Obesity Predisposing Genes in Drosophila Melanogaster

OBESITY PREDISPOSING GENES IN DROSOPHILA MELANOGASTER:

THE METABOLIC FUNCTIONS OF SPLIT ENDS

KELSEY ELIZABETH HAZEGH

B.S., University of Denver, 2012

A thesis submitted to the

Faculty of the Graduate School of the

University of Colorado in partial fulfillment

of the requirements for the degree of

Doctor of Philosophy

Molecular Biology

2017

This thesis for the Doctor of Philosophy degree by

Kelsey Elizabeth Hazegh

has been approved for the

Molecular Biology Program

Tom Evans, Chair

Tânia Reis, Advisor

Aaron Johnson

Emily Bates

Jed Friedman

Date: December 15, 2017

Hazegh, Kelsey Elizabeth (Ph.D., Molecular Biology)

Obesity Predisposing Genes in Drosophila melanogaster: The Metabolic Functions of Split ends

Thesis directed by Assistant Professor Tânia Reis

ABSTRACT

Obesity is a result of excess energy storage in the form of triglycerides (TAGs).

Preventing obesity requires a precise balance between deposition into and mobilization from fat stores, which is tightly controlled by metabolic enzymes and their regulators. Genetic background plays a major role in the predisposition to obesity, however it is estimated that

<2% of interindividual variation in BMI can be explained by the genes identified so far. A recent unbiased forward genetic screen by Reis et al. identified 66 genes that altered

Drosophila larval fat levels. Here we describe the identification of expression pattern for 33 of these genes to identify those that may regulate the storage or utilization of fat in the larval fat body (FB). Nineteen of the genes express in the FB, fourteen of which were individually depleted by RNAi expression in the FB and tested for changes in larval fat levels by means of a buoyancy assay. Depletion of fatty acid binding protein (Fabp) in the FB results in decreased fat levels, matching previous results in the mammalian literature and serving as a proof of principle for our screening methods. Nuclear factor of activated T cells (NFAT) and

Alan shepard (Shep) were identified as having novel pro-fat storage roles in the FB. Shep also serves an opposing role in the brain where it promotes the usage of organismal fat and is furthermore regulated by the nutritional intake of the larvae.

Split ends (Spen), an RNA-binding protein previously implicated in transcriptional control of conserved signaling pathways, was also identified in this screen. We found that

iii

Spen function is necessary and sufficient to promote fat depletion in the fat body in a cell autonomous manner. Interestingly, despite being fat, larvae in which Spen is depleted from the FB are sensitive to starvation, suggesting that these animals are incapable of using their excess fat stores. Consistent with this phenotype, metabolomics and RNA sequencing demonstrate metabolic alterations in Spen-depleted FBs indicative of a defect in mobilization of TAGs and utilization of other metabolites (proteins and carbohydrates) as primary sources of energy. We further find that another Spen family member Spenito (Nito) plays an opposing role in fat storage. FB overexpression of an N-terminal Spen fragment containing the RNA Recognition Motifs (RRMs) and undefined middle region of Spen causes a dominant-negative high-fat phenotype, whereas there was no effect of overexpression of a C- terminal fragment containing only the conserved Spen paralog and ortholog (SPOC) domain.

Thus, the RRMs or other undefined N-terminal domain are required for the ability of overexpressed full-length Spen to deplete fat stores, and when overexpressed alone may sequester important Spen binding partners into non-functional complexes. We propose that

Nito, which contains RRMs and a SPOC domain but is much smaller than Spen, may act as a negative regulator of Spen function. We further find that levels of the mammalian Spen and

Nito orthologues correlate with body weight in a diet-induced obese mice, supportive of a model where Spen and Nito act as a counterbalance to finely tune fat storage. No other study has implicated Spen or Nito in the regulation of metabolism or body fat control. Our work provides new directions for understanding metabolic disease as well as a molecular handle to generate novel mechanistic insights into conserved genetic causes of obesity.

The format and content of this abstract are approved. I recommend its publication. Approved: Tânia Reis

To Mom and Dad, for your endless love and encouragement

and

To Micah, for your constant support and for

helping me to find the fun and laughter in all things.

ACKNOWLEDGEMENTS

I couldn’t have made it to this point in my career without the help and support from people in all corners of my life. First, I want to acknowledge and thank my mentor Tânia for her guidance. She always pushed me to be my best self and believed in me even when I had doubts. Her passion for science is infectious and I gained new appreciation for genetics through her; like she always says, “genetics is always right!” I’d like also to thank my fellow lab members over the years: Nick Haynes and Jeremy Mosher for introducing me to lab and making me so comfortable upon starting; Lauren Schmitt for Denver Biscuit Co outings and for whipping our lab into shape; Claire Gillette for joining the lab and taking on the Shep project; Darcy Marceau, Shruthi Sivakumar, Johnny Nguyen, Tracey Nguyen, and Taylor

Tomita for helping us with our research, even for a little while; and everyone for their help with egg collections every single weekend. I want to especially thank my best friends Vevian

Zhang and Brenna Dennison(!) for margaritas, pad thai, and for making me smile, even on bad days. Thanks also to Chris and Michelle Boyd for zillions of dinners and saving the world one Faded at a time. You guys gave me respite when research got tough and I’m so thankful to have you all in my life.

A huge reason for who I am today is due to my family. Their love and support over the years has made me realize that I can accomplish anything that I set my mind to. Kim showed me how much you can achieve when you follow your dream. Dad gave me my very first appreciation and passion for science and taught me how to use a microscope for the first time. Mom has been ever supportive, reassuring, and encouraging as I took on all these challenges. My family has always been my biggest fans and I cannot thank them enough for all they’ve ever done for me.

Finally, I want to thank Micah. He has been my best friend and my rock through the good times and the bad. He has been there for me every step of the way and been supportive even when it was inconvenient. I could not have managed without him and am so excited for all our future adventures together.

I love you all, and thank you for being there for me.

vii

TABLE OF CONTENTS

CHAPTER

I. INTRODUCTION ...... 1

The Obesity Pandemic ...... 1

The Genetics of Obesity...... 6

Lipid Metabolism ...... 12

Drosophila as a Model for Metabolism ...... 18

Alan shepard ...... 24

Split ends ...... 29

Summary ...... 36

II. MATERIALS AND METHODS ...... 38

Fly Strains and Husbandry ...... 38

Immunohistochemistry ...... 40

Density Assay ...... 41

Gas Chromatography Mass Spectrometry ...... 51

Glycogen Quantification ...... 52

Feeding Assay ...... 52

Activity Assay ...... 53

Mosaic Analysis ...... 53

Starvation Assay ...... 54

viii

RNA Library Preparation and Sequencing ...... 55

Metabolomics ...... 55

Murine Analysis ...... 56

III. SCREENING FOR OBESITY PREDISPOSING GENES ...... 58

Abstract ...... 58

Introduction ...... 58

Results ...... 61

Discussion ...... 77

IV. AN AUTONOMOUS METABOLIC ROLE FOR SPEN ...... 83

Abstract ...... 83

Introduction ...... 83

Results ...... 86

Discussion ...... 120

V. DISCUSSION ...... 128

Conclusions and Discussion ...... 128

Future Directions ...... 133

REFERENCES ...... 142

APPENDIX

A. Supplemental Methods...... 166

B. Spen Conservation ...... 167

C. Unpublished Data...... 176

LIST OF TABLES

TABLE

2.1 RNAi directory...... 39

3.1 Candidate gene expression in third instar larvae ...... 62

4.1 Upregulation of Spen and starvation genes ...... 100

4.2 Downregulation of Spen and starvation genes ...... 101

LIST OF FIGURES

FIGURE

1.1 The pathway to obesity ...... 3

1.2 Lipogenesis and lipolysis ...... 14

1.3 Systemic metabolic signaling in Drosophila ...... 20

1.4 Shep is highly conserved between organisms ...... 25

1.5 Spen and Nito contain two domains conserved between family members and

organisms ...... 30

3.1 Workflow for identification of genes involved in fat regulation ...... 64

3.2 Fabp is necessary for fat accumulation in the fat body ...... 66

3.3 NFAT is necessary for fat accumulation in the fat body ...... 68

3.4 Shep is necessary for fat accumulation in the fat body ...... 71

3.5 Shep expression in the brain is necessary but not sufficient to decrease organismal fat

levels ...... 73

3.6 Shep expression in the brain varies based on nutrition ...... 76

4.1 Spen autonomously decreases fat levels in the fat body ...... 87

4.2 Knockdown of Spen results in a low-density phenotype ...... 89

4.3 Decreased larval density in Spen mutants ...... 91

4.4 Spen overexpression does not alter behavior or starvation response ...... 93

4.5 Spen autonomously regulates fat levels in the fat body ...... 95

4.6 Spen regulates the breakdown of fat ...... 98

xii

4.7 Spen manipulation results in abnormal metabolism ...... 103

4.8 Spen depletion or overexpression alters acyl-carnitine and amino acid levels ...... 104

4.9 Domain analysis of Spen and Nito function in metabolism ...... 106

4.10 Ectopic expression of truncated Spen or Nito does not alter behavior ...... 109

4.11 Expression of ΔSPOC results in inappropriate fat storage in other organs ...... 111

4.12 Ectopic expression of truncated Spen or Nito does not alter FB cell or

LD morphology ...... 113

4.13 Nito autonomously promotes fat accumulation in the fat body ...... 116

4.14 Nito autonomously regulates fat levels in the FB ...... 118

4.15 Spen and Nito transcript levels are modulated by body fat levels in mouse adipose

tissue ...... 121

4.16 Model: Spen family members counter-regulate metabolism ...... 126

B1 Domain conservation in Spen proteins ...... 167

C1 RalA is necessary for fat accumulation in the fat body ...... 176

C2 Dietary effects on Drosophila development ...... 177

C3 Spen depletion in the FB results in early onset of death ...... 178

xiii

CHAPTER I

INTRODUCTION

The Obesity Pandemic

All organisms must maintain a homeostasis between energy intake and utilization.

When utilization exceeds intake, the organism must have energy stores to maintain themselves throughout the duration of the energy expenditure. These stores of energy must be used to fuel body maintenance, all behaviors, and growth during development. However, when energy storage in the form of fat vastly exceeds necessity, this is known as obesity.

Over the last several decades, obesity has become a severe problem. Rates of obesity have increased from 13.4% in 1960 to 37.9% in 2014 (Fryar et al., 2016a). More than 70% of the American population is overweight or obese, while 7.7% of people exhibit extreme obesity (BMI ≥ γ5 kg/m2) (Fryar et al., 2016a). The problem effects children as well as adults. One of every six American children between the ages of six and 19 are obese (Fryar et al., 2016b). Although America has the highest rates of obesity in the world, it is not limited to America or even limited to first world countries. Dozens of countries throughout the world have obesity rates over 20%, including Mexico, Australia, Turkey, Latvia, South Africa, and many more (OECD, 2017). Obesity also increases the risk of many comorbidities. A meta- analysis of 89 different studies identified significant evidence for 18 comorbidities of obesity. These include: type 2 diabetes, all cancers (except esophageal in women), all cardiovascular diseases (except congestive heart failure), asthma, gallbladder disease, osteoarthritis, and chronic back pain (Guh et al., 2009). Often multiple comorbidities manifest in a single patient causing a slew of health problems and an increased chance of death. With so many Americans exhibiting obesity and the many comorbidities, the medical

1 costs associated with the disease rose to a staggering $147 billion as of 2008 (Finkelstein et al., 2009).

The drastic increase in obesity rates is not due to any one cause, but to a combination of several. According to Bouchard et al., among these culprits are the built environment, social environment, changing behaviors, and biology (Figure 1.1) (Bouchard, 2007). The built environment refers to the physical world that people have built around themselves. For example, the growth and expansion of cities have made travel by automobile by far the most common, while elevators have made the use of stairs much rarer. The social environment refers to the pressure people put on one another. The advertising campaigns by fast food and soda giants encourages people to overindulge and puts pressure on people to consume.

People have begun adopting obesogenic behavior with the development of new technologies.

Behaviors like watching television or playing video games rather than playing outside or like ordering a supersize meal because it is only $0.24 more make a significant contribution to society’s rise in obesity rates. The change in America’s economy away from manufacturing and toward the information sector results in more people spending all day sitting behind a desk and computer rather than up and doing physical labor. All combined, these result in a perfect storm for the development of obesity.

However, there is another huge factor at play: biology. There have been dozens of studies conducted to interrogate the contribution that our genetic backgrounds plays with our predisposition to and development of obesity. J. V. Neel proposed the “Thrifty Gene

Hypothesis” in 196β to try to explain the phenomenon (Neel, 1962). Neel suggested that the feast-and-famine cycle throughout the majority of human history has promoted the selection of genotypes that promote a quick insulin trigger and efficient fat storage. Our hunter-gather

Figure 1.1: The pathway to obesity. The modern built and social environments create an environment wherein obesogenic behavior is both easy to adopt and socially acceptable. Depending upon the genetic predisposition of the person adopting the obesogenic behaviors, varying levels of fat will be stored as a result, eventually leading to the development of obesity (Adapted from (Bouchard, 2007)).

3 and agriculturalist ancestors often had periods of famine that led to widespread starvation and lower fecundity. Those that were genetically primed to be resistant to these periods would have had greater success in outliving and out-reproducing others leading to a selection for the trait today (Neel, 196β, 1999; O’Dea, 1995; Stipp, β011). In today’s world of abundant food supply, this selection is no longer advantageous, leading to the risk and/or development of extreme obesity and diabetes (Neel, 196β, 1999; O’Dea, 1995; Pijl, β011). This genetic predisposition to store fat more easily along with the built and social environments of our age are widely believed to be the cause of the obesity pandemic (Figure 1.1).

Although the “Thrifty Gene” hypothesis has a lot of evidence and support, there are those that dispute the hypothesis. They claim that the rate of starvation during famines was too low to promote the selection of the “thrifty genes” and that there have been too few famines throughout human history to have had such a major impact on the selection of our genetics (Speakman, 2006). Furthermore, they claim that although fecundity is reduced during periods of famine for those that do not have a thrifty genotype, the periods following the famine show a compensation in the fertility rates that would wash out any effect the famine had on selecting thrifty genes (Speakman, 2008). Instead, they propose the “drifty gene” hypothesis. This hypothesis proposes that for the majority of human history, there were lower and upper weight bounds to protect against starvation and predation respectively.

In the advent of society and human-built fire two million years ago, the risk of predation decreased drastically thereby removing the upper weight limitation. This allowed for genetic drift to occur leading to some, but not all, people with a much higher risk for developing obesity (Speakman, 2008).

Although there is some dispute as to the mechanism of the inheritance of obesity- predisposing genetics, there is no question that genetics plays a very major role in the development of obesity. A number of studies have been done over several decades to determine contributions of nature (i.e. genetics) versus nurture (i.e. environment). Twin studies have provided a lot of evidence for the heritability of obesity with most studies showing 50-90% genetic involvement (Bouchard et al., 1985; Cornes et al., 2007; Hur, 2007;

Lajunen et al., 2009). Interestingly, the rate of heritability appears to change with age, with rates between 31-82% under 12 years of age and substantially higher rates of 81-90% between 12 and 19 years of age (Lajunen et al., 2009). This increase over time may be explained by the assertion of individuality and personal choice as the children age. This would better express the impact of their genetics compared to younger cohorts who rely more upon their parent or guardians’ decisions for food or physical activity. On the other hand, estimates of heritability among adolescents appear to be higher than those of adults (Maes et al., 1997). This is likely due to the common environment experienced by adolescent twin pairs raised in the same household and therefore partially the result of common environmental effects. Perhaps a more convincing group of studies was those of twins reared apart. These showed consistent heritability rates between 50-70% (Allison et al., 1994, 1996;

MacDonald and Stunkard, 1990; Maes et al., 1997; Price and Gottesman, 1991; Stunkard et al., 1990). A recent analysis of over 87,000 twin pairs from 45 different cohorts across the world confirmed these findings (Silventoinen et al., 2016). The proportion of BMI explained by genetic factors increased from ~40% in four-year-olds up to 75% in 19-year-olds.

Common environment effects on BMI were not observed after 15 years of age (Silventoinen et al., 2016). Despite large variations in BMI in twins across the world (e.g. America

5 compared to east Asia), the rate of heritability of obesogenic genes was similar (Silventoinen et al., 2016). These results confirm that genetics plays a major role in the predisposition to obesity in humans across the globe.

An alternate approach is adoption studies which are well suited to tease out the impact of the environment on BMI. Most adoption studies are partial, meaning only the adopted children and their adoptive parents are a part of the study. All the studies conducted found a significant correlation between parents and their biological children as well as between biological siblings. However none of the studies found any significant correlation between parents and adopted children or between non-biological siblings (Biron et al., 1977;

Hartz et al., 1977; Silventoinen et al., 2010; Withers, 1964). This is compelling evidence that environment plays a limited role in the heritability of obesity predisposition. Furthermore, complete studies that include both biological and adoptive parents of children agree that children gain the majority of risk of obesity from their biological parents rather than the environment of their adoptive home (Silventoinen et al., 2010; Sorensen et al., 1992).

Adoption studies concur with twin studies that the effect of common environmental factors disappear once the children reach adolescence and are at liberty to make their own decisions.

The Genetics of Obesity

Despite a clear indication that genetics plays a major role in the heritability of obesity predisposition, there are relatively few causal or contributing genes that have been identified.

Rare mutations and candidate gene studies have provided evidence that there are varying degrees of genetic determination of obesity, ranging from strong genetic predisposition to genetic resistance to obesity (Loos and Bouchard, 2003). There are also some cases of monogenic forms of obesity, where one gene mutation is dominantly inherited and manifests

6 as obesity, among other symptoms (Loos and Bouchard, 2003). For example, microdeletions within the promoter exon of the small nucleoriboprotein N (SNRPN) on chromosome 15 in humans results in Prader-Willi syndrome (PWS), which is the most common of human obesity syndromes (Ming et al., 2000; Ohta et al., 1999). This syndrome is autosomal dominant and is characterized by obesity as well as short stature, small hands and feet, mental retardation, and hyperphagia. This occurs in 1:25,000 people (Loos and Bouchard,

2003). Other monogenic forms of obesity include Albright hereditary osteodystrophy (AHO) derived from imprinted mutations in the guanine nucleotide-binding protein, α-stimulating activity polypeptide 1 (GNAS1) and Bardet-Biedl syndrome (BBS), caused in part by a single methionine to arginine mutation in the BBS1 gene (Loos and Bouchard, 2003). Despite knowing the causative genes in these syndromes, the mechanisms of the mutant gene products and their effect on energy imbalance have yet to be characterized. In an attempt to dissect the relationship between these obesity-causing syndromes and common obesity with clinically normal subjects, one group studied the chromosomal regions implicated in the syndromes in otherwise healthy sibling pairs with extreme obesity phenotypes. No correlation was found between the number of alleles shared between siblings within these chromosomal regions and their BMI, suggesting that these gene regions are not the same ones involved in common obesity (Reed et al., 1995).

Only rarely is a single gene causative of common obesity. Among these few is melanocortin 4 receptor gene (MC4R). In a study of 500 children with severe obesity, 5.8% had a mutation in MC4R (Farooqi et al., 2003). These mutations are dominantly inherited and complete loss of function results in a more severe obesity phenotype than partial loss of function (Farooqi et al., 2003). Other more rare forms of monogenic (non-syndromic) obesity

7 include leptin (LEP), leptin receptor (LEPR), perilipin (PLIN), fat mass and obesity associated (FTO), and others.

Leptin and leptin receptor were both found to be important genes for the development of obesity in 1973, although when identified they were called obesity (ob) and diabetes (db)

(Coleman, 1973, 1978; Coleman and Hummel, 1973). Mice homozygous for these mutations weighed two to three times more than wild type mice and stored five times as much fat, even when fed the same diet. Both of these mice are hyperphagic and develop severe diabetes

(Coleman, 1973, 1978; Coleman and Hummel, 1973; Friedman and Halaas, 1998; Friedman and Leibel, 1992). In some unique parabiosis experiments, wherein mutant animals (ob/ob or db/db) were surgically joined to either a wild type animal or another mutant (db/db or ob/ob respectively), Coleman determined that the db mice were overproducing a circulating satiety factor that they were unable to sense themselves. Conversely, the ob mice, though able to sense the satiety factor, were unable to produce it themselves (Coleman, 1978; Coleman and

Hummel, 1973; Grayson and Seeley, 2012). Nearly 20 years later, obesity was dubbed leptin and diabetes was dubbed leptin receptor (Zhang et al., 1994). A further 20 years after that, the Drosophila homolog of leptin was identified to be unpaired 2 (Upd2) (Rajan and

Perrimon, 2012). Patients that exhibit obesity due to a mutation in the leptin gene can be effectively treated with the daily administration of recombinant leptin (Licinio et al., 2004;

O’Rahilly et al., β00γ). Although leptin is a rare case wherein therapy can dramatically alleviate patients’ obesity, it gives hope that newly identified obesogenic genes will lead to the development of useful therapeutics in the future.

Perilipin (PLIN) is another example of a potential monogenic cause of obesity. The

PLIN family consists of five proteins (PLIN1-5) that coats lipid droplets (LDs) to aid in the

8 uptake of fatty acids (FAs) and the regulation of lipolysis. They are found mostly in adipocytes and skeletal muscle cells (Morales et al., 2017). In 2004, four PLIN1 single nucleotide polymorphisms (SNPs) were studied in 734 white subjects and it was determined that there was a significant correlation between two of the PLIN SNPs (SNP 5 and 6) and body fat percentage in women (Qi et al., 2004). Interestingly, there was no association found between PLIN polymorphisms and men. A follow up study by this group was able to show a similar correlation with PLIN SNP 6 polymorphisms (rs1052700) in Malaysian and Indian patients (Qi et al., 2005). This indicated that PLIN may be a genetic determinant for obesity.

More recently, >300 adolescents were studied to determine if PLIN polymorphisms influenced the risk for obesity. This group found that PLIN SNP 6 was significantly more common in obese adolescents (Tokgoz et al., 2017). Although in agreement that polymorphisms in PLIN affect the risk of obesity, a human trial on >100 subjects found a decrease in weight associated with PLIN SNPs (Soenen et al., 2009). It is unclear which of the above SNPs are hypo- or hyper-morphic. Conversely, a few studies have found opposing results. Perilipin knockout in mice results in 30% reduction in adipose tissue mass and elevated lipolysis (Tansey et al., 2001). Furthermore, perilipin-like protein LSD2 Drosophila mutants are lean (Grönke et al., 2003; Teixeira et al., 2003), whereas PLIN1 Drosophila mutants are hyperphagic and obese (Beller et al., 2010). These contradictions may be due to species differences in the role of perilipins, discrepancies in the types of measurements taken between the studies, or other mitigating genetic factors not controlled for. Regardless, it is agreed that PLIN is risk factor for obesity and further studies will help to elucidate the role it plays and potentially help to develop a therapeutic for those patients whose obesity is due to alterations in this gene.

Perhaps the most prevalent genetic contributor to common obesity found to date is the fat mass and obesity associated (FTO) gene. Genome-wide linkage scans found that

Chromosome 16q, the region that contains FTO, correlates with BMI (Geller et al., 2003; Wu et al., 2002). Several genome-wide association studies (GWAS) conducted in 2007 found a very significant correlation between SNPs in the FTO gene and BMI (Dina et al., 2007;

Frayling et al., 2007; Loos and Yeo, 2014; Scuteri et al., 2007). One of these groups estimates that polymorphisms in the FTO gene could be responsible for up to 22% of common obesity (Dina et al., 2007). Some studies have suggested that FTO has an essential role in lipolytic activity in fat cells or regulating cell growth while others suggest a role in food intake or preferences (Loos and Bouchard, 2008). Levels of FTO mRNA are positively correlated with levels of leptin, perilipin, and visfatin, all of which are risk factors for obesity in and of themselves (Zabena et al., 2009). The potential molecular mechanism by which

FTO mutations may be causing increased risk of obesity was elucidated in 2015. The risk allele (rs1421085-T to -C allele substitution) disrupted the binding of the transcriptional repressor ARID5B. This caused upregulation of IRX3 and IRX5, which in turn resulted in an increase in adipocyte size, decreased mitochondrial content, and decreased basal oxygen consumption rate (Claussnitzer et al., 2015). Despite these results, a follow up study was unable to find evidence that mutations in FTO reduce the fat oxidation within the adipocytes

(Blauw et al., 2017). FTO may likely be influencing obesity risk through other pathways than whole-body fat oxidation and more research will need to be done to determine the specific molecular mechanism of the association between FTO variants and obesity.

One promising avenue of research is FTO’s role in N(6)-methyladenosine (m(6)A) modification. This is the most abundant modification in mammalian mRNA and long non-

10 coding (lnc) RNA in which methyl groups are added by a methyltransferase complex to an

RNA molecule post-transcriptionally. This may have a profound impact on gene expression regulation, mRNA splicing, export, stability, and immune tolerance (Niu et al., 2013; Pan,

2013). FTO is the first identified gene that reverses this modification by functioning as a demethylase (Fu and He, 2012; Fu et al., 2014; Meyer and Jaffrey, 2014; Niu et al., 2013;

Pan, 2013). It performs this function in tandem with ALKBH5. Together, these demethylases along with methyltransferase complexes with METTL3, METTL14, WTAP, and Rbm15 can dynamically alter the “epitranscriptome” thereby effecting any number of cellular processes

(Meyer and Jaffrey, 2014; Niu et al., 2013; Zhao et al., 2014). In seeking to understand the commonalities between FTO’s role in obesity and FTO’s role in m(6)A modifications, several groups investigated how FTO might affect adipogenesis through this pathway. The results showed that FTO is necessary for adipogenesis and pre-adipocyte differentiation

(Merkestein et al., 2015; Zhang et al., 2015; Zhao et al., 2014). Furthermore, the catalytic demethylase activity of FTO was necessary for this function (Zhang et al., 2015; Zhao et al.,

2014) and regulated it through the splicing of RUNX1T1-S (Merkestein et al., 2015; Zhao et al., 2014). This provides evidence of at least one method through which FTO may be regulating fat levels and contributing to common obesity.

Although there are a number of examples of monogenic forms of obesity, common obesity is usually due to the heredity of multiple genes leading to the predisposition of obesity, and is therefore polygenic (Hainer et al., 2008). Genetic linkage studies, genome- wide association studies, and candidate gene studies are all being conducted in order to tease out many of the other more moderate players in the predisposition to obesity (Bell et al.,

2005; Comuzzie and Allison, 1998; Day and Loos, 2011; Loos, 2009). A growing list of

11 these genetic contributors was published as the Human Obesity Gene Map up until 2005 and the last update contained more than 600 genetic loci that had been implicated in obesity in some way (Rankinen et al., 2006). Since then, many more genes have been added to the list and a comprehensive map of genes involved in human obesity was published in 2016

(Jagannadham et al., 2016). Despite this huge list of genes involved in the predisposition to obesity, it is estimated that <2% of interindividual variation in BMI can be explained by the genes identified so far (Loos, 2009).

It is clear that genetics plays a major role in the predisposition and development of obesity. Although a number of obesogenic syndromes and monogenic forms of obesity have been identified, the causes and contributors to the majority of common obesity remain elusive. With rates of obesity continuing to increase and over 2/3 of Americans classified as overweight or obese, further work on this subject is crucial to identify additional genetic players. This will help to understand the origins and development of the disease more comprehensively and lead to increased potential for therapeutic treatments.

Lipid Metabolism

Obesity is caused by an excess of stored fat that is primarily stored in adipocytes.

Adipocytes are cells that make up the white adipose tissue (WAT) of mammals and the fat body (FB) of insects including Drosophila (Azeez et al., 2014). The main function of adipocytes is to maintain energy homeostasis by storing or mobilizing triglycerides (TAGs) depending on the availability of energy (Ali et al., 2013). When energy intake exceeds utilization, a number of events take place that lead to the accumulation of fat. These include the storage of dietary fatty acids or the de novo production of lipids, called lipogenesis.

Triglycerides (fatty acids joined by a glycerol backbone) are formed when excess glucose is

12 converted to acetyl-CoA through glycolysis (Figure 1.2A). The acetyl-CoA is then converted into malonyl-CoA by acetyl-CoA carboxylase (ACC), which is further processed by fatty acid synthase to create palmitate, a 16-carbon saturated fatty acid. This can be further processed to create other fatty acids, as well as to be processed into fatty acyl-CoA molecules and finally into triglycerides (Figure 1.2A) (Kersten, 2001). Lipogenesis is regulated by a couple of master genes including sterol regulatory element binding proteins (SREBPs) and peroxisome proliferator activated receptors (PPARs) (Ali et al., 2013; Kersten, 2001).

SREBPs are transcription factors that control the expression of enzymes involved in fatty acid, TAG, cholesterol, and phospholipid synthesis (Ali et al., 2013). PPARs are a family of ligand-activated transcription factors in the nuclear hormone receptor family. All four members of the family (α, , δ, and ) are involved in fatty acid metabolism, although PPAR is specifically involved in adipocyte differentiation and fatty acid storage (Tyagi et al., 2011).

These master regulators can be activated by molecules secreted as a result of food ingestion, including insulin, fatty acids, and leptin. Once upregulated, SREBP and PPAR can regulate the transcription of downstream genes and thereby increase lipogenesis and TAG storage.

Once created, TAGs are stored within lipid droplets (LDs) primarily in adipocytes where they serve as reservoirs for energy generation and other cellular processes like membrane synthesis (Walther and Farese, 2012). LDs are coated in LD-associated proteins like PLINs,

ATGL (a lipase homologous to brummer in Drosophila (Grönke et al., 2005)), and many others which help to form and stabilize the LD, regulate lipid storage and lipolysis, and contribute to intracellular signaling, protein degradation, and other cellular processes (Itabe et al., 2017; Xu et al., 2017). An excess of LD loading as in relation to obesity can lead to an increase in bioactive lipids like ceramides and diacylglycerols (DAGs) in the bloodstream,

Figure 1.2: Lipogenesis and lipolysis.

Figure 1.2: Lipogenesis and lipolysis. (A) Lipogenesis begins with a molecule of glucose which is processed through glycolysis to produce Acetyl-CoA. This is further processed by two enzymes to produce palmitate and thereby other fatty acids. Once conjugated to Coenzyme A, the fatty acyl-CoA is combined with glycerol-3-phospate to produce LPA. After a series of enzymatic modifications, TAG is formed and stored in LDs. (B) Lipolysis begins from a signal from AKH, which initiates a signaling pathway resulting in the activation of lipases like bmm and dHSL. These lipases break down TAGs into DAGs and FAs. DAGs can be further broken down in the hemolymph by lipoprotein lipase (adapted from (Trinh and Boulianne, 2013)). (C) To begin the process of -oxidation, FAs are taken up from the hemolymph and conjugated to carnitine by CPT1 to produce fatty acyl- carnitines. This allows the FAs to enter the mitochondrial matrix by CACT. CPT2 reverses the carnitine conjugation, allowing carnitine to be shuttled back into the cytoplasm by CACT. The FAs then enter a four-step enzymatic process to remove a two-carbon chain from the FA as acetyl-CoA, which is then broken down further by the TCA cycle. The remaining FA chain, now two carbons shorter, undergoes the process again until the whole molecule has been reduced (adapted from (Stephens et al., 2007)). Enzymes are depicted in red font. ACC, Acetyl-CoA Carboxylase; AGPAT, 1-Acyl-Glycerol-3-Phosphate Acyltransferase; DAG, Diacylglycerol; FA, Fatty Acid; FASN, Fatty Acid Synthase; GPAT, Glycerol-P Acyl- Transferase; LPA, Lyso-Phospatidic Acid; Mdy, Midway; PA, Phosphatidic Acid; PAP, PA Phosphatase; TAG, Triacylglycerol.

15 which has been connected to insulin resistance (Holland et al., 2007; Samuel and Shulman,

2012; Yu et al., 2002), a common comorbidity of obesity. This increase of TAG storage in the LDs of adipocytes is necessary for the development of obesity.

While excess fat can be a result of increased lipogenesis, a defect in TAG breakdown can similarly lead to excess fat storage. In mammals, this process, known as lipolysis, involves the activation of the hormone-sensitive lipase (HSL) and PLIN by phosphorylation by protein kinase A (PKA). With the phosphorylation of PLIN, activated HSL can translocate to the LD surface to begin the hydrolysis of TAGs (Holm, 2003). Mammalian adipose triglyceride lipase (ATGL) is a second crucial lipase necessary to mobilize the TAGs within the LDs for catabolism (Holm, 2003; Zimmerman et al., 2004). This process is similar in Drosophila. Lipolysis is stimulated by the release of adipokinetic hormone (AKH), a neuropeptide secreted from the corpora cardiaca (CC) in response to starvation (Figure 1.2B)

(Lee and Park, 2004). This signals the G protein-coupled receptor AKHR thereby increasing levels of cAMP and activating PKA. This signaling cascade can then activate lipases like dHSL and brummer (bmm), the Drosophila homolog of ATGL, to mediate lipid hydrolysis in the LDs (Figure 1.2B) (Grönke et al., 2005, 2007; Kühnlein, 2012). With a pivotal role in lipolysis, bmm is kept under strict transcriptional control. Levels of bmm are regulated by

FOXO in response to decreased insulin signaling when energy intake is low (Wang et al.,

2011), as well as by Target of rapamycin (TOR), a nutrient sensor in the fly (Luong et al.,

2006). A defect in lipolysis results in the inability to hydrolyze TAGs stored in the adipocytes and therefore no reduction in storage. In an instance where lipid storage is functioning well but lipolysis is defective, adipocytes will accumulate higher levels of fat, contributing to the development of obesity.

Once mobilized, there is a highly regulated process whereby TAGs are broken down to produce utilizable energy. In order to be made available to other tissues and cells in the body, the TAGs must be hydrolyzed into DAGs by lipases like ATGL and bmm. DAGs can be further reduced into free fatty acids (FFAs) in the fly by the action of lipoprotein lipase

(Figure 1.2B) (Trinh and Boulianne, 2013). Once mobilized, DAGs and FFAs can enter circulation to be taken up by other tissues for energy production. Upon uptake to the cell, the fatty acid forms a compound with coenzyme A (CoA), becoming a fatty acyl-CoA (Figure

1.2C) (Stephens et al., 2007). In order to be taken up into the mitochondria for -oxidation, the fatty acyl-CoA must then be conjugated to carnitine. This is accomplished by carnitine palmitoyltransferase 1 (CPT1), a transmembrane protein found on the outer mitochondrial membrane. CPT1 catalyzes the reversible conjugation between the fatty acyl-CoA and the carnitine, forming a fatty acylcarnitine (Murthy and Pande, 1987). This is the rate limiting step of -oxidation (Stephens et al., 2007). This conjugation then allows the fatty acid to be shuttled into the mitochondrial matrix by carnitine acylcarnitine translocase (CACT), whereupon the carnitine conjugation is reversed by carnitine palmitoyltransferase 2 (CPT2) to produce a fatty acyl-CoA and free carnitine (Figure 1.2C) (Bezaire et al., 2006; Pande,

1975). The free carnitine is shuttled back to the cytoplasm to repeat the process via CACT while the fatty acyl-CoA can continue to the -oxidation pathway. The breakdown of the fatty acid by -oxidation is a repeated four-step process. A series of four enzymes (acyl-CoA dehydrogenase, enoyl-CoA hydratase, 3-hydroxyacyl-CoA dehydrogenase, and -ketoacyl-

CoA thiolase) work together to break off a 2-carbon length of the fatty acid off from the long-chain fatty acyl-CoA in the form of acetyl-CoA (Figure 1.2C) (Bartlett and Eaton,

2004). This acetyl-CoA can then enter the TCA cycle in order to be broken down further

17 while the remainder of the fatty acid repeats the four-step enzymatic process until the entire length has been broken into multiple acetyl-CoA molecules.

This highly regulated process of lipid storage and hydrolysis must maintain a homeostasis in order to maintain a healthy weight and fat storage level. If either lipid storage or lipolysis becomes dysregulated, it will result in a higher amount of fat stored within the adipocytes. For example, mutations in the lipase bmm result in vastly larger lipid droplets within each cell of the fat body as well as significantly higher amounts of TAGs stored in the fly (Grönke et al., 2005). Conversely, overexpression of bmm depletes fat stores (Grönke et al., 2005). Dysregulations on the fat storage side can cause alterations in fat levels as well. In

Drosophila, the gene midway (mdy) encodes a diacylglycerol acyltransferase (DGAT), which converts DAGs into TAGs to be stored within LDs (Buszczak et al., 2002). Mutations in mdy results in a decrease in stored fat and overall lean phenotype (Buszczak et al., 2002).

Alterations in the expression of genes involved in either the storage or breakdown of fats or any master regulators thereof may result in a change in the levels of stored fat and result in an obese phenotype.

Drosophila as a Model for Metabolism

Drosophila have been used increasingly in the last several years as a tool to study questions in metabolism and human disease. Throughout its history as a model organism, flies have contributed to the discovery and characterization of several critical signaling pathways, including Notch, Wingless (Wg/Wnt), Hedgehog (Hh), Epidermal Growth Factor

Receptor (EGFR), Tumor Necrosis Factor Alpha (TNF-α), and others (Bier, 2005; Ruden and

Lu, 2006). Flies make an excellent model to examine the molecular mechanisms involved in these pathways as well as the developmental and pathological consequences of them.

Dysregulation of these central signaling pathways or of other genes often results in the development of human diseases. Approximately 75% of all human disease genes have a well conserved homologue in Drosophila (Bier, 2005; Chien et al., 2002; Fortini et al., 2000;

Reiter et al., 2001). Genes involved in more basic and conserved processes like metabolism are significantly more likely to be conserved between humans and flies (Bernards and

Hariharan, 2001). All of the basic metabolic and signaling pathways are conserved between

Drosophila and mammals making discoveries in flies easily applicable to mammalian systems.

Flies also have very similar physiology to mammals with similarly acting organs and cells. For example, the Drosophila FB performs the function of both the WAT and liver of a mammal by storing TAGs within LDs and serving as a sink for glycogen, the stored form of carbohydrates (Figure 1.3) (Padmanabha and Baker, 2014). Under starvation conditions, lipids are stored in specialized cells called oenocytes, which perform hepatocyte-like functions (Gutierrez et al., 2007). The midgut functions as the stomach and intestine and absorbs and digests nutrients (Baker and Thummel, 2007). Flies also have complex cross-talk pathways between different organs similar to mammals. Neuropeptides like Neuropeptide F

(NPF, homologue of Neuropeptide Y in mammals) are released from the brain in response to hunger to promote feeding and foraging behaviors (Figure 1.3) (Maule et al., 1995; Wu et al., 2005). Insulin signaling from the insulin producing cells (IPCs, analogous to mammalian pancreatic -cells) in the brain in the form of Insulin-Like Peptides 2, 3, and 5 (dILPs 2, 3, and 5) initiate the uptake of nutrients after a meal (Figure 1.3) (Brogiolo et al., 2001;

Rulifson et al., 2002). After reaching satiety, the FB signals to the brain to stop feeding behavior by releasing molecules like Upd2 and dILP6 (Figure 1.3)

Figure 1.3: Systemic metabolic signaling in Drosophila. In order to maintain energy homeostasis, the organs of the larva communicate through complex cross-talk pathways. In the absence of nutrient intake, AKH (Drosophila glucagon) is released from the corpora cardiaca (CC) and signals to the FB via the AKHR. This mobilizes stored glycogen and TAGs and results in the release of trehalose and DAGs into the hemolymph to be used for energy. In response to hunger, the brain releases NPF to initiate feeding. In response to feeding, dILPs 2, 3, and 5 are released from the insulin producing cells (IPCs) in the brain and signal to the FB to initiate the uptake of energy and the storage of TAGs and glycogen. Upon sensing high nutrient conditions, the FB signals to the brain by releasing Upd2 (Drosophila leptin) and dILP6. Under starvation conditions, TAGs are stored in oenocytes. (Adapted from (Padmanabha and Baker, 2014))

(Bai et al., 2012; Rajan and Perrimon, 2012). These are only some examples of cross-organ signaling pathways that regulate metabolism in Drosophila. Such signaling complexity is comparable to mammalian signaling and provides additional evidence that Drosophila are a useful source for learning about the molecular mechanisms behind metabolism.

While maintaining physiological complexity, flies provide a simpler model to study metabolic questions. The Drosophila genome is less complex than mammals and there is less gene redundancy (Liu and Huang, 2013). This is beneficial when dissecting out gene functions as there is less chance of other genes compensating for an intentionally manipulated gene. Furthermore, the model is easier and faster to work with than rodent models. The life cycle of flies is much shorter, at approximately 10 days from egg lay to adult eclosion when reared at 25°C (Ashburner et al., 2005; Greenspan, 2004). The animals also reach sexual maturity very quickly – within a few hours of eclosion – leading to a very quick generation time. Finally, there are incredible tools available for manipulating the

Drosophila genome in any spatial or temporal fashion, allowing for dissection of gene function with a level of resolution not possible in mammals (Baker and Thummel, 2007).

With improving technologies, studies in metabolomics and lipidomics are becoming more quantitative and detailed (Carvalho et al., 2012; Cox et al., 2017).

In wielding such a powerful model, progress is being made in the characterization of genes involved in metabolism and obesity. For example, adipose60 is a gene characterized in flies that results in increased TAG levels and starvation resistance when mutated (a naturally occurring mutation) and decreased fat levels when overexpressed (Hader et al., 2003). The significance of this gene in mammalian metabolism were confirmed a few years later when

Suh et al. confirmed that the mammalian ortholog adipose show similar fat storage

21 phenotypes as well as changes in insulin resistance to match (Suh et al., 2007). A number of full and partial genome-wide screens have been performed in Drosophila in the last decade to determine additional obesity-associated genes. Guo et al. asked what genes were involved in regulating LD formation and maintenance. In performing an RNA interference (RNAi) screen, they found that 1.5% of all genes were involved in this process in some way, bringing to light many new proteins that were unknown to have a role in lipid storage and mobilization (Guo et al., 2008). A couple years later, Pospisilik et al. tested over 11,000 heat shock (hs) inducible RNAi lines for their effect on the fat storage effects on adult

Drosophila. They found over 500 genes caused a significant change in the levels of stored triglycerides as determined by a colorimetric enzymatic assay, of which 62% had human orthologues (Pospisilik et al., 2010). Among their findings was hedhehog, which is the first time the gene has been found to be a major regulator of fat body TAG storage (Pospisilik et al., 2010). Around that same time, Reis et al. performed a partial genome wide screen using

P-element insertion lines to ask which genes caused an increase in fat levels in third instar larvae when disrupted. Approximately 900 lines that disrupted 500 distinct genes were tested.

Of these, 66 genes including Sir2 resulted in a significantly fat phenotype when disrupted as determined by larval buoyancy (Reis et al., 2010). Most recently, Baumbach et al. performed an RNAi screen in adult flies to find additional regulators of body fat. Nearly 6,800 genes were targeted for knock down in the fat body and 77 of these resulted in strong changes in adiposity, of which 75% had not been shown to have this role previously (Baumbach et al.,

2014). The group found several genes that were altered were components of the Store-

Operated Calcium Entry (SOCE) and went on to define intracellular calcium concentration as a determinant of body fat (Baumbach et al., 2014). This role for calcium signaling in

22 adiposity has been confirmed by several labs since (Bi et al., 2014; Moraru et al., 2017;

Subramanian et al., 2013).

Using Drosophila as a model to determine new genetic contributors to obesity has been and will continue to be fruitful. Owing to the similarities in physiology between flies and humans and the vast amount of conservation in human disease and metabolism genes, it is very likely that information obtained in flies will inform on the genetics of human obesity as well as to provide potential therapeutic targets to treat those with a genetically-based obese phenotype. The ease with which flies can be utilized to study genetics makes identification and characterization of these genes faster and cheaper than it would have otherwise been in mammalian models. With more studies being conducted on these questions around the world, it is clear that we are getting closer to understanding the genetic basis of fat storage and obesity. However, much remains to be done.

The genetic screens performed in flies mentioned above found a huge number of genes involved in fat regulation. However in each case, only a small number were followed up with characterization. Many genes remain to be dissected for their mechanism of action.

Included in this list are Alan shepard (Shep) and Split ends (Spen). Shep was identified as a regulator of obesity by Reis et al. in 2010. The homozygous viable mutant line tested resulted in a very strong floating score, indicating a higher level of stored fat (Reis et al., 2010).

Similarly, Spen was identified in the same screen when the hypomorph line tested resulted in positive buoyancy (Reis et al., 2010, unpublished). Furthermore, Spen was identified years later by Baumbach et al. in their genome-wide RNAi screen for adult fat level regulators

(Baumbach et al., 2014). Although identified as contributors, the molecular functions of these genes have not been characterized for their role in fat regulation. This work seeks to gain a

23 better understanding of the functions for Shep and Spen in regulating fat levels in Drosophila melanogaster.

Alan shepard

Alan shepard (Shep) encodes an RNA binding protein that is highly conserved between flies and mammals (Figure 1.4). It contains eight isoforms that result in six different protein products. These isoforms vary mostly at the N-terminus of the protein with minor variations at the C-terminal end as well. All isoforms contain two RNA Recognition Motifs

(RRMs) that allow Shep to bind to single-stranded RNA molecules (Figure 1.4). The human ortholog of Shep was the first to be identified by two different groups in 1994. These groups called the proteins scr2 and scr3 (suppressor of cdc2) and MSSP-1 and MSSP-2 (myc single stranded binding proteins) respectively (Kanaoka and Nojima, 1994; Negishi et al., 1994;

Takai et al., 1994). Scr2, MSSP-1, and MSSP-2 are now known to be three different isoforms of what is now called RBMS1 (RNA binding motif, single stranded interacting protein 1), while scr3 is the ortholog of the very similar family member RBMS2 (Fujimoto et al., 2001).

The Shep protein belonging to the fly was discovered several years later in a gravitaxis screen. The group found that mutations in the shep locus impeded the ability of the fly to perceive and respond to gravity (Armstrong et al., 2006). Based on this phenotype, Shep was named after the first American astronaut to go into space.

Since its discovery, Shep and its orthologues have been implicated in a number of different processes and cellular functions. The human ortholog of Shep regulates cell cycle progression from G1 to S phase through two different mechanisms. This regulation can take place through the indirect suppression of cdc2 mutations (Kanaoka and Nojima, 1994) as well as by forming a ternary complex with c-Myc and Max, thereby abrogating the ability of

Figure 1.4: Shep is highly conserved between organisms.

Figure 1.4: Shep is highly conserved between organisms. Shep has two mouse orthologues (Rbms1 and Rbms2) and two human orthologues (RMBS1 and RBMS2). These proteins are highly conserved, although Shep is longer with additional amino acids at the N-terminus. Shep family members contain two RNA-recognition motifs (RRMs) that are fully conserved across organisms (black circles).

26 c-Myc/Max to bind to the E-box sequence (Niki et al., 2000; Takai et al., 1994). Similarly,

RBMS induces apoptosis in a dose-dependent manner (Iida et al., 1997). The RRMs of

RBMS are necessary for its function in both of these processes (Iida et al., 1997; Niki et al.,

2000). Rbms1 also regulates c-Myc in mice, where it has been shown to regulate the release of estradiol from ovarian granulosa cells (Yin et al., 2012).

The best characterized function of Shep is neuronal development and remodeling.

Shep’s first identification in flies as a gene necessary for the response to gravity suggested a role in nervous system development (Armstrong et al., 2006). Since then, several groups have focused on characterizing Shep’s role in neurons. Shep is highly expressed in the central nervous system (CNS) of Drosophila. Shep has been shown to function as a negative regulator of gypsy insulator activity in the CNS (Matzat et al., 2012). gypsy and other chromatin insulators reorganize chromatin in a cell type-specific manner to control gene expression. Shep inhibits gypsy by altering the cellular localization of the insulator complex in the CNS, although Shep has no effect on the insulator complex in other tissues like muscle

(Matzat et al., 2012). Recently, Shep has been shown to regulate dendrite length and branching, axon organization in the ventral nerve cord, and organization of neuronal clusters in the peripheral nervous system (Schachtner et al., 2015), processes that take place during embryonic and early larval development. Moreover, Shep has been identified in two screens for nervous system expression within olfactory and peptidergic neurons (Chen et al., 2014;

Tunstall et al., 2012). Shep expression overlaps with neurons that secrete dILPs, Furin1

(Fur1), PHM, and other neuropeptides that are necessary for many different aspects of development and metabolism (Chen et al., 2014). Shep has also been directly linked to metabolism in a screen for obesity-predisposing genes, where it was determined that

27 mutations in Shep result in increased levels of larval fat as determined by larval buoyancy

(Reis et al., 2010).

Shep has been linked often to metabolism in humans as well. Within the last seven years, five different groups have found associations between the Shep (RMBS) locus and fat levels (Huffman et al., 2015; Mollah and Ishikawa, 2010; Sajuthi et al., 2016), type 2 diabetes (Sajuthi et al., 2016; Zhu et al., 2015), and dietary intake (Chu et al., 2013). In each case, mutations or deficiencies in the locus resulted in increased fat levels, increased BMI, or risk for type 2 diabetes. Taken together, this implicates Shep as an anti-obesity gene; when

Shep is fully functional, it decreases levels of stored fat, but when it is dysfunctional, it results in increases of fat levels and risk of comorbidities of obesity like type 2 diabetes.

Although it is unclear how Shep is controlling metabolism and stored fat levels, a meta- analysis of a population-based discovery cohort found an association with the genomic region nearby RBMS and increased caloric intake of carbohydrates (although a slight decrease in fat and protein intake) (Chu et al., 2013). This may suggest that Shep regulates levels of fat by modulating feeding behavior. This supposition is supported by the fact that

Drosophila Shep overlaps in expression with neuropeptide-secreting cells (Chen et al., 2014).

One likely candidate of overlapping expression is with NPF secreting cells, which control the feeding behavior of Drosophila. These peptidergic neurons are located on the dorsal surface of the lobes of the brain, in which there is ubiquitous Shep expression (Chen et al., 2014; Wu et al., 2003).

Although there is not yet a clear mechanism, it is clear that there is likely a conserved role for Shep in the regulation of fat levels and metabolism. Associations between Shep and fat storage have been made in flies (Reis et al., 2010), mice (Mollah and Ishikawa, 2010) and

28 humans (Chu et al., 2013; Huffman et al., 2015; Sajuthi et al., 2016; Zhu et al., 2015). Shep is therefore a strong candidate as an obesity associated gene in both flies and humans. Further studies in Drosophila will be critical to characterize how Shep effects fat storage and the development of obesity.

Split ends

Split ends (Spen) encodes a massive protein (600 kDa in Drosophila) that is highly conserved from flies to humans (Figure 1.5A). Consistent across species are two domains at each end of the protein. On the N-terminal side are three RNA Recognition Motifs (RRMs) that have been shown to recognize and bind to some RNA molecules (Newberry et al., 1999;

Shi et al., 2001). On the C-terminal side is a conserved protein-binding motif called the Spen

Paralog and Ortholog C-terminal (SPOC) domain (Kuang et al., 2000; Wiellette et al., 1999).

This is named after Spen and is conserved in all members of the protein family (Sánchez-

Pulido et al., 2004). The mammalian orthologues of Spen (MINT in mice, SHARP in humans) have two additional domains that are defined within the large middle region of the protein. These include a binding domain for nuclear receptors and RBPJ, the principal effector of the notch pathway (Oswald, 2002; Shi et al., 2001).

Spen has pleiotropic roles in the cell. It is a ubiquitous nuclear-localized protein that functions mainly as a transcriptional regulator for a number of key signaling pathways

(Ariyoshi and Schwabe, 2003; Kuroda et al., 2003; Li et al., 2005; Ludewig et al., 2004; Ma et al., 2007; Oswald, 2002; Oswald et al., 2005; Shi et al., 2001, 2002; Tsuji et al., 2007;

Vadlamudi et al., 2005; Yang et al., 2005). Spen can also influence alternative splicing and nuclear export (Hiriart et al., 2005). The protein was first identified over 20 years ago to affect the axonal outgrowth and guidance during nervous system development in Drosophila

Figure 1.5: Spen and Nito contain two domains conserved between family members and organisms. (A) Spen family members contain RNA-recognition motifs (RRMs) toward the N-terminus of the protein (black) as well as a Spen paralog and ortholog C-terminal (SPOC) domain at the C-terminal end of the protein (red). Spen othologues MINT (in mice) and SHARP (in humans) contain these same domains in addition to two additional protein- binding domains within the middle of the protein, including a nuclear receptor interaction domain (gray) and an RBP-J binding domain (blue). Numbers below the proteins indicate their length in amino acids. (B) Nito is a smaller Spen family member that also contains the RRMs (black) and SPOC domain (red). Nito has two mouse orthologues (Rbm15 and Rbm15B) and two human orthologues (OTT1 and OTT3) which are similar in size and makeup.

(Kolodziej et al., 1995) but was quickly thereafter identified as an enhancer of Raf signaling during eye development (Dickson et al., 1996), necessary for Hox signaling for head development and sclerite specification (Gellon et al., 1997; Wiellette et al., 1999), and an inhibitor of the entry into S-phase by opposing Cyclin E and E2F (Lane et al., 2000;

Staehling-Hampton et al., 1999). Since its initial discovery, Spen and its orthologues and family members have been studied extensively in many different pathways and biological functions.

Among the earliest described roles for Spen was within the receptor tyrosine kinase

(RTK) pathway. Members of this signaling pathway are critical regulators of many different cellular processes like cell survival and metabolism, cell cycle control, cell migration, and proliferation and differentiation (Lemmon and Schlessinger, 2010). Mutations in spen were found to modify mutations in Raf kinase (Dickson et al., 1996), kinase suppressor of ras (ksr)

(Therrien et al., 2000), and the transcription factor anterior open (aop/yan) (Chen and Rebay,

2000; Kuang et al., 2000; Rebay et al., 2000) resulting in a rough eye phenotype due to excess R7 ommatidial cells. Although the mechanism of Spen’s action within this pathway has yet to be elucidated, most of the evidence suggest a positive role for Spen in Ras1 signaling (Therrien et al., 2000). Spen also modulates the epidermal growth factor receptor

(EGFR) signaling pathway, a class of RTK signaling. The EGFR signaling pathway works to promote cell fate specification and differentiation in many different tissues of the fly and multiple stages of development (Schweitzer and Shilo, 1997). There is only one member of the EGF receptor tyrosine kinase family in Drosophila, called DER. The activation of the pathway is tightly controlled via the time and place of receptor activation as well as negative feedback loops to keep the process in check (Schweitzer and Shilo, 1997). Spen was

31 identified as a positive regulator of the protein tyrosine phosphatase corkscrew (csw) in eye and wing development (Firth et al., 2000) and a positive regulator of EGFR signaling through post-translational regulation of aop/yan, likely through activation of MAPK/ERK (Doroquez et al., 2007). Furthermore, mutations in spen resulted in defects in midline glial cells (MGCs) during Drosophila CNS development, a phenotype that was reminiscent of other DER/EGFR mutations (Chen and Rebay, 2000). These defects were rescued by increased Ras signaling specifically in MGCs (Chen and Rebay, 2000). This may provide evidence for the mechanism by which Spen regulates neuronal development as observed initially by

Kolodziej et al. (Kolodziej et al., 1995).

Spen also functions in other critical developmental pathways including the

Wnt/Wingless (Wg) pathway. Wg is well characterized to regulate tissue growth, polarity, and patterning (Swarup and Verheyen, 2012). Spen was identified in a screen to identify new

Wg partners that regulate transcription in different tissues. Spen mutations were able to suppress a Wg mutation in the Drosophila eye, increasing interommatidial bristle number

(Lin et al., 2003). This suggests that Spen is a positive regulator of Wg signaling. This group found that Spen was important to potentiate Wg signaling in larval eye, wing, and leg imaginal discs, but that it had no role in this pathway during embryogenesis (Lin et al.,

2003). Consistently, Spen’s human ortholog SHARP (SMRT/HDAC-1-associated repressor protein) was shown to have the same role in potentiating Wnt signaling in human RK3E cells. Overexpression of SHARP potently increased -catenin/TCF transcription of target genes while knockdown of SHARP with an RNAi inhibited the transcription of the target genes (Feng et al., 2007). SHARP expression is elevated in human and mouse colon and ovarian carcinomas that carry gene defects leading to -catenin dysregulation. Knockdown of

SHARP was able to inhibit the cancer cells’ transformed growth (Feng et al., 2007). The work from these two groups provides evidence that Spen has a conserved role in promoting transcription downstream of Wnt/Wg signaling.

The mammalian orthologues of Spen have been highly implicated in regulating Notch signaling as well. This pathway has a major role in regulating cell differentiation throughout the course of development in nearly every organism (Wang, 2011). Spen’s role in this pathway is as a transcriptional repressor. In human tissue culture, the absence of Notch intracellular domain (NICD) allows SHARP to bind to RBP-Jκ and corepressors CtIP and

CtBP, which together recruit histone deacetylase 1 (HDAC1) to prevent transcription of target genes (Oswald, 2002; Oswald et al., 2005). The mouse ortholog of Spen MINT (Msx2- interacting nuclear target protein) functions similarly by competing with NICD for RBP-J binding. Once bound, MINT repressed transcription of Notch target genes resulting in decreased B-cell (Kuroda et al., 2003) and thymocyte (Tsuji et al., 2007) differentiation.

MINT’s repression of Notch target genes necessary for hematopoiesis is likely due to chromatin modification (Jin et al., 2009). Interestingly, investigation into Spen’s role in

Notch signaling in Drosophila has yielded evidence that Spen is involved in cross-talk between Notch and EGFR signaling (Doroquez et al., 2007; Kuang et al., 2000; Price et al.,

1997), indicating that Spen may be the interface through which these two major signaling pathways interact.

The final major signaling pathway that Spen has been shown to regulate is the nuclear receptor pathway. Nuclear receptors regulate lipid and glucose homeostasis, cellular differentiation, and embryonic development, among other things (King-Jones and Thummel,

2005). The human ortholog SHARP acts as a transcriptional repressor for the estrogen

33 receptor (ER) and glucocorticoid receptor (GR) by binding and presumably sequestering the long non-coding RNA SRA, which activates the receptors (Shi et al., 2001). Additionally,

SHARP represses both liganded and non-liganded nuclear receptors by binding to corepressors SMRT, HDACs, and PPARδ (Shi et al., 2001, 2002). The mouse ortholog

MINT is able to both activate and repress transcription of nuclear receptors, where it serves as a nuclear matrix platform and binds to Msx2 and Runx2 to regulate osteocalcin activation during craniofacial development (Newberry et al., 1999; Sierra et al., 2004). Although Spen has not been shown to regulate nuclear receptors in Drosophila, one isoform of the C. elegans ortholog DIN-1 regulates this pathway. While most DIN-1 isoforms contain only

RRMs and a SPOC domain like Drosophila Spen, one isoform called DIN-1S uniquely contains a nuclear receptor interacting domain (RID) (Ludewig et al., 2004). This allows it to interact with the nuclear receptor DAF-12 (the ortholog of Drosophila DHR96) to initiate dauer formation, inhibit reproductive development, and increase fat storage under unfavorable conditions (Ludewig et al., 2004).

This work in C. elegans represents the first time that any Spen ortholog was linked to metabolism or fat storage. Since then, Spen has been identified in two separate screens in

Drosophila for obesity predisposing genes. The first was conducted in 2010 by Reis et al. when Spen hypormophic larvae had a mildly more buoyant phenotype than controls (Reis et al., 2010). Only a few years later, Baumbach et al. identified Spen in an RNAi screen in adult flies as an anti-obesity gene (Baumbach et al., 2014). Although there is not yet any known association between mammalian orthologues of Spen and metabolic disorders or increased fat storage, the high level of conservation between fly and mammalian Spen proteins suggests that there may be a conserved metabolic function. This work seeks to identify and

34 characterize how Spen functions in metabolism in flies and determine if there is likely to be a similar function in mammalian cells.

Spen has a smaller family member called Spenito (Nito). This protein is less than one-fifth the size of Spen, but also contains the RRM domains at the N-terminal end of the protein and the SPOC domain at the C-terminal end (Figure 1.5B). Like Spen, Nito is conserved in mammals with two mouse orthologues (Rbm15 and Rbm15B) and two human orthologues (OTT1 and OTT3), all of which contain the RRM and SPOC domains. Nito regulates transcription (Ma et al., 2007), alternative splicing (Hiriart et al., 2005; Lence et al.,

2016; Lindtner et al., 2006; Majerciak et al., 2014; Yan and Perrimon, 2015; Zhou et al.,

2002), and nuclear export (Hiriart et al., 2005; Uranishi et al., 2009; Zolotukhin et al., 2009) and has roles in sex determination (Lence et al., 2016; Yan and Perrimon, 2015), neuronal function (Lence et al., 2016), hematopoiesis (Raffel et al., 2007), and cancer (Hu et al.,

2016). Nito also functions alongside Spen in some instances. Spen and Nito act redundantly to promote Wg signaling (Chang et al., 2008) while they antagonize each other during eye development (Jemc and Rebay, 2006). Most recently, Spen and Nito were found to function synergistically in bursicon neurons to regulate neurite outgrowth (Gu et al., 2017). Both Spen and Nito have also been identified as regulators of X chromosome inactivation through direct interactions with the long noncoding RNA (lncRNA) Xist. Both Spen and Nito are required for the recruitment of various polycomb proteins bys Xist (Chu et al., 2015; McHugh et al.,

2015; Moindrot et al., 2015; Monfort et al., 2015; Patil et al., 2016; Roth and Diederichs,

2015). While all studies agree that Spen is necessary for Xist-mediated chromosome silencing, conflicting results have been obtained for the necessity of Nito in this regulation

(McHugh et al., 2015; Moindrot et al., 2015). Nito has not yet been implicated in a metabolic

35 role in any organism. With so many similarities between Spen and Nito in terms of structure and function, it is highly likely that Nito may function alongside Spen in regulating metabolism as implicated by previous studies (Baumbach et al., 2014; Ludewig et al., 2004;

Reis et al., 2010).

Summary

Obesity is an ever-increasing problem facing the world today. With rates of obesity and its comorbidities skyrocketing, it is imperative that more be learned regarding the causes of the disease. Obesity is caused by many different factors. Environment plays a huge role in defining behaviors involved in maintaining energy homeostasis including food intake and energy expenditure. While this is the largest factor at play in the obesity epidemic today, underlying genetics play a strong role in defining how much of an effect these behaviors have in terms of the amount of fat stored. Through decades of familial, twin, and adoption studies, there is very strong evidence that genetics contributes 40-70% to obesity predisposition. Many different genes involved in this process have been elucidated, including

FTO, PLIN, and Leptin, however there are far more that remain elusive. Drosophila provide an excellent model in which to study the question of the genetics of obesity due to the simplified yet comparable biology to mammals. Incredible tools exist for Drosophila that allow for both temporal and spatial manipulation of genes to identify the specific mechanism by which they regulate metabolism. A previous genetic screen of P-element insertion lines has identified many different potential obesity predisposing genes, including Alan shepard and Split ends (Reis et al., 2010). The goal of this work is to begin to understand which genes play a specific role in regulating fat levels in Drosophila and to begin to characterize their function in regulating metabolism. In characterizing these novel obesity predisposing genes,

36 new information regarding the genetics of obesity will be gained that may contribute to the understanding of the disease and the future development of therapeutics for those that exhibit obesity due to genetic mutations in these loci.

CHAPTER II

MATERIALS AND METHODS1

Fly Strains and Husbandry

W1118 (3605), w; dcg>Gal4 (7011), y1 sc v1; +; UAS-Fabp RNAi (34685), y1sc v1; +;

UAS-NFAT RNAi (51422), y1 sc v1; +; UAS-Shep RNAi (33996), y1 sc v1; +; UAS-Spen

RNAi (33398), y1 v1; UAS-Spen RNAi (50529), y1 sc v1; +; UAS-Nito RNAi (34848), y1 v1;

+; UAS-w RNAi (28980), y1 w; +; UAS-Shep (16366), y1 w; UAS-Spen (20756), w; UAS-

GFP (1521), w; +; UAS-GFP (1522), Spen14O1 (5808), Spen16H1 (5809), Spen3 (8735), and

Spen5 (8734) were obtained from the Bloomington stock center. Spen14O1 is an hypomorphic allele while Spen16H1 is a null (Dickson et al., 1996). Spen3 and Spen5 are null alleles, caused by small deletions in the Spen locus leading to truncations of the protein (Kuang et al., 2000). w; +; UAS Spen RNAi (48848), w; UAS Spen RNAi (45943), w; UAS Spen RNAi

(108828), and w; UAS w RNAi (30033) we obtained from the Vienna Drosophila Resource

Center (predicted off-targets in Table 2.1). The FlyTrap expression lines were created by the

Chia, Cooley, and Spradling labs and were obtained through the Bloomington stock center

(Buszczak et al., 2007; Kelso et al., 2004; Morin et al., 2001; Quinones-Coello et al., 2007). w; UAS-dcr2; elav>Gal4 was derived from w; +; elav>Gal4, which was a gift from Sarah

Siegrist of the Hariharan lab. w; UAS-SPOConly, w; UAS-ΔSPOC, w; UAS-Spen-FL, w;

UAS-Nito-ΔN, w; UAS-Nito-ΔC, and w; UAS Nito-FL were generous gifts from Ilaria

1 Portions of this chapter are published with permission from our previously published articles: Hazegh, K.E., and Reis, T. (2016). A Buoyancy-based Method of Determining Fat Levels in Drosophila. J. Vis. Med. 117, e54744. Hazegh, K.E., Nemkov, T., D’Alessandro, A., Diller, J.D., Monks, J., Mcmanaman, J.L., Jones, K.L., Hansen, K.C., and Reis, T. (2017). An autonomous metabolic role for Spen. PLoS Genet. 13, e1006859. 38

Table 2.1: RNAi directory. Predicted off-targets for the RNAi lines used and references for the figures in which they were used. RNAi Line Figure Predicted Off-Targets Fabp RNAi BL34685 3.2 CG6782 NFAT RNAi BL51422 3.3 None Shep RNAi BL33996 3.4, 3.5 None 4.1, 4.2A-C, 4.5, Spen RNAi BL33398 None 4.6, 4.7, 4.8 Spen RNAi BL50529 4.2D-F None Spen RNAi v48846 4.2G-H CG18740 Spen RNAi v49543 4.2G-H CG17834 Spen RNAi v108828 4.2G-H CG31517, CG32697 Nito RNAi BL34848 4.13, 4.14 None

Rebay and Ken Cadigan. ΔSPOC contains all but the last ~1500 amino acids of Spen (Chen and Rebay, 2000) while SPOConly contains only the last 936 amino acids of Spen as well as a nuclear localization signal (Lin et al., 2003). Nito-ΔC contains the first 593 amino acids of

Nito while Nito-ΔN contains only the last γββ amino acids of Nito (Jemc and Rebay, 2006).

Nito1 is a null mutant that was a generous gift from Norbert Perrimon (Yan and Perrimon,

2015). w; act>cd2>Gal4 UAS-GFP was obtained elsewhere (Neufeld et al., 1998). y1 sc v1;

+; UAS-Spen RNAi (33398) was used for all Spen KD experiments excluding those explicitly stated otherwise. y1 w; UAS-Spen (20756) is an EP overexpression line used for overexpression experiments including the initial density assay, RNA sequencing, and metabolomics analysis. w; UAS-Spen-FL was a generous gift from Bertrand Mollereau

(Querenet et al., 2015) and is a full-length Spen insertion used for subsequent overexpression experiments including truncation density and starvation assays and clonal analysis. Similar results were obtained with both the Spen-EP and Spen-FL lines. Animals were reared at 25°C unless otherwise specified and fed a modified Bloomington media (with malt) containing 35g yeast per liter. The diet food contained 70g yeast per liter (HYD), 35g yeast per liter (MYD), and 7g yeast per diet (LYD) respectively. Food was made fresh each week and used within the week. Eggs were collected on grape plates at 25°C and 50 first-instar larvae were transferred 22-24 hours later into a vial of food.

Immunohistochemistry

Wandering third instar larvae were dissected and fixed with 8% paraformaldehyde for

45 minutes and washed three times with 0.1% PBTriton. Carcasses were blocked with PAT for 2 hours at room temperature and then put on primary antibody at 4°C overnight. Rabbit anti-GFP (1:1,000) was from Roche. Guinea pig anti-Shep (1:2,000) was a generous gift

40 from Eugenia Olesnicky (Schachtner et al., 2015). Primary antibodies were used three times before disposal. Carcasses were then washed three times with PBTriton with 2% normal goat serum and put on secondary antibody. Goat anti-rabbit (1:5,000) and goat anti-guinea pig

(1:5,000) were conjugated to Alexa Fluor-488 from Invitrogen, Molecular Probes. FB, gut, salivary glands, brain, and imaginal discs were mounted in Slow Fade Gold mounting media and imaged with a Nikon i80 microscope to identify candidate gene expression. Brains measured for Shep expression were imaged with a Leica TCS SP5 laser-scanning confocal microscope with LASAF software. Z-stack projections were made for each brain and regions outlined to measure raw intensity. Background fluorescence was subtracted from each measurement. Unpaired t test was used to calculate statistical significance with Prism 6 software.

Density Assay

Density assays were performed as described below with 50 larvae per sample. For sex-specific density assays, two samples of 50 larvae each were collected, pooled, and sorted for sex prior to performing the assay. n=8-16. ANOVA was used to calculate statistical significance with Prism 6 software.

Introduction

Currently, there are many different quantitative methods of determining fat storage levels in Drosophila. The most widely used method is the coupled colorimetric assay (CCA)

(Hildebrandt et al., 2011; Tennessen et al., 2014). CCA was developed for determining TAG levels in human serum and operates on the premise that glycerol liberated from the backbone of triglycerides will undergo several reactions, ultimately resulting in a redox-coupled reaction generating a colored product. Absorbance of specific wavelengths of light is then

41 measured to determine the initial amount of glycerol present. However, glycerol can also be liberated from mono- and diacylglycerides in addition to TAG and therefore may not be an accurate measure of stored body fat (Tennessen et al., 2014). Furthermore, eye pigment of crushed adult flies can interfere with some absorbance readings and complicate results (Al-

Anzi and Zinn, 2010; Tennessen et al., 2014). Therefore, CCA must be accompanied and validated by thin layer chromatography (TLC), which allows for the separation of most lipid families that can be quantitated by densitometry (Al-Anzi and Zinn, 2010; Thanh et al.,

2000). However, some lipid classes like sterols cannot be analyzed and must be quantified a different way (Borrull et al., 2015). Mass spectrometry (MS) is an accurate way to quantitate all classes of major cellular lipids (Borrull et al., 2015; Shui et al., 2010). However, the lipid extraction procedures necessary to analyze by MS are both time consuming (most taking nearly a full day) and costly. Here we present an alternative method to quickly and cheaply determine organismal fat levels in the L3 larvae of Drosophila melanogaster.

The method presented below exploits the difference in density between fat tissue and lean tissue. Mammalian fat tissue has a density of approximately 0.9 g/mL (Farvid et al.,

2006) while skeletal muscle has a density of 1.06 g/mL (Urbanchek et al., 2001). This difference means that animals with higher stores of fat will have lower density than animals with lower stores of fat, which will allow them to float better in a solution of fixed density.

This property allows for extremely quick screening of a large number of animals while being both inexpensive and non-invasive. Buoyancy-based analysis has been used both to confirm the phenotypes of altering known regulators of fat levels as well as to identify new genetic and neurological regulators of obesity (Mosher et al., 2015; Reis et al., 2010).

Protocol

1. Collect eggs from flies with genotypes of interest.

Note: Ideal crosses consist of 150 virgin flies and at least 75 males. Stock collections

should consist of at least 200 flies.

1.1. Prepare grape plates.

1.1.1. Add 37.5 g agar to 1.5 L distilled water in a 4 L flask.

1.1.2. Autoclave water/agar mix for 50 minutes at 121 ºC (250 ºF).

1.1.3. While water/agar mix is autoclaving, add 50 g sucrose to 500 mL grape juice

and stir with a stir bar on a on a heated stir plate until sucrose is dissolved.

1.1.4. Turn off the heat and continue stirring to cool the grape juice/sucrose mixture

to below 70 ºC. Test the internal temperature with an alcohol thermometer.

1.1.5. Add 3 g Drosophila anti-fungal agent (e.g., Tegosept) to warm grape

juice/sucrose mixture and stir to dissolve.

1.1.6. When water/agar mixture is out of the autoclave, allow the flask to cool to the

touch by swirling occasionally.

1.1.7. Combine water/agar mixture and grape juice/sucrose mixture and stir until

well mixed.

1.1.8. Use a serological pipette to add 8-10 mL of the grape mixture to the lid of a

small petri plate (35 x 10 mm style). Allow the mixture to crown higher than the

height of the lid walls to form a convex shape.

1.1.9. Allow grape plates to cool for 2 hours at room temperature and then store in

an airtight container at 4 ºC.

1.2. Prepare yeast paste.

1.2.1. Add 6 mL phosphate buffer solution (PBS) to 4 g live active dried yeast in a

50 mL conical tube and mix with a spatula until smooth.

1.2.2. Store yeast paste at 4 ºC.

1.3. Add a small dollop (approximately 1/8 tsp) of yeast paste to the middle of a grape

plate. Smooth down any rough edges with a spatula.

1.4. Place grape plates in an incubator until they have reached ambient temperature.

1.5. Transfer flies of interest into an egg collection chamber and place grape plate with

yeast paste facing inwards on top. Tape grape plate to the egg collection chamber to

avoid letting the plate fall out. Make sure to write important genotype and date

information on the bottom of the grape plate.

1.6. Upend the egg collection chamber to allow the flies to lay eggs on the grape plate.

1.7. Place a box or cover over the egg collection chambers and place them in an incubator

at 25 ºC.

1.8. Allow flies to lay eggs for 4-6 hours.

1.9. End collection by taking the grape plate out of the egg collection chamber and

transferring the flies back into their bottle of food. Use a spatula to wipe off any

excess yeast paste from the grape plate. Store grape plate in a larger petri dish in a

humidified incubator at 25 ºC for ~24 hours.

2. Transfer larvae to experimental food.

2.1. Prepare experimental food.

Note: This food recipe is based on the Bloomington stock center recipe (2014).

Important changes include increased amount of yeast to optimize the nutritional

value of the food as assessed by timing of wandering stage (between 112-120 hours

44 after egg collection at 25ºC.) While we have found this recipe ideal for the health and development of our experimental larvae, any food recipe is acceptable given calibration with positive and negative controls, examples of which are listed in step

3.10.

2.1.1. Measure 1600 mL distilled water.

2.1.2. Mix 134 g cornmeal, 10.6 g Drosophila Agar Type II, and some of the water

(~500 mL) in a 1 L beaker. Blend for 2 minutes with a motorized hand mixer.

2.1.3. Mix 70 g live active dry yeast, 18.4 g soy flour, 84.8 g malt extract, and some

of the water (~500 mL) in another 1 L beaker. Blend for 2 minutes with a

motorized hand mixer.

2.1.4. Swirl each beaker and pour into a 4 L flask. Use remaining water to rinse out

the beakers and add into the flask.

2.1.5. Measure 140 mL light corn syrup and add to the flask, mix by swirling.

2.1.6. Autoclave for 50 minutes at 121 ºC (250 ºF) in the liquid setting.

2.1.7. After autoclaving, pour food into a 4 L beaker and allow to cool. Help the

cooling process by blending with a motorized hand mixer.

2.1.8. Once the flask has cooled enough to touch, add 8.8 mL propionic acid and

16.8 mL Drosophila anti-fungal agent/ethanol solution. Mix well. (Make

Drosophila anti-fungal agent solution by adding 380 g Drosophila anti-fungal

agent to 1 L 200 proof ethanol and stirring on a warm plate until dissolved,

making sure the solution does not exceed 70 ºC.)

2.1.9. Pour food into vials until the food is approximately 1 inch deep and allow to

cool at room temperature for 2-3 hours. Plug vials and store in an incubator at 25

ºC. Use within a week.

2.2. Plunge a spatula several times into the top layer of the vial of food to score it. This

will help the larvae to burrow and feed.

2.3. 22-24 hours after egg collection has ended, use a small paintbrush to collect 50 first

instar (L1) larvae of the same approximate size. Place larvae in the experimental

food. Clean the paintbrush on a laboratory wipe and ensure that there are no

remaining larvae on the paintbrush before continuing to another genotype.

2.4. Place vial of larvae in an incubator at 25 ºC and allow to develop.

3. Determine fat levels in larvae.

3.1. Prepare solutions.

3.1.1. Prepare 1 L 10X PBS stock: Add 80 g NaCl, 2 g KCl, 14.4 g Na2HPO4, and

2.4 g KH2PO4 to 800 mL water in a 2 L beaker containing a stir bar. Stir with no

heat until all salts are incorporated. Bring pH to 7.4. Increase volume to 1 L with

water.

3.1.2. Prepare 2 L 1X PBS. Add 200 mL 10X PBS stock solution to 1800 mL water.

3.1.3. Use this stock of 1X PBS to make 20% w/v sucrose.

3.2. Remove the vial of larvae from the incubator once there are 20-40 larvae wandering

up the sides of the vial (generally on the fifth day after egg collections). Record the

date and time of the assay as well as the number of wandering larvae.

3.3. Prepare experimental solution by adding 11.5 mL PBS and 9 mL 20% sucrose to a

50 mL conical tube.

3.4. Add 20% sucrose to the vial until it is 0.5-1 inch from the top of the vial.

3.5. Use a spatula to gently stir up the top layer of food to free larvae that are still

burrowing in the food. All larvae will rise to the surface of the sucrose solution.

3.6. Use forceps to gently transfer the larvae into the initial sucrose concentration from

step 3.3.

3.7. Use a spatula to gently stir the larvae in this solution.

3.8. Cap the conical tube and upend several times to thoroughly mix the solution.

3.9. Uncap the conical tube and swirl to create a gentle vortex.

3.10. Allow the larvae to settle, either floating to the top of the solution or sinking

to the bottom. Wait 2-5 minutes for larvae to settle.

Note: If there are still larvae that are not definitely either floating or sinking,

pick a line to use as a cutoff point to label the larvae as either a floater or a

sinker. Apply this same criterion to all genotypes. We use the angle at the

bottom of the conical tube as our boundary. adp or sir2 (Reis et al., 2010)

mutant larvae may be used as a positive control and lsd2 (Teixeira et al.,

2003) mutants may be used as a negative control for calibration of the assay.

3.11. Record the number of floating larvae and the specific concentration tested.

3.12. Add 1 mL 20% sucrose to the experimental solution to increase its density.

3.13. Repeat steps 3.7 to 3.10. Record the number of floating larvae and this new

concentration.

3.14. Continue steps 3.12 and 3.13 until at least 95% of the larvae are floating.

3.15. Use forceps to collect all larvae into a dissection dish filled with PBS.

3.16. Observe larvae under a microscope and note if there are any small larvae, pre-

pupae, pupae, or any other differences between genotypes.

Note: Ideally, all larvae should appear homogeneous and at the same

developmental stage. We have only observed consistent fat measurements

under these conditions.

3.17. Record the total number of larvae.

3.18. Using forceps, transfer 10 larvae to a laboratory wipe to dry. Label a

microcentrifuge tube and place 10 larvae in the tube.

3.19. Flash freeze the larvae in liquid nitrogen. Store the tube at -20 ºC.

Discussion

There are a multitude of techniques that have been developed to measure lipid levels

(Al-Anzi and Zinn, 2010; Hildebrandt et al., 2011; Tennessen et al., 2014). However, each of these methods comes with some major drawbacks that are addressed by the buoyancy-based assay outlined above. First, this assay is extremely quick. Testing a full concentration gradient takes no more than 30-60 minutes. This is a huge improvement on most of the techniques currently in use. For example, lipid quantitation by MS takes 7-9 hours to isolate the lipids and another several hours to analyze them. This is a strong deterrent when performing a screen on a large number of animals. By measuring a quick buoyancy phenotype, one can quickly focus on the genotypes that have the phenotype of interest.

Second, the density assay requires only common reagents such as salts (for PBS) and sucrose. This makes it much easier and cheaper to screen a large number of larvae or quickly test an interesting genotype. Finally, this protocol is non-invasive and the larvae are still alive at the end. This allows further experimentation to be performed on the animals such as

48 specific lipid quantitation or transcript or protein analysis. Additionally, recovered larvae may be allowed to continue growth to adulthood.

An additional benefit of this technique over others is increased sensitivity to small changes in fat levels in a population. By performing a whole concentration gradient on a population of 50 larvae, slight shifts in the population fat levels will be identified by a change in the intermediate concentrations, although the concentration at which the larvae begin floating and are totally floated may not be different than controls. This small shift in the population may not always be identified by other fat quantitation methods as they require a large group of larvae to analyze. The density assay on the other hand interrogates individuals within the population and can consequently identify these small changes or shifts in the overall population distribution.

As this technique relies on the density of the solution to produce reliable results, it is extremely important that the PBS and sucrose solutions are made consistently. We have found that different PBS recipes produce differences in solution density, leading to varying results. We use the recipe outlined above for consistent results. Furthermore, making the 20% sucrose from the same batch of 1X PBS is important as it results in the same background solution density. If the two solutions were made separately, slight variations in the PBS density could bring in additional density variable beyond the addition of sucrose to the experimental conical vial. Lastly, evaporation can be a problem, because over time the solutions will become more concentrated with the evaporation of water. It is therefore important to cap the bottles tightly and to remake the solution every week so that the experimental solutions remain consistent.

Among the steps outlined above, there are several critical steps that if altered, could result in inaccurate results by the buoyancy assay. First, it is important that the larvae transferred into the experimental food vial be the same age. Transferring larvae that vary in size will result in larvae at different developmental stages at the time the buoyancy assay is performed. This protocol has been developed to determine changes in fat levels in wandering

L3 larvae and will not work accurately for early L3 or pre-pupae or later. Early third instar larvae have a tendency to float independently of fat levels while pre-pupae and pupae sink even at the highest concentrations used. For this reason, the age of the L1 larvae transferred is critical. Similarly, the point at which the vial is taken to perform the buoyancy assay is important. For the same developmental timing reasons stated above, the vials should be taken only when 20-40 larvae are wandering to ensure accurate density measurements.

Additionally, the number of L1 transferred can affect results as well. The wide vials used in this protocol allow 50 larvae to eat without competition during development. A much larger number of larvae transferred than this will result in competition for food and may affect the amount of fat stored independent of genotype.

There may be genotypes with such extremely high fat levels that most or all of the larvae will float at the first concentration. In this case, the protocol may be modified by decreasing the concentration of the starting solution in order to obtain a full spectrum of the population density distribution. On the other hand, a specific genotype may be so lean that no larva floats in the initial sucrose. In this case, the initial steps may be omitted and a higher starting concentration used. This assay is very flexible and can be modified to better fit a wide range of fat levels. We advise the use of controls such as appropriate wild type animals

(controlled for genetic background), positive controls (established fat mutants, such as adp or

Sir2 (Reis et al., 2010)), and negative controls (published lean mutants such as lsd2 (Teixeira et al., 2003)). Appropriate use of this protocol will allow future screening of collections to determine new regulators of metabolism and obesity predisposition. Furthermore, this protocol provides an easy way for researchers in any field to test whether their genetic manipulation of interest causes a change in fat levels, without the need for complex protocols or expensive equipment. This protocol will help to elucidate the complex relationship between genetics and obesity.

Gas Chromatography Mass Spectrometry

Ten larvae from the group tested in the buoyancy assay (including both floaters and sinkers) were collected, frozen in liquid nitrogen, and weighed as a group. Larvae were homogenized with a motor and pestle. A solution of 2:1 chloroform:methanol with a known amount of the TAG standard tritridecanoin was added to the microcentrifuge tube with the homogenized larvae. Larvae and solution were poured into a glass vial and a total of 4 mL chloroform:methanol with tritridecanoin was added to the vial. Homogenized larvae were incubated on an orbital shaker at room temperature for one hour. Larval tissues were removed by centrifugation and NaCl was added to remove impurities. The chloroform:methanol solution containing the dissolved lipids was dried under bubbling compressed nitrogen. Lipids were re-dissolved in chloroform and run through a pre- equilibrated column to remove any remaining impurities. Lipids were eluted with chloroform and dried under bubbling compressed nitrogen. Methanol with 2.5% sulfuric acid was added to the vial and incubated for one hour at 80°C to create fatty acid methyl esthers (FAMEs).

Hexane and water were then added to the vial and placed in an ethanol dry ice bath to freeze the water. The neutral lipids were then extracted from the liquid hexane layer and analyzed

51 using a Thermo Fisher Trace 1300-ISQ GC/MS system. This protocol is as previously described (Perez and Van Gilst, 2008; Reis et al., 2010). n=8. ANOVA was used to calculate statistical significance with Prism 6 software.

Glycogen Quantification

Ten wandering third instar larvae were collected and frozen in liquid nitrogen. Larval samples were prepared using the Hexokinase (HK) Assay Kit (Sigma, St. Louis, MO) as described (Tennessen et al., 2014). Briefly, animals were homogenized and heat treated.

Sample was divided into two sets, one which was treated with amyloglucosidase to digest glycogen and one that was treated with PBS. These samples along with glycogen and glucose standards treated similarly were incubated for 1 hour at 37°C. 100 µL HK reagent was added to each standard and sample and measured for absorbance at 340 nm in 96-well plates using a

Cytation 3 plate reader (BioTek, Winooski, VT). Glycogen levels were determined by subtracting the absorbance measured for the untreated samples (basal glucose level) from the amyloglucosidase treated samples. n=17. ANOVA was used to calculate statistical significance with Prism 6 software.

Feeding Assay

Thirty early L3 larvae were collected, placed on a spot of yeast paste containing 0.5% food dye FD&C Red #40 on an agar plate at 25°C and the larvae allowed to eat for 30 minutes. Larvae were rinsed three times of any remaining external dye and frozen in batches of 20 in liquid nitrogen. Larvae were homogenized using a motor and pestle. PBS was added to the tube and a series of centrifugation steps were performed to remove the larval carcasses.

The supernatant was measured for its absorbance at 530 nm. This was performed as

52 previously described (Reis et al., 2010). n=4. ANOVA was used to calculate statistical significance with Prism 6 software.

Activity Assay

Pre-wandering L3 larvae were collected and tracked for movement as previously described (Mosher et al., 2015). Briefly, groups of 15 larvae were placed on a large agar plate on a light box. Three sets of two-minute videos were recorded with the larvae placed in the center of the plate at the beginning of each with a paint brush. The second video of the series was analyzed using the WrmTrck plugin for ImageJ (developed by Jesper S. Pedersen, http://www.phage.dk/plugins/wrmtrck.html). The following parameters were used to analyze the larval movements: minSize: 20, maxSize: 300, maxVelocity: 10, maxAreaChange: 40, minTrackLength: 80, bendThreshold: 2.0, binSize: 0.0, FPS: 30. n=4. Two-tailed unpaired t test was used to calculate statistical significance with Prism 6 software.

Mosaic Analysis

Wandering third instar larvae of the genotypes hs flp; act>cd2>gal4 UAS-GFP, hs flp; act>cd2>gal4 UAS-GFP UAS-Spen RNAi, hs flp; act>cd2>gal4 UAS-GFP UAS-Nito

RNAi, hs flp; act>cd2>gal4 UAS-GFP UAS-w RNAi, hs flp; act>cd2>gal4 UAS-GFP UAS-

Spen, hs flp; act>cd2>gal4 UAS-GFP UAS-Spen-ΔSPOC, hs flp; act>cd2>gal4 UAS-GFP

UAS-Spen-SPOConly, hs flp; act>cd2>gal4 UAS-GFP UAS-Spen-FL, hs flp; act>cd2>gal4

UAS-GFP UAS-Nito-ΔC, hs flp; act>cd2>gal4 UAS-GFP UAS-Nito-ΔN, and hs flp; act>cd2>gal4 UAS-GFP UAS-Nito-FL were dissected, fixed and stained with Nile Red

(Invitrogen), as described (Reis et al., 2010). Briefly, larvae were collected at wandering stage, inverted, and fixed in 8% paraformaldehyde for 45 minutes, covered at room temperature. Tissue was washed three times with PBTriton 0.1% for 20 minutes followed by

53 one wash with PBS for 20 minutes. Tissues were stained with 1:8,000 Nile Red in PBS for

30 minutes, covered at room temperature. Tissues were then washed three times with

PBTriton 0.1% for 20 minutes while covered. Fat bodies were dissected and mounted.

Stained tissues were imaged on a Leica TCS SP5 laser-scanning confocal microscope with

LASAF software.

Mitotic clonal analysis was performed using larvae of the genotype hs flp; FRT40A ubi>GFP / FRT40A and hs flp; FRT40A ubi>GFP / FRT40A Spen5. These animals were heat shocked directly after egg deposition for 3 hours at 37°C. Larvae were collected at wandering stage and dissected, fixed, and stained with Nile Red as above.

Clones were analyzed for LD size and intensity using an algorithm written for ImageJ by Nick Galati. Briefly, all clones were outlined and region location recorded. The FB tissue boundary was selected based on threshold. Once clone and tissue boundaries were defined,

LDs were automatically outlined based on intensity threshold of the LD and measured for size and average pixel intensity. LDs of each clone were then compared to the LDs from surrounding non-manipulated cells as well as to KD or OEX control clones. FB cell size was analyzed by manually outlining each cell within the clones and measuring for area. Two- tailed unpaired t tests were used to calculate statistical significance with Prism 6 software.

FB cell number was calculated by manually counting the number of cells within each clone.

ANOVA was used to calculate statistical significance with Prism 6 software.

Starvation Assay

Fifty larvae were placed in 20% sucrose/PBS and analyzed daily for survival. Dead larvae were removed immediately after scoring and the sucrose was changed daily. n=3. Log- rank test was used to calculate statistical significance with Prism 6 software.

RNA Library Preparation and Sequencing

Forty larval fat bodies were dissected for each genotype and total RNA was extracted using Trizol (Life Technologies) reagent following manufacturer's instructions. A total of

200-500 ng of total RNA was used to prepare the Illumina HiSeq libraries according to manufacturer’s instructions for the TruSeq Stranded mRNA Library Prep Kit. The mRNA template libraries were sequenced on the Illumina HiSeq4000 platform at the University of

Colorado’s Genomics and Sequencing Core Facility using a 1x50bp format.

Derived sequences were analyzed by applying a custom computational pipeline consisting of the open-source gSNAP (Wu and Nacu, 2010), Cufflinks, and R for sequence alignment and ascertainment of differential gene expression (Baird et al., 2014; Bradford et al., 2015; Cole et al., 2016; Henderson et al., 2015; Maycotte et al., 2015). Briefly, reads generated were mapped to the Drosophila genome by gSNAP (Wu and Nacu, 2010), expression (FPKM) derived by Cufflinks (Trapnell et al., 2010), and differential expression analyzed with

ANOVA in R. GO annotations were predicted using Panther 11.1 (Mi et al., 2017), Gene

Ontology versions 1.2, annotation 2017-04-24.

Metabolomics

Briefly, individual larvae (10 per sample, n=3 samples per genotype) were suspended in 1 ml of methanol/acetonitrile/water (5:3:2, v/v) pre-chilled to -20°C and vortexed continuously for 30 min at 4°C. Insoluble material was removed by centrifugation at

10,000xg for 10 min at 4°C and supernatants were isolated for metabolomics analysis by

UHPLC-MS. Analyses were performed as previously described (Nemkov et al., 2015, 2017;

Sun et al., 2016) using a Vanquish UHPLC system coupled to a Q Exactive mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA). Graphs, heat maps and

55 statistical analyses (either t-Test or ANOVA) were performed with GraphPad Prism 5.0

(GraphPad Software, Inc, La Jolla, CA).

Murine Analysis

All procedures involving animals were performed in accordance with published National

Institutes of Health Guidelines. The University of Colorado Anschutz Medical Campus

Institutional Animal Care and Use Committee approved this study and all procedures and housing conditions used to complete it. Mice were housed in facilities at the Anschutz

Medical Campus’s Center for Comparative Medicine with free access to food and water for the study’s duration (ββ–24°C; 14:10 h light-dark cycle). Female C57BL/6 mice were bred in house. At 8 weeks of age mice either continued on chow diet (Harlan 2920xi) or were placed on a defined high fat diet (60% kcal fat; Research Diets D12492i) for 30 weeks. Body weights were collected weekly. Upon completion of the feeding experiments, determination of the body composition of each animal was performed by quantitative magnetic resonance

EchoMRI-900 whole-body composition analyzer (Echo Medical Systems; Houston, TX). At termination, mice were euthanized with CO2, followed by heart puncture. Tissues were collected and immediately frozen in liquid N2. Total RNA was extracted from ~50 mg uterine adipose tissue for each mouse sample using Trizol (Life Technologies) reagent following manufacturer's instructions. RT was performed using Oligo d(T) 23 and M-MuLV

Reverse transcriptase (NEB) per manufacturer's instructions. qPCR was performed using

PowerUpTM SYBR® Green Master Mix (Applied Biosystems). Reactions were run in an

Applied Biosystems Step One Plus qPCR machine. Primer sequences: mSpen: 5’- ggctctggttctctacagcg-γ’ and 5’-ctccatgcagtgataaaatgcc-γ’ mNito: 5’-gcactggccaaatctgaagaag-

γ’ and 5’-tccatcagaggcccatgtaaac-γ’. mNito results were also confirmed with an independent

56 primer pair. Two technical repeats were performed on 5-7 biological replicates (standard chow n=5, HFD n=7). Percent body fat was calculated by dividing the fat mass by total body weight. P value and correlation coefficient was obtained by unpaired two-tailed t test from the average of the technical repeats using Prism 6 software.

Acknowledgements

The density assay was designed, tested, and optimized by Tânia Reis and originally published in (Reis et al., 2010).

CHAPTER III

SCREENING FOR OBESITY PREDISPOSING GENES

Abstract

Genetic background is an important predictor for the development of obesity.

Although many obesity predisposing genes have been identified, many more remain to be described. We searched for genes with roles in regulating fat levels in the Drosophila fat body from a list of potential obesity predisposing genes from Reis et al. We identify expression patterns for 33 genes and find that 19 are expressed in the larval fat body.

Fourteen of these fat body-positive genes were screened for changes in buoyancy upon depletion and four were found to cause fat body-specific alterations in fat storage, including

Split ends (Spen), fatty acid binding protein (Fabp), nuclear factor of activated T-cells

(NFAT), and Alan shepard (Shep). Fabp, NFAT, and Shep all cause a lean phenotype when depleted in the fat body despite minimal changes in metabolic behaviors, indicative of autonomous pro-fat storage roles in the fat body. In contrast, Shep has an opposing metabolic role in the brain and levels of the protein are manipulated upon changes in nutrition levels.

This provides unique insight into novel tissue-specific mechanisms of metabolic regulation in

Drosophila.

Introduction

Obesity prevalence has more than doubled since 1980 (World Health Organization,

2016). Many different factors contribute to the development of the disease, but one that underlies them all is biology. Genetic makeup determines 40-70% of the variance in body mass index (BMI) (Comuzzie and Allison, 1998; Hainer et al., 2008; Loos, 2009; Maes et al.,

1997; Silventoinen et al., 2010; Stunkard et al., 1986a, 1986b; Turula et al., 1990). Although

58 many contributing genes have been identified through genetic linkage studies, genome-wide association studies, and candidate gene studies, the vast majority of causal and contributing genes have yet to be identified. It is estimated that less than 2% of variation in BMI can be explained by the genes identified thus far (Loos, 2009). Identification of obesity predisposing genes is complicated by the multi-systemic nature of the disease. Drosophila is a widely used model to study metabolism and offers unique and powerful tools to isolate genetic functions within specific tissues (Baker and Thummel, 2007; Leopold and Perrimon, 2007; Schlegel and Stainier, 2007).

Drosophila store the majority of their energy in an organ called the fat body (FB) in the form of triglycerides (TAGs) and glycogen. The FB is the functional equivalent of mammalian white adipose tissue (WAT) and liver (Padmanabha and Baker, 2014). The storage and breakdown of energy stores is coordinated by the central nervous system (CNS) by way of controlling energy-related behaviors (food intake and energy expenditure) and by direct metabolic control (breakdown of storage molecules and endocrine signaling to the FB)

(Morton et al., 2006). During the larval stage of development, the balance between energy utilization and storage is tightly controlled. Larvae must build up fat stores within the FB to fuel metamorphosis during the pupal stage while utilizing energy to fuel larval growth and foraging behaviors (Baker and Thummel, 2007; Leopold and Perrimon, 2007; Liu and

Huang, 2013). This makes the larval stage particularly useful for identifying genes involved in maintaining energy homeostasis. Studying gene function within specific organs simplifies the multi-systemic aspects of the disease.

Metabolism is a highly conserved process throughout all phyla. While approximately

75% of all human disease genes are conserved in Drosophila, more are estimated to be

59 conserved in metabolism due to its fundamental nature (Bernards and Hariharan, 2001; Bier,

2005; Chien et al., 2002; Fortini et al., 2000; Reiter et al., 2001). Many different obesity predisposing genes have been identified in screens in Drosophila that have conserved orthologues in mammals (Baumbach et al., 2014; Beller et al., 2008; Guo et al., 2008;

Pospisilik et al., 2010). One such screen interrogated the adult Drosophila FB specifically and found 77 genes involved in promoting or protecting against excess storage of fat

(Baumbach et al., 2014). We previously identified 66 genes that were required to prevent excess fat accumulation in larvae (Reis et al., 2010). Here we follow up on many of the results of the screen to determine which have an autonomous metabolic role in the FB of the fly. Genes were initially tested for FB localization and then measured for fat storage functions within the FB. Among the genes interrogated are fatty acid binding protein (Fabp), nuclear factor of activated T-cells (NFAT), and Alan shepard (Shep), all of which displayed

FB-specific fat storage effects. Surprisingly, all of these gene manipulations resulted in lean organismal phenotypes in opposition to the fat phenotypes observed in the initial screen (Reis et al., 2010). We go on to examination the metabolic role of Shep in neurons as a likely additional method of controlling organismal metabolism and find modest effects on fat storage and foraging behaviors. Furthermore, we find evidence that levels of Shep in the brain are directly regulated by diet, suggesting that Shep may cooperate with Drosophila insulin-like peptides (dILPs) and may assist in the production or release of neuropeptides involved in the response to nutrition levels. The four genes investigated here represent potential new obesity-related genes with conserved orthologues in mammals. This will help us to understand the genetic contributions to obesity more fully.

Results

19 of 33 genes tested are expressed in the fat body

Sixty-six genes were identified to be potential fat regulators in the third instar (L3) stage of Drosophila development (Reis et al., 2010). The fat body (FB) is the main metabolic organ of the fly and plays a major role in regulating levels of stored fat in the larva (Baker and Thummel, 2007; Leopold and Perrimon, 2007; Liu and Huang, 2013). To determine which of the 66 genes regulates fat levels via the FB, we tested genes for expression in the fat body and other major organs of the larvae. This was accomplished with the use of FlyTrap expression lines in which P-elements expressing GFP are inserted into the genome to tag either a gene transcript (enhancer tag) or protein product (protein tag) (Buszczak et al., 2007;

Kelso et al., 2004; Morin et al., 2001; Quinones-Coello et al., 2007). Of the 66 genes with a potential role in regulating fat, 33 had at least one available FlyTrap line. Each line was immunostained with anti-GFP and tested for expression in the FB and the other major organs of the L3 larva including brain, gut, imaginal discs, and salivary glands. Of the 33 genes tested, 19 showed expression in the FB (Table 3.1, Figure 3.1). None of these FB-expressing genes expressed solely in the FB but rather expressed in at least one – and often several – other organs. These 19 genes are likely to have a metabolic role within the FB.

To identify which of the genes has a role within the FB, levels of the gene were knocked down in the FB using a FB-specific driver (dcg>Gal4 (Asha et al., 2003; Suh et al.,

2006)) and tested for its effect on larval fat levels. Of the 19 genes found to have expression in the FB, 14 were tested for FB-specific roles on fat regulation. Of these 14, 10 do not effect changes in fat storage in the FB (CG7530, cg, Corto, Dek, Fur1, Gdi, mub, Neur, Peb, and

Swip-1) (Figure 3.1). One gene, Split ends (Spen) was found to have a significantly fat

Table 3.1: Candidate gene expression in third instar larvae. Forty-three GFP expression lines representing 33 unique genes were tested for expression in the FB, gut, salivary glands, brain, and imaginal discs. The FlyTrap identifier is located in the Expression Line column. Insertion types include enhancer trap (enh), protein trap (prot), and enhancer opposite (enh opp). ? indicates unknown trap type. + indicates expression, - indicates lack of expression.

Gene CG Expression Type of FB Gut Salivary Brain Discs Name Number Line Insertion CG7530 CG7530 CA07233 enh? + - - + - CG9809 CG9809 CC06119 prot - - - - + ZCL2809 enh + + + + + CG15309 CG15309 ZCL2897 enh + + + + + CG32541 CG32541 YD0404 enh - - - - - CG32560 CG32560 CA06772 prot - - - + + cg CG8367 CC01469 prot + + + - - GO0440 enh + + + + + Corto CG2530 YD0831 enh + + + + + Crc CG9429 CA06507 prot + + + + + Dek CG5935 CA06616 prot + + + + + YB0141 enh - + - + + Dl CG3619 YD0952 enh - - - - + Eip75B CG8127 CB05160 enh - - - - - Esg CG3758 PO1986 enh - + - + + GO0240 prot + + + + + Fabp CG6783 ZCL1871 ? + + + + + Fas1 CG6588 CA06718 enh? - - - - - Fru CG14307 YD1008 enh + + + + + Fur1 CG10772 CB03489 enh + + - + + Gdi CG4422 CA07108 prot + + + + + grp CG17161 CB04894 enh? - - - - - hdc CG15532 CB05774 prot? - + - - + YD0842 enh - - - - - lola CG12052 YD1108 enh - + + + - CB02888 novel? - - - - - msn CG16973 CC00523 prot - - - - - YD0948 enh + + + + - mub CG7437 CC01995 prot - - - + - neur CG11988 YD1143 prot + + + + + NFAT CG11172 CA07788 prot + + + + + Peb CG12212 YB0041 enh + + + + + CA07483 enh - - - - - Psq CG2368 CC01645 prot? - - - - - Pum CG9755 CC00479 prot - - - + + Sea CG6782 YB0314 enh opp + + + + + GO0261 prot + + + - - Shep CG32423 CC00236 prot + + - + + Spen CG18497 YD0671 enh + + + + + Swip-1 CG10641 YB0230D enh + + + + - Trl CG33261 PO1651 ? + + + + + YB0239 enh - - - - - trn CG11280 CB05114 enh? - - - + +

62 phenotype when depleted in the FB. Further analysis of Spen’s function in fat regulation is discussed in Chapter IV. Three genes were found to produce a lean phenotype when depleted in the FB, and are presented below.

Fabp: a proof of principle

Among the FB positive genes was fatty acid binding protein (Fabp), the Drosophila ortholog of the mammalian Fabp family, particularly Fabp7 (Gerstner et al., 2011). Fabp proteins transport lipids intracellularly for the purpose of -oxidation (Veerkamp and

Maatman, 1995). With a known role in the fat breakdown process, we expected that Fabp would function to regulate fat levels in the FB. To test if Fabp is required in the FB to regulate fat storage, levels of the gene were depleted via FB-specific expression of an RNAi construct. Depletion of Fabp in the FB (dcg>iFabp) resulted in significantly higher density compared to a knockdown (KD) control (dcg>iw) and genetic background controls (iFabp/+ and iw/+) (Figure 3.2A). This is consistent with a lean phenotype (Hazegh and Reis, 2016;

Reis et al., 2010). This lean phenotype is in opposition to the fat phenotype observed for the

Fabp mutant in the initial P-element insertion screen (Reis et al., 2010). This suggests that

Fabp may have opposing roles in different organs of the fly, consistent with its ubiquitous expression.

Levels of stored fat can be influenced by metabolic behaviors like food intake and energy expenditure (Lakka and Bouchard, 2005; Morton et al., 2006). To identify if Fabp may affect fat storage levels through behavioral changes, we measured food intake and locomotion of the larvae in which Fabp had been depleted. KD of Fabp in the FB caused no significant changes in feeding behavior (Figure 3.2B), but caused a significant increase in the speed of the larvae (Figure 3.2C). While the increased locomotion aligns with the

Figure 3.1: Workflow for identification of genes involved in fat regulation. Sixty-six genes were identified by Reis et al. to have potential roles in regulating fat levels in Drosophila (Reis et al., 2010). Thirty-three of these genes have an available expression line and were tested for expression in the FB. Nineteen of these genes show expression in the FB, 14 of which were tested for buoyancy phenotypes after FB-specific depletion. Of these 14 lines tested for changes in buoyancy, three genes showed a lean phenotype and are discussed further herein.

64 decreased fat level, it is unclear whether it is one of the causes of the leanness or rather a compensatory behavior for a decrease in available energy stores.

With a goal to identify new obesity-related genes, we are pleased that a known metabolic regulator was identified by the initial screen and shown to have a FB-specific role in regulating fat levels. The decrease in fat storage observed upon KD of Fabp is consistent with previous studies in both mice and humans. While the closest homologue to Drosophila

Fabp, Fabp7, is expressed exclusively in the brain (Thumser et al., 2014), other very similar family members are expressed in adipocytes and hepatocytes (Storch and Thumser, 2010).

Germline knockout of the liver-expressed Fabp1 decreases liver TAG levels (Newberry et al.,

2003) and is required to maintain lipid droplets in hepatic stellate cells (HSCs) (Chen et al.,

2013). Adipocytes express Fabp4 and Fabp5 and knockout of both protects against insulin resistance and metabolic syndrome in mice (Maeda et al., 2005) and correlates to a QTL affecting fatness in pigs (Estelle et al., 2006). Similar trends have been found in humans, where levels of adipocyte-expressed Fabp4 are significantly correlated with levels of TAGs and BMI (Kaess et al., 2012). One potential point of complication is that the RNAi line used to test Fabp function has an off-target effect on scheggia (Sea), another result of the original screen (Table 3.1) which has non-overlapping open reading frames within the same genomic region as Fabp (Table 2.1). It is possible that the changes in fat level observed are due to Sea rather than Fabp, or possibly a combination of both. However, since the changes in fat regulation observed here are consistent with the Fabp literature, we believe it is due to the knockdown of Fabp. Our findings support a FB-specific role for Fabp in promoting fat storage, which is consistent with what is known for the mammalian homologues of Fabp and support our experimental methods in identifying obesity-related genes.

Figure 3.2: Fabp is necessary for fat accumulation in the fat body. (A) Percent of floating larvae in different density solutions. FB-specific Fabp KD (dcg>iFabp) compared to KD control (dcg>iw) and genetic background controls (iFabp/+ and iw/+). Fifty larvae per genotype per experimental replicate, n=8 biological replicates per genotype. Error bars represent SEM. P values represent results from ANOVA. (B) Absorbance at 530 nm as a measure of food intake, n=4. Error bars represent SD. (C) Average larval speed, pixels/sec. n=4. Error bars represent SEM. P values represent results from unpaired two-tailed t tests. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

NFAT: a likely target

Like Fabp, nuclear factor of activated T-cells (NFAT) has ubiquitous expression in the larval organs including the FB. This gene family in mammals is traditionally known for its role in immune responses (Hogan, 2017; Hogan et al., 2003), however the closest relative to Drosophila NFAT is mammalian NFAT 5, and both function to regulate osmotic balance

(Keyser et al., 2007; Lopez-Rodriguez et al., 2004; Miyakawa et al., 1999). To test the function of NFAT in regulating fat stores in the FB of the larva, levels were depleted using

RNAi. FB-specific KD of NFAT (dcg>iNFAT) resulted in significantly higher density than a

KD controls (dcg>iw) (Figure 3.3A). In testing the effect of the genetic backgrounds of these

RNAi lines on fat storage (iNFAT/+ and iw/+), we found that they were not a precise match and that the iNFAT insertion resulted in mildly decreased density compared to the iw insertion (Figure 3.3B). However, since expression of the NFAT RNAi by the dcg driver results in an opposite phenotype as the iNFAT insertion, the fold change when normalized to the genetic background results in a significantly decreased buoyancy indicative of a lean phenotype (Figure 3.3C). As NFAT was found to have increased buoyancy in the initial whole animal mutant screen, the FB-specific lean phenotype suggests roles elsewhere in the larva.

The lean phenotype observed with the KD of NFAT may be caused by changes in metabolism or due to alterations in metabolic-related behaviors like eating and locomotion.

We found that larvae in which NFAT was depleted in the FB tended to eat less than controls, although this was only significant compared to the iNFAT insertion control (Figure 3.3D).

Similarly, NFAT KD larvae tended to move slower than controls, although this was not quite a significant decrease compared to the iNFAT insertion controls (Figure 3.3E). While these

Figure 3.3: NFAT is necessary for fat accumulation in the fat body.

Figure 3.3: NFAT is necessary for fat accumulation in the fat body. (A) Percent of floating larvae in different density solutions. FB-specific NFAT KD (dcg>iNFAT) compared to KD control (dcg>iw). Fifty larvae per genotype per experimental replicate, n=8 biological replicates per genotype. (B) Genetic background controls (iNFAT/+ and iw/+) for (A). (C) As the NFAT insertion site results in a higher-buoyancy phenotype, KD animals were normalized to their genetic background. Error bars represent SEM. P values represent results from ANOVA. (D) Absorbance at 530 nm as a measure of food intake, n=4. Error bars represent SD. P values represent results from ANOVA. (E) Average larval speed, pixels/sec. n=4. Error bars represent SEM. P values represent results from unpaired two-tailed t tests. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

69 trends in altered behavior may affect or be affected by the decrease in fat, it is unlikely that they are the sole cause for the decrease in fat levels observed in these larvae since the behavior phenotypes are so mild. We therefore believe that NFAT has an autonomous pro-fat storage role in the FB as well as an additional opposing role elsewhere in the larva. With strong expression in the gut and the brain (Table 3.1), both with important roles in regulating metabolism, we believe NFAT may have additional unidentified metabolic roles in these tissues.

Shep promotes fat storage in the FB

Two independent protein-tagged expression lines for Shep were tested to determine

Shep expression in the larva. Although the two lines conflicted slightly in expression pattern in some organs, both lines showed expression of Shep in the FB (Table 3.1). Although the majority of studies on Shep in Drosophila have been focused in the CNS (Armstrong et al.,

2006; Chen et al., 2014; Matzat et al., 2012; Schachtner et al., 2015; Tunstall et al., 2012), human studies have found some connections between the Shep ortholog RBMS and changes in BMI (Huffman et al., 2015; Mollah and Ishikawa, 2010; Sajuthi et al., 2016), type 2 diabetes (Sajuthi et al., 2016; Zhu et al., 2015), and dietary intake (Chu et al., 2013). To test whether Shep has a role in regulating fat levels in the FB of Drosophila, levels of the gene were knocked down by RNAi. Depletion of Shep in the FB (dcg>iShep) resulted in a significantly lean phenotype compared to all controls (dcg>iw, iShep/+, and iw/+) (Figure

3.4A). Interestingly, Shep depletion in the FB caused no changes in the food intake or locomotion of the larvae, suggesting that Shep has an autonomous role in promoting fat storage in the FB (Figure 3.4B-C).

Figure 3.4: Shep is necessary for fat accumulation in the fat body. (A) Percent of floating larvae in different density solutions. FB-specific Shep KD (dcg>iShep) compared to KD control (dcg>iw) and genetic background controls (iShep/+ and iw/+). Fifty larvae per genotype per experimental replicate, n=8 biological replicates per genotype. Error bars represent SEM. P values represent results from ANOVA. (B) Absorbance at 530 nm as a measure of food intake, n=4. Error bars represent SD. (C) Average larval speed, pixels/sec. n=4. Error bars represent SEM. P values represent results from unpaired two-tailed t tests. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

Shep functions in the brain to inhibit fat storage

The contrast in fat storage phenotypes between the whole animal mutant and the FB- specific depletion of Shep prompted us to question the role of Shep in the brain. To test

Shep’s role in the brain for regulating fat storage, we depleted levels of Shep within neurons using the pan-neuronal driver elav (Campos et al., 1987). Dicer-2 (dcr2), a protein necessary for siRNA silencing (Lee et al., 2004), was co-expressed with the Shep RNAi to ensure a strong knockdown (Dietzl et al., 2007; Valakh et al., 2012). Pan-neuronal KD of Shep

(elav>iShep) resulted in a modest increase in larval buoyancy compared to a KD control

(elav>iw) (Figure 3.5A). This increase was not significant when analyzed by ANOVA but was slightly significant when analyzed by unpaired t-test. In contrast, the genetic backgrounds of both of these RNAi insertions (iShep/+ and iw/+) were very similar (Figure

3.5B), indicating that the slight increase in fat levels was due to the decrease in Shep expression. Given the whole animal mutant fat phenotype and the FB-specific KD lean phenotype, we had expected a much stronger increase in fat levels when Shep was knocked down in the brain. However, the slight increase in apparent fat levels in the elav>iShep larvae is more consistent with the increased buoyancy observed in the whole animal mutant screen performed previously (Reis et al., 2010). The inequality in the sum of accumulated fat from the FB-specific and brain-specific KDs to result in a whole animal fat phenotype may be due to an additional role for Shep in yet another organ. Both expression lines tested indicated strong expression of Shep in the gut, which has roles in metabolism. It is likely that pro-fat usage roles for Shep in the brain and gut may out-weigh the pro-fat storage role in the FB, resulting in the whole animal mutant fat phenotype observed.

Figure 3.5: Shep expression in the brain is necessary but not sufficient to decrease organismal fat levels.

Figure 3.5: Shep expression in the brain is necessary but not sufficient to decrease organismal fat levels. (A) Percent of floating larvae in different density solutions. Brain- specific Shep KD (elav>iShep) compared to KD control (elav>iw). Fifty larvae per genotype per experimental replicate, n=8 biological replicates per genotype. (B) Genetic background controls (iShep/+ and iw/+) for (A). Error bars represent SEM. P values represent results from unpaired two-tailed t tests. (C) Absorbance at 530 nm as a measure of food intake, n=4. Error bars represent SD. (D) Average larval speed, pixels/sec. n=4. Error bars represent SEM. P values represent results from unpaired two-tailed t tests. (E) As in (A), brain- specific overexpression of Spen (elav>Shep) compared to overexpression control (elav>GFP) and genetic background controls (Shep/+ and GFP/+). n=8. Error bars represent SEM. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

Genes regulating metabolism in the brain may do so by effecting behavioral aspects of energy homeostasis or by directly regulating the metabolic processes (Morton et al., 2006).

To test whether Shep KD in the brain results in changes in behavior we tested food intake and locomotion. Although larvae in which Shep was knocked down in the brain ate slightly more than KD controls, their feeding was indistinguishable from their genetic background and was not significantly different as determined by ANOVA analysis, indicating that Shep does not alter food intake behavior (Figure 3.5C). Conversely, larvae in which Shep was knocked down in the brain moved significantly slower than all controls, consistent with their modest increase in buoyancy (Figure 3.5D). It is unclear whether Shep’s role in regulating fat levels from the brain is due only to regulation of locomotion or if there is an additional direct role in affecting metabolism. To determine if Shep expression in the brain is sufficient to regulate organismal fat, levels of the gene were overexpressed using an EP overexpression line. Shep overexpression (elav>Shep) was indistinguishable from its genetic background

(Shep/+) suggesting that Shep may not be sufficient to perform this role in the brain (Figure

3.5E). However, the efficacy of this UAS-Shep EP line has not been tested and may not function as efficiently as a full-length overexpression. Several full-length overexpression lines exist for Shep and should be tested for a role in the brain before final conclusions can be drawn.

Shep expression in the brain is diet-dependent

Previous work has found that Shep expression in the brain overlaps with neurosecretory cells (NSCs) (Chen et al., 2014), some of which play a role in endocrine signaling during and after feeding events. Among these NSCs are the insulin producing cells

(IPCs) which secrete insulin-like peptides (dILPs) 2, 3, and 5 in response to food intake

Figure 3.6: Shep expression in the brain varies based on nutrition. (A) A schematic of the regions of the larval brain. L/R Lobe, left/right lobe; SOG, subesophageal ganglion; VNC, ventral nerve cord. (B) Anti-Shep immunostaining of w1118 larvae fed on high-, medium-, or low-yeast diets (HYD, MYD, LYD respectively). Region-specific expression was measured as well as the brain as a whole. LYD n=10, MYD n=9, HYD n=10. Error bars represent SD. P values represent results from unpaired two-tailed t tests. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

(Brogiolo et al., 2001; Rulifson et al., 2002). Release of the dILPs allows cells to uptake circulating sugars to metabolize them for energy or to store them as fat (Leto and Saltiel,

2012). The overlap of Shep expression with IPCs may suggest an interaction with dILP signaling. We therefore asked whether levels of Shep may be affected by nutrition intake similar to dILPs. To test this, wild type larvae (w1118) were fed on low-yeast (LYD), medium- yeast (MYD), or high-yeast (HYD) diets for the duration of their development from hatching from the egg to wandering stage. Larvae were then collected and stained with an anti-Shep antibody. Brains were mounted and analyzed for levels of Shep expression. Three different regions of the brain were interrogated separately, including the tip of the ventral nerve cord

(VNC), the sub-esophageal ganglion (SOG), and the lobes (Figure 3.6A). The expression of

Shep was also interrogated in the brain as a whole. Despite a very small n, there is a clear and significant trend of higher Shep expression in LYD and lower Shep expression in HYD

(Figure 3.6B). The VNC tip and lobes of the brain showed a very similar trend to the whole brain, although the change in expression in the lobes was not quite significant (Figure 3.6B).

The SOG showed a slightly different trend, with the highest level of Shep expression occurring in animals raised on MYD with decreased levels in LYD and HYD (Figure 3.6B).

Altogether, this indicates that Shep levels in the brain are modulated by the nutritional intake of the animal.

Discussion

The candidate genes tested herein represent many different potential regulators of body fat and obesity. Of the 66 genes that caused a fat larval phenotype when disrupted (Reis et al., 2010), 33 were tested for expression in larval organs (Table 3.1). Although this work only followed up on the 19 that showed expression in the FB, genes with different expression

77 patterns may still have roles in regulating organismal fat through roles in other organs. Both the brain and gut have roles in metabolism and may be the sites of action of some of these genes. For example, delta (dl), escargot (esg), and longitudinals lacking (lola) all have expression in both the brain and gut, while CG32560, headcase (hdc), pumilio (Pum), and tartan (trn) are expressed in one or the other. These genes may be good candidates to study regulation of fat levels in these tissues. FB expression was primarily sought in this research to identify genes with FB-autonomous metabolic roles. Although 19 genes with FB expression were identified, only half of the initial list of candidate genes had an available

FlyTrap expression line to test. Therefore 33 genes remain that may express in the FB and have FB-specific roles in regulating fat levels. Expression of these genes in different tissues may be determined through qPCR and result in a larger list of FB obesity regulator candidates.

While the majority of the 19 FB-expressing genes identified in the expression screen

(Table 3.1) were tested for fat regulation roles in the FB, there were five genes that remain untested including calreticulin (Crc), fruitless (fru), trithorax-like (Trl), sheggia (Sea), and

CG15309. Each of these genes has at least one available RNAi line, either through the

Transgenic RNAi Library (TRiP) or through the Vienna Drosophila Resource Center

(VDRC). These candidates may have high potential for regulation of body fat. Crc binds calcium and is involved in nervous system development and sleep (Harbison and Sehgal,

2008; Prokopenko et al., 2000). Crc has also recently been identified as a LD-associated protein in Drosophila, giving it high potential to have fat regulatory roles in the FB (Kolkhof et al., 2017). Fru is a transcription factor that contributes to sexual differentiation of neural circuits in flies (Yamamoto et al., 2014) and contributes to sex determination (Ryner et al.,

1996; Usui-Aoki et al., 2000). When the fruitless neurons are silenced, increased levels of organismal fat result (Al-Anzi et al., 2009). While this may be sufficient to explain the increased buoyancy phenotype observed initially, expression in the FB begs the question of whether Fru is playing a similar role in that tissue. Trl is a transcription factor involved in chromatin modification and dosage compensation (Greenberg et al., 2004). While genes involved in dosage compensation have been linked to obesity predisposition before (Hazegh et al., 2017), Trl may simply regulate the transcription of genes involved in metabolism in the

FB. Sea is a tricarboxylate carrier of the inner mitochondrial membrane and plays a central role in fatty acid synthesis (Carrisi et al., 2008). Although it has never been tested for its effect on organismal fat homeostasis, its role in fatty acid synthesis and overlapping genomic region with Fabp strongly suggests that it is a regulator of fat levels. Very little is known about CG15309, making its probability of regulating fat in the FB unknown.

Although there are still many avenues yet to be explored with the initial list of candidate genes, many were tested for FB-specific function in fat regulation. Fourteen genes were independently knocked down by RNAi expression in the FB and tested for changes in buoyancy. Most of the genes tested had no change in buoyancy phenotype, indicating that any fat regulatory role likely does not take place in the FB. However, the extent of each knockdown has not been tested and will need to be confirmed before the FB-roles of these genes are disregarded. One gene was found to have a fat phenotype when knocked down in the FB (discussed at length in Chapter IV (Hazegh et al., 2017)) and three genes were found to have lean phenotypes. Conspicuous among the genes causing a lean phenotype was Fabp.

Work in both mice and humans has shown that levels of hepatocyte- and adipocyte-expressed

Fabp are directly correlated to levels of organismal fat (Kaess et al., 2012; Maeda et al.,

2005; Newberry et al., 2003). Our genetic manipulations to knock down levels of Fabp in the fat body of Drosophila resulted in a similar lean outcome as expected (Figure 3.2A). This was a good proof of principle for our screening technique and gave us confidence in the significance of the other results of our screen.

Animals were similarly found to be lean upon the knockdown of NFAT, a transcription factor most well-known for its role in the immune response (Figure 3.3A-C)

(Hogan, 2017; Hogan et al., 2003). NFAT has been implicated in metabolic regulation only once before; mice mutant for two orthologues of NFAT (Nfatc2 and Nfatc4) show defects in fat accumulation and increased insulin sensitivity (Yang et al., 2006). Our results are consistent with this outcome and take it one step further by showing that the pro-fat storage role of NFAT is specific to fat cells in Drosophila. Furthermore, the initial low-density phenotype of the whole animal NFAT mutant fly indicates an opposing fat storage role elsewhere in the fly. NFAT also has roles in the nervous system where it controls neural excitability and regulates overall larval locomotion, among other things (Freeman et al.,

2011). While our results in the FB show slight trends in decreased larval locomotion, it is possible that manipulation of NFAT levels in the brain may strongly effect larval movement resulting in an overall fat phenotype as seen in the initial screen (Reis et al., 2010). NFAT has also been implicated in regulating leptin levels (Soudani et al., 2016; Yang et al., 2006), which may similarly impact larval food intake and thus impact fat storage levels. Our results are consistent with a pro-fat storage role for NFAT in the FB that are likely unrelated to behavioral changes (Figure 3.3D-E). However, there is evidence that NFAT has an opposing role elsewhere in the fly and likely in the brain. Further follow up work would help to elucidate the metabolic functions of NFAT.

The third gene found to have a lean phenotype in the FB was Shep, an RNA binding protein involved in neuronal development and remodeling (Armstrong et al., 2006; Matzat et al., 2012; Schachtner et al., 2015). This work represents the first time that Shep has been identified as having a role in the FB of Drosophila, where it causes a lean phenotype independent of food intake or energy expenditure (Figure 3.4). Previous associations between Shep and fat storage in mice and humans all indicate the opposite role for Shep, where Shep levels are inversely related with levels of stored fat (Huffman et al., 2015;

Mollah and Ishikawa, 2010; Sajuthi et al., 2016). The fat phenotype observed in the initial screen (Reis et al., 2010) as well as the opposing phenotypes in the literature prompted us to test the metabolic role of Shep in the brain. Pan-neuronal depletion of Shep resulted in a slight fat phenotype that was significant by unpaired t test but not by ANOVA (Figure

3.5A). Although minor, this change in fat accumulation matches the initial screen result and mammalian literature. The increase in buoyancy was accompanied with a decrease in larval movement (Figure 3.5D), which is consistent with the motor defects observed previously

(Armstrong et al., 2006) and may contribute to the increase in fat storage. However, Shep is not sufficient to deplete stored fat (Figure 3.5E) indicating that additional factors may be necessary.

Shep expression in the brain overlaps with many peptidergic neurons associated with metabolism including IPCs (Chen et al., 2014). Our results show that Shep expression in the brain is regulated by diet (Figure 3.6B). One explanation could be that Shep may sense and be regulated by the nutritional intake of the larva and regulate the transcription or secretion of dILPs and other neuropeptides, thereby altering organismal fat storage. This would predict that neuropeptides are altered upon Shep manipulation. Alternatively, altered dILP secretion

81 upon changes in nutritional intake may affect Shep expression in the brain, leading to an unidentified pathway of metabolic regulation. This model predicts that altering dILP secretion genetically would be sufficient to decrease neuronal Shep levels. Furthermore,

Shep may have additional roles in regulating the larval behavior that may contribute to the fat storage level of the larvae. This may be related to the overlap in expression between Shep and peptidergic neurons. Further experimentation will help to elucidate these possibilities.

Future studies will help to elucidate the mechanisms of the obesity predisposing genes presented here. This will both increase our understanding of the underlying genetics associated with obesity as well as the development of the disease. Further work may reveal potential therapeutic options to treat underlying genetic disorders of obesity.

Acknowledgements

We would like to thank the Bloomington Stock Center for fly stocks, Eugenia

Olesnicky for the Shep antibody, and Alejandro Gonzales for his help on the Shep IHC in the brain.

CHAPTER IV

AN AUTONOMOUS METABOLIC ROLE FOR SPEN2

Abstract

Preventing obesity requires a precise balance between deposition into and mobilization from fat stores, but regulatory mechanisms are incompletely understood.

Drosophila Split ends (Spen) is the founding member of a conserved family of RNA-binding proteins involved in transcriptional regulation and frequently mutated in human cancers. We find that manipulating Spen expression alters larval fat levels in a cell-autonomous manner.

Spen-depleted larvae had defects in energy liberation from stores, including starvation sensitivity and major changes in the levels of metabolic enzymes and metabolites, particularly those involved in -oxidation. Spenito, a small Spen family member, counteracted Spen function in fat regulation. Finally, mouse Spen and Spenito transcript levels scaled directly with body fat in vivo, suggesting a conserved role in fat liberation and catabolism. This study demonstrates that Spen is a key regulator of energy balance and provides a molecular context to understand the metabolic defects that arise from Spen dysfunction.

Introduction

Organisms strive to achieve homeostasis between energy intake and utilization, but also must maintain energy stores to survive when utilization exceeds intake. Demands for utilizable energy trigger hydrolysis of triglycerides stored in adipose cells to produce free

2 This chapter is published with permission from our previously published article Hazegh, K.E., Nemkov, T., D’Alessandro, A., Diller, J.D., Monks, J., Mcmanaman, J.L., Jones, K.L., Hansen, K.C., and Reis, T. (2017). An autonomous metabolic role for Spen. PLoS Genet. 13, e1006859. 83 fatty acids that are released into the circulatory system. Once within the energy-requiring cells, fatty acids must be conjugated first to coenzyme A and then to carnitine for transport across the inner mitochondrial membrane. During fasting, fat used for fuel is primarily derived from adipose tissue triglycerides, and the mobilization of fatty acids from triglyceride stores is a key regulatory step.

Obesity is caused by excess energy stored in the form of triglycerides (TAGs) (Guh et al., 2009). Genetic factors dictate 40-70% of the variance in body mass index (BMI) and obesity predisposition (Comuzzie and Allison, 1998; Hainer et al., 2008; Loos, 2009; Maes et al., 1997; Silventoinen et al., 2010; Stunkard et al., 1986a, 1986b; Turula et al., 1990) but understanding individual gene function in obesity is complicated by the multigenic and multi-systemic nature of the disease. Drosophila provides a powerful model to investigate mechanisms of energy storage and utilization (Baker and Thummel, 2007; Beller et al., 2010;

Grönke et al., 2003, 2005; Guo et al., 2008; Leopold and Perrimon, 2007; Reis et al., 2010;

Schlegel and Stainier, 2007). The fat body (FB) corresponds to mammalian liver and white adipose tissue (WAT) and stores glycogen and TAGs (Padmanabha and Baker, 2014).

Assessment of energy regulation during the larval stage is particularly informative, since energy is balanced between utilization (to fuel foraging behaviors and larval growth) and storage (to fuel later growth during the pupal stage) (Baker and Thummel, 2007; Leopold and

Perrimon, 2007; Liu and Huang, 2013). We previously identified 66 genes required to prevent excess fat accumulation in larvae, including many homologs of mammalian genes with established roles in energy balance (Reis et al., 2010). In addition, we identified a class of genes for which mammalian homologs have not yet been implicated in fat regulation.

These genes represent potential new directions in obesity research.

The Drosophila split ends (spen) gene is essential for viability and encodes an extremely large (>5,500 amino acids) RNA-binding protein known to regulate the transcription of key effectors of a number of signaling pathways. Spen promotes Wingless

(Wg) signaling in flies and the orthologous Wnt signaling pathway in mammals (Feng et al.,

2007; Lin et al., 2003), and suppresses Notch signaling in flies and mammals (Doroquez et al., 2007; Kuroda et al., 2003; Oswald, 2002; Oswald et al., 2005; Tsuji et al., 2007). Spen contains three RNA recognition motifs (RRMs) near its N terminus and, near the C terminus, the archetype Spen paralog and ortholog C-terminal (SPOC) domain (Wiellette et al., 1999).

Spenito (Nito), a much smaller (793 amino acids) Spen family member with RRMs and a

SPOC domain, acts redundantly with Spen to promote Wg signaling (Chang et al., 2008), whereas during eye development it acts antagonistically to Spen (Jemc and Rebay, 2006).

Nito has additional roles in sex determination (Lence et al., 2016; Yan and Perrimon, 2015) and neuronal function (Lence et al., 2016). Importantly, the mammalian homologs of both

Spen (SPEN/MINT/SHARP, hereafter mSpen) and Nito (Rbm15/OTT1, hereafter mNito) were recently found to be regulators of X chromosome inactivation via RRM-mediated interactions with the long, noncoding RNA (lncRNA) Xist (McHugh et al., 2015; Moindrot et al., 2015; Monfort et al., 2015; Roth and Diederichs, 2015). In addition to activation or repression of transcription, Spen family proteins influence alternative splicing (Hiriart et al.,

2005; Lence et al., 2016; Lindtner et al., 2006; Majerciak et al., 2014; Yan and Perrimon,

2015; Zhou et al., 2002) and nuclear export of RNAs (Hiriart et al., 2005; Uranishi et al.,

2009; Zolotukhin et al., 2009), and are commonly mutated in cancers (Feng et al., 2007;

Legare et al., 2015), but mechanistic details are lacking. Identification of spen hypomorphs in our unbiased screen for fat mutant larvae (Reis et al., 2010) represented the first evidence

85 that Spen family proteins have a role in organismal adiposity. spen was independently identified in a subsequent genome-wide RNAi-based screen for increased adiposity in adult flies (Baumbach et al., 2014). Mutation of the Spen homolog in C. elegans, Din-1, strongly increased stored fat, indicative of a conserved role in the regulation of fat storage (Ludewig et al., 2004). However, these studies did not determine in which tissue Spen or its homologs act to control fat storage, or what defects in metabolism resulted in (or were reflected by) the accumulation of stored fat.

Here we analyze Spen and Nito function in the regulation of body fat in Drosophila larvae using a combination of genetic, cell biological, and biochemical approaches. We further monitor adipose tissue expression of mSpen and mNito in response to diet-induced obesity. Our results suggest a conserved RRM-mediated role for Spen homologs in the control of energy metabolism in fat storage tissues.

Results

Spen is necessary and sufficient to reduce fat accumulation in the fat body

Third instar (L3) larvae homozygous for a hypomorphic P-element insertion allele in the spen locus float in a sucrose solution in which control larvae sink, indicative of lower overall density and consistent with elevated body fat (Hazegh and Reis, 2016; Reis et al.,

2010). Most larval fat is stored in the FB (Baker and Thummel, 2007; Leopold and Perrimon,

2007; Liu and Huang, 2013). To test if Spen is required specifically in the FB to prevent excess fat accumulation, we measured larval density upon FB-restricted (via a dcg>GAL4 driver (Asha et al., 2003; Suh et al., 2006)) expression of one of five distinct spen-targeting

RNAi constructs. In all five cases, FB knockdown of Spen (dcg>iSpen, hereafter referred to as Spen KD) resulted in lower density compared to both a knockdown control (dcg>iw) and

Figure 4.1: Spen autonomously decreases fat levels in the fat body.

Figure 4.1: Spen autonomously decreases fat levels in the fat body. (A) Percent of floating larvae in different density solutions. FB-specific Spen KD (dcg>iSpen) compared to KD control (dcg>iw) and genetic background controls (iSpen/+ and iw/+). Fifty larvae per genotype per experimental replicate, n=8 biological replicates per genotype. Error bars represent SEM. (B) Percent body fat (total neutral lipids divided by body weight) as measured by GCMS, n=8. Error bars represent SEM. (C) Absorbance at 340 nm as a measure of glycogen content, n=17. Error bars represent SEM. (D) As in Figure 4.1A, FB-specific overexpression of Spen compared to GFP overexpression and genetic backgrounds as controls. n=8. Error bars represent SEM. (E) Absorbance at 530 nm as a measure of food intake, n=4. Error bars represent SD. (F) Average larval speed, pixels/sec. n=4 biological replicates per genotype. Error bars represent SEM. P values represent results from unpaired two-tailed t tests. (G-J) Larval FB tissue ectopically expressing constructs along with GFP (green). Tissues stained with the lipophilic dye Nile Red to mark neutral lipids (red). Dotted white line outlines construct-expressing clones. (G) w-RNAi, (H) Spen-RNAi, (I) UAS-GFP, (J) UAS-Spen.

Figure 4.2: Knockdown of Spen results in a low-density phenotype. (A) Percent of floating larvae in different density solutions. FB-specific Spen KD (dcg>iSpen, BL33398) as in Figure 4.1A with additional dcg/+ background control. Fifty larvae per genotype per experimental replicate, n=8 biological replicates per genotype. (B) Percent of female only Spen KD larvae floating. (C) Percent of male only Spen KD larvae floating. (D) FB-specific Spen KD (dcg>iSpen, BL50529) with different insertion site as Spen KD in Figure 4.1A compared to KD control (dcg>iw). (E) Genetic background controls (iSpen/+ and iw/+) for (D). (F) As the Spen hairpin insertion site appears to result in a lean phenotype, KD animals were normalized to their genetic background. (G) As in (A), three additional independent Spen hairpin constructs (dcg>iSpen) tested in different density solutions and compared to KD control (dcg>iw). (H) Genetic background controls (iSpen/+’s and iw/+) for (G). P value obtained by ANOVA. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001. Error bars represent SEM.

89 genetic background controls (iSpen/+, iw/+, and dcg/+) (Figure 4.1A and 4.2), recapitulating the whole animal mutant phenotype. Buoyancy/density correlates strongly with adiposity as assessed directly via gas chromatography coupled with mass spectrometry (GC/MS) to measure levels of neutral lipids (Reis et al., 2010). We calculated percentage body fat in this way for the same animals tested by the buoyancy assay. KD of Spen in the FB increased body fat by ~18% (Figure 4.1B, mean ± SEM 8.0% ± 0.2% for Spen RNAi compared to

6.8% ± 0.1% for w RNAi control; P < 0.01 by ANOVA). Notably, although females of all genotypes stored more fat than males, for both sexes the increase in buoyancy resulting from

Spen depletion was similar (mean fold change for all sucrose concentrations ± SEM, 7.9 ±

1.5 for females and 6.0 ± 1.3 for males, P = 0.34 by unpaired t test) (Figure 4.2B-C).

Additionally, a trans-heterozygous combination of hypomorphic spen alleles (Dickson et al.,

1996) resulted in a similar density phenotype (Figure 4.3A). A smaller but significant decrease in density was also observed in larvae heterozygous for a null and a wildtype (WT) allele (Kuang et al., 2000) (Figure 4.3B). Levels of glycogen, the other major form of energy stored in the FB, were also decreased in larvae when Spen was depleted (Figure 4.1C). FB- restricted Spen overexpression (dcg>Spen) was sufficient to drive fat depletion (Figure

4.1D). Our findings thus support a FB role for Spen in control of fat storage.

Both food intake and energy expenditure can influence levels of stored fat (Lakka and

Bouchard, 2005; Morton et al., 2006). To ask if changes in feeding and foraging behaviors contributed to the increase of fat levels in the Spen KD larvae, we assessed food consumption and locomotion. Spen KD in the FB increased food intake in early L3 larvae compared to controls (Figure 4.1E). Furthermore, pre-wandering L3 larvae showed decreased locomotor activity (Figure 4.1F). Both behavioral changes align with the increased stored fat in these

Figure 4.3: Decreased larval density in Spen mutants.

Figure 4.3: Decreased larval density in Spen mutants. (A) Percent of floating larvae in different density solutions. Heterozygous Spen and Nito mutants compared to w1118 control. w1118 control and mutants were backcrossed to w; sco/cyo GFP for two generations. Fifty larvae per genotype per experimental replicate, n=8-16 biological replicates per genotype. Error bars represent SEM. (B) As in (A), additional heterozygous Spen null mutants and Nito mutant compared to w1118 control. w1118 control and mutants were backcrossed to w; sco/cyo GFP for six generations. (C-D) Larval FB tissue from heterozygous Spen5 mutant animals expressing WT (bright green) or fully mutant (no green) clonally. Tissues stained with the lipophilic dye Nile Red to mark neutral lipids (red). (C) WT control, (D) Spen5. Error bars represent SEM. (E-F) Lipid droplet (LD) size in WT control (E) or Spen5 (F) clones. GFP/GFP represents fully WT cells, GFP/WT or Spen5 represents heterozygous cells, and WT/WT or Spen5/Spen5 represents recombined mutant (or control) cells. (E) GFP/GFP n=143, WT/WT n=191, (F) GFP/GFP n=175, Spen5/Spen5 n=184. Error bars represent SEM. P value obtained by unpaired two-tailed t tests. (G-H) LD intensity in WT control (G) or Spen5 (H) clones. (E) GFP/GFP n=143, WT/WT n=191, (F) GFP/GFP n=175, Spen5/Spen5 n=184. Error bars represent SEM. P value obtained by unpaired two-tailed t tests. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

Figure 4.4: Spen overexpression does not alter behavior or starvation response. (A) Absorbance at 530 nm as a measure of food intake for Spen overexpression (dcg>Spen) compared to overexpression control (dcg>GFP) and genetic background controls (Spen/+ and GFP/+), n=4. P value obtained by ANOVA. Error bars represent SD. (B) Average larval speed, pixels/sec. n=4. P value obtained by unpaired two-tailed t test. Error bars represent SEM. (C) Larvae reared in amino acid-free media and tracked for survival. Fifty larvae per genotype per experimental replicate, n=3. P value obtained by Log-rank test. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

93 animals. By contrast, no change in food intake or locomotion accompanied the lean phenotype resulting from FB-restricted Spen overexpression (Figure 4.4A-B), indicating that behavioral changes did not contribute to the decrease in energy stored as fat. We conclude that overexpressed Spen acts autonomously in the FB to produce these effects.

Spen acts autonomously in fat body cells

Changes in levels of stored fat can result from changes in FB cell size or number

(Faust et al., 1978; Pospisilik et al., 2010) or lipid droplet (LD) morphology or density

(Walther and Farese, 2012). To better understand the effects of Spen manipulation, we generated by flp-mediated recombination (Neufeld et al., 1998; Pignoni and Zipursky, 1997) clones of FB cells in which Spen was either knocked down or overexpressed, surrounded by

WT FB cells. GFP was co-expressed in both conditions to mark construct-expressing cells, and LDs were labeled with the lipophilic Nile Red (Greenspan and Fowler, 1985) (Figure

4.1G-J). Spen depletion caused significantly larger and more intensely stained LDs compared to controls (Figure 4.1G-H and Figure 4.5A-B), although FB cell size and number were unaffected (Figure 4.5C-D). FB cells overexpressing Spen were smaller, with

LDs of normal size and staining intensity (Figure 4.1I-J and Figure 4.5E-G). As with Spen depletion, the number of FB cells per clone was unaffected by Spen overexpression Figure

4.5H). spen FB mutant clones resulting from flp-mediated mitotic recombination in a heterozygous background produced significantly larger and more brightly stained LDs compared to WT clones Figure 4.3 C-H). We conclude that Spen functions autonomously in

FB cells to regulate the amount of fat stored in LDs.

Figure 4.5: Spen autonomously regulates fat levels in the fat body.

Figure 4.5: Spen autonomously regulates fat levels in the fat body. (A) Lipid droplet (LD) size in Spen KD (iSpen) clones or KD control (iw) clones compared to non-clone cells (denoted as background). Spen KD n=91. W KD n=140. (B) LD intensity in Spen KD or KD control clones compared to non-clone cells. (C) FB cell size of Spen KD and KD control clones. (D) Percentage of numbers of cells within each clone of Spen KD compared to KD control. P value obtained by ANOVA. (E) As in (A), LD size in Spen OEX clones compared to OEX control (GFP) clones and non-clone cells (denoted as background). Spen OEX n=175. GFP n=220. (F) As in (B), LD intensity in Spen OEX clones compared to OEX control clones and non-clone cells. (G) As in (C), FB cell size of Spen OEX clones compared to OEX control clones. (H) As in (D), percentage of number of cells per clone of Spen OEX compared to OEX control. Error bars represent SEM. P values obtained by unpaired two-tailed t test. * P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

Defects in energy utilization upon Spen depletion

Despite their propensity to accumulate extra fat, larvae in which Spen was depleted from the FB died more rapidly than controls when reared from hatching in a sucrose solution, i.e., deprived of fats and amino acids (Figure 4.6A). Spen overexpression had no effect

(Figure 4.4C). If energy stores can be accessed normally, excess energy in the form of fat can provide a crucial advantage during starvation (Britton and Edgar, 1998; Reis et al.,

2010). On the other hand, the advantage is lost regardless of the abundance of stored energy if the mutant animals are unable to mobilize it. Larvae lacking Spen in the FB thus appeared to be defective in accessing energy stores and/or in extracting energy from a limited diet, and may be in a state of “perceived starvation”. Either defect could drive the overfeeding and lethargy that we observed with a regular diet (Figure 4.1E-F). Indeed, FB-specific Spen depletion also caused a one-day developmental delay (19.8 hours ± 1.3 hours as compared to dcg>iw), consistent with a dearth of available energy, although we cannot exclude other causes. These results point to a role for Spen in regulating the liberation of energy stored as fat in the FB.

Alterations in the expression of key metabolic enzymes upon Spen manipulation point to a role in energy catabolism

Spen and its homologs influence other pathways via control of transcription (Ariyoshi and Schwabe, 2003; Newberry et al., 1999; Shi et al., 2001). Accordingly, we suspected that the transcript levels of key metabolic enzymes would be affected by Spen manipulation in the

FB, and tested this prediction using RNA sequencing (RNAseq). 440 of the 516 genes whose levels significantly changed when Spen was KD in the FB were classified by the PANTHER system (Mi et al., 2017). 173 (39.3%) of the classified genes were categorized as being

Figure 4.6: Spen regulates the breakdown of fat. (A) Survival of larvae reared in amino acid-free media. FB-specific Spen KD (dcg>iSpen) compared to KD control (dcg>iw). Fifty larvae per genotype per experimental replicate, n=3. P value obtained by Log-rank test. (B- G) Gene regulation as a result of FB-specific Spen KD from RNA sequencing. Error bars represent SD. P values obtained by ANOVA. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

98 involved in a “metabolic process”, representing the largest functional “biological process” category (followed by “cellular process”, γ0.7%). We observed striking changes in transcripts encoding proteins involved in -oxidation, a process by which fatty acids are broken down to provide acetyl-CoA for the TCA cycle. Though redundant enzymes participate in -oxidation reactions, three key enzymes involved in this pathway were significantly downregulated in Spen KD larvae (Figure 4.6B-D), namely acyl-CoA dehydrogenase, enoyl-CoA hydratase, and 3-hydroxyacyl-CoA dehydrogenase. These enzymes participate in the release of a two-carbon chain from the fatty acid. Furthermore, significant downregulation of trehalase (Figure 4.6E) pointed to a potential blockage in disaccharide catabolism and, as a consequence, glycolysis.

With regard to the high-fat phenotype of Spen-depleted larvae, three lipases, potentially necessary for liberating stored fat, were downregulated (Figure 4.6F).

Additionally, PEPCK (phosphoenolpyruvate carboxykinase) was highly induced (Figure

4.6G), a hallmark of the starvation response (Palanker et al., 2009) that fits with the predicted state of ”perceived starvation” resulting from an inability to access stored fats or dietary energy. Furthermore, 39 of the 126 genes significantly upregulated in Spen-depleted FBs are induced by fasting/starvation (Table 4.1 and (Palanker et al., 2009)), and 69 of the 390 genes significantly downregulated in Spen-depleted FBs, are downregulated upon fasting/starvation

(Table 4.2 and (Palanker et al., 2009)), providing additional evidence of the similarities between Spen depletion and starvation. These findings provide strong support for the role of

Spen in modulating substrate utilization for catabolism and energy production.

Table 4.1: Upregulation of Spen and starvation genes. Genes that are commonly upregulated upon both Spen depletion in the FB and under starvation conditions (Palanker et al., 2009). Common Genes Upregulated Upon Spen KD and Starvation Gene FlyBase ID Gene FlyBase ID CG10799 FBgn0033821 wtrw FBgn0260005 Cyp4p1 FBgn0015037 CG32557 FBgn0052557 CG10924 FBgn0034356 Sp212 FBgn0053329 CG11529 FBgn0036264 CG4025 FBgn0025624 CG13704 FBgn0035583 pug FBgn0020385 PGRP-SC2 FBgn0043575 Idgf1 FBgn0020416 CG15065 FBgn0040734 eIF3-S9 FBgn0034237 CG15067 FBgn0034331 loco FBgn0020278 CG15068 FBgn0040733 sra FBgn0086370 CG16743 FBgn0032322 cert FBgn0027569 IM3 FBgn0040736 puc FBgn0243512 AdSS FBgn0027493 Strica FBgn0033051 Pepck FBgn0003067 CG8160 FBgn0034011 CG18107 FBgn0034330 CG8299 FBgn0034052 IMPPP FBgn0283462 CG8788 FBgn0028955 CG18473 FBgn0037683 Non1 FBgn0028473 CG1882 FBgn0033226 Cyp28a5 FBgn0028940 CG2017 FBgn0037391 Tsp29Fb FBgn0032075 Fer1HCH FBgn0015222 CR10102 FBgn0033927 Obp49a FBgn0050052

100

Table 4.2: Downregulation of Spen and starvation genes. Genes that are commonly downregulated upon both Spen depletion in the FB and under starvation conditions (Palanker et al., 2009). Common Genes Downregulated Upon Spen KD and Starvation Gene FlyBase ID Gene FlyBase ID CG10674 FBgn0035592 Dhpr FBgn0035964 Nedd8 FBgn0032725 CG4729 FBgn0036623 CG1092 FBgn0037228 Mtap FBgn0034215 Npc2g FBgn0039800 CG4995 FBgn0032219 CG11594 FBgn0035484 CG5026 FBgn0035945 GstE13 FBgn0033381 Sgt FBgn0032640 Mesh1 FBgn0039650 CG5515 FBgn0039163 CG12279 FBgn0038080 P5cr FBgn0015781 SmD2 FBgn0261789 spz FBgn0003495 Hrb87F FBgn0004237 CG6180 FBgn0032453 CCHa2 FBgn0038147 CHORD FBgn0029503 Cisd2 FBgn0062442 Bap55 FBgn0025716 CG14715 FBgn0037930 CG6805 FBgn0034179 CG1532 FBgn0031143 CG6908 FBgn0037936 CG15343 FBgn0030029 Cbp80 FBgn0022942 CG15369 FBgn0030105 LanB1 FBgn0261800 CG15717 FBgn0030451 eIF4AIII FBgn0037573 Roc1a FBgn0025638 DNaseII FBgn0000477 Tim8 FBgn0027359 janA FBgn0001280 CG17737 FBgn0035423 SLIRP2 FBgn0037602 CG2004 FBgn0030060 CG8417 FBgn0037744 CG2091 FBgn0037372 CG8778 FBgn0033761 Gip FBgn0011770 mRpL24 FBgn0031651 COX7C FBgn0040773 CG9034 FBgn0040931 CG2611 FBgn0032871 Ddx1 FBgn0015075 CG2767 FBgn0037537 DENR FBgn0030802 Cpr60D FBgn0050163 Trs23 FBgn0260861 CG30499 FBgn0050499 CG9344 FBgn0034564 CG31917 FBgn0031668 l(2)01289 FBgn0010482 CG32069 FBgn0052069 CG9667 FBgn0037550 CG3226 FBgn0029882 CG9853 FBgn0086605 CG3831 FBgn0034804 CG9914 FBgn0030737 CG3887 FBgn0031670 Cyp1 FBgn0004432 Got2 FBgn0001125 Surf1 FBgn0029117 CG4447 FBgn0035980

101

Abnormal metabolic profiles upon Spen manipulation

To define at a molecular level the metabolic defects accompanying Spen manipulation, we performed Ultra-High Pressure Liquid Chromatography (UHPLC)-MS- based metabolomic analysis on larvae in which Spen was knocked down or overexpressed in the FB, along with appropriate controls for each. We monitored 178 metabolites, and found that nearly every metabolite involved in glycolysis was significantly decreased in Spen- depleted larvae (Figure 4.7), consistent with a depletion in these animals of key sources of usable energy, and an accumulation of molecules in which energy is stored. As in most insects, trehalose is the primary circulating sugar in Drosophila, and is broken down to glucose to fuel cellular processes (Matsuda et al., 2015; Wyatt and Kalf, 1957). Consistent with the reduction of trehalase observed by RNAseq (1.9-fold decrease, P = 0.0009, Figure

4.6E), trehalose levels were significantly elevated in Spen KD larvae (Figure 4.7), indicating impaired conversion into glucose and thus decreased glycolytic intermediates.

In addition to defective mobilization of carbohydrate sources for energy production, we found clear defects in -oxidation. Acyl-carnitines are key intermediates of -oxidation that permit fatty acid transport into mitochondria (Stephens et al., 2007), which is the rate- limiting step of -oxidation. Spen KD larvae were significantly depleted of free carnitine as well as nearly every medium- and long-chain fatty acyl-carnitine (Figure 4.7 and Figure

4.8A, C), consistent with the observed decreases in -oxidation enzymes (Figure 4.6B-D) and suggestive of a defect in -oxidation. Finally, the levels of many free amino acids decreased in Spen KD larvae (Figure 4.8B), while markers of protein catabolism n- acetylmethionine and hydroxyproline were increased in Spen KD larvae and decreased in

Spen-overexpressing ones (Figure 4.7), consistent with increased proteolysis in response to

102

Figure 4.7: Spen manipulation results in abnormal metabolism. Quantification of selected metabolites by UHPLC. Spen overexpression (Spen OEX) compared to overexpression control (GFP) and Spen KD (Spen KD) compared to KD control (iw). Ten individual larvae per biological replicate tested per genotype per 3 biological replicates, n=30. Error bars represent SD. P values obtained by two-tailed t test. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

103

Figure 4.8: Spen depletion or overexpression alters acyl-carnitine and amino acid levels. (A) Levels of acyl-carnitines in Spen KD larvae (dcg>iSpen) compared to KD control (dcg>iw) as determined by UHPLC. (B) Levels of amino acids in Spen KD larvae compared to KD control as determined by UHPLC. (C) As in (A), Spen OEX (dcg>Spen) compared to OEX control (dcg>GFP). (D) As in (B), Spen OEX compared to OEX control. P values obtained by unpaired two-tailed t test. * P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001. Error bars represent SD.

104

Spen KD. Among the transcripts that increased significantly in Spen KD larvae are three predicted trypsin-family proteases (CG11529, CG31326, and CG8299, the latter increased

~80-fold) that may be good candidates to mediate elevated protein catabolism. These metabolic changes provide direct evidence of a defect in energy mobilization via catabolism of carbohydrate and lipid energy sources, and may indicate the use of amino acids as an energy source.

Importantly, FB overexpression of Spen had effects opposite to that of Spen depletion with regards to -oxidation, including increased levels of carnitine (Figure 4.7). Spen overexpression did not significantly alter glycolytic metabolites or acyl-carnitine levels, although the steady state of some TCA cycle intermediates and a few amino acids were slightly elevated (Figure 4.7 and Figure 4.8C-D). These findings further indicate that Spen regulates fat catabolism.

Both known domains of Spen are necessary for its function in metabolism

Despite the extreme size of the Spen protein, only the RRMs and SPOC domain have been functionally characterized in the context of other pathways. We obtained two Spen truncation alleles, one that lacks the C-terminal region including the SPOC domain but retains the RRMs (ΔSPOC), and one that retains only the C-terminal region and lacks the

RRMs (SPOConly) (Figure 4.9A). In other contexts, each allele can behave in a dominant- negative fashion. For example, expression of ΔSPOC in midline glial cells results in completely penetrant lethality (Chen and Rebay, 2000). Expression of SPOConly with an engrailed driver reduces or eliminates Senseless expression, suggesting an absolute requirement for this domain of Spen in its regulation of Wg signaling (Lin et al., 2003). To test for dominant negative effects in Spen regulation of fat storage, each of these alleles was

105

Figure 4.9: Domain analysis of Spen and Nito function in metabolism.

106

Figure 4.9: Domain analysis of Spen and Nito function in metabolism. (A) A schematic representation of Spen and Nito proteins and the location of the RNA-recognition motifs (RRMs) and SPOC domains in each. The truncation lines tested for each are pictured below. (B-C) Larvae reared at 16ºC and imaged at wandering stage. (B) dcg>GFP and (C) dcg>ΔSPOC. (D-E) Larvae reared at 16ºC and collected at wandering stage. Larval FB tissue stained with Nile Red. (D) dcg>GFP and (E) dcg>ΔSPOC. (F) As in Figure 4.1A, FB- specific expression of Spen truncation lines with UAS-GFP as a control, n=8. Larvae reared at 16ºC. P values obtained by ANOVA. Error bars represent SEM. (G) As in Figure 4.1A, FB-specific expression of Nito-ΔN with UAS-GFP as a control, n=8. P values obtained by ANOVA. Error bars represent SEM. (H) As in Figure 4.6A, survival rate at 18ºC upon amino acid starvation for FB-specific expression of Spen truncation lines with UAS-GFP as a control. P values obtained by Log-rank test. (I) As in Figure 4.6A, survival rate at 18ºC upon amino acid starvation for FB-specific expression of Nito truncation lines with UAS-GFP as a control. P values obtained by Log-rank test. (J-L) As in Figure 4.1G-J, larval FB tissue ectopically expressing constructs along with GFP (green). Tissues stained with the lipophilic dye Nile Red to mark neutral lipids (red). Dotted white line outlines construct-expressing clones. (J) UAS-GFP, (K) UAS-ΔSPOC, (L) UAS-Nito-ΔC. (M) Table summarizing the density, cellular, and starvation phenotypes for all Spen and Nito manipulations. Low density correlates with higher levels of fat while high density correlates with low levels of fat (Reis et al., 2010). LD: Lipid droplet. FB: Fat body. denotes early larval death (Density column) and dying cells (Cellular column). + denotes sensitivity to starvation, where more +’s indicate increased sensitivity to starvation. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

107 overexpressed in the FB. If the truncation had no effect, we predicted that overexpression would cause a lean phenotype, as observed with Spen overexpression using two independent constructs (Figure 4.1D and Figure 4.9F), whereas a dominant-negative effect would result in a similar phenotype to Spen depletion and elevate fat (Figure 4.1A).

ΔSPOC-overexpressing larvae were unable to survive at β5˚C or 18˚C, arresting at the Lβ stage. At 16˚C, where Gal4 is less active and levels of overexpression are lower

(Duffy, 2002), development was delayed by 11-13 days compared to controls (12 days 2.28 hours ± 23.3 hours as compared to dcg>GFP) and only 5-10% of larvae survived to L3.

Although we cannot exclude a neomorphic effect, we favor the interpretation that this developmental delay is an extreme version of the one-day delay observed upon Spen depletion, and thus is a manifestation of “perceived starvation” resulting from dominant inhibition of Spen function in catabolism. Lγ larvae obtained at 16˚C were tested by buoyancy and compared to larvae overexpressing GFP, SPOConly, or a full-length Spen construct (Spen-FL). Whereas Spen overexpression decreased larval buoyancy (Figure 4.1D and Figure 4.9F), expression of ΔSPOC strongly increased larval buoyancy, and expression of GFP or SPOConly had no effect (Figure 4.9F and Figure 4.10A). Analysis of feeding and activity showed no significant changes (Figure 4.10B-C). By staining isolated tissues of

ΔSPOC larvae with Nile Red to label neutral lipids (Greenspan and Fowler, 1985), we noticed a striking phenotype resulting from ΔSPOC overexpression. FBs in these larvae were almost non-existent (Figure 4.9B-C), but the FB tissue that remained stained much more brightly and contained very large LDs, some of which appeared to have “leaked” out of FB cells (Figure 4.9D-E). Unlike tissues from control animals, brighter staining was also observed in the brains, imaginal discs, and guts of larvae overexpressing Spen ΔSPOC in the

108

Figure 4.10: Ectopic expression of truncated Spen or Nito does not alter behavior.

109

Figure 4.10: Ectopic expression of truncated Spen or Nito does not alter behavior. (A) Percent of floating larvae in different density solutions. Genetic background controls for Figure 4.9F. Fifty larvae per genotype per experimental replicate, n=8 biological replicates per genotype. P values obtained by ANOVA. Error bars represent SEM. (B and E) Absorbance at 530 nm as a measure of food intake, n=4. (B) FB specific OEX of Spen-FL (dcg>Spen-FL) and SPOConly (dcg>SPOC-only) compared to OEX control (dcg>GFP) and genetic background controls (Spen-FL/+, SPOC-only/+, and GFP/+), (E) FB-specific OEX of Nito-ΔN (dcg>Nito-ΔN) compared to OEX control (dcg>GFP) and genetic background controls (Nito-ΔN /+ and GFP/+). P values obtained by ANOVA. Error bars represent SD. (C and F) Average larval speed, pixels/sec. n=4. P values obtained by unpaired two-tailed t test. Error bars represent SEM. (D) As in (A), genetic background controls for Figure 4.9G. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

110

Figure 4.11: Expression of ΔSPOC results in inappropriate fat storage in other organs. (A and B) Larvae reared at 16ºC and collected at wandering stage. (A) dcg>GFP and (B) dcg>ΔSPOC larval brains stained with Nile Red. (C and D) Imaginal discs. (E and F) Guts. (G and H) Salivary glands.

111

FB (Figure 4.11). The appearance of fat deposits in tissues where fat does not normally accumulate is consistent with the elevated body fat phenotype, and is reminiscent of similar effects documented in the Drosophila Seipin mutant lipodystrophy model (Pagac et al., 2016;

Tian et al., 2011).

Analysis of clones of FB cells expressing Spen-FL, ΔSPOC, or SPOConly along with

GFP revealed that, as with other full-length Spen overexpression constructs (Figure 4.1I-J),

Spen-FL overexpression resulted in smaller FB cells (Figure 4.12A-B, E-F, I-J). This particular Spen-FL transgene also decreased LD size, a stronger phenotype than observed with the Spen-OEX transgene (Figure 4.1J and Figure 4.5E-G). SPOConly overexpression resulted in normally-sized FB cells with no significant changes in LD or cell size or morphology (Figure 4.12A, C, G-J), indicating that the SPOC domain is required for the ability of Spen to deplete stored fat when overexpressed. ΔSPOC overexpression, on the other hand, caused a striking phenotype suggestive of catastrophic defects in metabolism.

Specifically, many of the clones consisted of a few extremely small cells containing nuclei

(marked with strong GFP signal) and little else (Figure 4.9J-K). ΔSPOC overexpression may cause pleotropic defects, including cell death. However, considering that similar effects have been previously documented for FB cells during starvation (Butterworth et al., 1965;

Neufeld, 2012; Scott et al., 2004), we favor a model in which the SPOC domain is required for normal Spen function in fat regulation and RRMs alone sequester crucial factors in a non- functional manner. Hence, overexpressing a version of Spen harboring the RRMs but lacking the SPOC domain perturbs the ability of full-length Spen to interact with such factors and carry out its normal function(s).

112

Figure 4.12: Ectopic expression of truncated Spen or Nito does not alter FB cell or LD morphology.

113

Figure 4.12: Ectopic expression of truncated Spen or Nito does not alter FB cell or LD morphology. (A-D) Larval FB tissue ectopically expressing constructs along with GFP (green). Tissues stained with the lipophilic dye Nile Red to mark neutral lipids (red). Dotted white line outlines construct-expressing clones. (A) UAS-GFP (as in Figure 4.9J), (B) UAS- Spen-FL, (C) UAS-Spen-SPOConly, (D) UAS-Nito-ΔN. Clones were obtained without heat shock induction of flp but instead from “leaky” flp expression during FB development. (E) Lipid droplet (LD) size in Spen-FL or OEX control (GFP) clones compared to non-clone cells (denoted as background). Spen-FL n=161. GFP n=147. (F) LD intensity in Spen-FL or control clones compared to non-clone cells. (G) As in (E), LD size in Spen-SPOConly clones compared to control clones and non-clone cells. Spen-SPOConly n=117. GFP n=147. (H) As in (F), LD intensity in Spen-SPOConly clones compared to control clones and non-clone cells. (I) FB cell size of Spen-FL, Spen-SPOConly, and control clones. (J) Percentage of numbers of cells within each clone of Spen-FL and Spen-SPOConly compared to control. P value obtained by ANOVA. (K) As in (E), LD size in Nito-ΔN clones compared to control clones and non- clone cells. Nito-ΔN n=136. GFP n=158. (L) As in (F), LD intensity in Nito-ΔN clones compared to control clones and non-clone cells. (M) As in (I), FB cell size of Nito-ΔN clones and control clones. (N) As in (J), percentage of number of cells within each clone of Nito-ΔN compared to control. Error bars represent SEM. P values obtained by unpaired two- tailed t test. * P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

114

FB overexpression of full-length Spen had no effect on survival during starvation

(Figure 4.9H). Both ΔSPOC and SPOConly were significantly more sensitive to starvation than controls (Figure 4.9H), very similar to Spen KD, although the ∆SPOC effect was far stronger. The ability of SPOConly overexpression to dominantly curtail survival during starvation contrasts with the lack of observed effects on buoyancy or LD appearance in FB cells, and suggests that the roles for Spen in fat storage and the starvation response are not strictly coupled. Phenotypes of all Spen truncation lines are summarized in Figure 4.9M.

Nito antagonizes Spen function in metabolism

In other pathways, Spen and Nito function either redundantly (e.g. Wg signaling

(Chang et al., 2008)) or antagonistically (e.g. EGFR pathway during eye development (Jemc and Rebay, 2006)). To determine the relationship between the two Spen family members in fat regulation, we first depleted Nito from the FB and tested buoyancy. Nito depletion caused a lean phenotype (Figure 4.13A), similar to Spen overexpression. Introducing one copy of a

Nito null allele (Yan and Perrimon, 2015) caused a very slight lean phenotype (Figure 4.3A) that was lost with further outcrossing (Figure 4.3B), hence in the absence of unknown background modifiers Nito is haplosufficient to promote normal fat storage. FB clones in which Nito was depleted had modestly smaller cells and lipid droplets, consistent with the observed lean phenotype (Figure 4.13C-D and Figure 4.14A, C). Cell number and LD intensity were not affected (Figure 4.14B, D). To ask if excess Nito inhibits Spen, we overexpressed full-length Nito in the FB. Larvae were unable to complete development even when reared at 16°C, a phenotype reminiscent of the developmental delays observed upon

Spen depletion or overexpression of Spen-ΔSPOC. Full-length Nito overexpression produced clones that consisted of tiny cells in which only the nucleus was discernable (Figure 4.13E-

115

Figure 4.13: Nito autonomously promotes fat accumulation in the fat body.

116

Figure 4.13: Nito autonomously promotes fat accumulation in the fat body. (A) As in Figure 4.1A, FB-specific Nito KD (dcg>iNito) compared to KD control (dcg>iw) and genetic background controls (iNito/+ and iw/+), n=8. P values obtained by ANOVA. Error bars represent SEM. (B) As in Figure 4.6A, survival rate at 25ºC upon amino acid starvation for FB-specific Nito KD (dcg>iNito) compared to KD control (dcg>iw). P value obtained by Log-rank test. (C-F) As in Figure 4.1G-J, larval FB tissue ectopically expressing constructs along with GFP (green). Tissues stained with the lipophilic dye Nile Red to mark neutral lipids (red). Dotted white line outlines construct-expressing clones. (C) w-RNAi, (D) Nito- RNAi, (E) UAS-GFP, (F) UAS-Nito-FL. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

117

Figure 4.14: Nito autonomously regulates fat levels in the FB. (A) Lipid droplet (LD) size in Nito KD (iNito) or control (iw) clones compared to non-clone cells (denoted as background). Nito KD n=204. w KD n=241. (B) LD intensity in Nito KD or control clones compared to non-clone cells. (C) FB cell size of Nito KD and control clones. (D) Percentage of numbers of cells within each clone of Nito KD compared to control. P value obtained by ANOVA. Error bars represent SEM. P values obtained by unpaired two-tailed t test. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

118

F), similar to the effects of Spen-ΔSPOC. While we cannot at this time rule out effects on cell survival factors unrelated to metabolism, this phenotype is consistent with FB cell death due to starvation (Butterworth et al., 1965; Neufeld, 2012; Scott et al., 2004). Finally, we asked if Nito depletion or overexpression affected sensitivity to starvation. Nito KD larvae died slightly earlier than controls (Figure 4.13B), as would be expected for lean animals with fewer fat stores to draw upon. Overexpression of full-length Nito caused premature death under sucrose-only conditions (Figure 4.9I), consistent with defects in utilization of energy from stores and/or imbalanced diets.

The N-terminal RRMs were required for these effects of Nito overexpression, as larvae overexpressing an N-terminally truncated version (Nito-ΔN) developed normally at all temperatures and were indistinguishable from controls with regard to buoyancy or other metabolic behaviors (Figure 4.9A, G, M and Figure 4.10D-F). On the other hand, overexpression of Nito-ΔC, which retains the RRMs but lacks the SPOC domain, caused Lβ arrest regardless of temperature. Nito-ΔN overexpression did not affect lipid storage, cell size, or cell number (Figure 4.12 A, D, K-N). In striking contrast, expression of a Nito-ΔC construct lacking the SPOC domain phenocopied overexpression of full-length Nito, with the majority of clones containing tiny cells (Figure 4.9L, M and Figure 4.13F). Importantly, the lack of phenotypes resulting from Nito-∆N did not reflect a failure to localize to the nucleus, as both Nito truncations localize appropriately (Jemc and Rebay, 2006). Nito-ΔC- overexpressing larvae were sensitive to starvation, similar to the effects of full-length Nito

(Figure 4.9I, M). Overexpression of Nito-ΔN caused starvation sensitivity that was milder than what we observed for full-length Nito or Nito-ΔC (Figure 4.9I), analogous to the effects of Spen-SPOConly overexpression (Figure 4.9H, M). Taken together, these data support a

119 model wherein Nito antagonizes Spen function in catabolism of stored energy in a mechanism that requires both the RRMs and SPOC domain, with SPOC-less Nito RRMs able to act in a potent dominant-negative manner.

Levels of mSpen and mNito directly correlate with fat levels in mammalian adipose tissue

If mSpen and/or mNito function is important in preventing excess fat accumulation in mammals, we predicted that driving fat accumulation via a high-fat diet (HFD) might trigger changes in the expression of these genes in mice. For individual mice fed either normal chow or a HFD for 30 weeks, we measured both body fat percentage (mass of isolated white adipose tissue (WAT) divided by body mass) and, via RT-qPCR, mSpen or mNito transcript levels in the isolated uterine WAT. The HFD increased body fat by ~2.6-fold on average

(54.2 ± 1.8%, n = 7 for HFD compared to 20.9 ± 2.9%, n = 5 for normal chow, unpaired t test

P < 0.0001). Strikingly, both mSpen and mNito transcript levels (normalized to levels of 4 housekeeping genes) correlated strongly with body fat percentage (Figure 4.15, R = 0.65, P

< 0.05 for mSpen and R = 0.74, P < 0.01 for mNito by unpaired two-tailed t test). While further studies will be required to determine how changes in mSpen and mNito expression in animals made obese by a HFD reflect the normal functions of these proteins, we take these data as evidence that Spen and Nito functions in fat storage are conserved from flies to mammals.

Discussion

Our work provides the first detailed investigation of a fat regulatory role for Spen in any organism, and the first evidence that Nito also functions in this process. Spen depletion in the FB drastically increased stored fat (Figure 4.1A-B). Spen has been implicated in multiple pathways involved in endocrine signaling, including Notch

120

Figure 4.15: Spen and Nito transcript levels are modulated by body fat levels in mouse adipose tissue. (A) mSpen levels with respect to percent body fat in mice fed either a normal chow (solid shapes, n=5) or HFD (open shapes, n=7). Two technical replicates per sample (circles vs. squares). P value and correlation coefficient obtained by unpaired two- tailed t test from the average of the technical repeats. P < 0.05, r = 0.65. (B) As in A, mNito levels with respect to percent body fat in mice. P < 0.01, r = 0.74.

121

(Jin et al., 2009; Zolotukhin et al., 2009), Wingless (Lin et al., 2003), and nuclear receptor signaling (Ludewig et al., 2004; Shi et al., 2001; Sierra et al., 2004). We find it unlikely that nuclear receptor pathways are relevant to the fat regulatory role we define, because we did not observe upon Spen depletion or overexpression consistent changes in the expression of genes that are targets of those pathways. Furthermore, the lack of phenotypes involving fat storage per se upon overexpression of Spen-SPOConly (Figure 4.9F) argues against a role for

Wg signaling, in which the same construct has potent dominant negative effects (Lin et al.,

2003). Conversely, whereas a C-terminally truncated version of mSpen has little effect on

Notch signaling (Kuroda et al., 2003), the strong fat phenotypes resulting from Spen-ΔSPOC overexpression suggest that Spen does not regulate fat via the Notch pathway.

Notably, Spen KD larvae also exhibited behavioral changes (increased food intake, decreased locomotion) that may have contributed to the fat increase (Figure 4.1E-F). Thus, in addition to direct roles in fat accumulation within fat storage cells, Spen may be involved in a cross-talk pathway between the FB and the brain. However, we strongly support a model wherein increased food intake is instead an attempt to compensate for a condition of

“perceived starvation” resulting from an inability to access energy stores. Similarly, a lack of available energy may restrict locomotion. This hypothesis is further strengthened by the observation that Spen overexpression was sufficient to deplete stored fat (Figure 4.1D) but did not cause opposing behavioral phenotypes (Figure 4.4A-B).

Mosaic analysis confirmed an autonomous role for Spen in FB cells. Spen KD in clones throughout the FB showed a striking increase in LD size (Figure 4.1H and Figure

4.5A). Larger LDs normally have lower surface tension, and the stored fat is easier to access

(Thiam et al., 2013). LD remodeling in WT animals is a highly regulated process involving

122 specific factors, some of which were identified in a genome-wide RNAi screen in cultured

Drosophila S2 cells (Beller et al., 2008; Guo et al., 2008; Lee et al., 2013). Notably, our

RNAseq data revealed that the products of several LD-regulating genes were significantly altered by Spen depletion, including l(2)01289 (~7-fold decreased, P < 0.0001 by unpaired two-tailed t test), CG3887 (1.3-fold decreased, P = 0.001), and eIF3-S9 (1.5-fold increased, P

= 0.0008). While it is unclear if these changes are direct effects of Spen depletion, they may explain why LDs in Spen KD larvae are large yet apparently inaccessible, resulting in starvation sensitivity.

Consistent with the observed changes in FB cell and LD morphology and starvation sensitivity, changes in metabolites and gene expression in Spen KD larvae pointed to a drastic defect in lipid catabolism. Defects in -oxidation were the most obvious, in part because the opposite effects were observed upon FB-restricted Spen overexpression. Spen depletion led to a decrease in the levels of free and acyl-conjugated carnitine, as well as of transcripts of three of the four enzymes necessary to break down acyl-carnitines into free fatty acids (Figure 4.6B-D and Figure 4.7). Three lipases were also downregulated (Figure

4.6F), which likely further contributes to an inability to convert energy stored as TAGs into usable forms. While an apparent upregulation of gluconeogenesis is evident, as supported by alterations in aspartate (Figure 4.7) and PEPCK expression (Figure 4.6G), these processes may be unable to completely compensate for decreased trehalose utilization, and these defects may contribute to the lethargy phenotype resulting from Spen KD. Consequently, surviving the loss of Spen may require breakdown of protein into free amino acids in order to anaplerotically replenish the TCA cycle, consistent with changes in expression of proteases, the observed decrease in many free amino acids (Figure 4.8B), as well as increases in protein

123 catabolism and collagen turnover markers (N-acetylmethionine and hydroxyproline) (Figure

4.7). Of note, sustained proteolysis is a marker of aging and inflammation, a phenotype that has been previously associated with decreased locomotion in human and mouse models of physical activity, suggesting potential future ramifications of Spen’s role in metabolism with respect to aging/inflammation research (Cavalli et al., 2017). Finally, the observed decrease in glycogen levels upon Spen KD (Figure 4.1C) supports a model wherein glycogen is used as a carbohydrate source (in lieu of decreased levels of trehalose) to fuel glycolysis (Figure

4.6E and Figure 4.7). The overall metabolic defects we describe are distinctly different from what has been observed upon manipulation of other fat regulators (e.g. Sir2 (Reis et al.,

2010)), suggesting that Spen operates in a previously undescribed pathway.

Our results with Spen and Nito truncations provide additional mechanistic insight into how these proteins function in fat regulation. Overexpressing Spen-ΔSPOC reversed the phenotype of full-length Spen overexpression, and instead resulted in similar phenotypes to

Spen depletion. Nito-∆C overexpression had the same effects: larvae arrested development and FB clones mimicked starvation even when dietary nutrients were abundant.

Overexpression of the Spen-SPOConly construct had no effect on FB cells, as was the case for

Nito-∆N. Thus only Spen harboring the RRMs and the SPOC domain was able to promote fat depletion when overexpressed. Conversely, only truncated forms of Spen or Nito that retain the RRMs dominantly perturbed both FB cell viability and organismal resistance to starvation.

Recent studies of X chromosome inactivation found that mSpen RRMs mediate binding to the lncRNA XIST (McHugh et al., 2015; Moindrot et al., 2015; Monfort et al.,

2015; Roth and Diederichs, 2015). Rbm15 (mNito) also binds XIST (McHugh et al., 2015;

124

Moindrot et al., 2015), and is required for N6-methyladenosine (m6A) modification of that lncRNA, which is in turn required for its ability to repress X chromosome transcription (Patil et al., 2016). Nito is a subunit of the Drosophila m6A methyltransferase complex and is required for RNA binding by that complex; Nito knockdown severely decreases global m6A modification of mRNA (Lence et al., 2016). Interestingly, the m6A demethylase

FTO/ALKBH9 was the first human obesity susceptibility gene identified by genome-wide association studies (Dina et al., 2007; Frayling et al., 2007; Scuteri et al., 2007), but the relevant nucleic acid target(s) remain unknown. Our work provides the first hint that an RNA bound by Spen and/or Nito may be a key FTO substrate.

These findings lead us to propose a model for Spen and Nito function in the regulation of fat storage (Figure 4.16). Spen binds via its RRMs to one or more RNAs and, via recruitment of other factors, promotes the expression of enzymes key for mobilization of energy stored as fat (e.g. lipases). The mechanism of activation may be direct or indirect, and via alternative splicing, activation/repression of transcription, or effects on RNA stability and/or translation. Moreover, RNA binding partners may be mRNA or non-coding RNA.

Future work will be required to make these distinctions. We propose that the Spen SPOC domain is critical for this function, but undefined domains in between the N-terminal RRMs and C-terminal SPOC domain are also important, and these are not shared with Nito. We propose that Nito binds via its RRMs the same or a largely overlapping set of RNAs, and also recruits additional factors via its SPOC domains, but – either because it fails to recruit specific factors recruited by Spen, or because it recruits other factors not recruited by Spen -

Nito ultimately inhibits/represses the same energy-storage-mobilizing enzymes that are activated by Spen (Figure 4.16). Overexpressed Spen or Nito fragments containing RRMs

125

Figure 4.16: Model: Spen family members counter-regulate metabolism. Our model predicts that Spen and Nito bind the same or similar RNAs via the RRMs as well as transcription factors via the SPOC domain. Spen acts to activate enzymes key for the mobilization of energy stored as fat while Nito antagonizes this function. Spen may achieve this activation by binding additional factors in the uncharacterized region between the RRMs and SPOC domain not found in Nito.

126 sequester target RNAs away from endogenous full-length Spen and the other effectors of fat storage control. Finally, our findings in mouse adipose tissue that mSpen and mNito both increase in expression when a HFD drives fat accumulation lead us to believe that in WT animals Nito acts as a counterbalance to Spen in order to fine-tune fat storage. Future studies delving into more mechanistic details may lead to treatments for obesity and related metabolic disorders that result from perturbation of the pathway that we elucidate here.

Acknowledgements

We would like to thank the Bloomington Stock Center and Vienna Drosophila

Resource Center for fly stocks, Ilaria Rebay, Ken Cadigan, Norbert Perrimon, and Bertrand

Mollereau for generously providing the Spen and Nito fly stocks, Nick Galati for writing the

ImageJ algorithm used to analyze the lipid droplets, and Michael McMurray, Sandy Martin, and Joan Hooper for their helpful comments on the manuscript.

127

CHAPTER V

DISCUSSION

Conclusions and Discussion

Obesity is a complex disease with many different factors that contribute to its development. Foremost among the contributing factors is genetics, which determines how much or little impact obesogenic behaviors will have on the amount of fat stored by a particular person. Common obesity is not caused by a single gene but rather from the combination of many. While many genes that contribute to the development of obesity are known, there are far more that have yet to be characterized. This thesis sought to identify novel obesity predisposing genes that have autonomous roles in the FB of Drosophila and to follow up on characterization to identify the pertinent mechanisms that may provide a therapeutic target in the future.

Previous work by Reis et al. had identified 66 genes that resulted in increases in fat levels when mutated (Reis et al., 2010). In Chapter III, I began by identifying which of the genes in this list had FB-specific roles in regulating fat levels. Relatively few of the genes tested (4/14) had FB-specific functions in regulating stored fat (Figure 3.1). Among these genes was Spen, which increased fat levels when knocked down in the FB. In contrast, Fabp,

NFAT, and Shep all reduced fat levels when knocked down in the FB, suggesting a pro-fat storage role in that tissue (Figure 3.2, 3.3, 3.4). Since the results of the initial screen by Reis et al. showed increases in fat levels in these mutants, I hypothesized that these genes may have opposing roles in other metabolic tissues in the larvae. I tested this hypothesis with

Shep by knocking down levels of the gene in the brain. This resulted in a mild fat phenotype along with a sharp decrease in locomotion and a slight (but non-significant) increase in food

128 intake (Figure 3.5). Therefore, the increase in fat levels upon Shep knockdown in the brain may be due wholly or in part to changes in behavior of the larvae. Furthermore, Shep levels in the brain are manipulated by the dietary intake of the larvae, with higher levels of nutrition resulting in lower expression of Shep (Figure 3.6). This may indicate that Shep acts as a nutrient sensor to regulate the production or secretion of dILPs to regulate energy intake and thereby fat storage. Alternatively, Shep levels may be influenced by dILP levels and may function to mediate the production or secretion of other neuropeptides that control metabolism or metabolic behaviors. Future work will help to tease out the mechanisms of

Shep in regulating body fat in both the brain and FB.

In Chapter IV, I characterize the catabolic role of Spen. Spen is both necessary and sufficient to deplete fat in the FB (Figure 4.1). Knocking down Spen in this tissue results in major changes to the metabolic profile of the larvae, including decreases in many metabolic enzymes including those involved in fat breakdown as well as a large number of metabolites

(Figure 4.6, 4.7). Both the RRMs and SPOC domains are necessary for Spen’s role in metabolism. Expressing a construct containing only the RRMs and middle of the protein but not the SPOC domain (ΔSPOC) results in severe metabolic defects including lethality, delayed development, starvation sensitivity, dying FB cells, and highly increased fat levels

(Figure 4.9). In contrast, expression of only the SPOC domain (SPOConly) causes no changes in fat levels, starvation sensitivity, or cellular phenotypes (Figure 4.9). If either domain were sufficient to perform Spen’s metabolic function, expression should have recapitulated Spen overexpression. Lacking this result, it is clear that both domains are necessary but not sufficient to promote fat catabolism.

129

Given that the ΔSPOC truncation was able to act as a potent dominant negative for fat levels but that the SPOConly truncation was not suggests that the SPOC domain is reliant on another to bind its partner. Otherwise, expression of SPOConly may also have sequestered necessary binding partners away from WT Spen in a similar manner to ΔSPOC. I favor a model wherein the RRM domains bind to a lncRNA or other molecule that upon binding allows the SPOC domain to bind its partner. This may be due to a conformational change in

Spen upon the RRM binding to its partner or alternatively, the molecule binding the RRMs may help to target a binding partner to the SPOC domain. Only when both domains are present can Spen bind all partners necessary to perform its function in regulating catabolism.

Interestingly, SPOConly is sufficient to cause starvation sensitivity despite no change in fat levels. This may suggest that Spen has a secondary role in regulating starvation that is unrelated to changes in fat levels. In this instance, the SPOC domain may be able to bind to the necessary partner for this function without the RRM binding to its partner first, causing sequestration and simulating Spen knockdown.

While I favor the model wherein the RRMs are the important binding domain for

Spen’s metabolic and starvation functions, it is very possible that there is an alternate undefined domain within the middle of Spen that is necessary. While the Drosophila Spen protein only has the RRMs and SPOC domains defined, the mouse and human orthologues of this protein have both a nuclear receptor binding domain and a Notch binding domain within the middle portion of the protein (Figure 1.5). Comparison between the various orthologues of Spen show high conservation between the RRMs and SPOC domains, but very little conservation surrounding the Notch and nuclear receptor binding domains (Figure B1). This

130 makes it unlikely that Drosophila Spen is regulating metabolism through these domains, however it does not preclude an alternative undefined domain from being involved.

Spen is the founding member of a family of SPOC-containing proteins. Nito, a smaller protein in the family, is also expressed in the FB of Drosophila. Nito appears to counteract Spen’s catabolic role as knockdown of the gene in the FB results in a striking lean phenotype in opposition to the fat phenotype caused by knockdown of Spen (Figure 4.13).

The RRM and SPOC domains in Nito are also necessary but not sufficient for its anabolic function and phenocopy the truncation mutants for Spen. The contrary function of Spen and

Nito may suggest that they counterbalance each other to regulate fat storage levels.

Since the truncation proteins for both Spen and Nito result in extremely similar phenotypes, it is likely that they bind to similar or the same partners. I favor a model wherein

Nito acts in opposition to Spen by sequestering binding partners necessary for Spen’s function. This model is based on the fact that Nito-FL and Nito-ΔC have nearly identical fat storage, starvation, and cellular phenotypes, which suggests that sequestration of the RRM

(or other domain) binding partners is sufficient to disrupt Spen’s function and that there is not an additional role for Nito in repressing the transcription of Spen’s target genes.

However, Nito is known for regulating transcription and may indeed be competing with Spen for binding to target genes to regulate transcription. The opposing roles of Spen and Nito may be due to Spen binding to an additional partner in the large undefined middle region that gives it an opposing transcriptional effect to Nito. Alternatively, Spen and Nito may not be affecting target genes through transcription at all. The human orthologues of Spen and Nito have both been implicated in nuclear export and alternative splicing (Hiriart et al., 2005), which may be the mechanism through which they regulate the expression of target genes to

131 regulate metabolism. It is also possible that Spen and Nito regulate metabolism through another mechanism entirely. Nito has recently been found as a member of the m6A methyltransferase complex (Lence et al., 2016). Intriguingly, FTO, a well-known obesity predisposing gene, acts as a demethylase for this process. Perhaps FTO and Spen function to demethylate a similar subset of m6A-modified transcripts leading to decreased fat storage while Nito methylates the same transcripts leading to increases in fat levels. Although Spen has not been implicated in m6A-demethylation, it has also never been directly tested. Future work will help to clarify the mechanism of antagonism between Spen and Nito.

Both Spen and Nito are highly conserved in mice and humans (Figure 1.5). To test if the metabolic function of these genes in the FB of Drosophila are conserved in mammalian adipose tissue, I measured levels of mSpen and mNito expression in WT normal weight and

HFD-fed obese mice and compared to the fat content of the mice. Both mSpen and mNito are positively correlated with the levels of fat in the WAT of the mice, which suggests that both genes function to regulate fat levels in adipose tissue. The correlate increase of both proteins may also provide evidence for counterbalancing each other in order to finely adjust fat storage in mice. While adipose tissue in mice is an obvious first place to look for conserved fat regulation roles for Spen and Nito, the Drosophila FB also performs the functions of the liver. Preliminary experiments have shown a different pattern of Spen and Nito expression when compared to fat levels in the livers of the same mice tested above. More experiments with mouse liver tissue and hepatic cell culture will help to elucidate the differences in Spen and Nito functions in these two tissue types.

Prior to this work, Spen family members had not been characterized as having a metabolic function in any organism. This work represents the identification and

132 characterization of Spen and Nito as novel obesity predisposing genes with conserved roles in mammalian tissue. This will further our understanding of the development of obesity in humans and may provide a therapeutic target in the future. Likewise, this work identified the first tissue-specific fat regulatory role for NFAT and Shep. Although both have been associated with changes in fat levels in mammals, the mechanisms have been lacking. This initial characterization will provide a baseline from which others can determine the precise metabolic function of these obesity predisposing genes.

An interesting larger question that this work presents is the role of RNA-binding proteins (RBPs) in metabolism. Three of the four genes identified as novel obesity predisposing genes in this work (Shep, Spen, Nito) contain RRM domains. There is some precedence for metabolic enzymes “moonlighting” as RBPs (Castello et al., 2015). However those that have been identified as “moonlighters” have primary roles in metabolism and are not upstream regulators of metabolism like Spen, Nito, and Shep appear to be. Some other

RBPs are known to regulate the expression or secretion of metabolic enzymes (Mobin et al.,

2016), but it is by no means a high proportion of known metabolic regulators. This work suggests that there may be a larger role for RBPs in regulating metabolism than has been previously identified.

Future Directions

While this work has answered many questions, many more have been brought to light. There are multiple avenues of inquiry remaining to be followed for every part of the project that will lead to valuable information on the genetic basis of obesity and the mechanisms of action for Shep and Spen family proteins. To begin, there are many more genes remaining from the initial screen by Reis et al. to be tested for a metabolic function in

133 the FB. While half of the genes were tested for expression using FlyTrap expression lines, 33 remain without an available expression line. These other genes can be tested for expression in dissected FB from WT flies by qPCR. For those that are found to have expression in the

FB, levels of the genes can be manipulated in the FB and tested for changes in buoyancy to determine if they have a role in regulating metabolism within the tissue. Although I had not confirmed expression in the FB, I tested the metabolic function of one of these genes Ras- like protein A (RalA). Knockdown of this gene in the FB produced a lean organismal phenotype with no changes in metabolic behaviors (Figure C1). Furthermore, there are five genes (CG15309, Sea, Trl, Crc, and Fru) with confirmed FB expression that have yet to be tested for a role in the FB (Figure 3.1). Like NFAT and Shep, RalA and the other untested genes all provide new potential avenues of research to determine their contribution to fat regulation.

Shep appears to be important for metabolism through its action both in the FB and in the brain. However, it is still unclear what changes in metabolism are taking place and how

Shep is contributing. A first interesting step to take next to define how Shep regulates metabolism is to test how metabolic enzymes are changing with Shep knocked down in the

FB and the brain. Since Shep’s regulation of fat levels appears to be in opposition in these tissues, perhaps the metabolic enzymes that are affected are also affected in opposite directions (e.g. glycolysis enzymes upregulated in the FB Shep KD but downregulated in the brain Shep KD). This could be easily done by knocking down Shep in each tissue individually and extracting RNA from the whole larvae to test by qPCR.

Next, the functional domains of Shep are yet unknown for this process. Shep has only been characterized as having two RRM domains, but it is likely that there are other parts of

134 the protein that are important to regulate metabolism. To test the functions of the different parts of the protein, truncation mutants could be created containing only small parts of the protein. Overexpression of these truncations could provide evidence for how important each area of the protein is. As there are not any truncation lines available for Shep, it would be a long process of cloning each desired truncation and injecting fly embryos to create the necessary fly lines. However, the information gained would be very interesting, particularly to investigate the importance of the RRM domains and potentially draw parallels to Spen and

Nito. Furthermore, these truncations could be tested for importance in both the brain and FB.

It is possible that Shep utilizes different parts of the protein in these tissues to perform its opposing roles. Determining the functional parts of the protein will also lead to other future experiments to determine the binding partners and targets of the protein, which will further help to identify Shep’s mechanism.

Perhaps one of the more interesting aspects of the Shep project is its expression regulation by diet in the brain. Further preliminary studies by Claire Gillette have also determined that Shep expression is not only altered by diet in the brain, but also in the same manner in the FB. dILP expression is well known to be regulated by nutritional intake.

Shep’s sensitivity to diet and regulation of metabolism suggests that it may affect or be affected by dILPs to perform this role. One possibility is that Shep alters dILP expression or secretion, which then regulates fat storage. This can be tested by manipulating Shep levels in either the brain or FB and measuring the changes in expression levels of dILP using immunohistochemistry (IHC). Traditionally, decreases in dILP signal are interpreted as increases in dILP secretion (Park et al., 2014). However, this does not distinguish between changes in insulin production from changes in insulin secretion. Secretion of dILP can be

135 specifically tested by extracting hemolymph from the larvae by means of puncture and centrifugation and measuring dILP levels by ELISA (Bai et al., 2012). An increase in levels relative to controls would suggest increased dILP secretion due to altered Shep expression.

Based on the decrease in Shep expression on HYD (Figure 3.6), I predict that Shep negatively regulates insulin secretion, which should increase upon higher nutritional intake.

An alternative possibility for a connection between Shep and dILPs is that Shep is altered by dILP expression or secretion. This can be tested by manipulating dILP levels in the

IPCs and measuring changes in Shep expression in the brain and FB. Preliminary results by

Claire Gillette show no significant changes in Shep expression upon dILP2 overexpression in the IPCs. However, overexpression does not necessarily mean increased secretion of dILP2.

To force secretion, both dILP2 and NachBach (NB), a sodium channel that activates neuronal secretion (Luan et al., 2006; Ren et al., 2001), could be expressed in the IPCs. If insulin secretion is inhibiting Shep expression, Shep expression should decrease upon dILP2 and NB co-overexpression.

The changes in behaviors upon Shep knockdown in the brain (Figure 3.5) suggest that Shep may be influencing fat storage through alterations in the secretion of neuropeptides involved in feeding and locomotion behavior. Shep is co-expressed with peptidergic neurons

(Chen et al., 2014). To test if Shep is altering neuropeptide release to regulate behavior, hemolymph can be extracted from larvae in which Shep has been knocked down in the brain and tested for changes in neuropeptides like NPF and hugin, which influence feeding behavior (Nassel and Winther, 2010) and tachykinin, which influences locomotion (Winther et al., 2006). Each of these experiments will help to elucidate how Shep functions to regulate fat storage in Drosophila.

136

Although much has been done to understand how Spen regulates fat catabolism, there are many more unanswered questions to be addressed. One such question is the molecular mechanism through which Spen causes these changes in metabolism. The work with the truncation lines has implicated the SPOC domain and another domain in the N- terminal side of the protein, potentially the RRMs. Performing immunoprecipitations (IPs) with the SPOConly truncation would help to identify binding partners to the SPOC domain.

Furthermore, CLIPseq would help to identify any RNAs that bind to the RRMs in Spen.

These experiments could also be performed using Nito truncations. I predict that Spen and

Nito are binding to a similar subset of partners since the phenotypes observed in their comparable truncation lines were nearly identical. Once the binding partners are found, knockdowns for each gene in the FB followed by buoyancy testing would help to determine if the binding partner is necessary for Spen’s role in metabolism. Binding partners found in both the Spen and Nito pulldowns should be prioritized. If Spen cannot promote catabolism without binding to a certain RNA or transcription factor, knockdown of the binding partner would result in a phenotype similar to Spen knockdown. If there is an additional domain in the middle of Spen that is necessary for its function, truncation lines containing only parts of the middle of the protein could be used to IP and identify binding partners. Together, these experiments would help to more fully characterize Spen’s mechanism in regulating metabolism.

The antagonism between Spen and Nito is another area that would be interesting to pursue. To determine which of the genes has the dominant phenotype, Spen and Nito double- knockdowns or double-overexpressions could be performed in the FB and the larvae tested for buoyancy. The correlate increase in expression levels of Spen and Nito in mammalian

137 tissue suggests that Spen and Nito may counterbalance each other to adjust fat storage

(Figure 4.15). This suggests that they might regulate the expression level of one other. This could be tested by clonally manipulating levels of Spen and measuring the expression changes of Nito in the clones as compared to the surrounding WT cells by IHC.

Alternatively, Nito could be clonally manipulated and Spen levels measured. However, existing Spen antibodies are very poor and this would likely be difficult to determine. An alternative to this experiment would be to knockdown or overexpress levels of Nito in FB, dissect the FB, and measure levels of Spen by qPCR. When compared to a knockdown or overexpression control, this would help to determine if Nito affects the expression of Spen.

However, it is possible that feedback regulation of Spen or Nito may not be a mechanism of control in fly cells. Nito was not found to be significantly altered by Spen manipulation in the

RNAseq data. This may instead be a potential feedback mechanism in mammalian tissues only, and can be tested by analyzing Nito transcript levels in mammalian adipocytes with overexpressed or knocked-down Spen. By determining whether Spen and/or Nito regulate expression of the other, the fine-tuning of fat catabolism will become more apparent.

Knockdown of Spen results in a unique metabolic phenotype that may implicate increased proteolysis in lieu of a decrease in fat breakdown (Figure 4.6, 4.7). This suggests that diet may help improve the outcome for larvae in which Spen is depleted. This could be easily tested by supplementing their diet with increased yeast, sugar, fat, or protein and measuring for an improvement or decline in Spen knockdown-related phenotypes.

Particularly interesting to follow would be fat levels of the larvae, developmental delays, and longevity. Spen knockdown causes ~1 day delay in development. Preliminary studies show that there might be a slight rescue in the development for Spen knockdown larvae fed on

138

HYD (the improvement in pupariation and eclosion time observed between HYD- and MYD- fed Spen knockdown larvae appears to be greater than the improvement observed between

HYD- and MYD-fed knockdown control larvae) (Figure C2). High sugar diet (HSD) and high fat diet (HFD) do not show significant changes (Figure C2). Furthermore, Spen knockdown flies have a severely decreased lifespan compared to control flies (Figure C3).

There is a well-known and strong correlation between obesity and aging (Ahima, 2009;

Bluher, 2008). If HYD is able to partially rescue developmental delays in Spen knockdown larvae, it is possible that it may help improve their longevity as well. HYD contains extra yeast, which is a major source of protein, but also a source of vitamins and some fats. To determine if the increased protein content is the important factor of the diet in improving

Spen knockdown larval development, high protein food (HPD) with added amounts of casein could be used instead. By testing the developmental and longevity effects of different diets for larvae with depleted Spen, valuable information can be gained to help the treatment of any people that suffer from obesity due to mutations in the Spen locus.

While all of these experiments will help to elucidate the functions of Spen and Nito in

Drosophila, a major remaining area of work is to determine whether the function and mechanism of these proteins is conserved in mammalian tissue. My results from the changes in expression in mSpen and mNito on mice fed a HFD suggest that these genes play a role in fat regulation in adipocytes (Figure 4.15). One relatively quick way to determine if Spen’s mechanism is highly conserved is to express the human or mouse ortholog of Spen in

Drosophila FB tissue and see if they can recapitulate the Spen-FL phenotype observed with the Drosophila orthologue. A more comprehensive way to test mSpen’s mechanism in mammalian tissue is through adipocyte tissue culture. Preliminary data with transient mSpen-

139

FL expression in 3T3-L1 adipocytes resulted in decreased LD size, intensity, and number, similar to phenotypes observed upon Spen-FL expression in the Drosophila FB. I am currently working on cloning several truncations for mSpen (FL, ΔRRM, ΔSPOC, RRMonly,

Middleonly, and SPOConly) into lentiviral vectors to create stable 3T3-L1 lines for each. This will allow us to test how the expression of each affects the storage of fat and the metabolism of the cell. Finally, given that Spen manipulation results in organismal changes to the fly, following up the characterization of mSpen in a mouse model would be highly interesting.

MINT (mSpen)-floxed mice have been developed that allow for the Cre/loxP-mediated conditional knockout of mSpen (Yabe et al., 2007). This would provide a system to test the adipocyte-specific functions of mSpen and the organismal consequences of its depletion.

MRI or DXA scanning could be used to determine increases in organismal fat levels and measuring blood glucose would give an indication of whether the mild diabetic phenotype observed in flies is also found in mammals. Blood samples could also determine the level of circulating lipids. Diet challenges to the mice would be a useful follow up to the diet experiments in flies to provide evidence on whether a dietary intervention would be helpful to improve outcome. All together, these experiments will help to clarify the metabolic role for Spen in mammalian tissue.

The data presented in this thesis bring the field of metabolism closer to understanding the genetic contributions to obesity. The secondary screen presented in Chapter III identified the expression pattern of many novel targets that may be regulating metabolism. NFAT and

Shep were furthermore characterized as having a pro-fat storage role in the FB, while Shep was shown to have an opposing role in the brain. Diet regulates expression of Shep in the brain, which implicates that it may be functioning in tandem with dILP. Further work will

140 elucidate the mechanism by which Shep regulates fat storage based on diet as well as the differences in its mechanisms between the brain and FB. Chapter IV presented a full characterization of the novel obesity predisposing gene Spen and how it promotes fat catabolism autonomously in the FB. The metabolic defects apparent in larvae with depleted

Spen are quite unique and appear to be conserved in mammalian tissue. Further work will identify the binding partners necessary for this function as well as the mechanism of action in mammalian tissues. Overall, this work presents many novel targets that act as obesity predisposing genes and improve the understanding of the development and genetic causes of obesity.

141

REFERENCES

Ahima, R.S. (2009). Connecting obesity, aging and diabetes. Nat. Med. 15, 996–997.

Al-Anzi, B., and Zinn, K. (2010). Colorimetric measurement of triglycerides cannot provide an accurate measure of stored fat content in Drosophila. PLoS One 5, e12353.

Al-Anzi, B., Sapin, V., Waters, C., Zinn, K., Wyman, R.J., and Benzer, S. (2009). Obesity- blocking neurons in Drosophila. Neuron 63, 329–341.

Ali, A.T., Hochfeld, W.E., Myburgh, R., and Pepper, M.S. (2013). Adipocyte and adipogenesis. Eur. J. Cell Biol. 92, 229–236.

Allison, D.B., Heshka, S., Neale, M.C., and Heymsfield, S.B. (1994). Race effects in the genetics of adolescents’ body mass index. Int. J. Obes. Relat. Metab. Disord. 18, 363–368.

Allison, D.B., Kaprio, J., Korkeila, M., Koskenvuo, M., Neale, M.C., and Hayakawa, K. (1996). The heritability of body mass index among an international sample of monozygotic twins reared apart. Int. J. Obes. Relat. Metab. Disord. 20, 501–506.

Ariyoshi, M., and Schwabe, J.W.R. (2003). A conserved structural motif reveals the essential transcriptional repression function of spen proteins and their role in developmental signaling. Genes Dev. 17, 1909–1920.

Armstrong, J.D., Texada, M.J., Munjaal, R., Baker, D.A., and Beckingham, K.M. (2006). Gravitaxis in Drosophila melanogaster: A forward genetic screen. Genes, Brain Behav. 5, 222–239.

Asha, H., Nagy, I., Kovacs, G., Stetson, D., Ando, I., and Dearolf, C.R. (2003). Analysis of Ras-induced overproliferation in Drosophila hemocytes. Genetics 163, 203–215.

Ashburner, M., Golic, K., and Hawley, R. (2005). Drosophila: A Laboratory Handbook (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press).

Azeez, O.I., Meintjes, R., and Chamunorwa, J.P. (2014). Fat body, fat pad and adipose tissues in invertebrates and vertebrates: the nexus. Lipids Health Dis. 13, 71.

Bai, H., Kang, P., and Tatar, M. (2012). Drosophila insulin-like peptide-6 (dilp6) expression from fat body extends lifespan and represses secretion of Drosophila insulin-like peptide-2 from the brain. Aging Cell 11, 978–985.

Baird, N.L., Bowlin, J.L., Cohrs, R.J., Gilden, D., and Jones, K.L. (2014). Comparison of varicella-zoster virus RNA sequences in human neurons and fibroblasts. J. Virol. 88, 5877– 5880.

Baker, K.D., and Thummel, C.S. (2007). Diabetic Larvae and Obese Flies-Emerging Studies of Metabolism in Drosophila. Cell Metab. 6, 257–266.

142

Bartlett, K., and Eaton, S. (β004). Mitochondrial -oxidation. Eur. J. Biochem. 271, 462– 469.

Baumbach, J., Hummel, P., Bickmeyer, I., Kowalczyk, K.M., Frank, M., Knorr, K., Hildebrandt, A., Riedel, D., Jäckle, H., and Kühnlein, R.P. (2014). A drosophila in vivo screen identifies store-operated calcium entry as a key regulator of adiposity. Cell Metab. 19, 331–343.

Bell, C.G., Walley, A.J., and Froguel, P. (2005). The genetics of human obesity. Nat. Rev. Genet. 6, 221–234.

Beller, M., Sztalryd, C., Southall, N., Bell, M., Jackle, H., Auld, D.S., and Oliver, B. (2008). COPI complex is a regulator of lipid homeostasis. PLoS Biol. 6, e292.

Beller, M., Bulankina, A. V, Hsiao, H.-H., Urlaub, H., Jäckle, H., and Kü Hnlein, R.P. (2010). PERILIPIN-Dependent Control of Lipid Droplet Structure and Fat Storage in Drosophila. Cell Metab. 12, 521–532.

Bernards, A., and Hariharan, I.K. (2001). Of flies and men--studying human disease in Drosophila. Curr. Opin. Genet. Dev. 11, 274–278.

Bezaire, V., Bruce, C.R., Heigenhauser, G.J.F., Tandon, N.N., Glatz, J.F.C., Luiken, J.J.J.F., Bonen, A., and Spriet, L.L. (2006). Identification of fatty acid translocase on human skeletal muscle mitochondrial membranes: essential role in fatty acid oxidation. Am. J. Physiol. Endocrinol. Metab. 290, E509-15.

Bi, J., Wang, W., Liu, Z., Huang, X., Jiang, Q., Liu, G., Wang, Y., and Huang, X. (2014). Seipin promotes adipose tissue fat storage through the ER Ca(2)(+)-ATPase SERCA. Cell Metab. 19, 861–871.

Bier, E. (2005). Drosophila, the golden bug, emerges as a tool for human genetics. Nat Rev Genet 6, 9–23.

Biron, P., Mongeau, J.G., and Bertrand, D. (1977). Familial resemblance of body weight and weight/height in 374 homes with adopted children. J. Pediatr. 91, 555–558.

Blauw, L.L., Noordam, R., Trompet, S., Berbee, J.F.P., Rosendaal, F.R., van Heemst, D., van Dijk, K.W., Mook-Kanamori, D.O., de Mutsert, R., and Rensen, P.C.N. (2017). Genetic variation in the obesity gene FTO is not associated with decreased fat oxidation: the NEO study. Int. J. Obes. 41, 1594–1600.

Bluher, M. (2008). Fat tissue and long life. Obes. Facts 1, 176–182.

Borrull, A., Lopez-Martinez, G., Poblet, M., Cordero-Otero, R., and Rozes, N. (2015). A simple method for the separation and quantification of neutral lipid species using GC-MS. Eur. J. Lipid Sci. Technol. 117, 274–280.

143

Bouchard, C. (2007). The biological predisposition to obesity: beyond the thrifty genotype scenario. Int. J. Obes. 31, 1337–1339.

Bouchard, C., Savard, R., Després, J., Tremblay, A., and Leblanc, C. (1985). Body Composition in Adopted and Biological Siblings. Hum. Biol. 57, 61–75.

Bradford, A.P., Jones, K., Kechris, K., Chosich, J., Montague, M., Warren, W.C., May, M.C., Al-Safi, Z., Kuokkanen, S., Appt, S.E., et al. (2015). Joint MiRNA/mRNA expression profiling reveals changes consistent with development of dysfunctional corpus luteum after weight gain. PLoS One 10, e0135163.

Britton, J.S., and Edgar, B. a (1998). Environmental control of the cell cycle in Drosophila: nutrition activates mitotic and endoreplicative cells by distinct mechanisms. Development 125, 2149–2158.

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, 213–221.

Buszczak, M., Lu, X., Segraves, W.A., Chang, T.Y., and Cooley, L. (2002). Mutations in the midway gene disrupt a Drosophila acyl coenzyme A: diacylglycerol acyltransferase. Genetics 160, 1511–1518.

Buszczak, M., Paterno, S., Lighthouse, D., Bachman, J., Planck, J., Owen, S., Skora, A.D., Nystul, T.G., Ohlstein, B., Allen, A., et al. (2007). The carnegie protein trap library: a versatile tool for Drosophila developmental studies. Genetics 175, 1505–1531.

Butterworth, F.M., Bodenstein, D., and King, R.C. (1965). Adipose Tissue of Drosophila melanogaster. I. An Experimental Study of Larval Fat Body. J. Exp. Zool. 158, 141–153.

Campos, A.R., Rosen, D.R., Robinow, S.N., and White, K. (1987). Molecular analysis of the locus elav in Drosophila melanogaster: a gene whose embryonic expression is neural specific. EMBO J. 6, 425–431.

Carrisi, C., Madeo, M., Morciano, P., Dolce, V., Cenci, G., Cappello, A.R., Mazzeo, G., Iacopetta, D., and Capobianco, L. (2008). Identification of the Drosophila melanogaster mitochondrial citrate carrier: bacterial expression, reconstitution, functional characterization and developmental distribution. J. Biochem. 144, 389–392.

Carvalho, M., Sampaio, J.L., Palm, W., Brankatschk, M., Eaton, S., and Shevchenko, A. (2012). Effects of diet and development on the Drosophila lipidome. Mol. Syst. Biol. 8, 600.

Castello, A., Hentze, M.W., and Preiss, T. (2015). Metabolic Enzymes Enjoying New Partnerships as RNA-Binding Proteins. Trends Endocrinol. Metab. 26, 746–757.

144

Cavalli, G., Justice, J.N., Boyle, K.E., D’Alessandro, A., Eisenmesser, E.Z., Herrera, J.J., Hansen, K.C., Nemkov, T., Stienstra, R., Garlanda, C., et al. (2017). Interleukin 37 reverses the metabolic cost of inflammation, increases oxidative respiration, and improves exercise tolerance. Proc. Natl. Acad. Sci. U. S. A. 114, 2313–2318.

Chang, J.L., Lin, H. V., Blauwkamp, T.A., and Cadigan, K.M. (2008). Spenito and Split ends act redundantly to promote Wingless signaling. Dev. Biol. 314, 100–111.

Chen, F., and Rebay, I. (2000). split ends, a new component of the Drosophila EGF receptor pathway, regulates development of midline glial cells. Curr. Biol. 10, 943–946.

Chen, A., Tang, Y., Davis, V., Hsu, F.-F., Kennedy, S.M., Song, H., Turk, J., Brunt, E.M., Newberry, E.P., and Davidson, N.O. (2013). Liver fatty acid binding protein (L-Fabp) modulates murine stellate cell activation and diet-induced nonalcoholic fatty liver disease. Hepatology 57, 2202–2212.

Chen, D., Qu, C., and Hewes, R.S. (2014). Neuronal remodeling during metamorphosis is regulated by the alan shepard (shep) gene in Drosophila melanogaster. Genetics 197, 1267– 1283.

Chien, S., Reiter, L.T., Bier, E., and Gribskov, M. (2002). Homophila: human disease gene cognates in Drosophila. Nucleic Acids Res. 30, 149–151.

Chu, A.Y., Workalemahu, T., Paynter, N.P., Rose, L.M., Giulianini, F., Tanaka, T., Ngwa, J.S., Qi, Q., Curhan, G.C., Rimm, E.B., et al. (2013). Novel locus including FGF21 is associated with dietary macronutrient intake. Hum. Mol. Genet. 22, 1895–1902.

Chu, C., Zhang, Q.C., Da Rocha, S.T., Flynn, R.A., Bharadwaj, M., Calabrese, J.M., Magnuson, T., Heard, E., and Chang, H.Y. (2015). Systematic discovery of Xist RNA binding proteins. Cell 161, 404–416.

Claussnitzer, M., Dankel, S.N., Kim, K.-H., Quon, G., Meuleman, W., Haugen, C., Glunk, V., Sousa, I.S., Beaudry, J.L., Puviindran, V., et al. (2015). FTO Obesity Variant Circuitry and Adipocyte Browning in Humans. N. Engl. J. Med. 373, 895–907.

Cole, J.B., Manyama, M., Kimwaga, E., Mathayo, J., Larson, J.R., Liberton, D.K., Lukowiak, K., Ferrara, T.M., Riccardi, S.L., Li, M., et al. (2016). Genomewide Association Study of African Children Identifies Association of SCHIP1 and PDE8A with Facial Size and Shape. PLoS Genet. 12, e1006174.

Coleman, D. (1973). Effects of Parabiosis of Obese with Diabetes and Normal Mice. Diabetologia 9, 294–298.

Coleman, D. (1978). Obese and Diabetes: Two Mutant Genes Causing Diabetes-Obesity Syndromes in Mice. Diabetologia 14, 141–148.

Coleman, D., and Hummel, K. (1973). The Influence of Genetic Background on the Expression of the Obese (Ob) Gene in the Mouse. Diabetologia 9, 287–293.

145

Comuzzie, A.G., and Allison, D.B. (1998). The Search for Human Obesity Genes. Science (80-. ). 280, 1374–1377.

Cornes, B.K., Zhu, G., and Martin, N.G. (2007). Sex differences in genetic variation in weight: A longitudinal study of body mass index in adolescent twins. Behav. Genet. 37, 648– 660.

Cox, J.E., Thummel, C.S., and Tennessen, J.M. (2017). Metabolomic Studies in Drosophila. Genetics 206, 1169–1185.

Day, F.R., and Loos, R.J.F. (2011). Developments in obesity genetics in the era of genome- wide association studies. J. Nutrigenet. Nutrigenomics 4, 222–238.

Dickson, B.J., Van Der Straten, A., Dominguez, M., and Hafen, E. (1996). Mutations modulating Raf signaling in Drosophila eye development. Genetics 142, 163–171.

Dietzl, G., Chen, D., Schnorrer, F., Su, K.-C., Barinova, Y., Fellner, M., Gasser, B., Kinsey, K., Oppel, S., Scheiblauer, S., et al. (2007). A genome-wide transgenic RNAi library for conditional gene inactivation in Drosophila. Nature 448, 151–156.

Dina, C., Meyre, D., Gallina, S., Durand, E., Körner, A., Jacobson, P., Carlsson, L.M.S., Kiess, W., Vatin, V., Lecoeur, C., et al. (2007). Variation in FTO contributes to childhood obesity and severe adult obesity. Nat. Genet. 39, 724–726.

Doroquez, D.B., Orr-Weaver, T.L., and Rebay, I. (2007). Split ends antagonizes the Notch and potentiates the EGFR signaling pathways during Drosophila eye development. Mech. Dev. 124, 792–806.

Duffy, J.B. (β00β). GAL4 system in Drosophila: a fly geneticist’s Swiss army knife. Genesis 34, 1–15.

Estelle, J., Perez-Enciso, M., Mercade, A., Varona, L., Alves, E., Sanchez, A., and Folch, J.M. (2006). Characterization of the porcine FABP5 gene and its association with the FAT1 QTL in an Iberian by Landrace cross. Anim. Genet. 37, 589–591.

Farooqi, I.S., Keogh, J.M., Yeo, G.S.H., Lank, E.J., Cheetham, T., and O’Rahilly, S. (β00γ). Clinical Spectrun of Obesity and Mutations in the Melanocortin 4 Receptor Gene. N. Engl. J. Med. 348, 1085–1095.

Farvid, M.S., Ng, T.W.K., Chan, D.C., Barrett, P.H.R., and Watts, G.F. (2006). Association of adiponectin and resistin with adipose tissue compartments, insulin resistance and dyslipidaemia. Diabetes. Obes. Metab. 7, 406–413.

Faust, I.M., Johnson, P.R., Stern, J.S., and Hirsch, J. (1978). Diet-induced adipocyte number increase in adult rats: a new model of obesity. Am. J. Physiol. 235, E279–E286.

146

Feng, Y., Bommer, G.T., Zhai, Y., Akyol, A., Hinoi, T., Winer, I., Lin, H. V., Cadigan, K.M., Cho, K.R., and Fearon, E.R. (2007). Drosophila split ends homologue SHARP functions as a positive regulator of Wnt/beta-catenin/T-cell factor signaling in neoplastic transformation. Cancer Res. 67, 482–491.

Finkelstein, E.A., Trogdon, J.G., Cohen, J.W., and Dietz, W. (2009). Annual medical spending attributable to obesity: payer-and service-specific estimates. Health Aff. (Millwood). 28, w822-31.

Firth, L., Manchester, J., Lorenzen, J.A., Baron, M., and Perkins, L.A. (2000). Identification of genomic regions that interact with a viable allele of the Drosophila protein tyrosine phosphatase Corkscrew. Genetics 156, 733–748.

Fortini, M.E., Skupski, M.P., Boguski, M.S., and Hariharan, I.K. (2000). A survey of human disease gene counterparts in the Drosophila genome. J. Cell Biol. 150, F23-30.

Frayling, T.M., Timpson, N.J., Weedon, M.N., Zeggini, E., Freathy, R.M., Lindgren, C.M., Perry, J.R.B., Elliott, K.S., Lango, H., Rayner, N.W., et al. (2007). A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity. Science 316, 889–894.

Freeman, A., Franciscovich, A., Bowers, M., Sandstrom, D.J., and Sanyal, S. (2011). NFAT regulates pre-synaptic development and activity-dependent plasticity in Drosophila. Mol. Cell. Neurosci. 46, 535–547.

Friedman, J.M., and Halaas, J.L. (1998). Leptin and the regulation of body weight in mammals. Nature 395, 763–770.

Friedman, J.M., and Leibel, R.L. (1992). Tackling a weighty problem. Cell 69, 217–220.

Fryar, C.D., Carroll, M.D., and Ogden, C.L. (2016a). Prevalence of Overweight, Obesity, and Extreme Obesity Among Adults Aged 20 and Over: United States, 1960-1962 Through 2013-2014.

Fryar, C.D., Carroll, M.D., and Ogden, C.L. (2016b). Prevalence of Overweight and Obesity Among Children and Adolescents Aged 2-19 Years: United States, 1963-1965 Through 2013-2014.

Fu, Y., and He, C. (2012). Nucleic acid modifications with epigenetic significance. Curr. Opin. Chem. Biol. 16, 516–524.

Fu, Y., Dominissini, D., Rechavi, G., and He, C. (2014). Gene expression regulation mediated through reversible m(6)A RNA methylation. Nat. Rev. Genet. 15, 293–306.

Fujimoto, M., Matsumoto, K.I., Iguchi-Ariga, S.M.M., and Ariga, H. (2001). Disruption of MSSP, c-myc single-strand binding protein, leads to embryonic lethality in some homozygous mice. Genes to Cells 6, 1067–1075.

147

Geller, F., Dempfle, A., and Gorg, T. (2003). Genome scan for body mass index and height in the Framingham Heart Study. BMC Genet. 4 Suppl 1, S91.

Gellon, G., Harding, K.W., McGinnis, N., Martin, M.M., and McGinnis, W. (1997). A genetic screen for modifiers of Deformed homeotic function identifies novel genes required for head development. Development 124, 3321–3331.

Gerstner, J.R., Vanderheyden, W.M., Shaw, P.J., Landry, C.F., and Yin, J.C.P. (2011). Fatty- acid binding proteins modulate sleep and enhance long-term memory consolidation in Drosophila. PLoS One 6, e15890.

Grayson, B.E., and Seeley, R.J. (2012). Deconstructing obesity: The face of fatness before and after the discovery of leptin. Diabetologia 55, 3–6.

Greenberg, A.J., Yanowitz, J.L., and Schedl, P. (2004). The Drosophila GAGA factor is required for dosage compensation in males and for the formation of the male-specific-lethal complex chromatin entry site at 12DE. Genetics 166, 279–289.

Greenspan, R. (2004). Fly Pushing: The Theory and Practice of Drosophila Genetics (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press).

Greenspan, P., and Fowler, S.D. (1985). Spectrofluorometric studies of the lipid probe, nile red. J. Lipid Res. 26, 781–789.

Grönke, S., Beller, M., Fellert, S., Ramakrishnan, H., Jäckle, H., and Kühnlein, R.P. (2003). Control of Fat Storage by a Drosophila PAT Domain Protein. Curr. Biol. 13, 603–606.

Grönke, S., Mildner, A., Fellert, S., Tennagels, N., Petry, S., Müller, G., Jäckle, H., and Kühnlein, R.P. (2005). Brummer lipase is an evolutionary conserved fat storage regulator in Drosophila. Cell Metab. 1, 323–330.

Grönke, S., Müller, G., Hirsch, J., Fellert, S., Andreou, A., Haase, T., Jäckle, H., and Kühnlein, R.P. (2007). Dual lipolytic control of body fat storage and mobilization in Drosophila. PLoS Biol. 5, 1248–1256.

Gu, T., Zhao, T., Kohli, U., and Hewes, R.S. (2017). The large and small SPEN family proteins stimulate axon outgrowth during neurosecretory cell remodeling in Drosophila. Dev. Biol. 0–1.

Guh, D.P., Zhang, W., Bansback, N., Amarsi, Z., Birmingham, C.L., and Anis, A.H. (2009). The incidence of co-morbidities related to obesity and overweight: A systematic review and meta-analysis. BMC Public Health 9, 88.

Guo, Y., Walther, T.C., Rao, M., Stuurman, N., Goshima, G., Terayama, K., Wong, J.S., Vale, R.D., Walter, P., and Farese Jr, R. V (2008). Functional genomic screen reveals genes involved in lipid-droplet formation and utilization Eukaryotic cells store neutral lipids in cytoplasmic lipid droplets. Nature 453, 657–661.

148

Gutierrez, E., Wiggins, D., Fielding, B., and Gould, A.P. (2007). Specialized hepatocyte-like cells regulate Drosophila lipid metabolism. Nature 445, 275–280.

Hader, T., Muller, S., Aguilera, M., Eulenberg, K.G., Steuernagel, A., Ciossek, T., Kuhnlein, R.P., Lemaire, L., Fritsch, R., Dohrmann, C., et al. (2003). Control of triglyceride storage by a WD40/TPR-domain protein. EMBO Rep. 4, 511–516.

Hainer, V., Zamrazilová, H., Spálová, J., Hainerová, I., Kunešová, M., Aldhoon, B., and Bendlová, B. (2008). Role of hereditary factors in weight loss and its maintenance. Physiol. Res. 57, S1–S15.

Harbison, S.T., and Sehgal, A. (2008). Quantitative genetic analysis of sleep in Drosophila melanogaster. Genetics 178, 2341–2360.

Hartz, A., Giefer, E., and Rimm, A.A. (1977). Relative importance of the effect of family environment and heredity on obesity. Ann. Hum. Genet. 41, 185–193.

Hazegh, K.E., and Reis, T. (2016). A Buoyancy-based Method of Determining Fat Levels in Drosophila. J. Vis. Med. 117, e54744.

Hazegh, K.E., Nemkov, T., D’Alessandro, A., Diller, J.D., Monks, J., Mcmanaman, J.L., Jones, K.L., Hansen, K.C., and Reis, T. (2017). An autonomous metabolic role for Spen. PLoS Genet. 13, e1006859.

Henderson, H.H., Timberlake, K.B., Austin, Z.A., Badani, H., Sanford, B., Tremblay, K., Baird, N.L., Jones, K., Rovnak, J., Frietze, S., et al. (2015). Occupancy of RNA Polymerase II Phosphorylated on Serine 5 (RNAP S5P) and RNAP S2P on Varicella-Zoster Virus Genes 9, 51, and 66 Is Independent of Transcript Abundance and Polymerase Location within the Gene. J. Virol. 90, 1231–1243.

Hildebrandt, A., Bickmeyer, I., and Kühnlein, R.P. (2011). Reliable Drosophila body fat quantification by a coupled colorimetric assay. PLoS One 6, e23796.

Hiriart, E., Gruffat, H., Buisson, M., Mikaelian, I., Keppler, S., Meresse, P., Mercher, T., Bernard, O.A., Sergeant, A., and Manet, E. (2005). Interaction of the Epstein-Barr virus mRNA export factor EB2 with human Spen proteins SHARP, OTT1, and a novel member of the family, OTT3, links Spen proteins with splicing regulation and mRNA export. J. Biol. Chem. 280, 36935–36945.

Hogan, P.G. (2017). Calcium-NFAT transcriptional signalling in T cell activation and T cell exhaustion. Cell Calcium 63, 66–69.

Hogan, P.G., Chen, L., Nardone, J., and Rao, A. (2003). Transcriptional regulation by calcium, calcineurin, and NFAT. Genes Dev. 17, 2205–2232.

149

Holland, W.L., Brozinick, J.T., Wang, L.-P., Hawkins, E.D., Sargent, K.M., Liu, Y., Narra, K., Hoehn, K.L., Knotts, T.A., Siesky, A., et al. (2007). Inhibition of ceramide synthesis ameliorates glucocorticoid-, saturated-fat-, and obesity-induced insulin resistance. Cell Metab. 5, 167–179.

Holm, C. (2003). Molecular mechanisms regulating hormone-sensitive lipase and lipolysis. Biochem. Soc. Trans. 31, 1120–1124.

Hu, M., Yang, Y., Ji, Z., and Luo, J. (2016). RBM15 Functions in Blood Diseases. Curr. Cancer Drug Targets 16, 579–585.

Huffman, J.E., Albrecht, E., Teumer, A., Mangino, M., Kapur, K., Johnson, T., Kutalik, Z., Pirastu, N., Pistis, G., Lopez, L.M., et al. (2015). Modulation of genetic associations with serum urate levels by body-mass-index in humans. PLoS One 10, e0119752.

Hur, Y.-M. (2007). Sex Difference in Heritability of BMI in South Korean Adolescent Twins. Obesity 15, 2908–2911.

Iida, M., Taira, T., Ariga, H., and Iguchi-Ariga, S.M. (1997). Induction of apoptosis in HeLa cells by MSSP, c-myc binding proteins. Biol. Pharm. Bull. 20, 10–14.

Itabe, H., Yamaguchi, T., Nimura, S., and Sasabe, N. (2017). Perilipins: a diversity of intracellular lipid droplet proteins. Lipids Health Dis. 16, 83.

Jagannadham, J., Jaiswal, H.K., Agrawal, S., and Rawal, K. (2016). Comprehensive Map of Molecules Implicated in Obesity. PLoS One 11, e0146759.

Jemc, J., and Rebay, I. (2006). Characterization of the split ends-like gene spenito reveals functional antagonism between SPOC family members during drosophila eye development. Genetics 173, 279–286.

Jin, L.H., Choi, J.K., Kim, B., Cho, H.S., Kim, J., Kim-Ha, J., and Kim, Y.-J. (2009). Requirement of Split ends for epigenetic regulation of Notch signal-dependent genes during infection-induced hemocyte differentiation. Mol. Cell. Biol. 29, 1515–1525.

Kaess, B.M., Enserro, D.M., McManus, D.D., Xanthakis, V., Chen, M.-H., Sullivan, L.M., Ingram, C., O’Donnell, C.J., Keaney, J.F., Vasan, R.S., et al. (β01β). Cardiometabolic correlates and heritability of fetuin-A, retinol-binding protein 4, and fatty-acid binding protein 4 in the Framingham Heart Study. J. Clin. Endocrinol. Metab. 97, E1943-7.

Kanaoka, Y., and Nojima, H. (1994). SCR: Novel human suppressors of cdc2/cdc13 mutants of Schizosaccharomyces pombe harbour motifs for RNA binding proteins. Nucleic Acids Res. 22, 2687–2693.

Kelso, R.J., Buszczak, M., Quinones, A.T., Castiblanco, C., Mazzalupo, S., and Cooley, L. (2004). Flytrap, a database documenting a GFP protein-trap insertion screen in Drosophila melanogaster. Nucleic Acids Res. 32, D418-20.

150

Kersten, S. (2001). Mechanisms of nutritional and hormonal regulation of lipogenesis. EMBO Rep. 2, 282–286.

Keyser, P., Borge-Renberg, K., and Hultmark, D. (2007). The Drosophila NFAT homolog is involved in salt stress tolerance. Insect Biochem. Mol. Biol. 37, 356–362.

King-Jones, K., and Thummel, C.S. (2005). Nuclear receptors - A perspective from Drosophila. Nat. Rev. Genet. 6, 311–323.

Kolkhof, P., Werthebach, M., van de Venn, A., Poschmann, G., Chen, L., Welte, M., Stuhler, K., and Beller, M. (2017). A Luciferase-fragment Complementation Assay to Detect Lipid Droplet-associated Protein-Protein Interactions. Mol. Cell. Proteomics 16, 329–345.

Kolodziej, P.A., Yeh Jan, L., and Nung Jan, Y. (1995). Mutations that affect the length, fasciculation, or ventral orientation of specific sensory axons in the Drosophila embryo. Neuron 15, 273–286.

Kuang, B., Wu, S.C., Shin, Y., Luo, L., and Kolodziej, P. (2000). split ends encodes large nuclear proteins that regulate neuronal cell fate and axon extension in the Drosophila embryo. Development 127, 1517–1529.

Kühnlein, R.P. (2012). Lipid droplet-based storage fat metabolism in Drosophila. J. Lipid Res. 53, 28–33.

Kuroda, K., Han, H., Tani, S., Tanigaki, K., Tun, T., Furukawa, T., Taniguchi, Y., Kurooka, H., Hamada, Y., Toyokuni, S., et al. (2003). Regulation of Marginal Zone B Cell Development by MINT , a Suppressor of Notch / RBP-J Signaling Pathway. Immunity 18, 301–312.

Lajunen, H.-R., Kaprio, J., Keski-Rahkonen, A., Rose, R.J., Pulkkinen, L., Rissanen, A., and Silventoinen, K. (2009). Genetic and environmental effects on body mass index during adolescence: a prospective study among Finnish twins. Int. J. Obes. (Lond). 33, 559–567.

Lakka, T.A., and Bouchard, C. (2005). Physical activity, obesity and cardiovascular diseases. Handb Exp Pharmacol 170, 137–163.

Lane, M.E., Elend, M., Heidmann, D., Herr, A., Marzodko, S., Herzig, A., and Lehner, C.F. (2000). A screen for modifiers of cyclin E function in Drosophila melanogaster identifies Cdk2 mutations, revealing the insignificance of putative phosphorylation sites in Cdk2. Genetics 155, 233–244.

Lee, G., and Park, J.H. (2004). Hemolymph sugar homeostasis and starvation-induced hyperactivity affected by genetic manipulations of the adipokinetic hormone-encoding gene in Drosophila melanogaster. Genetics 167, 311–323.

Lee, H., Peng, Y., and Guo, Y. (2013). Analysis of lipid droplet dynamics and functions in Drosophila melanogaster. Methods Cell Biol. 116, 53–69.

151

Lee, Y.S., Nakahara, K., Pham, J.W., Kim, K., He, Z., Sontheimer, E.J., and Carthew, R.W. (2004). Distinct roles for Drosophila Dicer-1 and Dicer-2 in the siRNA/miRNA silencing pathways. Cell 117, 69–81.

Legare, S., Cavallone, L., Mamo, A., Chabot, C., Sirois, I., Magliocco, A., Klimowicz, A., Tonin, P.N., Buchanan, M., Keilty, D., et al. (2015). The estrogen receptor cofactor SPEN functions as a tumor suppressor and candidate biomarker of drug responsiveness in hormone- dependent breast cancers. Cancer Res. 75, 4351–4363.

Lemmon, M.A., and Schlessinger, J. (2010). Cell signaling by receptor-tyrosine kinases. Cell 141, 1117–1134.

Lence, T., Akhtar, J., Bayer, M., Schmid, K., Spindler, L., Ho, C.H., and Kreim, N. (2016). m 6 A modulates neuronal functions and sex determination in Drosophila. Nat. Publ. Gr. 540, 242–247.

Leopold, P., and Perrimon, N. (2007). Drosophila and the genetics of the internal milieu. Nature 450, 186–188.

Leto, D., and Saltiel, A.R. (2012). Regulation of glucose transport by insulin: traffic control of GLUT4. Nat Rev Mol Cell Biol 13, 383–396.

Li, J., Li, J., Yang, X., Qin, H., Zhou, P., Liang, Y., and Han, H. (2005). The C terminus of MINT forms homodimers and abrogates MINT-mediated transcriptional repression. Biochim. Biophys. Acta - Gene Struct. Expr. 1729, 50–56.

Licinio, J., Caglayan, S., Ozata, M., Yildiz, B.O., de Miranda, P.B., O’Kirwan, F., Whitby, R., Liang, L., Cohen, P., Bhasin, S., et al. (2004). Phenotypic effects of leptin replacement on morbid obesity, diabetes mellitus, hypogonadism, and behavior in leptin-deficient adults. Proc. Natl. Acad. Sci. U. S. A. 101, 4531–4536.

Lin, H. V, Doroquez, D.B., Cho, S., Chen, F., Rebay, I., and Cadigan, K.M. (2003). Splits ends is a tissue/promoter specific regulator of Wingless signaling. Development 130, 3125– 3135.

Lindtner, S., Zolotukhin, A.S., Uranishi, H., Bear, J., Kulkarni, V., Smulevitch, S., Samiotaki, M., Panayotou, G., Felber, B.K., and Pavlakis, G.N. (2006). RNA-binding motif protein 15 binds to the RNA transport element RTE and provides a direct link to the NXF1 export pathway. J. Biol. Chem. 281, 36915–36928.

Liu, Z., and Huang, X. (β01γ). Lipid metabolism in Drosophila : development and disease Using Drosophila System to Study Lipid Metabolism Lipids Function in Drosophila Early Development. Acta Biochim Biophys Sin 45, 44–50.

Loos, R.J.F. (2009). Recent progress in the genetics of common obesity. Br. J. Clin. Pharmacol. 68, 811–829.

152

Loos, R.J.F., and Bouchard, C. (2003). Obesity--is it a genetic disorder? J. Intern. Med. 254, 401–425.

Loos, R.J.F., and Bouchard, C. (2008). FTO: the first gene contributing to common forms of human obesity. Obes. Rev. 9, 246–250.

Loos, R.J.F., and Yeo, G.S.H. (2014). The bigger picture of FTO: the first GWAS-identified obesity gene. Nat. Rev. Endocrinol. 10, 51–61.

Lopez-Rodriguez, C., Antos, C.L., Shelton, J.M., Richardson, J.A., Lin, F., Novobrantseva, T.I., Bronson, R.T., Igarashi, P., Rao, A., and Olson, E.N. (2004). Loss of NFAT5 results in renal atrophy and lack of tonicity-responsive gene expression. Proc. Natl. Acad. Sci. U. S. A. 101, 2392–2397.

Luan, H., Lemon, W.C., Peabody, N.C., Pohl, J.B., Zelensky, P.K., Wang, D., Nitabach, M.N., Holmes, T.C., and White, B.H. (2006). Functional Dissection of a Neuronal Network Required for Cuticle Tanning and Wing Expansion in Drosophila. J. Neurosci. 26, 573–584.

Ludewig, A.H., Kober-eisermann, C., Weitzel, C., Bethke, A., Neubert, K., Gerisch, B., Hutter, H., and Antebi, A. (2004). A novel nuclear receptor/coregulator complex controls C. elegans lipid metabolism, larval development, and aging. Genes Dev. 2120–2133.

Luong, N., Davies, C.R., Wessells, R.J., Graham, S.M., King, M.T., Veech, R., Bodmer, R., and Oldham, S.M. (2006). Activated FOXO-mediated insulin resistance is blocked by reduction of TOR activity. Cell Metab. 4, 133–142.

Ma, X., Renda, M.J., Wang, L., Cheng, E.-C., Niu, C., Morris, S.W., Chi, A.S., and Krause, D.S. (2007). Rbm15 modulates Notch-induced transcriptional activation and affects myeloid differentiation. Mol. Cell. Biol. 27, 3056–3064.

MacDonald, A., and Stunkard, A.J. (1990). Body-Mass Indexes of British Separated Twins. N. Engl. J. Med. 322, 1530.

Maeda, K., Cao, H., Kono, K., Gorgun, C.Z., Furuhashi, M., Uysal, K.T., Cao, Q., Atsumi, G., Malone, H., Krishnan, B., et al. (2005). Adipocyte/macrophage fatty acid binding proteins control integrated metabolic responses in obesity and diabetes. Cell Metab. 1, 107– 119.

Maes, H.H.M., Neale, M.C., and Eaves, L.J. (1997). Genetic and enviromental factors in relative body weight and human adiposity. Behav. Genet. 27, 325–351.

Majerciak, V., Lu, M., Li, X., and Zheng, Z.-M. (2014). Attenuation of the suppressive activity of cellular splicing factor SRSF3 by Kaposi sarcoma-associated herpesvirus ORF57 protein is required for RNA splicing. RNA 20, 1747–1758.

Matsuda, H., Yamada, T., Yoshida, M., and Nishimura, T. (2015). Flies without Trehalose. J. Biol. Chem. 290, 1244–1255.

153

Matzat, L.H., Dale, R.K., Moshkovich, N., and Lei, E.P. (2012). Tissue-Specific Regulation of Chromatin Insulator Function. PLoS Genet. 8, e1003069.

Maule, A.G., Halton, D.W., and Shaw, C. (1995). Neuropeptide F: a ubiquitous invertebrate neuromediator? Hydrobiologia 305, 297–303.

Maycotte, P., Jones, K.L., Goodall, M.L., Thorburn, J., and Thorburn, A. (2015). Autophagy Supports Breast Cancer Stem Cell Maintenance by Regulating IL6 Secretion. Mol. Cancer Res. 13, 651–658.

McHugh, C.A., Chen, C.-K., Chow, A., Surka, C.F., Tran, C., McDonel, P., Pandya-Jones, A., Blanco, M., Burghard, C., Moradian, A., et al. (2015). The Xist lncRNA interacts directly with SHARP to silence transcription through HDAC3. Nature 521, 232–236.

Merkestein, M., Laber, S., McMurray, F., Andrew, D., Sachse, G., Sanderson, J., Li, M., Usher, S., Sellayah, D., Ashcroft, F.M., et al. (2015). FTO influences adipogenesis by regulating mitotic clonal expansion. Nat. Commun. 6, 6792.

Meyer, K.D., and Jaffrey, S.R. (2014). The dynamic epitranscriptome: N6-methyladenosine and gene expression control. Nat. Rev. Mol. Cell Biol. 15, 313–326.

Mi, H., Huang, X., Muruganujan, A., Tang, H., Mills, C., Kang, D., and Thomas, P.D. (2017). PANTHER version 11: expanded annotation data from Gene Ontology and Reactome pathways, and data analysis tool enhancements. Nucleic Acids Res. 45, D183–D189.

Ming, J.E., Blagowidow, N., Knoll, J.H.M., Rollings, L., Fortina, P., McDonald-McGinn, D.M., Spinner, N.B., and Zackai, E.H. (2000). Submicroscopic deletion in cousins with Prader-Willi syndrome causes a grandmatrilineal inheritance pattern: Effects of imprinting. Am. J. Med. Genet. 92, 19–24.

Miyakawa, H., Woo, S.K., Dahl, S.C., Handler, J.S., and Kwon, H.M. (1999). Tonicity- responsive enhancer binding protein, a rel-like protein that stimulates transcription in response to hypertonicity. Proc. Natl. Acad. Sci. U. S. A. 96, 2538–2542.

Mobin, M.B., Gerstberger, S., Teupser, D., Campana, B., Charisse, K., Heim, M.H., Manoharan, M., Tuschl, T., and Stoffel, M. (2016). The RNA-binding protein vigilin regulates VLDL secretion through modulation of Apob mRNA translation. Nat. Commun. 7, 12848.

Moindrot, B., Cerase, A., Coker, H., Masui, O., Grijzenhout, A., Pintacuda, G., Schermelleh, L., Nesterova, T.B., and Brockdorff, N. (2015). A Pooled shRNA Screen Identifies Rbm15, Spen, and Wtap as Factors Required for Xist RNA-Mediated Silencing. Cell Rep. 12, 562– 572.

Mollah, M.B.R., and Ishikawa, A. (2010). A wild derived quantitative trait locus on mouse chromosome 2 prevents obesity. BMC Genet. 11, 84.

154

Monfort, A., Di Minin, G., Postlmayr, A., Freimann, R., Arieti, F., Thore, S., and Wutz, A. (2015). Identification of Spen as a crucial factor for Xist function through forward genetic screening in haploid embryonic stem cells. Cell Rep. 12, 554–561.

Morales, P.E., Bucarey, J.L., and Espinosa, A. (2017). Muscle Lipid Metabolism: Role of Lipid Droplets and Perilipins. J. Diabetes Res. 2017, 1–10.

Moraru, A., Cakan-Akdogan, G., Strassburger, K., Males, M., Mueller, S., Jabs, M., Muelleder, M., Frejno, M., Braeckman, B.P., Ralser, M., et al. (2017). THADA Regulates the Organismal Balance between Energy Storage and Heat Production. Dev. Cell 41, 72–81.e6.

Morin, X., Daneman, R., Zavortink, M., and Chia, W. (2001). A protein trap strategy to detect GFP-tagged proteins expressed from their endogenous loci in Drosophila. Proc. Natl. Acad. Sci. U. S. A. 98, 15050–15055.

Morton, G.J., Cummings, D.E., Baskin, D.G., Barsh, G.S., and Schwartz, M.W. (2006). Central nervous system control of food intake and body weight. Nat. Rev. 443, 289–295.

Mosher, J., Zhang, W., Blumhagen, R.Z., D ’alessandro, A., Nemkov, T., Hansen, K.C., Hesselberth, J.R., and Reis, T. (2015). Coordination between Drosophila Arc1 and a specific population of brain neurons regulates organismal fat. Dev. Biol. 405, 280–290.

Murthy, M.S., and Pande, S. V (1987). Malonyl-CoA binding site and the overt carnitine palmitoyltransferase activity reside on the opposite sides of the outer mitochondrial membrane. Proc. Natl. Acad. Sci. U. S. A. 84, 378–382.

Nassel, D.R., and Winther, A.M.E. (2010). Drosophila neuropeptides in regulation of physiology and behavior. Prog. Neurobiol. 92, 42–104.

Neel, J. V (196β). Diabetes Mellitus: A “Thrifty Genotypes Rendered Detrimental by ”Progress"? Am. J. Hum. Genet. 14, 353–362.

Neel, J. V (1999). The “Thrifty Genotype” in 1998. Nutr. Rev. 57, 2–9.

Negishi, Y., Nishita, Y., Saegusa, Y., Kakizaki, I., Galli, I., Kihara, F., Tamai, K., Miyajima, N., Iguchi-Ariga, S.M., and Ariga, H. (1994). Identification and cDNA cloning of single- stranded DNA binding proteins that interact with the region upstream of the human c-myc gene. Oncogene 9, 1133–1143.

Nemkov, T., D’Alessandro, A., and Hansen, K.C. (β015). Three-minute method for amino acid analysis by UHPLC and high-resolution quadrupole orbitrap mass spectrometry. Amino Acids 47, 2345–2357.

Nemkov, T., Hansen, K.C., and D’Alessandro, A. (β017). A three-minute method for high- throughput quantitative metabolomics and quantitative tracing experiments of central carbon and nitrogen pathways. Rapid Commun. Mass Spectrom. 31, 663–673.

155

Neufeld, T.P. (2012). Autophagy and cell growth--the yin and yang of nutrient responses. J. Cell Sci. 125, 2359–2368.

Neufeld, T.P., Flor, A., De La Cruz, A., Johnston, L.A., and Edgar, B.A. (1998). Coordination of Growth and Cell Division in the Drosophila Wing. Cell 93, 1183–1193.

Newberry, E.P., Latifi, T., and Towler, D.A. (1999). The RRM domain of MINT, a novel Msx2 binding protein, recognizes and regulates the rat osteocalcin promoter. Biochemistry 38, 10678–10690.

Newberry, E.P., Xie, Y., Kennedy, S., Han, X., Buhman, K.K., Luo, J., Gross, R.W., and Davidson, N.O. (2003). Decreased hepatic triglyceride accumulation and altered fatty acid uptake in mice with deletion of the liver fatty acid-binding protein gene. J. Biol. Chem. 278, 51664–51672.

Niki, T., Izumi, S., Saëgusa, Y., Taira, T., Takai, T., Iguchi-Ariga, S.M., and Ariga, H. (2000). MSSP promotes ras/myc cooperative cell transforming activity by binding to c-Myc. Genes to Cells 5, 127–141.

Niu, Y., Zhao, X., Wu, Y.-S., Li, M.-M., Wang, X.-J., and Yang, Y.-G. (2013). N6-methyladenosine (m6A) in RNA: an old modification with a novel epigenetic function. Genomics. Proteomics Bioinformatics 11, 8–17.

O’Dea, K. (1995). Overview of the thrifty genotype hypothesis. Asia Pac. J. Clin. Nutr. 4, 339–340.

O’Rahilly, S., Farooqi, I.S., Yeo, G.S.H., and Challis, B.G. (β00γ). Minireview: human obesity-lessons from monogenic disorders. Endocrinology 144, 3757–3764.

OECD (2017). Obesity Update 2017.

Ohta, T., Gray, T.A., Rogan, P.K., Buiting, K., Gabriel, J.M., Saitoh, S., Muralidhar, B., Bilienska, B., Driscoll, D.J., Horsthemke, B., et al. (1999). Imprinting-Mutation Mechanisms in Prader-Willi Syndrome. Am. J. Hum. Genet. 64, 397–413.

Oswald, F. (2002). SHARP is a novel component of the Notch/RBP-Jkappa signalling pathway. EMBO J. 21, 5417–5426.

Oswald, F., Winkler, M., Cao, Y., Astrahantseff, K., Bourteele, S., and Borggrefe, T. (2005). RBP-J/SHARP Recruits CtIP/CtBP Corepressors To Silence Notch Target Genes. Mol. Cell. Biol. 25, 10379–10390.

Padmanabha, D., and Baker, K.D. (2014). Drosophila gains traction as a repurposed tool to investigate metabolism. Trends Endocrinol. Metab. 25, 518–527.

156

Pagac, M., Cooper, D.E., Qi, Y., Lukmantara, I.E., Mak, H.Y., Wu, Z., Tian, Y., Liu, Z., Lei, M., Du, X., et al. (2016). SEIPIN Regulates Lipid Droplet Expansion and Adipocyte Development by Modulating the Activity of Glycerol-3-phosphate Acyltransferase. Cell Rep. 17, 1546–1559.

Palanker, L., Tennessen, J.M., Lam, G., and Thummel, C.S. (2009). Drosophila HNF4 Regulates Lipid Mobilization and -Oxidation. Cell Metab. 9, 228–239.

Pan, T. (2013). N6-methyl-adenosine modification in messenger and long non-coding RNA. Trends Biochem. Sci. 38, 204–209.

Pande, S. V (1975). A mitochondrial carnitine acylcarnitine translocase system. Proc. Natl. Acad. Sci. U. S. A. 72, 883–887.

Park, S., Alfa, R.W., Topper, S.M., Kim, G.E.S., Kockel, L., and Kim, S.K. (2014). A genetic strategy to measure circulating Drosophila insulin reveals genes regulating insulin production and secretion. PLoS Genet. 10, e1004555.

Patil, D.P., Chen, C.-K., Pickering, B.F., Chow, A., Jackson, C., Guttman, M., and Jaffrey, S.R. (2016). m6A RNA methylation promotes XIST-mediated transcriptional repression. Nature 537, 1–25.

Perez, C.L., and Van Gilst, M.R. (2008). A 13C Isotope Labeling Strategy Reveals the Influence of Insulin Signaling on Lipogenesis in C. elegans. Cell Metab. 8, 266–274.

Pignoni, F., and Zipursky, S.L. (1997). Induction of Drosophila eye development by decapentaplegic. Development 124, 271–278.

Pijl, H. (2011). Obesity: Evolution of a symptom of Affuence. How food has shaped our existence. Neth. J. Med. 69, 159–166.

Pospisilik, J.A., Schramek, D., Schnidar, H., Cronin, S.J.F., Nehme, N.T., Zhang, X., Knauf, C., Cani, P.D., Aumayr, K., Todoric, J., et al. (2010). Drosophila Genome-wide Obesity Screen Reveals Hedgehog as a Determinant of Brown versus White Adipose Cell Fate. Cell 140, 148–160.

Price, R.A., and Gottesman, I.I. (1991). Body fat in identical twins reared apart: roles for genes and environment. Behav. Genet. 21, 1–7.

Price, J. V, Savenye, E.D., Lum, D., and Breitkreutz, A. (1997). Dominant Enhancers of Egfr in Drosophila melanogaster: Genetic Links Between theNotch and Egfr Signaling Pathways. Genetics 147, 1139–1153.

Prokopenko, S.N., He, Y., Lu, Y., and Bellen, H.J. (2000). Mutations affecting the development of the peripheral nervous system in Drosophila: a molecular screen for novel proteins. Genetics 156, 1691–1715.

157

Qi, L., Shen, H., Larson, I., Schaefer, E.J., Greenberg, A.S., Tregouet, D.A., Corella, D., and Ordovas, J.M. (2004). Gender-specific association of a perilipin gene haplotype with obesity risk in a white population. Obes. Res. 12, 1758–1765.

Qi, L., Tai, E.S., Tan, C.E., Shen, H., Chew, S.K., Greenberg, A.S., Corella, D., and Ordovas, J.M. (2005). Intragenic linkage disequilibrium structure of the human perilipin gene (PLIN) and haplotype association with increased obesity risk in a multiethnic Asian population. J. Mol. Med. (Berl). 83, 448–456.

Querenet, M., Goubard, V., Chatelain, G., Davoust, N., and Mollereau, B. (2015). Spen is required for pigment cell survival during pupal development in Drosophila. Dev. Biol. 402, 208–215.

Quinones-Coello, A.T., Petrella, L.N., Ayers, K., Melillo, A., Mazzalupo, S., Hudson, A.M., Wang, S., Castiblanco, C., Buszczak, M., Hoskins, R.A., et al. (2007). Exploring strategies for protein trapping in Drosophila. Genetics 175, 1089–1104.

Raffel, G.D., Mercher, T., Shigematsu, H., Williams, I.R., Cullen, D.E., Akashi, K., Bernard, O.A., and Gilliland, D.G. (2007). Ott1(Rbm15) has pleiotropic roles in Hematopoietic Development. PNAS 104, 6001–6006.

Rajan, A., and Perrimon, N. (2012). Drosophila cytokine unpaired 2 regulates physiological homeostasis by remotely controlling insulin secretion. Cell 151, 123–137.

Rankinen, T., Zuberi, A., Chagnon, Y.C., Weisnagel, S.J., Argyropoulos, G., Walts, B., Pérusse, L., and Bouchard, C. (2006). The human obesity gene map: the 2005 update. Obesity 14, 529–644.

Rebay, I., Chen, F., Hsiao, F., Kolodziej, P.A., Kuang, B.H., Laverty, T., Suh, C., Voas, M., Williams, A., and Rubin, G.M. (2000). A Genetic Screen for Novel Components of the Ras/Mitogen-Activated Protein Kinase Signaling Pathway That Interact With the yan Gene of Drosophila Identifies split ends, a New RNA Recognition Motif-Containing Protein. Genetics 154, 695–712.

Reed, D.R., Ding, Y., Xu, W., Cather, C., and Price, R.A. (1995). Human obesity does not segregate with the chromosomal regions of Prader-Willi, Bardet-Biedl, Cohen, Borjeson or Wilson-Turner syndromes. Int. J. Obes. Relat. Metab. Disord. 19, 599–603.

Reis, T., van Gilst, M.R., and Hariharan, I.K. (2010). A buoyancy-based screen of drosophila larvae for fat- storage mutants reveals a role for Sir2 in coupling fat Storage to Nutrient Availability. PLoS Genet. 6, e1001206.

Reiter, L.T., Potocki, L., Chien, S., Gribskov, M., and Bier, E. (2001). A Systematic Analysis of Human Disease-Associated Gene Sequences In Drosophila melanogaster. Genome Res. 11, 1114–1125.

Ren, D., Navarro, B., Xu, H., Yue, L., Shi, Q., and Clapham, D.E. (2001). A prokaryotic voltage-gated sodium channel. Science 294, 2372–2375.

158

Roth, A., and Diederichs, S. (2015). Rap and chirp about X inactivation. Nature 521, 170– 171.

Ruden, D.M., and Lu, X. (2006). Evolutionary conservation of metabolism explains howDrosophila nutrigenomics can help us understand human nutrigenomics. Genes Nutr. 1, 75–83.

Rulifson, E.J., Kim, S.K., and Nusse, R. (2002). Ablation of insulin-producing neurons in flies: growth and diabetic phenotypes. Science 296, 1118–1120.

Ryner, L.C., Goodwin, S.F., Castrillon, D.H., Anand, A., Villella, A., Baker, B.S., Hall, J.C., Taylor, B.J., and Wasserman, S.A. (1996). Control of male sexual behavior and sexual orientation in Drosophila by the fruitless gene. Cell 87, 1079–1089.

Sajuthi, S.P., Sharma, N.K., Chou, J.W., Palmer, N.D., McWilliams, D.R., Beal, J., Comeau, M.E., Ma, L., Calles-Escandon, J., Demons, J., et al. (2016). Mapping adipose and muscle tissue expression quantitative trait loci in African Americans to identify genes for type 2 diabetes and obesity. Hum. Genet. 135, 869–880.

Samuel, V.T., and Shulman, G.I. (2012). Integrating Mechanisms for Insulin Resistance: Common Threads and Missing Links. Cell 148, 852–871.

Sánchez-Pulido, L., Rojas, A.M., van Wely, K.H., Martinez-A, C., and Valencia, A. (2004). SPOC: a widely distributed domain associated with cancer, apoptosis and transcription. BMC Bioinformatics 5, 91.

Schachtner, L.T., Sola, I.E., Forand, D., Antonacci, S., Postovit, A.J., Mortimer, N.T., Killian, D.J., and Olesnicky, E.C. (2015). Drosophila Shep and C. elegans SUP-26 are RNA- binding proteins that play diverse roles in nervous system development. Dev. Genes Evol. 225, 319–330.

Schlegel, A., and Stainier, D.Y.R. (β007). Lessons from Lower’’ Organisms: What Worms, Flies, and Zebrafish Can Teach Us about Human Energy Metabolism. PLoS Genet. 3, 2037– 2048.

Schweitzer, B., and Shilo, B.-Z. (1997). A thousand and one roles for the Drosophila EGF receptor. Trends Genet. 13, 191–196.

Scott, R.C., Schuldiner, O., and Neufeld, T.P. (2004). Role and regulation of starvation- induced autophagy in the Drosophila fat body. Dev. Cell 7, 167–178.

Scuteri, A., Sanna, S., Chen, W.-M., Uda, M., Albai, G., Strait, J., Najjar, S., Nagaraja, R., Orru, M., Usala, G., et al. (2007). Genome-wide association scan shows genetic variants in the FTO gene are associated with obesity-related traits. PLoS Genet. 3, e115.

Shi, Y., Downes, M., Xie, W., Kao, H.Y., Ordentlich, P., Tsai, C.C., Hon, M., and Evans, R.M. (2001). Sharp, an inducible cofactor that integrates nuclear receptor repression and activation. Genes Dev. 15, 1140–1151.

159

Shi, Y., Hon, M., and Evans, R.M. (2002). The peroxisome proliferator-activated receptor delta, an integrator of transcriptional repression and nuclear receptor signaling. Proc. Natl. Acad. Sci. U. S. A. 99, 2613–2618.

Shui, G., Guan, X.L., Low, C.P., Chua, G.H., Goh, J.S.Y., Yang, H., Wenk, M.R., Welte, M.A., Guan, X.L., Souza, C.M., et al. (2010). Toward one step analysis of cellular lipidomes using liquid chromatography coupled with mass spectrometry: application to Saccharomyces cerevisiae and Schizosaccharomyces pombe lipidomics. Mol. Biosyst. 6, 1008.

Sierra, O.L., Cheng, S.L., Loewy, A.P., Charlton-Kachigian, N., and Towler, D.A. (2004). MINT, the Msx2 interacting nuclear matrix target, enhances Runx2-dependent activation of the osteocalcin fibroblast growth factor response element. J. Biol. Chem. 279, 32913–32923.

Silventoinen, K., Rokholm, B., Kaprio, J., and Sorensen, T.I. (2010). The genetic and environmental influences on childhood obesity: a systematic review of twin and adoption studies. Int J Obes 34, 29–40.

Silventoinen, K., Jelenkovic, A., Sund, R., Hur, Y.-M., Yokoyama, Y., Honda, C., Hjelmborg, J. vB, Moller, S., Ooki, S., Aaltonen, S., et al. (2016). Genetic and environmental effects on body mass index from infancy to the onset of adulthood: an individual-based pooled analysis of 45 twin cohorts participating in the COllaborative project of Development of Anthropometrical measures in Twins (CODATwins). Am. J. Clin. Nutr. 104, 371–379.

Soenen, S., Mariman, E.C.M., Vogels, N., Bouwman, F.G., den Hoed, M., Brown, L., and Westerterp-Plantenga, M.S. (2009). Relationship between perilipin gene polymorphisms and body weight and body composition during weight loss and weight maintenance. Physiol. Behav. 96, 723–728.

Sorensen, T.I., Holst, C., and Stunkard, A.J. (1992). Childhood body mass index--genetic and familial environmental influences assessed in a longitudinal adoption study. Int. J. Obes. Relat. Metab. Disord. 16, 705–714.

Soudani, N., Ghantous, C.M., Farhat, Z., Shebaby, W.N., Zibara, K., and Zeidan, A. (2016). Calcineurin/NFAT Activation-Dependence of Leptin Synthesis and Vascular Growth in Response to Mechanical Stretch. Front. Physiol. 7, 433.

Speakman, J.R. (2006). Thrifty genes for obesity and the metabolic syndrome — time to call off the search? Diabetes Vasc. Dis. Res. 3, 7–11.

Speakman, J.R. (2008). Thrifty genes for obesity, an attractive but flawed idea, and an alternative perspective: the “drifty gene” hypothesis. Int. J. Obes. 32, 1611–1617.

Staehling-Hampton, K., Ciampa, P.J., Brook, A., and Dyson, N. (1999). A genetic screen for modifiers of E2F in Drosophila melanogaster. Genetics 153, 275–287.

Stephens, F.B., Constantin Teodosiu, D., and Greenhaff, P.L. (2007). New insights concerning the role of carnitine in the regulation of fuel metabolism in skeletal muscle. J. Physiol. 581, 431–444. ‐

160

Stipp, D. (2011). Linking nutrition, maturation and aging: From thrifty genes to the spendthrift phenotype. Aging (Albany. NY). 3, 85–93.

Storch, J., and Thumser, A.E. (2010). Tissue-specific functions in the fatty acid-binding protein family. J. Biol. Chem. 285, 32679–32683.

Stunkard, A.J., Foch, T.T., and Hrubec, Z. (1986a). A twin study of human obesity. JAMA 256, 51–54.

Stunkard, A.J., Sørensen, T.I.A., Hanis, C., Teasdale, T.W., Chakraborty, R., Schull, W.J., and Schulsinger, F. (1986b). An Adoption Study of Human Obesity. N. Engl. J. Med. 314, 193–198.

Stunkard, A.J., Harris, J.R., Pedersen, N.L., and McClearn, G.E. (1990). The Body-Mass Index of Twins Who Have Been Reared Apart. N. Engl. J. Med. 322, 1483–1487.

Subramanian, M., Jayakumar, S., Richhariya, S., and Hasan, G. (2013). Loss of IP3 receptor function in neuropeptide secreting neurons leads to obesity in adult Drosophila. BMC Neurosci. 14, 157.

Suh, J.M., Gao, X., McKay, J., McKay, R., Salo, Z., and Graff, J.M. (2006). Hedgehog signaling plays a conserved role in inhibiting fat formation. Cell Metab. 3, 25–34.

Suh, J.M., Zeve, D., McKay, R., Seo, J., Salo, Z., Li, R., Wang, M., and Graff, J.M. (2007). Adipose is a conserved dosage-sensitive antiobesity gene. Cell Metab. 6, 195–207.

Sun, K., Zhang, Y., D’Alessandro, A., Nemkov, T., Song, A., Wu, H., Liu, H., Adebiyi, M., Huang, A., Wen, Y.E., et al. (2016). Sphingosine-1-phosphate promotes erythrocyte glycolysis and oxygen release for adaptation to high-altitude hypoxia. Nat. Commun. 7, 12086.

Swarup, S., and Verheyen, E.M. (2012). Wnt/wingless signaling in drosophila. Cold Spring Harb. Perspect. Biol. 4, 1–15.

Takai, T., Nishita, Y., Iguchi-ariga, S.M. m, and Ariga, H. (1994). Molecular cloning of MSSP-2, a c-rnyc gene single-strand binding protein: Characterization of binding specificity and DNA replication activity. Nucleic Acids Res. 22, 5576–5581.

Tansey, J.T., Sztalryd, C., Gruia-Gray, J., Roush, D.L., Zee, J. V, Gavrilova, O., Reitman, M.L., Deng, C.X., Li, C., Kimmel, a R., et al. (2001). Perilipin ablation results in a lean mouse with aberrant adipocyte lipolysis, enhanced leptin production, and resistance to diet- induced obesity. Proc. Natl. Acad. Sci. U. S. A. 98, 6494–6499.

Teixeira, L., Rabouille, C., Rørth, P., Ephrussi, A., and Vanzo, N.F. (2003). Drosophila Perilipin/ADRP homologue Lsd2 regulates lipid metabolism. Mech. Dev. 120, 1071–1081.

Tennessen, J.M., Barry, W.E., Cox, J., and Thummel, C.S. (2014). Methods for studying metabolism in Drosophila. Methods 68, 105–115.

161

Thanh, N.T.K., Stevenson, G., Obatoml, D., and Bach, P. (2000). Determination of Lipids in Animal Tissues by High-Performance Thin-Layer Chromatography with Densitometry. J. Planar Chromatogr. 13, 375–381.

Therrien, M., Morrison, D.K., Wong, A.M., and Rubin, G.M. (2000). A genetic screen for modifiers of a kinase suppressor of RAS-dependent rough eye phenotype in drosophila. Genetics 156, 1231–1242.

Thiam, A.R., Farese, R. V, and Walther, T.C. (2013). The Biophysics and Cell Biology of Lipid Droplets. Nat. Rev. Mol. Cell Biol. 14, 775–786.

Thumser, A.E., Moore, J.B., and Plant, N.J. (2014). Fatty acid binding proteins. Curr. Opin. Clin. Nutr. Metab. Care 17, 124–129.

Tian, Y., Bi, J., Shui, G., Liu, Z., Xiang, Y., Liu, Y., Wenk, M.R., Yang, H., and Huang, X. (2011). Tissue-autonomous function of drosophila seipin in preventing ectopic lipid droplet formation. PLoS Genet. 7, e1001364.

Tokgoz, Y., Isik, I.A., Akbari, S., Kume, T., Sayin, O., Erdal, E., and Arslan, N. (2017). Perilipin polymorphisms are risk factors for the development of obesity in adolescents? A case-control study. Lipids Health Dis. 16, 52.

Trapnell, C., Williams, B.A., Pertea, G., Mortazavi, A., Kwan, G., van Baren, M.J., Salzberg, S.L., Wold, B.J., and Pachter, L. (2010). Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat. Biotechnol. 28, 511–515.

Trinh, I., and Boulianne, G.L. (2013). Modeling obesity and its associated disorders in Drosophila. Physiology (Bethesda). 28, 117–124.

Tsuji, M., Shinkura, R., Kuroda, K., Yabe, D., and Honjo, T. (2007). Msx2-interacting nuclear target protein (Mint) deficiency reveals negative regulation of early thymocyte differentiation by Notch/RBP-J signaling. Proc. Natl. Acad. Sci. U. S. A. 104, 1610–1615.

Tunstall, N.E., Herr, A., de Bruyne, M., and Warr, C.G. (2012). A screen for genes expressed in the olfactory organs of Drosophila melanogaster identifies genes involved in olfactory behaviour. PLoS One 7, e35641.

Turula, M., Kaprio, J., Rissanen, A., and Koskenvuo, M. (1990). Body weight in the Finnish Twin Cohort. Diabetes Res. Clin. Pract. 10 Suppl 1, S33-6.

Tyagi, S., Gupta, P., Saini, A.S., Kaushal, C., and Sharma, S. (2011). The peroxisome proliferator-activated receptor: A family of nuclear receptors role in various diseases. J. Adv. Pharm. Technol. Res. 2, 236–240.

162

Uranishi, H., Zolotukhin, A.S., Lindtner, S., Warming, S., Zhang, G.M., Bear, J., Copeland, N.G., Jenkins, N.A., Pavlakis, G.N., and Felber, B.K. (2009). The RNA-binding motif protein 15B (RBM15B/OTT3) acts as cofactor of the nuclear export receptor NXF1. J. Biol. Chem. 284, 26106–26116.

Urbanchek, M.G., Picken, E.B., Kalliainen, L.K., and Kuzon, W.M. (2001). Specific force deficit in skeletal muscles of old rats is partially explained by the existence of denervated muscle fibers. J. Gerontol. A. Biol. Sci. Med. Sci. 56, B191–B197.

Usui-Aoki, K., Ito, H., Ui-Tei, K., Takahashi, K., Lukacsovich, T., Awano, W., Nakata, H., Piao, Z.F., Nilsson, E.E., Tomida, J., et al. (2000). Formation of the male-specific muscle in female Drosophila by ectopic fruitless expression. Nat. Cell Biol. 2, 500–506.

Vadlamudi, R.K., Manavathi, B., Singh, R.R., Nguyen, D., Li, F., and Kumar, R. (2005). An essential role of Pak1 phosphorylation of SHARP in Notch signaling. Oncogene 24, 4591– 4596.

Valakh, V., Naylor, S.A., Berns, D.S., and DiAntonio, A. (2012). A large-scale RNAi screen identifies functional classes of genes shaping synaptic development and maintenance. Dev. Biol. 366, 163–171.

Veerkamp, J.H., and Maatman, R.G.H.J. (1995). Cytoplasmic fatty acid-binding proteins: Their structure and genes. Prog. Lipid Res. 34, 17–52.

Walther, T.C., and Farese, R. V (2012). Lipid Droplets And Cellular Lipid Metabolism. Annu. Rev. Biochem. 81, 687–714.

Wang, M.M. (2011). Notch signaling and Notch signaling modifiers. Int. J. Biochem. Cell Biol. 43, 1550–1562.

Wang, B., Moya, N., Niessen, S., Hoover, H., Mihaylova, M.M., Shaw, R.J., Yates, J.R. 3rd, Fischer, W.H., Thomas, J.B., and Montminy, M. (2011). A hormone-dependent module regulating energy balance. Cell 145, 596–606.

Wiellette, E.L., Harding, K.W., Mace, K. a, Ronshaugen, M.R., Wang, F.Y., and McGinnis, W. (1999). spen encodes an RNP motif protein that interacts with Hox pathways to repress the development of head-like sclerites in the Drosophila trunk. Development 126, 5373– 5385.

Winther, A.M.E., Acebes, A., and Ferrus, A. (2006). Tachykinin-related peptides modulate odor perception and locomotor activity in Drosophila. Mol. Cell. Neurosci. 31, 399–406.

Withers, R.F.J. (1964). Problems in the Genetics of Human Obesity. Eugen. Rev. 56, 81–90.

World Health Organization (2016). Obesity and Overweight.

Wu, T.D., and Nacu, S. (2010). Fast and SNP-tolerant detection of complex variants and splicing in short reads. Bioinformatics 26, 873–881.

163

Wu, Q., Wen, T., Lee, G., Park, J.H., Cai, H.N., and Shen, P. (2003). Developmental control of foraging and social behavior by the Drosophila neuropeptide Y-like system. Neuron 39, 147–161.

Wu, Q., Zhang, Y., Xu, J., and Shen, P. (2005). Regulation of hunger-driven behaviors by neural ribosomal S6 kinase in Drosophila. Proc. Natl. Acad. Sci. U. S. A. 102, 13289–13294.

Wu, X., Cooper, R.S., Borecki, I., Hanis, C., Bray, M., Lewis, C.E., Zhu, X., Kan, D., Luke, A., and Curb, D. (2002). A combined analysis of genomewide linkage scans for body mass index from the National Heart, Lung, and Blood Institute Family Blood Pressure Program. Am. J. Hum. Genet. 70, 1247–1256.

Wyatt, G.R., and Kalf, G.F. (1957). The Chemistry of Insect Hemolymph. Solutions 40, 833– 847.

Xu, S., Zhang, X., and Liu, P. (2017). Lipid Droplet Proteins and Metabolic Diseases. Biochim. Biophys. Acta.

Yabe, D., Fukuda, H., Aoki, M., Yamada, S., Takebayashi, S., Shinkura, R., Yamamoto, N., and Honjo, T. (2007). Generation of a conditional knockout allele for mammalian Spen protein Mint/SHARP. Genesis 45, 300–306.

Yamamoto, D., Sato, K., and Koganezawa, M. (2014). Neuroethology of male courtship in Drosophila: from the gene to behavior. J. Comp. Physiol. A. Neuroethol. Sens. Neural. Behav. Physiol. 200, 251–264.

Yan, D., and Perrimon, N. (2015). spenito is required for sex determination in Drosophila melanogaster. Proc. Natl. Acad. Sci. U. S. A. 112, 11606–11611.

Yang, T.T.C., Suk, H.Y., Yang, X., Olabisi, O., Yu, R.Y.L., Durand, J., Jelicks, L.A., Kim, J.-Y., Scherer, P.E., Wang, Y., et al. (2006). Role of transcription factor NFAT in glucose and insulin homeostasis. Mol. Cell. Biol. 26, 7372–7387.

Yang, X., Li, J., Qin, H., Yang, H., Li, J., Zhou, P., Liang, Y., and Han, H. (2005). Mint represses transactivation of the type II collagen gene enhancer through interaction with ??A- crystallin-binding protein 1. J. Biol. Chem. 280, 18710–18716.

Yin, M., Lü, M., Yao, G., Tian, H., Lian, J., Liu, L., Liang, M., Wang, Y., and Sun, F. (2012). Transactivation of microRNA-383 by steroidogenic factor-1 promotes estradiol release from mouse ovarian granulosa cells by targeting RBMS1. Mol. Endocrinol. 26, 1129– 1143.

Yu, C., Chen, Y., Cline, G.W., Zhang, D., Zong, H., Wang, Y., Bergeron, R., Kim, J.K., Cushman, S.W., Cooney, G.J., et al. (2002). Mechanism by which fatty acids inhibit insulin activation of insulin receptor substrate-1 (IRS-1)-associated phosphatidylinositol 3-kinase activity in muscle. J. Biol. Chem. 277, 50230–50236.

164

Zabena, C., Gonzalez-Sanchez, J.L., Martinez-Larrad, M.T., Torres-Garcia, A., Alvarez- Fernandez-Represa, J., Corbaton-Anchuelo, A., Perez-Barba, M., and Serrano-Rios, M. (2009). The FTO obesity gene. Genotyping and gene expression analysis in morbidly obese patients. Obes. Surg. 19, 87–95.

Zhang, M., Zhang, Y., Ma, J., Guo, F., Cao, Q., Zhang, Y., Zhou, B., Chai, J., Zhao, W., and Zhao, R. (2015). The Demethylase Activity of FTO (Fat Mass and Obesity Associated Protein) Is Required for Preadipocyte Differentiation. PLoS One 10, e0133788.

Zhang, Y., Proenca, R., Maffei, M., Barone, M., Leopold, L., and Friedman, J.M. (1994). Positional cloning of the mouse obese gene and its human homologue. Nature 372, 425–432.

Zhao, X., Yang, Y., Sun, B.-F., Shi, Y., Yang, X., Xiao, W., Hao, Y.-J., Ping, X.-L., Chen, Y.-S., Wang, W.-J., et al. (2014). FTO-dependent demethylation of N6-methyladenosine regulates mRNA splicing and is required for adipogenesis. Cell Res. 24, 1403–1419.

Zhou, Z., Licklider, L.J., Gygi, S.P., and Reed, R. (2002). Comprehensive proteomic analysis of the human spliceosome. Nature 419, 182–185.

Zhu, X.W., Deng, F.Y., Wu, L.F., Tang, Z.X., and Lei, S.F. (2015). Functional mechanisms for type 2 diabetes-associated genetic variants. J. Diabetes Complications 29, 497–501.

Zimmerman, R., Strauss, J.G., Haemmerle, G., Schoiswohl, G., Birner-Gruenberger, R., Riederer, M., Lass, A., Neuberger, G., Eisenhaber, F., Hermetter, A., et al. (2004). Fat Mobilization in Adipose Tissue is Promoted by Adipose Triglyceride Lipase. Science (80-. ). 306, 1383–1386.

Zolotukhin, A.S., Uranishi, H., Lindtner, S., Bear, J., Pavlakis, G.N., and Felber, B.K. (2009). Nuclear export factor RBM15 facilitates the access of DBP5 to mRNA. Nucleic Acids Res. 37, 7151–7162.

(2014). Current Bloomington Recipe for Drosophila Medium.

165

APPENDIX A

Supplemental Methods

Diet and Development

Diet food was based on a modified Bloomington media with malt as discussed in

Chapter II, Fly Strains and Husbandry. MYD contained 35g yeast per liter. HYD contained

70g yeast per liter. HFD contained 150g coconut oil per liter (15%). HSD contained 180g corn syrup per liter (1M). Larvae were reared in diets upon hatching and then counted three times daily for developmental progress. Fifty larvae transferred per genotype per diet. n=3.

Longevity

Upon eclosion, flies were allowed 48 hours to mate, whereupon females were separated and placed in new vials, 20 flies per vial. Vials were flipped every 2 days and scored for deaths once daily. Log-rank test was used to calculate statistical significance with

Prism 6 software.

166

APPENDIX B

167

168

169

170

171

172

173

174

Figure B1: Domain conservation in Spen proteins. The Drosophila (Spen), mouse (MINT), and human (SHARP) orthologues presented to show conservation between regions. Drosophila RRMs are outlined in black with no shading. Notch binding domains for MINT and SHARP are outlined with gray shading. Nuclear receptor interacting domains for MINT and SHARP are outlined with blue shading. The SPOC domain is outlined with red shading. Sequence alignment performed using Clustal Omega.

175

APPENDIX C

Figure C1: RalA is necessary for fat accumulation in the fat body. (A) Percent of floating larvae in different density solutions. FB-specific RalA KD (dcg>iRalA) compared to KD control (dcg>iw). Fifty larvae per genotype per experimental replicate, n=8 biological replicates per genotype. (B) Genetic background controls (iRalA/+ and iw/+) for (A). Error bars represent SEM. P values represent results from ANOVA. (C) Absorbance at 530 nm as a measure of food intake, n=4. Error bars represent SD. (D) Average larval speed, pixels/sec. n=4. Error bars represent SEM. P values represent results from unpaired two-tailed t tests. *P < 0.05, ** P < 0.01, ***P < 0.001, **** P < 0.0001.

176

Figure C2: Dietary effects on Drosophila development. (A-C) Wandering timeframe for larvae reared on MYD vs. HYD, HFD, or HSD, as indicated at the top of each column. (D- F) Pupariation timeframe for larvae reared on diets. (G-I) Eclosion timeframe for larvae reared on diets. Gray and light blue lines represent the developmental timing for larvae reared on MYD, or MD. Black and dark blue lines indicate the development timing for larvae reared on the supplemented diets: HYD, HFD, or HSD (collectively called HD). Fifty larvae per genotype per diet, n=3. Error bars represent SD. MYD, medium yeast diet; HYD, high yeast diet; HFD, high fat diet; HSD, high sugar diet; MD, medium diet; HD, high diets; AED, after egg deposition.

177

Figure C3: Spen depletion in the FB results in early onset of death. Adult flies reared in MYD and tracked daily for survival. dcg>iSpen, n=108. dcg>iw, n=159. P value obtained by Log-rank test. ***P < 0.001

178