What is the difference between cytoplasmic inheritance and genetic maternal effect?

Cytoplasmic Genetics

The clearest example of cytoplasmic inheritance in animal cells is the mitochondrial genome. The approximately 16,000 base-pair circular mitochondrial genome has genes for ribosomal RNAs, transfer RNAs, and approximately a dozen mitochondrial proteins, including a polymerase (Cummins, 1998). Although the vast majority of mitochondrial proteins are specified by chromosomal genes, the mitochondrial genome has considerable variability, and mutates more readily than chromosomal genes, in part due to poor proofreading during DNA synthesis (Cummins, 1998). The resulting mutations are the source of considerable (cytoplasmic) genetic disease, as well as phenotypic variation in normal mitochondrial function. Although there are rare exceptions (Cummins, 1998), for the most part mitochondrial genomes are inherited via the maternal ooplasm. The relatively few mitochondria introduced by the sperm usually degenerate, and in any case become so dilute that they usually would not end up in the relatively few cells of the blastocyst that differentiate into the resulting animal.

The maternal lineages of mitochondria in the recipient oocyte and in the donor nucleus often will be different. Simply by the process of dilution, the animal resulting from cloning procedures using cell fusion usually ends up with mitochondrial genes of the oocyte, not the donor nucleus. Of course, mitochondria will be heteroplasmic initially; sometimes this situation persists, and rarely the donor mitochondria out-compete the recipient oocyte mitochondria, but usually the mitochondria of the recipient oocyte prevail (Evans et al., 1999). It is also possible that donor mitochondria do not survive well in ooplasm because of the very specially differentiated state of mitochondria in oocytes (Cummins, 1998). Experiments to sort this out are progressing, and it is obvious that answers will have a statistical quality rather than a simple outcome. Note that, in the future, it is likely that heteroplasmy of mitochondrial genomes can be eliminated by selective elimination of donor or recipient mitochondria by chemical or other means.

Cytoplasmic inheritance also occurs for centrosomes, usually via the sperm in mammals (Stearns, 2001). To my knowledge, no cytoplasmic nucleic acid sequence information is involved, but the semiconservative nature of centriole duplication (Stearns, 2001) does have implications for cloning, particularly if procedures become more sophisticated. For example, centrosomal material might be provided from sperm parts rather than from donor cell cytoplasm. The sperm cytoplasm also contributes to embryonic development under some (most?) circumstances via microRNAs and phospholipase C zeta (Ito et al., 2011; Liu et al., 2012).

Cytoplasmic inheritance of viruses occurs in some situations. In these cases, there is a nucleic acid sequence specifying a cytoplasmic component.

Non-Mendelian Inheritance in Pelargonium zonale

Usually two men of science are quoted to have discovered non-Mendelian or cytoplasmic inheritance in plants at about the same time (1909): Erwin Baur and Carl Correns. However, it was definitely Baur who clearly drew attention to separate plastid inheritance whilst Correns developed a hypothesis that only the cytoplasm has changed but not the plastids themselves. The following quotations illustrate Baur's cautious, but clear-cut and entirely correct conclusions. On the very special case of biparental plastid genetics,Baur, 1909, pp. 349–350) wrote:

The zygote, arisen by uniting of a “green” and a “white” sexual cell, contains two different plastids, green and white ones. In the course of cell divisions forming the embryo, the plastids segregate to the daughter cells according to the laws of probability. If a daughter cell has only white plastids, all the cells derived from it will be white generating a white patch of cells. If the cell has only green plastids, a green complex of cells is produced. There is no need for me to further analyze (the point) that cells with both kinds of plastids will be able to continue to segregate.

…According to the present dominating opinion, the plastids of a zygote are derived solely from the mother. Whether this view is absolutely sure is not for me to decide…If, however, in contrast to the expert opinion so far, it can be shown that male gametes can also transmit plastids, the hereditary relations of the plants with the white edges will be entirely clear. Further studies will decide these questions.

Baur's analysis was proved to be fully correct by further research.

Genetic Effect and Utilization of Reciprocal Cross in Animals

Reciprocal crosses can be used to detect sex-linked inheritance and maternal or cytoplasmic inheritance in animals. For example, sex linkage was reported in 1910 by Thomas Morgan who studied the reciprocal crosses between white-eyed and red-eyed Drosophila. He found that white-eyed males crossed with red-eyed females produced only red-eyed offspring, while red-eyed males crossed with white-eyed females produced white-eyed male offspring and red-eyed female offspring. In this example, the recessive white-eyed allele r is assumed on Y chromosome; then the genotypes of red-eyed female and white-eyed male are XR XR and Xr Yr, respectively, in the first cross and the genotypes of white-eyed female and red-eyed male are Xr Xr and XR Yr, respectively, in the second cross (red-eyed allele R dominant to white-eyed allele r). From the assumption, the results of phenotypic eye color in the reciprocal crosses can be expected just as those indicated in Morgan’s experiment. Thus, the sex linkage of eye color was confirmed.

The example of maternal influence was reported by Arthur Boycott in 1930 who was studying morphological features of Lymnaea peregra. In L. peregra, the shell and internal organs can be arranged in one of two directions, right-handed (dextral) and left-handed (sinistral). Boycott crossed the two types of L. peregra in the form of reciprocal cross and found that the shell rotation of the F1 hybrids was the same as their respective mother and that of the F2 hybrids was all right-handed and segregation in the ratio of 3:1 occurred in F3. In this example, the maternal influence (due to female nuclear influence) was found. The trait is determined by both the nuclear genetic system (a single gene with right-handed allele dominant to left-handed allele) and the maternal nuclear influence. In this genetic system, the reciprocal F1’s should show their mothers’ phenotypes, that is, right-handed shell versus left-handed shell. If there is no maternal nuclear influence, a segregation ratio with 3 right-handed shell:1 left-handed shell in each of the two reciprocal F2’s should be expected. But there was no segregation in F2 and all F2 individuals of the two reciprocal crosses showed right-handed shell in Boycott’s experiment. This is because the maternal nuclear genotypes of the reciprocal F1’s are both heterozygous with their phenotype of right-handed shell performed in F2. The nuclear genotypes of the reciprocal F2’s are both segregated in 1:2:1 genotypic ratio which in turn performed in phenotypic ratio of 3 right-handed shell:1 left-handed shell in each of the two reciprocal F3’s, as was shown in Boycott’s experiment.

From the above, reciprocal cross can be used to detect both nuclear and cytoplasmic genetic systems, as well as their interaction mechanisms. In addition to the nuclear genetic system, the cytoplasmic genetic system is mainly transported by the DNA of extranuclear organelles, such as mitochondria and chloroplast. These organelles also contain genes which inherit uniparentally. Furthermore, in plant and animal breeding, reciprocal cross is widely used to identify cytoplasmic gene(s) resistant to diseases and to detect beneficial cytoplasmic–nuclear interaction for enhanced performance of some traits to be improved.

A. Non-Mendelian inheritance in ciliates

Long before the advent of molecular genetics, ciliates enjoyed a period of success during which they were strongly associated with non-Mendelian phenomena, and in particular with the question of cytoplasmic inheritance (Nanney, 1985; Preer, 1993). They had become the favorite models of a few biologists, such as T. M. Sonneborn, who believed that Morgan's chromosome theory fell short of explaining all heredity and could not by itself—with no role credited to the cytoplasm—account for complex phenomena such as development and evolution (Harwood, 1985). The early discovery of mating types (Sonneborn, 1937) had made Paramecium one of the first single-celled organisms in which genetic analyses could be conducted. Although many characters were found to follow Mendelian inheritance and were ascribed to nuclear genes, it was observed from the very beginning that a number of hereditary characters did not behave in a Mendelian way (Preer, 1993, 2000). One of these non-Mendelian characters, which could hardly have been ignored by the Paramecium geneticist, was the mating type itself. Opposite mating types were shown to develop from identical genotypes, and in some species mating types even proved to be cytoplasmically transmitted to sexual progeny. Other well-studied cases of cytoplasmic inheritance included the “killer” trait (production of a toxin that killed other cells) and the serotypes (antigenic variants of the cell surface).

To explain these observations, Sonneborn and others proposed around 1945 that such characters were determined by “plasmagenes,” cytoplasmic particles endowed with the genetic properties of reproduction and mutability (Sonneborn, 1948, 1949). Not all plasmagenes, however, were held to be completely independent from nuclear genes; indeed, Mendelian mutations could abolish their maintenance in the cytoplasm, and some plasmagenes even appeared to originate from nuclear genes. In its most general form, the plasmagene theory assumed that phenotypic characters can be determined by both nuclear genes and plasmagenes, and that their inheritance patterns in crosses depend on which component differs in the strains analysed (Sonneborn, 1950). Accumulating evidence, however, soon revealed that different mechanisms were at work in different cases, and the search for a unifying theory was abandoned. The killer trait was found to be determined by the presence of an endosymbiotic bacterium multiplying in the cytoplasm, and serotypes were rationalized as the stable inheritance of alternative patterns of expression for a set of nuclear genes encoding variant surface proteins. Unresolved cases of cytoplasmic inheritance, such as that governing Paramecium mating types, were confined to the periphery of mainstream research by the strong nuclear hegemony that had built on the remarkable success of Mendelian genetics. As a result of this initial emphasis on cytoplasmic inheritance, ciliates were increasingly perceived as being “anomalous” organisms, and despite Sonneborn's advocacy of the view that the peculiarities of any organism could only help define the universal principles of life (Schloegel, 1999), Paramecium genetics largely fell into oblivion (Preer, 1997).

But ciliate research survived, as these organisms continued to prove useful models in other domains, such as cellular morphogenesis or membrane excitability. In the late 1970s, molecular biology began to revive some interest in the genetics of ciliates through a variety of fundamental discoveries, including telomere structure, genome-wide DNA rearrangements, self-splicing introns, deviant genetic codes, telomerase, histone acetyltransferase, and novel histone modifications. Interestingly, some of the new molecular findings again revealed puzzling non-Mendelian phenomena, which have now been characterized as homology-dependent effects. As the Mendelian era is culminating with the first complete sequences of eukaryotic genomes, it is no longer possible to ignore the fact that dividing cells must inherit, in addition to the genome, essential regulatory information that is not to be found in the genome sequence itself, and much excitement is being redirected at some poorly understood epigenetic mechanisms of gene regulation. Homology-dependent effects, in particular, cannot be accounted for by classical paradigms of molecular genetics; yet there are now indications that their widespread occurrence in eukaryotes reflects the evolutionary conservation of a sophisticated machinery. It is the purpose of this review to present available information on homology-dependent effects in ciliates, and their relevance to known cases of non-Mendelian inheritance. The diversity and novel aspects of ciliate effects have the potential of exerting a significant impact in this domain. As in many eukaryotes, ciliate homology effects can lead to the specific silencing of almost any gene during vegetative growth, but in addition they have been shown to participate in the programming of genome rearrangements that occur during development. More generally, the evidence for maternal effects directing an epigenetic programming of the zygotic genome through homology-dependent mechanisms may turn out to be of interest in other systems. For a proper understanding of these maternal effects, some introduction to ciliate biology is necessary.

Reception of Symbiosis and Symbiogenesis in the Modern and Extended Synthesis

Ideas on symbiology first associated with sociopolitical ideologies and pre-evolutionary thought. After the introduction of natural selection theory, symbiology associated with vitalism, ecology, systems and hierarchy theory, cytoplasmic inheritance research, the biomedical sciences, and insight into the mechanisms of lateral gene transfer (understood as a form of hereditary symbiosis). These fields formed part of the ‘eclipse of Darwinism’ and developed in the margins of the Modern Synthesis that focused on selectionist, vertical-descent theories. From the onset, symbiologists have in addition adhered to holistic, inter- and transdisciplinary stances, that counter the mechanical and reductionist approaches that characterized the division of the sciences at the turn of the twentieth century.

Its early associations with Western socialist thought (including Marxism) is not to be underestimated as a ‘red flag’ for neoliberal sociopolitical and Darwinian thought. In biology, symbiosis and symbiogenesis have often been typified as ‘laws’ of nature that either complement or contradict the ‘laws’ or ‘mechanisms’ of natural selection. Both presumed ‘laws of nature’ have been interpreted either in terms of struggle and competition, or cooperation and socialism, leading to both laws being understood as mutually exclusive. Nonetheless, by emphasizing cooperation and ‘favoring’ symbiosis over competition, symbiology too has, like competitive natural selection theory, been used to justify false believes on eugenetics, racism, hegemony, and national-socialism in order to obtain a ‘higher good.’ Early symbiologists and especially their critics often defined symbiosis in terms of parasitism, or as ‘master–slave’ relations (Sapp, 1994). Scholars such as Kropotkin, Reinheimer, Merezhkowsky, and Wallin understood symbiosis as a natural law necessary for progress, and especially Reinheimer and Merezhkowsky also saw symbiosis as a means for acquiring a ‘higher good,’ a ‘better’ and ‘more cooperative’ society that could be obtained by eugenetics. Merezhkowsky (1920b), for example, saw in symbiogenesis a justification for ethnic cleansing in order to develop a ‘higher’ society where mutualism would only arise amongst a select and chosen group (Sapp et al., 2002).

Though both natural selection theory as well as theories on symbiosis and symbiogenesis find their historical roots in secular, Western sociocultural ideologies, both theories today are decoupled from such sociopolitical references. Nonetheless, the Serial Endosymbiogenetic Theory only became recognized post-synthetically, when molecular (phylo)genetics evidenced its basic morphologically obtained tenets.

Research on both symbiosis and symbiogenesis furthermore introduces new units and levels of evolution, including the superorganism (Spencer, 1876; Wheeler, 1928; Carrapiço, 2015), the holobiont (Margulis and Fester, 1991; Guerrero et al., 2013), symbiome (Sapp, 2003), symbiont (Gontier, 2007), and hologenome (Rosenberg et al., 2007), as well as new means to draw evolutionary phylogenies (Brucker and Bordenstein, 2012), which today designates the rising field of symbiomics (after Sapp, 2003).

Currently, scholars associated with these disciplines are either pleading for an extension of the Modern Synthesis that incorporates the findings of symbiology with those of the Neo-Darwinian paradigm, while others are arguing for, or, a rupture with the latter in favor of a new evolutionary biology. The debates remain unsettled, but it is certain that increased genetic evidence for the symbiogenetic origin of life is causing for symbiosis and symbiogenesis to have finally received the scientific attention they deserve (Figure 9).

Cytoplasmic inheritance

Most of the characters of an individual are governed by nuclear genes, however, some of the traits may be controlled by extra nuclear factors or genes and the inheritance of such traits is known as cytoplasmic inheritance (also called extra-nuclear or extra-chromosomal inheritance). The factors governing the cytoplasmic inheritance are called plasmon or plasmogenes, which are present in the chloroplasts (cp-DNA) or mitochondria (mt-DNA). The plant characters which are inherited by plasmogenes are inherited in uniparental fashion by the female (egg) plant. Therefore, it shows reciprocal differences in the F1 generation. The cytoplasmic male sterility (CMS) is the most extensively studied plant character of practical significance in plant breeding that is governed by plasmogenes located in the mt-DNA, which causes pollen abortion in higher plants. The CMS is observed in over 140 different plant species, and is useful in hybrid seed production of several important crop species such as maize, sorghum, pearl millet, rice, wheat, pigeonpea where it eliminates the process of hand emasculation.

Cytoplasmic Inheritance

After fertilization, the nuclear material of the sperm represents half of the genetic material in the zygote, but cytoplasmic organelles (e.g., plastids and mitochondria) vary between sperm and egg cells. The vast majority of angiosperms display uniparental, maternal cytoplasmic inheritance of plastids and mitochondria, consequently, all of the plastids and mitochondria found in the zygote originate from the egg cell. In some flowering plants, plastids have been detected in the sperm or generative cells (recall that the generative cell divides mitotically to yield the two sperm), representing the possibility of biparental cytoplasmic inheritance. Uniparental, paternal cytoplasmic inheritance, i.e., where the zygote contains mitochondria and plastids derived exclusively from the sperm cell, is rare among angiosperms, but common in many gymnosperms (e.g., Araucariaceae, Cephalotaxaceae, Cupressaceae, and Taxodiaceae).

1.2.7.1 PGC Fate Specification

The germline is the first distinct cell lineage established in the Drosophila embryo when the pole cells, or PGCs, form at the posterior of the embryo during nuclear cycle 8 (Figure 5). PGCs are specified by a unique mechanism that involves cytoplasmic inheritance of RNAs and proteins that are localized to the posterior germ plasm during oogenesis.

The existence of morphologically distinct types of cytoplasm in the posterior region of the egg has been recognized for well over 100 years (review: Mahowald, 1992, 2001). The importance of germ plasm in PGC specification was first described in the beetles Leptinotarsa and Calligrapha, where experimental removal of the posterior oocyte cytoplasm resulted in larvae lacking germ cells (Hegner, 1908, 1911). In Drosophila, cytoplasmic transplantation of posterior pole plasm to the anterior of a recipient egg is sufficient to induce ectopic pole cells that are capable of forming germ cells when implanted to a host posterior (Illmensee and Mahowald, 1974). These important experiments firmly established the role of the germ plasm as the morphogenetic determinant of PGC fate.

Germ plasm contains specialized organelles called polar granules in Drosophila – electron-dense fibrous bodies of RNA and protein found only in the posterior oocyte (or early embryonic) cytoplasm or in germ cells (review: Mahowald, 2001). Polar granules are thought to store and regulate the translation of maternal mRNA required for embryonic germ cell determination (see below). Some of the maternal-effect mutations that block PGC formation, namely capu, spir, osk, stau, valois (vls), tudor (tud), and vasa (vas), result in the disappearance of polar granules, supporting the direct involvement of the respective gene products in PGC specification. Although in most cases the molecular mechanisms by which polar granule components regulate PGC fate remain obscure, some polar granule components appear to have direct roles in defining the PGC lineage.

Polar granule construction begins with the posterior localization of the first components of polar granules, osk mRNA and Stau, and is completed late in oogenesis with the addition of late-localizing mRNAs like nos (review: Mahowald, 2001). The critical role for osk in forming PGCs was demonstrated by experiments in which a chimeric mRNA containing the osk coding region fused to the bcd 3′UTR was ectopically localized to the oocyte anterior. A high level of localized osk activity was thus produced at the anterior pole, sufficient to direct the assembly of ectopic polar granules, and resulting in the formation of functional PGCs at the anterior (Ephrussi and Lehmann, 1992), reminiscent of those produced by transfer of posterior cytoplasm (Illmensee and Mahowald, 1974). Formation of osk-induced ectopic PGCs depends on the function of vas and tud, but not capu, spir, or stau. This indicates that osk, vas, and tud are centrally important for PGC formation, and that the other three genes function indirectly, perhaps through involvement in localizing Osk. Later work confirmed that Stau functions directly in the transport and posterior anchoring of osk mRNA. As discussed in Section 1.2.6.4, capu and spir encode Actin-binding proteins, and mutations in these genes affect several localization processes. Once produced, Osk is phosphorylated by the Par-1 kinase to stabilize it at the posterior (Bullock and Ish-Horowicz, 2002; Riechmann et al., 2002).

Vas is also a key component of polar granules. vas mRNA is abundant and uniformly distributed in the cytoplasm of the nurse cells and the oocyte, but Vas protein accumulates at the posterior pole of the oocyte beginning at early stage 10 (Hay et al., 1990; Lasko and Ashburner, 1990). Osk is required for posterior Vas accumulation, and Osk and Vas can directly interact, suggesting a direct function for Osk in anchoring Vas to the posterior (Breitwieser et al., 1996). Vas arrives a the posterior pole by a mechanism distinct from that of Osk, which is dependent on a second Vas-interacting protein, Gustavus (Gus) (Styhler et al., 2002). Vas is stabilized in the pole plasm through an association with the deubiquitination enzyme Fat facets (Faf), which also localizes to the posterior pole dependent on Osk (Fischer-Vize et al., 1992; Liu et al., 2003).

Vas is a DEAD-box RNA helicase related to the translation factor eIF4A. In vas-null mutants, Grk protein levels are reduced (Styhler et al., 1998; Tomancak et al., 1998). Because Vas can also interact with the general translation factor, dIF2/eIF5B (Carrera et al., 2000), it is a candidate translational activator of grk and, by extension, at least one of the polar granule mRNAs (review: Johnstone and Lasko, 2001). However, a direct role for Vas in regulating translation has not been absolutely established for any mRNA, nor have any specific Vas-binding RNA sequences or structures yet been identified. Prior to its localization at the posterior, Vas accumulates in punctate perinuclear RNP structures, referred to as the nuage (the French word for “cloud”), near the nurse cell nuclear pores. Other polar granule components, such as Tud and Aub, colocalize with Vas in the nuage (Harris and Macdonald, 2001). Prior to its own accumulation into polar granules, Aub is involved in RNA interference-mediated translational repression of unlocalized osk mRNA, potentially implicating the nuage in this process (Kennerdell et al., 2002; Findley et al., 2003). Thus, the nuage may well be the site where mRNAs are assembled with the proteins that control their localization and translation, a critical step in the assembly of the transport and translational control machinery (Section 1.2.6.3).

Less is known about the third central gene implicated in germ cell specification, tud. PGCs do not form in tud mutants, even though some small polar granules have been observed in possibly hypomorphic alleles (Amikura et al., 2001). Tud contains 11 repeated copies of a domain of unknown function, termed the Tudor domain. Two related proteins in mouse, Mouse Tudor Repeat-1 (MTR-1), which is the probable Tudor ortholog, and the Survival of Motor Neuron (SMN) protein, have been implicated in direct protein–protein interactions with Sm proteins, the components of the small nuclear ribonucleoproteins (snRNPs) that are involved in RNA splicing (Selenko et al., 2001; Chuma et al., 2003, and references therein). The snRNPs are assembled in the cytoplasm, but are usually transported to and confined within the nucleus (Will and Luhrmann, 2001). In germline cells of other organisms, Sm proteins colocalize with vasa ortholog. This occurs in the perinuclear nuage of Panorpa nurse cells (Batalova and Parfenov, 2003), in C. elegans P-granules (Barbee et al., 2002), and with mouse Mtr-1 in chromatoid bodies, large RNA and protein complexes that probably represent the mouse nuage (Chuma et al., 2003). The distribution of Sm proteins in Drosophila nurse cells and oocytes has not been reported. Accumulation of Sm proteins in the cytoplasm is unusual, and may reflect an association with translationally silent mRNAs during oogenesis.

In many species including Drosophila, the germ plasm is distinguished not just by the presence of polar granules, but also by a high concentration of mitochondria. Individual mitochondria are often closely juxtaposed with polar granules (Mahowald, 1992, 2001). Surprisingly, two mitochondrial ribosomal RNAs, mtlrRNA and mtsrRNA, are transported from the mitochondria and localized to the surface of polar granules (Kobayashi et al., 1993; Kashikawa et al., 1999). While these RNAs are not readily amenable to genetic studies, several lines of evidence support a function for them in germ cell specification. For example, irradiation of pole plasm with UV-light inactivates it (Geigy, 1931), and this inactivation is relieved by injection of mtlrRNA (Kobayashi and Okada, 1989). Furthermore, injection into the pole plasm of a ribozyme that specifically cleaves mtlrRNA significantly reduces subsequent PGC formation (Kobayashi et al., 1993; Iida and Kobayashi, 1998). Accumulation of mtlrRNA on the polar granules is greatly reduced in tud mutants, leading to the suggestion that Tud might mediate the transport of mitochondrial ribosomal RNAs to the polar granules (Amikura et al., 2001).

1 Cytoplasmic Inheritance

“Omne vivum ex ovo: Every living thing comes from an egg.” As implied in William Harvey’s famous statement, the developmental fate of an embryo begins in the oocyte. The initial phase of embryonic development takes place during a period of genetic transcriptional silence until the activation of the embryonic genome. Prior to embryonic genome activation (EGA), the embryo depends entirely upon maternal RNAs, maternal DNA (in mitochondria), maternal organelles, proteins, substrates, and nutrients that have been deposited in the cytoplasm of the ovum during oogenesis. These maternal products control almost every aspect of early embryonic development. Collectively, they constitute an extraordinary maternal “cytoplasmic” inheritance. Variations in this cytoplasmic inheritance—in the “quality” of the oocyte—can have profound developmental consequences for offspring, both short and long terms. But with the exception of mitochondrial DNA (mtDNA), the effects of cytoplasmic inheritance are not due to offspring having inherited maternal (or paternal) genes.

I intentionally use the expression “cytoplasmic inheritance” in place of the more common “maternal inheritance,” to place the content of this chapter in historical context: in the early twentieth century, the rejection of the existence of cytoplasmic inheritance had important consequences. It represented the triumph of Mendelian inheritance and supported the development of the Modern Synthesis (Amundson, 2005). The view that inheritance is a matter of genes, not cytoplasm, became something of a dogma in genetics. With the denial of cytoplasmic inheritance came a corresponding diminishment of embryology as a scientific discipline (Gilbert, 1998). This dogma persists to this day, most notably in the approach that characterizes contemporary behavioral genetics: inheritance is a matter of the inheritance of alleles, and variation in alleles—generally, single nucleotide polymorphisms—can provide us with clues to understanding variation in complex phenotypes. Hence, embryology can still be ignored because when it comes to inheritance, what matters are genes and genes alone.

It is beyond any doubt today that cytoplasmic inheritance is a key component of human (biological) inheritance. With the growing awareness of the importance of cytoplasmic inheritance has come the resurrection of embryology. At least two contemporaneous forces have led to a growing awareness of the importance of cytoplasmic inheritance/embryogenesis. First is the continued elaboration of the molecular mechanics of epigenetics and the discovery that many of the earliest embryonic processes, those regulated by the maternal cytoplasm, are also epigenetic processes. Second is the global rise of artificial reproductive technologies (ARTs) as a means of conception. Current birth rates of ART-conceived children in a number of developed countries now range from 1% to 3% of all births. Concern about reports of increased risk of negative developmental outcomes among ART-conceived children has focused attention on the early stages of embryonic development prior to implantation. Investigations motivated by this concern have also pointed to the importance of epigenetic mechanisms in preimplantation development (PID). Likewise, research directed at the elaboration of epigenetic mechanisms and possible sources of negative ART outcomes have both pointed to the critical importance of the periconceptual environment(s) in developmental outcomes.

My intent in this chapter is to survey the current state of knowledge of cytoplasmic inheritance. Because what is known about this phenomenon (which is still very little) is enormously complex, I will need both to be selective and to engage in a certain degree of simplification, but hopefully not in a manner that distorts. Part of the elaboration of cytoplasmic inheritance involves a review of some known and conjectured developmental problems associated with ART. My purpose in discussing ART is in no way to pass judgment on the safety of ART procedures or to offer any recommendations regarding their improvement (which I am in no position to do). Rather, I discuss ART because of the insight it provides into PID, i.e., the period prior to the implantation of the blastocyst in the uterus. PID encompasses the period of cytoplasmic regulation of development and the complete activation (by cytoplasmic elements) of the embryonic genome.

Maternal (Non-Mendelian) Influence

In general, maternal influence means the effect that the maternal parent has on the offspring, and in this case we are interested in its effect on germination characteristics of the seeds. The mother plant may affect its seeds by one or more mechanisms: nuclear genetics, non-Mendelian (cytoplasmic) inheritance and/or through interactions with the environment (preconditioning). Some authors include all these mechanisms in discussions of the effects of the maternal parent on seeds (e.g., Roach and Wulff, 1987), but for the sake of increased clarity we have chosen to use a more restricted definition. In this chapter, maternal influence means information that is passed to the offspring in a more or less non-Mendelian fashion.

One way in which the maternal parent can influence offspring is through extranuclear or cytoplasmic inheritance. That is, genetic information in plastids and/or mitochondria is transferred to the offspring at the time of fertilization. Many angiosperms have maternal plastid inheritance (Birky, 1995, 2001Birky, 1995Birky, 2001), but some have both maternal and paternal plastid inheritance (Mogensen, 1996; McCauley et al., 2005, 2007McCauley et al., 2005McCauley et al., 2007; Azhagiri and Maliga, 2007; Pearl et al., 2009). Maternal mitochondria inheritance is known in a few angiosperms and is common in Pinaceae (gymnosperm) (Mogensen, 1996). In Nicotiana tabacum, there is maternal and paternal plastid inheritance, and also paternal mitochondrial DNA is transmitted to the offspring via the pollen (Svab and Maliga, 2007). Progeny from crosses between Passiflora oerstedii and P. retipetala had paternal plastid inheritance, while 12 of 15 progeny of P. costaricensis × P. costaricensis crosses had maternal plastid inheritance and three had maternal and paternal plastid inheritance (Hansen et al., 2007). In Linum usitatissimum, a small-seeded maternal parent had a stronger effect on cytoplasmic inheritance than a large-seeded one (Smith and Fitzsimmons, 1965). Cytoplasmic inheritance also may be important in determining seed size in Oryza sativa (Chandraratna and Sakai, 1960).

Maternal imprinting has been shown to occur in Secale cereale, when two B chromosomes occur in the maternal parent. DNA is replicated later during interphase in B than in regular (A) chromosomes (Jones and Rees, 1969), and transmission of B chromosomes occurs in a non-Mendelian fashion (Puertas et al., 1990). If a mother plant has two B chromosomes, its offspring can be recognized even if they do not have any B chromosomes (Puertas et al., 1990).

A second way in which information is passed from the mother to the offspring is via chemicals produced by the mother. This mechanism of maternal influence has been demonstrated in a number of animal systems and could be one explanation for maternal determination of dormancy in seeds. In the meal moth (Ephestia sp.), a hormone-like substance called kynurenin is involved in pigment synthesis in the skin of the larval stages. A larva with an AA genotype produces kynurenin and is black, while one with an aa genotype lacks kynurenin and is colorless. However, individuals with the aa genotype may have some pigmentation when they are young, but it eventually fades. Pigmentation in young aa individuals is explained by the fact that the mother had at least one A allele, which allowed the production of kynurenin. Thus, some kynurenin was in the cytoplasm of the egg, and it was enough to cause pigment development in the larva (see Strickberger, 1968).

The maternal influence in seed dormancy also may be the result of chemicals produced by the mother plant. Battle and Whittington (1971) suggested that the maternal influence in seed dormancy in Beta vulgaris is due to the formation of germination inhibitors by maternal tissues and that these compounds are transferred to the seeds. Since the genotype of the embryo in Trifolium subterraneum seeds is very important in determining dormancy (presumably physiological dormancy, but seeds also have physical dormancy), Morley (1958) hypothesized that a dormancy-causing chemical was produced in the embryo from substrate provided by the mother plant or that an inhibitor was deposited in the seeds by the mother plant. Higher germination percentages were obtained in hybrids between Morus alba and M. rubra, when M. alba was the maternal parent than when M. rubra was the maternal parent. Genotype of the paternal parent did not have an effect on germination percentages. The strong maternal influence suggested that there was a non-nuclear maternal effect, but the mode of action was not identified (Burgess and Husband, 2004). In seeds such as those of Trifolium subterraneum (Morley, 1958) and Brassica oleracea (Hodgkin and Hegarty, 1978), where the maternal genotype has a strong influence on dormancy, it has been observed that the environment in which the seeds mature may have an effect on dormancy.