Plant Cell Reports (1995) 14:694-698 PlantCell Reports �9 Springer-Verlag 1995

Biolistic transformation of chrysanthemum with the nucleocapsid gene of tomato spotted wilt virus

Luz Marcela Yepes 1, Veronica Mittak 1, Shenk-Zhi Pang 1, Carol Gonsalves ~, Jerry L. Slightom 2, and Dennis Gonsalves 1

1 Department of Plant Pathology, Cornell University, New York State Agricultural Experiment Station, Geneva, NY 14456, USA z Mol. Biology Unit 7242, The Upjohn Company, Kalamazoo, MI 49007, USA

Received 4 October 1994/Revised version received 30 January 1995 - Communicated by R. Gilbertson

Abstract In vitro regeneration and biolistic transformation procedures were developed for several commercial chrysanthemum Dendranthema grandiflora Tzvelev, syn. Chrysanthemum morifolium Ramat. cultivars using leaf and stem explants. Studies on the effect of several growth regulators and kanamycin on chrysanthemum regeneration were conducted, and a step-wise procedure to optimize kanamycin selection and recovery of transgenic plants was developed. A population of putative transformed chrysanthemum plants cvs. Blush, Dark Bronze Charm, Iridon, and Tara, was obtained after bombardment with tungsten microprojectiles coated with the binary plasmid pBIN19 containing the nucleocapsid (N) gene of tomato spotted wilt virus (TSWV) and the marker gene neomycin phosphotransferase (NPT lI). PCR analysis of 82 putative transgenic plants selected on kanamycin indicated that the majority of the lines (89%) were transformed and contained both genes (71%). However, some transgenic lines contained only one of the genes: either the NPT II (15%) or the TSWV (N) gene (14%). Southern blot analysis on selected transgenic lines confirmed the integrat ion of the TSWV (N) gene into the chrysanthemum genome. These results demonstrate the development of an efficient procedure to transfer genetic material into the chrysanthemum genome and selectively regenerate transgenic chrysanthemum plants at frequencies higher than previously reported.

Introduction Genetic engineering offers great potential for the improvement of chrysanthemum, one of the world's most important flower crops, allowing the introduction of new traits, such as: disease resistance, insect resistance, and novel flower color. Regeneration methods using somatic tissues have been documented for chrysanthemum (Fukai and O~ 1986, Fukai et al. 1987, Kaul et al. 1990, Lu et al. 1990, Roest and Bokelmann 1975, Urban et al. 1994). However, these regeneration protocols are genotype- dependent and difficult to adapt directly to new cultivars (de Jong et al. 1993, Fukai et al. 1987).

Transformation methods using A g r o b a c t e r i u m tumefaciens have been reported for chrysanthemum using marker genes (Ledger et al. 1991, Pavingerova et al. 1994, Renou et al. 1993, van Wordragen et al. 1991), chalcone synthase sense and antisense gene constructs (Courtney- Gutterson et al. 1994), and the nucleocapsid (N) gene of a tomato spotted wilt virus (TSWV) isolate from dahlia (Urban et al. 1994). In this last report, the frequency of 'putative' transformed regenerants cv. Iridon recovered after selection on kanamycin (50 mg/L) was 4-7%, but only 41% of them were confirmed as transformed following PCR analysis for the N-gene. Other groups have failed to recover transformed chrysanthemum plants using Agrobacterium-mediated transformation with genotypes that are not easily regenerated (de Jong et al. 1993, and

�9 Lowe et al. 1993), which emphasizes the need to optimize regeneration procedures.

Biolistic-mediated transformation (Sanford et al. 1993) has been used for several plant species to overcome the problem of strain specif ici ty encountered with Agrobacter ium. We report here the first successful biolistic transformation of chrysanthemum with the TSWV (N) gene. We have developed efficient and reproducible in vitro regeneration procedures for several commercial chrysanthemum cultivars, and increased the transformation frequency and recovery of transgenic plants using a step-wise kanamycin selection scheme.

Materials and Methods

Plant material Regeneration experiments were conducted using six chrysanthemum

cultivars: Blush, Dark Bronze Charm, Golden Polaris, Iridon, Polaris, and Tara. Our first objective was to identify the best cultivar and explant for transformation. The explants tested were leaf (7 ram2), petiole (3-4 mm), and stem segments (4-5 ram).

Plant culture media and growth conditions Stock mother plants were established using shoot-tips (1.0-2.0 cm in

length) from greenhouse plants; explants were cultured on the basal medium (BM) developed by E. Firoozabady at DNA Plant Technology (A. Mayers, California Plant Company, personal communication). BM contained the Murashige and Skoog (1962) salts, myo-inositol (100 mg/L), thiamine (0.4 mg/L), and sucrose (30 g/L). The pH of the medium was adjusted to 5.8 using 0.1 M KOH, before autoclaving for

Correspondence to: L.M. Yepes

15 min at 121 ~ and 1.1 kg cm -2. Shoot-tips were subcultured periodically from the established shoots, by transferring new tips to test tubes (25 x 150 mm) containing 10 ml of BM. One shoot tip was placed per tube up right in the medium. Cultures were incubated at 25 ~ under white fluorescent light (16-h photoperiod, 75 I.tEm-2s 1 light intensity).

Regeneration studies In preliminary experiments, the effect of auxin type [indolacetic

acid (IAA) versus naphthalenacetic acid (NAA)] and cytokinin concentration [benzyladenine (BA) at 1.0 or 3.0 mg/L] on regeneration was investigated. The hormone regimes tested were based on the regeneration studies of Ledger et al. (1991) [BA 3.0 mg/L and IAA 0.2 mg/L], and Kaul et al. (1990) [BA 1.0 mg/L and NAA 1.0 mg/L]. Since we did not observed any significant increase in regeneration by increasing the amount of BA from 1.0 to 3.0 mg/L, we chose BA 1.0 mg/L for our standard regeneration protocol. Similar regeneration frequencies were obtained for NAA 0.1 mg/L and IAA 0.2 mg/L (data not shown), so we chose to use NAA for our regeneration studies since it is not autoclave labile. From these studies, our standard regeneration medium (RM) consisted of: BM plus BA 1.0 mg/L and NAA 0.I mg/L. Elongation and rooting of regenerated shoots were done on BM (liquid or solid) with no hormones. Use of growth regulators, both BA and NAA, was minimized throughout our regeneration protocol to try to maintain elonal fidelity.

Kanamycin resistance Kanamycin (Kan) a widely used marker for plant transformation

can be phytotoxic and inhibit regeneration of transformed tissue (Yepes and Aldwinckle 1994). Experiments were conducted to determine the effect of Kan on chrysanthemum morphogenesis. Kan was filter-sterilized (Millipore | 0.22 I.tm pore size) and added to the medium after autoclaving. The effect of increasing Kan doses was evaluated on chrysanthemum using untransformed stem and leaf explants cultured on RM containing Kan at 0, 10, 15, 20, 25, 50, 75, and 100 mg/L (10 explants/treatment). All experiments were repeated at least twice.

TSWV (N) gene construct The N gene of the TSWV-BL strain (Wang and Gonsalves, 1990)

was engineered in our laboratory, and includes the open reading frame (771 nucleotides) plus part of the 3' end untranslated region (90 nucleotides) (Pang et al. 1992). The binary plant transformation vector pBIN 19 (Bevan, 1984) used in our study contained within the T-DNA borders the NPT II gene that confers resistance to Kan, and the TSWV-BL (N) gene under the control of the 35S Cauliflower Mosaic Virus (CaMV) promoter, the Cucumber Mosaic Virus-strain white leaf (CMV-WL)-untranslated leader (L) sequence, and the nopaline synthase (nos) terminator (Fig. 1). The (N) gene expression cassette was cloned into the Hind l[I cloning site of pBIN19.

m - -

X

: : ::T --RB---i~176 I NPT ,, lots I--]35 SlLIrswv-(.) I oos~---LB--

Fig. 1 T-DNA region of pBE'(19 containing the TSWV-BL (N) and NPT ]l genes. Restriction enzyme sites used in the Southern are shown.

Biolist& transformation Using a hel ium-driven biolistic apparatus, we attempted

transformation of four chrysanthemum cultivars with the plasmid pBIN19/TSWV-BL (N). Procedures for mieroprojectile coating and loading were as described (Sanford et al. 1993). Leaf and stem explants were bombarded with M10 Sylvania tungsten particles (average diameter=- 1 Ix_M) coated with the plasmid DNA.

For bombardment, new actively growing leaves were collected from in vitro shoot-tip cultures and placed (abaxial side up) on the central area (5 cm 2) of a 10 cm t)etri plate (6 to 8 leaves/plate) containing RM. Stem segments were collected from in vitro shoot-tip cultures and oriented basal side-up to cover the central area of the plate (20-25 segments/plate). For each explant type, the following treatments were included in each bombardment experiment: 1. Regeneration control (particles coated with buffer)- 1 plate 2. Transformation control (pBIN19)- 1 plate 3. TSWV (N) gene (pBIN19/TSWV-BL -N)- 10 plates.

Parameters used for bombardment were: a) whole leaves were bombarded once, at a helium pressure of 1,000 psi, a 1 cm flight

695

distance, and at 20-40% relative humidity (RH); b) stem segments were bombarded twice at a helium pressure of 1,400 psi, a 1 crn flight distance, and at 20-40% RH. After bombardment, stem explants were immediately transferred to RM without antibiotics (10-13 segments/ plate). Bombarded leaves were first sectioned in 7 mm 2 square segments, and then placed on the medium with the adaxial surface down (5 segments/plate).

Selection strategy Ten days after bombardment, selection of transformed tissue was

initiated by transferring the explants to RM plus Kan at 10 mg/L. One month after bombardment, explants were transferred to new medium and Kan was increased to 20 mg/L. Control explants bombarded with particles only or with particles coated with pBIN19 were treated in the same way. Some control explants were maintained under no selection to evaluate regeneration ability. Shoots (>1 cm in length) that regenerated on Kan-containing RM were excised and placed in test tubes on liquid EM plus Kan (20 mg/L) for further selection and elongation. After three months, any surviving bombarded explants were transferred individually to baby food jars containing liquid RM plus Kan at 50 or 75 mg/L. Shoots that regenerated at higher Kan selection levels were excised and placed on EM plus Kan. Surviving shoots were rooted by transferring them to individual test tubes containing solid BM with no hormones and no Kan. In the presence of Kan, rooting was possible only with the addition of NAA at 0.1 mg/L. Once roots developed, plantlets were transferred to .pots in the greenhouse. By five months after bombardment, several han resistant shoots had been recovered, but all bombarded explants had whitened and died during the step-wise selection on RM containing either Kan 20, 50, or 75 mg/L.

DNA isolation, PCR and Southern analysis DNA was isolated from actively growing leaves collected from

plants in the greenhouse. Leaf tissue was ground in liquid N 2 and the DNA was extracted using the procedure of Murray and Thompson (1980). Sodium bisulfite (3.9 g/L) was added to the extraction buffer to inhibit oxidation of phenolics. The presence of the transferred genes in putative transformed chrysanthemums generated in several biolistic experiments was assayed by the polymerase chain reaction (PCR) using oligonucleotides specific for the NPT II and TSWV (N) genes. The upstream and downstream primers of the TSWV (N) gene were respectively:

5' AGCTAACCATGGTTAAGCTCACTAAGGAAAGC 3' 5' AGCATTCCATGGq-TAACACACTAAGCAAGCAC 3'

PCR reactions were run using the Gene Amp| PCR Reagent Kit with Native Taq DNA polymerase following the instructions of the manufacturer (Perkin Elmer). In each reaction tube, the two NPT I1 and the two TSWV (N) primers were added in equimolar amounts for simultaneous template amplification.

Southern blot analysis was done with DNA extracted from some selected transgenic lines to confirm the integration of the TSWV (N) gene into the plant genome. Total DNA was digested with Xba I or Pst I, separated by electrophoresis on a 1% agarose gel, and blotted onto a Gene Screen Plus T M membrane (Du Pont de Nemours). Blots were hybridized using the 32p-labeled (Feinberg and Vogelstein 1984) TSWV (N) gene (0.9 kb) probe obtained by double digestion of the pBIN19/TSWV-BL (N) plasmid with Barn HI andXba I.

R e s u l t s

1. Chrysanthemum regeneration Regeneration experiments using different cultivars and

explants from leaves, petioles, or stems were conducted. Adventitious shoots developed from stem and leaf explants directly without callus formation (direct organogenesis) within two to three weeks. The highest regeneration efficiency was obtained using stem segments (Fig. 2A). No regeneration was observed from petioles. Stem segments cut from the proximity to the apical meristem were more competent for regeneration, as well as newly formed leaves. Regeneration ability decreased on stem segments cut from the base as the stem matured and from matured leaves. Regeneration ability was also cultivar dependent (Table 1). The highest regeneration frequencies with stem explants were obtained for Blush, Dark Bronze Charm, Iridon, and Tara, while Polaris and Golden Polaris were less responsive to regeneration.

696

Fig. 2 A) Stem regeneration; B) Kanamycin selection at 50 mg/L on stem regenerants after bombardment; C) Elongation on kanamycin 75 mg/L of a putative transgenic shoot; D) greenhouse establishment of transgenic chrysanthemums.

Tablel . Regeneration ability of six chrysanthemum culfivars

Cultivar explant %regeneration I #shoots/explant 2

Blush stem 83,0 +__ 8.2 >10.0 + 2.2 leaf 65,0 + 6.1 4.0 +_. 1,2

Dark Bronze stem 90.0 +__ 10 >10.0 + 2,0 leaf 60,0 +_ 5.3 3.0 + 1.0

Golden Polaris stem 70.2 + 3.0 3.0 + 0,5 leaf 62.2 + 2.2 2,0 + 1.3

Iridon stem 81.0 + %8 9.0 + 2,0 leaf 65.0 + 5.6 5.0 + 2.5

Polaris stem 78.6 + 3.0 6.0 + 1.8 leaf 60.0 +_ 6,2 3,0 + 0.5

Tara stem 93.0 + 6.2 6,0 +_ 1.8 leaf 68.0 + 4.2 2,0 + 1.0

1 Regeneration percent values represent the mean + SE. Ten explants were used per treatment and each experiment was repeated three times.

2 Average number of shoots/explant +_. SE,

Leaf explants were extremely sensitive to light intensity and temperature. Light intensity above 50 BEm-2s "1 and temperatures above 25 ~ induced release of phenolics and anthocyanins around the cut edges of leaf explants. Leaf regeneration increased by 30% when light intensities were maintained below 40 ~tEm-2s -1, and temperature was regulated to maximum 24 + 1 ~

2. Kanamycin sensitivity Differences in Kan sensitivity among cultivars were

detected over time. Control untransformed stem and leaf explants of Tara and Blush were very sensitive to Kan since no regeneration was observed even at Kan 10 mg/L. For Dark Bronze Charm and Iridon, increasing Kan concentrations reduced the regeneration frequency for untransformed stem explants from 80-90% at Kan (0) to 30-40% at Kan (10 mg/L), and to 15-20% at Kan (20-25 mg/L), Regeneration of untransformed Dark Bronze Charm and Iridon stem explants was completely inhibited at Kan 50 mg/L. Control unlransformed leaf segments of all cultivars tested regenerated only without Kan.

3. Biolistic transformation Iridon, Tara, Blush and Dark Bronze Charm gave the

best regeneration frequencies and were used to evaluate the effectiveness of biolistic-mediated transformation. Following bombardment, differences in Kan sensitivity among cultivars affected the selection level at which transgenic plants could be recovered. Tara and Blush were very sensitive to Kan, and regeneration was observed only upto Kan 20 mg/L (Table 2). Seventy to 90% of the shoots recovered at Kan 10 mg/L died when the level of selection was raised to Kan 20 mg/L. Increasing the Kan level to 20 mg~ reduced dramatically the number of escape shoots recovered. After two months of selection on Kan 20 mg/L, all bombarded Tara and Blush explants had bleached and died.

Table 2. Step-wise kanamycin selection following biolistic t r ans fo rmat ion of four c h r y s a n t h e m u m cutt ivars with the pB]:N19/TSWV (N) plasmid

Cultivar Kan level # explan~ %survival #Shoots Explant (rag/L) recovered a

IRIDON Leaf seg. 0 80 100.0

t0 80 100.0 77 20 44 55.0 44 50 3 3.0 4 75 1 1.2

Stem seg, 0 t45 100.0 10 145 100.0 315 20 92 63.4 200 50 21 14.5 96 75 9 6,2 25

I DARK BRONZE CHARM Stem seg. 0 40 90.0

10 40 90.0 103 20 31 70.0 63 50 7 15.9 25 75 3 6.8 11

BLUSH Stem ~g . 0 163 100.0

10 163 100.0 540 20 16 9.8 40

T A R A Stem seg. 0 120 100.0

10 120 I00.0 343 20 34 28.3 t00

a Regenerated shoots obtained from step-wise selection were excised from the explants, and selected individually on elongation medium plus Kan, Kan-resistant explants and shoots were transferred to new medium with higher Kan level at one month intervals. After 5 months, all bombarded explants had selected out,

697

In contrast, most bombarded explants of Iridon and Dark Bronze Charm remained green at Kan 20 mg/L, and had to be selected at higher Kan levels (50 and 75 mg/L) (Fig. 2B, Table 2).

Regenerants recovered at Kan 50 and 75 mg/L were excised and placed on EM with Kan (Fig. 2C). After five months of step-wise selection, all bombarded explants of Iridon and Dark Bronze Charm had bleached and died at either Karl 50 or 75 mg/L. Regenerated shoots recovered from explants of Iridon and Dark Bronze Charm on Kan 50 and 75 mg/L were elongated in liquid EM with Kan. All Kan-resistant shoots were rooted and plants were transferred to the greenhouse (Fig. 2D).

Regenerated shoots from different biolistic experiments of the cultivars Dark Bronze Charm, Iridon, Tara, and Blush were established in the greenhouse. Plants derived from stem and leaf explants had similar morphological characteristics to control plants (leaf size and shape, growth habitat, and flower color). Putative transgenic plants were screened by PCR to determine whether they had been transformed with the TSWV-BL (N) gene, the NPT II gene, or both.

4. Characterization of putative transgenic lines by PCR and Southern analysis

PCR products corresponding to the NPT II and TSWV- BL (N) genes were amplified from several putative transgenic lines established in the greenhouse (Fig. 3).

However, some lines contained only one gene, either the NPT II gene (15%) or the TSWV (N) gene (14%).

Table 3. PCR analysis of putative transgenic chrysanthemum plants in the greenhouse obtained after biolistic transformation

Selection # Plants PCR Genes detected tested +

NPT II & NPT II TSWV (IV) TSWV (N)

IRIDON Kan 75 17 16 9 2 5 Kan 50 12 8 5 3

DARK BRONZE CHARM Kan 75 7 7 7

BLUSH Kan 20 21 21 11 5 5

TARA Kan 20 25 21 20 1

Totals 82 73 52 11 10

Southern blot analysis of selected transgenic lines that were positive for PCR was conducted to corroborate the integration of the TSWV (N) gene (Fig. 4). A 0.9 kb Xba I plant DNA fragment hybridized to the 32p labeled TSWV (N) gene probe, indicating the presence of the TSWV (N) gene in the transgenic plants. Using Pst I, a single cutter of pBIN19/TSWV (N) (11.6 kb), different restriction patterns were obtained for the different transgenic clones depending on the insertion site of the transgene.

Fig. 3 PCR analysis of putative transgenic chrysanthemum lines. Ethidium bromide stained agarose gel (upper panel) with PCR amplified products for the NPT U and TSWV (N) genes. The first lane corresponds to the Molecular size marker 0X174 DNA/Hae 1TI (~), the next lane labelled (P) to the plasmid pBIN 19/TSWV (N) used for bombardment, followed by a non transformed chrysanthemum (lane 1), and several tmnsgenie lines of the cultivars: Iridon (lanes 2-3), Blush (lanes 4-5) and Tara (lane 6). The identity of the PCR amplified TSWV l~roducts was confirmed by Southern hybridization with the specific "32p labeled TSWV (N) gene probe (lower panel).

A summary of the PCR screening is given in Table 3. Most of the plants tested were PCR positive (89%), indicating that the Kan selection procedure effectively reduced the number of escapes to 11%. Out of the 73 PCR positive lines recovered, the majority (71%) contained both the NPT II and the TSWV (N) genes.

Fig. 4 Southern blot analysis on transgenic chrysanthemum using the 32p labeled TSWV-(N) gene as a probe. (A). Lanes t-3 were loaded with 90, 30 and 15 pg of the plasmid pBIN 19/TSWV-(N) digested with Xba I; lane 4 was loaded with DNA of a non transformed chrysanthemum; and lanes 5-9 contain DNA (7 ~tg) of transgenic lines of the cultivars: Iridon (lanes 4-5), Blush (lanes 6-7) and Tara (lane 8) digested with Xba L (B). Pst I linearized plasmid (P) produced a I 1.6 Kb fragment, absent in untransformed chrysanthemum (1). Two transgenic Iridon lines (labeled 2 and 3) have unique Psi I digestion patterns representing different insertion events. Pst I digested plant DNA was loaded at 10, 15, 25 btg/welL

698

Discussion We successfully transformed four chrysanthemum

cultivars with the TSWV (N) gene using a biolistic procedure, and optimized regeneration using stem explants. This is the first report of biolistic transformation of chrysanthemum. Although regeneration was significantly lower for leaf compared to stem segments in all six cultivars studied, we recovered some transformants from leaf explants, which demonslxates the high efficiency of our transformation and selection schemes.

High auxin/cytokinin ratios [2,4-D 1.0 mg/L or IAA 2.0 mg/L plus BA 0.1-0.2 mg/L] have been used to induced embryogenesis in chrysanthemum (May and Trigiano 1991, Pavingerova et al. 1994, Urban et al. 1994), while high cytokinin/auxin ratios [BA 3.0 mg/L plus NAA 0.1 mg/L] have been used to induce organogenesis (Ledger et al. 1991). Since somaclonal variation is undesirable in gene transfer studies, we used direct organogenesis minimizing the amount of growth regulators (BA 1.0 mg/L plus NAA 0.1 mg/L) to help maintain clonal fidelity. Our regeneration percent values for stem segments compare well with those reported b y Kaul et al. (1990), even though we tested different cultivars and used lower amounts of NAA (0.1 vs. 1.0 mg/L). However, our leaf regeneration results were more uniform among cultivars (60-70% vs. 0-90% for Kaul et al. 1990), but were not as high as those obtained for stem segments (70-100%). Higher regeneration percent and number of shoots per explant were consistently recovered for the cultivars Blush, Dark Bronze Charm, Iridon, and Tara compared to Polaris and Golden Polaris. Genotype dependency of regeneration methods using somatic tissues has been well documented for chrysanthemum (de Jong et al. 1993, Fukai et al. 1987, Kaul et al. 1990, Lu et al. 1990, Roest and Bokelmann 1975). Nevertheless, the regeneration method reported here has several advantages: 1) Direct organogenesis at high frequencies using low hormone regimes and vegetative explants from tissue culture; 2) This protocol is suitable for a range of commercially important chrysanthemum cultivars and can be adapted to both biolistic and Agrobacterium-mediated transformation studies [unpublished]; 3) It appears to produce true-to-type plants. However, this later aspect needs to be confirmed further in nursery studies.

Our study confirms previous reports indicating that chrysanthemum is very sensitive to Kan (Lowe et al. 1993, Renou et al. 1993), and demonstrates that sensitivity to Kan is cultivar dependent. We found two genotypes (Tara and Blush) that were extremely sensitive to Kan for which stringent selection occurred at Kan 20 mg/L, whereas two others (Iridon and Dark Bronze Charm) were less sensitive to Kan and could be selected at significantly higher levels (Kan 50 and 75 mg/L). Interestingly, our study is the second one to report recovery of transgenic chrysanthemum plants by selecting at levels higher than Kan 50 mg/L [Courtney-Gutterson et al. 1994, selected transgenic plants of the cv. Moneymaker at 100 mg/L Kan]. We delayed selection for 10 days and gradually increased the Kan concentration from 10 to 75 mg/L to obtain a larger number of transgenic plants, while minimizing the number of

escapes. Interestingly, Renou et al. (1993) also obtained a significantly higher number of transformants (34% vs. 8.6-14%) after delayed selection, but using different hygromycin selection protocols their frequency of escapes was very high: 66 to 93%.

So far, we have not detected expression of the TSWV (N) gene by ELISA on the transgenic chrysanthemum tested. Interestingly, Urban et al. (1994) were not able either to detect the TSWV (N) protein production in transformed Iridon plants either by Western blot or ELISA. Although we are testing other transgenic lines and other gene constructs, lack of expression may be due to the methylation of the transferred cassette after incorporation into the plant genome.

Evaluation of the resistance of the transgenic chrysanthemum population containing the TSWV-BL (N) gene against Tospoviruses is on going.

Acknowledgments We thank Dave Hummer for technical assistance, Sheri Ecker-Day

for maintaining the plants in the greenhouse, and Dr. Marc Fuchs for critical review of this manuscript. This research was supported by the California Plant Company, and by the Comell Biotechnology Program, which is sponsored by the New York State Science and Technology Foundation, a consortium of industry and the National Science Foundation. The chrysanthemum cultivars used in this study were sent to us by the California Plant Company.

References Bevan M (1984) NAR 12:8711-8721 Courtney-Gutterson N, Napoli C, Lemieux C, Morgan A, Firoozabady

E and Robinson KEP (1994) Bio/Technology 12:268-271 de Jong J, Rademaker W, van Wordragen MF (1993) Plant Cell Tissue

and Organ Culture 32:263-270 Feinberg AP, Vogelstein B (1984) Anal Biochem 137:266-267 Fukai S, O~ M (1986) Bull Osaka Agr Res Cent 23:25-31 Fukai S, Chen Z, O~ M (1987) Bull Osaka Agr Res Cent 24:55-58 Kaul V, Miller RM, Hutchinson JF, Richards D (1990) Plant Cell

Tissue and Organ Culture 21:21-30 Ledger SF, Deroles SC, Given NK (1991) Plant Cell Rep 10:195-199 Lowe J1M, Davey MR, Power JB, Blundy KS (1993) Plant Cell Tissue

and Organ Culture 33:171-180 Lu CY, Nugent G, Wardley T (1990) Plant Cell Rep 8:733-736 May RA, Trigiano RN (1991) J Am Soc Hort Sci 116:366-371 Murashige T, Skoog F (1962) Physiol Plant 15:473-497 Murray MG, Thompson WF (1980) Nucleic Acid Res 8:4321-4325 Pang SZ, Nagpala P, Wang M, Slightom JL, Gonsalves D (1992)

Phytopathol 82:1223-1229 Pavingerova D, Dostal J, Biskova R, Benetka V (1994) Plant Sci 97: 95-

101 Renou JP, Brochard P, Jalouzot R (1993) Plant Sci 89:185-197 Roest S, Bokelmann GS (1975) Scientia Hort 3:317-330 Sanford JC, Smith FD, Rusell JA (1993) In: Wu R (ed) Methods in

Enzymology, vol 217. Academic Press, Orlando, pp 483-509 Urban LA, Sherman JM, Moyer JW, Daub ME (1994) Plant Sci 98:

69 -79 van Wordragen MG, de Jong J, Huitema HBM, Dons HJM (1991)

Plant Cell Rep 9:505-508 Wang M, Gonsalves D (1990) Plant Dis 74: 154-158. Yepes LM, Aldwinckle HS (1994) Plant Cell Tissue and Organ Culture

37: 257-269.