Somatic Genome Variation: in Animals, Plants, and Microorganisms

Written by an international team of experts, Somatic Genome Variation presents a timely summary of the latest understanding of somatic genome development and variation in plants, animals, and microorganisms. Wide-ranging in coverage, the authors provide an updated view of somatic genomes and genetic theories while also offering interpretations of somatic genome variation. The text provides geneticists, bioinformaticians, biologist, plant scientists, crop scientists, and microbiologists with a valuable overview of this fascinating field of research.

• Language: English
• Publisher: Wiley
• Release date: Apr 20, 2017
• ISBN: 9781118647127

Book preview

List of Contributors

Jennifer L. Bandura

Biology Department

Roanoke College

Salem

USA

Benoit Bizimungu

Fredericton Research and Development Centre

Agriculture and Agri-Food Canada

Fredericton

New Brunswick

Canada

John R. Bracht

Department of Biology

American University

Washington, DC

USA

Normand Brisson

Department of Biochemistry

Université de Montréal

Montréal

Canada

Gregory G. Brown

Department of Biology

McGill University

Montreal

Canada

Christopher A. Cullis

Department of Biology

Case Western Reserve University

Cleveland

USA

Donglei Du

Quantitative Methods Research Group

Faculty of Business Administration

University of New Brunswick

Fredericton

Canada

Silvia Garagna

Laboratorio di Biologia dello Sviluppo

Dipartimento di Biologia e Biotecnologie

Università degli Studi di Pavia

Pavia

Italy

Andrey Golubov

Department of Biological Sciences

University of Lethbridge

Lethbridge

Canada

Étienne Lepage

Department of Biochemistry

Université de Montréal

Montréal

Canada

Xiu-Qing Li

Fredericton Research and Development Centre

Agriculture and Agri-Food Canada

Fredericton

New Brunswick

Canada

Paola Rebuzzini

Laboratorio di Biologia dello Sviluppo

Dipartimento di Biologia e Biotecnologie

Università degli Studi di Pavia

Pavia

Italy

Carlo Alberto Redi

Laboratorio di Biologia dello Sviluppo

Dipartimento di Biologia e Biotecnologie

Università degli Studi di Pavia

Pavia

Italy

Huaijun Si

Gansu Provincial Key Laboratory of Aridland Crop Science

Gansu Key Laboratory of Crop Genetic and Germplasm Enhancement

College of Life Science and Technology

Gansu Agricultural University

Lanzhou

People's Republic of China

Adrianna Skoneczna

Institute of Biochemistry and Biophysics Polish Academy of Sciences

Laboratory of Mutagenesis and DNA Repair

Warsaw

Poland

Marek Skoneczny

Institute of Biochemistry and Biophysics Polish Academy of Sciences

Department of Genetics

Warsaw

Poland

Jeramiah J. Smith

Department of Biology

University of Kentucky

Kentucky

USA

Samuel Tremblay-Belzile

Department of Biochemistry

Université de Montréal

Montréal

Canada

Sébastien Truche

Department of Biochemistry

Université de Montréal

Montréal

Canada

Tim Xing

Department of Biology

Carleton University

Ottawa

Canada

Éric Zampini

Department of Biochemistry

Université de Montréal

Montréal

Canada

Guodong Zhang

Gansu Provincial Key Laboratory of Aridland Crop Science

Gansu Key Laboratory of Crop Genetic and Germplasm Enhancement

College of Life Science and Technology

Gansu Agricultural University

Lanzhou

People's Republic of China

Norman Zielke

Genome-Scale Biology Research Program

Institute of Biomedicine

University of Helsinki

Helsinki

Finland

Maurizio Zuccotti

Dipartimento di Scienze Biomediche, Biotecnologiche e Traslazionali (SBIBIT)

Università degli Studi di Parma

Parma

Italy

Preface and Introduction

The success of genetic analysis and breeding using Mendel's laws, Johannsen's concepts of genotype and phenotype, Weismann's germ-plasm continuity hypothesis, Morgan's linear arrangement of genes on chromosomes, and Muller's mutation theory leads to a belief—all genes are given by parents and stay the same except for having mutation occasionally caused by external mutagens, somatic cells have no contribution to inheritance, the gene transcript abundance is purely determined by the promoter activity and RNA stability, and clones are genetically identical.

However, as shown in the present book, the somatic genome (the sum total of genetic materials in the cytoplasm and somatic nucleus) actually has environmental and developmental variations; for example: (1) many antibody genes are somatically produced; (2) some ciliate somatic genomes are generated using RNA templates and are therefore phenotypes of the germline genomes but are also the genotypes for many other traits; (3) gene transcription and some endogenous enzymes can induce mutation; (4) certain foods and drinks affect DNA stability and tumor growth; (5) the DNA fingerprint of an animal or plant has certain variations among somatic tissues; (6) various somagenetic and epigenetic variations are transgenerational, and some DNA is translocatable among cells; (7) some genes and repetitive DNA have copy number variation; (8) the chloroplast and mitochondrial genomes vary in ploidy and DNA amount; (9) some DNA sequences are functional through protein binding or DNA-fragment release; (10) DNA damage is sometimes not repaired; (11) clones are genetically mosaic to a certain degree; and (12) the average ploidy level varies among tissues.

This book attempts to assemble the evidence of somatic genome variation in microorganisms, plants, animals, and humans, update various basic concepts in genetics and breeding, determine the implications of somatic genome variation for human health and agriculture, and propose an updated synthesis of inheritance.

Acknowledgments

This book represents the cumulative work of 24 authors from 14 research institutes/universities in six countries. I wish to express my gratitude to all authors who have contributed manuscripts to this book. I thank my Research Centre and my family for their continued support. I would like to extend my very special thanks to the commissioning editor, Mr Justin Jeffryes (the Editorial Director, Natural Sciences, The Americas, at Wiley), Rebecca Ralf, Managing Editor of Life Sciences Books, Ms Divya Narayanan, the primary contact, Ms Bhargavi Natarajan, the former primary contact, Kavitha Chandrasekar, the production editor, Mrs Julie Musk, the copyeditor, and the entire production team at Wiley for their support and high professional standards.

Xiu-Qing Li

About the Editor

Dr Xiu-Qing Li has been a Research Scientist of Molecular Genetics at Agriculture and Agri-Food Canada, Government of Canada (the Fredericton Research and Development Centre, Fredericton, New Brunswick, Canada) since 1997, and an adjunct professor at the University of New Brunswick since 1998. Dr Li received his degree of Docteur d'Etat of France in natural science from Université de Paris-Sud (1987), was an associate professor of biotechnology at Peking University (1987–1993), Invited Professor of Genetics (Université de Paris-Sud, 1988), Research Scientist of Molecular Genetics (Chargée de recherche, level II) at the Centre National de la Recherche Scientifique (CNRS), Gif-sur-Yvette, France (1987), Visiting Professor at Purdue University (1991–1992), and researcher at McGill University (1992–1997). Dr. Li is an Academic Editor of PloS ONE, co-editor of the book Somatic Genome Manipulation (Li et al. eds, Springer, 2015), on the editorial boards of several other journals, and serves as the Communication Director of the Canadian Association of Plant Biotechnology. He is the organizer and chair for the annual Somatic Genome Workshops (San Diego, since 2010) and a co-organizer of the Genome Features and Chromosome Functionality Workshop (since 2016) at the International Plant and Animal Genome Conferences. Dr. Li has extensive research experience and numerous publications on genetics, genomics, RNA polyadenylation, plant somatic genome, plant genetic engineering, plant cell biology, somaclonal variation, chloroplasts, mitochondria, plant male sterility, plant carbohydrate metabolism, potato genetic improvement, bioinformatic analysis, and genome evolution.

Part I

Somatic Genome Variation in Animals and Humans

Chapter 1

Polyploidy in Animal Development and Disease

Jennifer L. Bandura¹ and Norman Zielke²*

¹Biology Department, Roanoke College, Salem, USA

²Genome-Scale Biology Research Program, Institute of Biomedicine, University of Helsinki, Helsinki, Finland

* Corresponding author: [e-mail] Bandura J.L. and Zielke N. (2017) Polyploidy in animal development and disease. In: Li X.-Q., editor. Somatic Genome Variation in Animals, Plants, and Microorganisms: Wiley-Blackwell, Hoboken, NJ, Ch. 1, pp. 3-44.

Abstract

Somatic polyploidization is a developmentally controlled process that can be found in many animal species, including mammals. Polyploidy is utilized as a mechanism to amplify gene expression via an increase in DNA copy number, to drive cellular growth, and to initiate cellular differentiation. Moreover, polyploidy has been implicated in the response to injury or disease. This chapter discusses the processes by which polyploid cells arise and the underlying molecular mechanisms, with a focus on endoreplication in Drosophila and mice. In addition, it summarizes recent progress on the related subjects of endomitosis, ploidy reversal, and gene amplification.

Keywords polyploidy; endoreplication; endomitosis; gene amplification

1.1 Introduction

A fundamental characteristic of the canonical eukaryotic cell cycle is that the genome is only replicated once per cycle, and becomes then equally distributed between two daughter cells during the following mitosis. This strict constraint on DNA replication ensures that each daughter cell receives exactly one diploid set of chromosomes. However, some cell types depart from this rule and differentiate into viable polyploid cells, in which the entire set of chromosomes has been multiplied. While the term polyploid refers to cells with virtually any chromosomal configuration, the term polytene refers to a subclass of polyploid cells in which sister chromatids remain closely associated in parallel arrays. The degree of polyploidy is indicated by the C-value, which represents the DNA content as a multiple of the normal haploid genome. Hence, tetraploid cells have a DNA content of 4 C, while octoploid cells are 8 C. Polyploidy is clearly distinguishable from aneuploidy, which refers to cells containing an aberrant number of chromosomes that is not a multiple of the haploid genome. Polyploidy is often confused with the term re-replication, which describes genome duplications originating from unscheduled initiation of DNA replication during S phase. DNA re-replication is an aberrant event that produces cells characterized by a heterogeneous DNA content, incomplete chromosome duplications, stalled replication forks, and DNA damage. In cells with a compromised DNA damage response, re-replication events can cause genomic instability and thereby promote cancer formation.

Polyploidy occurs in two forms that can be distinguished by their mechanisms of origin: Germline polyploidy refers to genome duplications in the germline, which are often caused by chromosome mis-segregation during meiosis. In this case, all the cells in the resulting progeny will be polyploid and, because the multiplied chromosomes may be inherited by future generations, germline polyploidization is generally thought to promote speciation. In this chapter, we focus on the more common somatic polyploidy, which is often intrinsically connected to cellular differentiation and is therefore also referred to as developmentally programmed polyploidy. Somatic polyploidy occurs frequently in ferns, flowering plants, mollusks, arthropods, amphibians, and fish, but is also found in a few specialized cell types in mammals. Somatic polyploidy is utilized to increase tissue size, to initiate cell differentiation, and to multiply gene copy number to enhance tissue-specific functions. Somatic polyploidy has also been implicated in tissue regeneration and the suppression of tumorigenesis.

1.2 Mechanisms Inducing Somatic Polyploidy

The formation of somatic polyploid cells relies on four distinct mechanisms: cell fusion, acytokinetic mitosis, endomitosis, and endoreplication (Figure 1.1; Table 1.1).

Image described by caption and surrounding text.

Figure 1.1 Mechanisms introducing somatic polyploidy. (A) Nuclear configurations in mitotic and polyploid cell types. C, C-value, multiples of the haploid DNA content arising from different mechanisms. (B) Overview of non-canonical cell cycle variants leading to somatic polyploidy. Hepatocytes (HPC) bypass cytokinesis, and as a consequence this cell cycle variant is called an acytokinetic mitosis. Megakaryocyte polyploidization occurs through endomitosis, which is another abortive mitosis lacking anaphase B, telophase, and cytokinesis. Mouse trophoblast giant cells (TGC) and Drosophila salivary glands (SG) both undergo endocycles, which are completely devoid of M phase. TGCs undergo full genome replications, while SGs exit S phase before completing late replication. (C) Schematic of an under-replicated region, which results from differential firing of replication origins and within the euchromatin usually encompasses 100–400 kb. (D) Genomic organization of amplified loci, which arise from reiterative origin firing, resulting in copy number gradients stretching approximately 100 kb.

Table 1.1 Polyploidization achieved by varying mechanisms across animal species.

Note: This is not a comprehensive list of all polyploid tissues in animals.

a There are 65 nuclei per hypodermis.

b Also undergo ploidy reversal.

c Additional unidentifiedmechanisms.

1.2.1 Cell Fusion

Cell fusion describes the fusion of two cells residing in G1 phase, which results in the formation of a multinucleated cell. A striking example is found during the development of nematodes such as Caenorhabditis elegans, whose hypodermis is a single syncytium that grows through successive cell fusion events (Flemming et al. 2000). Cell fusion also occurs during the fusion of skeletal muscle myoblasts into myotubes, when monocytes differentiate into osteoclasts, and during the formation of syncytiotrophoblasts in the human placenta (Cross 2005).

1.2.2 Acytokinetic Mitosis

Acytokinetic mitosis refers to an abortive cell cycle that includes mitosis but lacks cytokinesis and thereby leads to the formation of multinucleated cells. The stereotypical example of this mechanism happens during post-natal liver development, when the complex process of hepatocyte polyploidization is initiated by an acytokinetic division (Gentric et al. 2012). In addition, many cardiomyocytes undergo acytokinetic mitosis during early post-natal development to create binucleate polyploid cells (Soonpaa et al. 1996; Pandit et al. 2013).

1.2.3 Endomitosis

Endomitosis is a related to acytokinetic mitosis and also characterized by a truncated M phase, but in this case cells abort the cell cycle without completing anaphase or undergoing cytokinesis, resulting in the formation of a single polyploid nucleus. A well-known example of endomitosis takes place in the bone marrow of mice and humans when megakaryoblasts differentiate into polyploid non-proliferating megakaryocytes (Ravid et al. 2002; Bluteau et al. 2009).

1.2.4 Endoreplication

Endoreplication cycles, also referred to as endocycles, are the most extreme deviation from the canonical cell cycle, as endocycling cells execute multiple rounds of DNA replication without any intervening mitosis or cytokinesis (Edgar and Orr-Weaver 2001; Lilly and Duronio 2005; Lee et al. 2009; Ullah et al. 2009a; Ullah et al. 2009b; Fox and Duronio 2013; Zielke et al. 2013). Endocycles generally include Gap (G) phases between each S phase, which is the most obvious distinction from re-replication. By executing consecutive endocycles, cells can dramatically increase their nuclear DNA content and attain ploidies of up to 200,000 C. This mechanism of somatic polyploidy has been intensively studied in the fruit fly Drosophila melanogaster, where virtually all post-mitotic growth is achieved by endoreplication (Edgar and Nijhout 2004). In mammals, endoreplication cycles occur in the extra-embryonic tissue of the placenta, where they are involved in the formation trophoblast giant cells (TGCs) (Hu and Cross 2010). Endoreplication has also been observed in epithelial keratinocytes (Gandarillas 2012) and hepatocytes regenerating from damage after injury or stress (Duncan 2013; Pandit et al. 2013).

1.2.5 Gene Amplification

Gene amplification is a developmentally controlled process that results in massive expansion of specific genomic loci (Calvi and Spradling 1999; Calvi and Spradling 2001; Tower 2004; Claycomb and Orr-Weaver 2005; Calvi 2006). Amplification is achieved by reiterative firing of specialized origins of replication while origins throughout the remainder of the genome are silenced. This mechanism does not result in polyploidy as it only affects certain regions of the genome, but it was included because it represents a striking example of somatic genome variation. The regulation of gene amplification has been most thoroughly analyzed in the follicular epithelium of the Drosophila ovary, in which the chorion genes and a few other loci become amplified to facilitate eggshell formation (Calvi et al. 1998b; Claycomb et al. 2004a; Bandura et al. 2005; Kim et al. 2011). Gene amplification also occurs in the ciliate Tetrahymena thermophila, in which portions of the rDNA locus become amplified about 5000 times (Kapler 1993), as well as the II/9A-1 locus of the fly Sciara coprophila, which undergoes a 17-fold amplification (Gerbi et al. 1993).

1.2.6 Ploidy Reversal

Somatic polyploidy is generally thought to be an irreversible process that is crucial for differentiation into specialized cell types. However, certain polyploid cell types are capable of executing reductive divisions, a process that is also referred to as ploidy reversal. The biological relevance of this process is only poorly understood, but recent work has suggested that ploidy reversal could serve as a mechanism to increase genomic variation of somatic cells and thereby enhance the robustness of tissues and organs. On the other hand, ploidy reversal might represent a mechanism by which tumor cells to gain or lose chromosomes and thereby acquire a competitive advantage favoring tumor growth. Polyploid cells undergoing reductive division are prevalent among plants, but only rarely observed in animals. The existence of polyploid mitosis in insects has been known for decades (Berger 1938; Grell 1946), but remained largely unstudied until recently, when it was reported that the cells of the Drosophila rectal papilla enter mitosis after executing two or more endocycles (Fox et al. 2010). Further examples of insect cells undergoing ploidy reversal are found in the hindgut of the mosquito Culex pipens (Berger 1938; Grell 1946), the larval epidermis of another mosquito, Aedes aegypti (Risler 1959), the epidermis of the tobacco hornworm Manduca sexta (Kato et al. 1987), and the fat body of the kissing bug, Rhodnius (Wigglesworth 1967). Arguably the most well-understood example of ploidy reversal in animals is found in the liver of mice and humans where tetra- or octoploid hepatocytes can revert to the diploid or tetraploid state (Duncan et al. 2010; Duncan et al. 2012b).

1.3 The Core Cell Cycle Machinery

All mechanisms leading to somatic polyploidy, with the exception of cell fusion, are based on non-canonical cell cycles, which involve many of the known cell cycle regulators. Therefore, to ease the discussion of the mechanisms of these variant cell cycles it is imperative to include an overview of the core cell cycle machinery. However, the interested reader can find a more comprehensive review of this topic in Budirahardja and Gonczy (2009) and Nordman and Orr-Weaver (2012).

Cell cycle transitions in mitotic cells are largely regulated by the sequential activation of cyclin-dependent kinase (CDK) complexes. Entry into mitosis relies on Cdk1, which forms complexes with the mitotic cyclins Cyclin A (CycA), Cyclin B (CycB), or Cyclin B3 (CycB3) and the Stg/Cdc25 phosphatase. Progression through mitosis requires proteasomal destruction of several cell cycle proteins, including the mitotic cyclins, which are targeted by the anaphase-promoting complex/cyclosome (APC/C). The APC/C is large multimeric E3 ubiquitin ligase, whose activity depends on the adapter proteins Fizzy (Fzy/Cdc20) and Fizzy-related(Fzr/Cdh1).

M phase is divided into two periods: mitosis, in which the duplicated genome is subdivided into two daughter nuclei and therefore is also referred to as karyokinesis; and cytokinesis, which describes the division of the cytoplasm. Cytokinesis starts with the bundling of microtubules between the separating anaphase chromosomes (Carmena et al. 2012; Fededa and Gerlich 2012). The resulting structure is called the midzone and serves as a scaffold for the assembly of the chromosomal passenger complex, which includes Aurora B kinase, INCENP, and survivin, as well as the centralspindlin complex consisting of MKLP1 and MgcRacGAP. An actin–myosin ring mediates the ingression of the cleavage furrow. The formation and the contraction of the actin–myosin ring rely on the GTPase RhoA, whose activity is regulated by Aurora B and the GTPase-activating protein MgcRacGAP. RhoA promotes the movement of actin filaments by maintaining active myosin II, likely through its association with Anillin, a cytoskeletal scaffold protein which links RhoA, actin, and myosin. After remodeling of the cell cortex, the daughter cells split during the process of abscission, which involves disassembly of the actin–myosin ring and fusion of membranes.

In Drosophila, Cdk2 triggers S phase in conjunction with Cyclin E (CycE), whereas in mammals, Cdk2 associates with both CycE and CycA. Cdk2 activity is restricted by Cyclin-dependent kinase inhibitors (CKIs) of the CIP/KIP family, which consists of only Dacapo (Dap) in flies and p21/Cip1, p27/Kip1, and p57/Kip2 in vertebrates. In addition, the E2F transcription factor serves as an important regulator of G1–S transition. Different E2F family members act as either transcriptional activators or repressors and they control the transcription of many cell cycle genes.

S phase initiation in mitotic cycles is also controlled by the binding and activation of proteins on chromatin at origins of DNA replication. Prior to S phase, pre-replication complexes (pre-RCs) are assembled at DNA replication origins distributed throughout the genome. This process is referred to as ‘origin licensing’ and involves the sequential assembly of the origin recognition complex (ORC), Cdc6, and Cdt1, which together load the Mcm2-7 DNA helicase onto chromatin. Restricting genome duplication to once per cell cycle requires that cells do not reinitiate nuclear DNA replication within regions that have already been replicated until cell division is complete. Pre-RC assembly occurs during anaphase and G1 phase when CDK activity is suppressed. Activation of the pre-RC, by contrast, requires CDK and Dbf4-dependent Cdc7 kinase (DDK) activities, which are up-regulated during the G1 to S phase transition. To prevent premature initiation of DNA replication prior to cell division, both flies and mammals use CDKs and ubiquitin ligases and the inhibitory protein Geminin to inactivate one or more pre-RC proteins.

1.4 Genomic Organization of Polyploid Cells

Early studies of Drosophila polyploid tissues led to the realization that the genome is not entirely replicated during each endocycle S phase, as the heterochromatic regions flanking the centromere were not visible during cytological analyses of polytene chromosomes (Gall et al. 1971) and measurement of DNA content resulted in C-values that were not exact multitudes of the diploid genome (Hammond and Laird 1985a; Hammond and Laird 1985b; Smith and Orr-Weaver 1991). The invention of array-based comparative genomic hybridization (aCGH) allowed the systematic analysis of differential replication, and therefore led to the identification of under-replicated regions in a variety of Drosophila polyploid tissues, including larval salivary glands, fat body, and midgut (Nordman et al. 2011; Sher et al. 2012). Furthermore, these studies demonstrated that under-replication is not restricted only to heterochromatin, but also occurs in non-repetitive gene coding regions. The comparison of aCGH profiles from different Drosophila tissues revealed tissue-specific differences in the sites and the degree of under-replication, suggesting that differential replication may contribute to tissue differentiation. The tissue-specific differences in the extent of under-replication do not appear to correlate with the degree of polyploidization, however. For example, the larval fat body and midgut undergo the same number of endocycles, but differ signficantly in their aCGH profiles.

A crucial question that has yet to be fully answered is how differential replication is achieved at the molecular level. ChIP-chip analysis has been used to detect an enrichment of the repressive H3K27me3 mark in under-replicated regions, while binding of ORC protein was significantly reduced. Furthermore, the under-replicated regions are associated with the protein Suppressor of under-replication (Su(UR)) (Pindyurin et al. 2007). Mutation of Su(UR) results in full replication of the vast majority of under-replicated regions in Drosophila (Belyaeva et al. 1998; Nordman et al. 2011; Sher et al. 2012). However, tethering of Su(UR) to specific sites was insufficient to induce under-replication (Sher et al. 2012). The reduced association of ORC protein with chromatin in under-replicated regions was not restored in Su(UR)-deficient salivary glands, whereas the H3K27me3 mark was absent. Analysis of gene amplification in ovarian follicle cells revealed that loss of Su(UR) accelerates replication fork progression, and thus Su(UR) may reduce the replication efficiency in specific chromosomal regions. Altogether, these data suggest a model in which reduced ORC binding combined with impeded progression of replication forks initiated from distant origins give rise to chromosomal regions that are not replicated and thus are under-represented in Drosophila polyploid tissues.

Surprisingly, a recent study revealed that the genomes of megakaryocytes and TGCs are fully replicated, including heterochromatic sequences (Sher et al. 2013). These results demonstrate that the mechanisms of polyploidization differ between mice and flies, but at the moment we can only speculate about the reasons for this phenomenon. Gene expression profiling revealed that S phase genes are down-regulated in endoreplicating salivary glands, but were expressed at normal levels in TGCs (Maqbool et al. 2010; Sher et al. 2013). The reduced expression of replication factors may slow down S phase progression, and thus possibly cause under-replication of late replicating regions in Drosophila salivary glands, but not in TGCs. However, polyploid MKCs also showed reduced expression of S phase genes, albeit significantly higher than in salivary glands, suggesting that this cannot be the sole explanation. Alternatively, the discrepany could result from differences in genomic organization, as most of the heterochromatin of Drosphila is organized in large blocks, whereas in mice it is distributed throughout the genome.

1.5 Endoreplication: An Effective Tool for Post-Mitotic Growth and Tissue Regeneration

Endoreplication cycles are the best-understood mechanism for inducing somatic polyploidy and are often utilized to increase cell size under conditions prohibiting mitotic divisions (Edgar and Orr-Weaver 2001; Lilly and Duronio 2005; Lee et al. 2009; Zielke et al. 2013). The larval and adult growth of many invertebrate species, including the nematode Caenorhabditis elegans, the tunicate Oikopleura dioica, and the fruit fly Drosophila melanogaster, is largely achieved via endoreplication, and in these organisms cell size is roughly protional to ploidy. For example, a survey among nematode species revealed a correlation between body size and hypodermal ploidy, and thus gave rise to the hypothesis that the degree of polyploidy is a major determinant for the evolution of body size. Inhibition of endoreplication in the hypodermis of C. elegans by treatment with hydroxyurea or mutation of the essential S phase regulators CycE or MCM4 results in dwarfish worms (Lozano et al. 2006; Korzelius et al. 2011), demonstrating that endoreplication is essential for hypodermal growth. Likewise, a striking correlation betweencell size and endocycle progression was observed in the epithelium of Oikopleura, which is entirely comprised of polyploid cells but displays regional difference in the degree of polyploidization (Ganot and Thompson 2002). Many larval tissues of Drosophila, including the salivary glands, fat body, epidermis, midgut, trachea, and Malpighian tubules, are comprised of polyploid cells (Smith and Orr-Weaver 1991). Arguably, the best-studied example are the giant salivary gland cells, which undergo approximately 10 endocycles, resulting in a final ploidy approximating 1035 C (Hammond and Laird 1985b). Starvation experiments in the fat body revealed that endocycle progression in larval tissues is responsive to nutrient availability (Britton and Edgar 1998), underscoring the interdependency of polyploidization and cellular growth.

Endoreplication is also crucial for female germline development in Drosophila, as the nurse cells, which provide the maturating oocyte with mRNA and proteins, attain a final ploidy of approximately 1500 C (Hammond and Laird 1985a), and the somatic follicle cells that surround the egg become 16 C. Loss of the ability to undergo endoreplication in nurse cells results in sterility, indicating that endoreplication is essential for oogenesis (Lilly and Spradling 1996).

Importantly, a recent study also highlights the physiological relevance of polyploidization for cell growth during organ development. The subperineural glia (SPG), which encapsulate the neurons of the Drosophila brain, constitute an important part of the blood–brain barrier (Unhavaithaya and Orr-Weaver 2012). During larval development, the SPGs increase in size by polyploidization, which allows the SPGs to accommodate the growing neuronal mass without undergoing mitotic divisions in order to constantly maintain a barrier around the brain. Inhibition of SPG polyploidization disrupted the blood–brain barrier, demonstrating that polyploidization allows growth without losing tissue integrity. SPG polyploidization occurs partly through endoreplication, but also a large fraction of multinucleated cells was observed, implying that other mechanisms contribute to this process. This strategy may also be employed in vertebrate tissues, for example the trophoblasts that provide a barrier between embryonic and maternal tissues (Brandner et al. 2006), and the polyploid keratinocytes, which seal the skin epithelia (Wang et al. 2004; Gandarillas 2012).

As endoreplication is an effective strategy for post-mitotic growth it is often associated with cellular hypertrophy, which refers to volume increases of organs or tissues due to enlargement of its cells. Compensatory cellular hypertrophy has a pivotal role during liver regeneration (Pandit et al. 2013). After injury or damage the remaining hepatocytes re-enter the endocycle, and the accompanying increase in volume alleviates the loss of cell mass. Likewise, the polyploid cells of the Drosophila follicular epithelium compensate for the loss of defective, out-competed cells by increased endoreplication and cellular hypertrophy (Tamori and Deng 2013).

1.6 Initiation of Endoreplication in Drosophila

1.6.1 Endocycle Entry in Ovarian Follicle Cells

The mitotic-to-endocycle switch has been most extensively studied in ovarian follicle cells of Drosophila, where mitotic and endoreplicating cells are found alongside each other and can be easily distinguished by morphological criteria (Spradling 1993). A single fly ovary is composed of approximately 16 ovarioles, each of which is composed of several egg chambers of sequentially maturing developmental stages (Figure 1.2). Each egg chamber consists of an oocyte and 15 germline nurse cells surrounded by a layer of somatic follicle cells. These follicle cells undergo multiple mitotic divisions, and then undergo three endocycles after stage 6 of oogenesis.

Image described by caption and surrounding text.

Figure 1.2 Signaling pathways controlling cell cycle transitions in Drosophila ovarian follicle cells. (Top) A schematic depicts the development of the Drosophila egg chamber and delineates the points where follicle cells undergo cell cycle transitions. Each egg chamber consists of 16 germline cells, the diploid oocyte (red), and 15 polyploid nurse cells (orange), surrounded by a layer of somatic follicle cells (green). The follicle cells proliferate mitotically until stage 6, and then undergo three endocycles. At stage 10B, all genome-wide DNA replication ceases and amplification initiates at six specific loci. (Bottom left) Activation of the Notch signaling pathway in follicle cells is the major stimulus for the mitotic-to-endocycle transition. Notch signaling blocks mitosis in these cells by activating the transcriptional repressor Hnt. Hnt then represses the expression of Stg, as well as Cut, which triggers the degradation of the mitotic cyclins A and B by the APC/C–Fzr complex. In addition, Notch down-regulates the expression of Dap, preventing it from inhibiting S phase. CoREST is a transcriptional cofactor required for proper Notch signaling. (Bottom right) The endocycle-to-amplification transition requires the inactivation of Notch signaling to allow signaling through EcR. The crucial event downstream of both Notch silencing and EcR activation is increased expression of the transcriptional repressor Ttk69, which is required for the switch to amplification. miR-7 negatively regulates Ttk69 expression and can block amplification when overexpressed. Notch silencing also causes the down-regulation of Hnt expression, which allows Cut to be expressed. It is likely that additional signaling pathways are also involved in this cell cycle transition.

The transition from a mitotic cycle into an endocycle involves re-wiring of the cell cycle regulatory network, which is primarily achieved by eliminating the activity of mitotic cyclin/Cdk complexes by proteasomal degradation of the cyclins or termination of their transcription (Figure 1.2). In endoreplicating follicle cells the mRNAs encoding the mitotic cyclins are constantly expressed, but the cyclin proteins become immediately ubiquitinated by APC/C–Fzr and degraded by the proteasome (Schaeffer et al. 2004). Follicle cells lacking the APC/C activator Fzr accumulated high levels of the mitotic cyclins A and B and failed to execute endocycles (Schaeffer et al. 2004). Although these cells lacking Fzr accumulated high levels of CycA and CycB, the phase of mitotic proliferation was not extended beyond stage 6 of oocyte development. However, extra mitoses were induced in fzr mutants by ectopic expression of the Cdk1 phosphatase Stg, demonstrating that initiation of endoreplication requires both up-regulation of APC/C–Fzr activity and down-regulation of Stg (Schaeffer et al. 2004). In follicle cells, the repression of the CycE/Cdk2 inhibitor Dap also appears to be a component of the mitotic-to-endocycle switch, as ectopic expression of Dap prevented endocycle entry (Shcherbata et al. 2004). Follicle cells lacking Dap function displayed no obvious phenotypes, suggesting that Dap is not required for the follicle cell endocycle (Shcherbata et al. 2004; Hong et al. 2007).

1.6.2 Signaling Pathways Regulating Endocycle Entry in Follicle Cells

A key upstream regulator of the mitotic-to-endocycle transition in Drosophila follicle cells is the Notch signaling pathway (Deng et al. 2001; Lopez-Schier and St Johnston 2001). The oocyte expresses the Notch ligand Delta, which activates the Notch receptor and downstream signaling in the surrounding follicle cells. Deletion of Notch in follicle cells, or germline-specific deletion of Delta, prevented the initiation of endoreplication (Deng et al. 2001; Lopez-Schier and St Johnston 2001). Conversely, experimental activation of Notch signaling resulted in repression of Stg and Dap as well as up-regulation of Fzr (Schaeffer et al. 2004; Shcherbata et al. 2004). Moreover, deletion of Delta from the germline resulted in accumulation of Dap, demonstrating that activation of Notch signaling represses Dap in endoreplicating follicle cells (Shcherbata et al. 2004).

Notch mediates the mitotic-to-endocycle transition in follicle cells through a cascade of transcription factors (Figure 1.2). Based on multiple lines of evidence, Sun and Deng proposed that Notch signaling activates the transcriptional repressor Hindsight (Hnt), which subsequently suppresses the expression of Stg/Cdc25 as well as another transcription factor, Cut (Sun and Deng 2005; Sun and Deng 2007). The down-regulation of Cut at the mitotic-to-endocycle transition allows Fzr to accumulate, which in turn mediates the proteasomal degradation of CycA and CycB. It remains unclear, however, whether Fzr and Stg are direct targets of these transcription factors, and whether Hnt and Cut mediate the Notch-dependent down-regulation of Dap. In summary, these data suggest that endocycle entry in follicle cells occurs in two steps: First, the Hnt-mediated down-regulation of Stg arrests the cell at the G2/M transition, and then the down-regulation of Cut and subsequent de-repression of Fzr bypasses mitosis and pushes the cell directly into a G1-like state (Sun and Deng 2007).

A recent study has also identified the transcriptional cofactor CoREST as a positive modulator of Notch signaling during the mitotic-to-endocycle switch in follicle cells (Domanitskaya and Schupbach 2012). Notch signaling is faulty in CoREST mutant cells, which fail to up-regulate Hnt and down-regulate Cut and do not switch from the mitotic cycle to the endocycle on schedule. The authors of this study found that CoREST regulates Notch signaling downstream of Notch receptor proteolytic cleavage but upstream of Hnt expression. Further, CoREST mutant follicle cells experience increased histone H3K27 tri-methylation and H4K16 acetylation at the stage when they should be beginning endoreplication, suggesting that CoREST may epigenetically modify Notch target gene loci and thereby affect the transcription of those genes.

Intriguingly, the Notch pathway may also be important for the mitotic-to-endocycle transition in the ovarian follicle cells of another insect species, the beetle Tribolium castaneum (Baumer et al. 2012). Tribolium Notch is expressed ubiquitously in both germline and somatic cells of the ovary, and Delta, similar to Drosophila, is expressed only in the germline. However, in Tribolium ovaries that have been treated with either Notch or Delta RNAi, it appears that mitotic cycles switch prematurely to endocycles, indicating that Notch signaling in this context has the opposite effect to that in Drosophila in that it maintains mitotic proliferation.

Are other pathways besides Notch signaling involved in the mitotic-to-endocycle transition? Follicle cells mutant for patched (ptc), a negative regulator of the signaling molecule Hedgehog (Hh), fail to initiate endocycles, suggesting that the Hh pathway may antagonize the endocycle-promoting Notch signal (Zhang and Kalderon 2000). Sun and Deng (2007) have proposed a mechanism in which Hnt terminates the phase of mitotic proliferation by suppressing the transcription of the Hh-effector cubitus interruptus (Ci). Furthermore, genetic interaction experiments suggested that down-regulation of Ci by Hnt also requires the continuously expressed Zn-finger transcription factor Tramtrack69 (Tkk69) (Jordan et al. 2006; Sun and Deng 2007).

Additionally, the Salvador–Warts–Hippo (SWH) tumor-suppressor pathway has been demonstrated to play an important role in the entry into endocycles specifically in the posterior follicle cells, which are important for body axis formation (Meignin et al. 2007; Polesello and Tapon 2007; Yu et al. 2008). The core of the SWH pathway consists of the serine/threonine kinases Hippo (Hpo) and Warts (Wts) and the scaffold proteins Salvador (Sav) and Mats, which function to inactivate the transcription factor Yorkie (Yki). In the posterior follicle cells, mutation of hpo, wts, or sav, or overexpression of yki, disrupts Notch signaling. As a result, these cells do not express Hnt or repress the expression of Cut after stage 6 and they continue to divide mitotically.

1.6.3 Endocycle Entry in Other Tissues

The mechanisms of endocycle initiation have also been studied in the Drosophila salivary gland, which is technically challenging as the transition to the endocycle occurs during embryogenesis and involves only a relatively small number of cells (Smith and Orr-Weaver 1991). As in follicle cells, endocycle entry in salivary glands requires APC/C–Fzr-mediated degradation of mitotic cyclins (Sigrist and Lehner 1997), but in this case it also involves the transcriptional down-regulation of mitotic regulators (Zielke et al. 2008; Maqbool et al. 2010).

Recent work has demonstrated that suppression of mitotic regulators in endoreplicating salivary glands involves E2F-mediated repression (Zielke et al. 2011). The relatively small fly genome contains only a single activator E2F, E2F1, and a single repressor, E2F2, which both associate with the same dimerization partner, namely dDP (van den Heuvel and Dyson 2008). Salivary glands derived from E2F2 mutants, or from animals overexpressing E2F1, displayed ectopic expression of Cdk1, CycA, and CycB3 mRNAs (Zielke et al. 2011). Cdk1 and CycA have been identified as E2F targets in mitotic cells (Ishida et al. 2001; Ren et al. 2002), suggesting that E2F2 may act as a selectivity factor, preventing E2F1 from activating mitotic targets during endoreplication.

The decision to switch to endoreplication cycles also involves the Escargot (Esg) protein, which belongs to the family of Snail transcription factors. Esg is expressed in most imaginal tissues, but is absent in the polyploid tissues of the larvae (De Miglio et al. 1999; Polesello and Tapon 2007). Abdominal histoblasts lacking Esg become polyploid, whereas ectopic expression of Esg abolishes endoreplication in larval salivary glands (Fuse et al. 1994; De Miglio et al. 1999). The esg phenotype resembles that of Cdk1 mutants, which also undergo ectopic endocycles, and further experiments revealed that Esg indeed genetically interacts with Cdk1 (Stern et al. 1993; Hayashi 1996; Weigmann et al. 1997). Tellingly, the abdominal histoblasts in hypomorphic esg mutants showed reduced levels of CycA protein (Hayashi 1996). These findings suggest that Esg is required in diploid cells to sustain mitotic Cyclin/Cdk activity and that its function is dampened when cells switch to endoreplication cycles. Altogether, evidence from several Drosophila cell types indicates that the suppression of mitotic regulators upon entry into endocycles involves multiple mechanisms that act at both the transcriptional and post-transcriptional level.

1.7 Mechanisms of Endocycle Oscillations in Drosophila

1.7.1 An Autonomous Oscillator Drives Endocycling in the Salivary Gland

Endoreplicating cells utilize the same set of S phase genes as mitotic cells, but upon endocycle entry the underlying regulatory network becomes reorganized to adjust to the fact that mitosis is suppressed. The expression of many genes involved in DNA replication culminates right before S phase, and relies, as in all higher eukaryotes, on the activity of the transcriptional activator E2F1 (Duronio et al. 1995). Deletion of E2F1 impairs the polyploidization of larval tissues, whereas mild overexpression of E2F1 promotes endoreplication in the salivary gland, resulting in hyperpolyploidy (Royzman et al. 1997; Zielke et al. 2011). Like in mitotic cells, E2F1 protein accumulates during G phase, but is rapidly destroyed upon initiation of S phase (Asano et al. 1996; Weng et al. 2003; Reis and Edgar 2004). The S phase-specific degradation of E2F1 is mediated by the CRL4-Cdt2 E3 ligase and relies on the fact that activation of CRL4-Cdt2 requires chromatin-bound PCNA, which exists only at active replication forks (Shibutani et al. 2007; Shibutani et al. 2008; Havens and Walter 2009; Zielke et al. 2011). Substrates of CRL4-Cdt2 are characterized by the PCNA-interacting protein (PIP) motif, which is crucial for the association with PCNA (Arias and Walter 2006; Havens and Walter 2009). Consequently, mutation of the N-terminal PIP box in E2F1 inhibits its turnover (Shibutani et al. 2008). Tellingly, either depletion of CRL4-Cdt2 or stabilization of E2F1 by deletion of the PIP box blocks endocycle progression in salivary glands (Zielke et al. 2011). Altogether, these data demonstrated that endocycling in Drosophila salivary glands relies on an autoregulatory feedback loop comprised of E2F1 and CRL4-Cdt2, and thus provides the basis for the periodic accumulation of S phase genes (Figure 1.3).

A plot and flow diagram each for endocycle oscillations of (a) Fly SG and (b) Mouse TGC. The plots have three curves plotted and labeled for E2F1, CycE, and Geminin (a) and p57, CycE, and CycA Geminin (b).

Figure 1.3 Endocycle oscillations. (A) Model of the autonomous oscillator that drives the salivary gland endocycles. The transcriptional activator E2F1 accumulates during the late G(ap) phase and stimulates CycE transcription, which triggers S phase in conjunction with its kinase partner Cdk2. Ongoing DNA replication activates the E3 ubiquitin ligase CRL4-Cdt2, which marks E2F1 for proteasomal degradation. The CRL1–Ago complex continuously targets CycE for degradation, and hence this negative feedback loop causes a drop in CycE levels, which is a prerequisite for origin licensing during the next G phase. CycE/Cdk2 also inhibits APC/C–Fzr activity, thereby allowing accumulation of the licensing inhibitor Geminin. (B) The regulatory network that controls the endocycle progression in trophoblast giant cells (TGC). S phase in TGCs relies on Cdk2, which associates with either CycE or CycA. During late G phase a peak of CycE/Cdk2 activity triggers S phase and concomitantly inhibits the ACP/C–Fzr complex, thereby promoting the accumulation of CycA and Geminin. CycA/Cdk2 phosphorylates CycE, which in turn will be recognized by E3 ligase CRL1-Fbw7 and thus marked for proteasomal degradation. The cessation of CycE/Cdk2 activity permits accumulation of CKI p57, which is no longer antagonized by the CRL1–Skp2 complex. The synergistic action of three factors, CRL1-Fbw7, CRL3, and p57, ensures low levels of CycE/Cdk2 activity during G phases, thereby allowing origin licensing to occur.

E2F1 is essential for both the expression and oscillation of CycE protein in Drosophila during both mitotic cell cycles and endocycles (Duronio and O'Farrell 1995; Duronio et al. 1996). In CycE-deficient embryos, DNA synthesis is abolished in cells undergoing either mitotic cell cycles or endocycles (Knoblich et al. 1994), and endocycles in salivary glands are eliminated by tissue-specific removal of Cdk2/cdc2c (Zielke et al. 2011). CycE is an unstable protein (Lilly and Spradling 1996; Weng et al. 2003), whose proteasomal degradation relies on the E3 ubiquitin ligase CRL1-Ago and the specificity factor minus (mi) (Moberg et al. 2001; Szuplewski et al. 2009). So far, there is no evidence that CRL1-Ago activity oscillates in endoreplicating cells. Therefore its main function may be to render CycE constitutively unstable, rather than acting as part of the endocycle core oscillator. Salivary glands and follicle cells deficient for either ago or mi and salivary gland cells continuously overexpressing CycE fail to undergo DNA replication (Follette et al. 1998; Weiss et al. 1998; Shcherbata et al. 2004; Szuplewski et al. 2009; Zielke et al. 2011), implying that endocycle progression depends on the periodic accumulation and degradation of CycE/Cdk2 activity. The inhibitory effect of CycE is unique to endoreplicating tissues, as ectopic expression of CycE did not interfere with DNA synthesis in mitotically proliferating wing discs (Neufeld et al. 1998) or follicle cells undergoing gene amplification (Calvi et al. 1998a).

A critical CycE/Cdk2-target is the APC/C–Fzr complex, whose activity is essential for endocycle progression in salivary glands and follicle cells (Narbonne-Reveau et al. 2008; Zielke et al. 2008). High levels of CycE/Cdk2 activity trigger S phase and concomitantly inhibit APC/C–Fzr activity (Sigrist and Lehner 1997; Reber et al. 2006), thereby promoting the accumulation of Geminin and Orc1. Low CycE/Cdk2 activity during G phase, by contrast, releases the APC/C–Fzr complex, which subsequently degrades Geminin and thus allows assembly of pre-RCs. Consistently, it was found that depletion of APC/C–Fzr activity stabilizes Geminin and blocks endocycle progression (Zielke et al. 2008). Although Geminin is essential for cell proliferation in flies (Quinn et al. 2001), it appears to be dispensable for normal endoreplication in salivary glands, suggesting that CycE/Cdk2 prevents re-replication in endoreplicating salivary glands through multiple redundant mechanisms (Zielke et al. 2011).

1.7.2 Alternative Modes of Endoreplication

Currently, the autoregulatory feedback loop based on E2F1 and CRL4-Cdt2 has only been demonstrated in the Drosophila salivary gland, but considering the diversity of species undergoing endocycles, it seems likely that a number of variant mechanisms have evolved. Interestingly, the mechanisms of endoreplication appear to vary even among Drosophila tissues. The endocycles of the ovarian nurse cells represent a notable exception, as these cycles involve the CKI Dap, which is dispensable in most other endoreplicating tissues including salivary glands, follicle cells, and socket and shaft cells of the bristle lineage (Shcherbata et al. 2004; Audibert et al. 2005; Zielke et al. 2011). In endoreplicating nurse cells, Dap accumulates only in cells expressing high levels of CycE/Cdk2, which leads to a subsequent decrease of CycE/Cdk2 activity (de Nooij et al. 2000). Late replicating heterochromatic sequences are under-represented in many polyploid Drosophila tissues, including the ovarian nurse cells (Nordman et al. 2011; Sher et al. 2012), but in Dap-deficient nurse cells where CycE/Cdk2 activity is extended, S phase is lengthened and these sequences are nearly fully replicated (Hong et al. 2007). It has been hypothesized that the negative feedback loop between CycE/Cdk2 activity and Dap creates a window of low CDK activity during the next G phase, and thus allows pre-RC formation. However, Dap-mutant nurse cells can undergo fairly normal endocycles (Hong et al. 2003), suggesting that redundant mechanisms safeguard pre-RC formation during the Gap phase and that Dap is only required for ensuring proper replication timing.

The endoreplicating cells of the mechanosensory bristle lineage employ another endocycle variant that involves oscillations of CycA protein (Salle et al. 2012). The presence of CycA in endoreplicating shaft and socket cells is surprising, as the expression of many mitotic regulators, including CycA, is dampened in salivary glands (Zielke et al. 2008; Maqbool et al. 2010). In shaft and socket cells, CycA protein specifically accumulates during late S phase, and both loss and gain of function results in decreased ploidy, delayed S phase, and dramatically affected association of Orc2 with heterochromatic regions (Salle et al. 2012), suggesting that CycA/Cdk1 regulates Orc2 localization and thereby modulates replication timing in endoreplicating bristle cells.

1.8 Gene Amplification in Drosophila Follicle Cells

1.8.1 Molecular Mechanism of Gene Amplification

During Drosophila oogenesis, six loci within the ovarian follicle cells undergo developmentally programmed DNA amplification by repeated firing of origins of replication. Amplification at these sites is required to meet the demand for high levels of expression of proteins needed for eggshell synthesis. Indeed, a reduction in amplification leads to thin eggshells and female sterility. These six loci are known as Drosophila amplicons in follicle cells (DAFCs) (Claycomb et al. 2004b; Kim et al. 2011), and they are named by their cytological positions in the genome. The amplicons DAFC-7F and DAFC-66D correspond to clusters of the chorion (eggshell) genes on the X and 3rd chromosomes, respectively (Spradling and Mahowald 1980; Claycomb et al. 2004a). Several genes expressed in the follicle cells of late-stage egg chambers are encoded in the DAFC-30B and DAFC-62D amplified loci, including the gene yellow-g, which is likely required for proper vitelline membrane formation (Claycomb et al. 2004a). The genomic regions amplified in DAFC-22B and DAFC-34B have been identified most recently, and DAFC-34B contains the gene Vm34Ca that encodes a structural component of the vitelline membrane (Kim et al. 2011).

DAFC-66D, the amplicon that achieves the highest increase in DNA copy number, begins to amplify periodically during the asynchronous S phases that occur during the three follicle cell endocycles, attaining an approximately four-fold increase in copy number by the end of endoreplication (Calvi et al. 1998a). All six amplified loci then synchronously enter into continuous amplification at stage 10B, immediately after endocycles have completed (Calvi et al. 1998a). The maximal increases in DNA copy number that have been observed by quantitative real-time PCR are 30-fold for DAFC-66D, 14-fold for DAFC-7F, and 4- to 6-fold for DAFC-30B, DAFC-62D, DAFC-22B, and DAFC-34B (Claycomb et al. 2002; Claycomb et al. 2004a). The amplified region at each locus is a gradient with the highest DNA copy number near the origins of replication and each spans in total 75–100 kb (Spradling and Mahowald 1980; Spradling 1981; Claycomb et al. 2004a; Kim et al. 2011). For DAFC-7F, DAFC-66D, and DAFC-30B, all DNA replication initiation events have completed by stage 10B or 11, and after that, further increases in DNA copy number only occur at greater distances from the origins by elongation by existing replication forks (Claycomb et al. 2002; Claycomb et al. 2004a). In contrast, DAFC-62D undergoes a late replication initiation event between stages 12 and 13, resulting in a relatively small maximally amplified region (Claycomb et al. 2004b).

Amplification can be visualized by a punctate pattern of BrdU incorporation in follicle cells from stage 10B on (Calvi et al. 1998a). In addition, amplifying loci can be visualized by immunostaining for several proteins that bind DNA at origins of replication, such as ORC1, ORC2, ORC5, CDC45, DUP/Cdt1, PCNA, and the MCM2-7 complex, as the factors known to function at all origins of DNA replication also function at the DAFCs (Asano and Wharton 1999; Austin et al. 1999; Royzman et al. 1997; Loebel et al. 2000; Whittaker et al. 2000; Claycomb et al. 2002). As further evidence that DNA replication factors are required for amplification of the DAFCs, hypomorphic mutations in genes encoding necessary DNA replication proteins such as double-parked (dup/cdt1), origin recognition complex subunit 2 (orc2), chiffon (chif, dbf4-like), proliferating cell nuclear antigen (pcna, mus209), minichromosome maintenance factor 6 (mcm6), and humpty dumpty (hd, fs(3)272) result in reduced amplification, thin eggshells, and female sterility (Underwood et al. 1990; Landis et al. 1997; Landis and Tower 1999; Henderson et al. 2000; Whittaker et al. 2000; Schwed et al. 2002; Bandura et al. 2005). Moreover, amplification is under the control of the S phase regulators E2F and CycE/Cdk2. The levels of both E2F1 protein and CycE/Cdk2 activity oscillate in mitotic and endoreplicating follicle cells, but coincident with the onset of amplification both markers become evenly distributed specifically within the follicle cells undergoing amplification (Calvi et al. 1998a; Sun et al. 1997). Either inhibition of CycE/Cdk2 with Dap or mutation of E2f1, E2f2, Dp, or Rbf selectively disrupts amplification (Calvi et al. 1998b; Royzman et al. 1997; Bosco et al. 2001; Cayirlioglu et al. 2001).

If the same factors required for DNA replication at all origins are also utilized at the chorion origins and other DAFCs during amplification, what instructs the amplifying origins to fire repeatedly while all other origins in the genome remain silent? The answer to this question has not been fully elucidated, but epigenetic regulation appears to be part of the explanation. For example, it has been demonstrated that histone acetylation at the DAFCs is crucial for activity of these origins during amplification. Hyperacetylation of histones H3 and H4 has been observed specifically at the DAFCs in amplification-stage follicle cells (Aggarwal and Calvi 2004; Hartl et al. 2007; Liu et al. 2012). Further, increasing the levels of acetylation throughout follicle cell nuclei by either mutating or inhibiting histone deacetylases (HDACs) resulted in extra genomic DNA replication during amplification stages (Aggarwal and Calvi 2004). Consistent with these data, tethering a histone acetyltransferase (HAT) to an amplification reporter containing the chorion origin at DAFC-66D increased amplification, whereas tethering an HDAC reduced it. A recent study has extended these results by identifying two HATs, Chameau/HBO1 and CBP/Nejire, that bind to DNA at DAFC-66D and contribute substantially to the acetylation seen at all DAFCs (McConnell et al. 2012). In addition, the Myb-MuvB (MMB) or dREAM complex binds to the regulatory DNA sequences of at least one of the DAFCs and constrains genomic DNA replication during amplification (Beall et al. 2002; Beall et al. 2004; Beall et al. 2007). MMB/dREAM recruits chromatin modifiers and is believed to act as either an activator or repressor of replication, depending on developmental context and genomic location (Beall et al. 2002; Beall et al. 2004; Korenjak et al. 2004; Lewis et al. 2004; Beall et al. 2007; Georlette et al. 2007). This is consistent with earlier work that had demonstrated roles for both E2F2 and Rbf in restraining genomic DNA replication in amplification-stage follicle cells, as both E2F2 and Rbf are members of MMB/dREAM (Cayirlioglu et al. 2001; Cayirlioglu et al. 2003).

1.8.2 The Endocycle-to-Amplification Switch

It is currently only poorly understood how the switch from endocycles to amplification is regulated. Continuous Notch activation suppresses gene amplification in follicle cells, which instead execute an additional round of endoreplication, suggesting that down-regulation of Notch signaling is a prerequisite for the endocycle-to-amplification switch (Sun et al. 1997). In addition, the endocycle-to-amplification switch coincides with the up-regulation of the insect hormone ecdysone, and ectopic expression of a dominant-negative ecdysone receptor (EcR) prevents the proper initiation of gene amplification (Sun et al. 1997). Thus both down-regulation of Notch and activation of EcR are essential for the endocycle-to-amplification switch (Figure 1.2). Increased expression of the transcriptional repressor Ttk69 at the endocycle-to-amplification transition appears to be the crucial event downstream of Notch down-regulation and EcR activation, and expression of Ttk69 at this transition is further modulated by the microRNA miR-7 (Sun et al. 1997; Huang et al. 2013). The combination of Notch down-regulation and activation of EcR, or direct up-regulation of Ttk69, were not sufficient to induce a premature transition to amplification, and so other signaling pathways are likely involved in this process (Sun et al. 1997). Further, the mechanism by which these signaling events may lead to chromatin modifications specifically at the DAFC origins has yet to be illuminated.

1.9 Endocycle Entry in the Trophoblast Lineage

TGCs are the paradigm for endoreplication in mammals and contain, like the endoreplicating tissues of Drosophila, large polytene chromosomes with DNA contents up to 512 C (Barlow and Sherman 1974; Varmuza et al. 1988; Sher et al. 2013). TGCs are derived from the trophectoderm of blastocysts, the outer layer of epithelial cells that give rise exclusively to the various cells comprising the placenta. The large size of the TGCs facilitates their function as a barrier between the maternal blood supply and the embryo proper (Rossant and Cross 2001; Cross 2005; Hu and Cross 2010). TGCs originate from trophoblast stem cells (TSCs) and initiate endoreplication without undergoing mitotic divisions in the absence of fibroblast growth factor 4 (FGF4) (Tanaka et al. 1998). TSCs can be maintained in culture and differentiated into endoreplicating TGCs, but under these conditions their DNA content rarely exceeds 64 C (Tanaka et al. 1998). Although cultivated TGCs only execute a limited number of endocycles, they are currently the vertebrate cell type most suitable for in vitro studies on endoreplication.

Similar to endocycle entry in Drosophila, the onset of endoreplication in TGCs requires the suppression of mitotic CDK activity, while the activity of S phase Cdks is not affected (Ullah et al. 2008). Endoreplication in TSCs is triggered by withdrawal of FGF4, which induces the expression of the CKIs p57 and p21 and thereby prevents the activation of Cdk1 and initiation of mitosis (Ullah et al. 2008). The switch to endoreplication also involves Checkpoint Kinase-1 (Chk1), a kinase that is part of the DNA damage checkpoint (Ullah et al. 2011). In proliferating TSCs, Chk1 phosphorylates and thereby targets p57 and p21 for proteasomal degradation. The expression of Chk1 ceases when FGF4 is withdrawn from TSCs, thus allowing unphosphorylated p57 and p21 proteins to accumulate and the subsequent switch to endoreplication. However, only p57 is essential for the mitotic to endocycle transition, as FGF4 deprivation induces endoreplication in p21-deficient TSCs, whereas p57-depleted TSCs undergo acytokinetic mitosis instead (Ullah et al. 2008). Treatment with the Cdk1 inhibitor RO3306 induces endoreplication in wild-type and p57-deficient TSCs, demonstrating that the main function of p57 is to constrain Cdk1 activity and that its activity is negligible after the onset of endoreplication (Hochegger et al. 2007; Ullah et al. 2008). Examination of p57 knockout mice revealed multiple developmental abnormalities, including placentomegaly and hyperplasia of labyrinthine trophoblasts and spongiotrophoblasts, but the DNA content of the TGCs was not significantly changed (Zhang et al. 1998; Takahashi et al. 2000; Kanayama et al. 2002), suggesting that trophoblasts can initiate endoreplication by alternative, p57-independent mechanisms. Knockout of the APC/C activator protein Cdh1 results in embryonic lethality due to defective endoreplication in extraembryonic tissues (Garci-Higuera et al. 2008; Li et al. 2008), and a recent study revealed that deletion of Cdh1 prevents the onset of endoreplication in cultivated TSCs,