Plant Cell Reports

Plant Cell Reports (1996) 15:672-676

9 Springer-Verlag1996

Further evidence of a cybridization requirement for plant regeneration from citrus leaf protoplasts following somatic fusion * J.W. Grosser 1, E G. Gmitter, Jr. x, N. Tusa 2, G. Reforgiato Recupero a, and P. Cucinotta 3 1 University of Florida, IFAS, Citrus Research and Education Center, 700 Experiment Station Road, Lake Alfred, FL 33850, USA 2 Centro Miglioramento Genetico Degli Agrumi, Palermo, Italy 3 Istituto Sperimentale per L'Agrumicoltura, Acireale, Italy Received 2 August 1995/Revised version received 20 November 1995 - Communicated by G. C. Phillips

Summary. Somatic hybridization experiments in Citrus that involve the fusion of protoplasts of one parent isolated from either nucellus-derived embryogenic callus or suspension cultures with leaf-derived protoplasts of a second parent, often result in the regeneration of diploid plants that phenotypically resemble the leaf parent. In this study, plants of this type regenerated following somatic fusions of the following three parental combinations were analyzed to determine their genetic origin (nuclear and organelle): (embryogenic parent listed first, leaf parent second) (1) calamondin (C. microcarpa Bunge) + 'Keen' sour orange (C. aurantium L.), (2) Cleopatra mandarin (C. reticulata Blanco) + sour orange, and (3) 'Valencia' sweet orange (C. sinensis (L.) Osbeck) + 'Femminello' lemon (C. limon (L.) Burro. f.). Isozyme analyses ofPGI, PGM, GOT, and IDH zymograms of putative cybrid plants, along with RFLP analyses using a nuclear genome-specific probe showed that these plants contained the nucleus of the leaf parent. RFLP analyses using mtDNA-specific probes showed that these plants contained the mitochondrial genome of the embryogenic callus donor, thereby confirming cybridization. RFLP analyses using cpDNAspecific probes revealed that the cybrid plants contained the chloroplast genome of either one or the other parent. These results support previous reports indicating that acquisition of the mitochondria of embryogenic protoplasts by leaf protoplasts is a prerequisite for recovering plants with the leaf parent phenotype via somatic embryogenesis following somatic fusion. Abbreviations:

cp-chloroplast; GOT-glutamateoxaloacetatc transaminase; IDH-isocitrate dehydrogenase; mt-mitochondria; PEG-polyethylene glycol; PGIphosphoglucose isomerase; PGM-phosphoglucomutase; RFLP-restriction fragment length polymorphism Key words: chloroplast, embryogenesis, tissue culture

mitochondria,

somatic

Introduction Somatic hybridization has become an integral part of research programs in citrus genetics and cultivar development. More than 100 citrus somatic hybrids have been reported to date (for a partial review, see Grosser and Gmitter, 1990). Somatic hybridization of citrus has been achieved most frequently by fusing protoplasts from nucellar-derived embryogenic callus or suspension cultures (embryogenic parent) with non-regenerable mesophyll protoplasts (leaf parent), in the presence of polyethylene glycol (PEG). However, in several somatic hybridization experiments, diploid plants with morphology and zymotype of the presumably non-embryogenic leaf parent were recovered (Ohgawara et al. 1989; 1991; Tusa et al. 1990; Grosser and Gmitter 1992; Louzada et al. 1992; Saito et al. 1993). Diploid plants of this nature have been recovered in our program from at least 12 different leaf parents (J. W. Grosset, unpublished data). Tusa et al. (1990) speculated that these plants may have regenerated directly from unfused mesophyll protoplasts and suggested a role for co-culturing with an embryogenic callus line. An alternative hypothesis, originally put forth by Dr. Aliza Vardi of the Volcani Center, Israel (personal communication, 1990), was that such diploid plants regenerated from leaf-derived protoplasts could be cybrids. This latter hypothesis was based in part on the production of five Citrus cybrids by a donor-recipient method (Vardi et al. 1987; 1989). The method of Vardi et al. was based on selection for comptimentation of nuclear and cytoplasmic functions, and recovered cybrid plants contained the nuclear genome of the recipient and the cytoplasmic genome of the donor, mtDNA recombination was also reported in these cybrids. More specifically, the mitochondria have been implicated in somatic embryogenesis in vitro from Citrus. Kobayashi et al. (1991), and Ohgawara et al. (1994) analyzed the mitochondrial genomes of two different

Correspondence to: J. W. Grosser * Florida Agricultural Experiment Station Journal Series No. R-04631.

673 somatic hybrids and found that each possessed the mitochondria o f the embryogenic parent, but not the mesophyll parent. Saito et al. (1993) reported the recovery o f diploid leaf parent-like plants following electro fusion of Citrus sudachi Hort. ex. Shirai embryogenic protoplasts with mesophyll protoplasts of lemon (C. limon [L.] Burro. f.) and lime (C. aurantifolia Swing.). RFLP analysis o f mtDNA revealed that these diploid plants were cybrids because their mitochondria were identical to the C. sudachi embryogenic callus parent. Similarly, Yamamoto and Kobayashi (1995) reported the recovery o f a diploid leaf-parent like plant following electrofusion o f Citrus unshiu Marc cv. 'Juman' with mesophyll protoplasts o f Citrus sinensis (L.) Osbeck cv. 'F.N. Washington' navel orange, and again mtDNA analysis confirmed that this plant was a cybrid that contained m t D N A identical to C. unshiu. Thus, it appears that mitochondria o f the embryogenic callus lines play a critical role in recovery o f Citrus somatic hybrids and cybrids. By contrast, Kobayashi et al. (1991) and Ohgawara et al. (1994) reported that chloroplasts in somatic hybrid plants obtained from the fusion o f navel orange (C sinensis L. Osb.) with 'Murcott' tangor (putative hybrid o f C. sinensis and C. reticulata Blanco) and 'Trovita' sweet orange (C. sinensis L. Osb.) with trifoliate orange (Poncirus trifoliata (L.) Raf., randomly segregated from either one or the other parent in both cases, and no plants possessed chloroplasts from both. Vardi et al. (1989) analyzed chloroplast D N A in embryos and regenerated plants following cybridization o f Microcitrus (a related genus) with rough lemon (C. jambhiri Lush) or sour orange (C. aurantium L.). Although recovered embryos contained chloroplasts of either or both parents, regenerated cybrid plants o f both combinations had chloroplasts o f either one parent or the other, but not both. Herein, we report on investigations to characterize the nuclear and cytoplasmic genomes o f putative cybrid plants regenerated following fusion o f embryogenic with mesophyll protoplasts o f three parental combinations (embryogenic parent listed first): Cleopatra mandarin (C. reticulata) + Sicilian sour orange (C. aurantium); calamondin (C. microcarpa Bunge) + Keen sour orange (C. aurantium); and 'Valencia' sweet orange (C. sinensis) + 'Femminello' lemon (C. limon). Putative cybrid plants exhibiting the phenotype o f the leaf parent were examined by isozyme and RFLP analyses to identify the nuclear origin, and by using nat and cp specific probes to determine the genetic origin o f the organelles. Materials and Methods

Plant material The putative cybrid plants used for this study were produced by polyethylene glycol (PEG)-induced somatic fusion of nucellus-derived embryogenic protoplasts with nucellar seedling leaf-

derived protoplasts, as described by Grosser and Gmitter (1990). The putative 'Femminello' lemon cybrid was reported by Tusa et al. (1990); the others were recovered by Tusa (unpublished results). The putative cybrid plants were selected on the basis of their morphological similarities to the leaf donor parents.

Isozyme analyses. Isozyme banding patterns were analyzed using crude leaf extracts sampled as in Torres et al. (1978). PGI (phosphoglucose isomerase) and PGM (phosphoglucomutase) were separated using horizontal starch gel electrophoresis on 10% starch gels and the pH 7.5 histidine-citrate buffer of Cardy et al. (1981). For GOT (glutamateoxaloacetate transaminase), the gel buffer was tris-citrate pH 8.2 (0.03 M) and the electrode chamber buffer was pH 8.7 (0.37 M) sodium borate (Torres et al. 1978). For IDH (isocitrate dehydrogenase), the gel buffer was tris-citrate pH 7.0 (0.03 M) and the electrode chamber buffer was tris-citrate pH 7.0 (0.3 M). Electrophoresiswas carried out at 4~ for 3 h and constant 45 mA current for PGI and PGM, for 4 h and constant 35 mA current for GOT, and for 7 h and constant 70 rnA current for IDH (Torres et al. 1982); staining recipes were from Vallejos (1983). DNA extraction. Total genomic DNA was isolated from Citrus leaves of regenerated plants using the protocol of Doyle and Doyle (1987), modified by the addition of 1% polyvinylpyrrolidone and 1% 2mercaptoethanol to the isolation buffer. Alternatively, total DNA was extracted using a 'Genomics Kit' (Talent; Trieste, Italy) based on initial lysis of the sample via a cationic detergent, followed by a single extraction step with chloroform. The phase separation was simplified greatly by the presence of a physical barrier between phases. RFLP/Southern Blot Hybridization. Two/~g of total genomic DNA were digested with 10 U ofEco RI and HindIII. The digested products were electrophoresed on 1% agarose gels with TAE buffer and detected by ethidium bromide staining. Following electrophoresis, gels were treated with alkali and blotted onto a nylon filter (Hybond N+, Amersham International plc; Amersham, England) by the Southern method. Cloned inserts (probes) were isolated using mini-preps methods (modificationof the method of Birnbhoimand Doly 1979;Ish-Horowiczand Burke 1981), and purified by Qiaex gel extraction kit (Genenco; Hilden, Germany). Probes were labelled using a 'Prime-a-gene' labelling system kit (Promega; Madison, WI, USA). Typical reactions included 25 /zg of isolated insert and 50 /xCi a[32P]dCTP according to Feinberg and Vogelstein (1984). Probes. Fragment11-73(1.6 Kb), selected from a genomic library made from a zygotic hybrid of C. latipes (Swing.) Tan. x C. aurantium, was used to determine the origins of nuclear genomes. For mtDNA analysis, the followingprobes were kindly provided by Gallerani (Bari University, Italy) from the COX sunflower mtDNA library: MI, a 450 bp fragment cloned in pUC8; CI P(3)l, a 1.2 Kb fragment cloned in pUC19; CI21 Cox III, a 900 bp fragment cloned in pUC8; and M, a 1.1 Kb fragment cloned in pBlueScript. The following probes were kindly provided by Sugiura (Nagoya University, Japan) from a rice cpDNA library, to analyze putative cybrid chloroplasts: PRB 3-1 (3.4 Kb fragment), PRB 3-2, (2.5 Kb), and PRB 3-3 (1.5 Kb), all cloned in pBlueScript KS(+). Results and Discussion

Results of isozyme analyses conducted on diploid regenerated plants resembling the seedling leaf parent following somatic fusion are summarized in Table 1. These results clearly indicate that the origin of nuclear DNA in all tested putative cybrid plants was the leaf parent. Further evidence that the nuclei in these plants came from the leaf parent was obtained from RFLP analysis using the citrus genomic D N A fragment II-73, as a probe o f Southern blots; cybrids from all three parental

674 combinations produced banding patterns identical to the corresponding leaf parents (data not shown).

cultures may be involved. warranted.

RFLP analyses, shown by Southern blot hybridization of total DNA from cybrids and parental types digested with either Eco RI or Hind III and probed with the mtDNAspecific probe CI 21 Cox III or CI P(3)1 of sunflower, are provided in Figs. 1-3. All cybrids tested with these probes (two from calamondin + Keen sour orange, five from Cleopatra + sour orange, and 17 [15 diploid and two tetraploid] from 'Valencia' sweet orange + 'Femminello' lemon), show clearly mtDNA patterns identical to that of their corresponding embryogenic parent. Southern blot hybridization of total DNA of five cybrids from Cleopatra + sour orange parentage, digested with either Eco Pd or

Some experiments to determine the chloroplast origin in confirmed cybrids were also conducted, and the results of RFLP analyses are provided in Fig. 5. Four cybrids of calamondin + Keen sour orange produced cpDNA restriction patterns identical to the embryogenic parent calarnondin. However, two of five Cleopatra + sour orange cybrids had cpDNA patterns of sour orange, the leaf parent. This latter result suggests that acquisition of the chloroplast genome from the embryogenic parent is not required for subsequent plant recovery via somatic embryogenesis. These results are in agreement with the chloroplast inheritance in somatic citrus hybrids reported by Kobayashi et al. (1991) and Ohgawara et al. (1994), who showed that individual somatic hybrid plants contained the chloroplast genome of either one or the other parent.

Table 1. Zymotypes of parents and putative cybrids for different isozyme staining systems. Allelic designations (F, S,

W, I, M) arc according to Soost and Tortes (1981) for PGI, PGM, GOT-I and IDH.

Isozyme Plant I.D.

PGI

PGM

GOT-1

IDH

Calamondin(E)

FS

FF

Keen sour orange (L)

WS

FS

Cybrids (3 tested)

WS

FS

Cleopatra mandarin(E)

FF

FF

Sour orange (L)

WS

FS

Cybrids (5 tested)

WS

FS

'Valencia' sweet orange (E)

SS

1M

'FemmineUo' lemon(L)

FS

IS

Cybrids (13 tested)

FS

IS

L = leaf parent, E = embryogenicparent Hind III and probed with CI-P(3)I mtDNA probe, also

produced the same banding pattern as the Cleopatra embryogenic parent (Fig. 4). The results of these experiments prove that the diploid regenerants with leaf parent morphology are cybrids, and they provide additional evidence that the acquisition of mitochondria from embryogenic cells by leaf-derived protoplasts is a prerequisite for subsequent plant regeneration via somatic embryogenesis in citrus. It is currently not known how mitochondria enable cybrid cells to undergo somatic embryogenesis. The phenomenon does not appear to be genotype specific. It is possible that the physiological status or elevated numbers of mitochondria provided to cybrids by protoplasts isolated from embryogenic cell

Further investigation is

Cybrid cells may have originated from successful protoplast fusion accompanied by failed nuclear fusion and the subsequent loss of the nucleus of the embryogenic parent. This mechanism seems to be the most likely for the Cleopatra + sour orange cybrid, because some cybrids were found to have mtDNA of the embryogenic line but cpDNA from the leaf parent; obviously, the distinctive cytoplasms of the parents were combined at some point. The other cybrids (with mtDNA and cpDNA from the embryogenic parent only) may have originated in like fashion, followed by subsequent sorting out and selection in favor of the cytoplasm of the embryogenic parent. The incorporation into mesophyll protoplasts of mitochondria released from ruptured embryogenic cells may be another mechanism for cybridization. Additional work will be required to determine the actual mechanism(s) underlying such cybridization events. The practical value of citrus cybrids is currently unknown, because no horticulturally important traits have been shown to be encoded by organelle DNA. Proven and putative cybrids with leaf parent morphology (including cybrids resembling 'Femminello' lemon, 'Valencia' sweet orange, 'Murcott' tangor, 'Dancy' and 'Ponkan' mandarins, and 'Duncan' grapefruit) have been propagated for field testing to determine if they can provide any useful variation. Fruit quality, maturity date, and disease resistance of these cybrids will be subsequently compared with the original cultivars. Cybrids exhibiting the sour orange phenotype are being assayed for resistance to quick-decline, a disease caused by citrus tristeza virus that has limited the use of this otherwise desirable citrus rootstock. It remains to be proven what value may be derived from direct utilization of citrus cybrids in commercial citriculture. However, the development of citrus cybrid

675 plants and, more importantly, ofembryogenic citrus cybrid callus provides some new potential applications of somatic fusion for citrus cultivar development. For example, somatic hybrids of lime with lemon were produced by Saito et al. (1994) by fusion of protoplasts from the limesudachi cybrid callus (Saito et al. 1993) with lemon mesophyll protoplasts; the resulting somatic hybrid possessed mtDNA from the original sudachi embryogenic line. Previously, Citrus embryogenic cultures were initiated only from nucellar tissues or from embryos cultured in vitro. This tissue source limitation effectively excluded valuable monoembryonic Citrus selections as partners in fusion experiments (except as leaf parent donors), because genetically true-to-type embryogenic callus cultures could not be obtained. It should now be possible to induce embryogenesis from mesophyll protoplasts ofmonoembryonic selections by cybridization, thus expanding the range of potential fusion partners for somatic hybridization. Further, diploid cybrid embryogenic callus of monoembryonic citrus may be utilized in genetic transformation schemes, to improve such cultivars. Continued research emphasis on cybridization in Citrus should provide additional basic information regarding somatic embryogenesis and determine the applied value of cybridization to citrus cultivar improvement.

Fig. 1. Southern blot hybridization of Eco RI (lanes 1-7) and Hind III (lanes 9-16) digests of total DNA of cybrids and parental types (parental combination: Calamondin + Keen sour orange) probed with CI 21 Cox III mtDNA specific probe of sunflower (lanes 1-7) or CI-P(3)I (lanes 916). 1. Calamondin (embryogenic parent); 2. Keen sour orange (leaf parent); 3. cybrid no. 2; 4. cybrid no. 3; 5. Keen sour orange; 6. cybrid no. 2; 7. cybrid no. 3; 8. DNA ladder; 9. calamondin; 10. Keen sour orange; 11. cybrid no. 2; 12. cybrid no. 3; 13. 'Keen' sour orange; 14. cybrid no. 2; 15. cybrid no. 3; 16. calamondin.

Fig. 2. Southern blot hybridization of Eco RI (lanes 1-7) and Hind III (lanes 8-11) digests of total DNA of cybrids and parental types (parental combination: Cleopatra mandarin + sour orange) labelled with CI 21 Cox III mtDNA specific probe of sunflower. 1. Sour orange (leaf parent); 2. Cleopatra (embryogenic parent); 3. cybrid no. 5; 4. Cybrid no. 1; 5. Cybrid no. 4; 6. Cybrid no. 2; 7. cybrid no. 3; 8. sour orange; 9. Cleopatra; 10. cybrid no. 5; 11. cybrid no. 1.

Fig. 3. Southern blot hybridization of Eco RI (Fig. 3a) and Hind III (Fig. 3b) digests of total DNA cybrids and parental types (parental combination: 'Valencia' sweet orange + 'Femminello' lemon) probed with CI 21 Cox III mtDNA specific probe of sunflower. Left to right: 1. DNA ladder; 2. 'Femminello' lemon (leaf parent); 3. 'Valencia' sweet orange; 4. cybrid no. 6; 5. cybrid no. 24; 6. cybrid no. 49; 7. cybrid no. 7; 8. cybrid no. 2; 9. cybrid no. 15; 10. cybrid no. 5; 11. cybrid no. 12; 12. cybrid no. 40; 13. cybrid no. 31; 14. cybrid no. 14; 15. cybrid no. 1; 16. eybrid no. 29; 17. cybrid no. 23; 18. cybrid no. 11; 19. cybrid no. 47; 20. 'Femminello' lemon; 21. 'Valencia' sweet orange.

676

Fig. 4. Southern blot hybridization of total DNA from cybrids and parental types (parental combination: Cleopatra + sour orange) digested with either Eco RI (lanes 2-8) or Hind III (lanes 10-16) and probed with CI-P(3)I mtDNA probe. 1. DNA ladder; 2. and 10. Cleopatra mandarin (embryogenic parent); 3. and II. sour orange (leaf parent); 4. and 12. cybrid no. 1; 5. and 13. cybrid no. 2; 6. and 14. cybrid no. 3; 7. and 15. cybrid no. 4; 8. and 16. cybrid no. 5.

Fig. 5. Southern blot hybridization of total DNA digested with Eco RI and probed with PRB 3-1 cpDNA probe (lanes 2-7) or PRB 3-3 cpDNA probe (lanes 9-15), that indicates the segregation of cpDNA among cybrids. 1. DNA ladder (1 Kb); 2. calamondin (embryogenic paren0; 3. Keen sour orange (leaf parent); 4. calamondin + Keen cybrid no. 1; 5. calamondin + Keen cybrid no. 2; 6. calamondin + Keen cybrid no, 3, 7. calamondin + Keen cybrid no, 4; 8. DNA ladder (1 Kb); 9. Cleopatra mandarin (embryogenic parent); 10. sour orange (leaf parent); 11. Cleopatra + sour orange cybrid no. 1; 12. Cleopatra + sour orange cybrid no. 2; 13. Cleopatra + sour orange cybrid no. 3; 14. Cleopatra + sour orange cybrid no. 4; 15. Cleopatra + sour orange cybrid no. 5.

References Birnbhoim HC, Doly J (1979) Nucleic Acids Res 7:1513 Cardy BJ, Stuber CW, Goodman MM (1981) Inst Statist Mimeo Serv, No. 1317, North Carolina State Univ, Raleigh Doyle JJ, Doyle JL (1987) Phytochem Bull 19:11-15 Feinberg AP, Vogelstein B (1984) Annals Biochem 137:266-267 Grosser JW, Gmitter FG Jr (1990) Plant Breed Rev 8:339-374 Grosser JW, Gmitter FG Jr (1992) J Amer Soc Hort Sci 117:168-173 Ish-Horowicz D, Burke JF (1981) Nucleic Acids Res 9:2989 Kobayashi S, Ohgawara T, Fujiwara K, Oiyama I (1991) Theor Appl Genet 82:6-10 Louzada ES, Grosser JW, Gmitter FG Jr, Deng XX, Tusa N, Nielsen B, Chandler JL (1992) HortScience 27:1033-1036 Ohgawara T, Uchimiya H, Ishii S, Kobayashi S (1994) In: Bajaj YPS (ed) Biotechnology in Agriculture and Forestry Vol 27, Sprinter-Verlag, Berlin, pp 439-454 Ohgawara T, Kobayashi S, Ishii S, Yoshinaga K, Oiyama I (1989) Theor Appl Genct 78:609-612

Ohgawara T, Kobayashi S, Ishii S, Yoshinaga K, Oiyama I (1991) Theor Appl Genet 81:141-143 Saito W, Ohgawara T, Shimizu J, Ishii S, Kobayashi S (1993) Plant Sci 88:I95-201 Salto W, Ohgawara T, Shimizu J, Kobayashi S (1994) Plant Sci 99:8995 Soost RK, Torres AM (1981) Proc Int Soc Citriculture, pp 7-10 Tusa N, Grosser JW, Gmitter FG Jr (1990) J Amer Soe Hort Sci 115:1043-1046 Vallejos CE (1983) In: Tanksley SA, Orton TJ (eds). Isozymes in plant genetics and breeding, part A, Elsevier, Amsterdam, pp 469516 Vardi A, Breiman A, Galun E (1987) Theor Appl Genet 75:51-58 Vardi A, Arzee-Gonen P, Frydman-Shani A, Bleichman S, Galun E (1989) Theor Appl Genet 78:741-747 Yamamoto M, Kobayashi S (1995) Plant Tissue Culture Letters 12:131137