sign in sign up
<<
Home , Apomixis

3

TßENDSin Genetics vo1.17 No.10 October 2001 597 Developmental genetics of gametophytic apomixis OlivierÍ Leblanc, Enrico Perotti and Ueli Grossniklaus Some higher plants reproduce asexually by apomixis, a naturalway of revealed an intriguing paradox: although the through seeds. Apomictic plants produce progeny that are an exact genetic developmental mechanisms underlying apomixis replica ofthe mother plant.The replication is achieved through changes in the are complex, most genetic analyses suggest that female reproductive pathway such thatfemale gametes develop without apomixis is inherited as a simple mendelian trait, and develop without fertilization. Although apomixis is a complex and results from one or a few mutations that affect developmentalprocess, genetic evidence suggests that it might be inheritedas a the normal course of sexual reproductiongc. simple mendeliantrait -a paradox that could be explained by recent data derived The growing popularity of apomixis research has from apomictic and model sexual organisms. The data suggest that brought developmental biologists into the field. Their apomixis might rely more on a global deregulation of sexual reproductive perception is that apomixis is not a true novelty in developmentthan on truly newfunctions,and molecular mechanismsfor such a plant development, but rather has evolved through global deregulation can be proposed. This new understanding has direct the rearrangement of the subprograms that consequences for the engineering of apomixis in sexual crop species. an constitute a normal sexual pathwaqr'~~. application that could have an immense impact on agriculture. Understanding apomixis will probably depend on gaining a broader understanding of how these Sexual in plants entails a series of programs are regulated in sexual plants, and developmental steps that culminate in the formation investigations of sexual development are making of the seed (Box l)*,z.The process is highly regulated important contributions to apomixis research. In this and most aberrations result in abortion2. review, we have attempted to summarize these recent Nevertheless, plants have retained considerable contributions, and in particular to show how they plasticity in the possible outcomes of reproduction. might help to revise current models of the genetic (see One example is APOMIXIS Glossary), where control of apomixis. meiosis and fertilization of the egg by male gametes are bypassed to result in the production of clonal Developmental aspects of apomixis progeny without a paternal contribution. Apomixis There has been independent of a therefore allows perpetuation of a fixed genotype remarkable diversity of seemingly unrelated through generationsg6. The ability to fix indefinitely apomictic mechanisms=. In this review, we focus on even highly complex genotypes, including high- one class of apomixis, GAMETOPHYTIC APOMIXIS,in which yielding hybrids, through apomixis would have the originates from an unreduced gamete. tremendous advantages in plant breeding and seed Reviews on a second class of apomictic development, production (Box 2).Apomixis does not occur in the ADVENTITIOUS EMBRYONY, in which the embryo major crop species, but is found in many wild differentiates directly from a somatic cell, can be species. The genetics of apomixis in these species found elsewhere=. has been under scrutiny for several decades, with Gametophytic apomixis requires profound the aim of transferring apomixis to, or inducing modifications of the processes that govern sexuality. apomixis in, their crop relatives. These studies have Whereas in animals gametes differentiate directly Daniel GrimaneIli. Olivier Leblanc Enrico Perotti Glossary IRD-CIMMYT, Apomixis ~~ Apomixis: through seeds, often referred to as agamospermy. Project, carretera Mexico Adventitiousembryony:Asubcategoryof apomixis, in which theembryodifferentiates directlyfrom a somaticcell intheovule, without Veracruz, Km 45.5, formation a megagametophyte; often referred to as sporophytic apomixis. The sexual pathway remainsfunctional and the apomicitic Batan.Texcoco. of EI embryo relies on sexually produced . Therefore, the seeds can contain apomiciticand sexual embryos. Ed Mexico, Mexico. Gametophyticspomixis: The second subcategory of apomixis comprising diplosporous and aposporous apomixis, but excluding '%-mail: adventitious embryony. lngametophytic apomixis, the megagametophyte (embryo sac) arises from an unreduced spore. DGrimanelli @cgiar.org Diplosporous apomixis: The megagametophyte develops from an unreduced megaspore. Ueli Grouniklaua Aposporous epomixis:The megagametophyte develops from a somatic cell within the . Dept of Plant Apomeiosis: Genericterm describing the failure of meiosis, via diplospory or apospory, ingametophytic apomicts. Development, Institute of Pseudogamousapomictt: Apomictic plants in which endosperm development depends on the fertilization of the central cell while the Plant Biology, embryo develops parthenogenetically. developwithoutfertilization,through Zollikerstrasse 107, Autonomous apomicts: Apomictic plants in which boththe embryo and the endosperm University ofZiirich, . Parthenogenesis:Activation of embryo (or endosperm) development in the absence offertilization. CHdOO8Zii~ic_h,I Switzerland 7-- ... i I 1

I 0168-952501/5-seefrontmetterO2001 ElsevferScienceLtd.All rightsr

O 10026364 I. 1 _. I 598 TRfNDSin Genetics Vol.1 No.10 October 2001

Box 1. in higher plants The lifecycle of higher plants (Fig. la) several cell divisions and extensive cell Megasporogenesis, or female meiosis alternates between a diploid sporophytic differentiation.The nuclei in red are occurs within a specialized organ of the phase, which producesthe spores through diploid (unreduced) and the nuclei in , the ovule. It starts with the meiosis (the sporophytic phase, dark purple are haploid (reduced). differentiation of the megaspore mother blue), and a haploid gametophytic phase Fig. Ib shows sporogenesis and seed cell,which undergoes meiosis to produce (blue),which produces the gametes after development in higher plants. a tetrad of megaspores, only one of which usuallysurvives to form the functional megaspore. Microsporogenesis (male meiosis) occurs within theanthers, and producesfour , all ofwhich survive. The differentiation of the spores marksthe beginning of the gametophytic generation. Both the micro- and megaspores divide Microgametophyte (n) V mitoticallyto form the multicellular : the microgametophyte (pollen grain) on the male side and the megagametophyte (embryo sac) on the Fertilization femaleside. The mature megagametophyte is embedded in Megaspore Zygote CQ) '6 mother cell \Br mother cell /y maternal sporophytic tissue of the ovule, and typically comprises seven cells. Two ofthose cells, the egg cell and the binucleatecentral cell participate in double fertilization that is a unique feature ofthe angiosperms. The mature 0-4 two Sporogenesis Megagametogenesls Seed development microgametophyte contains sperm cells. Onefertilizes the egg, to produce the zygote and eventually the embryo. The second sperm fertilizes the central cell, producing the endosperm. Hence, the mature seed contains three types of tissue: the mlaternal tissues, and two hybrid tissues of different levels-the diploid embryo and thetriploid in Fig. I 1 TRENDS Genelics endosperm.

from the meiotic products, gametogenesis in plants is and germination8, although its relative importance preceded by formation of multicellular gametophytes varies greatly hom species to species. For example, (Box 1). In the female reproductive organ (the ovule), the endosperm never forms in certain orchids and is a single cell normally becomes committed to the transient in Arabidopsis thdianu,but constitutes sexual pathway, undergoes meiosis and forms a 80-90% of mature cereal seeds. tetrad of reduced spores. One of these spores divides The switch from a normal sexual pathway (Box 1) mitotically to form a multicellular female to an apomictic pathway entails at least three major , the embryo sac, which contains the steps (Box 3): female gametes. The three other spores degenerate. circumvention of meiosis (a process called In the male reproductive organs (the anthers), all APOMEIOSIS); meiotically produced spores undergo mitotic * development of the embryo independently of development to form mature male gameto hytes - fertilization (i.e. PARTHENOGENESIS); the pollen grains. Apollen grain usually CO sists of formation of a functional endosperm. two sperm cells contained within a large v getative These requirements are met by a variety of means in cell. Fertilization involves two pairs of ga etes. One apomictic plants, and as a result the mechanisms of sperm fuses with the egg cell to form the z gote and apomixis are numerous (Box 3), although they share eventually the embryo, and the second spe m fuses common characteristics. First, most if not all with the binucleate central cell to form the apomicts are polyploid. Second, apomixis affects only endosperm. "he endosperm has important nutritive the female reproductive pathway and male gametes and physiological functions during seed development1 are still produced through meiosis. Third, most

http:/ltig.trends.com Box 2. Apomixis and its potential in agriculture Apomixis can be viewed as a short-circuited Gametophytic Sexual sexual pathway, where meiosis and apomixis reproduction fertilizationare initiated before completion oftheprevious developmental step. Fig. Ila comparesthe steps in sexual and apomictic two I development, highlighting major elements of apomixis-apomeiosis and Apomeiosis parthenogenesis. Endosperm formation I and apomixis can be autonomousor rely on Reduced /‘ Reduced fertilizationofthecentral cell. egg celi polar nuclei Cloning is no less a hot topic in plants than it is in the animal world, and apomixis I Autonomy hasattracted mud7 attention from Parlheno- academics and commercial researchers. For the academic, apomixis offers an ideal opportunityto study an efficient natural P cloning system, and investigate a wide t rangeof biological questions,from the molecular basis of to the evolutionary role of sex. However, the current wave of apomixis research is largely driven by marked interest from the private biotechnologysector and the seed industry. Apomixis is potentially a valuable means of crop improvement, one application being hybrid seed production (Fig.Ilb). In crops, the production of hybrids is complex and expensive, requiring large- scale crossing schemesunder controlled conditions. Apomixis còuld reduce the cost I of seed production, while ensuring purity. In addition, apomixis should allowfarmers, especially in the developing world, to save seed from hybrid plantsforthe next cropping season and retain the benefits of hybrid vigor, which is normally lost in the next generation because of segregation. Unreduced cells are denoted by rectangles, Apomixis and reduced cells are denoted as ovals. Key

Fig. II TRENDS In Genetics developmental eventsthat areaffected in apomictic species are highlighted in yellow.

apomicts are facultative, in the sense that a temporal and spatial constraints on sexual proportion of the progeny still results from sexual developmental processes7. Asexual pathways are reproduction. Hence, apomixis does not replace built by reassembling, in space and time, the sexuality; rather, it coexists with sexual development elements of ‘normal’ sexual reproductive pathways. within the same plant. Finally, the precocious initiation of a developmental step before completion of Genetic analysis of apomixis: few genes or many genes? the previous one is a hallmark of apomixis; that is, Apomixis is a heritable trait, but its genetic control is apomixis corresponds to a ‘short-circuiting’ of sexual unclear. How many genes control apomixis, and are pathways, where gamete formation occurs without those genes the same in the various forms of meiosis and embryogenesis without fertilization4. In apomixis? Apomixis is found almost exclusively in certain types of apomixis, the fate of cells within the polyploid, highly heterozygous and genetically poorly ovule is altered, and somatic cells that are usually not characterized species, making its genetic dissection committed to the reproductive pathway take part in The development of female gametes apomictic reproduction (Box 3).Apomixis can be without meiosis (i.e. apomeiosis) is the element of viewed, therefore, as the result of a relaxation of apomixis that has attracted the most interest.

http:lltig.trends.com Box 3. Mechanisms of apomictic development Gametophytic apomixis can follow one of genome ofthe mother (a,b,c). The Many variations have been observed two types of pathway: OIPLOSPOROUSor unreduced nuclei are shown in red, and in both diplosporous and aposporous APOSPOROUS (Fig. 111). In diplosporous reduced nuclei in purple. lnaposporous pathways. In the Antennariatype of pathways the gametophyte is derived pathways the megaspores are derived diplospory (a),the unreduced spore is from the megaspore mother cell, and the from somatic cells within the ovule that formed without undergoing meiosis (a), megaspores result from an aberrant or develop directly into a megagametophyte, whereas in the type, it results modified meiosis that restores the bypassing meiosis (d,e). from the restitution of the nucleus at meiosis I (b). In the type, meiosis Time is normal but preceded by an extra round ofDNA replication before meiosis I (c). In 'Seed all these examples, the egg cell forms an Meiosis Gametogenesis formation embryo parthenogenetically, that is, without fertilization by a male sperm. Variations also exist for apospory. lnthe Sex Panicum type, the megagametophyte is mature after only two mitoses and hence containsonlyfour nuclei (d).in the Nieracium type, three mitoses occur, and the embryo sac contains eight nuclei, closely resembling the sexual one (e). Another fundamental difference (not illustrated) among apomicts relates tothe formation of the endosperm. In autonomous apomicts, both embryo and endosperm develop parthenogenetically. In contrast, pseudogamous apomicts still require fertilization of thecentral cell forthe formation of a hybrid endosperm to support seed development. The mode of endosperm development does not correlatewiththe modality of apomeiosis: some aposporous plants have autonomous endosperm while Fig. Ili %ENDS in GenefiG others are pseudogamous, and the same is true for dipllosporousspecies.

Genetic analyses conducted in tetraploid Panicum This suggests a functional' relationship between the maximum, a forage grass, and Ranunculus traits, in spite of their genletic independence. auricornus, a dicotyledonous plant, have shown that The contradiction between these data and apomeiosis is inherited as a single dominant observations in Panicum and Ranunculus illustrates a mendelian This finding has often been taken characteristic of apomixis genetics: major conclusions as evidence for monogenic inheritance, although a are drawnfrom a few case studies and attempts to mendelian trait can encompass anything from a derive general principles often prove unsatisfactory. single gene to an entire (e.g. the Comparative mapping studies of different forms of Y-chromosome of humans). apomixis in various grass species helps to explain why In the same experiments, apomeiosis and this is so. The genome of the grasses has been parthenogenesis were shown to co-segregate strictly, remarkably stable during evolution, and gene content suggesting that these two components rely on the and gene order along the is well same genetic control, or that parthenogenesis is a conserved between homologous chromosomes, even pleiotropic consequence of apomeiosis5-G.Recent among distantly related ~peciesll-'~.Hence, reports in another dicotyledonous plant, Erigeron comparative mapping among grasses of apomixis loci, annuusgand in dandelions ()I0, which are always inherited as simple traits, offers a however, have shown that in some species way to determine whether different apomictic apomeiosis and parthenogenesis can segregate processes rely onidentical! or different loci. Attempts to independently, suggesting that they probably rely on map apomixis with common molecular probes in different genes. Note, however, that although several species, including Dipsacum14, Brawhizrials apomeiosis can occur in plants in which and PaspalumI6, three forage grasses,have shown the parthenogenesis does not, the reverse is not true. genomic regions that regulate apomixis to be distinct

http:/ttig.trends.com (i.e. nonhomologous) in those species. This means that Apomixis and genomic imprinting apomixis, in its various forms, probably arose in Apomixis is quite fiequent in floweringplankP. By dflerent grass species through the action of different contrast, mammals do not have apomixis-related genetic loci. phenomena, such as parthenogenesis, because of genomic imprinting. Genomic imprinting refers to Master regulators or complex loci? parent-of-origin specific gene expression, and it renders Apomixis in these different species, although relying maternal and paternal genomes functionally different on different loci, is always inherited in a simple to each other. It can afFect entire genomes, manner, as one or two mendelian traits at most. The chromosomes or individual loci (reviewed in Refs 22,23). presence of such apparent simplicity in several In mammals, genomic.imprintingrenders the maternal different systems implies that plant sexual pathways and paternal genomes complementary for genes that can be ‘short-circuited‘ towards apomixis relatively are essential to embryo development.23-25,and thus it easily, by means of simple alterations of single loci, ensures that both genomes are present in the zygote. and at many different points in the pathway with The relatively &quent occurrence of apomixis indicates similar effects. that embryo development in plants must be governed Apossible explanation is that the loci correspond to by radically different rules. In particular, it suggests master regulatory factors. Peacock proposed that that the presence of a paternal genome might not be an apomixis might result from the ectopic expression of absolute requirement for embryo development. the transcriptional cascade that induces embryo sac Vielle-Calzada and colleagues recently offered a devel~pment’~.Expressed normally &er meiosis and possible explanation for this observation”. In a megaspore differentiation, the cascade’s expression in survey of 20 genes in Arabidopsis thaliana, no the early germ cell or in somatic cells would lead to paternally derived transcripts were detectable in the embryo sac formation before meiosis. A global inducer developing seed (embryo or endosperm) for the first of gametogenesis or embryogenesis has yet to he few days after fertilization. Their observation identified in plants, but in budding yeast most of the suggests that this early phase of development could postmeiotic events depend on the expression of Ndt80, be largely under maternal control, through a ’ a meiosis-specific transcription factor’8J9. When combination of maternal products stored in the ectopically expressed in vegetative cells, Ndt80 induces gametes before fertilization and uniparental the expression of almost the entire set of sporulation expression of some genes after fertilization; that is, genes. The ectopic or heterochronic expression of a genomic imprinting. This indicates that, in apomicts, master regulatory gene such as Ndt8O in an unreduced the fundamental developmental processes of early plant cell, or before meiosis, could cause apomeiosis in embryo growth are essentially unchanged from those a sexual plant. However, no homologs of Ndt80 have in sexual plants; in both instances, to a large extent, yet been reported in higher eukaryotes. only the female genome is involved. Hence, the lu contrast to this simple control mechanism. difference between apomictic and sexual seed recent reports indicate that the ‘apomixis genes’ development should lie in the mechanisms that might be rather complex loci. In tetraploid ?kipsacurn regulate the activation of the corresponding program, dactyloides, a wild relative of maizeI4, apomeiosis not the program itself. This would support a simple segregates as a dominant mendelian trait. Linkage inheritance model for parthenogenesis: although the maps for the segment that controls apomeiosis in program is necessarily complex, its activation might tetraploid apomicts were used to compare the same rely on a limited number of regulatory factors. This linkage group in diploid sexual plants. The data show resembles the hypothesis put forward earlier for that recombination is strongly suppressed in the gametogenesis, and is equally hypothetical. segment that controls apomeiosis, which therefore Nevertheless, several regulatory factors have been behaves as a single genetic unit. This unit represents identified in plants that, when mutated or expressed 40 cM on the map developed for sexual Dipsacurn; ectopically, induce partial embryo development or this block probably contains several hundred genes, of embryo-specific gene e~pression~~.~~.The which an unknown number participate in apomixis. identillcation of proteins that regulate or are targets Similar results were found in unrelated species. In of such genes might shed light on the mechanisms of Penniseturn squamulatum2Q-21,Paspalurn simplex16, embryo initiation, and therefore on the mechanisms and Erigeron annu$, the trait was mapped to a that are altered in apomictic plants. segment showing no recombination. Although these data do not indicate that apomixis relies on complex Dosage effects and endosperm development in genetic control, they show that a simple genetic apomictic plants control cannot be inferred from segregation analyses. Endosperm development has rarely been considered So, how many genes control apomixis? The question by those studying the inheritance of apomixiss. The remains open, but our understanding of sexual endosperm is equally important to apomictic and reproduction has increased in recent years, offering sexual seed development, but its development differs some hypotheses to explain two aspects of apomixis: significantly between the two pathways. In some embryo development and endosperm development. apomicts (such as Erigeron or %ramcum) the

hnp:lhig.trcnds.com * 4. a 602

endosperm develops autonomously; that is, istiation of endosperm development has to rely on parthenogenetically without fertilization (AUTONOMOUS alternative pathways. Screens for mutants that allow APOMICTS).In others (such as Panicum, Pennisetulnor fertilization-independent seed development in Dipsacurn), it still depends onfertilization of the &abidopsis thaliana have identified a class of central cell by a male sperm (PSEUDOGAMOUS hpo”). mutations that partially allow autonomous Both these types of apomict differ from their endosperm de~elopment~~.The three genes of the FIS sexual counterparts in the relative contributions to (FERTILIZATION-INDEPENDENTSEED) class, 35), the endosperm of the maternal and paternal FIS2 (Ref. FIE (FERTILIZATION- genomes. In diploid plants, the embryo is typically INDEPENDENT ENDOSPERM)36 and MEA diploid while the endosperm is triploid, receiving two (MEDEA)37,repress endosperm formation in the maternal genomes and one paternal genome (2m:lp). absence of fertilization. Molecular cloning showed In apomicts, the relative genomic contributions of the that the genes of the FISclass share structural and maternal and paternal genomes are altered, because functional similarities with factors thought to control the central cell contains unreduced nuclei, whereas higher-order clxomatin structure in animals: both the male gametophytes are normal. For example, MEA and FIE encode polycomb group (PcG) proteins, tetraploid autonomous apomicts produce endosperm and FIS2 encodes a zinc finger protein, some of which with a 8m:Op ratio, whereas a tetraploid are involved in PcG complex formatiorP. Endosperm pseudogamous apomict produces a pentaploid development in autonomous apomicts probably endosperm witharatioof 8m:lp. requires the specific inactivation of the repressive In many plants, genome dosage is critical to seed PcG complexes. Note, however, that fis class mutants development. In maize and probably most other allow only partial development of the endosperm. cereals, normal development of the endosperm Recent results show that autonomous endosperm requires a maternal to paternal genome ratio of development progresses further if fie is combined 2m:lp; deviations lead to seed abortionz9.This with genome-wide hypomethylati~n~~,indicating that requirement has been regarded as a strong barrier to the actual mechanisms operating in apomicts rely on the emergence of apomixis during evolution. and to the deregulation of a larger number of genes. introgression of the trait in crop specie^^,^. In Moreover, it has been shown that MEA and possibly apomicts, the requirement is rela~ed~~.~’or met by, FIS2 are regulated by genomic imprinting; that is, only modifications of gametogenesis or fertilization3JQ1. In the maternally inherited alleles are expressed after Panicum and many aposporous plants, the apomictic fertilizati~n~~~.This regulation is reminiscent of the forms have a modified embryo sac: the central cell of mechanisms described by VielleCahada and which contains only a single unreduced polar nucleus. colleagues in the embryo. Nevertheless, the FIS This nucleus is fertilized to produce an endosPerm mutants allow partial endosperm development, but no with a 2ni:lp ratio“-”.Apomictic ?F~~SUCILO~and embryo development. ‘l’his is important. because it I’aspuli~mare apparently relatively immune to suggests that activation of the egg and activation of‘the dosage de~iations”~*~~.Note that a strict requirement central cells rely on different processes, and hence that for a genome ratio of 2m:1p, which is particularly in autonomous apomicts parthenogenetic development pronounced in maize, might not be a general feature of the embryo occurs independently of parthenogenetic of sexual plants. For example, seed viability in development of the endosperm. Arabidopsk is much less susceptible to variation from the 2m:lp ratio.32But in a crop plant such as maize Afunctional role for with marked susceptibility to dosage effects in the The studies that we have summarized illustrate the endosperm, inducing apomixis would probably create complexity of the regulatory pathways that need to problems with seed viability. Indeed. lines where be altered ifwe are to develop apomictic seeds. introgression of apomixis into sexual crops from wild Overall, they suggest that apomixis might require relatives has been attempted typically have a high the coordinated deregulation of several genes level of seed abortion. This occurs because the involved in reproduction. Polyploidy is a possible mechanisms responsible for relaxation of the dosage route to such deregulation. constraints were apparently not transmitted with Complete or segmental polyploidy is apparently apomeiosis and parthenogenesis“. The genetic bases ubiquitous in gametophytic apomicts. Rare instances of dosage response in the endosperm remain of diploid apomicts have been reported3, for example in unknown, but it is clear that they represent an Arabis holboellii, a relative ofArabidopsis, but as the essential aspect of the genetics of apomixis. recent release of the Arabidopsis genome sequence revealed, those might well be paleopolyploids4g. Initiation of endosperm development in apomicts Several authors have proposed that apomixis might In sexual plants, endosperm develops in response to eventually be expressed in diploids, but that apomixis fertilization. Pseudogamous apomicts are similar to alleles are not transmitted through haploid gametes, their sexual counterparts in that respect, and do not or are lethal in diploid progenies4M6,suggesting that require specific modifications. The situation is more the relationship between apomixis and polyploidy complex for autonomous apomixis, where the might be structural rather than functional. Other

hnp:lhig.trends.com ~ ~ ~~

. .- T/?€A!DSinGenetics Vo1.17 No.10 October 2001 603

Box 4. Polyploidy and apomixis Several authors have proposed that induce apomixis (seetext). Two andthe genome asynchrony model polyploidy (complete or segmental possible models are shown here: (Fig. 1Vb). duplication of the genome) might the ploidy regulation model (Fig. ¡Va) The ploidy-induced apomixis hypothesis postulates that alleles of (a) The ploidy regulation model some reproductive genes are differentiallyexpressed depending on Complete or segmental genome duplication ploidy levels. A rise in ploidywould therefore induce a modification of the A: expression profile of those genes. premeiotic genes Fig. IVa showstheoretical expression w"mx 8: meiosis genes I C: gametogenesisgenes profiles overtime of genes with a specific D. - post-fertilization genes role in regulating premeiotic events A), !&.:$,$y(ì)>.?; ' , (gene meiosis (gene E), I . '..-..n.:*,,.h I. I gametogenesis (gene C) and Pre- Sporo- Mega- Post- Gametogenesis without meiosis meiosis genesis gametogenesis fertilization -postfertilization events (gene D). The red profiles indicate the window of of (b) The genome asynchrony model gene expression. In the event ploidy-induced apomixis, partial or Ecotype 1 complete genome duplication would deregulate meiosis-specific profiles, or 88 A etwa- induce ectopic expression of 8 gametogenesis genes, resulting in Kim&, *- C apomeiosis. Resulting allopolyploid 1 The 'genome asynchrony' hypothesis postulates that two ecotypes with x"wa- ' X --+- divergent reproductive behavior are .¶mmm" brought together by hybridization. In Ecotype 2 Fig. IVb, ecotype2 has a much shortë& meiosis than the ecotype 1. In the Gametogenesis without meiosis resulting allopolyploid, asynchronous - expression of meiosis and gametogenesis-specific genes leads to . the ectopic expression of gametogenesis, TRENDS in Genelics without meiosis.

evidence, however, suggests a functional role for directly influences gene expression in, for example, partial or complete genome duplications. In yeast49,maizea and Arabidopsisj'. In Arabidopsis, Paspalum notatum, for example, simple comparative profiling of cDNAs in two diploid species chromosome doubling of diploid plants produces and the corresponding allotetraploid showed that up apomictic autotetraploids, indicating that the alleles to 20 of 700 genes surveyed in the allotetraploids are for apomixis are present in the diploids, and that eventually silenced; this results in considerable their expression is ploidy dependent4' (Box 4). variation in morphology, flowering time and Similarly, the 'genome asynchrony hypothesis' put fertility5'. In yeast, a genome-wide comparison of forward by Carman4Bproposes that polyploidy, when gene expression profiles between isogenic strains at originating from genetically divergent genomes, different ploidy levels, from haploid to tetraploid, induces apomixis. Hybridization of divergent indicates ploidy regulation of essential genes, genotypes could cause asynchrony in the expression including key regulators of the cell cycle49. The maize of the regulatory genes that control reproductive dataa show, moreover, that segmental polyploidy can programs, leading to the concurrent asynchronous induce trans-effects, affecting genes that are not part expression of essentially unaltered developmental of the segment of higher ploidy. Polyploidy, however, programs. For example, if embryo sac is clearly not a sufficient condition for apomixis. developmental programs are superimposed on Indeed, polyploids account for more than 50% of megasporogenesis, meiosis could be omitted. angiosperm species, most of them sexual - a Is it plausible that polyploidy per se induces the disproportionately large percentage when compared deregulation of genes crucial for reproduction, thus with apomictic species. The model remains giving rise to apomixis in some specific situations?? speculative, but a trend for polyploidy to play a part in Several recent studies have shown that polyploidy the regulation of apomixis is clearly discernable.

hnp:lhig.trends.com

... . Conclusions In apomicts, such alterations allow the mangementin Our understanding of apomixis is changing rapidly. time and space of the programs that ccanstitute a sexual Until recently, it was widely accepted that apomixis was pathway. Therefore, understanding the regulation of Acknowledgements the result of a few mutations within the reproductive apomixis will depend on a better understanding of the We apologize to our colleagues whose pathways, but the results of molecular mapping in basic process of sexual , about publications we failed to apomicts challenge this assumption. Also, mapping which fundamental questions remain unanswered. quote owing to space shows that naturally occurring forms of apomixis have What are the signals that commit a cellto the meiotic limitations. We thank probably evolved from distinct genetic bases. Defining pathway? What factorscontrol the initiation of Dave Polland for help with ' editing, and Markus general principles for apomixis genetics might therefore megagametogenesis? How is the egg cdlactivated? Affolter, Maggy Collinge, be an impossible task. Rather, the mechanisms of What factors control the expression ofimprinted genes? Francoise Jardinaud and apomixis are interesting because of their diversity. They At the same time, the diversity of apomictic phenomena Pierre Dubreuilfor offers a unique opportunity to study and understand commeots on the show that sexual reproduction in plants has remained manuscri pi. highly plastic, and can be alteredin many different ways. the plasticity of reproductive development in plants.

References 19 Chu,S.etal. (1998) The transcriptional programof 36 Ohad,N,etal.(1999) Mutationsin FIE. aWD 1 Grossniklaus, U. and Schneitz, K. (1998) Genetic sporulation in budding-yeast. Science282.699-705 polymmb group gene. allow endosperm development and moleculnr control of ovule development and 20 Ozias-Akins. P. el al. (1998) Tightclustering and without fertilization. PhnfCdll. 407-416 megagametogenesis. Seminars in Cell t hJl. hemixygosity ofapomixis-linked molecular 37 Grossniklaus. U. e/ al. (1998)MaLernalcontrol of Bio¿.9% 227-238 markers in Pennise~u.rnsqu.arnirlalu.cn implies embryogenesis by MEDEA. apolycoinb group 2 Yang, W.C. and Sundaresan. V. (2000) Geneticsof genetic control of apospory by a divergent locus genein Arabidopsis. Sckflce280.446-450 gametophyte biogenesis in Arabidopsis. Curc that may have 110 allelic form in sexual genotypes. 38 Vinkenoog. R. el al. (2000) Hypomethylation OpNin. PlanL Biol. 3,5.3-.57. Proc. Na11 Acad. Sci. U. S. A. 28.51274132 promotes autonomous endosperm development 2 3 Nogler. C.A. (1984) Gametophytic apomixis. In 1 Roche, D. e/ al. (1999) An apospory-specific and rescues postfertilizationlethality in fi(! Embryology ofAng¿osperms (Johri. ß.M.. ed.). genomic region is conserved beLween buffelgrass mutants. Plan1 Cell 12.2271-2282 pp. 475-518. Springer-Verlag (Ccnchnrs siliaris L.) and Pecwiselum 39 I

hnp:/ltig.trends.com