Miekevd.home.xs4all.nl

Selective Embryo Abortion Hypothesis Revisited ± G. Korbecka, P. G. L. Klinkhamer, and K. Vrieling Institute of Evolutionary and Ecological Sciences, Section Plant Ecology, Leiden, The Netherlands Received: October 9, 2001; Accepted: February 20, 2002 Abstract: Many plant species abort a large fraction of their em- otypes with a potential low quality later in life such that an in- bryos. It has often been suggested that embryos of genotypes crease in the level of abortion leads to an increase in offspring that would perform worse later in life are preferentially abort- quality. In evolutionary theory the hypothesis is important in ed. Such selective embryo abortion would lead to investment relation to explanations for low seed to ovule ratios and the of resources only in the offspring with the highest potential fit- optimal allocation of resources to male and female reproduc- ness. Many studies have shown that otherwise viable embryos are aborted. However, only few manipulative studies have in- deed shown a correlation between the level of abortion and The SEA hypothesis is based on the following premises: i) offspring quality and these studies have been challenged for plants abort a substantial part of the embryos, ii) otherwise their experimental design. Molecular techniques open new op- viable embryos are aborted, iii) the probability of being abort- portunities to study selective embryo abortion. Non-random ed depends on the genotype of the embryo. In contrast to abortion at the level of molecular markers can be observed as a ample theoretical attention given to the SEA hypothesis (e.g., deviation from Mendelian segregation: over- or under-repre- Kozøowski and Stearns, 1989[56]; Latta, 1995[61]; Burd, 1998[16]), sentation of markers in the offspring. Subsequently, the over- there are only a few experimental studies in which all condi- or under-represented markers can be related to offspring quali- tions, mentioned above, were considered. The SEA hypothesis ty later in life. We reviewed the literature on the genetic maps received a lot of attention in the 1980s (Stephenson, 1981[105]; of intraspecific crosses of wild plant species and the selection Wilson and Burley, 1983[122]; Casper, 1988[19]; Lee, 1988[63]; An- of cultivated species. The level of non-Mendelian segregation dersson, 1990[1]; Andersson, 1993[2]) but, probably due to diffi- we found in these maps is high. On average, 11.5% of the tested culties with the interpretation of the results of experiments in markers in the genetic maps of wild species and 14.6% in the which the abortion level was manipulated, empirical research cultivated ones, show a departure from Mendelian segregation.
on the topic has drastically reduced. Most of these problems From six studies, providing sufficient data, it was calculated can now be overcome using molecular techniques. In this pa- that in 68% of loci segregating in non-Mendelian fashion post- per we review the work on this hypothesis and discuss the fertilization selection is involved. We propose that the devia- possibilities and difficulties that are presented by the use of tion from Mendelian segregation can be partly explained by se- molecular methods to shed new light on the topic.
lective embryo abortion. We describe an experimental design that allows for attributing selective embryo abortion to the In crossings, SEA shows up as a deviation from Mendelian seg- non-Mendelian segregation that is found in a genetic map.
regation for some molecular markers. We will review data on the genetic maps of plants to judge whether there is a potential Key words: Selective embryo abortion, non-Mendelian segrega- for SEA to be detected by means of marker segregation analy- tion, distorted segregation, genetic maps, seed±ovule ratios, sis. However, SEA is not the only mechanism leading to devia- tion from Mendelian segregation. We will discuss how we may distinguish SEA from other mechanisms that include, e.g., meiotic drive, gametophytic selection, and seedling death.
Selective embryo abortion (SEA) is the phenomenon that some Two explanations have been proposed for the selective abor- genotypes are aborted more frequently than others. The Selec- tion of particular genotypes: 1. Maternal control, 2. Embryo tive Embryo Abortion hypothesis proposes that the fitness of a female plant can be increased by the selective abortion of gen- Since, potentially, SEA may be maternally controlled, it has of- ten been discussed as an aspect of female choice together with pre-fertilization processes, such as selective inhibition of pol-  Georg Thieme Verlag Stuttgart ´ New York len germination and pollen tube growth (Wilson and Burley, 1983[122]; Marshall and Folsom, 1991[73]). Theoretically, in an- Selective Embryo Abortion Hypothesis Revisited ± A Molecular Approach giosperms maternal control can be through the endosperm, onic viability is often assessed in relation to early-acting in- the tissue that nourishes the embryos, because it contains breeding depression. On the basis of information on seed pro- two copies of the maternal and one copy of the paternal genes, duction after selfing and outcrossing, the number of so-called also abortion may be related to an interaction of the maternal lethal equivalents is estimated. A lethal equivalent is a lethal and paternal genome. However, Marshall and Folsom (1991[73]) gene or a number of deleterious genes that make up one lethal concluded in their review on mate choice in plants that there is gene. According to Lynch and Walsh (1998[69]), the number of little evidence to prove that specific maternal mechanisms lethal equivalents per gamete affecting early embryonic sur- produce sorting among compatible donors. The problem with vival varies approximately from 1.7 to 5.0 for conifers and from the assumption of maternal control is how the link with off- 0.4 to 0.91 for short-lived angiosperms. However, for consider- spring quality later in life is brought about, in other words: ing SEA it is essential to make a distinction between lethal and how can the mother plant ªknowº which embryos will give deleterious alleles. If abortion results from recessive lethal al- the highest fitness contribution? Moreover, it is technically leles, both the level of abortion and the direction of selection extremely difficult to experimentally test this hypothesis. If are fixed. The embryos, which possess lethal alleles, will die changes in the maternal tissue of the seed (nucellus and in- irrespective of the conditions they encounter during develop- teguments) precede changes in embryo and endosperm devel- ment. In such cases it is not likely that the level of inbreeding opment, this could point to maternal control. If the order of depression during seed set will be correlated with the level of changes is the reverse, this would point to embryo competi- inbreeding depression later in life. This may explain the ab- tion (Marshall and Folsom, 1991[73] and refs. therein). As yet sence of such a correlation in the studies of Husband and there is very little evidence to decide for either of the two pos- Schemske (1995[40]) or Koelewijn et al. (1999[54]). On the other hand, if the alleles on which embryo abortions depend are de- leterious, embryo abortion may be selective and depend on the More generally accepted is the idea that SEA is brought about conditions the embryo encounters. Remington and OMalley through competition among embryos. Some embryos may be (2000[91]) studied early acting inbreeding depression in lob- better competitors for resources than others, either because lolly pine (Pinus taeda) using information from a genetic map.
they present a larger sink or they may even release chemical They estimated that in this species 19 loci have moderately substances which are most probably indole compounds that deleterious or lethal embryonic effects. Moreover, most of inhibit the sucrose uptake of siblings (Mohan Raju et al., the alleles reducing viability are recessive and for 3 loci over- 1996[80]; Krishnamurthy et al., 1997[57]; Arathi et al., 1999[3]).
dominance was found. There is also another study (Melser et In this scenario, maternal ªrecognitionº of the embryos is not al., 1997[75]) suggesting that embryo abortion may not be a necessary. The mother plant can influence offspring quality in- result of action of recessive deleterious alleles. On the basis of directly by controlling the level of resources and thus setting comparing seed production after selfing and outcrossing in the selective arena for the embryos. It has even been suggested E. vulgare (after sufficient amount of pollen was applied), that endosperm reduces embryo competition since it is more they found that some individual plants aborted more selfed frequently observed in the species with multiovulated ovaries embryos and others more outcrossed ones. Melser et al.
compared to those with uniovulated ones, and it is found more (1997[75]) concluded that in E. vulgare the effects of the deleter- often in the species with multiovulated species that experi- ence less abortion (Uma Shaanker et al., 1996[111]).
One can imagine that embryo abortion may be influenced by a However, from an evolutionary ecological perspective, the number of (mildly) deleterious alleles that each by themselves mechanism leading to SEA is not as interesting as the fact have only a small effect and are therefore not easily purged whether or not it can increase offspring quality. The increase of offspring quality through SEA would mean that embryo abortion is potentially adaptive. The correlation between the In the remaining part of the paper we will first present the abortion level and offspring quality is, under the assumption more traditional phenotypic approach to study SEA and then of embryo competition, brought about by genes that control we will discuss how molecular techniques can be used to e.g., basic metabolic processes that are important, both during embryo development and during later life, or by genes that have pleiotropic effects. Goldberg et al. (1989[34]) summarize in their review: ªMore than 90 % of the 15000 diverse mRNAs present in mid-maturation stage embryos are represented in both cotyledon stage and fully differentiated, mature embryos.
Most of these mRNAs are also present in post-germination Flowering plants commonly produce more ovules than seeds.
cotyledons and in the mature plant leaf.º The fact that most In many angiosperm species ovules may not develop into of the genes that are expressed in embryonic stage are also seeds due to pollen limitation (Wolfe, 1983[123]; Zimmerman expressed later in life, gives ample opportunity for embryo and Pyke, 1988[126]; and see Burd, 1994[15] for a review) or be- abortion to have an effect on offspring quality later in life.
cause they are involved in self-incompatibility mechanisms (Waser and Price, 1991[118]; Seavey and Carter, 1996[98]). How- Most evidence for the fact that abortion depends on the geno- ever, even after successful fertilization, a considerable propor- type of the embryo comes from studies on inbreeding depres- tion of the ovules fail to produce seed in many species. Wiens sion. If selfed embryos have a higher chance of being aborted (1984[119]) estimated that seed±ovule ratio equals, on average, than outcrossed ones (Montalvo, 1992[81]; Gibbs and Sassaki, about 85% for annuals and 50 % for perennials. Wiens data are 1998[33]), this suggests that recessive deleterious or lethal al- based on developing fruits. If ovules in undeveloped fruits are leles may influence competitive strength of embryos. Embry- also included, the seed ovule ratios may be even lower. Dissec- G. Korbecka, P. G. L. Klinkhamer, and K. Vrieling tions of ovaries showed that a large fraction of embryos are produced more leaves, and later in life more inflorescences, aborted, for example in Prunus cerasus (Bradbury, 1929[12]), Ox- flowers and matured more seeds compared to the treatment alis magnifica (Guth and Weller, 1986[36]) and Epilobium angu- after random destruction of ovules (Stephenson and Winsor, stifolium (Wiens et al., 1987[120]). Some species show extremely 1986[106]). In a study on patterns of seed abortion in P. cocci- high abortion rates. In Dedeckera eurekensis, the seed±ovule neus, Rocha and Stephenson (1991[96]) found that ovules at ratio equals 2.5%, although about 90 % of the ovaries initiate the basal end of the ovary are more likely to abort, due to the growth, indicating that fertilization took place (Wiens et al., fact that they lag behind in development because they were 1989[121]). In Asclepias speciosa approximately only 3.8% of the pollinated later, which may result in reduced nutrient avail- ovaries develop into mature fruits, although 82.4% of them ability. Destroying the ovules on the stylar end increased the were fertilized (Bookman, 1984[10]).
probability of seed maturation on the basal end. The progeny that resulted from this treatment was significantly less suc- Gymnosperms also abort many seeds. In Pinus sylvestris on cessful compared to the control treatment with regard to ger- average 30% seeds are aborted (Karkkainen et al., 1999[49]).
mination time, vegetative growth, flowering time and number The level of embryo abortion is, however, higher due to poly- embryony. The most common form is simple polyembryony with independent fertilizations of more than one archegonium Is abortion dependent on embryo genotype? within the same ovule of which usually only one develops into a seed (Sorensen, 1982[103]; Willson and Burley, 1983[122]; Haig, Stephenson (1981[105]) and Lee (1988[63]) show in their reviews that in many plant species the chance for an embryo of being aborted depends on factors like time of initiation, position In some cases it has been argued that embryos are not viable within ovary, resource availability and pollen source. Even if because of high genetic load, as in the D. eurekensis example the level of embryo abortion is high and viable embryos are (Wiens et al., 1989[121]). However, as we will discuss later, even aborted, abortions do not necessarily depend on the genotype viable embryos are often aborted at a very high rate.
of the embryo and may not lead to selection. Both single pollen donor and mixed pollen donor experiments have been used to Are potentially viable embryos aborted and study the relationship between abortion rate and genotype.
does abortion increase offspring quality? The easiest way to detect selection is when each flower of a In some species with linearly arranged ovules, developing em- plant receives pollen from a single pollen donor only. One can bryos in the basal end of the ovary are more likely to abort.
then compare the siring success of different potential fathers Nakamura (1988[83]) described a successful in vitro culture of directly, by counting the seeds in the flowers, without the use embryos from the basal end in Phaseolus vulgaris. In Dalberia of genetic markers. With this approach, it is unlikely that pol- sisso, Ganeshaiah and Uma Shaanker (1988[31]) cut off two dis- len tube competition influences differences among fathers.
tal seeds and implanted the remaining pod in agar to complete The disadvantage of single pollen donor experiments is that maturation of the rest of the seeds. This treatment resulted in competition among the embryos within a flower cannot be de- an abortion rate in the basal end of the pod as low as in the dis- tal end of intact, control pods. Both Nakamura (1988[83]) and Ganeshaiah and Uma Shaanker (1988[31]) did not relate abor- Bertin (1982[8]) studied the self-incompatible trumpet creeper (Campsis radicans) and applied pollen of different fathers and found that the pollen donors that were favoured by particular To our knowledge, only four papers present evidence that recipients were usually those whose pollinations resulted in abortion can increase offspring quality. These experiments fruit with many and large seeds. Although prezygotic mecha- compared offspring quality after natural abortion and after nims were not all properly excluded, the author concluded that random thinning of the ovaries. In contrast to the first, the lat- fruit abortion seems to have been more important in donor ter is not selective. Species of the family of the Boraginaceae al- selectivity than prezygotic phenomena.
ways produce four ovules in each flower arranged in equal po- sitions, in a square. Although pollen is not limiting seed pro- Most single pollen donor experiments aim at comparing abor- duction, on average there are fewer than four seeds per flower tion after self- and outcross pollination, or comparing outcross found in many species of the Boraginaceae (e.g., Cynoglossum pollination with close and distant donors. Such comparisons officinale ± Jong and Klinkhamer, 1989[48]; Echium vulgare ± are interesting, especially because Husband and Schemske Klinkhamer et al., 1994[53] and Cryptantha flava (Casper, (1996[41]) showed that embryo development is one of the most 1988[19]). In Cr. flava (Casper, 1988[19]) and in Cy. officinale (Mel- important life stages in which inbreeding depression can act.
ser et al., 2001[76]) the random destruction of three ovules in a flower resulted in doubling of the chance of maturation for the For the self-compatible Aquilegia caerulea Montalvo (1992[81]) remaining ovule, compared to a control treatment with all found that the abortion rates for selfing were, on average, 38% ovules intact. This shows that in the control treatment a large higher compared to abortion for outcrossing, while there were fraction of the aborted embryos was potentially viable. In Cr.
no significant differences in fertilization rate for both pollina- flava seeds from the control group with natural abortion tion types. For E. vulgare, Melser et al. (1997[75]) found that in showed higher emergence and survival during two years of some individuals self-pollen was relatively more successful growth (Casper, 1988[19]). Melser and Klinkhamer (2001[76]) compared to outcrossed pollen, while in others the outcrossed found that natural abortion resulted in higher offspring sur- pollen was more successful. Pollen donors did not differ in pol- vival in Cy. officinale. In Lotus corniculatus, offspring produced len viability, pollen germination and pollen tube growth.
after natural embryo abortion showed better germination, Therefore, Melser et al. (1997[75]) concluded that differences Selective Embryo Abortion Hypothesis Revisited ± A Molecular Approach in siring success of different pollen donors were most likely Crushing ovules does not always reduce abortion levels. One can explain Caspers (1988[19]) and Melser and Klinkhamers (2001[76]) results by assuming that the resources not used by Gibbs and Sassaki (1998[33]) found for Dalbergia miscolobium in the destroyed ovules are allocated to the remaining ovules the field that 30.0% of crossed flowers and 3.6% of selfed flow- within the same flower, thereby increasing the chance for mat- ers developed mature fruits. This difference was mainly caused uration. If the experimental treatment is applied to only a part by abortion of selfed embryos because, in the ovules dissected of the flowers, it is possible that resources that would be used 4±6 days after pollination, embryos were found in similar fre- by crushed ovules are divided among all ovules of the plant quency and condition for both treatments.
and not only among those that remained in the hand-thinned flowers. In such a case, the difference among the treatments Marshall and Whittaker (1989[74]) studied effects of identity of would be small and could go undetected. Perhaps this may at a pollen donor on offspring quality in Raphanus sativus. They least partly explain the negative results found in two studies found significant paternal effect on the number of leaves and on Anchusa officinalis (Andersson, 1990[1]) and Achillea ptarmi- weight of offspring after eight weeks of growing in a green- ca (Andersson, 1993[2]). The difficulties in the interpretation of house. The effects of pollen donor were more pronounced if results from the experiments discussed above can be avoided maternal plants were grown in water stress conditions. Their if a single treatment is applied to a whole plant and the same results suggest that the processes that sort among potential genotypes are used in different treatments (Melser et al., fathers during pollination, fertilization and seed filling may Decreased offspring quality after random crushing of ovules may Multiple donor experiments, where a mixture of pollen from be an artefact. In experiments based on ovule destruction, in- different genotypes is applied to a single flower, combined ferior offspring do not necessarily result from genetic differ- with paternity analysis, can also provide information about ences but may be caused by subtle effects of the mechanical SEA. The advantage of multiple donor experiments is that se- damage itself. Casper (1988[19]) cautions: ªPrematurely remov- lection among pollen donors within flowers can be detected.
ing some reproductive structures might upset initial source± The disadvantage is that, if it is not possible to analyse aborted sink relationships and thus plant±resource levels, adversely af- embryos for their paternity, an appropriate method has to be fecting seed quality. In addition, forcing a flower to distribute found to separate the effects of pollen tube competition from resources to an ovule that it normally would not mature might SEA. Marshall and Ellstrand (1988[72]) carried out a multiple itself result in an inferior seed.º Moreover, developmental irre- donor experiment on Raphanus sativus under stress condi- gularities of the flower can influence the competitive strength tions. Early water stress can affect both fertilization and early seed abortion. In contrast, late water stress can only influence seed abortion. The contribution to the progeny of the three An experiment, as described above, is therefore not sufficient pollen donors differed from the control in the late stress treat- to prove that SEA occurs. The best way to show that SEA can ment but not after early stress. Apparently, only late abortions increase offspring quality in ovule destruction experiments is provide the opportunity to select in this case.
to collect genetic evidence as well. We will therefore discuss in the remaining of the paper how molecular data can be used Attributing the abortion rate to the origin of pollen in some to overcome the problems caused by the traditional approach gymnosperms is even easier since they have poorly developed prezygotic selection mechanisms (Willson and Burley, 1983[122]). For example, Karkkainen et al. (1999[49]) determined the abortion rate in Pinus sylvestris as a proportion of empty seeds, because seed coat formation in this species is an effect If embryo abortion is selective, certain alleles will be under- of pollination. They found that frequency of abortion increases or over-represented in the offspring, compared to Mendelian with the proportion of self pollen applied to the flowers. The segregation. The upswing in molecular methods in the last proportion of empty seeds ranges from 23% after outcrossing decade has led to easy access to abundant molecular markers in almost every organism (e.g., AFLP). Such molecular markers might be a powerful tool to detect and assess the adaptive value of SEA. Using molecular markers avoids the limitation of pollination experiments because selection among offspring Missing information about the selection among genetically dif- of a single pollen donor can be detected. Even if the plant is ferent offspring sired by the same father. Pollination experi- self-pollinated, selection among embryos may be observed in ments can show that selective abortion exists only if siring the loci for which the parent plant was heterozygous. So far, success of different fathers is compared, either after single selection among the offspring of a single father has largely donor pollination or after mixed donor pollination combined been ignored. This may have caused an under-estimation of paternity analysis. Moreover, it is necessary to eliminate that prezygotic mechanisms that may play a role. The big disadvan- tage of this approach is that a part of post-fertilization selec- The second advantage of using molecular markers to test the tion, which may occur among genetically different offspring SEA hypothesis is that the presence or absence of alleles that of the same father, cannot be observed. Only molecular tech- are under- or over-represented in the offspring, compared to Mendelian segregation, can be related to offspring perform- ance in later life. This would be a much better way of assessing the selective advantage of embryo abortion compared to tradi- G. Korbecka, P. G. L. Klinkhamer, and K. Vrieling tional methods, because no manipulations of flowers or plants artificial selection and inbreeding were minimal. We expected (e.g., destroying ovules) are needed and because selection can that in genetic maps of cultivated species non-Mendelian seg- be directly linked to the genotype of the offspring.
regation is found more often because mapping populations are often derived from crosses between different inbred line vari- Selection among embryos can be presented at the level of DNA eties or come from distinct geographical areas (e.g., Loarce et as a deficiency or excess of certain genotypes among the off- al., 1996[68]; Jenczewski et al., 1997[43]; Liu et al., 1997[67]; Qi et spring that successfully went through seed maturation, com- al., 1998[89]). It may happen that genes from one inbred line/ pared to expected Mendelian segregation. We reviewed genet- variety do not function properly when combined with genes ic maps of plants in order to determine the potential for SEA. If the percentage of molecular markers showing non-Mendelian segregation found in genetic maps of plants is as low as ex- Results of the literature survey are presented in Tables 1 and 2.
pected due to chance alone, we have to conclude that SEA is The percentage of markers showing non-Mendelian segrega- not an important process. This argument, however, cannot be tion differs significantly from 5% for the 59 analysed species reversed. If many markers show non-Mendelian segregation, (t = 9.143; df = 58; p < 0.001). It ranges from 0.1±40.82 % (aver- that could be due to SEA but other selective mechanisms can- age: 14.6) for cultivars (Table 1) and from 0±41.0% (average: not be excluded. For instance, meiotic drive and gametophytic selection can also lead to non-Mendelian segregation (Appen- dix). The difficulty in distinguishing the cause of non-Mende- The difference in the average percentage of markers showing lian segregation is a disadvantage of this method. An appropri- non-Mendelian segregation between cultivated and wild spe- ate experimental design should be used to study segregation cies is not significant (F = 1.099; df = 1,57; p = 0.299).
in plants, with different treatments leading to differences in the level of abortion, as will be discussed later.
Distinguishing between biological phenomena Is non-Mendelian segregation common in plants? Sometimes it is argued that sampling error or irreproducibility Data about non-Mendelian segregation in plants can be found of the techniques can be responsible for a high percentage of in genetic maps. In almost all genetic maps of plants we have molecular markers showing non-Mendelian segregation. Here reviewed, authors refer to a statistically significant departure we will consider the importance of those problems.
from Mendelian segregation as distorted segregation, although they usually do not present any evidence for the presence of Inconsistent PCR amplification can cause irreproducibility of segregation distorter genes sensu Lyttle (1991[70]). Lyttle de- the method and hence a detection of apparently higher non- fines segregation distorters as genetic elements that exhibit Mendelian segregation. RAPD is known as a technique that meiotic drive. That is why, when we consider a statistically sig- does not always give fully reproducible results (Jones et al., nificant departure from Mendelian segregation, we will use 1997[45]). We therefore compared the level of non-Mendelian the more neutral term: non-Mendelian segregation.
segregation detected in genetic maps using three techniques: AFLP, RAPD and RFLP. None of the techniques gave significantly It is common practice to test, by means of a chi square test at a higher level of non-Mendelian segregation (paired samples 5% significance level, whether or not segregation of a certain test results for: RAPD vs. RFLP: df = 10, p = 0.433; RFLP vs. AFLP: marker deviates from the expected ratio. If all markers are in- df = 4, p = 0.222; RAPD vs. AFLP: df = 4, p = 0.386), although herited independently, 5% of all markers should show non- PAGE gels used in AFLP give much higher resolution than agar- Mendelian segregation, if no selection occurs. However, it is ose gels used commonly in RAPDs. Note, however, that the extremely difficult to determine the expected fraction of mar- tests are based on a small number of comparisons. The con- kers showing non-Mendelian segregation under the null hy- stant warning (e.g., Jones et al., 1997[45]) that RAPDs are not pothesis, that no selection occurs. Firstly, non-Mendelian seg- fully reproducible may have caused a severe selection against regation can be over- or under-estimated when judged from markers giving non-Mendelian segregation previously used in the number of loci with a significant non-Mendelian segrega- mapping. Many authors using RAPD markers for the construc- tion because, in a distorted region of the genetic map, the den- tion of a genetic map, only include markers which are effi- sity of mapped molecular markers may differ from the aver- ciently amplified and exhibit unambiguous polymorphism.
age. Secondly, an unknown percentage of DNA markers is lo- Jenczewski et al. (1997[43]) write: ªWhen such precautions are cated in non-functional regions (e.g., not or loosely linked to taken, RAPD does not induce higher levels of distortion than functional regions). For such markers, only non-Mendelian restriction fragment length polymorphisim (RFLP).º Discard- segregation due to chance is expected. Nevertheless, Tables 1 ing markers before use in mapping, although to a smaller ex- and 2 provide useful information because, averaged over all tent, may have happened in these other techniques as well.
species, the first problem should disappear as we have no rea- Tables 1 and 2 may therefore underestimate the level of non- son to assume that the density of molecular markers is higher or lower in the region where selection occurs. The second problem can only lead to an under-estimation of selection. Un- Other sources of artefacts can be homoplasy, which is the am- fortunately, we do not know the quantitative importance of plification of two fragments of the same length from non-alle- lic regions, low resolution of agarose-gels, and co-migrating and overlapping polymorphic fragments. However, we expect We searched for genetic maps based on intraspecific crosses of these explanations to have only a minor influence on the level cultivated and wild species. Wild species were defined in the of non-Mendelian segregation. Rieseberg (1996[95]) tested the broadest sense possible. The basic criterion we used was that homology of 220 RAPD co-migrating fragments in three close- Selective Embryo Abortion Hypothesis Revisited ± A Molecular Approach Table 1 Percentage of non-Mendelian segregation found in genetic maps of cultivated species$ No. of loci showing nMS for each type of marker $ We have searched for the genetic map of cultivated species using the keywords: nMS: non-Mendelian segregation, BC: backcross, RIL: recombinant inbred lines, BIL: ªgenetic mapº or ªlinkage mapº and ªplantºin the journal ªTheoretical and Applied Geneticsº from volume 93 (year 1996) till volume 97 (year 1998). We used Win- a % of markers showing nMS only in the map (the number of loci that showed nMS spirs 2.0 to search in the Current Contents database. The search resulted in 222 and that are not linked in the genetic map could not be retrived from the article), records. Data from 33 out of 222 papers could be included in the table. A paper b only markers showing nMS given for the probability level p < 0.01 were presented, was included if the number of loci with significant non-Mendelian segregation for c the type of markers that showed nMS could not be indentified in a paper, the genetic map could be calculated. Partial genetic maps and maps based on d two mapping populations combined together, doubled haploids or intraspecific cross were not included.
# this category of markers may have included the following markers: isoenzymes, Many doubled haploid lines are derived from the pollen of one parent. Analysing minisatellites, microsatellites, IGS, SCAR, CAPS, PCR markers, rDNA, STS, morpho- these lines yields the segregation directly without the necessity of crossing. After pollen germination and regeneration of haploid plants, chromosome doubling oc- * indication of a genetic map that could be used for analysis of distribution of loci curs spontaneously or it is induced chemically (by colchicine). The plants grow and with non-Mendelian segregation along linkage groups. This analysis is described then the material is sampled for DNA analysis. During the germination of pollen in a chapter: ªdistinguishing between biological phenomena and technical prob- and while the plants are growing in in vitro culture, selective mortality may occur.
This mortality might explain the relatively high levels of non-Mendelian segregation found in the doubled haploid offspring (Xu et al., 1997[125]). Since the offspring did not develop from embryos, the genetic maps based on double haploid populations were not reviewed in this study. We did not include maps derived from interspe- cific and intergeneric crosses because, in such wide crosses, chromosome pairing and other phenomena ± that are not related to selective embryo abortion ± may G. Korbecka, P. G. L. Klinkhamer, and K. Vrieling Table 2 Percentage of non-Mendelian segregation found in genetic maps of wild species$$ No. of loci showing nMS for each type of marker Abbreviations as in Table 1, M.: megagametophytes; search resulted in 1275 records. We scanned the abstracts of all articles to find ge- $$ We searched for genetic maps of wild species by means of Winspirs 2.0 in the Cur- netic maps of intraspecific crosses of wild plant species. The criteria of incuding rent Contents database until August 2000. We used the same keywords as in the the data from a genetic map into this review were the same as for cultivated spe- search for genetic maps of cultivated species, but in all available journals. The ly related species of sunflowers and found that 91% of frag- A strong argument for the fact that the non-Mendelian segre- ments are homologous. This means that artefacts like homo- gation is not found by chance or sampling error is the repeat- plasy and wrong scoring due to low resolution of agarose gels ability of finding skewed markers in the same species in many may be responsible for only 9% of co-migrating fragments.
crosses, with a different set of parents, or in the same chromo- However, this number would be much lower if individuals somal regions. Xu et al. (1997[125]) mapped chromosomes of from the same mapping population derived from an intraspe- rice using many types of crosses: inter-subspecific crosses, doubled haploid and recombinant inbred lines. They detected a non-Mendelian segregation by means of RFLP in all types of Moreover, artefacts mentioned above cannot explain why crosses, ranging from 17% for one of the inter-subspecific non-Mendelian segregation often occurs in clustered loci. We crosses, to 70% for one of the doubled haploid populations.
screened the reviewed genetic maps for the distribution of 227 distorted markers were clustered in 17 chromosomal re- markers showing non-Mendelian segregation. On the basis of gions, and nine of these regions were associated with segre- the data from 30 maps (these maps are indicated in Tables 1 gation distortion in more that one population. Repeatability and 2) we found that 56% of 633 loci segregating in non-Men- of non-Mendelian segregation in the same region of genetic delian fashion formed clusters of two or more markers.
Selective Embryo Abortion Hypothesis Revisited ± A Molecular Approach The difficulty with isolation of embryos and the very small amount of material may limit the feasibility of this method.
PCR-based techniques, like microsatellites, can be a better al- ternative to allozyme analysis since they require much a smal- ler amount of plant material. Hufford et al. (2000[39]) have shown that aborting embryos of Platypodium elegans can be successfully genotyped by means of microsatellites. Reusch (2000[92]) also used microsatellites to genotype developing embryos in Zostera marina. However, isolation of embryos at the stage when they are large enough for analysis makes it im- possible to investigate effects of very early stages of abortion.
When one cannot analyse the aborting embryos for their pa- ternity, it is rather difficult to judge what was the cause of ob- served non-Mendelian segregation that is already detected in a map. An attempt to separate different causes has been made by Pham et al. (1990[85]), who determined whether selection before or after fertilization took place on the basis of segrega- tion analysis of isozyme loci in an F2 generation in several crosses of rice. They used successive c2 tests for 18 loci in which non-Mendelian segregation was found. Firstly, the equi- frequency of alleles (p, q) was tested. Secondly, a c2 test was made to test if the distribution of genotype frequencies fits p2:2pq:q2 (based on the observed allele frequencies p and q) (see Fig.1). Since, for most of the tested skewed loci, the fre- quency of alleles was not equal and genotype frequencies fit- ted the p2:2pq:q2 distribution, Pham et al. (1990[85]) conclud- ed that prefertilization (gametophytic) selection was respon- sible for non-Mendelian segregation. However, for other loci, Fig.1 Key for determining whether gametophytic or post-fertilisa- evidence for post-fertilization selection was found. Guiderdoni tion selection takes place in a segregating F2 population analysed (1991[35]) reported similar results for crosses between different with co-dominant markers (based on Pham, 1990[85]).
We performed such an analysis for the reviewed genetic maps maps derived from different crosses of the same species was that reported segregation data for the F2 generation for co- also reported by Price and Tomos (1997[88]).
dominant markers. We found segregation data for such mar- kers only in 6 genetic maps (Mukai et al., 1995[82]; Baudracco- How to distinguish between SEA and other biological Arnas and Pitrat, 1996[5]; Katzir et al., 1996[50]; Chen et al., explanations for non-Mendelian segregation? 1998[20]; Korzun et al., 1998[55]; Vanhala, unpublished data).
Chi square analysis of 56 loci revealed that, in 31 loci, post-fer- There is a long list of biological explanation for non-Mendelian tilization selection affecting heterozygotes took place, while in segregation in plants, it includes, apart from SEA, meiotic 7 loci, zygotic or a combination of selection before and after drive, gametophytic selection, selective germination and seed- fertilization occurred. In 15 loci gametophytic selection was ling death, B chromosomes, cytoplasmic inheritance, endo- detected. In 3 loci the stage of selection could not be deter- phytes and chromosomal rearrangements (Appendix). There- mined by means of subsequent c2 tests. This analysis shows fore, it is very difficult to separate SEA and other causes of that post-fertilization selection occurs in the majority of cases non-Mendelian segregation. Below we will present three ways (68%) involved in non-Mendelian segregation.
Another method to track down the cause of non-Mendelian segregation is the analysis of segregation of molecular markers in reciprocal crosses. Korzun et al. (1998[55]) performed such Direct evidence for SEA may be presented by molecular marker crosses in rye (Secale cereale) and found, in one cross, 7 loci segregation analysis of aborting embryos in comparison to showing non-Mendelian segregation, while in the other cross mature seeds. Rigney (1995[94]) performed a successful pater- such a skewed segregation was found for 9 loci. Only 2 of those nity analysis for aborting embryos by means of the MDH allo- loci were common for both crosses and they are potential loci zyme marker. Rigney (1995[94]) removed embryos that were in which post-fertilization selection could take place. Asym- being aborted from a plant and analysed their paternity in metry of segregation data in reciprocal crosses could be caused Erythronium grandiflorum. Selfed embryos were more likely by post-fertilization selection due to an interaction between to abort than outcrossed ones. Moreover, the progeny fertilized nuclear and cytoplasmic genes which is different, depending by nearby donors are aborted more often than those sired by on which plant is used as a female in a cross, or gametophytic selection affecting either male or female function of one of the G. Korbecka, P. G. L. Klinkhamer, and K. Vrieling parents. The distinction between pre- and post-fertilization increases offspring quality is still weak. The molecular geno- selection in the latter case could be made if reciprocal back- typic approach links SEA to the presence and absence of par- crosses to both parents are performed. Faris et al. (1998[28]) ticular alleles, which is why offspring quality can be related to compared non-Mendelian segregation in 4 such crosses in specific alleles and, therefore, manipulations, such as crushing Aegilops tauschii. They attributed nearly the whole observed non-Mendelian segregation on chromosome 5D to gameto- phytic selection affecting male function, however, they could The level of non-Mendelian segregation found in the published not exclude nuclear±cytoplasmic interaction in one region on genetic maps suggests that there is ample opportunity to de- tect SEA. An appropriate experimental design, which we pro- pose in this paper, would not only detect SEA and provide the Possible experimental design to test SEA hypothesis possibility to relate this to offspring quality, but it would also provide more information about the genetic mechanisms con- Attributing non-Mendelian segregation, in crosses used to make genetic maps, to one of the explanations given in the ap- pendixcan be done in some cases if an experiment is carefully planned or additional research is conducted. Some of the ex- planations (other than SEA) for non-Mendelian segregation We thank Eddy van der Meijden, Tom de Jong, Russell Lande, can be ruled out by additional studies such as: chromosome Brenda Casper, Leo Beukeboom, Jan Kozlowski for their com- counting and observation of pollen germination (Appendix).
ments on the earlier versions of the manuscript. The work The best way to separate the effect of SEA from other explana- was supported by Life Sciences Foundation (SLW), which is tions listed in appendixwould be to find an experimental subsidized by the Netherlands Organisation for Scientific Re- treatment with which the level of abortion is manipulated without influencing other processes. Nutrient stress would be a good candidate: it is known to influence abortion levels while there are no reports that it influences e.g., meiotic drive.
If the deviation from Mendelian segregation for certain mo- Biological explanations for non-Mendelian segregation in lecular markers is positively correlated with the level of em- genetic maps of plants. Meiotic drive. Lyttle (1991[70]) de- bryo abortion this would imply that, indeed, embryo abortion fines meiotic drive as ªmechanics of meiotic division that is selective. Using the same genotype (clone) in all treatments cause one member of a pair of heterozygous alleles or hetero- could further strengthen the argument because meiotic drive, morphic chromosomes to be transmitted to progeny in excess B chromosomes, cytoplasmic inheritance and chromosomal of the expected Mendelian proportion of 50%º. A number of rearrangements can be excluded, since the same nuclear genes meiotic drive systems are described in detail for animals. How- (chromosomes) are in the same cytoplasmic environments in ever, little is known about meiotic drive in plants. In most all nutrient treatments. If clones are grown in the same envi- flowering plants, megasporogenesis may lead to meiotic drive ronment, they could possibly also contain the same endo- because of an obvious asymmetry of meiotic division: only one of the four haploid cells develops into a functional egg and this cell may contain preferentially transmitted alleles or chromo- An alternative experiment would be to compare the segrega- somes. An example of such a process is the preferential trans- tion of molecular markers in the offspring coming from two mission of chromosomal knobs (large clusters of repetitive treatments performed on self-pollinated clonal replicates of DNA) on chromosome 10 into viable megaspores in maize one genotype of a plant. The first treatment would have the (Buckler et al., 1999[13]). Another example of meiotic drive, al- ovules randomly crushed, while the control treatment would though of interspecific origin, is the preferential transmission not be manipulated. If selective abortion is playing a role, then of alien chromosomes. Such chromosomes, common in Triti- it is expected that more loci in the offspring of control plants ceae, are called ªcuckooº chromosomes (Miller, 1983[78]). Finch would show non-Mendelian segregation.
et al. (1984[29]) described the effects of one chromosome com- ing from Aegilops sharonensis in wheat (Triticum aestivum) The advantages of these two experimental designs are that: plants. Such monosomic plants have abnormal female and 1. non-Mendelian segregation can be attributed to SEA, male meiosis, only meiospores containing the alien chromo- 2. it can be established if SEA leads to higher offspring quality, some develop into normal gametophytes. Only such a cyto- and linked to the genotype of the offspring, logical analysis combined with mapping would allow for attri- 3. at least for the control treatment, SEA can be studied in un- buting non-Mendelian segregation found in the map to meio- 4. non-Mendelian segregation can also be studied among the Gametophytic selection includes all phenomena that cause 5. it is possible to find markers for SEA that can be used on differential success of pollen from different donors or pollen from the same donor but bearing different alleles. Gametophy- 6. major loci controlling SEA can be detected.
tic selection may occur, for example, during pollen germina- tion and pollen tube growth. Germination of pollen in vitro is a standard method used to assess both its viability and pollen tube growth. However, only in a very few cases are such tests The traditional experimental phenotypic approach to test the combined with data on segregation of molecular markers (Lin SEA hypothesis has the disadvantage that the treatment itself et al., 1992[65]; Sari-Gorla et al., 1992[97]). Often, gametophytic can be a source of artefacts. That is why the evidence that SEA selection is assumed to occur on the basis of allele frequencies Selective Embryo Abortion Hypothesis Revisited ± A Molecular Approach in the offspring for the loci in which non-Mendelian segrega- of Smith (1988[102]). Nineteen out of 60 angiosperms showed tion was found. Under-representation of one of the alleles is at least occasional biparental inheritance. Less examples (only then attributed to gametophytic selection in one of the parents four species) are available for biparental inheritance of mito- (see e.g., Wagner et al., 1992[114]). Such studies neglect the fact chondria for two reasons. Firstly, this phenomenon has not re- that post-fertilization selection affecting homozygotes also in- ceived much attention (Reboud and Zeyl, 1994[90]). Secondly, it may occur less often. Species with biparental inheritance of plastids may have strictly maternal transmission of mitochon- Selective germination and seedling death. Kuang et al.
(1998[59]) linked non-Mendelian segregation to seedling death in Pinus radiata. A comparison of the segregation of RAPD mar- Endophytes. A diversity of organisms, like bacteria and fungi, kers was made for megagametophytes, for surviving seedlings are known to live inside and among plant tissues (Carroll, and those that died within the first month after germination in 1988[18]; Clay, 1988[21]; Hallmann et al., 1997[38]). The DNA from order to find markers for which segregation was significantly endophytes may be extracted together with plant DNA and skewed in opposite directions in both groups. A null allele of eventually give the same effect as contamination. Cytoplasmic one locus was over-represented in dead seedlings while it inheritance and endophytes can potentially be observed as dis- was strongly under-represented in the seedlings that were still torted unlinked markers. However, molecular markers for or- alive. The authors suggested that an allele closely linked to this ganelle DNA will never be linked to markers for nuclear genes null allele is responsible for the seedlings death. Moreover, a and, if there are two polymorphic markers for organelle DNA, segregation analysis at the same locus for unsown seeds they will be 100% linked to each other because of a lack of re- showed that the null allele was over-represented in this stage.
combination. Molecular markers for eucaryotic endophyte Kuang et al. (1998[59]) gave two possible explanations: selec- DNA may appear in a map (resulting in more groups than chro- tion favouring this allele prior to germination or a sampling mosomes). However, they will never be linked to the markers error. If the allele responsible for seedling death is indeed that are known to be developed for plants e.g., morphological favoured during embryo maturation, this would present a case opposite to that is predicted by the SEA hypothesis.
Chromosome rearrangements, such as translocation and du- The elimination of selective seed germination and seedling plication, are often suggested causes of non-Mendelian segre- death as the explanation for non-Mendelian segregation found gation found in genetic maps (e.g., Vaillancourt and Slinkard, in the map can be done if seed that did not germinate and dead 1992[112]). However, genetic mapping alone is not sufficient to seedlings are included into the segregation analysis.
link non-Mendelian segregation directly to translocation. Be- lay and Merker (1998[7]) analysed inheritance of the transloca- B chromosomes are supernumerary chromosomes that are tion on chromosome 6B in tetraploid wheat (Triticum turgi- not essential for growth and reproduction of organisms. They dum) by cytogenetics. They observed C-banding patterns in have been described in more than 1000 species of plants (see the F2 generation derived from two crosses, both with one par- Jones and Rees, 1982[47] for a review). B chromosomes can be ent homozygous for a translocation and one parent homozy- distinguished from normal (A) chromosomes because they gous for the lack of translocation. In both crosses, non-Men- are usually smaller and consist of large amounts of heterochro- delian segregation was observed. Homozygotes containing 2 matin. Their number may be variable even within the same in- chromosomes without translocation were over-represented.
dividuals due to nondisjunction in anaphase of mitosis. B chro- mosomes are inherited in a non-Mendelian fashion which, ac- cording to Jones (1991[46]), can be caused by their nondisjunc- tion in female or/and male meiosis, nondisjunction in mitosis 1 Andersson, S. (1990) No evidence for selective seed maturation in during development of the male gametophyte, or preferential Anchusa officinalis (Boraginaceae). Oikos 57, 88±93.
fertilization by B-containing sperm. B chromosomes usually 2 Andersson, S. (1993) The potential for selective seed maturation consist of repetitive DNA and some of such repeats were found in Achillea ptarmica (Asteraceae). Oikos 66, 36±42.
to code for ribosomal RNA (Camacho et al., 2000[17] and refer- 3 Arathi, H. S., Ganeshaiah, K. N., Uma Shaanker, R., and Hegde, S. G.
ences therein). Theoretically, markers of B chromosomes may (1999) Seed abortion in Pongamia pinnata (Fabaceae). Am. J. Bot.
appear in genetic maps as single, unlinked markers or separate linkage groups. A way to avoid this possibility is careful selec- Barreneche, T., Bodenes, C., Lexer, C., Trontin, J. F., Fluch, S., Streiff, tion of parents without B chromosomes (by chromosome R., Plomion, C., Roussel, G., Steinkellner, H., Burg, K., Favre, J. M., Glossl, J., and Kremer, A. (1998) A genetic linkage map of Quercus counting), for the mapping population.
robur L. (pedunculate oak) based on RAPD, SCAR, microsatellite, minisatellite, isozyme and 5S rDNA markers. Theor. Appl. Genet.
Cytoplasmic inheritance via plastids and mitochondria can show up as non-Mendelian segregation in molecular markers 5 Baudracco-Arnas, S. and Pitrat, M. (1996) A genetic map of melon in the case of biparental inheritance of organelles. In angio- (Cucumis melo L.) with RFLP, RAPD, isozyme, disease resistance sperms, inheritance of plastids is predominantly maternal and morphological markers. Theor. Appl. Genet. 93, 57±64.
and in gymnosperms ± paternal. However, cases of biparental 6 Beaumont, V. H., Mantet, J., Rocheford, T. R., and Widholm, J. M.
inheritance of those organelles were also described. Examples (1996) Comparison of RAPD and RFLP markers for mapping F2 are alfalfa (Medicago sativa), evening primrose (Oenothera) generations in maize (Zea mays L.). Theor. Appl. Genet. 93, 606± and Pelargonium cultivars (Mogensen, 1996[79]). Ten out of 68 angiosperm species and 3 out of 11 gymnosperms listed by 7 Belay, G. and Merker, A. (1998) Cytogenetic analysis of a sponta- Reboud and Zeyl (1994[90]) are classified as having biparental neous 5B/6B translocation in tetraploid wheat landraces from inheritance of plastids. A similar picture arises from a review Ethiopia, and implications for breeding. Plant Breeding 117, G. Korbecka, P. G. L. Klinkhamer, and K. Vrieling 8 Bertin, R. I. (1982) Paternity and fruit production in trumpet 30 Freyre, R., Skroch, P. W., Geffroy, V., Adam-Blondon, A. F., Shirmo- creeper (Campsis radicans). Am. Nat. 119, 694 ±709.
hamadali, A., Johnson, W. C., Llaca, V., Nodari, R. O., Pereira, P. A., 9 Blanco, A., Bellomo, M. P., Cenci, A., De Giovanni, C., DOvidio, R., Tsai, S. M., Tohme, J., Dron, M., Nienhuis, J., Vallejos, C. E., and Iacono, E., Laddomada, B., Pagnotta, M. A., Porceddu, E., Sciancale- Gepts, P. (1998) Towards an integrated linkage map of common pore, A., Simeone, R., and Tanzarella, O. A. (1998) A genetic linkage bean. 4. Development of a core linkage map and alignment of map of durum wheat. Theor. Appl. Genet. 97, 721±728.
RFLP maps. Theor. Appl. Genet. 97, 847±856.
10 Bookman, S. S. (1984) Evidence for selective fruit production in 31 Ganeshaiah, K. N. and Uma Shaanker, R. (1988) Seed abortion in wind-dispersed pods of Dalbergia sissoo: maternal regulation or 11 Boyko, E. V., Gill, K. S., Mickelson-Young, L., Nasuda, S., Raupp, W.
sibling rivalry? Oecologia 77, 135±139.
J., Ziegle, J. N., Singh, S., Hassawi, D. S., Fritz, A. K., Namuth, D., La- 32 Gianfranceschi, L., Seglias, N., Tarchini, R., Komjanc, M., and Gess- pitan, N. L. V., and Gill, B. S. (1999) A high density genetic linkage ler, C. (1998) Simple sequence repeats for the genetic analysis of map of Aegilops tauschii, the D-genome progenitor of bread apple. Theor. Appl. Genet. 96, 1069±1076.
wheat. Theor. Appl. Genet. 99, 16±26.
33 Gibbs, P. and Sassaki, R. (1998) Reproductive biology of Dalberia 12 Bradbury, D. (1929) A comparative study of the developing and miscolobium Benth. (Leguminosae-Papilionideae) in SE Brazil: aborting fruits of Prunus cerasus. Am. J. Bot. 16, 525±543.
The effects of pistillate sorting on fruit set. Ann. Bot. 81, 735±740.
13 Buckler, E. S., Phelps-Durr, T. L., Buckler, C. S. K., Dawe, R. K., Doeb- 34 Goldberg, R. B., Barker, S. J., and Perez-Grau, L. (1989) Regulation ley, J. F., and Holtsford, T. P. (1999) Meiotic drive of chromosomal of gene expression during plant embryogenesis. Cell 56, 149±160.
knobs reshaped the maize genome. Genetics 152, 415±426.
35 Guiderdoni, E. (1991) Gametic selection in anther culture of rice 14 Butcher, P. A., Moran, G. F., and Bell, R. (2000) Genetic linkage (Oryza sativa L.). Theor. Appl. Genet. 81, 406±412.
mapping in Acacia magnum. 1. Evaluation of restriction endonu- 36 Guth, C. J. and Weller, S. G. (1986) Pollination, fertilization and cleases, inheritance of RFLP loci and their conservation across ovule abortion Oxalis magnifica. Am. J. Bot. 73, 246±253.
species. Theor. Appl. Genet. 100, 576±583.
37 Haig, D. (1992) Brood reduction in gymnosperms. In Cannibal- 15 Burd, M. (1994) Batemans principle and plant reproduction: the ism ± Ecology and Evolution among Diverse Taxa (Elgar, M. A.
role of pollen limitation in fruit and seed set. Bot. Rev. 60, 83±139.
and Crespi, B. J., eds.), Oxford: Oxford University Press, pp. 63±84.
16 Burd, M. (1998) ªExcessº flower production and selective fruit 38 Hallmann, J., Quadt-Hallmann, A., Mahaffee, W. F., and Kloepper, J.
abortion: A model of potential benefits. Ecology 79, 2123±2132.
W. (1997) Bacterial endophytes in agricultural crops. Canadian 17 Camacho, J. P. M., Sharbel, T. F., and Beukeboom, L. W. (2000) B- Journal of Microbiology 43, 895±914.
chromosome evolution. Philosophical Transactions of the Royal 39 Hufford, K. M., Kochert, G., and Hamrick, J. L. (2000) Microsatellite primers and amplification of aborted embryos in Platypodium ele- 18 Carroll, G. (1988) Fungal endophytes in stems and leaves: from la- gans J. Vogel (Fabaceae, Papilionoideae). Molecular Ecology 9, tent pathogen to mutualistic symbiont. Ecology 69, 2±9.
19 Casper, B. B. (1988) Evidence for SEA in Cryptantha flava. Am. Nat.
40 Husband, B. C. and Schemske, D. W. (1995) Magnitude and timing of inbreeding depression in diploid population of Epilobium angu- 20 Chen, C., Sleper, D. A., and Johal, G. S. (1998) Comparative RFLP stifolium (Onagraceae). Heredity 75, 206±215.
mapping of meadow and tall fescue. Theor. Appl. Genet. 97, 41 Husband, B. C. and Schemske, D. W. (1996) Evolution of the mag- nitude and timing of inbreeding depression in plants. Evolution 21 Clay, K. (1988) Fungal endophytes of grasses: a defensive mutual- ism between plants and fungi. Ecology 69, 10±16.
42 Jaing, Q. and Gresshoff, P. M. (1997) Classical and molecular ge- 22 Cloutier, S., Cappadocia, M., and Landry, B. S. (1997) Analysis of netics of the model legume Lotus japonicus. Molecular Plant-Mi- RFLP mapping inaccuracy in Brassica napus L. Theor. Appl. Genet.
43 Jenczewski, E., Gherardi, M., Bonnin, I., Prosperi, J. M., Olivieri, I., 23 Costa, P., Pot, D., Dubos, C., Frigerio, J. M., Pionneau, C., Bodenes, C., and Huguet, T. (1997) Insight on segregation distortions in two in- Bertocchi, E., Cervera, M. T., Remington, D. L., and Plomion, C.
traspecific crosses between annual species of Medicago (Legumi- (2000) A genetic map of Martime pine based on AFLP, RAPD and nosae). Theor. Appl. Genet. 94, 682±691.
protein markers. Theor. Appl. Genet. 100, 39±48.
44 Jermstad, K. D., Bassoni, D. L., Wheeler, N. C., and Neale, D. B.
24 Crouzillat, D., Lerceteau, E., Petiard, V., Morera, J., Rodriguez, H., (1998) A sex-averaged genetic linkage map in coastal Douglas-fir Walker, D., Phillips, W., Ronning, C., Schnell, R., Osei, J., and Fritz, (Pseudotsuga menziesii [Mirb.] Franco var ªmenziesiiº) based on P. (1996) Theobroma cacao L.: a genetic map and qualitative trait RFLP and RAPD markers. Theor. Appl. Genet. 97, 762±770.
loci analysis. Theor. Appl. Genet. 93, 205 ±214.
45 Jones, C. J., Edwards, K. J., Castaglione, S., Winfield, M. O., Sala, F., 25 Dirlewanger, E., Pronier, V., Parvery, C., Rothan, C., Guye, A., and vandeWiel, C., Bredemeijer, G., Vosman, B., Matthes, M., Daly, A., Monet, R. (1998) Genetic linkage map of peach Prunus persica Brettschneider, R., Bettini, P., Buiatti, M., Maestri, E., Malcevschi, (L.) [Batsch] using morphological and molecular markers. Theor.
A., Marmiroli, N., Aert, R., Volckaert, G., Rueda, J., Linacero, R., Vaz- quez, A., and Karp, A. (1997) Reproducibility testing of RFLP, AFLP 26 Dufour, P., Deu, M., Grivet, L., Dhont, A., Paulet, F., Bouet, A., La- and SSR markers in plants by a network of European laboratories.
naud, C., Glaszmann, J. C., and Hamon, P. (1997) Construction of a composite sorghum genome map and comparison with sugar- 46 Jones, R. N. (1991) B-chromosome drive. Am. Nat. 137, 430±442.
cane, a related complexpolyploid. Theor. Appl. Genet. 94, 409± 47 Jones, R. N. and Rees, H. (1982) B chromosomes. London: Aca- 27 Echt, C. S. and Nelson, C. D. (1997) Linkage mapping and genome 48 Jong, T. J. de and Klinkhamer, P. G. L. (1989) Limiting factors for length in eastern white pine (Pinus strobus L.). Theor. Appl. Genet.
seed production in Cynoglossum officinale. Oecologia 80, 167± 28 Faris, J. D., Laddomada, B., and Gill, B. S. (1998) Molecular map- 49 Karkkainen, K., Savolainen, O., and Koski, V. (1999) Why do plants ping of segregation distortion loci in Aegilops tauschii. Genetics abort so many developing seeds: bad offspring or bad maternal genotypes? Evol. Ecol. 13, 305±317.
29 Finch, R. A., Miller, T. E., and Bennet, M. D. (1984) ªCuckooº Aegi- 50 Katzir, N., Danin-Poleg, Y., Tzuri, G., Karchi, Z., Lavi, U., and Cregan, lops addition chromosome in wheat ensures its transmission by P. B. (1996) Length polymorphism and homologies of micro- causing chromosome breaks in meiospores lacking it. Chromoso- satellites in several Cucurbitaceae species. Theor. Appl. Genet.
Selective Embryo Abortion Hypothesis Revisited ± A Molecular Approach 51 Kennard, W., Phillips, R., Porter, R., and Grombacher, A. (1999) A ni, S., Schmidt, H., Tartarini, S., Verhaegh, J. J., Vrielink-van Ginkel, comparative map of wild rice (Zizania palustris L. 2n = 2x= 30).
M., and King, G. J. (1998) Aligning male and female linkage maps of apple (Malus pumila Mill.) using multi-allelic markers. Theor.
52 King, J. J., Bradeen, J. M., Bark, O., McCallum, J. A., and Havey, M. J.
(1998) A low-density genetic map of onion reveals a role for tan- 72 Marshall, D. and Ellstrand, N. C. (1988) Effective mate choice in dem duplication in the evolution of an extremely large diploid wild radish: evidence for selective embryo abortion and its genome. Theor. Appl. Genet. 96, 52±62.
53 Klinkhamer, P. G. L., Jong, T. J. de, and Nell, H. W. (1994) Limiting 73 Marshall, D. L. and Folsom, M. W. (1991) Mate choice in plants: an factors for seed production and phenotypic gender in the gyno- anatomical to population perspective. Annu. Rep. Ecol. Syst. 22, dioecious species Echium vulgare (Boraginaceae). Oikos 71, 469± 74 Marshall, D. L. and Whittaker, K. L. (1989) Effects of pollen donor 54 Koelwijn, H. P., Koski, V., and Savolainen, O. (1999) Magnitude and identity on offspring quality in wild radish, Raphanus sativus. Am.
timing of inbreeding depression in scots pine (Pinus sylvestris L.).
75 Melser, C., Rademaker, M. C. J., and Klinkhamer, P. G. L. (1997) Se- 55 Korzun, V., Malyshev, S., Kartel, N., Westermann, T., Weber, W. E., lection on pollen donors by Echium vulgare (Boraginaceae). Sex.
and Borner, A. (1998) A genetic map of rye (Secale cereale L.). The- 76 Melser, C. and Klinkhamer, P. G. L. (2001) Selective seed abortion 56 Kozøowski, J. and Stearns, S. C. (1989) Hypotheses for the produc- increases offspring survival in Cynoglossum officinale L. (Boragi- tion of excess zygotes: models of the bet-hedging and SEA. Evolu- naceae). Am. J. Bot. 88, 1033±1040.
77 Menendez, C. M., Hall, A. E., and Gepts, P. (1997) A genetic linkage 57 Krishnamurthy, K. S., Uma Shaanker, R., and Ganeshaiah, K. N.
map of cowpea (Vigna unguiculata) developed from a cross be- (1997) Seed abortion in an animal dispersed species, Syzygium cu- tween two inbred, domesticated lines. Theor. Appl. Genet. 95, minii (L.) Skeels (Myrtaceae): the chemical basis. Current Science 78 Miller, T. E. (1983) Preferential transmission of an alien chromo- 58 Krutovskii, K. V., Vollmer, S. S., Sorensen, F. C., Adams, W. T., Knapp, somes in wheat. In Kew chromosome conference 2 (Brandham, P.
S. J., and Strauss, S. H. (1998) RAPD genome maps of Douglas Fir. J.
E. and Bennett, M. D., eds.), London: George Allenand Unwin, 59 Kuang, H., Richardson, T. E., Carson, S. D., and Bongarten, B. C.
79 Mogensen, H. L. (1996) The hows and whys of cytoplasmic inheri- (1998) An allele responsible for seedling death in Pinus radiata tance in seed plants. Am. J. Bot. 83, 383±404.
D. Don. Theor. Appl. Genet. 96, 640±644.
80 Mohan Raju, B., Uma Shaanker, R., and Ganeshaiah, K. N. (1996) 60 Kuittinen, H., Sillanpaa, M. J., and Savolainen, O. (1997) Genetic Intra-fruit seed abortion in wind dispersed tree, Dalbergia sissoo basis of adaptation: flowering time in Arabidopsis thaliana. Theor.
Roxb: proximate mechanisms. Sex. Plant Reprod. 9, 273±278.
81 Montalvo, A. M. (1992) Relative success of self and outcross pollen 61 Latta, R. G. (1995) The effects of embryo competition with mixed comparing mixed- and single donor pollinations in Aquilegia mating on the genetic load in plants. Heredity 75, 637±643.
62 Laucou, V., Haurogne, K., Ellis, N., and Rameau, C. (1998) Genetic 82 Mukai, Y., Suyama, Y., Tsumura, Y., Kawahara, T., Yoshimaru, H., mapping in pea. 1. RAPD-based genetic linkage map of Pisum sati- Kondo, T., Tomaru, N., Kuramoto, N., and Murai, M. (1995) A link- vum. Theor. Appl. Genet. 97, 905±915.
age map for sugi (Cryptomeria japonica) based on RFLP, RAPD, and 63 Lee, T. D. (1988) Patterns of fruit and seed production. In Plant Re- isozyme loci. Theor. Appl. Genet. 90, 835±840.
productive Ecology (Doust, J. L. and Doust, L. L., eds.), New York, 83 Nakamura, R. R. (1988) Seed abortion and seed size variation Oxford: Oxford University Press, pp.179±202.
within fruits of Phaseolus vulgaris: pollen donor and resource lim- 64 Lerceteau, E. and Szmidt, A. E. (1999) Properties of AFLP markers itation effects. Am. J. Bot. 75, 1003±1010.
in inheritance and genetic diversity studies of Pinus sylvestris L.
84 Paglia, G. P., Olivieri, A. M., and Morgante, M. (1998) Towards sec- ond-generation STS (sequence tagged sites) linkage maps in co- 65 Lin, S. Y., Ikehashi, H., Yanagihara, S., and Kawashima, A. (1992) nifers: a genetic map of Norway spruce (Picea abies K.). Mol. Gen.
Segregation distortion via male gametes in hybrids between Indi- ca and Japonicca or wide varieties of rice (Oryza sativa L.). Theor.
85 Pham, J. L., Glaszmann, J. C., Sano, R., Barbier, P., Ghesquiere, A., and Second, G. (1990) Isozyme markers in rice: genetic analysis 66 Lin, S. Y., Sasaki, T., and Yano, M. (1998) Mapping quantitative trait and linkage relationships. Genome 33, 348±359.
loci controlling seed dormancy and heading date in rice, Oryza 86 Plomion, C., Bahrman, N., Durel, C. E., and OMalley, D. M. (1995) satva L., using backcross inbred lines. Theor. Appl. Genet. 96, Genomic mappig in Pinus pinaster (martime pine) using RAPD and protein markers. Heredity 74, 661±668.
67 Liu, K. D., Wang, J., Li, H. B., Xu, C. G., Liu, A. M., Li, X. H., and Zhang, 87 Plomion, C., Durel, C. E., and OMalley, D. M. (1996) Genetic dis- Q. F. (1997) A genome-wide analysis of wide compatibility in rice section of height in maritime pine seedlings raised under acceler- and the precise location of the S5 locus in the molecular map. The- ated growth conditions. Theor. Appl. Genet. 93, 849±858.
88 Price, A. H. and Tomos, A. D. (1997) Genetic dissection of root 68 Loarce, Y., Hueros, G., and Ferrer, E. (1996) A molecular linkage growth in rice (Oryza sativa L.). 2: mapping quantitative loci using map of rye. Theor. Appl. Genet. 93, 1112±1118.
molecular markers. Theor. Appl. Genet. 95, 143±152.
69 Lynch, M. and Walsh, B. (1998) Genetics and Analysis of Quantita- 89 Qi, X., Stam, P., and Lindhout, P. (1998) Use of locus-specific AFLP tive Traits. Sunderland, Massachusetts: Sinauer Associates, markers to construct a high-density molecular map in barley.
70 Lyttle, T. W. (1991) Segregation distorters. Ann. Rev. Genet. 25, 90 Reboud, X. and Zeyl, C. (1994) Organelle inheritance in plants.
71 Maliepaard, C., Alston, F. H., Arkel, G. van, Brown, L. M., Chevreau, 91 Remington, D. L. and OMalley, D. M. (2000) Whole-genome char- E., Dunemann, F., Evans, K. M., Gardiner, S., Guilford, P., van Heus- acterisation of embryonic stage inbreeding depression in a selfed den, A. W., Janse, J., Laurens, F., Lynn, J. R., Manganaris, A. G., den loblolly pine family. Genetics 155, 337±348.
Nijs, A. P. M., Periam, N., Rikkerink, E., Roche, P., Ryder, C., Sansavi- G. Korbecka, P. G. L. Klinkhamer, and K. Vrieling 92 Reusch, T. B. H. (2000) Pollination in the marine realm: micro- 114 Wagner, H., Weber, W. E., and Wricke, G. (1992) Estimating link- satellites reveal high outcrossing rate and multiple paternity in age relationship of isozyme markers and morphological markers eelgrass Zostera marina. Heredity 85, 459 ±464.
in sugar beet (Beta vulgaris L.) including families with distorted 93 Richter, K., Schondelmaier, J., and Jung, C. (1998) Mapping of segregation. Plant Breeding 108, 89±96.
quantitative trait loci affecting Drechslera teres resistance in bar- 115 Wang, G., Mahalingam, R., and Knap, H. T. (1998) (C-A) and (G-A) ley with molecular markers. Theor. Appl. Genet. 97, 1225 ±1234.
anchored simple sequence repeats (ASSRs) generated poly- 94 Rigney, L. P. (1995) Post fertilization causes of differential success morphism in soybean, Glycine max (L.) Merr. Theor. Appl. Genet.
of pollen donors in Erythronium grandiflorum (Liliaceae): nonran- dom ovule abortion. Am. J. Bot. 82, 578±584.
116 Wang, Y. H., Thomas, C. E., and Dean, R. A. (1997) A genetic map of 95 Rieseberg, L. H. (1996) Homology among RAPD fragments in in- melon (Cucumis melo L.) based on amplified fragment length terspecific comparisons. Mol. Ecol. 5, 99±105.
polymorphism (AFLP) markers. Theor. Appl. Genet. 95, 791±798.
96 Rocha, O. J. and Stephenson, A. G. (1991) Effects of nonrandom 117 Wang, Z. M., Devos, K. M., Liu, C. J., Wang, R. Q., and Gale, M. D.
seed abortion on progeny performance in Phaseolus coccineus L.
(1998) Construction of RFLP-based maps of foxtail millet, Setaria italica (L.) P. Beauv. Theor. Appl. Genet. 96, 31±36.
97 Sari-Gorla, M. E. P. E., Mulcahy, D. L., and Ottaviano, E. (1992) Ge- 118 Waser, N. M. and Prince, M. V. (1991) Reproductive costs of self- netic dissection of pollen competitive ability in maize. Heredity pollination in Ipomopsis aggregata (Polemoniaceae) are ovules usurped? Am. J. Bot. 78, 1036±1043.
98 Seavey, S. R. and Carter, S. K. (1996) Ovule fates in Epilobium ob- 119 Wiens, D. (1984) Ovule survivorship, brood size, life history, cordatum (Orangaceae). Am. J. Bot. 83, 316±325.
breeding systems, and reproductive success in plants. Oecologia 99 Shappley, Z. W., Jenkins, J. N., Meredith, W. R., and McCarty, J. C.
(1998) An RFLP linkage map of Upland cotton, Gossypium hirsu- 120 Wiens, D., Calvin, C. L., Wilson, C. A., Davern, C. I., Frank, D., and tum L. Theor. Appl. Genet. 97, 756 ±761.
Seavey, S. R. (1987) Reproductive success, spontaneous embryo 100 Skov, E. and Wellendorf, H. (1998) A partial linkage map of Picea abortion, and genetic load in flowering plants. Oecologia 71, abies clone V6470 based on recombination of RAPD-markers in haploid megagametophytes. Silvae Genetica 47, 273±282.
121 Wiens, D., Nickrent, D. L., Davern, C. I., Calvin, C. L., and Vivrette, N.
101 Skov, E. (1998) Mendelian inheritance and tissue expression of J. (1989) Developmental failure and loss of reproductive capacity RAPD-markers in Picea abies (L.) Karst. Silvae Genetica 47, 262± in the rare palaeoendemic shurb Dedeckera eurekensis. Nature 102 Smith, S. E. (1988) Biparental inheritance of organelles and its im- 122 Wilson, M. F. and Burley, N. (1983) Mate Choice in Plants. Prince- plications in crop improvement. Plant Breed. Rev. 6, 361±393.
ton, New Jersey: Princeton University Press.
103 Sorensen, F. C. (1982) The roles of polyembryony and embryo via- 123 Wolfe, L. M. (1983) The effect of plant size on reproductive char- bility in the genetic system of conifers. Evolution 36, 725±733.
acteristics in Erythronium americanum (Liliaceae). Can. J. Bot. 61, 104 Sondur, S. N., Manshardt, R. M., and Stiles, J. I. (1996) A genetic linkage map of papaya based on randomly amplified polymorphic 124 Wu, R. L., Han, Y. F., Hu, J. J., Fang, J. J., Li, L., Li, M. L., and Zeng, Z. B.
DNA markers. Theor. Appl. Genet. 93, 547±553.
(2000) An integrated genetic map of Populus deltoides based on 105 Stephenson, A. G. (1981) Flower and fruit abortion: proximate amplified fragment length polymorphisms. Theor. Appl. Genet.
causes and ultimate functions. Ann. Rev. Ecol. Syst. 12, 253±279.
106 Stephenson, A. G. and Winsor, J. A. (1986) Lotus corniculatus regu- 125 Xu, Y., Zhu, L., Xiao, J., Huang, N., and McCouch, S. R. (1997) Chro- lates offspring quality through selective fruit abortion. Evolution mosomal regions associated with segregation distortion of mo- lecular markers in F2, backcross, doubled haploid, and recombi- 107 Tan, X. L., Vanavichit, A., Amornsilpa, S., and Trangoonrun, S.
nant inbred populations in rice (Oryza sativa L.). Mol. Gen. Genet.
(1998) Genetic analysis of rice CMS-WA fertility restoration based on QTL mapping. Theor. Appl. Genet. 97, 994 ±999.
126 Zimmerman, M. and Pyke, G. H. (1988) Reproduction in Polemo- 108 Teulat, B., This, D., Khairallah, M., Barries, C., Ragot, C., Sourdille, P., nium: assessing the factors limiting seed set. Am. Nat. 131, 723± Leroy, P., Monneveux, P., and Charrier, A. (1998) Several QTLs in- volved in osmotic-adjustment trait variation in barley (Hordeumvulgare L.). Theor. Appl. Genet. 96, 688±698.
109 Travis, S. E., Ritland, K., Whitham, T. G., and Keim, P. A. (1998) Ge- netic linkage map of Pinyon pine (Pinus edulis) based on amplified Institute of Evolutionary and Ecological Sciences fragment length polymorphisms. Theor. Appl. Genet. 97, 871± Tuberosa, R., Sanguineti, M. C., Landi, P., Salvi, S., Casarini, E., and Conti, S. (1998) RFLP mapping of quantitative trait loci controlling abscisic acid concentration in leaves of drought-stressed maize (Zea mays L.). Theor. Appl. Genet. 97, 744±755.
111 Uma Shaanker, R., Ravishankar, K. V., Hegde, S. G., and Ganeshaiah, K. N. (1996) Does endosperm reduce intra-fruit competition among developing seeds? Pl. Syst. Evol. 201, 263±270.
112 Vaillancourt, R. E. and Slinkard, A. E. (1992) Inheritance of new ge- netic markers in lentil (Lens Miller). Euphytica 64, 227±236.
113 Vaz Patto, M. C., Martin, A., Lindhout, P., Stam, P., and Niks, R. E.
(2001) A genetic map of wild barley (Hordeum chilense) based on amplified fragment length polymorphisms. In The genetics andmechanism of avoidance of rust infection in Hordeum chilense.
University of Wageningen: PhD thesis.

Source: http://miekevd.home.xs4all.nl/PDF/74.pdf