CORE Metadata, citation and similar papers at core.ac.uk

Provided by Horizon / Pleins textes

Reproductive Behavior in - Polyhaploid : Implications for the Transfer of Apomixis Into Maize

O. Leblanc, D. Grimanelli, N. Islam-Faridi, J. Berthaud, and Y. Savidan

Expression of gametophytic apomixis in grasses is restricted to polyploid plants. In our effort to transfer apomixis from diplosporous tetraploid Tripsacum (2n = 4x = 72) into maize (2n = 20) through conventional backcrosses, we produced poly- haploid plants combining one complete set of chromosomes from each genus. These polyhaploid plants were totally male sterile. By contrast, viable seeds were produced apomictically when they were pollinated using maize. Apomictic repro- duction in such polyhaploids, which show a diploid-like chramosomal complement, suggests that diplosporous apomixis and polyploidy are not totally linked and that diploid crops such as maize might reproduce apomictically.

Gametophytic apomixis, an asexual repro- mentioned fertile 28-chromosome plants duction- through seeds (Nogler 1984a), derivating from maize-Tripsacum hybrids, commonly occurs in Angiosperms such as but neither mode of reproduction nor wild grasses and Asteraceae (Asker and chromosomal complement were clearly Jerling 1992). Its introduction into crops documented. would open new areas in breeding The aims of this study were to docu- r and agriculture, especially in Third World ment the chromosome complement and countries (Savidan and Dujardin 1992). In the reproductive behavior in one of these agamic complexes, diploids are typically maize-Tripsacum polyhaploids from a re- sexual, whereas polyploids reproduce search program designed to move gene(s) apomictically (Stebbins 1950). Most host responsible for diplosporous apomixis crop species, however, are diploid. Obsta- into maize. Possibilities of expression of cles to the transfer of apomixis through apomixis in diploid maize plants are also conventional backcrossing from wild-relat- discussed. ed species into crops are numerous (e.g., genomically distant apomictic donor to Materials and Methods the host species, male sterility in wide hy- brids, detection of apomixis in wide hy- Plant Material brids and backcross progeny), but this One 28-chromosome polyhaploid plant, 6- strong relationship between gametophytic 529, was obtained through a complex apomixis and polyploidy represents one of backcrossing scheme from Fl-16, a maize- the most important. Tripsacum F, hybrid. FI-,, was 2n = 46- From diplosporous tetraploid Tripsa- i.e., 36 chromosomes from Tripsacum (Tr) cum, the closest wild apomictic relative of and 10 from maize (M). It was obtained maize (2n = 20) with x = 18 (Harlan et al. using a maize plant (2n = 20) from CIM- 1970; Reeves and Mangelsdorf 1935), we MYT population 34 as the pistillate parent established a collection of several gener- and CIMMYT Tripsacurn dactyloides acces- ations of maize-Tripsacurn hybrids. From sion 65-1234, an apomictic tetraploid (2n ' From the French Scientific Research lnstitute for De- apomictic 2n = 56 hybrids (i.e., 20 maize = 72) from Evergleades (Florida, U.S.A.), i velopment through Cooperation (ORSTOM) (Leblanc, Grimanelli, Berthaud, Savidan) and The lnternational chromosomes + 36 from the wild spe- as the male parent. Because of complete Maize and Wheat Improvement Center (ClMMYq cies), we produced apomictic polyhaploid male sterility in F,-,, and its backcross de- (Leblanc, Grimanelli, Islam-Faridi, Berthaud, Savidan), plants (2n = 28) combining one haploid rivatives, plant 6-529 was derived from Apartado Postal 6-641,06600 México D. F., México. O. Leblanc and J. Berthaud are now at ORSTOM, Labor- set of chromosomes from each maize and FI-,, after six generations of backcross US- atoire de Ressources Génétiques et d'Amélioration des Tripsacurn. Although apomictic polyhap- ing pollen from maize individuals from Plantes Tropicales, BP 5045, 34032 Montpellier Cedex, loid plants have been already reported in CIMMYT population 21 (Table 1A). 6-529 France. N. Islam-Faridi is now at the Department of Soils &Crop Sciences, Texas A & M University, College aposporous agamic complexes, they were was then pollinated with a maize hybrid Station, Texas. generally weak and sterile (Nogler 1984a). (CML62 x CML135, CIMMYT lines), result- Journal of Heredity 1996;87:108-111; 0022-1503/96/$5.00 De Wet et al. (1973), however, already ing in a progeny of 64 plants.

108 . Table 1. Maize-Tripsocum dactyloides 65-1234 hybrid generations from which 2n = 28 polyhaploid ment of a reduced gamete (i.e., plant 6-529 plant, 6-529, originated from 5-8 progeny). Generations of backcrosses

FI GI G2 G3 G4 G5 GG Chromosomal Analyses A Plant 6-529 and 52 of its offspring were es- Plant no. 16 1,001 2,017 3,031 4,031 5-8 6-529 timated to have 2n = 28 using flow cyto- No. of chromosomes 2n = 46 2n = 56 2n = 56 2n = 56 2n = 56 2n = 56 2n = 28 metry. The 12 remaining plants were esti- Origin 2n+n 2n+O 2n+O 2n+O 2n+O n+O B mated to have 2n = 38. In situ hybridiza- 2n = 46 (36Tr + 10M) 6 ------tion in plant 6-529 and the 28-chromosome y 2n = 56(36Tr+20M) 1 15 9 1 29 1 - 6 plants from its progeny we analyzed 2n = 66 (36Tr + 30M) ------2n = 28 (18Tr + 10M) - - - - - 1 - revealed that all these hybrids combined 10 and 18 chromosomes respectively from 4 A Hybrids were backcrossed using maize pollen (n = 10). Chromosome number and origin of each genotype maize and Tripsacum (Figure la). The 38- involved in the production of plant 6-529. B Chromosome numbers detected in the progeny of each genotype involved in the production of plant 6-529. Tr = Tripsacum chromosomes; M = maize chromosomes. chromosome offspring showed a 20M + 18Tr complement (Figure lb). This confirmed that 28-chromosome Chromosomal Analyses analyses, pistils were recleared for inter- plants combine one complete haploid set Chromosome numbers of FI-,, and deriva- ference contrast microscopy using a ben- of chromosomes from Tripsacum 65-1234 tives in the five first backcross genera- zyl benzoate-dibutyl phthalate procedure and one from maize, and therefore have a tions were determined from root tips (pre- (Crane and Carman 1987) for internal de- polyhaploid structure similar to that of a fixed at 4°C in saturated S-hydroxyquino- tail observations within ovules. diploid. line for 4 h, fixed in Carnoy's solution for Fingerprinting analyses for 6-529 progeny 24 h, and stored at -4°C in 70% ethanol) testing. Total genomic DNA was extracted squashed and observed by phase-contrast according to Hoisington et al. (1994) from Morphology and Fertility microscopy. Flow cytometry was used to the Tripsacum parent 65-1234, plant 6-529, 6-529 and its 28-chromosome offspring estimate the chromosome numbers in and 13 offspring (10 morphologically sim- were morphologically uniform and char- plant 6-529 and its 64 offspring. Leaf sam- ilar to 6-529 and three morphologically acterized by a particular phenotype re- ples were prepared according to Galbraith different, including the four previously sembling Tripsacum 65-1234 but quite dis- et al. (1981) and analyzed with a PARTEC analyzed by in situ hybridization). Ten mi- tinct from accessions of the maize-Tripsa- CA-11 cytometer (Partec GmbH, Münster, crograms of DNA were digested with re- cum hybrid collection derived from it: Germany). We also performed fluorescing striction endonuclease HindIII, electropho- plant height was shorter, leaves narrower, in situ hybridization to determine the resed, and transferred onto nylon mem- and shorter, carrying both chromosomal complement of plant 6-529 brane. Hybridizations, stringency washes, fertile pistillate and atrophied staminate and four offspring (three morphologically detection, and exposure were made ac- . 38-chromosome plants strongly similar to the mother plant and one mor- cording to Hoisington et al. (1994). We differed from the wild type by recovering phologically different). Chromosome prep- chose maize probes UMC3, UMC8, UMC38, several traits from the crop: solid single arations were made following Jewel1 and UMC 310, and BNL16-06 because they were stem generally with one or more weak til- Islam-Faridi (1994), and DNA labeling and known to give multiple band patterns from lers, leaves borne alternately on either detection procedures according to Islam- a previous survey of an ORSTOM-CIMMYT side of the stem at each node, and male Faridi and Mujeeb-Kazi (1995). Tripsacum collection (unpublished data). and femal spikelets borne in separate in- florescences. All 28- and 38-chromosome Determining Mode of Reproduction plants were male-sterile. Anthers were not ResuIts Megasporogenesis analyses. Callose depo- dehiscent and pollen completely aborted, sition patterns throughout megasporogen- Chromosome numbers and morphology in possibly as a result of meiotic irregulari- esis are reliable indicators of the mode of progenies of plant FI-,, and derivatives, ties due to the presence of two nonhom- ologous genomes. In contrast, viable reproduction in Tripsacum spp. (Leblanc 1001, 2017, 3031, 4031, and 5-8, indicated et al. 1995b): callose synthesis, that is ob- that they reproduced through facultative seeds were produced when the plants served in and around cell walls of meiotic apomixis (Table 1B). Two types of proge- were pollinated with maize pollen. megasporocytes and their derived mega- nies were produced: maternal types (2n + spores, does not occur during diplospo- O) that showed both chromosome number Mode of Reproduction rous development. For callose analyses, and morphology similar to the mother Megasporogenesis in 6-529 and in its 28 30 pistils from 6-529 and 15-25 from 10 off- plant, and offtypes that differed in mor- and 38chromosome offspring was charac- spring (eight morphologically similar to 6- phology and chromosome number from terized by the complete failure of meiosis, ') 529 and two morphologically different, in- the mother plant. These offtypes were as revealed by callose analyses and inter- cluding the four previously analyzed by in classified (a) as 2n t n offtypes when ference-contrast observations. Among all situ hybridization) were first optically chromosome numbers corresponded to the pistils analyzed, no callose was detect- cleared using three sucrose solutions the fertilization of an unreduced femelle ed during megasporogenesis, and enlarged (200, 500, and 800 g/L, respectively) sup- gamete (i.e., 2n = 56 = 36Tr + 20M indi- megasporocytes developed directly into fe- plemented with aniline blue (100 mg/L), vidual in FI-,, progeny, 2n = 66 = 36Tr + male gametophytes through mitoses. This and examined using ultraviolet vertical 30M individuals in 4031 progeny), or (b) reproductive pathway is similar to the dip- epifluorescence microscopy (Leblanc et as polyhaploid (n + O) when they corre- losporous development observed in Z dac- al. 1995b). In addition, to confirm callose sponded to the parthenogenetic develop- tyloides 65-1234, the apomictic parent from

Leblanc et al - Apomixis in Maize-Tripsacom Polyhaploids 1O9 Figure 1. Root tip metaphases from 28- and 38-chromosome maizeTripsacum hybrids after genomic fluorescing in situ hybridization, using total genomic DNA from 7: dactyloides 65-1234 as a probe (yellow-labeled chromosomes are from the wild species, while the red-labeled ones are from maize). (a) Diploid-like structure in 28-chromosome polyhaploid hybrids that combine 10 maize chromosomes and 18 from the wild species. @) A 38-chromosome plant, combining one haploid chromosome set from Tripsacum and the complete chromosome stock from maize. which these plants were derived (Leblanc loid plant 6-529 also reproduced through from 2n = 56 hybrids, and are expected to et al. 1995b). facultative apomixis with the following segregate for mode of reproduction. All 28-chromosome plants from 6-529 patterns typical of apomictic reproduction Therefore, using this pathway (46 -+ 56 + progeny produced the same restriction (Bashaw 1980): (a) meiosis failure and ex- 38 + addition lines) and given the low - fragment patterns as the mother plant pression of diplospory; (b) maternal off- rate of residual sexuality in apomictic 56- when hybridized with the five selected spring (81.5%) at the morphological and chromosome hybrids, producing and iden- maize probes, suggesting that neither re- moleculal' levels; (c) presence of 2n f n tifying apomictic 38-chromosome hybrids combination nor egg cell fertilization had offtypes (i.e., 2n = 20M + 18 Tr) in a pro- requires the capacity to analyze a large occurred. Plants with 38 chromosomes portion (18.5%) commonly observed in number of progenies from 2n = 56 plants. showed additional bands which probably apomictic maize-Tripsacum hybrids de- When backcrossed to maize, apomictic 28- originated from the maize pollen thus con- rived from plant 65-1234 (Leblanc 1995). chromosome polyhaploids are thus a firming fertilization of some of the unre- Pathways of genetic transfer from Trip- good source of apomictic 38-chromosome duced gametes of 6-529 (2n + n offspring). sacum to maize through conventional plants, due to the frequency of 2n + n off- 2n = 28 polyhaploids have been also backcrossing have been already described types (-20%) in their progeny. This type found at a low frequency (<0.2%) in prog- by Harlan and de Wet (1977). Maize plants of backcross is also of great interest for eny of other BC1 plants derived from 65- possessing Tripsacum introgressions can cytogenetic analyses: recovering a 20- 1234. Those investigated for mode of re- be recovered in four steps from maize- chromosome apomictic plant from an ap- production also reproduced apomictically Tripsacum BC1 hybrids (2n = 56) through omictic maize-Tripsacum addition line as revealed by progeny tests based on sexual reproduction. However, because would imply that exchanges of genetic ma- both chromosome numbers and morphol- our goal is to transfer apomixis into maize terial do occur at meiosis. The particular ogy (at least 25 offsprings per mother and since male sterility strongly affects chromosome constitution of these poly- plant) and by cytoembryological analyses maize-Tripsacum F, hybrids and their 56- haploids appears to facilitate the study of (at least 15 pistils per mother plant). chromosomes derivatives, hybrids that the possible pairings between maize and ' contribute to the next step must be off- Tripsacum chromosomes. Although pair-

types arising through residual sexuality. ings and exchanges between maize and ( Discussion The frequency of such offtypes was esti- Tripsacum have already been reported 28-chromosome lpolyhaploid plant 6-529 mated to be -3% in T. dactyloides 65-1234 (Kindiger and Beckett 1990; Maguire originated through parthenogenetic devel- and its 2n = 56 derivatives (Leblanc 1960), the chromosome complement of opment of a reduced gamete in 56-chro- 1995). Putative apomictic addition lines the polyhaploids may allow researchers to mosome apomictic plant 5-8. Observa- that combine 20 chromosomes from maize determine whether preferential pairing ex- tions of megasporogenesis and results and a few from Tripsacum will result from ists and, if so, which chromosomes are in- from progeny testing (based on morphol- residual sexuality in apomictic 38-chro- volved. ogy, chromosome numbers and molecular masome plants (18Tr + 20M). 2n = 38 hy- In species that reproduce following the . data) clearly indicated that the polyhap- brids can be n + n offtypes in progenies aposporous type of gametophytic apomix-

11O The Journal of Heredity 1996:87(2) , is, nonreduction is controlled by one dom- been already reported in Ranunculus ar- Harlan JR, de Wet JMJ, Naik SM, and Lambert RI, 1970. Chromosome pairing within genomes in maize-Tripsa- inant allele A (reviewed by Asker and Jer- gouiensis (Nogler 1984a), and in two deriv- cum hybrids. Science 1671247-1248. ling 1992). Although very few data are atives (2n = 21 and 2n = 23) from inter- Hoisington D, Khairallah M, and Gonzaez-de-León D, available for diplosporous species, diplo- specific hybrids between Pennisetum 1994. Laboratory protocols: CIMMYT Applied Molecu- spory also appears to be simply inherited glaucum L. (x = 7) and wild Pennisetum lar Genetics Laboratory, 2nd ed. Mexico D.F.: CIMMYT. in Eragrostis curvula (Voigt and Bashaw species (x = 9) (Dujardin and Hanna 1986; Islam-Faridi MN and Mujeeb-Kaz¡ A, 1995. Visualization of Secale cereale DNA in wheat germplasm by in situ 1972), and in Tripsacum dactyloides (Le- Hanna et al. 1992). However, the origin of hybridization. Theor Appl Genet 90:595-600. blanc et al. 1995a). All models describing the chromosomes in these Pennisetum po- Jewel1 DC and Islam-Faridi N, 1994. A technique for so- the evolution of apomixis show that, un- lyhaploids could not be determined. In re- matic chromosome preparation and C-banding of t der this single-dominant-gene hypothesis, gard to the genetic factors responsible for maize. In: The maize handbook (Freeling M and Walbot V, eds). New York: Springer-Verlag; 484-493. apomixis will become fixed unless it dra- apomixis, this does not clearly indicate Kindiger B and Beckett JB, 1990. Cytological evidence matically reduces the fitness of its carriers nonpolyploid conditions for their expres- supporting a procedure for directing and enhancing r! (Marshall and Brown 1981; Pernès 1972). sion: both 2n = 21 and 2n = 23 polyhap- pairing between maize and Tripsacum. Genome 33:495- Apomixis is usually not expressed at the loids combined more than two Pennisetum 500. diploid level in agamic complexes; but dip- genomes. Therefore, triploid condition Leblanc O, 1995. Modes de reproduction dans le com- plexe agamique des Tripsacum et analyse d’hybrides loid genetic pools do make a large genetic still may occur, more particularly for the m~s-Tripsacumen vue du transfert de I’apomixie chez contribution to such complexes by pro- locus or loci involved in apomixis expres- le maïs (phD dissertation). Paris: INA-PG/University of ducing new variability through recombi- sion. In the 2n = 28 chromosome plants Paris IV/University of Paris XI. reported here that combine one genome Leblanc O, Grimanelli D, González de León D, and Sav- nation and, at the same time, by exchang- idan Y, 1995a. Detection of the apomictic mode of re- ing genetic material with polyploids (de from each maize and the wild species, di- production ìn maize-Tripsacum hybrids using maize Wet 1968; Savidan and Pernès 1982). The plosporous apomixis is normally ex- RFLP markers. Theor Appl Genet 90:119&1203. strong interrelationship between apomixis pressed and functional in a diploid-like Leblanc O, Peel MD, Carman JG, and Savidan Y, 1995b. expression and polyploidy in wild agamic condition, suggesting that apomixis might Megasporogenesis and megagametogenesis in several be expressed in a diploid crop like maize. Tripsacum species (). Am J Bot 8257-63. complexes is not well understood yet, and Maguire MP, 1960. A study of pachytene chromosome mechanisms of protection against apomic- pairing in maize-Tripsacum hybrid derivative. Genetics tic reproduction at the diploid level still 45651-664. References have to be elucidated. Dihaploids ob- Marshall DR and Brown HD, 1981. The evolution of apo- tained from apomicts in the Bothriochloa- Asker S and Jerling L, 1992. Apomixis in plants. Boca mixis. Heredity 471-15. Raton, Florida: CRC Press. Nogler GA, 1984a. Gametophytic apomixis. In: Embry- Dichanthium complex (de Wet 1968) and in Bashaw EC, 1980. Apomixis and its implication in crop ology of angiosperms (Johri BM, ed). Berlin: Springer- - Panicum maximum (Savidan 1982) were ei- improvement. In: Hybridization of crop plants (Fehr Verlag; 475-518. ther sexual and fertile or potentially apo- WR and Hadley HH, eds). Madison, Wisconsin: Ameri- can Society of Agronomy; 45-63. Nogler GA, 1984b. Genetics of apospory in apomictic mictic, as aposporous embryo sacs were Ranunculus auricomus. V. Conclusion. Bot Helv 94411- Crane CF and Carman JG, 1987. Mechanisms of apo- 422. observed, but sterile. Genetic models mixis in Elymus recfisetus from Eastern Australia and Noirot M, 1993. Allelic ratios and sterility in the agamic New Zealand. Am J Bot 74:477-496. such as the allelic ratio hypothesis, which complex of the Maximae (): evolutionary postulates that sterility arises when the de Wet JMJ, 1968. Diploid-tetraploid-haploid cycles and role of the residual sexuality. J Evo1 Bio1 6:95-101. the origin of variability in Dichanthium agamospecies. proportion of A to a alleles in a given in- Pernb J, 1972. Organisation évolutive d’un groupe aga- Evolution 22:394-397. dividual is >0.25 explain sterility in dihap- mique: la section des Maximae du genre Panicum (Gra- de Wet JMJ, Harlan JR, Engle LM, and Grant CA, 1973. minées) (PhD dissertation). Paris: University of Paris loids from the Panicum maximum complex Breeding behavior of maize-Tripsacum hybrids. Crop XI. Sci 13254-256. (Noirot 1993); however, they do not fit the Reeves RG and Mangelsdorf PC, 1935. Chromosome 2n = 28 and 38 maize-Tripsacum hybrids Dujardin M and Hanna WW, 1986. An apomictic poly- numbers in relatives of Zea mays L. Am Nat 69633-635. haploid obtained from a pearl millet x Pennisetum described here, for which the allelic ratios squamulatum apomictic interspecific hybrid. Theor Savidan Y, 1982. Embryological analysis of facultative are expected to be 0.5 and 0.33, respec- Appl Genet 723336. apomixis in Panicum maximum Jacq. Crop Sci 22467- 469. tively (the Tripsacum parental genotype is Galbraith DW, Harkins KR, Maddox JM, Ayres NM, Shar- Savidan Y and Dujardin M, 1992. Apomixie: la prochai- simplex for the allele conferring diplo- ma DP, and Firoozabady E, 1981. Rapid flow cytometric analysis of the cell cycle in intact plant tissues. Science ne revolution verte? Recherche 241:326-334. spory; Leblanc et al., 1995a). Besides, a 220:1049-1051. Savidan Y and Pernès J, 1982. Diploid-tetraploid-dihap few diploid species (largely from dicoty- Hanna WW, Dujardin M, Ozias-Akins P, and Arthur L, loid cycles and the evolution of Panicum maximum Jacq. Evolution 36:596-600. ledonous taxa) reproduce apomictically 1992. Transfer of apomixis in Pennisetum. In: Interna- (Asker and Jerling and apospory tional Workshop on Apomixis in Rice, Changsa, Peo- Stebbins GL, 1950. Variation and evolution in plants. 1992), ple’s Republic of China, Jan. 13-15 1992 (Wilson KI, New York Columbia University Press. was also expressed in a trisomic hybrid ed). Canberra, Australia: Center for the Application of Molecular Biology to International Agriculture (CAM- Voigt PW and Bashaw EC, 1972. Apomixis and sexuality 6 form (2n + 1) of Ranunculus (Nogler in Eragrostis curvula. Crop Sci 128434347, 1984b). BIA); 23-29. Harlan JR and de Wet JMJ, 1977. Pathways of genetic Received November 7, 1994 . The expression of apomixis in partially transfer from Tripsacum to Zea mays. Proc Nat1 Acad Accepted September 6, 1995 I’ fertile polyhaploids from apomicts has Sci USA 743494-3497. Corresponding Editor: Gary E. Hart

Leblanc et al *Apomixis in Maize-TripsacmPolyhaploids 111 REPRODUCTIVE BEHAVIOR IN MAIZE-TRIPSACUM POLYHAPLOID PLANTS: IMPLICATIONS FOR THE TRANSFER OF APOMIXIS INTO MAIZE

O. J. LEBLANC,D. GRIMANELLI,N. ISLAM-FARIDI, BERTHAUD, AND Y. SAVIDAN

i

i

Made in United States of America Reprinted from THEJOURNAL OF HEREDITY Vol. 87, No. 2, MarchIApril 1996 Q 1996 The American Genetic Association