Nutritional plasticity and evolutionary divergence in the Drosophila ovary

Cláudia Carolina de Almeida Mendes Dissertation presented to obtain the Ph.D degree in Developmental Biology Instituto de Tecnologia Química e Biológica António Xavier | Universidade Nova de Lisboa Oeiras, May 2015 Nutritional plasticity and evolutionary divergence in the Drosophila ovary Cláudia Carolina de Almeida Mendes Dissertation presented to obtain the Ph.D degree in Developmental Biology Instituto de Tecnologia Química e Biológica António Xavier | Universidade Nova de Lisboa Oeiras, May 2015 Nutritional plasticity and evolutionary divergence in the Drosophila ovary Research work coordinated by: Mendes, C.C. Nutritional plasticity and evolutionary divergence in the Drosophila ovary PhD thesis, Instituto Gulbenkian de Ciência, Universidade Nova de Lisboa, 2015 In English, with summary in Portuguese This thesis has been scanned for plagiarism on April 29th 2015 and there was no conflict with published works. Cover image by the author © 2015 Drawings of adult ovaries of Drosophila. Printed by 2002 Estúdio Gráfico, Lda., Odivelas, Portugal. Licensed under creative commons (BY-SA-NC) 3.0 This thesis is dedicated to My parents, António Mendes and Luciana Cruz, for always encouraging my creativity and curiosity. Declaração/Declaration Declaro que esta dissertação é o resultado do meu próprio trabalho desenvolvido entre Abril de 2011 e Janeiro de 2015 no laboratório da Dra. Christen Mirth, Instituto Gulbenkian de Ciência em Oeiras, Portugal, com co-orientação do Dr. Élio Sucena do Instituto Gulbenkian de Ciência, Oeiras, Portugal. Este doutoramento foi realizado no âmbito do Programa Gulbenkian de Doutoramento (edição 2010-2011). Parte do capítulo 1 foi publicado no Frontiers in Physiology como “Mechanisms regulating nutrition-dependent developmental plasticity through organspecific effects in insects”, T. Koyama, C.C. Mendes and C.K. Mirth (2013). O capítulo 2 e 3 estão integrados num manuscrito submetido para publicação com autoria de C.C. Mendes e C.K. Mirth. O capítulo 4 está integrado num manuscrito em preparação com autoria C.C. Mendes, E. Sucena e C.K. Mirth. I declare that this dissertation is a result of my own research carried out between April 2011 and January 2015 in the laboratory of Dr. Christen Mirth, Instituto Gulbenkian de Ciência in Oeiras, Portugal, with the co-supervision of Dr. Élio Sucena of the Instituto Gulbenkian de Ciência, Oeiras, Portugal. Part of chapter 1 has been published in Frontiers in Physiology entitled “Mechanisms regulating nutritiondependent developmental plasticity through organ-specific effects in insects”, T. Koyama, C.C. Mendes and C.K. Mirth (2013). Chapter 2 and 3 are part of a manuscript submitted for publication, authored by C.C. Mendes and C.K. Mirth. Chapter 4 is part of a manuscript in preparation, authored by C.C. Mendes, E. Sucena and C.K. Mirth. Apoio Financeiro/Financial Support Esta dissertação teve o apoio financeiro da Fundação para a Ciência e a Tecnologia, bolsa de doutoramento #SFRH/BD/51624/2011 e da Fundação Calouste Gulbenkian. This dissertation had the financial support from Fundação para a Ciência e a Tecnologia, doctoral fellowship #SFRH/BD/51624/2011 and Fundação Calouste Gulbenkian. This thesis was written in L A TEX. All images were made with Inkscape, ImageJ (NIH) and Adobe Photoshop (Adobe Systems). LIST OF FIGURES 1.1 The relantionship between phenotype and the environment. 7 1.2 The IIS pathway in D. melanogaster.............. 11 1.3 The activation and derepression functions of ecdysone signalling............................. 13 1.4 Organs differ in their sensitivity to nutrition. . . . . . . . . 15 2.1 Changes in nutrition during the first phase of sensitivity have greater effects in ovariole number than in the second phase of sensitivity. . . . . . . . . . . . . . . . . . . . . . . . 29 2.2 Ovariole number is positively correlated with early female fecundity. ............................ 31 2.3 Ovary development during L3 larval stages under optimal nutritional conditions. . . . . . . . . . . . . . . . . . . . . . 32 2.4 Distinct stage-specific developmental processes during ovary development are regulated by nutrition. . . . . . . . . 34 2.5 TF formation and ovary growth respond differently to preand post-critical weight nutrition. . . . . . . . . . . . . . . . 36 2.6 Critical weight separates two phases of sensitivity to nutrition in ovariole number. . . . . . . . . . . . . . . . . . 37 2.1 Control test for En antibody. . . . . . . . . . . . . . . . . . 40 2.2 Larval nutrition affects development time and female weight. 41 3.1 Experimental design of nutritional manipulations. . . . . . . 48 3.2 traffic jam-GAL4 is expressed in ovarian somatic cells during L3 larval stages. . . . . . . . . . . . . . . . . . . . . 51 v 3.3 traffic jam-GAL4 is expressed in neuroblasts and glial cells in the larval brain. . . . . . . . . . . . . . . . . . . . . . . . 52 3.4 Manipulating IIS or ecdysone signalling in the larval ovary reduces adult ovariole number and female weight. . . . . . . 54 3.5 Role of IIS during ovary development. . . . . . . . . . . . . 56 3.6 Role of ecdysone signalling during ovary development. . . . 58 3.7 Feeding wild-type larvae with 20E-supplemented sucroseonly food increases TF number and ovary volume. . . . . . 60 3.8 The interplay between IIS and ecdysone signalling pathways. 62 4.1 Distribution of the four D. mojavensis subspecies. . . . . . 73 4.2 Ovariole number in females of D. mojavensis wrigleyi is reduced relative to other D. mojavensis subspecies. . . . . . 79 4.3 Daily egg production in D. moj. sonorensis and D. moj. wrigleyi.............................. 81 4.4 Adult body size and duration of L3 development in D. moj. sonorensis and D. moj. wrigleyi. ............... 82 4.5 The dynamics of TF formation and ovary growth in D. moj. sonorensis and D. moj. wrigleyi. ............... 84 4.6 The effects of larval nutrition in developmental time, female body size and ovariole number in D. moj. sonorensis and D. moj. wrigleyi......................... 86 4.7 Ovariole number and female body size in F1 and F2 hybrids. 88 4.8 Changes in distinct developmental processes underlie nutritional-induced and subspecies-specific variation in ovariolenumber. ........................ 90 5.1 A hypothetical threshold of ecdysone sensitivity controlled byIIS...............................101 5.2 Alternative mechanisms might underlie differences in the onset of TFC differentiation between the two D. mojavensis subspecies. ...........................103 5.3 The relationship between IIS activity and nutritional sensitivity.............................105 vi LIST OF TABLES 3.1 Pairwise comparisons of the rate of ovary growth in larvae fed on standard food. . . . . . . . . . . . . . . . . . . . . . . 63 3.2 Pairwise comparisons of the rate of TF formation in larvae fed on standard food. . . . . . . . . . . . . . . . . . . . . . . 63 3.3 Pairwise comparisons of the rate of TF formation in larvae fed on sucrose alone. . . . . . . . . . . . . . . . . . . . . . . 63 4.1 Species stocks used in this study. . . . . . . . . . . . . . . . 75 4.2 Generalized linear model (family=Poisson) for developmental time. . . . . . . . . . . . . . . . . . . . . . . 86 4.3 Two-away ANOVA model for female body size. . . . . . . . 86 4.4 Two-away ANOVA model for ovariole number. . . . . . . . 87 vii viii SUMMARY The environment can modify developmental trajectories and generate a range of distinct phenotypes without altering an organism’s genome, a widespread phenomenon called developmental plasticity. The past decades have seen a resurgent interest in understanding how developmental plasticity contributes to evolutionary processes, as it can produce phenotypic variation among individuals and facilitate diversification among populations that inhabit distinct ecological niches. To better understand the importance of plastic responses for evolutionary change, we need to explore how the environment alters development to produce phenotypic variation and then compare this to how genetic variation influences these same developmental processes. My thesis work explored the developmental mechanisms underlying both plasticity and subspecies-specific variation in ovariole number, a major determinant of female reproductive capacity, in Drosophila. Ovariole number is determined during third instar (L3) larval stages and begins with the differentiation of terminal filament cells (TFCs) that gradually intercalate into stacks called terminal filaments (TFs). The number of TFs at pupariation directly determines the number of ovarioles. The developmental processes underlying TF formation are know to vary both with environmental conditions, like nutrition, and between species. I first addressed how nutrition influences ovariole number in D. melanogaster. By manipulating nutrition at specific stages during L3 larval development, I found that ovariole number exhibits two phases of sensitivity to nutrition (Chapter 2). These two phases are separated by ix the nutrition-dependent developmental transition known as critical weight. When larvae are poorly fed during the first phase of sensitivity, ovary growth arrests and the onset of TFC differentiation is strongly delayed, resulting in a severe reduction in ovariole number. On the other hand, the effects on ovariole number in larvae that are malnourished during the second phase are more modest; ovary growth and the formation of new TFs continue, although at a reduced rate relative to well-fed larvae. Secondly, I determined the role of two hormonal pathways, the insulin/insulin-like growth factor signalling (IIS) and ecdysone signalling pathways, in regulating the nutritional sensitivity of the ovary (Chapter 3). My results indicate that both pathways regulate the nutritional-sensitive onset of TFC differentiation, with ecdysone signalling playing a pivotal role in this process. Conversely, IIS, and to a lesser extent, ecdysone signalling coordinate the rate of TF formation and of ovary growth with nutritional conditions. Lastly, I investigated the developmental changes that give rise to differences in ovariole number between two subspecies of D. mojavensis, the D. moj. sonorensis and D. moj. wrigleyi (Chapter 4). As these subspecies inhabit geographically isolated areas and breed in distinct host cacti, they provide a unique opportunity to investigate the early events associated with morphological diversification. Based on my detailed characterizations of ovary development, I found that differences in the rate of ovary growth can explain much of the variation in ovariole number between D. moj. sonorensis and D. moj. wrigleyi. From these findings, I propose that evolutionary changes in the activity of IIS could underlie the differences in ovary growth, and consequently ovariole number, between these subspecies (Chapter 5). In summary, my results underscore the importance of hormonal pathways in coordinating stage-specific developmental processes with environmental conditions, and specifically suggest that changes in the activity of hormonal pathways can account for plastic responses, and potentially also for evolutionary diversification. The powerful developmental approach embraced in this thesis may be useful to investigate how other morphological traits respond to environmental variation, and could provide significant insights to unravel one of the most captivating mysteries of biology; that of the origin of diversity in living things. x SUMÁRIO O ambiente pode alterar o desenvolvimento de um organismo e criar uma variedade de fenótipos sem alterar o seu genoma. Este fenómeno, extremamente comum na natureza, é denominado plasticidade. Nas últimas décadas, o interesse em compreender como a plasticidade pode contribuir para os processos evolutivos tem vindo a crescer. É um interesse preenchido ao facto de a plasticidade ser capaz de produzir variação fenotípica entre indivíduos e facilitar a diversificação entre populações que habitam diferentes nichos ecológicos. Para melhor compreender a importância da plasticidade na evolução, é necessário explorar de que forma o ambiente altera o desenvolvimento para produzir diversos fenótipos, e identificar se alterações semelhantes no desenvolvimento são responsáveis pela variação fenotípica entre espécies. Esta tese teve como objetivo interpretar os mecanismos de desenvolvimento, que estão na base de, quer de respostas plásticas no número de ovaríolos de Drosophila, quer de diferenças neste número resultante de variação genética entre subespécies de Drosophila. O número de ovaríolos influencia a capacidade reprodutiva da fêmea e é determinado durante o terceiro estágio larvar. O processo de formação dos ovaríolos começa com a diferenciação de ‘células dos filamentos terminais’ (TFCs), que gradualmente se intercalam, formando pilhas de células denominadas ‘filamentos terminais’ (TFs). O número de TFs no momento da pupariação iguala o número de ovaríolos no adulto. Este processo pode variar com as condições ambientais da larva, como, por exemplo, com a nutrição, e com a variação genética entre espécies. xi Primeiramente investiguei como a nutrição influencia o número de ovaríolos em D. melanogaster. Ao manipular a nutrição em diferentes alturas do terceiro estágio larval, demonstrei que o número de ovaríolos exibe dois períodos sensíveis à nutrição (Capítulo 2). Estes são separados pela transição de desenvolvimento conhecida como peso crítico. Quando as larvas são submetidas a um défice alimentar durante o primeiro período sensível, o crescimento do ovário é reprimido e o início da diferenciação dos TFCs é extremamente atrasado, resultando daí, uma severa redução no número de ovaríolos. Por outro lado, os efeitos no número de ovaríolos em, larvas que são mal nutridas, durante o segundo período sensível à nutrição são mais moderados; o crescimento do ovário e a formação de novos TFs continua. No entanto, a sua taxa de progressão é reduzida relativamente a larvas bem nutridas. De seguida, explorei a função de duas vias de sinalização hormonal, a via da insulina e a via da ecdisona, na regulação da resposta nutricional do ovário (Capítulo 3). Estes resultados indicam que ambas as vias regulam o início da diferenciação dos TFCs, tendo a via da ecdisona um papel fulcral neste processo. Contrariamente, ambas as vias, a da insulina, e, em menor grau a da ecdisona, regulam as taxas de formação de TFs e do crescimento do ovário em resposta às condições nutricionais. Finalmente, explorei possíveis alterações no desenvolvimento que pudessem explicar as diferenças observadas no número de ovaríolos entre duas subespecies de D. mojavensis, a D. moj. sonorensis e a D. moj. wrigleyi (Capítulo 4). Estas subespecies habitam áreas geograficamente isoladas e desenvolvem-se em cactus distintos, providenciando uma oportunidade única para investigar os primeiros eventos associados com a diversificação morfológica. Baseado nas caracterizações detalhadas do desenvolvimento ovárico que efetuei, demonstrei que as diferenças na taxa de crescimento do ovário podem explicar, em grande parte, as diferenças no número de ovaríolos entre a D. moj. sonorensis e a D. moj. wrigleyi. Tendo em conta esta observação, propus que mudanças evolutivas na atividade da via da insulina poderão estar na base das diferenças da taxa de crescimento do ovário, e, consequentemente, no número de ovaríolos entre as duas subespecies (Capítulo 5). Resumindo, estes resultados revelam a importância das vias de sinalização hormonal na regulação de processos que ocorrem em períodos específicos do desenvolvimento, e na sua coordenação com as condições ambientais. Além disso, sugerem também que, mudanças na atividade de vias de sinalização hormonal são responsáveis, tanto pela resposta plástica como, potencialmente também pela diversificação evolutiva. xii A poderosa abordagem focada no desenvolvimento, que foi utilizada nesta minha tese poderá ser útil para investigar como outras características morfológicas respondem à variação ambiental. Desta forma, poderá auxiliar a resolver um dos mistérios mais cativantes da biologia: o da origem da diversidade de formas de vida. xiii xiv Chapter 1 Figure 1.1: The relantionship between phenotype and the environment. (A) Average size values of two traits (blue and red lines) are different due to genetic variation. Plasticity is absent. (B) Both traits respond to environmental variation and the degree of their plastic response is similar. Plasticity is present, but genetic variation in plasticity is absent. For example, wing and body size in D. melanogaster. (C) The slopes of the reaction norms are different, suggesting that the two traits respond differently to the same environmental cue. Plasticity and genetic variation in plasticity are present. For example, male genitalia and wing size in D.melanogaster. (D) A given trait may show distinct plastic responses depending on the environmental cue. For example, ovariole number in D.melanogaster. See text for more information. 7 Chapter 1 small eyespots. Temperature shift experiments at specific developmental points revealed that the critical period in which temperature can induce changes in wing pattern occurs late in larval development (Kooi and Brakefield, 1999). Further examples have been described in other polyphenic butterflies (Nijhout, 2003b). These studies are beginning to allow us to piece together how critical periods of environmental sensitivity are regulated. 1.4 Hormonal mechanisms of developmental plasticity Recent research has revealed that environmental factors often modulate developmental trajectories to produce distinct phenotypes by controlling which, when, and how strongly genes are expressed during development. The action of hormones is perhaps one of the best-understood mechanisms mediating developmental plasticity. Hormones are known to integrate information from the external environment and regulate multiple developmental processes throughout the entire organism: some bind directly to transcription factors and activate the expression of specific genes (Baniahmad and Tsai, 1993), others activate a series of intercellular signalling cascades that regulate growth (Wu and Brown, 2006) and even others change the DNA methylation profiles in the genome, regulating which genes are expressed (Beldade et al., 2011; Snell-Rood et al., 2013). An integrated picture of how hormones link environmental variation with developmental changes has largely been drawn from studies on the regulation of body and organ size in one of the most diverse animal groups: the insects. Much of the morphological diversity seen across insect species is generated by changes in organ size and shape relative to the whole body (Shingleton et al., 2007, 2008). In holometabolous insects, body size is a function of the larval rearing environment, in which nutritional conditions play a major role. As adults have a sclerotized outer skeleton that prevents further growth, the adult body size is fixed once larvae stops feeding at the onset of metamorphosis. Moreover, many adult organs develop inside the larval body as imaginal discs and respond to the same cues that control 8 Chapter 1 whole body growth. Therefore, like adult body size, the size of many adult organs is determined by the amount of growth that imaginal discs achieve during development. Although significant progress has been made in understanding the hormonal mechanisms underlying nutritional plasticity of organ size in non-model insects (Beldade et al., 2011), recent advances in D. melanogaster have opened up unique opportunities to generate insight into the hormonal mechanisms through which nutrition changes organ size and produces novel and diverse morphologies. In D. melanogaster, like many holometabolous insects, three developmental hormones – the insulin-like peptides, juvenile hormone (JH), and the steroid moulting hormone ecdysone – translate signals from the nutritional environment to regulate body and organ growth (Mirth and Shingleton, 2012; Nijhout, 2003a). Although JH is a key regulator of growth in the tobacco hornworm Manduca sexta (Nijhout and Williams, 1974) and the dung beetles Onthophagus taurus (Emlen and Nijhout, 1999), its role in growth D. melanogaster was, until recently, controversial (Flatt, 2005; Mirth et al., 2014; Riddiford and Ashburner, 1991; Riddiford et al., 2010). In the following pages, I will therefore focus on what is known about the role of the D. melanogaster insulin-like peptides (dILPs), the insulin/insulinlike growth factor signalling (IIS) pathway, and the ecdysone signalling pathway in regulating nutritional plasticity in body and organ size. 1.4.1 Nutrition and the insulin/insulin-like growth factor signalling (IIS) pathway In D. melanogaster, and many other animals, nutrition modifies body and organ growth through the action of the IIS pathway (Figure 1.2). In rich nutritional environments, neurosecretory cells in the brain, the insulinproducing cells, synthetize and secrete high amounts of dILPs. Only three 9 Chapter 1 of the eight dILPs – dILP2, dILP3, and dILP5 – are exclusively expressed in the insulin-producing cells (Ikeya et al., 2002; Rulifson et al., 2002). The expression of these dILPs is nutrient dependent; starvation represses both their synthesis and secretion (Brogiolo et al., 2001; Ikeya et al., 2002). Further, ablation of the insulin-producing cells reduces adult body size in a similar fashion to starvation (Rulifson et al., 2002). These findings indicate that most of the nutrition-dependent growth is presumably regulated by the dILP production in the insulin-producing cells. The additional dILPs are expressed in several different tissues, including the imaginal discs, the mid gut, and the ventral nerve cord, and are thought to have systemic effects on growth (Brogiolo et al., 2001; Colombani et al., 2012; Garelli et al., 2012). After being released into the insect bloodstream, dILPs act on target tissues by binding to the insulin receptor (InR) (Brogiolo et al., 2001). Once InR is activated, a highly conserved phospho-kinase signal transduction cascade, the IIS, is induced ultimately regulating cell growth and division. This is mainly achieved by activating positive growth regulators, such as the protein kinase Akt, and suppressing negative growth regulators, such as the transcription factor Forkhead Box class (FOXO) and the Tuberous Sclerosis Complex 1 and 2 (TSC1/2) (reviewed in (Taniguchi et al., 2006). The suppression of TSC1/2 allows an additional nutrient-sensitive pathway, the target of rapamycin (TOR) signalling pathway, to remain active. The TOR pathway responds directly to intracellular amino acid concentrations via the TOR complex and regulates a number of cellular processes to promote growth (Gao et al., 2002; Sarbassov et al., 2005). In addition, the TOR complex itself regulates the IIS pathway by activating Akt (Sarbassov et al., 2005), which illustrates the extensive crosstalk between the two nutrition-sensitive pathways. Suppressing any component in the IIS pathway slows growth and results in smaller adults in a similar manner as starvation (Britton et al., 2002; Brogiolo et al., 2001). Combined, these findings illustrate that the circulating levels of dILPs and the IIS pathway coordinate growth rate with nutritional inputs. Additionally, the IIS pathway also controls the length of the growth 10 Chapter 1 Figure 1.2: The IIS pathway in D. melanogaster The secretion of dILPs by the insulin-producing cells in the brain depends on nutrition. (A) Under high nutritional conditions, high levels of dILPs are expressed and activate the IIS pathway, promoting cell growth and division primarly by activating Akt and supressing the activitity of FOXO. (B) In contrast, when larvae are malnourished, low levels of dILPs are expressed, resulting in a reduced activity of IIS pathway, and consequently, reduced growth. 11 Chapter 1 period – another crucial determinant of body and organ size in insects. The IIS pathway controls the growth period primarily by regulating the timing of the pulses of the steroid hormone ecdysone at specific stages in development (Koyama et al., 2014). How does ecdysone, in turn, regulate the duration of the growth period and ultimately body and organ size? 1.4.2 Nutrition and the ecdysone signalling pathway Ecdysone is synthetized and secreted by the prothoracic glands in a series of discrete peaks throughout larval and pupal development. This periodic release of ecdysone together with a temporaland tissue-specific expression of the ecdysone receptor complex, a heterodimer between Ecdysone Receptor (EcR) and Ultraspiracle (Usp), orchestrate many aspects of larval development: from larval molts and metamorphosis to growth and differentiation of target tissues (reviewed in (Yamanaka et al., 2013). Several studies have uncovered that one small peak of ecdysone that occurs early in the third and final instar (L3) larvae is sensitive to nutritional conditions (Caldwell et al., 2005; Colombani et al., 2005; Koyama et al., 2014; Layalle et al., 2008; Mirth, 2005). This small peak of ecdysone reaches its maximum at around 8 to 10 h after third larval ecdysis (AL3E) (Koyama et al., 2014; Warren et al., 2006) and induces a key developmental transition, critical weight. Critical weight determines when to end growth and initiate metamorphosis, thereby regulating body and organ size (Koyama et al., 2014; Mirth and Riddiford, 2007; Mirth and Shingleton, 2012). Starving larvae before reaching critical weight significantly delays the timing of the ecdysone peak, which in turn, delays the onset of metamorphosis (Beadle et al., 1938; Mirth et al., 2005; Shingleton et al., 2005; Stieper et al., 2008) and delays patterning of the presumptive adult tissues, the imaginal discs (Mirth et al., 2009). Conversely, starvation after critical weight accelerates the onset of metamorphosis and no longer prevents continued patterning and growth of the imaginal discs in the absence of nutrition (Beadle et al., 1938; Mirth et al., 2005; Mirth et al., 2009; Shingleton et al., 2005). 12 Chapter 1 Figure 1.3: The activation and derepression functions of ecdysone signalling. Ecdysone binds to its receptor, a heterodimer between EcR and Usp, to induce two types of functions.(A) Activation function: ecdysone binds to EcR-Usp and directly activates gene transcription. (B) Derepression function: ecdysone binds to EcR-Usp and relieves the repressive action of the EcR-Usp, allowing gene transcription. Knocking down either EcR or Usp partially activates gene transcription, while overexpressing a dominant negative form of EcR with a mutated ligand-binding domain (EcRDN) prevents gene transcription. Ecdysone exerts its effects by binding to the EcR/Usp heterodimer complex. This complex represses the transcription of a subset of ecdysone target genes in the absence of ecdysone (Figure 1.3B) (Brown et al., 2006; Cherbas, 2003; Schubiger and Truman, 2000; Schubiger et al., 2005). Once ecdysone binds to EcR/Usp, it induces target gene transcription either by direct activation via EcR/Usp (Figure 1.3A) or by relieving the repressive action of the EcR/Usp (Figure 1.3B). Several genetic tools in D. melanogaster allow us to explore the specific roles of ecdysone signalling in body and organ growth. For instance, knocking down either EcR or 13 Chapter 1 Usp, using RNAi, eliminates the repressive function of EcR/Usp, thereby partially inducing ecdysone function (Figure 1.3A) (Brown et al., 2006; Cherbas, 2003; Mirth et al., 2009; Schubiger et al., 2005). Conversely, overexpressing a dominant negative form of EcR with a mutated ligandbinding domain abolishes both the derepression and activation functions of ecdysone (Figure 1.3B)(Brown et al., 2006; Cherbas, 2003; Hu et al., 2003). In Chapter 3, I took advantage of these two well-described genetic tools to investigate the role of ecdysone signalling in regulating nutritional plasticity in organ size. 1.4.3 Organ-specific sensitivities to nutrition If the levels of circulating dILPs reflect the nutritional status of an insect, how do different organs respond differentially to nutritional variation? As discussed above, the size of the male genitalia and the CNS is relatively invariant across nutritional conditions (Cheng et al., 2011; Shingleton et al., 2005; Tang et al., 2011). This low sensitivity to nutrition is achieved through different mechanisms. In the case of the CNS, InR-independent activation of the IIS pathway allows the CNS to maintain its growth rate even when circulating dILPs are low (Cheng et al., 2011). Alternatively, the male genitalia reduces its plasticity in response to nutrition by expressing low levels of foxo mRNA (Tang et al., 2011). When circulating dILPs and the activity of the IIS pathway are reduced, FOXO remains in the nucleus and supresses growth (Jünger et al., 2003). As the male genitalia expresses low levels of foxo, it is able to maintain its size even when larvae are malnourished (Figure 1.4A) (Shingleton et al., 2005, 2009; Tang et al., 2011). Overexpressing FOXO in the male genitalia increases its sensitivity to nutrition and results in smaller genitalia (Figure 1.4A) (Tang et al., 2011). Despite the differences in mechanisms between the CNS and the male genitalia, ultimately these organs are protected from the effects of poor nutrition by retaining high levels of activity of the IIS pathway irrespective of nutritional conditions. Changing the level of activity of the IIS pathway in organs that scale proportionally with body size, as the wing discs, can result in 14 Chapter 1 Figure 1.4: Organs differ in their sensitivity to nutrition. (A) The male genital disc maintains its size even when larvae are poorly fed. This reduction in nutritional sensitivity is achieved by reducing the levels of foxo mRNA and retaining high IIS activity in low nutritional environments (grey line). Overexpressing FOXO in the male genitalia increases its sensitivity to nutrition (blue line). (B, C) Nutrition affects the size of the wings in proportion with body size (grey line). (B) An increase of InR expression results in an increase in nutritional sensitivity by enhancing wing size in large flies (red line). (C) An increase of foxo expression enhances the nutritional sensitivity of the wing by supressing wing size in small individuals (blue line). (D) The sensitivity to nutrition of the wing discs varies with developmental time. Before critical weight, starvation severely reduces the growth of the wing disc. On the other hand, discs grow considerably even in post-critical weight larvae that are malnourished.CW:critical weight. Adapted from (Shingleton and Frankino, 2013; Shingleton and Tang, 2012; Shingleton et al., 2008) . 15 Chapter 1 an exaggerated response to nutrition (Shingleton and Frankino, 2013; Shingleton and Tang, 2012). Overexpressing either FOXO or InR specifically in the wing disc increases its sensitivity to nutrition making it hyperallometric (i.e. disproportionally larger) in relation to body size (Figure 1.4B, C). However, this hyperallometry is achieved through different ways: increasing InR expression resulted in an exaggerated increase in the wing size of larger individuals, but has little or no effects in the wing size in smaller individuals (Figure 1.4B) (Shingleton and Tang, 2012). On the other hand, an increase in foxo expression led to a disproportionally small wing size in smaller individuals, but almost no effect in larger individuals (Figure 1.4C) (Shingleton and Tang, 2012). Thus it appears that organs can display exaggerated responses to nutrition by modulating the IIS pathway at several levels of its action (Shingleton and Frankino, 2013). Organs can also change their sensitivity to nutrition with developmental time. For instance, starving pre-critical weight larvae compromises wing disc growth and differentiation, but after critical weight starvation has a more modest effect on the development of the wings discs; that is, discs grow considerably and continue to differentiate even when post-critical weight larvae are poorly fed (Figure 1.4D) (Mirth et al., 2009; Shingleton et al., 2008). This switch in sensitivity to nutrition at critical weight seems to be mediated by changes in the IIS pathway. Supressing the IIS pathway just after critical weigh abolishes body growth, but the wing discs continue to grow presumably until their size is appropriate for the much reduced body size (Figure 1.4D)(Shingleton et al., 2005, 2008). These findings have led some authors to hypothesize that an intrinsic growth rate that does not require nutritional inputs may enable further growth of the developing organs when nutrition, and accordingly the IIS, is severely reduced (Nijhout et al., 2014; Shingleton et al., 2008). The activity of ecdysone signalling at critical weight may account for the switch in sensitivity to nutrition of developing organs. Knocking down EcR specifically in the wing discs of starved pre-critical weight larvae allows differentiation of the wing disc to proceed (Mirth et al., 2009). A similar observation was reported for the optic lobe of the Drosophila CNS 16 Chapter 2 2.1 Introduction Developmental plasticity, the ability of an organism to adjust its developmental trajectory in response to environmental variation, is a seemingly universal property of all multicellular organisms. Often, the extent of developmental plasticity depends not only on the traits and environmental conditions considered (Mirth and Shingleton, 2012), but also on the existence of phases of environmental sensitivity, commonly referred as critical periods, during which developmental processes can respond plastically (Koyama et al., 2013; Nijhout, 2003a). In the most extreme cases, an environmental cue within a critical period triggers a developmental switch between alternative developmental trajectories, giving rise to distinct phenotypes, such as dramatic seasonal differences in the pigmentation of butterfly wing patterns and different body sizes and shapes in honeybee castes (Brakefield et al., 1996; Smith et al., 2008). Understanding how developing organs change their sensitivity to environmental conditions, and how this influences their plastic response, is an important step towards a comprehensive knowledge of how the environment generates new phenotypic variants. Nutritional status is one of the major regulators of body and organ growth and its effects have been extensively studied in insects, in particular, in the fruit fly, Drosophila melanogaster. In D. melanogaster, like many insects, nutrition regulates growth by regulating a key developmental transition, critical weight, which occurs around 10 h after moulting to the third and final instar larvae (L3) (Beadle et al., 1938; Koyama et al., 2014; Mirth et al., 2005; Nijhout and Williams, 1974; Shingleton et al., 2005; Stieper et al., 2008). Starving larvae before critical weight causes them to significantly delay the onset of metamorphosis (Beadle et al., 1938; Mirth et al., 2005; Shingleton et al., 2005; Stieper et al., 2008), whereas starving larvae after critical weight induces early metamorphosis (Beadle et al., 1938; Mirth et al., 2005; Stieper et al., 2008). As critical weight determines when to end growth and initiate metamorphosis, the underlying mechanisms regulating its attainment have been studied extensively (Beadle et al., 1938; Koyama et al., 2014; Mirth et al., 2005; Nijhout and Williams, 1974; Shingleton et al., 2005; Stieper et al., 2008). 23 Chapter 2 Critical weight also regulates the sensitivity of developing organs to nutrition over developmental time. Larvae starved before reaching critical weight delay the patterning of their presumptive adult tissues, the imaginal discs (Mirth et al., 2009). Conversely, starvation after critical weight allows continued patterning and growth of the imaginal discs (Mirth et al., 2005, 2009; Shingleton et al., 2008). This additional role of critical weight has been overlooked in current research, and importantly, whether critical weight determines periods of nutritional sensitivity has not yet been fully investigated. In this chapter, I attempt to elucidate how developing organs change their sensitivity to nutrition over developmental time, with special emphasis on the potential role of critical weight in mediating nutritional sensitivity. To address this issue, I used ovariole number in D. melanogaster as a model. Ovarioles are egg-producing structures in the insect ovary that directly affect female reproductive capacity (Boulétreau-Merle et al., 1982; R’ kha et al., 1997; Klepsatel et al., 2013b,a). Although little is known about the genetic cascades involved in ovariole development (Cheng et al., 2011; Forbes et al., 1996; Godt and Laski, 1995; Patel et al., 1989; SahutBarnola et al., 1995; Sarikaya and Extavour, 2015), the cellular events mediating this process are better characterized. Ovariole development occurs during the third instar (L3) larval and early pupal stages (Kerkis, 1931; King, 1970; King et al., 1968) through the intercalation of terminal filament cells (TFCs) into stacks of seven to ten flattened cells, called terminal filaments (TFs) (Godt and Laski, 1995; Sahut-Barnola et al., 1995, 1996). Each TF defines the position of one ovariole and thus, the number of TFs at pupariation is equivalent to the number of ovarioles in the adult (Godt and Laski, 1995; Hodin and Riddiford, 1998; SahutBarnola et al., 1995; Sarikaya et al., 2012). Nutritional conditions during larval stages regulate ovariole number (Bergland et al., 2008; Hodin and Riddiford, 2000; Sarikaya et al., 2012; Tu and Tatar, 2003). Previous studies of the developmental effects of nutrition on ovariole number have shown that diluting the food on which larvae were raised altered ovariole number by changing the total number of TFCs (Sarikaya et al., 2012) or the rate of TF formation in late L3 larvae (Hodin 24 Chapter 2 and Riddiford, 1998). However, it was unclear whether ovary development exhibits critical periods of nutritional sensitivity, and importantly, how the developmental processes are modified by nutrition at different periods of sensitivity. I therefore examined whether changes in nutrition at specific stages during L3 larvae influence the plastic response of ovariole number. I further investigated how distinct stage-specific developmental processes during ovary development respond to changes in nutrition and account for nutritional-induced differences in ovariole number. 2.2 Materials and Methods 2.2.1 Fly stock To assess the effects of larval nutrition on ovariole number, I used an outbred population (wild type) of Drosophila melanogaster established in the laboratory of Dr. Élio Sucena in 2007, originating from 160 fertilized females collected in Azeitão, Portugal (Martins et al., 2013). The population was kept in laboratory cages with high census (>1500 individuals) and maintained at constant temperature (25°C) on standard fly food (4.5% molasses, 7.2% sugar, 7% cornmeal, 2% yeast extract, 1% agar and 2.5% Nipagin solution). 2.2.2 Larval staging and dietary manipulations Adults were allowed to lay eggs for two to six hours on fresh food plates (60 ×15 mm Petri dish). Egg density was controlled to prevent overcrowding (approximately 200 eggs per plate). Larvae were selected 0-2 hours after ecdysis to L3 (AL3E) and transferred onto new food plates (40-60 larvae per plate) to feed until they reached the appropriate age. To determine critical periods of sensitivity to nutrition in ovariole number, 20-30 larvae of the appropriate age were transferred to vials containing either 20% sucrose on 0.5% agar medium (sucrose-only food) or standard fly food (standard food) until the end of the feeding period (Figure 2.1A). On sucrose-only food, most larvae survived until pupariation and adult 25 Chapter 2 eclosion. To obtain L3 ovaries, larvae of the appropriate age were dissected and processed for immunocytochemistry (Figure 2.1C, 2.3A, 2.5A). All experiments were performed at 25°C. 2.2.3 Measurements of life-history traits: developmental time, female weight, early female fecundity and ovariole number To determine the average time to pupariation, newly ecdysed L3 larvae were transferred to vials (20-30 larvae per vial) containing standard food. The number of larvae pupariating (immobile larvae with evaginated spiracles) was counted in 2 h intervals until all larvae pupariated. I used pharate weight as a proxy of adult body size (Mirth et al., 2005). Pharate adults were collected from food vials and food residuals were carefully cleaned off from the pupal cases using distilled water and a paintbrush. I distinguish females from males by the presence or absence of malespecific sex combs through the pupal case. Female pharate adults were individually weighed on a Sartorius SE2 ultramicrobalance. To determine early fecundity, newly eclosed females were individually maintained in vials on standard food with one male of the same food/time point. Individuals were transferred to fresh vials every day during the first three days after eclosion. All eggs were counted daily. To count adult ovariole number, newly eclosed flies were maintained in vials (ten females and five males per vial) on standard food until the time of dissection (4-6 days after eclosion) (Figure 2.1A). Ovaries were dissected in cold phosphate buffered saline containing 1% Triton X-100 (PBT) and ovarioles were teased apart and counted under a dissecting microscope. 2.2.4 Immunocytochemistry Female larvae were selected by the small size of their gonads located in the posterior third of the fat body. Larvae were dissected in cold phosphate buffered saline (PBS) and fixed in 4% formaldehyde in PBS for 30 minutes at room temperature. Larvae were then washed three times for 20 minutes 26 Chapter 2 with PBT and blocked in 2% normal donkey serum in PBT for 30 minutes. Primary antibody incubation in mouse anti-Engrailed (Developmental Studies Hybridoma Bank 4D9, 1:40) diluted into 2% normal donkey serum in PBT was conducted overnight at 4°C. After washing three times for 20 minutes in PBT, larvae were incubated in the dark with goat anti-mouse Alexa 568 (Invitrogen, 1:200) and TRICT-Phalloidin (Sigma, 1:200) diluted into 2% normal donkey serum in PBT overnight at 4°C. Larvae were rinsed with PBT and ovaries were mounted on a poly-Llysine-coated coverslip using Fluoromount-G (SouthernBiotech). 2.2.5 Image Acquisition and Analysis Samples were imaged using a Zeiss LSM 510 Meta confocal microscope using a 40x 1.3NA oil objective lens. During confocal image acquisition, the detection parameters were adjusted to avoid underor overexposed pixels, and images were acquired through the full thickness of the ovary at 1 µm. Images were processed and analysed using ImageJ (NIH) and Adobe Photoshop (Adobe Systems). For each time point/genotype/food treatment, terminal filament cells (TFC) were identified by Engrailed expression. Forming terminal filaments (TFs) were identified by the presence of TFC in stacks with the characteristic flattened cell morphology, and total number of forming TFs were counted. For ovary volume, the ImageJ Volumest plugin was used (Merzin, 2008). 2.2.6 Statistical Analysis All experiments were replicated at least twice. The preliminary experiment described in Figure 2.1C, D was performed one time with small sample size. The distribution of residuals was tested for normality using Q-Q plots and the appropriate statistical test was applied. For multiple comparisons, ANOVAs were performed followed by Tukey’s multiple comparison test to evaluate pairwise differences. Welch t-test (parametric) and Wilcoxon rank sum test (non-parametric) were used to test differences in mean values between two samples. To determine differences in the rate of TF formation and of ovary growth, slopes were compared using the 27 Chapter 2 function ‘sm.ancova’ under ‘sm’ library. All data analyses and statistics were conducted using R v3.1.2 (R Development Core Team, 2014). Plots were made using GraphPad Prism v6 (GraphPad Software). p-values are indicated in the text and figures. 2.3 Results 2.3.1 Two phases of sensitivity to nutrition regulate the plastic response of ovariole number To determine critical periods of sensitivity to nutrition in ovariole number, I fed L3 larvae either on standard food or on sucrose-only food at timed intervals starting between 0 h to 30 h AL3E until the end of the feeding period (Figure 2.1A). Larvae fed on sucrose-only food are starved of protein, lipids and other micronutrients present in yeast, yet show higher rates of survival than when starved completely. Overall, larvae transferred to sucrose-only food between 0 and 25 h AL3E showed a significant reduction in ovariole number when compared to the controls transferred to standard food (Figure 2.1B). In contrast, transferring larvae to sucroseonly food at 30 h AL3E did not cause a significant reduction in ovariole number (Figure 2.1B). Interestingly, the effect of the sucrose-only food in ovariole number depended on the timing at which larvae were transferred to the sucroseonly food (Figure 2.1B). To test for a significant change in the response to sucrose-only food over time, I applied a bi-segmental linear regression model to the data and tested for a significant change in slope. The relationship between ovariole number and the age at transfer to sucroseonly food (in h AL3E) has a significant change in slope around a single breakpoint (Davies’ test for a change in slope, p<0.0001) at 11.5 h AL3E (95% CI: 9.37–13.64 h AL3E) (Davies, 1987; Muggeo, 2003, 2007). This estimated breakpoint coincides with the attainment of critical weight, suggesting that pre-critical weight ovaries are more sensitive to changes in nutrition than post-critical weight ovaries. 28 Chapter 2 Figure 2.1: Changes in nutrition during the first phase of sensitivity have greater effects in ovariole number than in the second phase of sensitivity.(A) Experimental design to determine critical periods of sensitivity to nutrition in ovariole number. Only the first two time points are shown (0 h and 5 h AL3E). (B) Adult ovariole number from larvae transferred either to standard food (yellow circles) or to sucrose-only food (blue circles). Dashed lines show the best fitted-lines from the segmental regression analyses. n≥30 for all treatments. (C) Experimental design to determine whether the length of exposure to sucrose-only food influences ovariole number. (D) Adult ovariole number from larvae fed on standard food (yellow circles); larvae transferred to sucroseonly food either at 5 h AL3E (light blue circles) or at 20 h AL3E (dark blue circles) and larvae fed on sucrose-only food for a 20 h interval either between 0 h to 20 h AL3E (open blue circles) or between 20 h to 40 h AL3E (open dark circles). Plotted values represent means and error bars show 95% confidence intervals of means. Two-way ANOVA using Tukey’s test: *p<0.05, ***p<0.001, ns non-significant. L3: third instar larvae; AL3E: after L3 ecdysis. 29 Chapter 2 Nevertheless, the effects of the sucrose-only food in ovariole number could also be a direct consequence of different lengths of exposure to the sucroseonly food. To test this hypothesis, I performed a preliminary experiment where L3 larvae were fed on sucrose-only food for 20 h starting either at 0 h AL3E or at 20 h AL3E and then returned to standard food until the end of the feeding period (Figure 2.1C). As described above, ovariole number is severely reduced when larvae were transferred to sucrose-only food at 5 h AL3E (Figure 2.1B, D). Surprisingly, ovariole number in larvae fed on sucrose alone for a short period between 0 h to 20 h AL3E was similar to standard food control (Figure 2.1D). In contrast, when larvae were fed on sucrose-only food from 20 h to 40 h AL3E, ovariole number was significantly reduced (Figure 2.1D). This reduction in ovariole number was similar when compared to larvae transferred to sucrose alone at 20 h AL3E until the end of development (Figure 2.1B, D). These observations corroborate a previous study where re-feeding pre-critical weight larvae after a brief period of starvation delays pupariation for longer than the length of the starvation period, but does not affect final body size. After critical weight, when the duration of the larval growth period is fixed, short periods of starvation have no effect on the timing of pupariation and thus, larvae are unable to reach their optimal body size even after re-feeding (Beadle et al., 1938). Ovariole number is positively correlated with egg production rate and therefore is closely related to fitness (Boulétreau-Merle et al., 1982; R’ kha et al., 1997; Klepsatel et al., 2013b,a). As expected, I found that females that were fed on sucrose-only food as larvae at timed intervals starting between 5 h to 25 h AL3E until the end of the feeding period laid fewer eggs in the first three days after eclosion relative to standard food control (Figure 2.2). When I plotted ovariole number against daily egg production, I found that differences in ovariole number correlated with differences in the number of eggs laid (Figure 2.2). Thus, these results confirm that ovariole number is a good proxy for early female fecundity in D. melanogaster. 30 Chapter 2 Figure 2.2: Ovariole number is positively correlated with early female fecundity. Number of eggs laid was counted in the first three days after eclosion (diamond: 1st day after eclosion; square: 2nd day after eclosion; circle: 3rd day after eclosion) from females fed on standard food as larvae (yellow symbols) and females fed on sucrose-only food as larvae at timed intervals starting between 5 h to 25 h AL3E (symbols with different shades of blue) until the end of the feeding period. Plotted values represent means and error bars show 95% confidence intervals of means. L3: third instar larvae; AL3E: after L3 ecdysis. 2.3.2 Ovary development during L3 larval stages To further understand how nutrition regulates ovariole number, I first analysed ovary development in L3 larvae raised in standard food. When TFCs differentiate from the surrounding ovarian somatic cells, they upregulate expression of the transcription factor Engrailed (En) (Patel et al., 1989). Thus, I used En as a marker for TFC differentiation and TF formation. Consistent with previous studies, TFCs were not observed in pre-critical weight ovaries (from 0-10 h AL3E) (Figure 2.3A, B) (Godt and Laski, 1995). At 15h AL3E, TFCs appeared in the medial side of the ovary and few short TFs were visible (Figure 2.3A, B). New TFCs continued to emerge from the surrounding ovarian somatic cells and gradually intercalated into forming TFs. The formation of new TFCs occurs in a lateral direction (Figure 2.3A, B) (Godt and Laski, 1995; Sahut-Barnola et al., 1995, 1996). At the end of L3, all of the approximately 18-22 TFs 31 Chapter 2 Figure 2.3: Ovary development during L3 larval stages under optimal nutritional conditions.(A) Schematic drawings representing ovary development in L3 larvae reared in standard food. Terminal filaments (TFs) are represented as dark grey symbols. Axis are presented as A-P, anterior-posterior; D-V, dorsal-ventral; M-L, medial-lateral. Pictures show developing ovary during L3 larval stages under standard food. Engrailed (grey) marks terminal filament cells (TFCs). Scale bar: 20µm. (B) Number of forming terminal filaments (TFs). (C) Ovary volume. Plotted values represent means and error bars show 95% confidence intervals of means. L3: third instar larvae; AL3E: after L3 ecdysis. have formed (Figure 2.3A, B) (Godt and Laski, 1995; Hodin and Riddiford, 1998; Sarikaya et al., 2012). To assess the dynamics of ovary growth, I measured ovary volume over development time. Ovary volume increased exponentially throughout L3 larval development (Figure 2.3C), confirming results previously found in (Kerkis, 1931). 32 Chapter 2 modifies the growth trajectories of the wing imaginal discs (Garcia-Bellido and Merriam, 1971; Martin, 1982; Bryant and Levinson, 1985). After attainment of critical weight, wing discs have an intrinsic growth rate that promotes considerable growth under poor nutritional conditions. This intrinsic growth is not present before critical weight, and thus, wing discs arrest growth when pre-critical weight larvae are poorly fed (Shingleton et al., 2008). In light of these observations, I propose that developing ovaries may also have an intrinsic growth rate after critical weight that allows progression of growth in poorly-fed larvae. 2.5 Conclusions The results described in this chapter revealed that critical weight plays a fundamental role in reprograming the developing ovary’s response to nutrition. Furthermore, this work contributes to a better understanding of the developmental processes that regulate ovariole number, and provides the developmental tools that will be used throughout this thesis. Acknowledgements I would like to thank Vítor Faria and Nelson Martins for taking care of the outbred populations of D. melanogaster. Christen Mirth for teaching me how to count ovarioles and for commenting and proofreading this chapter. Élio Sucena for discussions throughout this project and for comments on an early version of this chapter. The Unit of Imaging at IGC for all the assistance during image acquisition and analysis. The Fly Facility, in particular Liliana Vieira, for making fly food every week. 39 Chapter 2 Supplementary Figures S 2.1: Engrailed is expressed in the wing discs of larvae reared in sucroseonly food. To confirm that the absence of TFCs (Engrailed-positive cells) was due to the diet manipulation and not to the immunocytochemistry protocol, I dissected wing discs or central nervous system (CNS) and analysed them together with the larval ovaries. Engrailed is expressed in the posterior compartment of the wing disc and in the neuroblasts in the CNS. Shown is a wing disc of larvae fed on sucrose-only food from 5 h AL3E expressing Engrailed. Phalloidin marks F-actin to outline cell membranes. Scale bar: 20µm. L3: third instar larvae; AL3E: after L3 ecdysis. 40 Chapter 2 S 2.2: Larval nutrition affects development time and female weight.((A) Development time represented in hours after third instar ecdysis (h AL3E) to pupariation. Wilcoxon rank test using Holm’s p-value adjustment: ***p<0.001, ns non-significant. (B) Female weigh. Welch Two sample t-test using Holm’s p-value adjustment: ***p<0.001. Larvae fed on standard food (yellow circles) and larvae transfer to sucrose-only food either at 5h AL3E (light blue circles) or at 15h AL3E (dark blue circles). Plotted values represent means and error bars show 95 confidence intervals of means. L3: third instar larvae; AL3E: after L3 ecdysis. 41 Chapter 2 42 3 HORMONAL SIGNALLING REGULATES PLASTICITY IN OVARIOLE NUMBER IN DROSOPHILA MELANOGASTER “I mean, how could they know that because of their little dance the world lives? But it does. By simply doing what they’re designed to do, something large and magnificent happens.” –from the film Adaptation (2002) 43 Chapter 3 Abstract Hormones coordinate body and organ growth with environmental conditions. Several studies have uncovered that changes in the timing and amount of hormone production are associated to numerous plastic responses in morphology, behaviour and physiology. However, how hormones act at specific stages in development to modify the sensitivity of a developing organ to environmental variation is poorly understood. In the previous chapter, I have shown that critical weight separates two phases of sensitivity to larval nutrition in ovariole number. Here, I set out to test whether two hormonal signalling pathways – the insulin/insulinlike growth factor signalling (IIS) and ecdysone signalling pathways – act at critical weight to control the response of the developing ovary to nutritional conditions and thus, regulate the plastic response of ovariole number. Indeed, I found that both IIS and ecdysone signalling pathways change the developing ovary’s sensitivity to nutrition by regulating distinct stage-specific developmental processes that I have shown to be modified by nutritional conditions. These results enhance our understanding of the stage-specific action of hormones in regulating plastic responses. Publication Chapter 2 and 3 are part of a manuscript submitted for publication, authored by C.C. Mendes and C.K. Mirth. Authors’ contributions Christen Mirth and Cláudia Mendes conceived this study. Cláudia Mendes performed and analysed all experiments and wrote this chapter. 44 Chapter 3 3.1 Introduction Environmentally-induced phenotypes are often induced by hormone signals that integrate cues from the external environment and coordinate developmental processes throughout the whole organism. This integration of environmental information is often reflected in changes in the timing and/or amount of hormone production during specific stages in development. These changes in hormone production can then affect the growth and differentiation of developing tissues and produce a wide range of distinct morphologies (Beldade et al., 2011; Koyama et al., 2013; Nijhout, 2003a). Moreover, changes in the activity of hormonal pathways have been shown to underlie differences in nutritional sensitivity between different organs (Tang et al., 2011). However, there is still a fundamental gap in our understanding of how hormones act at specific stages in development to change the sensitivity of a developing organ to environmental variation. In the previous chapter, I have shown that a key developmental transition, critical weight, appears to reprogram the sensitivity of the developing ovary to nutritional conditions (see Chapter 2). While starving larvae before critical weight strongly compromises terminal filament cell (TFC) differentiation and ovary growth, starvation after critical weight has a more modest effect on ovary development; that is, ovary growth and the formation of new TFs continue, although at slower rates, in post-critical weight larvae that are poorly fed. Such changes in nutritional sensitivity greatly influence the plastic response of ovariole number. Interestingly, wing disc development shows a similar change in nutritional sensitivity at critical weight (Mirth et al., 2009; Shingleton et al., 2008). This switch in the nutritional sensitivity of the wing is regulated by two hormonal pathways, the insulin/insulin-like growth factor signalling (IIS) and ecdysone signalling pathways. At critical weight, the prothoracic glands produce a small peak of ecdysone (Koyama et al., 2014). Once released, ecdysone binds to its receptor, a heterodimer between Ecdysone Receptor (EcR) and Ultraspiracle (Usp), in the wing discs and activates the transcription of target genes, allowing 45 Chapter 3 the progression of differentiation to continue even in the absence of nutrition (Mirth et al., 2009). In addition, the growth of wing discs becomes less sensitive to nutrition after critical weight (Shingleton et al., 2008). Because IIS regulates body and organ growth in response to nutrition, the pulse of ecdysone at critical weight is thought to make the growth of the disc less sensitive to IIS (Koyama et al., 2013; Stieper et al., 2008). Interestingly, both IIS and ecdysone signalling pathways regulate ovariole number (Gancz and Gilboa, 2013; Gancz et al., 2011; Green and Extavour, 2012, 2014; Hodin and Riddiford, 1998), and IIS is known to underlie the plastic response of ovariole number to larval nutrition (Green and Extavour, 2014). These studies reported that manipulating IIS in the larval ovary affects ovary volume and total TF number, but has no effect on the timing of the onset of TFC differentiation (Gancz and Gilboa, 2013; Green and Extavour, 2012, 2014). Repressing ecdysone signalling in the developing ovary delays TFC differentiation and reduces ovary volume (Gancz et al., 2011; Hodin and Riddiford, 1998). However, these studies have not explored the effects of IIS and ecdysone signalling pathways with sufficient temporal resolution to dissociate their roles in regulating TFC differentiation, the rate at which new TFs are formed, or the rate of ovary growth. Furthermore, the relative roles of IIS and ecdysone signalling pathways in altering developmental processes in response to nutrition have not been explored. Here, I investigated whether IIS and ecdysone signalling pathways act during critical weight to regulate the developing ovary’s sensitivity to nutritional conditions, and ultimately, influence the plastic response of the ovary. I first examined the relative contributions of each signalling pathway in regulating three developmental processes during ovary development: the onset of TFC differentiation, the rate of TF formation, and the rate of ovary growth in well-fed larvae. I next explored whether the nutritional sensitivity of these developmental processes is mediated by IIS and ecdysone signalling pathway. 46 Chapter 3 3.2 Material and Methods 3.2.1 Fly Stocks To genetically manipulate IIS and/or ecdysone signalling pathways, I used traffic jam-GAL4 (tj) [yw;PGawBNP1624/CyO; a gift of Lilach Gilboa] to drive expression in somatic cells of the larval ovary (Gancz et al., 2011). This line was crossed to w1118 (used as a control; w1118, tj or tj>), w; UAS EcRA.W650A TP3 (>EcRwDN), w; UAS EcR RNAi CA104 (>EcRi), yw;+;UAS PTEN (>PTEN), yw flp; +; UAS InR29.4 (>InR), w; UAS EcRA.W650A TP3; UAS InR29.4 (>EcRi, InR) or w; UAS EcRA.W650A TP3; UAS InR29.4 (>EcRDN, InR) . To characterize the expression patterns of traffic jam-GAL4, the following lines were used: w; +; UAS GFP (>GFP), w; traffic jam-GAL4; UAS GFP (tj>GFP), w; elav-GAL4 (drives expression in neuroblast and glial cells in the larval brain; elav>) and w; elav-GAL80, traffic jam-GAL4 (inhibits traffic jamGAL4 expression in the neuroblast and glial cells in the larval brain; elav80, tj>). Fly stocks were maintained at 22°C in bottles on standard fly food (4.5% molasses, 7.2% sugar, 7% cornmeal, 2% yeast extract, 1% agar and 2.5% Nipagin solution). For ecdysone feeding experiments, I used an outbred population (wild type) of D. melanogaster established in the laboratory of Dr. Élio Sucena in 2007, originating from 160 fertilized females collected in Azeitão, Portugal (Martins et al., 2013). The population was kept in laboratory cages with high census (>1500 individuals) and maintained at constant temperature (25°C) on standard fly food (4.5% molasses, 7.2% sugar, 7% cornmeal, 2% yeast extract, 1% agar and 2.5% Nipagin solution). 3.2.2 Larval staging, dietary manipulations and ecdysone feeding experiments Adults were allowed to lay eggs for two to six hours on fresh food plates (60 ×15 mm Petri dish). Egg density was controlled to prevent overcrowding (approximately 200 eggs per plate). To determine the effects of suppressing IIS and/or ecdysone signalling pathways in developing 47 Chapter 3 Figure 3.1: Experimental design of nutritional manipulations.(A) Experimental design to determine the effects of supressing IIS or ecdysone signalling on ovary development and ovariole number in larvae fed on standard food (yellow line). (B) A similar experimental design was performed to examine the effects of activating IIS and/or partially activating ecdysone signalling on ovary development. In this set of experiments, larvae were transferred at 5 h AL3E either to standard food (yellow line) or sucrose-only food (blue line). Dissection times are marked with red crosses.L3: third instar larvae; AL3E: after L3 ecdysis. ovaries, larvae were selected 0-2 hours after ecdysis to L3 (AL3E) and transferred onto new plates (40-60 larvae per plate) with standard fly food (standard food) until the time of dissection (15 h, 29 h and 39 h AL3E) (Figure 3.1A). To determine the effects of activating IIS and/or partially activating ecdysone signalling pathways in developing ovaries, larvae were collected as above in 2 h intervals from ecdysis and fed on standard food. At 5 h AL3E, larvae were transferred to vials (20-30 larvae per vial) containing either 20% sucrose on 0.5% agar medium (sucrose-only food) or standard food. The sucrose-only food allowed most larvae to survive until pupariation and adult eclosion. Larval ovaries were dissected at four time points (5 h, 15 h, 29 h and 39 h AL3E) (Figure 3.1B). For ecdysone feeding experiments, I supplemented 0.15 mg/mL of the active ecdysone metabolite 20-hydroxyecdysone (20E; SciTech Chemicals, Dejvice-Hanspaulka, Czech Republic) to 1 g of either standard food or sucrose-only food. The food was then well mixed and spun down a day before use. Newly ecdysed L3 larvae were collected as above and fed on 48 Chapter 3 I first analysed ovary development in tj>PTEN larvae reared in standard food conditions. Ovaries from tj>PTEN larvae showed a moderate delay in the onset of TFC differentiation (Figure 3.5A, B). Further, both rate of TF formation and the rate of ovary growth were greatly reduced in tj>PTEN ovaries when compared to control ovaries (w1118; tj) (Figure3.5A-A”, B-B”, D, E). Conversely, activating IIS in ovarian somatic cells by overexpressing Insulin Receptor (tj>InR) did not affect the timing of TFC differentiation (Figure-3.5A, C), although it increased the rate of TF formation and of ovary growth (Figure 3.5A-A”, C-C”, D, E). Overall, these results suggest that IIS regulates all three developmental processes in well-fed conditions. I next sought to test whether activation of IIS was sufficient to overcome the effects of poor nutrition in larvae fed on sucrose alone. In control larvae fed on sucrose, I failed to detect any TFCs before 39 h AL3E (Figure 3.5F’, F”). In contrast, I detected TFCs in tj>InR ovaries from larvae fed on sucrose already at 15 h AL3E (Figure 3.5G’). Such timing of TFC onset in sucrose fed tj>InR larvae was similar to that of well-fed controls (Figure 3.5A, C, G’, H) and new TFs were still forming at 39 h AL3E (Figure 3.5G”, H). The rate of TF formation was significantly higher in tj>InR larvae fed on sucrose alone than in similarly treated control larvae. Further, at the time of transfer to sucrose-only food (5 h AL3E), the ovary volume in tj>InR larvae was significantly bigger than that of ovaries from control larvae (p<0.001; Kruskal-Wallis test). However, when tj>InR larvae were fed on sucrose-only food between 5 and 39 h AL3E, ovary growth was completely arrested (Figure 3.5I). In summary, activation of IIS in ovarian somatic cells of poorly-fed larvae is sufficient to induce a precocious onset of TFC differentiation and to accelerate the rate of TF formation, but has no effect on ovary growth. 3.3.3 Role of ecdysone signalling during ovary development My previous manipulations of larval nutrition revealed that the onset of TFC differentiation is highly sensitive to changes in nutrition during the 55 Chapter 3 Figure 3.5: Role of IIS during ovary development. (A-G”) Shown is terminal filaments (TFs) marked with En (grey).(A-C”, F, G) Ovaries from larvae reared on standard food: (A-A”, F) w1118, tj (control), (B-B”) tj>PTEN and (C-C”, G) tj>InR.(D) Number of forming TFs and (E) ovary volume of ovaries from larvae reared on standard food. (F’-F”, G’-G”) Ovaries from larvae transferred to sucrose-only food at 5 h AL3E: (F’-F”) w1118, tj (control) and tj>InR. (F’-F”).(H) Number of forming TFs and (I) ovary volume of ovaries from larvae fed on sucrose-only food. n ≥8 ovaries for all genotypes. Plotted values represent means and error bars show 95% confidence intervals of means. In some cases, error bars are too small to be seen. ANCOVAs: **p<0.01, ***p<0.001, ns non-significant. L3: third instar larvae; AL3E: after L3 ecdysis. Scale bar: 20µm. 56 Chapter 3 pre-critical weight phase. Once TFC differentiation is initiated around the timing of critical weight, TF formation proceeds at a reduced rate in poorly-fed larvae (see Chapter 2). Critical weight itself is regulated by a small nutrition-sensitive ecdysone peak that occurs at around 8 h AL3E (Koyama et al., 2014; Mirth et al., 2005; Warren et al., 2006). Moreover, both EcR and Usp are expressed in ovarian somatic cells during L3 larval stages (Gancz et al., 2011; Hodin and Riddiford, 1998). Thus, I reasoned that ecdysone is likely to induce TFC differentiation. Under standard food conditions, control ovaries (w1118, tj) showed TFCs and a few forming TFs at 15 h AL3E (Figure 3.6A-A”, D). However, I did not detect any TFCs in tj>EcRDN ovaries from well-fed larvae until 39 h AL3E (Figure 3.6B-B”, D). Accordingly, the rate of TF formation was severely reduced in tj>EcRDN ovaries (Figure 3.6D). Thus, ecdysone signalling is necessary to induce the timely onset of TFC differentiation and promote subsequent formation of new TFs. In addition, I found that the rate of ovary growth was significantly reduced in tj>EcRDN ovaries (Figure 3.6E), suggesting that basal levels of ecdysone are likely required to promote ovary growth. I then tested whether a partial activation of ecdysone signalling in ovarian somatic cells is sufficient to induce the onset of TFC differentiation, to increase the rate of TF formation, and promote ovary growth in precritical weight larvae fed on sucrose-only food. To do this, I used the traffic jam-GAL4 driver line to overexpress an RNAi construct against EcR under the control of UAS (tj>EcRi). In well-fed conditions, tj>EcRi ovaries showed similar timing in their onset of TFC differentiation as well as rate of TF formation relative to controls (Figure 3.6A, C, D). The rate of ovary growth was reduced in tj>EcRi ovaries (Figure 3.6E). When tj>EcRi larvae were fed on sucrose-only food between 5 and 15 h AL3E, most ovaries had TFCs (Figure 3.6G’, H). In contrast, TFCs were not detectable in control (w1118, tj) larvae fed on sucrose-only food until 39 h AL3E (Figure 3.6F’, H). Also, the rate of TF formation, but not of ovary growth, increased in tj>EcRi larvae fed on sucrose alone (Figure 3.6H, I). Together, these results indicate that knocking down EcR in the ovarian somatic cells of larvae fed on sucrose alone is sufficient to induce precocious 57 Chapter 3 Figure 3.6: Role of ecdysone signalling during ovary development. (A-G”) Shown is terminal filaments (TFs) marked with En (grey).(A-C”, F, G) Ovaries from larvae reared on standard food: (A-A”, F) w1118 , tj (control), (B-B”) tj>EcRDN and (C-C”, G) tj>EcRi.(D) Number of forming TFs and (E) ovary volume of ovaries from larvae reared on standard food. (F’-F”, G’-G”) Ovaries from larvae transferred to sucrose-only food at 5 h AL3E: (F’-F”) w1118 , tj (control) and (G’-G”) tj>EcRi. (H) Number of forming TFs and (I) ovary volume of ovaries from larvae fed on sucrose-only food. n ≥8 ovaries for all genotypes. Plotted values represent means and error bars show 95% confidence intervals of means. In some cases, error bars are too small to be seen. ANCOVAs: **p<0.01, ***p<0.001, ns non-significant. L3: third instar larvae; AL3E: after L3 ecdysis. Scale bar: 20µm. 58 Chapter 3 onset of TFC differentiation and accelerate the rate of TF formation when compared to similarly treated controls. However, knocking down EcR in either standard or sucrose-only conditions reduces ovary growth. Knocking down EcR in the larval ovaries induces the derepression, but not the activation function of ecdysone signalling (Brown et al., 2006; Schubiger and Truman, 2000; Schubiger et al., 2005). To investigate the full role of ecdysone signalling in regulating ovariole number plasticity, I fed wild-type larvae from 5 to 29 hours AL3E on either standard food or sucrose-only food supplemented with 0.15 mg/mL of the active ecdysone metabolite 20-hydroxyecdysone (20E). Adding 20E to the standard food had no effect on TF number (Figure 3.7A, C, E). However, larvae fed on 20E-supplemented sucrose-only food had significantly more TFs at 29 hours AL3E than larvae fed on sucrose-only food plus solvent (ethanol) (Figure 3.7B, D, E). Moreover, ovary volume significantly increased in larvae fed on both standard and sucrose-only foods containing 20E relative to ethanol controls (Figure 3.7F). This experiment confirms that ecdysone is sufficient to induce TFC differentiation when pre-critical weight larvae are fed on sucrose alone. Moreover, it also suggests that ecdysone regulates ovary growth through its activation function. 3.3.4 The interplay between IIS and ecdysone signalling pathways Both IIS and ecdysone signalling pathways regulate the onset of TFC differentiation, the rate of TF formation, and the rate of ovary growth, suggesting that nutrition influences ovariole number through both signalling pathways. However, the relative contribution of these signalling pathways in regulating each development process appears to be different, with ecdysone signalling playing a more prominent role in regulating TFC differentiation and IIS contributing more to the rates of ovary growth. To test this hypothesis, I first upregulated IIS, using InR, while inhibiting ecdysone signalling, via EcRDN, in ovarian somatic cells (tj>EcRDN, InR) of well-fed larvae. Ovaries from tj>EcRDN, InR larvae delayed the onset of TFC differentiation and reduced the rate of TF formation when compared 59 Chapter 3 Figure 3.7: Feeding wild-type larvae with 20E-supplemented sucrose-only food increases TF number and ovary volume. (A-D) Shown is terminal filaments (TFs) marked with En (grey). Ovaries from larvae reared on standard food: (A) plus ethanol (control) or (C) plus 20E (+20E). Ovaries from larvae reared on sucrose-only food: (B) plus ethanol (control) or (D) plus 20E (+20E). Larvae were dissected at 29 h AL3E. Scale bar: 20µm. (E) Number of forming TFs and (F) ovary volume of ovaries from larvae fed either on standard food plus ethanol (control) or on 20Esupplemented standard food (+20E) (yellow points) and larvae fed either on sucrose alone plus ethanol (control) or on 20E-supplemented sucrose-only food (+20E) (blue points). Plotted values represent means and error bars show 95% confidence intervals of means. In some cases, error bars are too small to be seen. Welch Two sample t-test: *p<0.1, ***p<0.001, ns non-significant. 60 Chapter 3 to control larvae (w1118; tj) (Figure 3.8A-A”, 3.8B-B”, D). Thus, to induce TFC differentiation, IIS requires ecdysone signalling to be intact. Nevertheless, the rate of ovary growth increased relative to control and tj> EcRDN larvae (Figure 3.8E and Table 3.1). Taken together, these results demonstrate that ecdysone signalling is essential in regulating the timing of the onset of TFC differentiation. Furthermore, these results also show that IIS can overcome the growth defects arising from disrupting ecdysone signalling, suggesting that IIS plays a primary role in controlling ovary growth. I next tested the effects of upregulating both signalling pathways on each developmental process. I partially activated ecdysone signalling, using EcRi, while upregulating IIS, with InR, in ovarian somatic cells (tj>EcRi, InR). Interestingly, TFCs were observed at 5 h AL3E in ovaries from tj>EcRi, InR fed on standard food (Figure 3.8G). This onset of TFC differentiation was not only earlier than that of control larvae (w1118; tj) (Figure 3.8), it was also significantly earlier than the onset of TF formation in tj>InR and tj>EcRi ovaries (p<0.0001, χ2= 45, df = 3, Chi-Square Test) (Figure 3.5G and Figure 3.6G). The rate of TF formation and of ovary growth was faster in ovaries from tj>EcRi, InR larvae than ovaries from either control or tj>EcRi larvae (Figure 3.8D, E and Table 3.2). Remarkably, even though ovaries grew at the same rate in both tj>EcRi, InR and tj>InR larvae, the rate of TF formation was faster in tj>EcRi, InR ovaries (Table 3.2). In summary, activating both IIS and ecdysone signalling pathways in ovarian somatic cells of well-fed larvae induced an earlier onset of TFC differentiation, and promoted a greater increase in the rate of TF formation than activating each signalling pathway individually (Table 3.1 and 3.2). This led me to hypothesize that activating both signalling pathways may overcome most of the effects of poor nutrition. When tj>EcRi, InR larvae were fed on sucrose-only food between 5 and 15 hours AL3E, the rate of TF formation was significantly faster than the control (w1118; tj) ovaries (Figure 3.8H). In fact, the rate of TF formation in ovaries from tj>EcRi, InR larvae was faster than all previous genetic manipulations in poorly fed larvae (Table 3.3). While at 5 hours AL3E 61 Chapter 3 Figure 3.8: The interplay between IIS and ecdysone signalling pathways. (A-G”) Shown is terminal filaments (TFs) marked with En (grey).(A-C”, F, G) Ovaries from larvae reared on standard food: (A-A”, F) w1118, tj (control), (B-B”) tj>EcRDN, InR and (C-C”, G) tj>EcRi, InR.(D) Number of forming TFs and (E) ovary volume of ovaries from larvae reared on standard food. (F’-F”, G’-G”) Ovaries from larvae transferred to sucrose-only food at 5 h AL3E: (F’-F”) w1118 , tj (control) and (G’-G”) tj>EcRi, InR. (H) Number of forming TFs and (I) ovary volume of ovaries from larvae fed on sucrose-only food. n ≥8 ovaries for all genotypes. Plotted values represent means and error bars show 95% confidence intervals of means. In some cases, error bars are too small to be seen. ANCOVAs: **p<0.01, ***p<0.001, ns non-significant. L3: third instar larvae; AL3E: after L3 ecdysis. Scale bar: 20µm. 62 Chapter 3 the ovary volume of tj>EcRi, InR ovaries was bigger than the control (p<0.001, Kruskal-Wallis test), I did not observe any further increase in ovary volume after transferring tj>EcRi, InR larvae to sucrose-only food (Figure 3.8I). Together, these results show that, in poor nutritional conditions, activating both IIS and ecdysone signalling pathways in ovarian somatic cells increases the rate of TF formation, but not ovary growth. Table 3.1: Pairwise comparisons of the rate of ovary growth in larvae fed on standard food. Shown are p-values (** p<0.01, ***p<0.001, ns non-significant; ANCOVAs using Holm’s p-value adjustment) for each pairwise comparison. Genotypes tj>EcRDN, InR tj>EcRi, InR w1118,tj 0.0001 *** 0.0001 *** tj>EcRDN 0.0014 ** - tj>EcRi - 0.0007 *** tj>InR 0.0014 ** 0.8450 ns tj>EcRi, InR 0.0005 *** - Table 3.2: Pairwise comparisons of the rate of TF formation in larvae fed on standard food. Shown are p-values (** p<0.01, ***p<0.001, ns non-significant; ANCOVAs using Holm’s p-value adjustment) for each pairwise comparison. Genotypes tj>EcRDN, InR tj>EcRi, InR w1118,tj 0.0026 ** 0.0029 ** tj>EcRDN 0.0055 ** - tj>EcRi - 0.0055 ** tj>InR 0.0030 ** 0.0055 ** tj>EcRi, InR 0.0029 ** - Table 3.3: Pairwise comparisons of the rate of TF formation in larvae fed on sucrose alone. Shown are p-values (** p<0.01, ***p<0.001, ANCOVAs using Holm’s p-value adjustment) for each pairwise comparison. Genotypes tj>EcRi, InR w1118,tj 0.0001 *** tj>EcRi 0.0017 ** tj>InR 0.0017 ** 63 Chapter 3 3.4 Discussion Hormonal regulation underlies most, if not all, well-studies cases of developmental plasticity (Beldade et al., 2011). However, how hormones change an organ’s response to environmental conditions over developmental time and how this influences their plastic response are poorly understood. In this chapter, I provide evidence that two hormonal signalling pathways – the IIS and ecdysone signalling pathways – act during critical weight to change the sensitivity of the developing ovary to nutritional conditions and thus, regulate the plastic response of ovariole number. In the previous chapter, I have shown that pre-critical weight larvae fed on sucrose-only food delay the onset of TFC differentiation (see Chapter 2). Both IIS and ecdysone signalling pathways regulate the onset of TFC differentiation; altering either IIS or ecdysone signalling in the ovary altered the timing of the onset of TFC differentiation. However, the effects of IIS on the onset of TFC differentiation depended on ecdysone signalling. Ovaries in which IIS was upregulated while ecdysone signalling was repressed delayed the onset of TFC differentiation as much as ovaries in which only ecdysone signalling was repressed. Even though ecdysone signalling was required to induce the onset of TFC differentiation, the two pathways appear to interact in a complex manner to regulate this process. Simultaneously upregulating both IIS and ecdysone signalling in the ovary resulted in earlier onset of TFC differentiation than upregulating either pathway on its own. Potentially, nutrition, via IIS, may modify the sensitivity of the ovary to ecdysone signalling. Under high levels of IIS, the ovary may require lower levels of ecdysone signalling to induce the onset of TFC differentiation, resulting in earlier onset. Further studies will be required to fully understand the nature of the link between IIS and ecdysone signalling in this developmental process. Additionally, my results suggest that unliganded EcR/Usp represses genes involved in the onset of TFC differentiation. Knocking down EcR in the ovary eliminates the repressive function of the receptor, and induces 64 Chapter 4 4.1 Introduction In recent years, it has become increasingly clear that the environment not only selects among phenotypes, but it also alters developmental trajectories, sometimes in profound ways, to generate astonishing variation in morphology, behaviour, physiology and life history (Beldade et al., 2011; Pfennig et al., 2010; Stearns, 1989). Such environmentally-induced traits may allow the persistence of a population in a novel and challenging environment, and can gradually become stabilized and integrated with the rest of the phenotype by means of genetic accommodation (WestEberhard, 2003). A possible outcome from such adaptive refinement of environmentally-induced phenotypes is that the developmental processes that generate phenotypic variation among populations should be to some degree similar to the developmental changes that emerge as a response to environmental variation (Wund, 2012). The remarkable plastic capacity, paired with a substantial genetic variation both within and between species, makes ovariole number in Drosophila an exciting model to address this issue. In previous chapters, I examined how larval nutrition shapes ovariole number in an outbred population of D. melanogaster. A detailed characterization of ovary development during third instar (L3) larval stages revealed that nutrition influences ovariole number by altering three distinct developmental processes during ovary development: i) the onset of terminal filament cell (TFC) differentiation, which marks the beginning of ovariole development, ii) the rate at which new terminal filaments (TF) are formed, and iii) the rate of ovary growth. Importantly, the nutrition-dependent developmental transition, critical weight, alters the types of developmental processes nutrition can affect: while starving larvae before critical weight delays the onset of TFC differentiation and ovary growth, changes in nutrition after critical weight reduce the rates of ovary growth and of TF formation. Here, I sought to explore whether changes in one or all of the three developmental processes are responsible for the divergence in ovariole number among Drosophila species. The intra/interspecific variation in ovariole number is correlated with the colonization of diverse ecological niches. In general, Drosophila species 71 Chapter 4 that feed on a narrow range of food sources have fewer ovarioles than species with a generalist diet (Green and Extavour, 2014; Kambysellis and Heed, 1971). Moreover, ovariole number shows substantial latitudinal and altitudinal clinal variation on different continents, providing further evidence that ovariole number is under selection (Capy et al., 1993; Gibert et al., 2004; Wayne et al., 2005). The repleta group is one of the largest species group in Drosophila genus comprising over 100 species with striking morphological and behavioural diversity (Durando et al., 2000; O’Grady and Markow, 2012; Smith et al., 2012). Particularly interesting is the cactophilic D. mojavensis, which has emerged as an important system for understanding the genetic basis of local adaptation, providing significant insights into the early events associated with ecological speciation. D. mojavensis is found in four geographically isolated areas in the arid regions of southwestern USA and northwestern Mexico and utilizes necrotic tissues of distinct local host cactus for both feeding and breeding (Figure 4.1). In the mainland Sonoran Desert, D. mojavensis uses the organ pipe cactus, Stenocereus thurberi, and occasionally shares the columnar cina cactus, S. alamosensis, with sister species D. arizonae. Although organ pipe cactus is abundant in Baja California, D. mojavensis utilizes almost exclusively the pitaya agria cactus, S. gummosus, in this region. The population in Mojave Desert uses red barrel cactus, Ferocactus cylindraceous, while on Santa Catalina Island, where columnar cacti are absent, the prickly pear cactus species, Opuntia demissa and O. littoralis, serve as the host (Fellows and Heed, 1972a; Ruiz and Heed, 1988). The different cacti used by each population differ in a variety of chemical compounds, including free fatty acids and sterols, and in the bacterial and yeast communities that colonize and decompose the cactus tissue into food for these flies (Fogleman and Starmer, 1985; Fogleman et al., 1981; Starmer and Fogleman, 1986). Besides differences in chemical composition, the total size of the plant and their necroses also vary among cactus hosts. Larger cacti, such as the organ pipe cactus, contain stable necroses of large dimensions, whereas in small cacti, such as Opuntia sp., necroses are small and ephemeral (Breitmeyer and Markovw, 1998). Each cactus 72 Chapter 4 Figure 4.1: Distribution and cactus host use of the four D. mojavensis subspecies. Map of the southwestern United States and northwestern Mexico showing the geographical distributions of D. mojavensis baja,D. mojavensis wrigleyi,D. mojavensis mojavensis and D. mojavensis sonorensis and their respective host cactus (adapted from (Pfeiler et al., 2009)). thus offers a distinct ecological niche to the resident flies, which may drive genetic divergence between populations of D. mojavensis. Studies using cytological and molecular data have clearly shown strong genetic differentiation between the four populations. Yet, the evolutionary relationships between them, and which population is ancestral, have proven difficult to resolve (Machado et al., 2007; Reed et al., 2007; Ruiz et al., 1990; Smith et al., 2012). In addition to genetic variation, several physiological, morphological, and behavioural adaptations have been reported between populations, which has led to them being recently described as subspecies: D. mojavensis baja,D. mojavensis sonorensis, D. mojavensis mojavensis, and D. mojavensis wrigleyi (Figure 4.1) (Date et al., 2013; Etges et al., 2010; Pfeiler et al., 2009; Richmond et al., 2012). Previous studies have described that ovariole number shows minor variations (i.e. ovariole number ranges from 25 to 33) among populations 73 Chapter 4 of D. mojavensis collected from the field or reared in laboratory cactus (Markow, 1996). I therefore tested whether ovariole number varies between subspecies of D. mojavensis reared under standard laboratory conditions. I confirmed that ovariole number varies considerably between the four D. mojavensis subspecies. This finding prompted me to focus on two D. mojavensis subspecies that display the greatest divergence in ovariole number, D. moj. sonorensis and D. moj. wrigleyi. I first examined whether differences in ovariole number were associated with changes in three life history traits: female fecundity, female body size, and developmental time. Next, I compared the ovary development in each subspecies to test whether developmental processes known to regulate ovariole number in response to nutrition also underlie differences in ovariole number between the two subspecies. Finally, I investigated how environmental and genetic variation shape ovariole number in the two D. mojavensis subspecies. 4.2 Material and Methods 4.2.1 Species stocks The following species were obtained from the Drosophila Species Stock Center (University of California, San Diego), with the exception of one D. arizonae line which was kindly provided by Janelia Farm Research Campus (Ashburn, Virginia): D. arizonae,D. mojavensis baja,D. mojavensis mojavensis,D. mojavensis sonorensis,D. mojavensis wrigleyi and D. mulleri (Table 4.1). Fly stocks were maintained at 22°C in bottles on standard fly food (4.5% molasses, 7.2% sugar, 7% cornmeal, 2% yeast extract, 1% agar and 2.5% Nipagin solution). 4.2.2 Larval staging Adults were allowed to oviposit for 4-6 hours on fresh plates (60 ×15 mm Petri dish) containing standard fly food. Often, females would lay multiple eggs in very close proximity. When I controlled for overcrowding 74 Chapter 4 Table 4.1: Species stocks used in this study. Stock numbers from Drosophila Stock Center (DSSC) with expection of one D. arizonae line from Janelia Farm Research Campus (JFRC). Species/Subspecies Stock Number D. arizonae JFRC D. arizonae 15081-1271.29 D. moj. baja 15081-1351.30 D. moj. baja 15081-1352.34 D. moj. mojavensis 15081-1352.00 D. moj. mojavensis 15081-1352.01 D. moj. sonorensis 15081-1352.26 D. moj. sonorensis 15081-1352.32 D. moj. wrigleyi 15081-1352.14 D. mulleri 15081-1371.01 D. mulleri 15081-1379.30 by performing the same protocol as described in Chapters 2 and 3, this would favour bacterial contamination during the first and second larval instar, as larvae tended to aggregate in large groups while foraging and burrowing into a small region of the food plate. As larval density was shown to influence developmental time and thorax length in different populations of D. mojavensis (Etges and Heed, 1987), I controlled larval density at several time points throughout larval development to avoid both overcrowding and bacteria contaminations. This was done by cutting the food containing larvae into smaller pieces every other day and gently transferring them to fresh food plates (60 ×15 mm Petri dish). After egg laying (AEL), plates were maintained at 25°C in a 12 h light-dark cycle with 70% humidity. In these conditions, both D. moj. sonorensis and D. moj wrigleyi exhibit similar generation times (15-16 days from egg to adult). At day 6 AEL, larvae were selected 0-2 hours after ecdysis to L3 (AL3E) and transferred onto new food plates (40-60 larvae per plate) until the time of dissection [0 h, 20 h, 40 h, 60 h, 80 h AL3E and at pupariation (D. moj. sonorensis, 110 h AL3E; D. moj. wrigleyi, 104 h AL3E)]. Each time point was replicated at least twice. 75 Chapter 4 4.2.3 Immunocytochemistry, imaging and analysis Immunocytochemistry was performed as previously described (see Chapter 2). The primary antibody mouse anti-Engrailed (Developmental Studies Hybridoma Bank 4D9, 1:40) was used to identify terminal filament cells (TFCs). The following secondary reagents were used: Alexa 568 (Invitrogen, 1:200) and TRICT-Phalloidin (Sigma, 1:200). Samples were mounted on a poly-L-lysine-coated coverslip using FluoromountG (SouthernBiotech) and imaged using a Zeiss LSM 510 Meta confocal microscope. Measurements of total number of forming terminal filaments (TFs) and ovary volume were performed as previously in Chapter 2 and 3 described using ImageJ (NIH) and Adobe Photoshop (Adobe Systems). 4.2.4 Dietary manipulations Well-fed flies were allowed to deposit eggs overnight on plates containing standard fly food. Egg lays were repeated in two independent days with different parental flies to minimize possible parental effects on larval development. Eggs were collected and groups of 50 eggs were randomly transferred to vials containing one of the following food concentrations: standard fly food (100%), standard fly food diluted with 1% agar in a ratio of 1:2 (50%), 1:4 (25%) or 1:8 (12.5%). The total volume of food was the same in all food concentrations. Four replicates were performed for each food concentration and subspecies. Developing larvae were reared at 25°C in a 12 h light-dark cycle with 70% humidity until all individuals eclosed or died. Eclosed flies were transferred to fresh vials with standard fly food. 4.2.5 Experimental crosses to generate F1 and F2 hybrids Ten virgin females from D. moj. wrigleyi were grouped with five males from D. moj. sonorensis in vials on standard food (n=5 vials). The reciprocal cross was also made, but only four replicates laid fertile eggs (n=4 vials). Once females were sexually matured (around 8-10 days of age), they were left to lay eggs in food vials for 24 h. Overcrowding 76 Chapter 4 was controlled as previously mentioned. This procedure was repeated in three independent days to maximize the number of offspring. Eclosed F1 hybrids from each replicate were grouped (ten females and five males per vial) in fresh vials until sexually mature. F1 hybrids were then left to lay eggs using the same procedure as described above. A subset of F1 hybrid females were frozen after 8-10 days for ovariole number counts. A similar protocol was implemented to obtain ovaries from F2 hybrid females. 4.2.6 Measurements of life history traits: developmental time, adult body size, ovariole number and female fecundity The average time to pupariation was measured by transferring newly ecdysed L3 larvae to food vials (10-20 larvae per vial) and counting the number of larvae pupariating (immobile larvae with evaginated spiracles) every three hours on day 10-11 AEL. To quantify adult body size, pharate adults were individually weighed on a Sartorius SE2 ultramicrobalance. Males of D. mojavensis lack sex combs, so I was unable to distinguish males from females at this stage. After weighing, I kept each pharate adult inside an eppendorf until eclosion and females and males were identified based on their external genitalia. Ovariole number was counted for both ovaries from mated females (8-10 days of age) as previously described (see Chapter 2). To determine female fecundity in each subspecies, I implemented two experimental designs: i) three virgin females were grouped with two virgin males (grouped females, n=10 vials/subspecies), and ii) one virgin female was grouped with one virgin male (single females, n=20 vials/subspecies). Flies were transferred to fresh vials every day during the first 20 days after eclosion. The total number of eggs in each vial was counted daily. Variation between biological replicates differed significantly in single females in both subspecies (p<0.0001, D. moj. sonorensis and D. moj. wrigleyi, Kruskal-Wallis). 4.2.7 Statistical Analysis The distribution of residuals was tested for normality using Q-Q plots and the appropriate statistical test was applied. Differences in ovariole 77 Chapter 4 number among pairs of species were tested with ANOVA followed by Tukey’s multiple comparison test. To evaluate differences in total number of eggs laid during 20 days after eclosion, a Poisson regression model (using the R function ‘glm’) was performed. Correlations between female body size and ovariole number were tested using the Pearson’s Correlation coefficient. The distribution of residuals was non-normal for body size and developmental time, and thus, a Wilcoxon rank test was used for pairwise comparisons. Differences in the timing of the onset of TFC differentiation were tested with a Chi-squared test. ANCOVAs were used to evaluate differences in the rates of TF formation and of ovary growth. To test whether survival in four food concentrations was different among subspecies, Kruskall-Wallis rank test was performed. The same test was used to evaluate whether survival within each subspecies differs among food concentrations. To compare the effects of nutrition on ovariole number and female body size between subspecies, ANOVAs were performed. The distribution of residuals for developmental time was nonnormal. For this dataset, differences between subspecies were tested using a generalized liner model with Poisson distribution. All data analyses and statistics were conducted using R v3.1.2 (R Development Core Team, 2014). Plots were made using GraphPad Prism v6 (GraphPad Software). Statistical tests and p-values are indicated in the text and figures. 4.3 Results 4.3.1 Ovariole number diversity in species of the Drosophila mulleri subgroup To evaluate the diversity in ovariole number among D. mojavensis subspecies, I counted the number of ovarioles in females of each subspecies and of two additional cactophilic species of the repleta group, D. arizonae and D. mulleri. These two species are closely related to D. mojavensis, but utilize different cactus hosts and have broader geographical distributions (Oliveira et al., 2012; Reed et al., 2007). Additionally, I collected data from two independent lines of each species/subspecies, with exception 78 Chapter 4 Figure 4.2: Ovariole number in females of D. mojavensis wrigleyi is reduced relative to other D. mojavensis subspecies.Adult ovariole number in females of D. moj. baja (two lines, 15081-1351.30 and 15081-1352.34, n=14 and 27), D. moj. mojavensis (two lines, 15081-1352.00 and 15081-1352.01, n=10 and 31), D. moj. sonorensis (two lines, 15081-1352.32 and 15081-1352.26, n=70 and n=69) and D. moj. wrigleyi (one line 15081-1352.14, 14, n=69). D. arizonae (two lines, JFRC and 150811271.29, n=17 and 19) and D. mulleri (two lines, 15081-1371.01 and 15081-1379.30, n=21 and 21). The numbers in the graph legend represent the last digits from the stock number of each line used. from the Plotted values represent means, and error bars show 95% confidence intervals of the means. One-way ANOVA using Tukey’s pairwise test comparison (bars with same letters are not significantly different). of D. moj. wrigleyi, to validate that my measurements of ovariole number were species-specific (Table 4.1). Indeed, the variability between lines in each species/subspecies was not significantly different, ensuring that ovariole number is robust within a given species/subspecies (Figure 4.2). Conversely, ovariole number differed considerably among different members of the repleta group. The island subspecies, D. moj. wrigleyi, had the lowest number of ovarioles (28.1±3.3) when compared either to other D. mojavensis subspecies (with exception of D. moj. baja 15081-1351.30) or to other repleta species, D. arizonae and D. mulleri (Figure 4.2). On the other hand, the subspecies inhabiting the Sonoran Desert, D. moj. sonorensis, showed the highest number of ovarioles among D. mojavensis subspecies (39.6±2.8 in D. moj. sonorensis 15081-1352.26 and 40.1±3.8 in D. moj. sonorensis 15081-1352.32). Both D. arizonae (only line 15081-1271.29) and D. mulleri showed a similar ovariole number as both of the D. moj. sonorensis lines and D. moj. mojavensis line 150811352.00. Based on these observations, I focused on two D. mojavensis 79 Chapter 4 subspecies that display the greatest divergence in ovariole number, D. moj. wrigleyi and D. moj. sonorensis 15081-1352.32 (henceforth referred as D. moj. sonorensis and D. moj. wrigleyi). 4.3.2 Divergence of life history traits in D. moj. sonorensis and D. moj. wrigleyi Female fecundity Studies in several species of the melanogaster group demonstrated that the number of ovarioles correlates positively with the number of eggs females lay (see Chapter 2) (Klepsatel et al., 2013b,a; R’ kha et al., 1997; Kambysellis and Heed, 1971). I therefore asked whether females of D. moj. sonorensis, which have a higher number of ovarioles, lay significantly more eggs than D. moj. wrigleyi females. Preliminary observations revealed that flies of D. moj. sonorensis tend to aggregate in one side of the food plate and lay multiple eggs in a small region of the food plate. This oviposition behaviour was also observed in flies of D. moj. wrigleyi, although it was less conspicuous. Hence, to fully investigate whether female fecundity differs between the two D. mojavensis subspecies, I counted the number of eggs laid in each subspecies in two experimental groups: grouped females (three females and two males) and single females (one female and one male). The dynamics of egg laying in grouped females were erratic in both D. mojavensis subspecies, yet an interesting pattern emerged (Figure 4.3A). Eggs were laid in clutches in given days, with almost no eggs laid in the intervening days. This irregular pattern of egg laying was more noticeable in single females (Figure 4.3B). In two biological replicates, in particular, individual females laid more than one hundred eggs in a single day and very few eggs were oviposited in the remaining days (higher peaks in light blue in Figure 4.3B). Even though the dynamics of egg laying was similar in the two D. mojavensis subspecies, the total number of eggs laid over 20 days was significantly lower in D. moj. wrigleyi relative to D. moj. sonorensis, and this was seen in both grouped and single females (p<0.0001, Poisson regression model) (Figure 4.3C, D). 80 Chapter 4 Table 4.4: Two-away ANOVA model for ovariole number. Source of variation df F p-value Subspecies 1 335.4036 <0.0001 *** Food concentration 1 29.5408 <0.0001 *** Subspecies x Food concentration 1 0.4847 0.6966 ns Error 74 - - species of the melanogaster group (Bergland et al., 2008; Orgogozo et al., 2006; Wayne and McIntyre, 2002; Wayne et al., 2001). However, only a small number of candidate genes have been identified and functionally investigated (Green and Extavour, 2012, 2014; Orgogozo et al., 2006). Green and Extavour (2012) advocate that the identification of relevant candidate genes will benefit from a better understanding of the developmental mechanisms that generate diversity in ovariole number (Green and Extavour, 2012). Throughout this thesis, I extensively investigated the developmental mechanisms that contribute to variation in ovariole number. In this final section, I show my preliminary exploration of the genetic basis of intraspecific variation in ovariole number and female body size. For this purpose, I examined ovariole number and female body size in F1 and F2 hybrids between D. moj. sonorensis and D. moj. wrigleyi. In both reciprocal crosses, F1 hybrids showed intermediate values in ovariole number, suggesting that variation in ovariole number involves alleles with opposite dominant effects or without any dominance at all (Figure 4.7A, B). Female body size in F1 hybrids was skewed towards the smaller parental line, D. moj. wrigleyi (Figure 4.7C, D), which may indicate that alleles involved in regulating smaller body size are dominant over those that determine larger body size. However, even though larval density was controlled as carefully as possible, the number of larvae in each bottle varied substantially, which could contribute to a reduction in female body size in F1 hybrids. Thus, I cannot exclude the possibility that the smaller adult size of F1 hybrids is a consequence of the larval rearing environment. Interestingly, in the F2 hybrids the variance for ovariole number and female body size was similar to that observed in the F1 hybrids (Figure 4.7). 87 Chapter 4 Figure 4.7: Ovariole number and female body size in F1 and F2 hybrids. Relative frequency distribution (in percentages) of ovariole number (n=20-45) (A, B) and female body size (n=20-90) (C, D) in the parental lines, D. moj. sonorensis (light blue line) and D. moj. wrigleyi (light orange line), and the F1 (black line) and F2 hybrids (grey line) from both reciprocal crosses between D. moj. sonorensis and D. moj. wrigleyi. 88 Chapter 4 Thus, it is likely that few loci contribute to divergence in ovariole number and female body size between D. moj. sonorensis and D. moj. wrigleyi. More generations would be required to further break up possible linkage associations and obtain a broader range of phenotypes. 4.4 Discussion 4.4.1 Plastic responses and evolved variation in ovariole number By examining the ovary development in both subspecies, I found that all three developmental processes differ between D. moj. sonorensis and D. moj. wrigleyi (Figure 4.8A). The onset of TFC differentiation was delayed in D. moj. wrigleyi relative to D. moj. sonorensis. Such delay is not due to differences in developmental time between subspecies, since D. moj. wrigleyi larvae are the ones that develop faster. Even though the timing of the onset of TFC differentiation was different between species, the rate of TF formation was reduced only at the end of L3 development in D. moj. wrigleyi when compared with D. moj. sonorensis. This contrasts with the nutritional-induced differences during ovary development (Figure 4.8B) (see Chapter 2). Starving pre-critical weight larvae delays the onset of TFC differentiation and reduces the rate of TF formation throughout the post-critical weight phase of development. This correlation is further observed when either insulin/insulin-like growth signalling (IIS) or ecdysone signalling pathways are suppressed in the larval ovary of well-fed larvae (see Chapter 3). How is it that changes in the onset of TFC differentiation have no effect on the rate of TF formation between subspecies? A possible explanation is that once the first TFCs differentiate from the surrounding somatic cells in D. moj. wrigleyi, they rapidly intercalate into stacks and hence the number of forming TFs quickly increases. Additionally, the relative number of TFCs that initially emerged at the onset of TFC differentiation may also differ between subspecies. Perhaps at the onset of TFC differentiation, a higher relative number of TFCs is produced in D. moj. wrigleyi ovaries, 89 Chapter 4 Figure 4.8: Changes in distinct developmental processes underlie nutritional-induced and subspecies-specific variation in ovariole number. (A) Changes in the rates of TF formation and of ovary growth are responsible for ovariole number divergence between D. mojavensis subspecies. (B) Before critical weight, starvation delays the onset of TFC differentiation and arrests ovary growth. In contrast, the rates of TF formation and of ovary growth continue when post-critical weight larvae are poorly fed. However, these rates are considerably reduced, resulting in fewer ovarioles than well-fed larvae. which could contribute to a faster formation of new TFs in a short period of time. Even though the rate of TF formation is initially similar, at the end of L3 development the rate of TF formation is reduced, contributing to the divergence in ovariole number between D. moj. sonorensis and D. moj. wrigleyi (Figure 4.8A). Changes in the rate of TF formation were also observed when post-critical weight larvae were poorly fed (Figure 4.8B) (see Chapter 2). In addition, Hodin and Riddiford (2000) found that rearing D. melanogaster larvae on diluted medium only reduced the rate of TF formation at the end of the L3, similar to what I observed in D. 90 Chapter 4 moj. wrigleyi. The rate of ovary growth was considerably reduced in D. moj. wrigelyi when compared with D. moj. sonorensis. The total number of TFs greatly depends on the number of TFCs that differentiate from the somatic cell pool (Sarikaya et al., 2012). Therefore, a slower rate of ovary growth either in D. moj. wrigleyi or in starved post-critical weight larvae may gradually constrain the number of cells available to form TFCs, thereby reducing the rate of TF formation. Taken together, I propose that ovary growth rates have diverged between D. moj. wrigleyi and D. moj. sonorensis, generating differences in ovariole number by limiting the number of somatic cells available to form TFCs. In a previous study by Hodin and Riddiford (2000), the authors concluded that nutritionally-induced variation in ovariole number is restricted to alterations in the rate of TF formation, while changes in a broader range of developmental processes can account for variation in ovariole number both between populations and between species (Hodin and Riddiford, 2000). My results differ from this study in several ways. First, to investigate the developmental processes that account for nutritionalinduced differences, I reared L3 larvae in a medium containing only sugar and agar, which significantly reduced both ovariole number and body size and delayed developmental time. Under these conditions, the onset of TFC differentiation, the rate of TF formation and the rate of ovary growth were all affected. In contrast, Hodin and Riddiford (2000) reared larvae from egg to adult in 50% diluted food from its original concentration in order to avoid any change in body size and developmental time (Hodin and Riddiford, 2000). Such diluted food provides a less stressful environment for larvae to develop, and only altered the rate of TF formation at the end of larval development. Taken together, the range of developmental processes altered by nutrition depends partly on the severity of the nutritional conditions. Second, my analysis was performed with significant resolution throughout L3 larval development, which allowed me to detect differences in the onset of TFC differentiation, the rate of TF formation, and the rate of ovary growth between D. mojavensis subspecies and between nutritional 91 Chapter 4 responses. Hodin and Riddiford (2000), by contrast, inferred differences in the onset of TFC differentiation by counting the number of TFs at 24 h AL3E without any analysis on previous time points (Hodin and Riddiford, 2000). Moreover, the rates of TF formation were only compared during the wandering stage (i.e. when larvae leave the food to find a pupariation site) and no comparisons were performed amongst rates of ovary growth (Hodin and Riddiford, 2000). Third, Hodin and Riddiford (2000) compared five species in the melanogaster subgroup with a much greater range of ovary sizes than those studied here (Hodin and Riddiford, 2000). D. sechellia has severely reduced ovariole number relative to D. melanogaster (R’ kha et al., 1997; Orgogozo et al., 2006; Hodin and Riddiford, 2000; Green and Extavour, 2014). These differences in ovariole number are due to a smaller size of ovarian primordium and slower rates of TF formation in D. sechellia(Green and Extavour, 2012; Hodin and Riddiford, 2000). Hence, the difference in ovariole number between these two species is defined by changes in ovary development since embriogenesis. In my study, I focused on two subspecies of D. mojavensis with more moderate differences in ovary size. Here, I found that ovary size is indistinguishable between the two races at the onset of the L3. Differences in ovary development, including ovary growth rates and the rate of TF formation, arise only in the final larval stage. Thus, it would seem that much like for the plastic response, the number of developmental processes underlying intra/interspecific genetic variation depends on the magnitude of the difference in ovary size. Although differences in ovariole number both across species and across populations often correlate with observed plastic responses, these are not the only mechanisms known to generate variation in ovariole number. Green and Extavour (2012) found that differences in ovariole number between two strains of D. melanogaster resulted not from differences in ovary growth but from differences in the allocation of somatic cells, which have the potential to become TFC, to different somatic cell fates (Green and Extavour, 2012). In strains with lower ovariole numbers, a greater proportion of the somatic cells adopted a swarm cell fate than in strains 92 Chapter 4 with higher ovariole numbers. Whether this developmental difference corresponds to a plastic response to another type of environmental cue, such as temperature or hypoxia, remains to be discovered. 4.4.2 The relationship between ovariole number and female body size I have shown that variation in ovariole number between the two D. mojavensis subspecies is positively correlated with variation in female body size. However, the relationship between ovariole number and female body size is not always clear (Bergland et al., 2008; Green and Extavour, 2012; Hodin and Riddiford, 2000; Kambysellis and Heed, 1971; Robertson, 1956; Santos et al., 1992; Wayne et al., 1997). The strength of this relationship appears to vary with environmental factors, including nutrient availability and larval competition (Bergland et al., 2008; Kambysellis and Heed, 1971; Santos et al., 1992). For instance, in some specialist species, such as the cactophilic D. buzzatii and the Hawaiian D. mimica, a positive correlation between thorax length (a proxy index for body size) and ovariole number is found in wild caught flies, but not in laboratory-reared flies (Kambysellis and Heed, 1971; Santos et al., 1992). Furthermore, larvae of D. melanogaster reared in low yeast concentrations show a strong positive correlation between thorax length and ovariole number. This correlation is lost when larvae are fed on optimal nutritional conditions (Bergland et al., 2008; Wayne et al., 1997). Perhaps the most interesting case is that of temperature-induced differences in ovariole number and body size in populations of D. melanogaster. Larvae reared at higher temperatures produce smaller adults with reduced ovariole number. In contrast, adult from larvae grown at lower temperatures have larger body size, but similar reduction in ovariole number (Klepsatel et al., 2013a; Mirth and Shingleton, 2012). Future studies examining more species and a greater range of environmental conditions will lead to vast improvements in our understanding of the relationship between ovariole number and female body size. 93 Chapter 4 4.4.3 Generating a hypothesis on egg laying behaviour Ovariole number is an important determinant of female reproductive capacity; females with higher number of ovarioles can potentially lay more eggs (see Chapter 2) (Boulétreau-Merle et al., 1982; Klepsatel et al., 2013b; R’ kha et al., 1997). Indeed, D. moj. sonorensis females have a greater reproductive capacity than females of D. moj. wrigleyi. However, this difference in total number of eggs laid among the two subspecies of D. mojavensis only became apparent in the last days of the analysis, suggesting that other factors, such as rate of oogenesis or egg retention, influence female reproductive output. Moreover, substantial differences in male reproductive traits could also exist between the two D. mojavensis subspecies, which might contribute to the observed differences in the total number of eggs laid. Interestingly, both subspecies exhibited an erratic dynamic of egg laying, in which numerous eggs were laid in given days. The rate of egg production may offer a possible explanation for this egg laying behaviour. Ovaries can be defined as synchronous when all ovarioles have at least one mature egg (Kambysellis and Heed, 1971). One such example is found in D. mulleri, in which mature eggs develop synchronously and are oviposited in large clutches in a single day (Markow and O’Grady, 2008). A close examination of the number of eggs per ovariole may elucidate whether the observed egg laying behaviour in D. mojavensis subspecies is a direct result of a synchronous ovary. Alternatively, the abundance and distribution of suitable oviposition sites may influence the frequency of oviposition (Kambysellis and Heed, 1971). Females of D. mojavensis oviposit in necrotic cacti, which offers a moist and nutritious environment for larval development (Fellows and Heed, 1972b; Ruiz and Heed, 1988). The patchy distribution of cactus hosts and the ephemeral nature of necroses require that D. mojavensis efficiently disperse great distances to find the appropriate breeding site (Breitmeyer and Markovw, 1998; Pfeiler and Markow, 2011). In their natural setting, D. mojavensis females may retain mature eggs until a potential oviposition site is found, which might lead to an irregular egg laying behaviour. However, if postponing oviposition is advantageous when 94 Chapter 4 suitable resources are infrequently encountered, why is this behaviour maintained in laboratory conditions? The physiological mechanisms regulating the rate of egg laying could be genetically fixed and thus, even in the presence of abundant food resources, the dynamics of egg laying is unchanged. Finally, the standard fly food provided in laboratory conditions might not be optimal for D. mojavensis species. Although D. mojavensis is relatively easy to maintain under laboratory conditions, the necrotic cacti provides a complex nutritious environment with specific olfactory cues (Date et al., 2013) that may influence the frequency of oviposition. Further studies using extracts of necrotic cactus could provide new insights on egg laying behaviour in D. mojavensis subspecies. 4.5 Conclusions The aim of this chapter was to elucidate the developmental mechanisms underlying variation in ovariole number between subspecies of Drosophila mojavensis. Coupled with the extensive studies regarding their ecology, D. mojavensis subspecies represent an attractive model to address the early events leading to evolutionary diversification in ovariole number. Importantly, the results described in this chapter open up new avenues of research in diverse disciplines, including ecology (e.g. what underlies irregular egg laying behaviour?), cellular biology (e.g. what are the underlying mechanisms regulating proliferation, differentiation and intercalation of TFCs) and physiology (e.g. how different traits respond differently to nutrition?). Acknowledgments I would like to thank Christen Mirth and Nelson Martins for advice on statistical analyses. Liliana Vieira for making the four food concentrations. Christen Mirth for commenting and proofreading this chapter. The Unit of Imaging at IGC for all the assistance during image acquisition and analysis. The Fly Facility, in particular Liliana Vieira, for making fly food every week. 95 Chapter 4 96 Chapter 5 Figure 5.2: Alternative mechanisms might underlie differences in the onset of TFC differentiation between the two D. mojavensis subspecies. (A) The small ecdysone peak is secreted at different time points, resulting in differences in the timing of both the attainment of critical weight and the onset of TFC differentiation. (B) Critical weight is attained at the same time. However, ovaries from D. moj. sonorensis and D. moj. wrigleyi may show substantial differences in IIS activity, and thus, different thresholds of ecdysone sensitivity could exist between the two subspecies. As a result, the onset of TFC differentiation would occur at different time points. in harsh conditions) in the nematode Caenorhabditis elegans (Schaedel et al., 2012). Here, the authors described the existence of a responsive period during which worms can respond to the hormone dafacronic acid and induce alternative developmental programs. Both crowding and the presence of dauer pheromone raises the threshold of sensitivity to dafacronic acid necessary to induce adult development(Schaedel et al., 2012). Studies in the dung beetle Onthophagus taurus have elucidated how thresholds of hormone sensitivity and the responsive period can be altered by social conditions (Emlen and Nijhout, 1999). The Drosophila ovary now provides an opportunity to uncover the molecular mechanisms underlying these types of phenomena. 5.3 Plasticity and evolution in ovariole number and body size Not all traits show the same sensitivity to larval nutrition (Shingleton et al., 2009). One mechanism by which organs can change their nutritional sensitivity is by changing the levels of IIS activity (Tang et al., 2011; 103 Chapter 5 Shingleton and Tang, 2012; Koyama et al., 2013; Shingleton and Frankino, 2013)). For instance, Green and Extavour (2014) found that the ovaries of D. sechellia and D. melanogaster differ in the levels of phosphorylated Akt, and propose that this variation in the levels of IIS underlies the difference in ovariole number between the two species. Changes in the levels of FOXO activity, a negative regulator of IIS, are also known to mediate the nutritional sensitivity of developing organs (Tang et al., 2011). At low levels of foxo mRNA expression, growth of the wing discs is not inhibited even in larvae that are poorly fed, allowing flies to maintain a large wing size when nutrition and IIS is low (Figure 5.3A, B). Whilst moderate levels of foxo increase the nutritional sensitivity of the wing, at very high levels of foxo expression the wing shows reduced sensitivity to nutrition, and thus, flies bear small wings even when nutrition and IIS are high (Figure 5.3A, B) (Tang et al., 2011). Thus, plotting the degree of nutritional plasticity against IIS activity generates a bell-shaped curve. Using a similar logic to that used by Tang et al. (2011), I hypothesize that the levels of IIS activity regulate the nutritional sensitivity of ovariole number and female body size in the two subspecies of D. mojavensis. When larvae were reared across a range of nutritional conditions, female body size in D. moj. sonorensis was more sensitive to nutritional variation when compared with D. moj. wrigleyi (i.e. slopes of reaction norms were significantly different).Moreover, females of D. moj. sonorensis have higher number of ovarioles and larger body size relative to D. moj. wrigleyi females. I therefore propose that whole body of D. moj. sonorensis females might display moderate levels of IIS signalling, resulting in an increase in its sensitivity to larval nutrition (Figure 5.3C, D). Conversely, females of D. moj. wrigleyi might show low levels of IIS activity throughout the body, making them both smaller and less sensitive to nutritional variation (Figure 5.3C, D). In terms of their ovariole number, both subspecies show similar nutritional sensitivity (i.e. slopes of reaction norms were equal). Ovary growth rate in D. moj. wrigleyi is significantly reduced in comparison with D. moj. sonorensis, and hence, levels of IIS activity in the developing ovary are likely different between the two subspecies. This could occur if the levels of IIS activity were at 104 Chapter 5 Figure 5.3: The relationship between IIS activity and nutritional sensitivity. (A) Changes in the levels of FOXO activity can account for differences in nutritional sensitivity. Either low or high levels of foxo mRNA in the wing discs reduce their sensitivity to nutritional variation. In contrast, wing discs expressing moderate levels of foxo mRNA increase their nutritional sensitivity. (B) Therefore, the relationship between nutritional sensitivity and IIS activity shows a bell-shape curve. (C) Different plastic responses and genetic variation for body size between D. moj. sonorensis and D. moj. wrigleyi. (D) D. moj. sonorensis might show moderate levels of IIS activity, resulting in higher nutritional sensitivity for body size. (E) Similar plastic responses, but genetic variation for ovariole number between the two subspecies. (F) Levels of IIS activity may be at opposite extremes of the bell-shape curve, resulting in similar levels of nutritional sensitivity, but different ovary sizes. 105 Chapter 5 opposite extremes of the bell-shaped curve, resulting in similar levels of plasticity but different ovary sizes (Figure 5.3E, F). My work has focussed on the developmental and physiological mechanisms underlying variation in ovariole number. However, we are beginning to understand the genetic basis of organ and body size. Green and Extavour (2014) further demonstrated that changes in the activity of IIS can account for differences in ovariole number between D. sechellia and D. melanogaster: high IIS activity in the developing ovary of D. melanogaster can promote the formation of a higher number of ovarioles (Green and Extavour, 2014), presumably by increasing the rate of ovary growth. Furthermore, introducing one copy of D. melanogaster InR into a D. sechellia background increases the number of ovarioles in this species (Green and Extavour, 2014). A similar evolutionary change in IIS activity between D. moj. sonorensis and D. moj. wrigleyi could underlie differences in ovary growth, which ultimately result in differences in ovariole number. Together with the current literature on plasticity and evolution of organ size, my work underline that changes in the activity of a major hormonal pathway, the IIS, can account for plastic responses, and potentially facilitate evolutionary diversification among populations and species. Understanding the genetic mechanisms underlying evolutionary changes in the activity of IIS could be a fruitful avenue of research. Acknowledgments I would like to thank Christen Mirth for all the brainstorming until the last minute, which was crucial to clarify my random hypotheses that were discussed in this chapter, and for commenting and proofreading this chapter. 106 BIBLIOGRAPHY Baniahmad, a. and Tsai, M. J. (1993). Mechanisms of transcriptional activation by steroid hormone receptors. J Cell Biochem, 51(2):151– 156. Bateman, J. M. and McNeill, H. (2004). Temporal control of differentiation by the insulin receptor/tor pathway in Drosophila. Cell, 119(1):87–96. Beadle, G. W., Tatum, E. L., and Clancy, C. W. (1938). Food Level in Relation To Rate of Development and Eye Pigmentation in Drosophila Melanogaster. Biol. Bull., 75(3):447–462. Bear, A. and Monteiro, A. (2013). Male courtship rate plasticity in the butterfly Bicyclus anynana is controlled by temperature experienced during the pupal and adult stages. PLoS One, 8(5):e64061. Beldade, P., Mateus, A. R. a., and Keller, R. a. (2011). Evolution and molecular mechanisms of adaptive developmental plasticity. Mol. Ecol., 20(7):1347–63. Bergland, A. O., Genissel, A., Nuzhdin, S. V., and Tatar, M. (2008). Quantitative trait loci affecting phenotypic plasticity and the allometric relationship of ovariole number and thorax length in Drosophila melanogaster. Genetics, 180(1):567–82. Boulétreau-Merle, J., Allemand, R., Cohet, Y., and David, J. R. (1982). Reproductive strategy in Drosophila melanogaster: Significance of a genetic divergence between temperate and tropical populations. Oecologia, 53(3):323–329. 107 Brakefield, P. M., Gates, J., Keys, D., Kesbeke, F., Wijngaarden, P. J., Monteiro, A., French, V., and Carroll, S. B. (1996). Development, plasticity and evolution of butterfly eyespot patterns. Nature, 384(6606):236–42. Breitmeyer, C. and Markovw, T. (1998). Resource of availabity and population size in cactophilic Drosophila. Func. Ecol., 12:14–21. Britton, J. S., Lockwood, W. K., Li, L., Cohen, S. M., and Edgar, B. a. (2002). Drosophila’s Insulin/PI3-Kinase Pathway Coordinates Cellular Metabolism with Nutritional Conditions. Dev. Cell, 2(2):239–249. Brogiolo, W., Stocker, H., Ikeya, T., Rintelen, F., Fernandez, R., and Hafen, E. (2001). An evolutionarily conserved function of the Drosophila insulin receptor and insulin-like peptides in growth control. Curr. Biol., 11(4):213–21. Brown, H. L. D., Cherbas, L., Cherbas, P., and Truman, J. W. (2006). Use of time-lapse imaging and dominant negative receptors to dissect the steroid receptor control of neuronal remodeling in Drosophila. Development, 133(2):275–85. Bryant, P. J. and Levinson, P. (1985). Intrinsic growth control in the imaginal primordia of Drosophila, and the autonomous action of a lethal mutation causing overgrowth. Dev. Biol., 107(2):355–363. Caldwell, P. E., Walkiewicz, M., and Stern, M. (2005). Ras activity in the Drosophila prothoracic gland regulates body size and developmental rate via ecdysone release. Curr. Biol., 15(20):1785–95. Capy, P., Pla, E., and David, J. (1993). Phenotypic and genetic variability of morphometrical traits in natural populations of Drosophila melanogaster and D simulans. I. Geographic variations. Genet. Sel. Evol., 25(6):517–536. Charlesworth, B. (1996). The good fairy godmother of evolutionary genetics. Curr. Biol., 6(3):220. Cheng, L. Y., Bailey, A. P., Leevers, S. J., Ragan, T. J., Driscoll, P. C., and Gould, A. P. (2011). Anaplastic lymphoma kinase spares organ growth during nutrient restriction in Drosophila. Cell, 146(3):435–47. Cherbas, L. (2003). EcR isoforms in Drosophila: testing tissuespecific requirements by targeted blockade and rescue. Development, 130(2):271–284. Chevin, L. M., Lande, R., and Mace, G. M. (2010). Adaptation, plasticity, and extinction in a changing environment: Towards a predictive theory. PLoS Biol., 8(4). 108 Colombani, J., Andersen, D. S., and Leopold, P. (2012). Secreted Peptide Dilp8 Coordinates Drosophila Tissue Growth with Developmental Timing. Science (80-. )., 336(6081):582–585. Colombani, J., Bianchini, L., Layalle, S., Pondeville, E., DauphinVillemant, C., Antoniewski, C., Carré, C., Noselli, S., and Léopold, P. (2005). Antagonistic actions of ecdysone and insulins determine final size in Drosophila. Science, 310(5748):667–70. Date, P., Dweck, H. K. M., Stensmyr, M. C., Shann, J., Hansson, B. S., and Rollmann, S. M. (2013). Divergence in olfactory host plant preference in D. mojavensis in response to cactus host use. PLoS One, 8(7):e70027. Davidson, A. M., Jennions, M., and Nicotra, A. B. (2011). Do invasive species show higher phenotypic plasticity than native species and, if so, is it adaptive? A meta-analysis. Ecol. Lett., 14(4):419–431. Davies, R. B. (1987). Hypothesis testing when a nuisance parameter is present only under the alternative. Biometrika, 74(1):33–43. Delpuech, A. J.-m., Moreteau, B., Chiche, J., Pla, E., and David, J. R. (2011). Phenotypic Plasticity and Reaction Norms in Temperate and Tropical Populations of Drosophila melanogaster : Ovarian Size and Developmental Temperature Published by : Society for the Study of Evolution Stable URL : http://www.jstor.org/stable/2410320 . PHE. Society, 49(4):670–675. Durando, C. M., Baker, R. H., Etges, W. J., Heed, W. B., Wasserman, M., and DeSalle, R. (2000). Phylogenetic analysis of the repleta species group of the genus Drosophila using multiple sources of characters. Mol. Phylogenet. Evol., 16(2):296–307. Emlen, D. J. and Nijhout, H. (1999). Hormonal control of male horn length dimorphism in the dung beetle Onthophagus taurus (Coleoptera: Scarabaeidae). J. Insect Physiol., 45(1):45–53. Etges, W. J., De Oliveira, C. C., Noor, M. A. F., and Ritchie, M. G. (2010). Genetics of incipient speciation in Drosophila mojavensis. III. Life-history divergence in allopatry and reproductive isolation. Evolution (N. Y)., 64(12):3549–3569. Etges, W. J. and Heed, W. B. (1987). Sensitivity to larval density in populations of Drosophila mojavensis: Influences of host plant variation on components of fitness. Oecologia, 71(3):375–381. Feinberg, A. P. (2007). Phenotypic plasticity and the epigenetics of human disease. Nature, 447(7143):433–440. 109 Fellows, D. and Heed, W. (1972a). Factors Affecting Host Plant Selection in Desert-Adapted Cactiphilic Drosophila. Ecology, 53(5):850. Fellows, D. P. and Heed, W. B. (1972b). Factors Affecting Host Plant Selection in Desert-Adapted Cactiphilic Drosophila. Ecology, 53(5):850–858. Flatt, T. (2005). The evolutionary genetics of canalization. Q. Rev. Biol., 80(3):287–316. Fogleman, J. C. and Starmer, W. T. (1985). Analysis of the community structure of yeasts associated with the decaying stems of cactus. III.Stenocereus thurberi. Microb. Ecol., 11(2):165–73. Fogleman, J. C., Starmer, W. T., and Heed, W. B. (1981). Larval selectivity for yeast species by Drosophila mojavensis in natural substrates. Proc. Natl. Acad. Sci. U. S. A., 78(7):4435–4439. Forbes, A. J., Lin, H., Ingham, P. W., and Spradling, A. C. (1996). hedgehog is required for the proliferation and specification of ovarian somatic cells prior to egg chamber formation in Drosophila. Development, 122(4):1125–1135. Gancz, D. and Gilboa, L. (2013). Insulin and Target of rapamycin signaling orchestrate the development of ovarian niche-stem cell units in Drosophila. Development, 140(20):4145–54. Gancz, D., Lengil, T., and Gilboa, L. (2011). Coordinated regulation of niche and stem cell precursors by hormonal signaling. PLoS Biol., 9(11):e1001202. Gao, X., Zhang, Y., Arrazola, P., Hino, O., Kobayashi, T., Yeung, R. S., Ru, B., and Pan, D. (2002). Tsc tumour suppressor proteins antagonize amino-acid-TOR signalling. Nat. Cell Biol., 4(9):699–704. Garcia-Bellido, A. and Merriam, J. R. (1971). Parameters of the wing imaginal disc development of Drosophila melanogaster. Dev. Biol., 24(1):61–87. Garelli, A., Gontijo, A. M., Miguela, V., Caparros, E., and Dominguez, M. (2012). Imaginal discs secrete insulin-like peptide 8 to mediate plasticity of growth and maturation. Science, 336(6081):579–82. Gibert, P., Gibert, P., Capy, P., Capy, P., Imasheva, A., Imasheva, A., Moreteau, B., Moreteau, B., Morin, J. P., Morin, J. P., David, J. R., and David, J. R. (2004). Comparative analysis of morphological traits among. Genetica, 120:165–179. Gibson, G. and Dworkin, I. (2004). Uncovering cryptic genetic variation. Nat. Rev. Genet., 5(9):681–690. 110 Godt, D. and Laski, F. a. (1995). Mechanisms of cell rearrangement and cell recruitment in Drosophila ovary morphogenesis and the requirement of bric à brac. Development, 121(1):173–87. Green, D. a. and Extavour, C. G. (2012). Convergent evolution of a reproductive trait through distinct developmental mechanisms in Drosophila. Dev. Biol., 372:120–30. Green, D. A. and Extavour, C. G. (2014). Insulin signalling underlies both plasticity and divergence of a reproductive trait in Drosophila. Proc. Biol. Sci. / R. Soc., 281(1779):20132673. Hodin, J. (2009). She shapes events as they come: plasticity in female insect reproduction. In Whitman, D. W. and Ananthaskishnan, T., editors, Phenotypic Plast. insects Mech. Consequences. Hodin, J. and Riddiford, L. M. (1998). The ecdysone receptor and ultraspiracle regulate the timing and progression of ovarian morphogenesis during Drosophila metamorphosis. Dev. Genes Evol., 208(6):304–17. Hodin, J. and Riddiford, L. M. (2000). Different Mechanisms Underlie Phenotypic Plasticity and Interspecific Variation for a Reproductive Character in Drosophilids ( Insecta : Diptera ). Evolution (N. Y)., 54(5):1638–1653. Hu, X., Cherbas, L., and Cherbas, P. (2003). Transcription activation by the ecdysone receptor (EcR/USP): identification of activation functions. Mol. Endocrinol., 17(4):716–31. Ikeya, T., Galic, M., Belawat, P., Nairz, K., and Hafen, E. (2002). Nutrient-dependent expression of insulin-like peptides from neuroendocrine cells in the CNS contributes to growth regulation in Drosophila. Curr. Biol., 12(15):1293–300. Kambysellis, M. and Heed, W. (1971). Studies of Oogenesis in Natural Populations of Drosophilidae . I . Relation of Ovarian Development and Ecological Habitats of the Hawaiian Species. Am. Nat., 105(941):31–49. Kerkis, J. (1931). The growth of the gonads in Drosophila melanogaster. Genetics. King, R. C. (1970). Ovarian development in Drosophila melanogaster. Academic Press. King, R. C., Aggarwal, S. K., and Aggarwal, U. (1968). The development of the female Drosophila reproductive system. J. Morphol., 124(2):143–166. 111 Klepsatel, P., Gáliková, M., De Maio, N., Huber, C. D., Schlötterer, C., and Flatt, T. (2013a). Variation in thermal performance and reaction norms among populations of Drosophila melanogaster. Evolution (N. Y)., 67(12):3573–87. Klepsatel, P., Gáliková, M., De Maio, N., Ricci, S., Schlötterer, C., and Flatt, T. (2013b). Reproductive and post-reproductive life history of wild-caught Drosophila melanogaster under laboratory conditions. J. Evol. Biol., 26(7):1508–20. Kooi, R. E. and Brakefield, P. M. (1999). The critical period for wing pattern induction in the polyphenic tropical butterfly Bicyclus anynana (Satyrinae). J. Insect Physiol., 45(3):201–212. Koyama, T., Mendes, C. C., and Mirth, C. K. (2013). Mechanisms regulating nutrition-dependent developmental plasticity through organ-specific effects in insects. Front. Physiol., 4(September):263. Koyama, T., Rodrigues, M. a., Athanasiadis, A., Shingleton, A. W., and Mirth, C. K. (2014). Nutritional control of body size through FoxOUltraspiracle mediated ecdysone biosynthesis. Elife, 3:1–20. Laland, K., Uller, T., Feldman, M., Sterelny, K., Müller, G., Moczek, A., Jablonka, E., Odling-Smee, J., Wray, G., Hoekstra, H.E.Futuyma, Richard E. Lenski, D. J., Mackay, T. F. C., Schluter, D., and Strassmann, J. (2014). Does evolutionary theory need a rethink ? Nature, 514:161–164. Lanet, E., Gould, A. P., and Maurange, C. (2013). Protection of neuronal diversity at the expense of neuronal numbers during nutrient restriction in the Drosophila visual system. Cell Rep., 3(3):587–94. Lanet, E. and Maurange, C. (2014). Building a brain under nutritional restriction: insights on sparing and plasticity from Drosophila studies. Front. Physiol., 5(March):117. Layalle, S., Arquier, N., and Léopold, P. (2008). The TOR pathway couples nutrition and developmental timing in Drosophila. Dev. Cell, 15(4):568–77. Lengil, T., Gancz, D., and Gilboa, L. (2015). Activin signaling balances proliferation and differentiation of ovarian niche precursors and enables adjustment of niche numbers. Development. Machado, C. a., Matzkin, L. M., Reed, L. K., and Markow, T. a. (2007). Multilocus nuclear sequences reveal intraand interspecific relationships among chromosomally polymorphic species of cactophilic Drosophila. Mol. Ecol., 16(14):3009–24. 112 Truman, J. W., Hiruma, K., Allee, J. P., Macwhinnie, S. G. B., Champlin, D. T., and Riddiford, L. M. (2006). Juvenile hormone is required to couple imaginal disc formation with nutrition in insects. Science, 312(5778):1385–1388. Tu, M.-p. and Tatar, M. (2003). Juvenile diet restriction and the aging and reproduction of adult Drosophila melanogaster. Aging Cell, 2:327– 333. Waddington, C. (1959). Canalization of development and genetic assimilation of acquired characters. Nature, 183(4673):1654–1655. Warren, J. T., Yerushalmi, Y., Shimell, M. J., O’Connor, M. B., Restifo, L. L., and Gilbert, L. I. (2006). Discrete pulses of molting hormone, 20-hydroxyecdysone, during late larval development of Drosophila melanogaster: correlations with changes in gene activity. Dev. Dyn., 235:315–26. Wayne, M. L., Hackett, J. B., Dilda, C. L., Nuzhdin, S. V., Pasyukova, E. G., and Mackay, T. F. (2001). Quantitative trait locus mapping of fitness-related traits in Drosophila melanogaster. Genet. Res., 77(1):107–116. Wayne, M. L., Hackett, J. B., Mackay, T. F. C., and Mackay, T. E. C. (1997). Quantitative Genetics of Ovariole Number in Drosophila melanogaster. I. Segregating Variation and Fitness. Evolution (N. Y)., 51(4):1156–1163. Wayne, M. L., Korol, A., and Mackay, T. F. C. (2005). Microclinal variation for ovariole number and body size in Drosophila melanogaster in ’Evolution Canyon’. Genetica, 123(3):263–270. Wayne, M. L. and McIntyre, L. M. (2002). Combining mapping and arraying: An approach to candidate gene identification. Proc. Natl. Acad. Sci. U. S. A., 99(23):14903–14906. West-Eberhard, M. J. (2003). Developmental Plasticity and Evolution. Oxford University Press, New York. Wheeler, D. E. and Nijhout, H. (1983). Soldier determination in Pheidole bicarinata: Effect of methoprene on caste and size within castes. J. Insect Behav., 29(11):847–854. Wheeler, D. E. and Nijhout, H. (1984). Soldier determination in Pheidole bicarinata: Inhibition by adult soldiers. J. Insect Physiol., 30(2):127– 135. Whitman, D. W. and Agrawal, A. A. (2009). What is phenotypic plasticity and why is it important? In Whitman, D. W. and Agrawal, A. A., editors, Phenotypic Plast. Insects Mech. Consequences. 119 Wu, Q. and Brown, M. R. (2006). Signaling and function of insulin-like peptides in insects. Annu. Rev. Entomol., 51:1–24. Wund, M. a. (2012). Assessing the impacts of phenotypic plasticity on evolution. Integr. Comp. Biol., 52(1):5–15. Yamanaka, N., Rewitz, K. F., and O’Connor, M. B. (2013). Ecdysone control of developmental transitions: lessons from Drosophila research. Annu. Rev. Entomol., 58:497–516. 120