenic lines and the positive control, but not from untransformed WT plants (Fig. 4a). Southern blot analysis of three putative transgenic plants confirmed the stable integration of nptII into the genome of oriental melon as a single copy (Fig. 4b).

Fig. 4
figure 4

PCR and Southern blot analysis of transgenic oriental melon. PCR analysis of putative transgenic plants (a). Genomic DNAs isolated from putative transgenic plants were subjected to PCR amplification with the neomycin phosphotransferaseII (nptII) gene. Lane L: 1 kb ladder (Gibco-BRL Life Technologies, Grand Island, NY), P: positive control (plasmid), W: wild-type plant, lanes 1–4: putative transgenic plants. Southern blot analysis of transgenic plants (b). Genomic DNAs were digested with HindIII and hybridized with the nptII probe. Lane L: 1 kb ladder, W: wild-type plant, lanes 1–3: transgenic plants

Full size image

To summarize, we have compared the efficiency of two well-known antibiotics, Km and Gt, as selection agents for the selection and regeneration of transgenic shoots of oriental melon. While a larger number of shoots were regenerated from the explants cultured in SIM containing 100 mg L−1 Km than from those cultured in SIM containing 30 mg L−1 Gt, a large number of the former were found to be non-transformed or escaped shoots. In comparison, the selection process in SIM containing 30 mg L−1 Gt reduced non-transgenic shoot production by about 30 % compared to Km. Integration of the exogenous gene into the genome of oriental melon was confirmed by GUS, PCR, and Southern blot analysis. These results suggest that Gt is the more suitable of the two agents for the production of transgenic plants of oriental melon (C. melo L. var. makuwa).