Review TRENDS in Plant Science Vol.12 No.9

Towards the molecular basis of heterosis

Frank Hochholdinger and Nadine Hoecker

University of Tuebingen, Center for Plant Molecular Biology, Auf der Morgenstelle 28, 72076 Tuebingen, Germany

Heterosis describes the superior performance of (midparent) value of the two parental inbred lines heterozygous hybrid plants over their homozygous (MPH = F1 À [(P1 +P2)/2]). BPH on the other hand parental inbred lines. Despite the rediscovery of this indicates that a hybrid trait performs significantly better phenomenon a century ago and its paramount agro- than the better (Pb) of the two homozygous parental inbred nomic importance, the genetic and molecular basis of lines (BPH = F1 À Pb). heterosis remains enigmatic. Recently, various pioneer studies described differences in genome organization Genetic hypothesis to explain heterosis and gene expression of hybrids and their parental inbred Despite the rediscovery of heterosis about a century ago lines. At the genomic level, a significant loss of colinear- and the suggestion of various genetic models to explain this ity at many loci between different inbred lines of maize phenomenon, little consensus has yet been reached about was observed. At the level of gene expression, complex the genetic basis of heterosis [10–13]. The most prominent transcriptional networks specific for different develop- genetic hypotheses to explain heterosis are the ‘dominance’ mental stages and tissues were monitored in maize (Zea (see Glossary) and ‘overdominance’ (see Glossary) hypoth- mays), rice (Oryza sativa) and Arabidopsis (Arabidopsis eses. Both describe nonadditive phenotypic behaviour as a thaliana). Integration of this complex expression data consequence of genetic differences between distinct homo- might contribute to improve our understanding of the zygous parental inbred lines and their heterozygous molecular basis of heterosis. hybrids. The dominance hypothesis explains heterosis by the complementing action of superior dominant alleles The phenomenon of heterosis from both parental inbred lines at multiple loci over the Heterosis (see Glossary), or hybrid vigour, describes the corresponding unfavourable alleles leading to improved superior performance of heterozygous F1-hybrid plants in vigour of hybrid plants [14–17]. The overdominance hy- terms of increased biomass, size, yield, speed of develop- pothesis [5] attributes heterosis to allelic interactions at ment, fertility, resistance to disease and to insect pest, or to one or multiple loci in hybrids that result in superior traits climatic rigours of any kind compared to the average of compared to the homozygous parental inbred lines. their homozygous parental inbred lines [1,2]. Self-pollina- Finally, the epistasis (see Glossary) hypothesis [18], con- tion of hybrids over several generations leads to a gradual siders epistatic interactions between non-allelic genes at reduction of heterozygosity and vigour in these plants, a two or more loci as main factor for the superior phenotypic phenomenon that is known as inbreeding depression (see expression of a trait in hybrids. However, it has recently Glossary). Hence, heterosis and inbreeding depression are been demonstrated in tomato introgression lines that het- different aspects of the same phenomenon [3]. Heterosis erosis is manifested even in the absence of epistasis [19]. was first described by Charles Darwin in 1876 after he In summary, all these genetic hypotheses make the observed that progeny of cross-pollinated maize (Zea mays) combination of a considerable number of genes responsible were 25% taller than progeny of inbred maize [4]. The for the more vigourous phenotypes of hybrids over inbred phenomenon was rediscovered independently by George H. Shull and Edward M. East in 1908 [5,6]. Since then, heterosis has been extensively exploited in plant breeding, Glossary particularly in maize. Today, ca. 95% of the maize acreage Colinearity: Genetic concept that genes correspond to each other within in the USA and 65% of the maize acreage worldwide is different individuals of a species and are arranged in the same linear sequence. planted to hybrids [7]. Heterosis is most evident for adult Dominance: Term from quantitative genetics explaining improved vigour of traits (Figure 1a) but is already manifested during embryo hybrid plants by the complementing action of superior dominant alleles from both parental inbred lines at multiple loci over the corresponding unfavourable [8] and early seedling development [9] (Figure 1b). The alleles. terms midparent heterosis (MPH) and best parent hetero- Epistasis: Term from quantitative genetics explaining the superior phenotypic expression of a trait in hybrids by interactions between non-allelic genes at two sis (BPH) describe the degree of phenotypic difference of a or more loci. trait in a hybrid (F1) compared to its parental inbred lines Heterosis: Improved agronomic performance of heterozygous F1-hybrids in comparison to their genetically different homozygous parents. (P1,P2). MPH indicates that a trait displays a hybrid Inbreeding depression: Gradual reduction of heterozygosity and vigour in performance that is significantly better than the average hybrid plants after self-pollination of these plants over several generations. Overdominance: Term from quantitative genetics attributing superior hybrid Corresponding author: Hochholdinger, F. traits compared to their parental inbred lines to allelic interactions at one or ([email protected]). multiple loci. Available online 27 August 2007. www.sciencedirect.com 1360-1385/$ – see front matter ß 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.tplants.2007.08.005 428 Review TRENDS in Plant Science Vol.12 No.9

Figure 1. Phenotypic manifestation of heterosis. (a) Heterosis is typically detected in adult traits such as yield or cob size. (b) Heterosis is already manifested during seedling development. lines. This also implies that positive and negative effects of level between different inbred lines of maize. First, it was various loci might compensate each other, which makes it demonstrated that among ten genes in the bz region of the difficult to support one hypothesis over the other [20].Itis McC inbred line, only the proximal six have counterparts in important to keep in mind that the quantitative genetics the B73 inbred line [22]. Similarly, whereas 22 gene copies of terms ‘dominant’ and ‘overdominant’ cannot be directly the a-zein storage protein subfamily z1C were detected in associated with the quantitative behaviour of phenotypic the BSSS53 inbred line, only 15 z1C genes were present in traits or with molecular principles. the B73 genomic region [23]. Only seven of the BSSS53 and six of the B73 z1C genes had an intact coding region, and Molecular approaches to study heterosis only two intact genes were present in both inbred lines. With novel molecular tools at hand, several laboratories Finally, among 72 genes identified in various chromosomal have recently started to analyze different molecular regions of the inbred lines B73 and Mo17, 27 genes were aspects of heterosis. In the past, quantitative trait loci absent in one of the inbred lines [24]. (QTL) analyses were a first step towards the molecular In maize, many genes are members of small gene understanding of heterosis [21]. Although these studies families. Therefore, gene deletions in inbred lines might were able to demonstrate that heterosis is defined by a have only minor quantitative effects on plant performance limited number of individual genes inherited in a complex because these genes might often be functionally compen- way, none of the QTLs associated with heterotic traits have sated by duplicate copies elsewhere in the genome [22]. yet been cloned. However, with novel molecular tools that However, hemizygous complementation of many genes allow for a comprehensive QTL-based phenotyping fol- with minor quantitative effects in hybrids might lead to lowed by map-based cloning, it might be possible to identify a significantly increased performance of hybrid plants and loci controlling heterotic traits in the near future [21]. would be consistent with the dominance hypothesis [22] More recently, several studies initiated the dissection of (Figure 2). This concept would also explain inbreeding heterosis at the gene expression and genome organization depression by the loss of functional genes in subsequent level. These analyses can be classified in three categories. generations when hemizygous genes get lost after several First, genome organization has been analyzed in maize at rounds of self-pollination of the hybrids (Figure 2). The various loci for different inbred lines, and expression of a high degree of non-colinearity in the genomes of different limited number of genes in these regions has been tested. inbred lines of maize might thus, in part, explain the Second, transcriptome-wide gene expression profiles have exceptionally high degree of heterosis in this species. been determined for different developmental stages and However, it is probable that other molecular mechanisms tissues in different species including maize, rice (Oryza might also be involved in heterosis because it is unlikely sativa) and Arabidopsis (Arabidopsis thaliana). Finally, that all species that display heterosis contain a degree of allele-specific contribution to gene expression has been non-colinearity in their genome as high as that of maize. analyzed for a selected number of genes. These studies will be summarized in the following sections. Heterosis-associated gene expression in maize Recently, several studies have analyzed heterosis- Maize genomes of different inbred lines display associated gene expression in maize, rice and Arabidopsis significant aberrations from genetic colinearity by comparing expression patterns of selected genes in One of the basic notions of genetics is the concept of genetic inbred lines and hybrids [8,23,25], or by performing colinearity (see Glossary), which describes the observation high-throughput gene expression analyses via microarray that genomes of individuals in a given species contain the profiling or GeneCalling [7,26–31]. These studies analyzed same gene content. Recently, several studies observed sig- various aspects of plant development including very nificant aberrations from colinearity on the microcolinearity young embryos (six and 19 days after pollination) [8,27], www.sciencedirect.com Review TRENDS in Plant Science Vol.12 No.9 429

patterns were observed in these gene expression studies. Whereas in some studies nonadditive gene expression (Figure 3a) was prevalent between inbred lines and hybrids [23,25,26,28,30], additive gene expression (Figure 3a) was observed in other studies for most of the genes [7,8,27,31]. Finally, in one analysis, a similar number of genes followed additive and nonadditive expression [29] (Table 1). The discrepancy of these results might be the result of significant differences in genotypes, plant material, experimental designs and statistical procedures applied in the various studies. However, it might also be an indication that in different tissues or developmental stages different global expression patterns might prevail, which might neverthe- less be related to heterosis. This notion is supported by the observation that different tissues and organs within a hybrid plant display significant differences in their degree of heterosis [20]. Remarkably, it was also observed that triploid hybrids, in general, displayed higher expression levels compared to their diploid counterparts, although the transcript levels were nonadditive [25], thus underscor- ing the importance of genomic dosage for gene expression in hybrids. Another important finding in those studies that analyzed more than one inbred versus hybrid combination [28,29] was that no consensus gene set was found that was differentially expressed between all inbred/hybrid combi- nations. This might indicate that it was not possible to identify any key genes of heterosis in these studies and that global trends of gene expression might correlate with het- erosis. Although global trends of gene expression appeared to be different from tissue to tissue, it was observed in at least one study that the proportion of allelic additivity was positively associated with hybrid yield and heterosis for immature maize ears [29]. Moreover, most heterosis-related gene expression analyses have not revealed a predominant functional category to which differentially expressed genes belong. This might be because only a small fraction of the 25 500 Arabidopsis [32] or the estimated 59 000 maize or rice genes [33] were profiled (Table 1). Therefore, it is remarkable that an open-end differential screening of maize genes revealed that 25% of the genes differentially expressed between maize inbred lines and hybrids during Figure 2. Hemizygous complementation in maize hybrids. Different maize inbred early embryo development were associated with signal lines are characterized by a considerable loss of genetic colinearity, that is, the loss of particular genes. It is hypothesized that gene loss can be partly compensated in transduction [8]. hybrids by hemizygous complementation, which also leads to an improved Finally, more significant expression differences are performance of hybrids versus inbred lines. The colored boxes A to D represent found between parental inbred lines than between reci- different genes. Different members of gene families are indicated by the same colour (e.g. A1 and A2 belong to a gene family). The loss of a member of a gene procal hybrids. For example, whereas 4–18% of the genes family can be partly compensated for by the action of other members of this were differentially expressed between two inbred lines in gene family and leads only to minor phenotypic effects in inbred lines. When a various tissues, no differentially expressed genes were hybrid is selfed over several generations inbreeding depression is observed because not all genes present in the hybrid will be maintained. To date, partial loss identified in these tissues between the reciprocal hybrids of colinearity has only been observed in maize. In other species as well as in maize, [27]. This is not surprising given that reciprocal hybrids additional factors might also contribute to heterosis. are genetically identical and that gene expression differ- ences in reciprocal hybrids are mainly attributed to epi- endosperm (10, 14, 18 and 21 days after pollination) genetic effects. In the case of reciprocal hybrids selected for [23,26], 11 and 14 day-old seedlings [7,27], 21–23 day- cold germination and desiccation tolerance, only 1–2% of old shoot apical meristems [28], immature ear tissue the genes were differentially expressed [34]. In summary, [27,29], young and adult leaves [25,30], and rice panicles these gene expression profiling studies represent a first [31] in a wide variety of different genetic backgrounds step towards the definition of the complex gene expression (Table 1). Besides the differences in the tissues and networks that might be relevant in the context of heterosis. genotypes analyzed, these studies also differed signifi- However, there is currently no direct link between the cantly in their experimental design and their statistical classical genetic hypothesis and these gene expression procedures. In summary, no uniform global expression profiles. This might be attributed to the complex molecular www.sciencedirect.com 430 Review TRENDS in Plant Science Vol.12 No.9

Figure 3. Relative gene expression levels in hybrids and regulation of allele-specific gene expression in hybrids. (a) Gene expression levels in hybrids are not strictly related to the genetic concepts of dominance and overdominance. Therefore, a systematic terminology that was previously suggested [27] should be used to describe gene expression in hybrids A Â B (yellow and orange) compared to their parental inbred lines A and B (red). Gene expression in a hybrid gene can be manifested in an additive manner (orange) as the average expression of the parental inbred lines (no differential expression). Alternatively, gene expression can be significantly different from the midparent value either between the high and low parent value, or above the high or below the low parent value (yellow). (b) Effects of cis- and trans-regulation on the allelic contribution to gene expression in hybrids. Genes that are completely subject to cis-regulation reflect the relative expression levels of the parental inbred lines in the allelic ratio of gene expression in the hybrid. Genes that are exclusively regulated by trans-acting factors show equal expression of the two alleles in the hybrid. Genes that are subject to cis- and trans-regulation fall in between the two classes, that is, the relative allelic contribution to gene expression in hybrids for this class of genes neither displays the relative expression levels of the parental inbred lines nor an equal expression of both alleles. Genes are indicated by gray bars. The dark grey 5’ end of the genes represents the regulatory promoter and/or enhancer region, whereas the light-grey part corresponds to the transcribed region of the gene. Distinct trans-acting factors (e.g. transcription factors, repressors) are indicated as blue or yellow circles binding to the promoter and/or enhancer region. Cis-differences of genes (e.g. SNPs, insertions, deletions) are indicated as red bars in the grey gene structure. Size of green and orange arrows indicate expression levels of the allele. Green and orange areas in the histograms indicate allelic contribution to gene expression in the inbred lines and hybrids.

www.sciencedirect.com Review TRENDS in Plant Science Vol.12 No.9 431

Table 1. Heterosis related gene expression studiesa Plant organ Developmental Approach Genetic background Global expression Refs stage trend Maize Embryo 6 DAP 12K cDNA microarrays UH005 Additivity [8] SSH UH301 qRT-PCR Endosperm 10, 14, 21 DAP GeneCalling 7 Pioneer inbred Nonadditivity [26] lines Endosperm 18 DAP RT-PCR B73 Nonadditivity [23] BSSS53 Embryo 19 DAP 13.5K microarrays Mo17 Additivity [27] Seedling 11 DAG B73 Immature ear Seedling 14 DAG 14K cDNA microarrays Mo17 Additivity [7] qRT-PCR B73 Shoot apical meristem 21–23 DAP 12K cDNA microarrays UH002 Nonadditivity [28] qRT-PCR UH005 UH250 UH301 Adult leaves of di- and triploids Quantitative Northern Mo17 Nonadditivity [25] blotting B73 Immature ear GeneCalling 17 Pioneer inbred Additivity and [29] lines nonadditivity Arabidopsis First leaves 21, 24 DAG 6K cDNA microarrays Col Nonadditivity [30] Ler Cvi Rice Panicle Stage III, IV, V 9K cDNA microarrays Zhenshan97 Additivity [31] Minghui63 aAbbreviations: DAP, days after pollination; DAG, days after germination; SSH, suppressive subtractive hybridization. regulation that is required for the manifestation of even a expression, implying only minimal parent-of-origin effects. single phenotypical trait on the level of gene expression. Differential response of the two alleles in a hybrid to These regulatory networks could include genes that are up- environmental stress implied unequivalent functions of and downregulated in the hybrid relative to their parental the different alleles that might have an impact on hetero- inbred lines, although they are related to the manifestation sis. Finally, allelic expression of the ZmGrp3 gene (a of the same phenotypical trait. For example, if a gene that marker for root initiation) in six reciprocal hybrids was encodes for a repressor of a particular pathway is down- assayed but no significant deviation from a 1:1 expression regulated, other genes in that pathway could be de- ratio of the two alleles in the hybrids was detected [37].In repressed and thus be upregulated. conclusion, allele-specific gene expression was frequently observed in recent studies and might thus represent Allele-specific gene expression another regulatory mechanism in the context of the Global gene expression analyses via microarray technology phenotypic manifestation of heterosis. Interestingly, do not allow for the quantification of allele-specific contri- parent-of-origin effects apparently play only a minor role bution to gene expression. Allele-specific analyses can help for allele-specific gene expression whereas environmental to distinguish between cis- and trans-acting regulation of factors might have a more severe influence on allelic gene expression. A gene that is completely subjected to contribution to gene expression [27] and, thus, heterosis. trans-regulation is expected to provide a similar contri- bution of both alleles to gene expression in the hybrid Concluding remarks whereas genes subjected to cis-regulation will show Recently, several heterosis-related molecular studies have unequal expression of the two alleles in the hybrid [35] been performed. First, a considerable loss of gene colinearity (Figure 3b). In this context, allele-specific gene expression in the maize genome indicates that, at least in maize, these of 32 genes whose expression in the hybrid deviated from genomic differences might be a key to some aspects of the midparent value was analyzed via the quantitative heterosis. The sequencing of complete genomes of various SNP-based Sequenom technology [27]. For 18 genes cis- genotypes in different species [38–40] will help to better regulation was prevailing, whereas only one gene was understand the relevance of genome organization in relation classified as trans-regulated. The remaining 13 genes were to heterosis. Moreover, various gene expression studies classified as a combination of cis- and trans-regulation. In including allele-specific expression analyses revealed sig- another study of parental transcript accumulation via nificant differences in global gene expression patterns in allele specific RT-PCR, 11 of 15 maize genes displayed different tissues and developmental stages. Detailed expres- an unequal expression of the two alleles [36]. Maternal sion profiling experiments under standardized conditions or parental transmission had little effect on allele-specific of experimental design, microarray platforms and data www.sciencedirect.com 432 Review TRENDS in Plant Science Vol.12 No.9 analysis in a multitude of genotypes, developmental stages 15 Keeble, F. and Pellow, C. (1910) The mode of inheritance of stature and and organs will further refine the insight in genes that might time of flowering in peas (Pisum sativum). J. Genet. 1, 47–56 16 Bruce, A.B. (1910) The Mendelian theory of heredity and the be differentially expressed between inbred lines and hybrids augmentation of vigor. Science 32, 627–628 and, therefore, play a role during heterosis manifestation. 17 Jones, D.F. (1917) Dominance of linked factors as a means of These experiments might also be helpful for heterosis accounting for heterosis. Genetics 2, 466–475 predictions based on expression profiles. The functional 18 Powers, L. (1945) An expansion of Jones’s theory for the explanation of characterization of such differentially expressed genes will heterosis. Am. Nat. 78, 275–280 19 Semel, Y. et al. (2006) Overdominant quantitative trait loci for yield determine their relevance in the context of heterosis. and fitness in tomato. Proc. Natl. Acad. Sci. U. S. A. 103, 12981– Finally, future progress in the integration of genomic tools 12986 and the mapping and cloning of complex heterosis associ- 20 Melchinger, A.E. (1999) Genetic diversity and heterosis. In The ated QTLs via association mapping and map-based cloning Genetics and Exploitation of Heterosis in Crops (Coors, J.G. and might allow to map and clone simultaneously multiple key, Pandey, S., eds), pp. 99–118, American Society of Agronomy, Inc and Crop Science Society of America, Inc genetically-unrelated genes of heterosis. A major challenge 21 Lippman, Z.B. and Zamir, D. (2007) Heterosis: revisiting the magic. in this context is to monitor and quantify adequately the Trends Genet. 23, 60–66 complex heterotic phenotypes that are related to almost all 22 Fu, H. and Dooner, H.K. (2002) Intraspecific violation of genetic heterotic traits. In conclusion, the molecular data available colinearity and its implications in maize. Proc. Natl. Acad. Sci. U. S. A. 99, 9573–9578 thus far, do not indicate a simple correlation between one of 23 Song, R. and Messing, J. (2003) Gene expression of a gene family in the genetic hypotheses and the molecular events leading to maize based on nonlinear haplotypes. Proc. Natl. Acad. Sci. U. S. A. heterosis. Thus, the availability of novel genetic and geno- 100, 9055–9060 mic tools that allow for the integrated study of the complex 24 Brunner, S. et al. (2005) Evolution of DNA sequence nonhomologies interactions between genome organization and expression among maize inbreds. Plant Cell 17, 343–360 25 Auger, D.L. et al. (2005) Nonadditive gene expression in diploid and might contribute to a better understanding of heterosis. triploid hybrids of maize. Genetics 169, 389–397 26 Guo, M. et al. (2003) Genome-wide mRNA profiling reveals Acknowledgement heterochronic allelic variation and a new imprinted gene in hybrid This work was supported in part by the DFG framework program maize endosperm. Plant J. 36, 30–44 SPP1149 ’Heterosis in plants’ (award HO2249/6). 27 Stupar, R.M. and Springer, N.M. (2006) Cis-transcriptional variation in maize inbred lines B73 and Mo17 leads to additive expression References patterns in the F1 hybrid. Genetics 173, 2199–2210 1 Shull, G.F. (1952) Beginnings of the heterosis concept. In Heterosis 28 Uzarowska, A. et al. (2007) Comparative expression profiling in (Gowen, J.W., ed.), pp. 14–48, Iowa State College Press meristems of inbred-hybrid triplets of maize based on morphological 2 Falconer, D.S. and Mackay, T.F.C., eds (1996) Introduction to investigations of heterosis for plant height. Plant Mol. Biol. 63, 21–34 quantitative genetics (4th edn), Longman 29 Guo, M. et al. (2006) Genome-wide transcript analysis of maize hybrids: 3 Zirkle, C. (1952) Early ideas on inbreeding and crossbreeding. In allelic additive gene expression and yield heterosis. Theor. Appl. Genet. Heterosis (Gowen, J.W., ed.), pp. 1–13, Iowa State College Press 113, 831–845 4 Darwin, C.R., ed. (1876) The effects of cross- and self-fertilization in the 30 Vuylsteke, M. et al. (2005) Genetic analysis of variation in gene vegetable kingdom (1st edn), John Murray expression in Arabidopsis thaliana. Genetics 171, 1267–1275 5 Shull, G.F. (1908) The composition of a field of maize. Rep. Am. Breed. 31 Huang, Y. et al. (2006) Heterosis and polymorphisms of gene expression Assoc. 5, 51–59 in an elite rice hybrid as revealed by a microarray analysis of 9198 6 East, E.M. (1908) Inbreeding in corn. Conn. Agric. Exp. Sta. Rpt. 1907, unique ESTs. Plant Mol. Biol. 62, 579–591 419–428 32 The Arabidopsis Genome Initiative (2000) Analysis of the genome 7 Swanson-Wagner, R.A. et al. (2006) All possible modes of gene action sequence of the flowering plant Arabidopsis thaliana. Nature 408, are observed in a global comparison of gene expression in a maize F1 796–815 hybrid and its inbred parents. Proc. Natl. Acad. Sci. U. S. A. 103, 6805– 33 Messing, J. et al. (2004) Sequence composition and genome 6810 organization of maize. Proc. Natl. Acad. Sci. U. S. A. 101, 14349–14354 8 Meyer, S. et al. (2007) Heterosis associated gene expression in maize 34 Kollipara, K.P. et al. (2002) Expression profiling of reciprocal maize embryos 6 days after fertilization exhibits additive, dominant and hybrids divergent for cold germination and deciccation tolerance. Plant overdominant pattern. Plant Mol. Biol. 63, 381–391 Physiol. 129, 974–992 9 Hoecker, N. et al. (2006) Manifestation of heterosis during early maize 35 Wittkopp, P.J. et al. (2004) Evolutionary changes in cis and trans gene (Zea mays L.) root development. Theor. Appl. Genet. 112, 421–429 regulation. Nature 430, 85–88 10 Lamkey, K.R. and Edwards, J.W. (1998) Heterosis: Theory and 36 Guo, M. et al. (2004) Allelic Variation of Gene Expression in Maize estimation. In Proceedings 34th Illinois Corn Breeders’ School pp. Hybrids. Plant Cell 16, 1707–1716 62–77 Urbana, IL 2–3 March 1998 37 Woll, K. et al. (2006) ZmGrp3: Identification of a novel marker for root 11 Stuber, C.W. (1999) Biochemistry, molecular biology, and physiology of initiation in maize and development of a robust assay to quantify heterosis. In The Genetics and Exploitation of Heterosis in Crops allele-specific contribution to gene expression in hybrids. Theor. Appl. (Coors, J.G. and Pandey, S., eds), pp. 173–184, American Society of Genet. 113, 1305–1315 Agronomy, Inc and Crop Science Society of America, Inc 38 Martienssen, R.A. et al. (2004) Sequencing the maize genome. Curr. 12 Birchler, J.A. et al. (2003) In search of the molecular basis of heterosis. Opin. Plant Biol. 7, 102–107 Plant Cell 15, 2236–2239 39 Messing, J. and Dooner, H.K. (2006) Organization and variability of the 13 Birchler, J.A. et al. (2006) Unraveling the genetic basis of hybrid vigor. maize genome. Curr. Opin. Plant Biol. 9, 157–163 Proc. Natl. Acad. Sci. U. S. A. 103, 12957–12958 40 Rabinowicz, P.D. and Bennetzen, J.L. (2006) The maize genome as a 14 Davenport, C.B. (1908) Degeneration, albinism and inbreeding. model for efficient sequence analysis of large plant genomes. Curr. Science 28, 454–455 Opin. Plant Biol. 9, 149–156

www.sciencedirect.com