Plant Cell, Tissue and Organ Culture 77: 231–243, 2004. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.
231
Review of Plant Biotechnology and Applied Genetics
Use of tissue culture and biotechnology for the genetic improvement of watermelon Michael E. Compton1,∗ , D.J. Gray2 & Victor P. Gaba3 1 School of Agriculture, University
of Wisconsin-Platteville, 1 University Plaza, Platteville, Wisconsin 53818, USA; Research and Education Center, University of Florida, Institute of Food and Agricultural Sciences, 2725 Binion Road, Apopka, Florida 32703-8504, USA; 3 Department of Virology, Institute of Plant Protection, The Volcani Center, Bet Dagan, Israel (∗ requests for offprints: Fax: +1-608-342-1395; E-mail:
[email protected])
2 Mid-Florida
Received 12 September 2003; accepted in revised form 2 December 2003
Key words: Agrobacterium tumefaciens, Citrullus lanatus, plant breeding Abstract Watermelon is an important vegetable crop world-wide with over 81 million metric tons produced annually. Despite these high production figures, million of metric tons of fruit are lost in fields to disease. Genetic improvement through tissue culture and biotechnology offer potential routes of improving fruit harvest by offering higher quality products, like seedless fruit, or by introducing recombinant genes or generating somaclonal variants with improved resistance to biotic or abiotic stresses. The purpose of this review is to highlight how tissue culture and biotechnology have been used for the genetic improvement of watermelon and provide suggestions for future application of these methods to facilitate further genetic improvement. Introduction Most consumers enjoy watermelon, Citrullus lanatus (Thunb.) Matsum. and Nakai, because of its sweetness and flavor. However, watermelon is important throughout the world because of its high vitamin (25 and 20% of the USA recommended daily allowance of vitamins C and A, respectively, per 0.280 g) and nutrient content (8% US-RDA of potassium, 4% USRDA iron and 2% US-RDA of calcium per 280 g). Watermelon flesh is also rich in lycopene, a potent antioxidant that has been shown to reduce human risk to cancer of the prostrate, pancreas and stomach (Garster, 1997; Anonymous, 2003a). The origin of Citrullus species has been traced to tropical Africa (Decoteau, 2000). However, evidence exists that supports a separate site of origin on the North American Continent. China produces most of the worlds watermelon crop, accounting for about 71% (57.65 metric tons) of global production (Anonymous, 2003b) followed by Turkey (3.9 million metric tons), Iran (1.9 million metric tons), the US (1.78 million metric tons) and Egypt (1.45 million metric tons). In the US, watermelon is cultivated
mainly in the southeastern (Florida and Georgia), southwestern (California, Texas and Arizona) and central (Indiana) states. The 2002 watermelon crop in the US was valued at over 326 million US dollars. Fruit of watermelon cultivars vary in size, shape, rind color and pattern, flesh and seed color, and maturity date. Most watermelon cultivars are diploid and produce fruit that have a striped rind, red flesh with small black seeds, and weigh 20–30 lbs at maturity. However, triploid seedless cultivars have been available for over 50 years (Kihara, 1951; Andrus et al., 1971) and are becoming more prevalent (Lucier and Lin, 2001). Seedless cultivars are preferred by most consumers because of their sweeter taste and lack of hard seeds (Marr and Gast, 1991). Tetraploid breeding lines have been created for use in producing triploid hybrids (Andrus et al., 1971).
Types of plant regeneration systems Watermelon plants have been propagated using clonal micropropagation of shoot tips and nodes, adventitious shoot regeneration from cotyledon pieces and
232 somatic embryogenesis. The following sections explain the application of each method and summarizes the protocols that have been developed. Shoot tip micropropagation The use of shoot tip (∼1 cm) explants for clonal propagation has potential application for the propagation of elite triploid cultivars and tetraploid watermelon breeding lines. Micropropagation of triploid cultivars has potential use in the vegetable industry as a means of offsetting the high cost and difficulties often encountered when germinating triploid seed (Kihara, 1951; Elmstrom and Maynard, 1992). Several attempts have been made to adapt tissue culture procedures for the propagation of triploid genotypes (Barnes et al., 1978; Barnes, 1979; Anghel and Rosu, 1985; Adelberg and Rhodes, 1989). However, most have failed because of low shoot proliferation rates coupled with poor rooting and acclimatization. Compton and Gray (1992) observed that 1 cm shoot tips of triploid genotypes produced 2–4 axillary shoots every 21 days when cultured in MS medium containing 1 µM BA. Approximately 70–90% of shoots formed roots when transferred to medium containing 1 µM IBA or 0.1 µM NAA with 60–85% of plantlets surviving acclimatization to greenhouse and field conditions (Compton et al., 1993a). Using this protocol about 1.2 million plants could be obtained in just over 6 months from an initial 100 shoot tip explants. Mechanization has been developed for handling of triploid transplants from seeds (Swaider and Ware, 2002) and could be used to transplant micropropagated plantlets. One of the highest costs associated with micropropagation of any species is labor. Tedious handling of shoot tips is typically associated with most micropropagation systems. Many hours are spent by employees to precisely cut and transfer shoot tips to fresh medium on a regular basis. To reduce tedious handling of microshoots, Alper et al. (1994a, b) developed a method for non-selective cutting and handling of watermelon microshoots. Shoot cultures of diploid ‘Charlee’ were transferred to a ‘unitizing wire cutter’ containing a rectangular grid and cutting block adjoined to a handle fitted with a wire mesh. Shoots were randomly cut when the upper handle was pressed against the cutting block. Although 67–73% fewer shoots were obtained using the semi-automated device over hand cutting, less time was required to perform transfers (5% of time for unitized cutting v.s. 48% of total time for hand cutting). This method combined
with mechanical transplanting should greatly reduce the costs associated with micropropagating triploid genotypes. Clonal propagation of tetraploid genotypes for the use of breeding triploid hybrids currently represents the greatest potential use of micropropagation for watermelon (Gray and Elmstrom, 1991). Tetraploid watermelon genotypes are produced by treating diploid seedlings with aqueous colchicine (Kihara, 1951). Tetraploid plants commonly exhibit lower fertility compared to their diploid counterparts, especially during the first five generations of selfing, making it difficult to produce the amount of tetraploid seed required to produce a sufficient quantity of triploid hybrid seed (Kihara, 1951; Compton and Gray, 1992; McCuistion and Elmstrom, 1993; Compton et al., 1993a). Gray and Elmstrom (1991) developed and patented a process for micropropagating tetraploid genotypes for use in producing triploid hybrid seed. In this process, shoot tips (∼2 cm) harvested from stems of putative tetraploids identified in the greenhouse or field are trimmed and surface-disinfested in a bleach solution (2.6% NaOCl containing about 0.1% surfactant) for 3 min followed by several rinses in sterile distilled water. The apical meristems were excised and cultured in vessels (100 mm × 15 mm tubes or 100 mm × 15 mm Petri dishes) containing 20–25 ml (tubes or Petri dishes, respectively) of MS medium with (per liter) 30 g sucrose, 4.44 µM BA and 7 g TC agar (Phytotechnology, Kansas City, MO) at a pH of 5.7–5.8. Using this medium, meristems developed into masses containing numerous axillary buds (Figure 1A), which were transferred to medium (MS with sucrose and agar as above but with 9.3 µM kinetin and 45.7 µM IAA instead of BA) that stimulated shoot elongation (Figure 1B). Plants obtained from this method were easily acclimatized to greenhouse conditions (Figure 1C) and transferred to the field where they produced normal fruit. No somaclonal variants were observed among the R0 plants or their progeny. Further study indicated that elongation medium could be eliminated by initially culturing meristems in medium (MS with sucrose and agar as above) with reduced (1 µM) BA (Compton and Gray, 1992; Compton et al., 1993a). Shoots easily rooted once they obtained a length of at least 2 cm (Gray and Elmstrom, 1991) and over 90% of plantlets survived acclimatization in a high humidity chamber (Compton et al., 1993a). Mature plants (100%) obtained using this procedure retained their initial ploidy without any observed somaclonal variations.
233
Figure 1. Micropropagation of watermelon from shoot tip explants. (A) Freshly-plated shoot-tips consisting primarily of the apical meristem (bottom row) enlarge and develop axillary shoots on medium containing BA (top row). (B) Long-term maintenance of shoot tip derived tissue results in a mass of nodal tissue and organized callus from which shoots can be recurrently harvested for rooting. (C) Rooted shoots develop into fertile plants that can be established in pots and eventually transferred to the field.
Clonal propagation of diploid and tetraploid malesterile genotypes would be the best application of micropropagation for watermelon. Currently two male sterile genotypes have been identified (Zhang and Rhodes, 1992; Fehér, 1993; Zhang et al., 1994c). Glabrous male sterile (msg) is a line of diploid malesterile watermelon in which the trait for glabrous foliage (lacking trichomes) is closely linked or pleiotropically affected by the male-sterile locus (Watts, 1967). Although this line produces male flowers with sterile pollen, plants also display reduced fertility of female flowers. A second male-sterile genotype was created by Xia et al. (1988) following the identification of two spontaneous male-sterile mutants in a planting of self-pollinated ‘Nongmei 100’ watermelon in northeast China. The mutants were crossed with siblings to develop the breeding line G17AB, also know as ‘Chinese male-sterile’. In addition to possessing more agronomically desirable traits than the msg breeding line, male-sterility in G17AB is controlled by a single locus designated ms (Xia et al., 1988; Zhang and Wang, 1990). Use of male-sterility could greatly reduce the cost of diploid and triploid hybrid seed production by eliminating the need for controlled hand pollination. The problem with malesterile genotypes is that they require heterozygous maintenance lines (MS/ms) to produce homozygous recessive (ms/ms) male-sterile plants. Use of maintenance lines can be eliminated by employing the micropropagation scheme patented by Gray and Elmstrom
(1991) to produce ms/ms plants. Currently male-sterile genotypes with high female fertility only exists in the diploid state. Jaworski and Compton (1997) generated putative tetraploid plantlets from cotyledons of diploid msg seedlings. Unfortunately, none of the plants produced seed when crossed with selected diploid cultivars. Adventitious shoot organogenesis Regeneration of adventitious shoots has been reported from a wide range of diploid and tetraploid watermelon cultivars (Srivastava et al., 1989; Dong and Jia, 1991; Compton and Gray, 1993a). In all reports, cotyledons of in vitro-germinated seedlings were the best source of explants. To establish seedlings in vitro, embryos are extracted from quiescent seeds, disinfested for 10–25 min in 1–1.3% NaOCl followed by 3–6 rinses with sterile distilled water. Cultivars with small seed may be disinfested satisfactorily for 10– 15 min in 1% NaOCl, whereas large-seeded cultivars have required longer bleach treatment (20–25 min) and higher bleach (1.3% NaOCl) concentrations. Some have found that embryo dissection can be facilitated by soaking seeds in 2.6% NaOCl followed by a 15 h imbibition in sterile distilled water before extracting embryos (Compton and Gray, 1993a). Two- to five-day-old seedlings with their cotyledons in close contact have displayed the greatest organogenic competence (Compton and Gray, 1993a;
234 Choi et al., 1994). Embryos may be germinated in light or darkness. However, improved organogenic competence of cotyledons was reported when seedlings were germinated in darkness (Compton, 1999). The proximal end of cotyledons generally has the greatest regeneration rate (Compton, 2000). Either whole cotyledon bases or basal halves have been used as explants (Compton and Gray, 1993a; Choi et al., 1994). However, some cultivars, especially those with poor regeneration competence, regenerate best when whole cotyledon bases are used (Compton, 2000). Shoot regeneration from immature cotyledons has been reported (Zhang et al., 1994a). However, immature seeds at the correct stage of development are more difficult to obtain because of the need of mature plants in the field. Shoot regeneration from 0.5 to 1 cm hypocotyl sections has been reported but the regeneration rate was poor (Srivastava et al., 1989). Regeneration of shoots from leaves of greenhouse or field-grown plants, or micropropagated shoots would be a more desirable source of explants because of clonal uniformity. Unfortunately, attempts to regenerate shoots from leaf explants have not been successful (Compton and Gray, 1993a). All reports of watermelon shoot organogenesis involved the use of media formulated with the basic macro and micro salts plus vitamins as outlined by Murashige and Skoog (1962) with additions of (per liter) 0.1 g myo-inositol and 30 g sucrose (Srivastava et al., 1989; Dong and Jia, 1991; Choi et al., 1994; Compton, 1999, 2000). Optimum shoot regeneration has been achieved by adding 4.4–10 µM benzyladenine as the only plant growth regulator (Srivastava et al., 1989; Compton and Gray, 1993a; Choi et al., 1994; Jaworski and Compton, 1997; Compton, 1999). However, Dong and Jia (1991) reported that adding 2.85 µM indole-3-acetic acid improved the number of shoots produced per explant for some genotypes. Substituting 5–40 µM kinetin, zeatin or 2-isopentenyladenine, or 0.1–10 µM thidiazuron for BA reduced organogenic competence (Dong and Jia, 1991; Compton and Gray, 1993a). Watermelon cotyledon pieces reportedly regenerate best when cultured in medium solidified with 7–10 g TC agar (Compton and Gray, 1993a), 5 g AgarGel (Compton, 1999) or 4 g Gelrite (Choi et al., 1994) per liter. Plant regeneration declines from 75 to 50% when explants are incubated in liquid medium. Watermelon shoots form directly from buds initiated along the cut edge of cotyledon pieces (Figure 2A). Watermelon shoots appear to exhibit strong
Figure 2. Shoot organogenesis in watermelon. (A) Adventitious shoot meristems formed at the proximal end of a watermelon cotyledon 10 days after culture initiation. (B) Shoots formed from meristems 4 to 6 weeks after culture initiation; scale bar = 1 mm. From Compton and Gray (1993a) with permission.
apical dominance as only about 3–5 of the buds formed develop into shoots (Figure 2B). However, additional shoots develop after the first shoots are harvested, allowing for multiple harvests of shoots for as long as 18 months (data not shown). Shoot elongation has been a problem when cotyledon pieces are continuously cultured on medium with cytokinin. Transfer of cotyledons with shoots and buds to medium without plant growth regulators (Compton and Gray, 1993a) or 0.92 µM kinetin (Dong and Jia, 1991) has improved shoot elongation for many genotypes (Compton and Gray, 1993a, 1994). Shoots greater than 15 mm can be easily rooted in medium with 1 µM indole-3-butyric acid (Compton and Gray, 1993a) or 0.54 µM naphthaleneacetic acid (Dong and Jia, 1991). Similar results were obtained for other cucurbit spe-
235
Figure 3. Abnormalities in somatic embryos of watermelon; fused cotyledons (A), fused embryos (B), lack of an apical meristem (C) and precocious germination; scale bar = 1 mm (From Compton and Gray (1993b) with permission.).
cies (Compton et al., 2001). Acclimatization of plantlets obtained through shoot organogenesis is similar to plantlets derived from micropropagated shoot tips as described above. Acclimatization of watermelon regenerants has been relatively trouble free when the optimized conditions are provided. Plantlets should be well developed (minimum 16 mm stem) possessing a minimum of one tap root at least 1 cm in length and placed in small containers (3.3 cm × 5.1 cm plugs) containing soilless medium amended with an equal volume of course grade vermiculite or perlite (Compton et al., 1993a). The medium should be moistened and containers covered with a clear plastic humidity dome several hours before transplanting. Plantlets should be incubated in conditions similar to in vitro cultures (16-h photoperiod at 50 µm m−2 s−1 and 25◦ C). Acclimatizing plantlets under mist in the greenhouse is not recommended because plantlets can become too saturated and rapidly rot (Gray and Elmstrom, 1991). Using these conditions, about 75–90% of plantlets should survive acclimatization. The time frame required from seedling germination to acclimatization of the first plants is about 14 weeks (Compton and Gray, 1993a, 1994; Compton et al., 1996; Jaworski and Compton, 1997).
To date, all watermelon genotypes tested have demonstrated some degree of shoot regeneration competence from cotyledon explants (Srivastava et al., 1989; Dong and Jia, 1991; Compton and Gray, 1993a, 1994; Choi et al., 1994; Zhang et al., 1994a; Compton et al., 1996; Compton 1999, 2000). However, slight adjustments in the general protocol may be necessary when experimenting with new genotypes. Shoot regeneration from cotyledons of diploid seedlings has been reported to be two- to three-fold greater than from similar explants prepared from tetraploid genotypes (Compton and Gray, 1993a). Adventitious shoot regeneration from triploid cultivars has been poor (
