Characterization of Novel Solute Carriers in Humans, Mice and Flies

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Pharmacy 287

Solute Carriers in a Broad and Narrow Perspective

MIKAELA CEDER

ACTA UNIVERSITATIS UPSALIENSIS ISSN 1651-6192 ISBN 978-91-513-0980-4 UPPSALA urn:nbn:se:uu:diva-416506 2020 Dissertation presented at Uppsala University to be publicly examined in A1:107a, Biomedicinskt centrum (BMC), Husargatan 3, Uppsala, Friday, 16 October 2020 at 09:15 for the degree of Doctor of Philosophy (Faculty of Pharmacy). The examination will be conducted in English. Faculty examiner: Universitetslektor Anita Öst (Linköpings universitet).

Abstract Ceder, M. 2020. Characterization of Novel Solute Carriers in Humans, Mice and Flies. Solute Carriers in a Broad and Narrow Perspective. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Pharmacy 287. 81 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-513-0980-4.

The solute carrier family is the largest family of membrane-bound transporters in humans, with 430 members divided into 65 subfamilies. They transport various substrates across lipid barriers and are vital for absorption, distribution, metabolism and excretion in all cell types in the body. Despite being involved in vital functions, and their effect on both physiology and pathophysiology, many transporters are not characterized. The aim of this thesis was to study newly identified putative solute carriers of which little is known. In Paper I, the relationship of solute carriers in humans and fruit flies was studied. The study revealed that 54 of the 65 subfamilies in humans have one or more orthologues in fruit flies, and a total of 381 orthologues were identified in fruit flies. In Paper II, a comprehensive study of the putative solute carriers and their response to different sugar concentrations were performed. Several, but not all, putative solute carriers were altered in cell cultures maintained in media containing low or no glucose, and the expression normalized upon refeeding with glucose. Similar results were observed in fruit flies subjected to complete starvation or diets with varying sugar concentrations. Last, in Paper III and IV, characterization of one putative solute carrier, UNC93A, was performed. The studies revealed that UNC93A was a conserved protein with an abundant expression in the body of mice but with a restricted expression in fruit flies. The protein was found to possibly be expressed at, or close to, the plasma membrane of cells and to co-localize with Twik-Acid sensitive potassium channels. UNC93A was found to be important for the renal function in fruit flies and to affect survival and membrane potentials in cells. The findings of this thesis establish a high conservation of several putative solute carriers and that they have a highly dynamic regulation during fluctuating energy and glucose availability. Further, while several clear biological aspects of UNC93A was identified, the exact function of transporter proteins is cumbersome to find and more research about these transporters is needed to fully understand their mechanistic role and their association and/or involvement in health and sickness.

Keywords: Solute carrier, SLC, Putative SLCs, Major facilitator superfamily, MFS, Drosophila melanogaster, Glucose metabolism, UNC93A, CG4928

Mikaela Ceder, Department of Pharmaceutical Biosciences, Box 591, Uppsala University, SE-75124 Uppsala, Sweden.

ISSN 1651-6192 ISBN 978-91-513-0980-4 urn:nbn:se:uu:diva-416506 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-416506) To my husband, daughter, and son, with you I can accomplish everything

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Ceder M.M., Fredriksson R (2020). Bridging the gap: The human and D. melanogaster repertoire of Solute Carriers. Submit- ted manuscript

II Ceder M.M.*, Lekholm E.*, Klaesson A., Tripathi R., Schweizer N., Weldai L., Patil S., Fredriksson R. (2020) Glucose availability alters gene and protein expression of several newly classified and putative Solute Carriers in mice cortex cell culture and D. melanogaster. Frontiers in Cell and Developmental Biol- ogy, 8(579)

III Ceder M.M., Lekholm E., Hellsten S.V., Perland E., Fredriksson R. (2017) The neuronal and peripheral expressed membrane- bound UNC93A respond to nutrient availability in mice. Front Mol Neurosci, 10:351

IV Ceder M.M., Aggarwal T.*, Hosseini K.*, Patil S., Perland E., Maturi V., Williams M.J., Fredriksson R. (2020) CG4928 is vital for renal function in fruit flies and membrane potential in cells: Characterizing the putative Solute Carrier UNC93A. Submitted manuscript

*Equal contribution

Reprints were made with permission from the respective publishers. Additional papers

V Perland E., Lekholm E., Eriksson M.M., Bagchi S., Arapi V., Fredriksson R. (2016) The putative SLC transporters Mfsd5 and Mfsd11 are abundantly expressed in the mouse brain and have a potential role in energy metabolism. PloS one, 11(6):e0156912

VI Roshanbin S., Lindberg F.A., Lekholm E., Eriksson M.M., Perland E., Åhlund R., Raine A., Fredriksson R. (2016) Histo- logical characterization of orphan transporter MCT14 (SLC16A14) shows abundant expression in mouse CNS and kidney. BMC Neurosci, 17(1):43

VII Williams M.J., Perland E., Eriksson M.M., Carlsson J., Er- landsson D., Laan L., Majebali T., Potter E., Fredriksson R., Benedict C., Schiöth HB. (2016) Recurrent sleep fragmentation induces insulin and neuroprotective mechanisms in middle-aged flies. Front Aging Neurosci, 8:180.

VIII Perland E*., Hellsten S.V*., Lekholm E., Eriksson M.M., Arapi V., Fredriksson R. (2017) The novel membrane-bound proteins MFSD1 and MFSD3 are putative SLC transporters affected by altered nutrient intake. J Mol Neurosci, 61(2):199- 214

IX Lekholm E., Perland E., Eriksson M.M., Hellsten S.V., Lind- berg F.A., Rostami J., Fredriksson F. (2017) Putative membrane-bound transporters MFSD14A and MFSD14B are neuronal and affected by nutrient availability. Front Mol Neurosci, 10:11

X Hellsten S.V., Eriksson M.M., Lekholm E., Arapi V., Perland E., Fredriksson R. (2017) The gene expression of the neuronal protein, SLC38A9, changes in mouse brain after in vivo starvation and high-fat diet. PLoS One, 12(2):e0172917 XI Hellsten S.V., Hägglund M.G., Eriksson M.M., Fredriksson R. (2017) The neuronal and astrocytic protein SLC38A10 transports glutamine, glutamate and aspartate, suggesting a role in neurotransmission. FEBS Open Bio, 7(6):730-746

XII Hellsten S.V., Tripathi R., Ceder M.M., Fredriksson R. (2018) Nutritional stress induced by amino acid starvation results in changes for Slc38 transporters in immortalized hypothalamic neuronal cells and primary cortex cells. Front Mol Biosci, 5:45

XIII Aggarwal T., Patil S., Ceder M.M., Hayder M., Fredriksson R. (2020) Knockdown of SLC38 transporter orthologue – CG13743 reveals a metabolic relevance in Drosophila. Front Physiol, 10(1592)

*Equal contribution

Contents

Introduction ...... 13 Transporters ...... 13 Solute Carriers ...... 15 The evolution of Solute Carriers ...... 16 Putative Solute Carriers ...... 16 Solute Carriers’ role in metabolism ...... 17 Solute Carriers are important therapeutic targets ...... 21 Methodology ...... 23 In silico models to understand the evolutionary connection among species and predict protein structures ...... 23 Hidden Markov Model ...... 23 Protein alignment ...... 24 Phylogenetic trees ...... 24 Modeling and structure prediction of proteins ...... 26 In silico predictions of transcription factor binding sites for transcription factors ...... 27 In vitro and in vivo models in research ...... 28 In vitro models and their advantages and disadvantages ...... 28 In vivo models and their advantages and disadvantages ...... 29 D. melanogaster as a model organism ...... 30 D. melanogaster as a model for human diseases and pharmacology ...... 33 Aim ...... 36 Summary of research papers ...... 38 Paper I: Bridging the gap – the human and D. melanogaster repertoire of Solute Carriers ...... 38 Methodological considerations ...... 39 Key findings ...... 40 Paper II: Glucose availability alters gene and protein expression of several newly classified and putative Solute Carriers in mice cortex cell culture and D. melanogaster...... 4 1 Methodological considerations ...... 45 Key findings ...... 47 Paper III and Paper IV: Characterization of to the putative Solute Carrier UNC93A...... 48 Methodological considerations ...... 53 Key findings ...... 55 Conclusions ...... 56 Perspectives ...... 57 Populärvetenskaplig sammanfattning ...... 60 Acknowledgements ...... 63 For My and Mio ...... 67 References ...... 68 Abbreviations

AAR Amino acid response AARE Amino acid response element ADP Adenosine diphosphate AMP Adenosine monophosphate AMPK AMP-activated protein kinase APC Amino acid-polyamine-organocation superfamily ATF4 Activating transcription factor 4 ATP Adenosine triphosphate BBB Blood-brain barrier BLAST Basic local alignment search tool C/EBP CCAAT/enhancer-binding protein ChoRE Carbohydrate response element ChREBP Carbohydrate-responsive element binding protein CPA/AT Cation:proton antiporter/anion transporter DAMS Drosophila activity monitor system DMT Drug/Metabolite Transporter Superfamily ER Endoplasmic reticulum ENaC Epithelial sodium channel FDA Food and Drug administration flyPAD Fly proboscis and activity detector Gal4 Galactose-responsive transcription factor GAL4 GCN2 General control nonderepressible 2 GFP Green fluorescent protein GPCR G-protein-coupled receptor HGNC Human gene nomenclature committee HMM Hidden markov model ICC Immunocytochemistry IHC Immunohistochemistry IT Ion transporter superfamily KCNK Potassium channel subfamily K MFS Major facilitator superfamily MFSD Major facilitator superfamily domain MLX Max-like protein MLXIP MLX interacting protein MLXIP MLX interacting protein-like mRNA Messenger RNA MT Malpighian tubules mTOR Mammalian target of rapamycin mTORC12 Mammalian (mechanistic) target of rapamycin complex 1–2 NHE Sodium/proton exchanger NKCC Sodium/potassium/chloride cotransporter NSRE12 Nutrient sensing response element 1–2 Pfam Protein family PLA Proximity ligation assay PMDA Pharmaceuticals and Medical Device Agency PSI-BLAST Position specific iterative BLAST qRT-PCR Quantitative real-time polymerase chain reaction RNA Ribonucleic acid RNAi RNA interference SLC Solute Carrier SPNS Sphingolipid transporters SSM Solid supported membrane SV2 Synaptic vesicle glycoprotein 2 SVOP/SVOPL SV2 related proteins TASK TWIK related acid sensitive potassium channel TCA Tricarboxylic acid cycle TCDB Transporter classification database TF Transcription factor TFBS Transcription factor binding site TLR Toll-like receptor TMS Transmembrane segment tRNA Transfer RNA TSS Transcription start site UAS Upstream activator sequence Unc-93 Uncoordinated-93 protein Introduction

The cell keeps the fundamental molecules of life within its lipid barrier, which also protects the inner milieu from the surrounding. This is true not only for the cell itself, but also organelles that lies within the cellular compartment. For the inner milieu to stay connected with the outer environment the membranes contain membrane-bound proteins that acts as “gate-keepers” [1, 2]. Mem- brane-bound proteins can roughly be divided into peripheral and integral membrane proteins [3]. Peripheral membrane proteins attach to the membrane temporarily; either to an integral membrane protein or directly to the lipid barrier through different anchors [4]. These proteins play vital roles as regulators both of transporters and receptors but also in cellular processes such as prote- olysis and signaling [4, 5]. Integral membrane proteins, e.g. receptors, channels and transporters, are located within the membrane [6]. Approximately 30 % of the human proteome encodes membrane-bound proteins [7], and 2000 genes are estimated to be transporters or transporter-related proteins [8]. Due to their vital functions, many membrane transporters are linked to diseases affecting almost all organs in the body [8-10], and around half of all pharmaceutical drugs targets membrane-bound proteins [11].

Transporters Transporters enables water soluble molecules across the lipid membranes [12]. They are important for nearly all aspects of physiology, both in homeostasis, establishing electrochemical gradients and membrane potentials [8], and can be divided into passive and active transporters. Passive transporters, also known as facilitators or uniporters, transport substrates along a concentration gradient and will do so until the gradient is eliminated [13]. Uniporters are usually ion channels or carrier proteins that open upon stimulation by voltage, pressure or ligands [14]. Some examples of uniporters are the ion channels for sodium, calcium and potassium that are involved in the transmission of nerve impulses along neurons [15]; along with facilitative sugar transporters [16, 17] and nucleoside transporters [18]. Active transporters, on the other

13 hand, move substrates across a membrane against a gradient and they are usually associated with accumulation of molecules crucial for the cell. Active transporters can either be primary active or secondary active. The primary active transporters use adenosine triphosphate (ATP) as energy to move molecules against the electrochemical gradient [12]. Sodium, potassium, protons, magnesium and calcium are common molecules transported by primary active transporters of ATPase-type e.g. the Na+/K+-ATPase [19] that maintain membrane potentials. Other examples of primary active transporters are found in the mitochondrial electron transport chain, whose transport is driven by redox energy [20], and ATP-binding cassette (ABC) transporters that translocate substrates across membranes [21]. Secondary active transporters, also known as solute carriers (SLCs), rely on the existing electrochemical gradient to couple transport of substrates against the concentration gradient. They can either act as symporters or antiporters [22-24], moving one molecule with the electrochemical gradient to transfer another molecule against a concentration gradient. Symporters move two substrates in the same direction, e.g. the glucose symporter, while antiporters move two molecules in opposite directions, e.g. Na+/Ca2+-exchanger, Figure 1. The most common ion used for coupled transport in mammals is sodium, whose electrochemical gradient is then used to complete the transport, while for transporters in bacteria, and also in the mitochondria, hydrogen is the most common co-transported ion [25].

Figure 1. Transporters allow movement of molecules; some requires ATP (primary active transporters), while others do not (secondary active and passive transporters). Primary active transporters are pumps and ABC-transporters. SLCs are secondary active transporters and act as antiporters, symporters and uniporters. Passive transports are ion channels and aquaporins. Credit to Somersault 18:24 (www.somer- sault1824.com) for figure components, shared under creative common license.

14 Solute Carriers SLCs are transmembrane proteins composed of hydrophobic alpha helices connected by hydrophilic extra- and intracellular loops [12, 26]. They are located in the plasma membrane [17, 27-29] and the membranes of mitochondria [30], endoplasmic reticulum (ER) [31], Golgi [32], proteolytic granule (e.g. lysosomes) [33] and vesicles [34-36]. SLCs were originally defined as ATP-independent transporters and protein sequences were sorted into the SLC superfamily based on their function rather than their structure and sequence identity. The SLC superfamily is the second largest family of membrane-bound proteins after GPCRs but are the largest family of transporters in humans [7, 37]. Current practice assigns a new protein sequence to a SLC family if it shares at least 20 % amino acid sequence identity to one other member of the family. This has resulted in a great degree of variation regarding structure and primary sequence compared to other large membrane-bound protein families e.g. GPCRs and voltage-gated ion channels [38, 39]. Today the SLC superfamily comprises 439 members divided into 65 subfamilies, SLC1–SLC65, listed in the SLC table (slc.bioparadigms.org). Approximately two thirds of the SLCs are considered characterized and information regarding their expression, function and/or structure is available. SLCs have diverse functions and they transport numerous compounds, both charged and uncharged, across membranes. They are capable of transporting larger molecules e.g. peptides, amino acids, sugars, neurotransmitters and therapeutic drugs, but also single ions. The transport is dependent on the concentration and electrochemical gradients, which make the transport highly connected to the concentration of substrates available in the extra- and intracellular compartments [12]. For example, there are transporters that move the same substrate in different tissues with different affinity, e.g. the SLC2 family (Glucose transporters, GLUTs) [16]; and families with a diverse range of substrates but all members are expressed in the same organelle, e.g. SLC25 family of mitochondrial transporters [30]. Not all SLCs transport solutes on their own, some are required as supporting subunits e.g. the SLC3 family [40]. In- terestingly, new functions have recently been assigned to the SLC proteins and a few members are now suggested to both function as transporters and as sensors/regulators for pathways involved in nutrient availability and transport e.g. SLC38A9 [41-43]. The functional diversity among the SLCs is a possibility to why there are multiple classification systems in use; i) the Human Gene Nomenclature Commit- tee (HGNC), ii) the Transport Classification Database (TCDB) and iii) the

15 Protein Family Database (Pfam). The human SLCs are annotated and named according to a root system set by HGNC [44, 45]; “the SLC nomenclature”. TCDB uses phylogenetic and functional data to categorize all recognized transporters into superfamilies [46, 47], while Pfam aims to cluster all transporters into protein families (clans) based on sequence similarities [48]. Therefore, SLCs populate several Pfam clans, where seven, so far, contain more than one SLC family; (I) Major Facilitator Superfamily (MFS) clan (CL0015), (II) Amino acid/Polyamine/organocation (APC) clan (CL0062), (III) Cation:Proton antiporter/Anion Transporter (CPA/AT) clan (CL0064)) (IV) Drug/Metabolite Transporter Superfamily (DMT) clan (CL0184), (V) Ion Transporter Superfamily (IT) clan (CL0182), (VI) MtN3-like clan of vesicle-trafficking cargo receptors (CL0141) and (VII) Mvin, MATE-like Super- family clan (CL0222).

The evolution of Solute Carriers The physiological importance of transporters is evident as SLCs and SLC- related proteins are found in archaea, bacteria, plants and eukaryotes [37]. Several SLCs families are considered evolutionary old and have identified homologues in bacteria [37, 49, 50], e.g. SLC2, SLC22 and SLC25, and ancient members of algae [51], e.g. the SLC32, SLC36 and SLC38. Meanwhile, other families are more recent and are mainly identified in Animalia [37], e.g. SLC5, SLC6, SLC8 and SLC18. Some of these families are important for the developed nervous system, suggesting that these families have evolved paral- lel to the development of a more complex central nervous system. Protein sequences that presumably belong to the SLC superfamily are still being identified and several protein sequences have not yet been assigned to a SLC superfamily and are instead named putative (“atypical”) SLCs [52, 53].

Putative Solute Carriers Currently 20 out of the 65 SLC families are categorized into the MFS clan (see slc.bioparadigms.org). In the early 1990s, members of the MFS clan were believed to be facilitated sugar transporters [49, 50, 54]. Today, it is known that these transporters have a diverse substrate profile and perform their transport as uniporters and cotransporters [54]. The MFS is the largest group of transporter proteins known, with over 10,000 sequenced members [55],

16 present in both ancient life forms and mammals [56]. The amino acid sequence identity among SLCs and putative SLCs belonging to the MFS clan is low compared to other families, but their protein folding structure is conserved [57, 58]. The common structure for MFS proteins are an even number of alpha helices arranged in monomer or polymers, Figure 2, creating a globular protein with 12–14 transmembrane spanning helices [57, 58].

Figure 2. Members of the MFS usually have an even number of transmembrane helices arranged in monomers or polymers. Illustration exemplifies the general secondary structure of MFS transporter with 12 alpha helices spanning the membrane.

Several putative SLCs belonging to the MFS clan were identified during genome annotation [59]. Most of them are classified into the MFS domain (MFSD#) nomenclature, however also previously named genes such as SV2, UNC-93 and SPNS belong among these putative SLCs. Recently, MFSD2A and MFSD2B (SLC59); MFSD3 (SLC33); MFSD4A and MFSD4B (SLC60); MFSD5 (SLC61); MFSD10 (SLC22); SV2A, SV2B, SV2C, SVOP and SVOPL (SLC22); and SPNS1, SPNS2 and SPNS3 (SLC63) were sorted into SLC families (slc.bioparadigms.org) and more of these putative SLCs will probably be sorted into the SLC superfamily in the near future. Information regarding putative SLCs has increased over the past years and now the gene and protein expression in mouse are established for many of them [60-65].

Solute Carriers’ role in metabolism The process of life is dependent on access to sufficient levels of essential solutes, e.g. amino acids, fatty acids, ions, nucleotides and sugars to maintain fundamental function within each cell, and to maintain homeostasis of important molecules. Mechanisms to sense and react to fluctuations in solutes have evolved to keep circulating nutrients within a narrow range both in unicellular and multicellular organisms [66-68]. SLCs make up a highly dynamic

17 interface to supply the cell with major components needed for acidification, glycolysis, TCA cycle, neurotransmission and nucleic acid synthesis [66], Figure 3.

Figure 3. SLCs are important for several aspects of the cellular metabolism and maintain influx and efflux of substrates that contribute to acidification, protein and nucleic acid synthesis, glycosylation, glycolysis and TCA cycle. Some examples of SLCs important for these processes are: glucose transport (SLC2 and SLC5), amino acid transport (SLC1, SLC7:SLC3 and SLC38); mitochondrial transporters (SLC25;, mon- ocarboxylate transporters (SLC16); fatty acid transporters (SLC27); ion transporters (SLC9 and SLC30); nucleoside-sugar transporters (SLC35) and nucleobase transporters (SLC29). Credit to Somersault 18:24 (www.somersault1824.com) for figure components, shared under creative common license.

Amino acids are important for protein synthesis and catabolism during feeding and starvation, and they are essential as neurotransmitters and for food intake [68]. The levels of amino acids are important to keep constant and the levels are regulated through amino acid response (AAR), general control nonderepressible 2 (GCN2) and the mechanistic target of rapamycin (mTOR) pathways [68]. The AAR pathway is activated by low levels of essential amino acids and lead to an upregulation of the activating transcription factor 4

18 (ATF4), which binds genes with amino acid response elements (AARE) and nutrient sensing response elements (NSRE-1 and 2), resulting in a reduced protein synthesis [69-71]. The GNC2 pathway is triggered by the accumulation of uncharged transfer RNAs (tRNAs) (normally covalent linked to a certain amino acid) that occurs during decreased levels of free amino acids. The activation of GNC2 leads to inhibitory phosphorylation of the essential activator of translation initiation, eukaryotic translation initiator factor 2α (eIF2α), also resulting in reduced protein synthesis [72]. The mTOR pathway, with its two distinct protein complexes, mTOR complex 1 (mTORC1) and mTOR complex 2 (mTORC2), integrates both extra- and intracellular signals to regulate anabolic and catabolic processes to coordinate nutrient levels [73, 74]. Each complex, mTORC1 and mTORC2, phosphorylates various effectors as a reaction to the varying energy demands of the cell. The mTORC1 is involved in promoting anabolic processes and limiting catabolic processes as a response to e.g. availability of nutrients such as amino acids and glucose, while mTORC2 promotes proliferation and survival as a response to growth factors [73, 75]. Leucine is an essential amino acid to initiate protein synthesis and cell growth through mTORC1 [76, 77]; as such, several SLC family members, e.g. SLC1A5, SLC3A2 and SLC7A5, are implemented in regulation of this complex since they maintain sufficient leucine levels upstream of this particular pathway. The mTORC1 is not an amino acid sensor itself, but recently the lysosomal arginine transporter, SLC38A9, was found to act as a sensor for mTORC1 by conveying the arginine and leucine availability directly to the complex [41-43]. Sugar, especially glucose, is a main source of energy for the body; and, there are several SLCs families that are specialized to help the cell to maintain glucose intake, storage, mobilization and breakdown. Glucose is found, be- sides in glycolysis and gluconeogenesis, in several fundamental processes in the cell including the pentose phosphate pathway, where it generates NADPH and ribose-5-phosphate; the hexosamine pathway; protein glycosylation; ser- ine biosynthesis as well as purine and glutathione biosynthesis [68, 78, 79]. Therefore, one can assume that SLCs that transport sugars and sugar metabolites are implemented in different steps in all of these different pathways. Glucokinase, a hexokinase, catalyze the initial steps of glycogen synthesis and glycolysis, and is an important glucose sensor. Compared to other hexokinases, which mainly acts as phosphorylation machines, glucokinase has a low affinity for glucose and is therefore only activated by high levels of glucose. During conditions with low glucose, glucokinase help the body maintain transport of glucose, through SLC families, to organs with high metabolic rate

19 such as the brain, kidneys and muscles [68]. Another important glucose sensor is the bi-directional SLC2A2 (GLUT2), which sense extracellular glucose levels. During hyperglycemia, GLUT2 mediates influx for storage and energy production, while during hypoglycemia it mediates efflux of glucose to the circulation [80]. In addition, members of the SLC5 family, in particularly SLC5A1 and A2, are cotransporters for glucose and sodium that provide the cell with glucose needed during glycolysis [66, 81]. The AMP-activated protein kinase (AMPK) and mTORC1 complex are also important elements for glucose sensing [68]. AMPK is activated by increasing levels of AMP and ADP, which induce glucose and fatty acid uptake and oxidation to increase ATP production [68]. The mTORC1 is regulated by glucose availability through the activity of the Rag GTPases, however, the mechanisms behind this are less clear [68]. Similar to the AAR pathway, there are carbohydrate response elements-binding proteins (ChREBP) that binds to carbohydrate response elements (ChoRE) to alter gene regulatory pathways in a glucose-dependent manner [82-86]. These elements mainly regulate genes involved in glycolytic and lipogenic processes, but also in insulin dependent pathways. However, compared to the AAR pathway and the AAREs, the mechanisms are less clear. It is suggested that Max-like protein X (MLX) and MLX interacting protein (MLXIP, MondoA) form a complex that is regulated by glucose. It translocate from the outer mitochondrial membrane to the nucleus to enhance transcription of glycolytic target genes and regulators for glucose fluxes [82, 84, 86]. Similar theories have been proposed for the MLX:ChREBP complex [82-84]. SLC9A3 and SLC16A1 are reported to be important components when it comes to controlling the acidification of the cell [66]. SLC9A3 acts as a sodium:proton exchanger that prevents acidification in the gut [87, 88] and SLC16A1 mediates outward transport of lactate and protons produced by glycolysis [66, 89]. The mitochondrial carrier SLC25A1 exchanges malate and citrate:proton over the inner mitochondrial membrane to maintain the TCA cycle, whereas another member of the same family, SLC25A8 is responsible for transferring protons from the inner to the outer mitochondrial membrane for energy metabolism [30, 66]. The SLC27 and SLC59 families provide transport of fatty acids [90, 91] that instead of glucose can be used to yield ATP for the cell. Fatty acids are also important components of the membranes of the cell. In addition to the macronutrients, there are several SLC families that provide the cell with other essential molecules e.g. SLC39 family [92] transport zinc that is important for secretion of secretory granules; SLC29 family [18] carries nucleobases over the nucleus membrane that are needed

20 for the formation of nucleic acids and the SLC35 family [93] transports UDP- GlcNAc and UDP-galactose, which are biochemical precursors for polysac- charides.

Solute Carriers are important therapeutic targets The list of diseases that SLCs are involved in is long, which come as no sur- prise due to their diverse expression and function within the body. Except hav- ing key roles in important cellular processes to maintain health, they are also the cause of disorders spread throughout the body [66]. Several SLCs are co- expressed in the same tissue and it is speculated that when this healthy co- expression of SLCs is perturbed in the body, it becomes sick [66, 94]. The SLC superfamily is highly complex and several families are involved in a va- riety of acute and chronic diseases. For example malfunction in amino acid/neurotransmitter transporters (SLC1, SLC6, SLC17), sugar transporters (SLC2) and ion transporters (SLC9, SLC12) are all linked to a spectrum of brain disorders (Alzheimer’s disease, autism, depression, epilepsy, Parkin- son’s disease and schizophrenia), lung diseases (cystic fibrosis and sleep-dis- ordered breathing), liver diseases (neonatal diabetes and Rotor syndrome) and conditions linked to the kidney (hypertension, renal tubular acidosis, Hartnup disorder and gout) [66]. As therapeutic targets, the drug can act on the SLC itself or the SLCs can be used as mediators for drug distributions. Within the SLC superfamily, members of the SLC6 family are the most exploited drug targets for therapeutic treatment of depression [95]. Other drug targets include the SLC5, SLC12, and SLC18 families [95]. Two good examples of SLCs acting as mediators of drug transport are the SLC21 (also known as SLCO) and SLC35 families; where SLCO1B1 mediates sufficient drug distribution of statins in the liver and SLC35F2 is crucial for the delivery of a cancer drug, YM155, into tumor cells [94]. Currently, a little more than 40 % of all SLCs have been linked to a disease. Through the advancement in sequencing techniques the contribution of genetic variance of SLCs is clearer and the linkage to disorders are constantly increasing. However, compared to other membrane-bound proteins, e.g. GPCRs, the number of compounds targeting the SLCs is far lower [94, 96, 97]. Even with numerous SLCs linked to diseases only 12 SLCs are known to be utilized as drug targets [66, 95] compared to GPCRs with 475 approved drugs [96]. Between the years 2015 and 2018 only five drugs targeting SLCs

21 were approved by the Food and Drug Administration (FDA) and the Japanese Pharmaceuticals and Medical Devices Agency (PMDA), compared to the GPCRs that in the end of 2017 had approximately 321 drug candidates in clinical trials [95-98]. There are several possible reasons for the delay in drug development towards SLCs, where the major challenge has been the technical barrier to characterize SLCs [94, 99, 100]. For examples, cell-based assays suffer problems due to overexpression difficulties and compensatory mechanism; the lack of high-quality antibodies gives partial understanding of the subcellular localization; cumbersome transport assays fail in screening cotransport and gradient necessary to a complete functional characterization. In addition, structure analysis of the integral membrane proteins is notoriously difficult [100]. In recent years there have been an increase in the number of determined protein structure of SLCs and several transport mechanisms, e.g. rocker-switch and gated-pore mechanisms [100] have been established, which most likely can help the pharmaceutical industry understand and design appropriate ligands. Hopefully, the increase in research regarding SLCs will contribute to the development of more direct treatments as well as synergistic treatments to increase the efficacy of other pharmaceuticals through better distribution through targeted SLCs.

22 Methodology

Imagine a world where no animal or cell-based testing was used and instead drug candidates were tested on humans directly. The time required for a new compound to reach the market could probably be cut in half, but along the way unpredicted adverse effects, that could even be fatal, would most likely occur. Clearly, this is both unethical and impractical. Therefore, models are used in research to provide a system, a simple reflection of a complex reality, to define, simulate, visualize and quantify a question to understand e.g. physiological phenomena and processes. Modelling by e.g. programmed mathematical calculations in computers (in silico) and/or living material such as cell cultures (in vitro), vertebrates and invertebrates (in vivo) are all used in the pharmaceutical industry to find the best candidates for new drugs and to understand biological mechanisms. No model is superior over another, instead the key behind choosing a model is to consider the purpose of the research; which model that is most suitable to answer the research question and/or hypotheses, and also to keep in mind what the limitations are with each model.

In silico models to understand the evolutionary connection among species and predict protein structures A majority of current classifications systems are based on phylogeny, the evolutionary relationship between organisms. The evolutionary relationship can be predicted by using mathematical models to calculate similarities among genomes and proteomes, which then can be presented in phylogenetic trees.

Hidden Markov Model The Hidden Markov model (HMM) is a tool for homology searches to identify protein sequences that are evolutionary conserved within a set of protein sequences. The model is based on the Markov process, where the probability calculations of possible events are dependent on previous observations (states). In the HMM, the Markov process has unobservable (hidden) states,

23 but the parameters are still known (observed) [101, 102]. For example, in Pa- per I, the observed part of the model is the alignment of the human protein sequences used to build the HMM. Meanwhile the unobservable part of the model is the proteins not yet identified in a given proteome, in this case the proteome of D. melanogaster. The model calculates the probability of amino acid shifts over time and calculates the likelihood of how similar the assem- bling of protein sequences is between species.

Protein alignment Protein alignments are used to arrange sequences to identify regions of similarity, which can indicate conserved function, structure and/or provide information regarding evolutionary relationship.

Pairwise and multiple alignments Sequences alignment such as protein sequences can be performed on one sequence against another (pairwise) to search for identical and similar regions within the sequence, either over the sequence as whole (global) or in specific parts of the sequence (local). It can also be applied to several sequences at once (multiple alignment), where some of the tools available today perform a so-called progressive alignment, e.g. MAFFT [103] (Paper I–II and IV) and t- coffee [104, 105] (Paper III–IV). A progressive alignment is based on a hier- archal technique where the multiple alignment is built by performing several pairwise alignments starting with the most similar pair of the sequences and moves on to the least similar sequences. These methods are usually not opti- mal for a global alignment, but rather focus on conserved motifs within parts of the sequences [106].

Phylogenetic trees A multiple alignment can later be used to build phylogenetic trees, which aim to reconstruct an evolutionary relationship among species. The outline of the phylogenetic tree is based on gathered information from groups of species, e.g. protein sequences, and represent the most likely hypothesis on how the species have evolved from one common ancestor. The different parts of the phylogenetic tree provide us with important information. The root (the beginning of the tree) tells us about the most recent and common ancestor of the tree, while the branches (the horizontal lines) represent events, but not time directly, and a series of ancestors, resulting in the species at the end. The pattern of the connection between branches represent the hypothesis on how the

24 species have evolved from the ancestors. Each branch point, also referred to as an internal node, indicate an event leading to the split of one group into two [107, 108]. The phylogenetic tree will tell us which groups e.g. protein sequences that are more related to each other, Figure 4. However, even if we know that modern organisms have Figure 4. Phylogenetic trees can tell us about the relationship be- evolved from ancient life-forms, we cannot tween the gathered data to pro- specify the exact paths of evolution. There- vide insight in family assortment fore, a tree only displays information-based or evolution. Group A, B, C and D have a common ancestor; how- hypotheses of relatedness and not definitive ever, group B, C and D are more facts. The more we uncover about a lineage related to each other compared the more accurate the tree becomes. to group A. There are several different approaches to construct a tree, where maximum parsimony and maximum likelihood are traditional alternatives. A tree built with maximum parsimony represent the simplest explanation supported by the given evidence to minimize the homoplasy (evolutionary events). The sim- plicity of this method fails to consider factors that affect how sequences might have evolved over time result in poor grouping. Meanwhile, when using maximum likelihood a model of evolution will be incorporated, where the highest probability of explaining the data will be the base for the construction of the tree [109]. Compared to maximum parsimony, maximum likelihood considers substitution of substrates e.g. amino acids as an evolutionary event, hence explaining the phylogenetic relationship in a more correct way. Bayesian interference [110], used in all papers presented in this thesis, also applies a model of evolution. However, the Bayesian interference is more efficient to construct phylogenetic trees compared to the more traditional techniques. This model has advanced compared to the regular maximum likelihood and has become more complex and efficient at the same time as it quantifies and addresses the source of uncertainty, prior and posterior probability [111]. Both the traditional maximum likelihood and Bayesian interference require a great main memory capacity and are usually less suited to handle a larger set of sequences. However, Bayesian interference has been speculated to produces “too good to be true” result in contrast to bootstrap, a computer-based technique for assessing the accuracy of a statistical estimation, used in maximum parsimony and maximum likelihood [112-114]. On the other hand, bootstrap used in phylogeny has been criticized to be too conservative when providing assessments of confidence levels of observed clades in a tree [115].

25 Another popular software for phylogenetic analysis is RAxML (Random- ized Axelerated Maximum Likelihood), which also uses maximum likelihood and bootstrap calculation to generate phylogenetic trees of larger datasets using less main memory and time [116, 117]. This software was used in Paper I to construct trees for the larger Pfam clans: MFS and APC.

Modeling and structure prediction of proteins The amino acid sequence and protein folding reveals a lot about its function. There are several methods to model macromolecules and these are deposit in the archive for macromolecular crystal structures, the Protein Data Bank (PDB) [118]. Large integral membrane proteins such as the SLCs have been difficult to model and only a few solved crystal structures exist [119]. A number of bacterial homologues [120-123] and some human SLCs e.g. SLC2A1 [124], SLC2A3 [125] and SLC42A3 [126] have solved protein structures but the number of solved structures is increasing. These structures have provided insight into the general structure of SLCs, with α-helices spanning the membrane connected by loops. More solved structures have also resulted in the development of online model and prediction software that can be used to esti- mate the structure of orphan proteins. Two such applications are PROTTER [127], used in Paper I and IV, and Phyre2 [128], used in Paper III and IV. PROTTER is a web-based application that predicts the secondary structure of proteins by gathering information about protein topology, annotated features such as post-transcriptional modifications, and experimental proteomic data from several repositories [129]. Phyre2 is also a web-based tool, but this suite uses a template-based modeling approach, meaning that the sequence of interest is aligned to a sequence of known structure, i.e. homology modeling, which due to the growing availability of data regarding evolutionary relationship and sequencing projects together with the improved computer capacity has led to great success for this method. A model prediction by Phyre2 requires first information about homologous sequences, from which it predicts a secondary structure and builds a HMM that is used in a fold library scan, which later is used to build a backbone, predict loops and side chains to reveal a tertiary protein structure. However, this model has its limitations. For instance, modeling will be impossible or unreliable if no homologous template is detected [128]. Only a small part of the deposited protein sequences in the Uni- ProtKB/TrEMBL database has their experimental structure solved [130]. De- spite the increased quality of in silico models challenges with the methods still

26 remains, and usually only fractions of the user-supplied protein sequence are represented. The computational software assemble proteins complexes by combining the fractions with information from different templates and structural domains, leaving the model with some uncertainty and inaccuracy [131]. Still, protein structure prediction and modelling are very useful and aid in the progression of the biological research field [130].

In silico predictions of transcription factor binding sites for transcription factors The advancement in genomics, transcriptomic and epigenomics has provided great opportunities to understand regulation of cellular transcription. Instead of focus on gene expression patterns that are associated with particular conditions, the research has advanced to interpret signaling pathways that regulates gene expression. Genes are enhanced or silenced by transcription factors (TFs), e.g. ATF4 [71, 132, 133] and MLX [82, 83], that reacts to cellular events such as fluctuations in nutrients, as described earlier, and binds to transcription factor binding sites (TFBS) in the promoter region of the gene and/or within the gene itself. These binding sites can be predicted by entering the gene sequence and its promoter region into computer-based tools such as the Eukaryotic promoter database (EPD) [134], used in Paper II–III. The search motif tools available via EPD is used to scan promoter regions with position weight matrices (PWM) of TFs; modelling binding specificity of the TFs and core promoter elements to find putative binding sites. The method is used to report TFBS by scanning a sequence for the presence of a specific motif that have greater similarity to the PWM than to the background. PWMs are partially based on position frequency matrices (PFMs), which display the frequency of each nucleotide at each position, and position probability matrices (PPMs) [135]. These are created by e.g. literature curation and experimental determination and the constructed PWMs are listed in open databases such as JASPER [136, 137]. During the past years, the PWM scanning method has improved, but still a remaining issue with the method is that it reports false positive predictions [138], most likely because the TFBS sequences are too short and too variable [135]. It is therefore very common that every gene in the genomes will have one match to the PWM of almost all TFs. Furthermore, the PWM prediction does not consider chromatin structure or epigenetic regulation of the gene, meaning that it cannot provide information

27 if the TFBS is accessible to the transcriptional machinery [139]. However, the PWM scanning method is a good tool to use to generate a list of preliminary putative binding sites that can further be interpreted.

In vitro and in vivo models in research In vitro refers to experimental procedures that are performed under controlled conditions outside a living organism, while in vivo refers to experimental procedures where a living organism is used e.g. animal studies. In vitro models make an excellent tool to explore mechanisms on the cellular level, as well as refine ideas and theories that later will be applied in in vivo models. Both in vitro and in vivo work is important to gain an overall picture of factors relevant for alterations in metabolism, gene and protein expression and to establish a proteins function.

In vitro models and their advantages and disadvantages Human cell lines are a good alternative for reducing the use of multicellular organisms such as mice. The cell culture condition can be strictly controlled, are cost effective, easy to use and provide unlimited supply of material, and of course, by-pass ethical concern. They also provide a pure population which minimize biological variations, gives constant samples and usually gives re- producible results. Immortalized cell lines are today used in several vital scientific areas such as vaccine production, drug metabolism and toxicity studies, gene function studies and synthesis of biological compounds such as antibodies and therapeutic proteins [140]. However, cell lines are often genetically manipulated, which could cause alterations in their function and how they react to stimuli compared to a normal cell. Also, microbial infections can go unnoticed for a longer time-period in cell lines, which could lead to altered behavior and gene expression. Hence, great care needs to be taken when using cell lines and validation is needed to support key findings. Without knowledge from unicellular life forms there would probably be more unsolved questions about the human body, its physiology and disease mechanisms unsolved today [141]. However, even if the unicellular system can be strictly controlled and provides good research reproducibility, a limitation with the system is the lack of complexity the in vivo systems have.

28 In vivo models and their advantages and disadvantages Model organisms are considered a precious resource in research to understand processes involved in human physiology and disease. Mice, for example, have a high genetic and physiological similarities to humans and are invaluable as a model to understand evolutionary conserved biological processes. However, despite the value of these models, it is important to remember the impact evolution has had on our and the model organism’s development and adaptation to survive in different environment [142]. There are also fundamental differences between the model organisms and humans, e.g. the number of chromosomes, differences in anatomy, organ function and physiology. In addition, the fruit fly, for example, lack or differs greatly in important physiological functions compared to humans, such as the adaptive immune system, cardio- vascular system and the blood-brain barrier (BBB). In these regards, both mice and flies can be questioned as reliable models to study human diseases since it is differences in the linking between genes along with different efficacy of pharmaceutical drugs between the species. However, several findings from these model organisms have pioneer the field of medicine and pharmacy pointing to their importance in research [142-145]. Disadvantages of using mice compared to fruit flies are increased costs, the time consumed to perform a study and the need of ethical permission, hence, there are less possibilities to perform certain studies in mice. However, in comparison to mice, the fruit flies have the disadvantages of increased spon- taneous mutations within stocks, gene redundancy and a single fly is usually not sufficient to generate material for several molecular assays. In addition, the genetic similarity and protein identity between human and mouse make it possible to study protein localization using antibody-related assays such as immunohistochemistry (IHC), immunocytochemistry (ICC), proximity ligation assay (PLA) and western blot (WB). Protein detection relies on antibodies: a primary antibody recognizing a specific antigen epitope in the sample that is visualized with a labeled secondary antibody that amplifies a fluorescent or chemiluminescent signal corresponding to the protein. In Paper II, WB was used to investigate total mTOR, to provide information about the general metabolic state of the cell, and ICC was used to investigate the protein expression and localization of six putative SLCs (MFSD1, MFSD4A, MFSD11, MFSD14A, MFSD14B and UNC93A). In Paper III and IV, WB was used to confirm antibody specificity and insert of a CG4928 expression vector, re- spectively. ICC was used to visualize UNC93A expression and PLA was used to study co-localization between CG4928/UNC93A and TWIK related acid- sensitive K+ (TASK) channels, but no immunostaining was performed in flies.

29 There are antibodies available for several targets related to human disease genes that can be used in fruit flies, but to study the protein localization of an orphan gene can be challenging, since these antibodies usually need extensive verifications to demonstrate that the immunostaining is correct.

D. melanogaster as a model organism D. melanogaster has been used as a model organism for over 100 years, and is still widely used in genetic, immune system, development, physiologic and behavior research. There are several features that has made this insect a popular model organism; i) easy and cheap maintenance, ii) short generation time, iii) high fecundity, iv) easy to morphologically separate by eye, v) studies do not require ethical per- mits and vi) its complete genome was sequences in 2000 [146]. The size of the D. melanogaster genome is about 5 % of the size of the human genome, with approximately 15,500 genes Figure 5: The life span of D. melanogaster is three spread on four chromo- weeks at 25 ºC. The fly goes through six stages of life: embryo, first, second and third instar larva, pupa and somes: three autosomes adult. The embryo stage is around one day before the and one sex chromosome. first instar larva emerges. All three larva stages are Around 4 % of all genes are four to five days prior pupation. Metamorphosis takes four days before the adult fly emerge. Credit to suggested to encode trans- Andreas Prokop for images of adult flies [147]. porter and/or transporter- related protein [148]. The complete life cycle of the fruit fly is around three weeks and the development (egg to adult) is approximately 8.5 days at 25 ºC, Figure 5. However, the life cycle and developmental timeline is highly affected by temperature. Briefly, the fertilized eggs (embryos) hatch after 12 to 15 h, and the first instar larvae emerge. The larvae stage is around four days, where molting occurs twice, approximately 24 h (second instar) and 48 h (third instar) after hatching,

30 before pupation. After four days of metamorphosis, the adult fly emerges [149]. There are several genetic tools that are currently used to influence the genetic patterns of D. melanogaster; galactose-responsive sequence, transposon mutagenesis and CRISPR/CAS9 system [150]. The Gal4-UAS system is a tool to regulate gene expression in D. melanogaster. The system is dependent on two parts: the Gal4 gene, a transcription factor from yeast [151], and an Up- stream Activator Sequence (UAS), a cis-acting regulatory sequence (CGG-

N11-CCG), that Gal4 binds to and promote gene transcription [152]. To obtain flies that express the combination of Gal4 and UAS parental flies containing one part each will be mated, rendering their offspring to contain both, Figure 6. In the fruit fly, Gal4 is usually positioned under the control of a promoter or a tissue specific enhancer, which enables studies in specific subsets of cells and tissues, e.g. specific neurons, ubiquitous in the whole fly or a single gene of interest. The UAS, on the other hand, are present next to the gene of interest and can be used to express reporter genes such as green fluorescent protein (GFP) or interfering RNA (RNAi).

Figure 6. The Gal4-UAS system is a tool used to regulate and study genes in D. melanogaster. It is composed of two parts, the Gal4 transcription factor and a regulatory sequence, UAS. Gal4 is positioned under the control of a promoter and the UAS is introduced near a gene. This technique can be used to drive the expression of a reporter gene e.g. GFP or to perform gene silencing using RNAi.

When placing the UAS next to a RNAi sequence, a double-stranded RNA is synthesized upon activation of the UAS. The produced RNA is complemen- tary to the gene of interest and is recognized as exogenous genetic material that will be eliminated by activating the RNAi pathway [153]. However, this

31 technique has disadvantages. Research has shown that overexpression of Gal4 can lead to non-specific side-effects, usually linked to the immune system and stress responses [154], which must be considered when using the system. It has been reported that a percentage of the lines produce non-specific phenotypes, hence it has become important to verify the experimental model [155]. For example, so-called GD RNAi lines have a transgene insertion that is P-element based and the insertion site is random [156]. KK RNAi lines were constructed to increase the efficiency and specificity of genetic phenotypes via a targeted insertion of the transgene, a hairpin construct, to a single location. However, 80 % of the GD lines are estimated to produce potent and gene- specific silencing, while only 75 % of the KK lines are predicted to produce a gene-specific phenotype [157, 158]. The genetic power of being able to control the gene expression has also led to the possibility to study alleles that are fatal in homozygous condition. These fly lines are kept alive by genetic markers, also called balancer chromosomes, to prevent recombination and death, but also to make it possible to separate the genetic modified offspring by their appearance.

Figure 7. Genetic markers are common to use in D. melanogaster. They prevent recombination and are used to keep stocks of lethal alleles heterozygous. They can also be used to easily separate flies that express the genotype of interest. Some common balancers are Cy1, Sb1, w1 and y1. Credit to Andreas Prokop for images of flies [147]. Balancer chromosomes have three important properties; i) to carry dominant markers, ii) to make homozygosity affect reproduction negatively and iii) to suppress recombination with their homologues. Some common balancers are Cy1 (Curly wings), Sb1 (short and thick bristles), w1 (eyes that lack pigmenta- tions, white) and y1 (body and wing pigmentation, yellow), Figure 7.

32 D. melanogaster as a model for human diseases and pharmacology Even if the size difference is substantial between humans and D. melanogaster, we are more alike than we think. Biochemical pathways, as well as biological and physiological features are conserved between D. melanogaster and mammals [159] and around 60 % of the genes in human are conserved in D. melanogaster [160]. A large set (75 %) of all known human disease genes have a homologue in D. melanogaster [160], hence, this model organism is promising to study numerous disease and to perform preclinical research. Many similarities exist between the human and fruit fly body, and several human organs have a counterpart in the embryo, larvae and adult body of D. melanogaster. The hindgut and Malpighian tubules (MTs) of the fruit fly fulfil similar functions as the human kidneys, which has provided a good system to study genetic and molecular basic functions of the excretory system. Excretion of waste and homeostasis of ions, water and sugars is a common and important biological function in both species. The important parts of the human kidneys are the nephrons, while the renal system of D. melanogaster consists of two pairs of tubules [161]: MTs and nephrocytes that reside close to the esophagus and heart. MTs work similar to the tubular parts of the nephron, and are responsible for absorption, reabsorption and secretion [162, 163], while nephrocytes remove waste products from the hemolymph in similar manners as the glomerulus in the kidneys [164]. Several SLCs families that are essential for excretion (SLC5, SLC6, SLC12 and SLC21) are conserved in the excretory system of the fruit fly [165]. The growth and development of all multicellular organisms depends on nutrient availability, hence both humans and fruit flies have developed sensory systems to monitor nutrients; amino acid and sugar homeostasis pathways [165]. SLC7, SLC13 and SLC29 orthologues, with the same or similar functions, are present in D. melanogaster, where all orthologues are linked to normal body and organ growth as well as survival. The anatomic divergence between the human and D. melanogaster brain are apparent, which make it difficult to recapitulate morphological features observed in neurological diseases such as Alzheimer’s disease. Furthermore, the BBB of D. melanogaster is much simpler compared to humans, and it is only formed by glial cells surrounding the nervous system. However, the BBB of fruit flies does share similar functions and basic principles as the human BBB and several SLC families in this complex structure are well-preserved between the two species. Further- more, pathways are conserved e.g. neurotransmitters and mechanisms for

33 neurotransmission storage, release and recycling are conserved [165-167]; hence, some defined phenotypes of human neurological disease can be reproduced and used to identify and understand underlying genetic and molecular pathways. In addition, a majority of the metabolic pathways in humans are conserved in D. melanogaster, and recently, studies have highlighted that same or similar phenotypes observed in humans are found in flies when these metabolic pathways are challenged [168, 169]. Sometimes a prominent phenotype; seizures, disrupted movement patterns, obesity, edema and severe effects of pharmaceutical drugs can be seen by the naked eye. Usually, however, this is not the case and therefore behavioral assays have been used with great success in D. melanogaster. Several sophisticated techniques have been developed for example the Drosophila Activity Monitor System (DAMS) [170], the fly Proboscis and Activity Detector (fly- PAD) [171] and Ramsay Assay [172]. DAMS is a useful system when studying general activity, sleep patterns and circadian rhythm as well as starvation resistance and survival. DAMS au- tomatically record the movement of single flies every time they pass a laser- beam. This system has been used to study the consequences of disrupted sleep- wake patterns [173]. The findings have established the importance of sleep and has mapped specific genes involved in the sleep-wake behavior. It has also been used to study the activity of flies with altered biological pathways important for movement [174] and circadian rhythm [175]. To monitor changes of metabolic pathways flyPAD has been used with success. It measures the number of meals (sips), the sizes of meals and the duration and the system has contributed to our understanding of genes and their connection to feeding. In addition, there are assays developed to measure circulating and stored macronutrients; something that is usually altered during obesity and diabetes. Together these two methods can provide information about genes and their impact on specific parts of metabolism. It is of great importance to maintain the ionic and osmotic homeostasis during changes in external conditions by modulating the renal epithelial ion and water transport. The MTs of D. melanogaster offer a good model to study molecular mechanisms of the renal ion transport by using Ramsay assay. The Ramsay assay is used to measure fluid secretion rate per minute and by using electrophysiology the ion fluxes in each droplet can be measured [172]. Taken together, these similarities and techniques prove D. melanogaster to be a promising model to study the fundamental molecular pathways causing diseases linked to transporters and to identify novel pharmaceutics and therapeutic strategies [165]. The research and development processes in the

34 pharmaceutical industry are costly and time consuming. Hence, this model organism enables an easy and cheap first stage research platform to characterize orphan transporters, to identify transporter that mediates the cellular influx and efflux of new and current pharmaceutics and how the genetic variants in membrane transporters contribute to pharmacokinetics and pharmacodynam- ics.

35 Aim

Despite the important impact SLCs have on fundamental processes of life, research lags far behind other protein families and many members are still orphan. There have been technical barriers limiting the research on these integral membrane proteins, however, the advancement of today’s technologies and tools create a promising future for further research. SLCs are evolutionary old and well conserved in several model organisms. During the past years several new SLCs have been identified and there is a need to update the repertoire of orthologues proteins and create new databases. Therefore, a detailed mining of the complete SLC family in D. melanogaster was performed in Paper I. Through phylogenetic analysis, putative SLCs have been identified and recently a few of those were classified into existing or new SLCs families. Little is known about these putative transporters. Several of them have been found to be affected by changes in amino acid level and during complete starvation. However, little has been reported about what impact varying glucose levels have on their expression and therefore in Paper II changes of several putative SLCs were studied as a response to glucose availability in cortex cell cultures from mice and D. melanogaster. Several assumed putative SLCS have not yet been classified to a SLC family since the knowledge about them are limited. Characterization of one orphan transporter was conducted to reveal its expression and function in Paper III and Paper IV.

The specific aims for each paper were:

Paper I: The aim was to investigate the evolutionary relationship of SLCs in human and D. melanogaster, and to identify protein-protein orthologues, thereby bridging the gap of the repertoire of SLCs in these two species.

Paper II: The first aim was to perform a focused study on the gene regulation of putative and not yet annotated SLCs in response to limited glucose levels in cortex cell cultures. The second aim was to investigate gene expression alterations of putative SLCs in D. melanogaster

36 subjected to starvation and different sugar concentration, and to see if the genes were regulated in similar patterns between the two model systems used. The third aim was to distinguish if the protein expression of the putative SLCs changes, i.e. if the transporters localization shifts from one compartment to another, as a response to glucose deprivation and starvation.

Paper III: The first aim was to investigate the protein expression profile and cellular localization of UNC93A in brain tissue and in primary cell cultures from mice. The second aim was to investigate gene expression of Unc93a in the central nervous system and peripheral organs of mouse and how the gene is affected by food restrictions in both an in vivo and in vitro model.

Paper IV: Here the aim was to begin to functionally characterize the orphan transporter UNC93A, by studying its orthologue, CG4928, in D. melanogaster, by knockdown techniques to discover the function of the orthologue.

37 Summary of research papers

Paper I: Bridging the gap – the human and D. melanogaster repertoire of Solute Carriers Phylogenetic analysis and protein alignment can be used to determine the evolutionary relationship between various biological species based on similarities and differences in their genetic characteristics. Previously, 348 orthologous proteins in D. melanogaster were identified as members of the SLC superfamily ([176], flybase.org). However, the focus of that study and the combined list on flybase.org were not to study the fly proteome in detail and search for specific protein orthologues. Some of the SLC classifications are based on assumptions and function rather than phylogenetics. Here, HMM for each

Pfam cluster of SLCs was built using HMMBUILD. Related proteins in D. melanogaster were identified by running the entire D. melanogaster promote against the human Pfam HMMs with HMMSEARCH [102, 177]. Meanwhile, global protein alignment was used to investigate if the identified orthologues meet the current criterion of 20 % amino acid sequence identity. We identified 381 orthologous proteins in D. melanogaster and delineated them to their respective SLC family, Figure 8. We found that in general the SLC superfamily is well conserved between these two species, highlighting their importance in cellular processes, but that there are 11 families that lack orthologues in D. melanogaster. This suggest that these 11 families have evolved later, probably to encounter the increased demands of a more complex biological system. Compared to other large protein families, e.g. GPCRs, the protein identity between the 65 SLCs families are relatively low, and we found that the total sequence identity of all human SLCs and their orthologues in flies was below 30 %. The finding reveals the diversity within the family, where sometime the only connection among family members is that they transport similar solutes in an energy independent manner. This reflects the difficulties in classifying the SLCs and why the current classification system was quite recently adopted, leading to a delay in characterization. Hopefully, the annotation of SLC protein sequences in D. melanogaster and assigning them to families can

38 promote further research ideas and usage of D. melanogaster in human disease models and drug development.

Figure 8. The SLC family is well conserved between humans and flies. There are a total of 442 SLCs and putative SLCs present in the human proteome, while in D. melanogaster 381 orthologues protein sequences to the superfamily were identified. The pie-charts illustrate the clustering of the SLC members in the different clans in humans (left) and in flies (right).

Methodological considerations D. melanogaster has been used as a model organism for a long time and many orthologues SLCs have already been identified in multiple publications and a summary of these has been generated at flybase.org. The human SLC family has grown since 2010 and 55 new SLC protein sequences have been added to the superfamily without using HMM. Instead alternative ways to reveal the evolutionary relationship between human and fly have been performed. In this study alternative methods could also have been used instead of HMM. For example, protein alignments between all human SLCs and the SLCs listed on flybase.org could have been used to reveal protein-protein orthologues, ex- cluding the HMMs. However, there are no publications that have performed a detailed mining to identify the complete SLC family and related, not yet annotated “atypical” SLCs in D. melanogaster. In addition, compared to the clas- sical protein-protein basic local alignment search tool (BLAST), the HMM allows a less time-consuming and more complex search for protein orthologues, usually resulting in a more confident and greater output. We showed that the HMMs successfully identified 55 new protein sequences in D. melanogaster that are related to the SLC superfamily. These sequences would have been missed if the search was only based on the existing data

39 summarized on flybase.org. However, protein-protein alignment was used to identify possible orthologues to the SLC families that are not assigned to a SLC family (No clan). A HMM for each family could have been an alternative, but the advantages would be few. Moreover, one HMM for each of the 65 families could have been used instead of the approached used here, but then it is a risk that the models become restricted and protein sequences belonging to the SLC superfamily might not be identified. The alignments that the HMMs are built on in this paper are based on the Pfam classification system, a system that also is based on multiple sequence alignments and HMMs. Hence, the observed parameter in our model becomes the conserved regions of the SLCs. Some of the Pfam clans contain only proteins belonging to the SLC family, while others have both SLCs and other protein families with similar conserved transmembrane regions. In addition, the number of members in each SLC family varies tremendously, which could affect the sensitivity of the model resulting in false-negative hits. For example, the MFS clan contains the largest number of SLCs, both regarding families and protein sequences, and the HMM failed to identify two already listed orthologues to the SLC29 family. Meanwhile, only two SLC families, SLC50 and SLC54, together with other protein families populate the MtN3-like clan. It is possible that the alignment, which this clan is based on, varies in a higher degree compared to e.g. the alignment of the MFS clan. Furthermore, SLC50 has one member, while SLC54 family has three members, if the transmembrane regions between this two families varies, it is less likely that the HMM will be based on the SLC family with fewer members. If this would be the case, it could explain why the model successfully identified four new SLC orthologues to the SLC54 family, while it failed to identify the only and already known orthologue to the SLC50 family.

Key findings . There are 381 orthologous SLC proteins in D. melanogaster. . 55 protein sequences were identified in flies that have not previously been classified as SLCs. . 54 families have at least one orthologue in flies. . The largest cluster of SLC families in flies populates the MFS clan. . The SLC25 family was identified to be the largest SLC family in D. melanogaster. . The amino acid sequence identity within the SLC families are low compared to other superfamilies of membrane-bound proteins.

40 Paper II: Glucose availability alters gene and protein expression of several newly classified and putative Solute Carriers in mice cortex cell culture and D. melanogaster. The body is dependent on anabolism and catabolism of nutrients to produce energy, where carbohydrates are one macronutrient with great importance for several metabolic pathways and for energy homeostasis [178]. Glucose is the main source of energy for several organs in the body, however, several other macronutrients such as fats can also be utilized to provide the body with energy during periods of starvation [79, 179]. Each tissue has their own specific metabolic need and regulatory pathways and mechanisms (gene regulatory networks, hormonal regulation and signaling pathways) have developed to coordinate the metabolic processes [68, 79]. The brain consumes around 20 % of the ingested glucose even if it only makes up a small part of the total body weight, but it can also use ketone bodies during fluctuating glucose concentrations [78]. A couple of years ago, a group of putative (“atypical”) SLCs were presented [52, 53], and recently some of these have been included in the SLC superfamily (slc.bioparadigms.org). The majority of the putative SLCs belong to the MFS, an evolutionary old superfamily that from the beginning was believed to be sugar transporters [50, 54]. Unfortunately, the knowledge about these putative SLCs are limited and most of them have unknown substrate profile but for a few the messenger RNA (mRNA) expression and protein localization have been established [60-65, 91]. Interestingly, some of these putative SLCs were found to be affected on transcriptional levels by amino acid starvation in vitro [52] and by complete starvation in vivo [60, 62, 64, 65]. However, even if several putative SLCs are predicted sugar transporters or predicted to be closely related to sugar and its transport, information regarding transcriptional and protein changes as a response to sugar starvation is missing for these transporters. As mentioned before, gene regulatory systems have developed to regulate intracellular metabolism [68] and these pathways are conserved. In D. melanogaster the glucose sensing key transcription factors MondoA (Mlxip) and Max-like protein X (Mlx) are conserved and have a vital role in survival on carbohydrate enriched diets and regulating feeding [84]. However, several steps in the glucose sensing pathway, both in vertebrates and invertebrates, are still unknown. In addition, putative SLCs: MFSD1 (mrva, CG12194), MFSD6 (jef), MFSD8 (Cln7), SLC22A32 (rtet), MFSD11 (CG18549),

41 MFSD14AB (CG11537, CG5078, CG18281, CG17637), SLC63A1A3 (spin), SLC22B13 (CG3168), SLC22B45 (CG4324) and UNC93A (CG4928), are conserved in D. melanogaster. Hence, flies subjected to complete starvation and diets with varying sugar concentrations were used to study the transcriptional effects on the conserved putative SLCs. To map the gene alterations, relative mRNA expression was measured by quantitative real time polymerase chain reaction (qRT-PCR). After 3 h of starvation approximately one third of all monitored genes were altered and many of them, both in mouse and fly, returned to the same expressional levels as controls upon glucose refeeding, Figure 9.

Figure 9. mRNA alterations in cortex cell cultures and fruit flies subjected to 3 h of glucose starvation and refeeding. Gene names in similar color indicate groups of genes that are regulated in the same way. Blue display downregulated genes and red indicate upregulated genes. More than half of the monitored genes were not altered, displayed in the grey box.

After 12 h of starvation 18 of 46 monitored genes were found to be altered and approximately 61 % of them returned to the same expressional levels as controls upon glucose refeeding. Meanwhile, 39 % was found to be both altered by starvation and glucose refeeding, and three genes were only affected by glucose refeeding, Figure 10. Interestingly, the observed alterations between the two timepoints differed. For example, at timepoint 12 h, a greater number of genes were found to be altered both after starvation and refeeding, and more genes were regulated in the same direction. Moreover, there are genes that were altered after 3 h but not 12 h and vice versa. Only Mfsd8 and Mfsd10 had similar mRNA alteration

42 after glucose starvation in both primary cortex cells and D. melanogaster, which could indicate that these genes have a conserved function.

Figure 10. mRNA alteration followed by 12 h of glucose starvation and refeeding in primary cortex cells and D. melanogaster. Approximately half of the monitored genes were found to be altered. Gene names in similar color indicate groups of genes that are similarly regulated. Blue display downregulated genes and red indicate upregulated genes.

After glucose deprivation, 14 of the putative SLCs displayed alterations in mRNA expression in mice cortex cultures. In D. melanogaster, seven of the putative SLCs were altered due to a low sugar diet. Glucose deprivation lead to an early transcriptional change among both putative SLCs and the general metabolic targets that were chosen, Figure 11. It was not possible to compare the transcriptional outcome between deprivation and starvation in D. melanogaster since the flies were deprived of sugars for a much longer period of time, five days, before genes were monitored. However, it is interesting that some of the putative SLCs reacts to both glucose deprivation and starvation with similar alterations in both mouse cortex cells and D. melanogaster, e.g. Mfsd1 and CG12194; Mfsd11 and CG18549; and Mfsd14a/b and CG11537. This could suggest that these genes have similar functions in both species. How- ever, more research is needed to evaluate these findings. In cortex cultures subjected to glucose starvation, the total mTOR expression was altered in similar patterns that has been reported before during starvation and refeeding [180], and the observed response indicate that the metabolic machinery controlled by mTOR is activated. Glucose starvation altered the protein expression level of a few putative SLCs, but the observed regulation was not directly linked to the mRNA expression changes. MFSD14A was the only putative SLC that both had an increased protein expression and

43 changed protein expression pattern compared with the control, indicating either that it moves or that it is expressed in other cellular compartments but not used when cells are maintained in high glucose concentrations. MFSD14A also remained upregulated in the core cytoplasm and was higher expressed in both the whole and outer cytoplasm after glucose refeeding.

Figure 11. Comparison between glucose deprivation and starvation and between the two timepoints, 3 h and 12 h, for both conditions. Several genes were found to be altered during a specific timepoint e.g. Slc60a2 and Mfsd9. A few genes were also found to be altered by both conditions during at least two or more timepoints e.g. Mfsd14a and Mfsd14b. A majority of the orthologous genes in D. melanogaster were found to be altered after both 3 h and 12 h of starvation. Deprivation (low sugar diet, light brown box) in flies were performed for a longer time period and hence cannot be compared with the deprived cortex cultures.

Mechanisms are initiated in the cortex cells to maintain homeostasis during glucose starvation, but most likely other substrates are utilized to produce energy. Even if the function of many putative SLCs are unknown, we observe clear changes for many of them during deprivation and starvation on both mRNA and protein level. During deprivation, the most transcriptional changes are observed. Since the cells still have access to glucose but in a lower extent, it could be that the expression of transporters needed to increase sugar intake are altered to maintain the intracellular glucose homeostasis. However, it is most likely that these putative SLCs are not the primary units needed for early

44 regulation since most alterations occur late (12 h after deprivation). When there is no glucose available neuronal cells need to rely on other substrates to produce energy as well as release of glucose and lactate from astrocytes [78, 79], the gene and protein regulations observed suggest that some putative SLCs are involved in alternative pathways to maintain the cellular metabolism. Since the focus of this study was putative SLCs, and their reaction to glucose availability, it is most likely that alterations of several crucial known SLCs and other genes that could explain which cellular processes that are initiated and maintained due to glucose fluctuations were missed. Some general metabolic genes and SLCs known to be involved in cellular processes linked to metabolism were monitored; Gapdh and Slc2a3 were induced and Slc16a3 was reduced during glucose deprivation, suggesting that the cultures still perform glucose metabolism, while no alterations are observed for neither gene during glucose starvation. In theory, the Slc16a3 could be regulated, since there are evidences that lactate is used by the brain as an energy reservoir during starvation [79]. However, since the cortex cultures have both neurons and glial cells, it is possible that cell specific alterations are lost.

Methodological considerations The methods are applied on both unicellular and multicellular models, which could aid in the overall understanding about the role and involvement of the putative SLCs in glucose metabolism. However, it is important to note that neither of these two systems reflects the metabolic processes that occurs in the human body. The cell cultures lack important inputs that have great impact on initiation and termination of metabolic processes such as hormones and nutrient storages. Also, several elements, e.g. B27, added to the culturing media have been found to protect neurons from metabolic stress [181]. Another limitation about the study is that the results are always compared to standard culturing protocols; the cells are grown in a media with high glucose concentration. Therefore, the results obtained for glucose deprivation does not represent a physiological glucose deprivation in humans. Instead the glucose deprivation represents the normal physiological glucose concentration. Gene and protein levels that we observed could therefore indicate the “normal”, healthy state. However, once more the cell cultures lack vital factors that affects gene and protein expression, but overall, the results give an indication on the regulation that occurs during variations in glucose concentrations.

45 Many metabolic pathways are conserved between man and fruit fly; hence fruit flies make an excellent model when studying the connection between metabolism and specific genes, but it has its limitations. For instance, not all putative SLCs are conserved which makes it impossible to study some of the genes that were found to be affected in cell cultures from mice. There are also anatomical and physiological differences, meaning that a monitored result in fruit flies do not necessarily reflect the function of the human orthologue. In addition, man and fruit fly have evolved to survive in different environments suggesting that not all conserved genes execute the exact same function. Moreover, there are differences in the experimental setups, making it more difficult to translate the findings to the physiological role of several of the putative SLCs. There are several methods to retrieve information about mRNA expression in tissues, e.g. in situ hybridization, microarrays, RNA sequencing and RT- qPCR. Here, RT-qPCR has been used to determine the relative mRNA expression in cortex cells and fruit flies because of its sensitivity to detect these putative genes and due to other factors such as high output, less time-consuming and cheaper compared to in situ hybridization. To gain information about the relative mRNA expression of a target gene, good and reliable reference (housekeeping) genes (RG or HKG) are needed [182, 183]. There are several HKGs used in the molecular biology field of research, which are assumed to have stable mRNA expression level in all cell types during all conditions. Lately several HKGs have been discredited [184] and found to be more or less stable depending on the samples used [185, 186]. The dynamic of the cell is ever-changing and depending on the conditions the cell is exposed to the genetic information is processed in various ways. The HKG Gapdh has long been widely used due to its ubiquitous expression in different tissues, and it has been found to be a highly stable HKG in e.g. HEK293T cells. Gapdh is involved in cellular metabolism, meaning that this HKG can be and is most likely affected by the nutritional state of the cell. Gapdh was not used for normalization in this study and we confirmed the sus- picion when Gapdh was found to be altered after glucose starvation and deprivation. Moreover, there are two orthologous genes to Gapdh in flies, and we found that one copy (Gapdh1) was affected by starvation and sugar diets, but not the other copy (Gapdh2). Some argues that if the same HKG is used to normalize all experimental samples it should not affect the result. This is true when focusing on normalized mRNA levels and you want to know the difference between samples that have been treated the same way; but to be able to know the accurate mRNA levels the expression level of the HKG will

46 definitely matter. Algorithms have been developed to manage this issue, for example the GeNorm protocol [187], which was used in this study. The GeNorm software calculates HKG variability between samples from a set of tested candidate genes and determine the most stable ones. Furthermore, genomic DNA (gDNA) can be used to determine the total mRNA expression of a target gene during a certain condition. Immunostaining is a way to visualize the protein expression and localization in tissues and within cells. The techniques rely on the specificity of antibodies, which also is the limiting step of this technique. Antibodies are produced to recognize a small part of a target protein, the epitope, consisting of approximately four to six amino acids. Antibodies tend to cross-react with epitopes of similar shape as the target epitope, which lead to misleading immunostaining results if the study is not carefully conducted. The more used an antibody is, the more validation has been performed to demonstrate the specificity of that antibody. Unfortunately, this means that for less studied proteins such as the putative SLCs there are few or no antibody studies available, and some putative SLCs do not even have a targeting antibody. Therefore, not all putative SLCs have been studied.

Key findings . The putative SLCs have binding site for transcription factors known to regulate gene expression during nutrient fluctuations. . Glucose starvation affected mRNA and protein expression for MFSD14A and UNC93A in cortex cells. . In flies, glucose starvation altered all monitored genes except mrva and CG4928. . Glucose refeeding normalized the mRNA and protein expression of all putative SLCs, except for MFSD11, in cortex cells. . In flies subjected to glucose refeeding, the mRNA expression was altered for Cln7, rtet, CG18549 and CG11537. . Glucose deprivation altered the mRNA and protein expression for MFSD1, SLC60A1, MFSD11, UNC93A. . Low sugar diet altered all orthologues except jef in fruit flies. . MFSD11 was the only putative SLC that was found to be altered by all conditions in both cortex cells and D. melanogaster. . MFSD14A was the only putative SLC found to have both increased protein expression and the potential to move from the suggested cellular location (Golgi) as a response to glucose fluctuations.

47 Paper III and Paper IV: Characterization of to the putative Solute Carrier UNC93A. The putative membrane-bound UNC93A was recently found to be closely related to SLCs [52, 53, 65]. Similar to SLCs of MFS type, UNC93A is conserved in several kingdoms [52, 65]. Unc-93 was first identified in a genome- wide screening in C. elegans. It is suggested to be a regulatory protein of potassium channels and is one of five genes identified to be involved in muscle coordination in C. elegans [188, 189]. In mammals, there are two UNC93 proteins, UNC93A and UNC93B1 [53]. UNC93B1 is reported to be a regulator of toll-like receptor (TLR) signaling and important for the trafficking of TLRs from ER to endolysosomes [190]. For example, UNC93B1 is reported to be important for TLR5 expression and function at the plasma membrane [191]. There have been several attempts to discover the role of UNC93A. For example, Liu et al. (2002) studied an allele loss on chromosome 6 linked to sporadic ovarian cancers where UNC93A was identified within the interval of the allele loss but was concluded not to be a tumor suppressor gene. However, they discovered that UNC93A was located to the plasma membrane [192]. UNC93A has also been identified in a family-based association study, as a candidate for pulmonary function in a northeast Asian population [193]. How- ever, exactly how UNC93A affects the pulmonary function is unclear. A recent study identified UNC93A as a metabolite-associated locus in Chronic Kidney Disease patients [194], but also here the role of UNC93A still remains uncertain. A couple of years ago, Unc93a expression was found to be altered in mouse hypothalamic N25/2 cell lines subjected to amino acid starvation [52]. Moreover, in A. aegypti, Campbell et al. (2011) studied conserved endosomal proteins involved in infection and found that viral RNA accumulates in UNC93A-silenced mosquitoes [195]. In D. melanogaster, the homologues protein to UNC93A (CG4928) has been found to be one of several genes that confers resistance against manganese toxicity [196]; are linked to the amount of developed hemocytes [197] and to be highly expressed in the MTs [198]. However, no further studies on mammals or insects have investigated the expression and function of UNC93A in detail. In Paper III, an expressional characterization of UNC93A was performed in mice, Figure 12, and in Paper IV a functional characterization in D. melanogaster and human cell lines was performed. UNC93A was confirmed to be conserved in several proteomes and it was predicted to have twelve transmembrane helices packed to form a globular protein similar to other SLCs of MFS type [57].

48 UNC93A displayed a dotted staining pattern in cell bodies and projec- tions in brain tissue from mice and in the human cell lines. The dotty staining pattern resembles the im- munoreactivity of UNC93B1, which could suggest that UNC93A, like UNC93B1, is involved in protein trafficking; maybe retrieving TASK channels [199]. However, the expression pattern observed for UNC93A is different compared to the TASK channels [200, 201], indicating that UNC93A could have another role in murine neurons sep- arated from TASK channels. In addition, Unc93a was found to be expressed both in central and peripheral tissues in mice, while it was only found peripheral in D. melanogaster. This could suggest that one of the orthologous proteins have, after the evolutionary separation between insects and mammals, evolved and perhaps adapted to a more complex system and hence de- Figure 12. In paper III mice were used to veloped additional functions. study gene and protein expression of In Paper IV, CG4928 was indeed, UNC93A. It was found to be expressed both as stated by earlier reports, found in neuronal and peripheral, with the highest expression in kidneys and intestine, and the MTs of fruit flies. It was mainly was also found to be altered after 3 and 12 located to “larger” cells in the lower h of altered amino acid availability in pri- and main segment of MT: segments mary cortex cells. UNC93A displayed expression in neurons close to the plasma responsible for secretion of water, membrane but was not found to overlap sodium, potassium and chloride, as with the synaptic membrane marker well as reabsorption of water and so- SNAP25. dium [202, 203]. Similar expression is observed for two Gal4 lines expressed in the Principal cells [204], suggesting that CG4928 is expressed in the Principal cells. To further study the function of CG4928, knockdown with RNA interference was performed, which

49 reduced the expression with 50 %. Knockdown of CG4928 caused edema, Figure 13, accumulation of a yellow-colored pigment in the MT and reduced the secretion rate in the adult fly, all pointing to a malfunctioning MT.

Figure 13. Both male and females flies with reduced CG4928 gene expression (CG4928 knockdown) developed edema three days after eclosion compared with controls. The figure depicts five days old males (top row) and females (bottom row).

In addition, lower levels of potassium in the hemolymph and MTs were observed, establishing that the potassium balance was disturbed. Reduction in secretion rate has also been observed for two potassium channels, Irk1 and Irk2, channels that are responsible for transepithelial potassium fluxes in the principal cells [205]. This suggest that CG4928 could be important for proper potassium flow since this ion is of great significance for the MTs to function properly [206]. The edema was minimized using amiloride, but not mannitol and furosemide. Amiloride blocks epithelial sodium channels (ENaC), thereby inhibits sodium reabsorption leading to excretion of sodium and water from the body, while reducing the potassium excretion [207]. Amiloride has been shown to have similar effects in D. melanogaster via the Na+/H+ exchanger (NHE) family [208, 209]. It is likely that the phenotype was minimized by inhibiting the transepithelial sodium secretion and thereby affecting fluid secretion. Manni- tol is a poorly soluble sugar alcohol that can be metabolized in flies, but not

50 in humans. It is therefore possible that, in higher concentrations, mannitol act as a weak osmotic molecule within the digestive tract, increasing the excreted volume of water via feces. Furosemide causes sodium, potassium and chloride loss in the urine, leading to lowered water reabsorption by inhibiting the Na+/K+/Cl- cotransporter (NKCC) [207]. In flies, NKCC is located on the ap- ical membrane compared to mammals where it is located on the basal membrane. In flies, the inhibition of NKCC lead to less potassium uptake into the MTs [210] and, in contrast to mammals, primary urine production in flies is potassium dependent, hence the fluid secretion will be decreased, and the edema will remain. Immunostaining and PLA measurement revealed that UNC93A and the TASK channels KCNK3 and KCNK9 co-localized in wild-type and CG4928:Flag transfected cells, indicating an interaction. These results suggest that UNC93A potentially have a conserved function as a regulatory protein of TASK channels. CG4928 overexpressing cells were used to determine substrate profile and other potential functions. Six different sugars and 12 different amino acids/polypeptides were tested as potential substrates. No transport could be observed for the transfected cells, suggesting that it is not a cotransporter of sugars or amino acids. On the other hand, transfected cells were found to decease more rapidly compared with controls, and measurements revealed that transfected cells had high apoptosis and cytotoxicity, and low viability.

Figure 14. CG4928 overexpressing cells were found to have induced cell death (apoptosis) and cytotoxicity when cultured in regular DMEM media. Apoptosis was hindered by high sodium media, worsen by high potassium concentration and low sodium levels and unchanged by a low potassium concentration. Cytotoxicity was hindered by high sodium levels, worsen by high potassium levels and unchanged by low sodium and potassium concentrations.

51 To mitigate the cell death, overexpressing cells were cultured in media with high and low concentrations of either sodium or potassium. Both apoptosis and cytotoxicity could be hindered by a high concentration of sodium, while higher concentration of potassium induced the apoptosis and cytotoxicity even more, Figure 14. These findings formed a second theory, that CG4928, and maybe UNC93A, affects the membrane potential. Indeed, the basal membrane potential was found to be disturbed in CG4928 overexpressing cells, Figure 15.

Figure 15: The membrane potential of CG4928 transfected cells were found to be reduced, indicating that the cells are less depolarized. The fluorescent intensity indi- cates the strength in membrane potential. The results indicate that CG4928, and possibly also UNC93A, is important for maintenance of potassium concentrations inside and outside the cell by regulating potassium channels, Figure 16.

Figure 16: The hypothesis regarding the function of CG4928, and possibly UNC93A. Most likely CG4928 is a regulatory protein of the TASK channels, and hence indi- rectly affect the potassium flows over the plasma membrane in the cell. Credit to Som- ersault 18:24 (www.somersault1824.com) for figure components, shared under creative common license.

52 Unfortunately, it still remains unclear how the regulatory function between CG4928 and TASK-1 (KCNK3)/TASK-3 (KCNK9) occurs. It could be that CG4928 regulate these channels by affecting their i) function directly by acting as a subunit of the channels in a similar fashion as the SLC3 family does for the SLC7 family; ii) localization by transporting them similar as UNC93B1 transports TLRs from the ER to the plasma membrane; iii) insertion in the membrane and/or retention to the membrane acting as a anchoring protein.

Methodological considerations There are limitations with this characterization of UNC93A and more experi- ments are needed to understand its function. In Paper III, mRNA expression was performed on in vivo and in vitro samples. The set-up of the starved conditions was different between the two systems and the results were not com- parable. It would have been interesting to see if the gene regulation occurs in both unicellular and multicellular models. In addition, several important aspects of the physiological state of the mice that were subjected to 24 h of starvation was not measured and therefore information regarding stress level and immune response is missing and we cannot tell if the gene regulation observed is only due to the metabolic challenge. The protein localization was studied using tissue sections and primary cultures from mice. Two antibodies were used, and the results obtained were similar for both antibodies. However, the observed immunostaining was not completely in line with earlier published work where UNC93A was established to be expressed in the plasma membrane. The difference observed could be due to the set-up of the protein visualization. When transfecting cells with an engineered gene, it is a risk that the protein ends up in organelles and membranes where it is not expressed [211]. Furthermore, HeLa cells were used when detecting the plasma membrane staining, an immortalized cell line, while we used primary cells from mice. Since cell lines can have alterations in phenotype compared to primary cell lines, it is possible that this result is specific for that experimental setup. There is also the possibility that the protein expression pattern differs between human and mice. Moreover, it is reported that the location of the epitope affects the immunostaining [211]. This was also seen in Paper III where the antibody targeting the C-terminal gave a better signal compared with the antibody targeting the N-terminal. In Paper IV, D. melanogaster and two human cell lines were used to study the function of UNC93A. In this study there are several limitations that should be considered. First, the phenotype of the flies occurred

53 when maintaining the flies at a temperature higher than 25 °C. Higher temper- atures usually cause dehydration, which alters the ion and fluid homeostasis of the flies. Therefore it is impossible to tell if the phenotype is a combination of dehydration and CG4928 loss; only a cause due to CG4928 loss or a sys- temic response to survive the reduction of CG4928 expression. Also, the phenotype observed for fruit flies has not been reported in humans and UNC93Ahas not been linked to renal function in humans so far, hence it is possible that the phenotype is fly specific and that humans have other genes that can compensate the loss of UNC93A. Second, the biological variation within samples used when performing measurements of ions and the RNA sequencing could possibly lead to inaccurate results and biological differences could be missed. Third, the protein localization of CG4928 in the Malpighian tubules could not be examined, without knowing the exact location it is difficult to pinpoint its function. Fourth, the cell lines used in the paper are immortalized cell lines, meaning that the expression and function of proteins might be altered. Fifth, we could not identify a substrate profile of CG4928 or UNC93A even though it has similarities, structure wise, to other SLCs. Information about substrate profile and transport mechanisms can shed some light on the role in maintaining substrate homeostasis in organs and cells. Cells do not only express one SLC that perform the transport of specific substrates, the cell expresses several, and many are even co-localized to the same compartment of the cell and needed to be co-expressed to function properly. It is therefore crucial to verify findings from transport assays and to use good controls. A traditional way to study the substrate profile of transporters has been to overexpress the protein in oocytes from X. laevis and perform radiolabeled uptake and efflux assay and electrophysiology. This has been done for several amino acid transporters from the SLC38 family, where members expressed both in the plasma membrane and organelles have been studied [212-214]. Another way to investigate the substrate profile of transporters is SSM-based electrophysiology. This is a sensitive, “cell-free” and useful method to determine the substrate profile of electrogenic transporters positioned in membrane fragments from the plasma membrane and/or intracellular membranes. Compared to oocyte-based electrophysiology, this method is less time-consuming, and several substrates can be tested relatively fast and not as much material is needed. The SSM-based electrophysiology does not gener- ally have limitations regarding the type of the transporter, but other techniques can have advantages in contrast to this technique. In comparison to clamping techniques, a potential cannot be applied and characterization of transporters that rely on a membrane potential is troublesome. Also, artificial currents can

54 arise during solution exchange [215] and by charged substrates e.g. some amino acids. Even though advancements to determine the substrate profile have been done, it is still difficult and time-consuming to determine substrates for putative transporters especially if there is no hypothesis about the transporters function. Both the CG4928 vector and control vector were designed to express GFP to validate transfection, but none of them expressed GFP. Instead, insertion was confirmed using anti-GFP immunostaining that showed a higher GFP signal in transfected cells. In addition, WB showed that the anti-UNC93A immu- noreactivity was higher in the cells transfected with the CG4928 vectors. The lack of GFP signal made it difficult to perform patch-clamp electrophysiology on living cells. It is still possible that UNC93A is a transporter that is either ion:ion coupled or a facilitator. Last, CG4928 was determined to co-localize with two TASK channels and the membrane potential was found to be disturbed, suggesting that the theory about UNC93A being a regulatory protein holds true. However, we were not able to determine the exact regulatory mechanism and further assays need to be conducted. It is plausible that the tools available today are not enough to unravel the true function of UNC93A, but hopefully in the future we will know what it does and its importance in human physiology.

Key findings . UNC93A is evolutionary conserved. . UNC93A has twelve transmembrane helices and folds like other MFS proteins. . UNC93A is expressed in the brain and peripheral tissues. . Unc93a respond to nutrient availability in in vivo and in vitro. . The promoter region of Unc93a express binding sites for transcription factors that are active during nutritional challenges. . Knockdown of CG4928 cause renal dysfunction. . Potassium levels are affected by reduction of CG4928. . UNC93A co-localize with KCNK3 and KCNK9. . CG4928 overexpressing cells have induced apoptosis and cytotoxicity, which could be hindered when cells were maintained in high sodium media. . CG4928 does not act as a cotransporter for sugars and amino acids. . CG4928 overexpression disturb the normal membrane potential.

55 Conclusions

Today’s technology has increased the number of publications about SLCs. D. melanogaster is an old model organism that has been extensively used in many fields of research including genetics, cancer and neurobiology. This system can be of great importance also in the study of orphan SLCs, but also to find other cellular pathways that already characterized SLCs are involved in. The SLCs are well conserved in D. melanogaster, and 83 % of all SLC families have at least one orthologue that can be used with the sophisticated tools available today to break the path for new findings about the SLCs. The superfamily of SLCs has grown over the past years, and it is highly expected that more protein sequences will be added in the future. Some of the putative SLCs have already, within a short time frame, been classified as SLCs. So far there are a few publications about gene expression and protein localization. The metabolic study demonstrates that some of these putative SLCs are of significance for nutrient homeostasis and that they have a potential role in metabolic pathways, while others probably are needed for other important cellular processes that have not yet been uncovered. However, more research is needed to classify these putative transporters as SLCs. We have begun to understand the importance of UNC93A in unicellular systems and in D. melanogaster, where it is established to be of importance for the osmotic and potassium balance. Further research is needed to reveal the physiological function of UNC93A in humans, to know if it is linked to disorders and can be used as a therapeutic target for treatment. However, while several clear biological aspects of UNC93A was identified, the exact function of transporter proteins is cumbersome to find and more research about these transporters is needed to fully understand their mechanistic role and their association and/or involvement in health and sickness.

56 Perspectives

A lot about the human body is known but there are still matters to unravel and obstacles to overcome to completely understand the physiological network that operates non-stop throughout our lifetime. To understand the importance of orphan transporters and their involvement in cellular processes, we need to discover more about them. For the putative SLCs that are not yet classified as SLCs this is a challenge. Putative and orphan SLCs are evolutionary conserved [52, 59], and functional studies can be performed in model organisms. A good starting point to study some of these putative SLCs could be to perform knockdown/knockout methods. In addition, there are modern tools available to uncover molecular pathways and diseases the putative SLCs are associated and/or involved in. Expression and substrate profiles are of great importance to be able to understand their role in cellular processes and therefore it is often required to combined different models and techniques to understand the putative SLCs. We have showed that it is a vital protein for normal membrane potential and normal renal function in flies, and regardless the importance of these findings we still do not know how UNC93A exerts its function. Does UNC93A cause the same phenotype in other model organisms that are closer to humans; is it a regulatory protein that upon binding actives and control TASK channels; does it, similar to UNC93B1, transport TASK channels to their correct position and retain them there; does it affect all TASK channels the same; is UNC93A a transporter and if so what does it transport? There are many questions to be answered and one desire I have is to discover the detailed function of UNC93A and fully characterize and deorphanize this transporter. In an attempt to answer those questions, I would first perform other types of transport assays to confirm if UNC93A does or does not act as a transporter. So far both oocyte-based and SSM-based electrophysiology have failed to reveal a substrate profile of UNC93A. However, it is still possible that it is an electroneutral uniporter, which most likely are not detected by the SSM-based electrophysiology. There are possibilities to detect a pre-steady-state current that is needed to measure alterations in the kinetics of uniporters. If possible, there would also be of great use to produce a viable cell line that expresses

57 GFP-tagged UNC93A. This cell line could be used to perform live-cell patch- clamp electrophysiology to study if UNC93A is an ion transporter. If UNC93A is a transporter on its own, the substrate profile can aid in understanding several of the findings in Paper II–IV, and it can give us directions in further studies, knowledge about roles in diseases and etcetera. Third, I would like to verify the findings of Paper IV in primary cells derived from kidney. Next, I would like to reduce the expression of UNC93A using siRNA and/or CRISPR/CAS9 editing in both cell lines and primary cell cultures to study what happens with the TASK channels when lacking UNC93A and also investigate how the membrane potential, potassium flow and cell survival are affected. It would be interesting to see if the cells display an opposite behavior, i.e. increased cell survival, reduced cytotoxicity and higher membrane potential. By using immunocytochemistry and image analysis the protein expression of TASK channels can be studied and reveal if the expression increase or decrease at the plasma membrane or intracellularly. To use more organelle and cell structure protein specific markers and perform double immunostaining with UNC93A and TASK channels in cells subjected to different electrogenic stimulus such as varying salt concentrations in the culturing media could also help us understand the linkage between these proteins. Such findings would confirm that the interaction is not only occurring in immortalized cell lines and could provide more information about the role of UNC93A in the regulation of TASK channels and if these proteins are regulated and move together in the cell. The interaction, and how UNC93A regulates TASK channels, would be of a great interest to study. To identify UNC93A-associated polypeptides in cell cultures that stably express tagged UNC93A and a mutant UNC93A could be produced and used to conduct a preparative immunoprecipitation of the tagged proteins. This could tell us if the TASK channels coimmunoprecipitate with UNC93A. To verify interactions tagged TASK channels can be generated and co-expressed with the UNC93A and mutant UNC93A. Immunoblotting on the immunoprecipitate from these cells can confirm findings from the previous described experiment [190]. It would also be interesting to figure out which part of the TASK channel that UNC93A interacts with. This is a more com- plicated experiment but can be done by inducing point mutations within the sequence of a certain TASK channel or generating a chimeric TASK channel; if UNC93A fails to interact with e.g. TASK-5, parts of its protein sequence can be used to alter specific regions of the other TASK channels. Such results would help us to understand if and how UNC93A interacts with TASK channels. To answer the question if the phenotype can be reproduced in other

58 animal models or if other phenotypes occurs, a mouse model of UNC93A could be produced. With this mouse model UNC93A can be studied by performing knockout ubiquitously and organ and/or cell-type specific to reveal the physiological role of UNC93A. Last, a very important finding, which would help us understand the potential role UNC93A can have in diseases and evaluate if it can be used as a pharmaceutical compound, is to uncover its crystal structure using e.g. crys- tallography. My wish is that the work presented in this thesis can be used further to increase our knowledge about the SLC family, as well as increase the knowledge about the physiological role many of these putative SLCs have, including UNC93A. The importance of the putative SLCs are indicated by their cellular localization in various cell types of the body and evolutionary history, additional knowledge about their function could provide to the bigger picture of understanding the human body in health and sickness. I remain op- timistic, that one day, the scientific area will unravel these questions and I hope that the discoveries will lead to better health care systems and therapeutic treatments.

59 Populärvetenskaplig sammanfattning

Den mänskliga kroppen består av organ och vätskor som bildar komplexa system och kretslopp. Organen i sin tur består av specialiserade små enheter som kallas celler, och inuti dessa finns små ”inre organ” som benämns organeller. Cellerna och organellerna har olika form, funktion och livslängd, men vad de har gemensamt är att de utför komplicerade och energikrävande processer da- garna i ända. Cellen, och alla organeller, omsluts av en barriär (membran) som beskyddar cellen från den yttre (extracellulära) miljön och infektioner, samt hjälper cellen att bibehålla ett balanserat inre (intracellulärt). Även organellerna omges av ett skyddande membran. Det är viktigt att det sker ett utbyte av information mellan celler (signalering), upptag av näringsämnen så som kolhydrater (socker), protein (aminosyror och polypeptider), fett (fettsyror), läkemedel (till exempel antibiotika) och salter (joner), en ämnesomsättning och utsöndring av avfallsprodukter och giftiga ämnen. Detta upprätthålls via ”dörrvakter”, så kallade membranbundna proteiner, vilka sitter förankrade i membranet. Dessa proteiner hjälper cellen att skicka och ta emot signaler (re- ceptorer), ta upp och utsöndra ämnen (transportörer), samt bryta ner och bygga upp byggstenar (enzymer). Många membranbundna proteiner är associerade med sjukdomar till exempel cancer, diabetes och neurologiska sjukdomar. Det är därför en stor del av tillgängliga läkemedel har sin verkan via dessa membranbundna proteiner. Transportörer förflyttar olika molekyler över membranet. Detta sker an- tingen passivt, genom att förbruka energi eller via koncentrationsskillnader över det cellulära membranet. Transportörer som utnyttjar koncentrationsskillnader kallas för sekundärt aktiva transportörer. I oss människor är ”solute carriers” (SLC) den största familjen av sekundärt aktiva transportörer med över 430 medlemmar fördelade i 65 familjer. Trots att dessa har en otroligt viktig funktion i cellen så saknar vi fortfarande kunskap om många av dem. Till exempel så vet vi inte var i kroppen de finns, vad de gör eller hur de ser ut. Det är därför mycket viktigt att ta reda på så mycket som möjligt om dessa transportörer, vilket ökar vår kunskap om deras roll i sjukdomar och ifall att de kan utgöra en viktig del av den framtida läkemedelsutvecklingen.

60 För att lära oss mer om SLC-familjer krävs modeller som representerar den inre miljön i oss människor, till exempel cellkulturer och modeller. Möss och bananflugor är två exempel på viktiga modeller som används i forskningen om kroppens funktioner (fysiologi) och sjukdomar samt läkemedelsanvänd- ning och effekter. Möss har trots sin storlek och utseende stora likheter med människan vad avser ärftlighet (genetik) och fysiologi. De har därför varit ovärderliga när det kommer till forskning som undersöker evolutionärt beva- rade biologiska processer i kroppen. Bananflugor har använts som modell- organism i över 100 år, och cirka två tredjedelar av människans arvsmassa (generna) återfinns i bananflugan. Detta har lett till flera viktiga framsteg inom forskningen om bland annat utvecklingen av embryot under graviditeten och underliggande orsaker till sjukdomar. Det är dock viktigt att komma ihåg att evolutionen har haft stort inflytande på hur vi har utvecklats och anpassats för att leva i helt olika miljöer. Dessu- tom kan både möss och bananflugor anses vara mindre pålitliga modeller vad gäller mänskliga sjukdomar på grund av att kopplingen mellan gener och hur dessa utvecklar sig till en sjukdom kan skilja sig åt. Därtill har det förekommit att läkemedel visats ha goda effekter i modellorganismerna medan de i män- niska har gett förödande biverkningar, och tvärtom. Sedan finns det grundläg- gande skillnader mellan människan och dessa modellorganismer, till exempel inom anatomin, organfunktion, fysiologi och antalet kromosomer som förva- rar arvsmassan. Dessutom saknar bananflugor viktiga fysiologiska kompo- nenter och funktioner som människan har, till exempel ett adaptivt immunför- svar, hjärtkärlsystem och en blodhjärnbarriär. Trots dessa begränsningar har dessa modeller varit banbrytande för den medicinska forskningen, vilket visar att modeller behövs för att kunna förstå hur våra kroppar fungerar. I den här avhandlingen har SLC-familjer och transportörer som sannolikt till hör SLC-familjen, men som ännu inte inkluderats i familjen, studerats. Fo- kuset har varit att studera deras evolution, uttryck och funktion utifrån både ett brett och ett smalt perspektiv genom att använda olika modeller såsom da- torbaserade modeller, cellkulturer, möss och bananflugor. Vi har visat bland annat att SLC-familjen är en stor och bevarad genfamilj i bananflugor där 381 gener hade en eller flera motsvarigheter i människan. Vi studerade även hur transportörer som nyligen har inkluderats eller som ännu inte har inkluderats i SLC-familjen påverkas av varierande sockernivåer i celler och bananflugor. Socker är en viktig del av vår ämnesomsättning. Det tas upp av våra celler och omvandlas till energi, byggmaterial och annat som cellerna behöver för att vi ska må bra. Genom att ändra sockernivåerna som cellerna och bananflugorna fick så fann vi att flera av dessa transportörers påverkades på gen- och

61 proteinnivå av den rådande sockernivån. Slutligen så studerades en individuell transportör, UNC93A, mer ingående för att ta reda på dess funktion. UNC93A är en av många transportörer som har stora likheter med andra medlemmar i SLC-familjen, men som ännu inte inkluderats i familjen. Den tidigare forskning på UNC93A är begränsad, men man tror att det är ett regulatoriskt protein, det vill säga att UNC93A påverkar funktionen hos ett eller flera andra proteiner, till exempel andra membranbundna proteiner, genom att interagera med det. Vi visade att UNC93A är evolutionärt bevarad och finns i flera arter av däggdjur men även hos insekter och fiskar. Genom att studera denna trans- portör i möss hittade vi att den finns i flera vävnader i kroppen, framförallt i njurar och tarm men även i hjärnan där den återfinns i nervceller. Vi hittade även att denna transportör regleras av matintag. I bananflugor har UNC93A en intressant roll då den verkar ha en roll i regleringen av salt- och vätskeba- lansen. Bananflugor som uttrycker färre genetiska kopior av UNC93A sväller i kroppen (ödem) och har minskad kaliumtransport, problem att göra sig av med urin och bygger upp ansamlingar av ämnen i njurarna. Genom att an- vända cellkulturer så visade vi att UNC93A troligen interagerar med kalium- kanaler som är viktiga för cellens normala funktion och överlevnad. Det krävs dock mer forskning för att fastställa funktionen hos UNC93A. Har vi rätt an- gående att UNC93A reglerar kaliumkanalerna så krävs det dessutom mer forskning för att förstå hur detta går till. SLC-familjen och andra transportörer är viktiga för vår överlevnad. De har en betydelsefull roll i kroppens funktion, hälsa och sjukdom, samt utgör viktiga läkemedelsmål. Många av medlemmarna i SLC-familjen har även en viktig roll i kroppens metabolism där variationer i matintag, till exempel under svält, påverkar hur mycket det finns av varje transportör. Detta påverkar tro- ligtvis i sin tur cellens upptag, nedbrytning, omvandling och utsöndring av molekyler. Resultaten som presenteras i denna avhandling visar hur viktigt det är med grundläggande forskning för att kunna förstå den mänskliga kroppen. Tyvärr finns det hinder i dagens forskningsutrustning och forskningskapacitet för att kunna kartlägga funktionen hos alla transportörer. Trots det utgör de en mycket viktig grupp av gener att studera för att kunna ge oss en helhetsbild av hur vår kropp fungerar, både när vi är friska och sjuka.

62 Acknowledgements

These projects have been performed at Uppsala University, the department of Pharmaceutical Biosciences, unit Molecular Neuropharmacology. The studies were supported by the Swedish Research Council, the Swedish Brain Foun- dation, the Swedish Society for Medical Research, the Novo Nordisk Founda- tion, Åhlen Foundation, Olle Engkvist Foundation, and Magnus Bergvall Foundation.

First of all, I would like to thank my supervisor Robert Fredriksson. Thank you for giving me the opportunity to pursuit my dream. Since I started working with you in 2013, I have learned so much. I appreciate that you do not mind doing basic lab work when someone needs help. I still remember the day when I was supposed to dissect brain regions for the first time, alone, you immedi- ately put on a protection coat (that was way too big and almost dragged in the floor) and happily did it together with me. Thank you for being that kind of group leader and professor. Also, thank you for allowing me to grow and test my abilities as a researcher: to outline and conduct my projects on my own, but when needed you were there to provide guidance. I have learned that I am more that capable to perform scientific research both during success and strug- gles. Thank you for all years (seven in total) as my supervisor, from bachelor student to PhD.

Thank you to my co-supervisors Helgi Schiöth and Michael Williams at the Department of Neuroscience. Thank you Helgi for giving me the opportunity to do my bachelor thesis in your lab and for introducing me to my main supervisor Robert. Thank you for always pushing me to become better, for introducing me to the research environment, for always coming with good scientific input and discussion. I would probably not have come so far without you. Mike, I do not know where to start and I will probably not be able to convey what you have meant for me and my scientific journey so far, thank you for everything. Thank you for the small everyday discussion and the guidance through my projects. To know that you believe in me mean a lot. With

63 your sense of humor and support, you have been able to turn a bad day into a better one. I am very grateful that I have meet you.

To all past and present members and students of the Molecular Neuropharma- cology, THANK YOU, without you this experience would not have been as good and fun as it was and I wish all of you the best in life. Emelie, I still remember the first day I met you. Early on, you guided me, and I joined you in some of your projects. Your brilliance, joy and enthusiasm for your work is something I admire. Thank you for being an awesome tutor during my first years in research, all the effort you did to make my research, my texts and my planning better paid off big times. I do not only feel comfortable with performing research but also to write in English something I feared for long. Emi- lia, at first we did not have that much contact, our lab-relationship was something that grew over time and now when writing the acknowledgement I do not exactly now where to begin. Just thank you for being you, for becoming not only an amazing colleague but also a great friend, for all the support during my seven years, for being there during mental breakdowns and joy, laughter and struggle. I really, really enjoy doing lab work with you no matter if the experimental set-up work or not. Hopefully, we can continue to conduct studies together in the future as well. Frida, my sweet and strong friend, thank you for being the awesome, crazy and wonderful person you are. I am so happy that you joined as a PhD in Robert’s group. You and I share the longest time together and most likely also the funniest, and probably also our less bright, moments together. We laugh, we cry, we are grumpy and happy together. Your support and the fact that you believe in me means so much, thank you! Thank you to Tanya and Rekha for all the help, laughter, support and Bolly- wood, for always listening when I needed someone to talk to and for just being fantastic co-workers. Erica and Kimia, thank you for all the help, laughter and support, I know both of you will do great! Thank you, Karin for being awesome just the way you are, your positive attitude and just to see the effort you put into your work are so inspiring. Sofie, thank you for all the early lunches (before 10.30), for the daily exercise (I barely can keep your pace and definitely not talk at the same time), for sharing your experience and teaching me new techniques. Also, thank you for always being positive, supporting and helpful, I appreciate that so much, and I am grateful that you have included me in many of your projects!

To everyone at the GU unit, thank you for these years of teaching. I have had so much fun with you and I have learned a lot. Ingrid Nylander (thank you so

64 much for being my examiner), Erika Roman, Anne-Lie Svensson (thank you for always keeping your door open) and Åsa Konradsson-Geuken, you all in- spire me regarding teaching and science. I have never met so many women that have had the impact on me as a person as all of you have had. Fredrik, thank you for all the help regarding teaching and for being the one who han- dled all big and small issues that arose along the way. Linnea, thank you for all the help with Försäkringskassan when I applied for parental leave, and for being a wonderful person. Stina, you are my comfort zone. When I first saw you on my first day at FarmBio, I was filled with relief. You are truly inspiring and have been my rock during these five years, I have learned so much with and from you. You are not only one of the best colleagues I have, you have also become one of my closest friends, and my daughter still asks for the lady with the unicorn hair. She adores you. Nikita, you are such a joy to be around, your laughter and smile are contagious, continue to be exactly as you are! To Lova, I miss bumping into you in the corridor, thank you for all discussions and fun stories, and for you calm and awesome person. I hope you found another pillow and that you get to have it for yourself.

To past and present Postdocs, PhD students and students at the department of Neuroscience, thank you, my first years as a project student and PhD student would not have been the same without you; Vanni, Lyle, Galina, Mathias, Anders, Linda, Philip, Marcus, Markus, Misty, Sofia, Sandra, Anna, Amina, Azra, Jinar, Mehwish and Megha. Especially thanks to Giulia and Nadine for all the help and support during my years as a PhD student, you both are amazing.

Thank you to all PhD students I have met during my journey at the Faculty of Pharmacy. All of you made my “second home” (BMC) feels like home. To all PhD students I have met through my engagement in FDR, thank you for being so ambitious, inspiring and hard-working, whatever comes ahead we have stand strong. To Rebecka, Sara, Jonas, Albin, Karin, Rekha and Stina, we did great!

To all the hard-working administrative and technical staff at the department. Thank you, without you these five years would not have run as smoothly as they did. There is not a too big or too little issue that you could not handle and you always provided help when needed no matter what.

Till mina bästa vänner Sanna och Ina. Tack för allt, tack för att ni lyssnar, stöttar och bryr er om mig. Ni betyder så mycket för mig, och jag är så glad

65 att jag har er i mitt liv. Även om vi inte hörs varje dag, så vet jag ändå var jag har er och att jag alltid kan skriva eller ringa och veta att ni är där. Moa, Emil, Tilde och lillebror, jag är så tacksam och glad över att vi lärt känna varandra. Att My har fått en sådan fin vän och trygghet i er känns underbart. Det har varit och är fantastiskt att tillsammans med er kunna koppla bort forskningen och bara ägna sig åt familjen, vänner och lekplatser.

Mamma och pappa, det är inte alltid vi har tid för varandra på senaste tiden – livet och avståndet har kommit emellan, men ni ska veta att jag älskar er. Tack för alla timmar ni har lagt på att hjälpa mig med läxläsning, att plugga till diverse tentor i ämnen som ni inte kan, för alla kramar jag har fått, allt stöd ni ger trots att ni säger att ni inte förstår vad jag håller på med och tack för att ni finns. Tack till mina syskon och deras familjer – Nina, Patrik, Oliver, Joakim, Johanna, Sofia, bror Mikael och Annica (+1).

Till mina svärföräldrar, Sonja och Örjan. Tack! Tack för alla otaliga timmar ni har tagit hand om My under denna covid-19 pandemi – utan er hade troligen inte den här avhandlingen blivit färdig. Tack för att ni finns, för att ni stöttar mig och min lilla familj avsett vad, och för att ni får mig att känna mig som en viktig del av er familj. Tack för allt ni gör för både My och Mio. Jag blir varm i hjärtat av att tänka på det och se hur mycket de betyder för er.

Sist vill jag tacka min familj – min man och mina två barn. Mikael, min Mi- kael. Du är min största idol, du är den som gör att jag håller båda fötterna kvar på jorden. Tack för att du älskar mig precis som jag är, för att du alltid försöker ditt bästa i att stötta mig oavsett vilket humör jag har, du accepterar mig för den jag är både under bra och dåliga dagar. Vi är inne på vårt elfte år tillsammans – 11 år, det är en lång tid, mer än en tredje del av våra liv. För varje dag som går så älskar jag dig mer och mer. Du och jag, vi är ett team, ett bra team och även om vi inte alltid är sams så delar vi det bästa med varandra – livet. Att få se våra barn växa upp tillsammans och bli gammal och grå med dig är något jag ser framemot. My och Mio, ni betyder så mycket för mig att ni har fått en egen sida i min avhandling.

To each and every one of you, I am so grateful for everything you have done for me, I could not have done this without you. Thank you so so much!

With love, Mikaela

66 For My and Mio

My och Mio, mina älskade små. Jag älskar er så. Ni är min sol, mitt universum, min mittpunkt i vardagen. Utan er skulle inte de vakna timmarna vara så ro- liga, underbara, utmanade och fantastiska som de är. Att återupptäcka världen med er är bland det roligaste jag vet. Att få krypa ner i sängen och ha er nära när vi sover är det mysigaste jag vet. Att få kittla, krama och pussa på er och att få tillbaka en kram, en puss, ett leende eller ett skratt är det bästa jag vet. Att få vara just mamma till er är det mest fantastiska jag vet. Jag älskar er mest av allt nu och för alltid.

67 References

1. Alenghat, F.J. and D.E. Golan, Membrane protein dynamics and functional implications in mammalian cells. Curr Top Membr, 2013. 72: p. 89-120. 2. Singer, S.J. and G.L. Nicolson, The fluid mosaic model of the structure of cell membranes. Science, 1972. 175(4023): p. 720-31. 3. Cho, W. and R.V. Stahelin, Membrane-Protein Interactions in Cell Signaling and Membrane Trafficking. Annual Review of Biophysics and Biomolecular Structure, 2005. 34(1): p. 119-151. 4. Johnson, J.E. and R.B. Cornell, Amphitropic proteins: regulation by reversible membrane interactions (review). Mol Membr Biol, 1999. 16(3): p. 217-35. 5. Goni, F.M., Non-permanent proteins in membranes: when proteins come as visitors (Review). Mol Membr Biol, 2002. 19(4): p. 237-45. 6. Wallin, E. and G. von Heijne, Genome-wide analysis of integral membrane proteins from eubacterial, archaean, and eukaryotic organisms. Protein Sci, 1998. 7(4): p. 1029-38. 7. Almen, M.S., et al., Mapping the human membrane proteome: a majority of the human membrane proteins can be classified according to function and evolutionary origin. BMC Biol, 2009. 7: p. 50. 8. Sahoo, S., et al., Membrane transporters in a human genome-scale metabolic knowledgebase and their implications for disease. Front Physiol, 2014. 5: p. 91. 9. Urban, T.J., et al., Functional genomics of membrane transporters in human populations. Genome Res, 2006. 16(2): p. 223-30. 10. Zhang, Y., et al., The SLC transporter in nutrient and metabolic sensing, regulation, and drug development. J Mol Cell Biol, 2018. 11. Overington, J.P., B. Al-Lazikani, and A.L. Hopkins, How many drug targets are there? Nat Rev Drug Discov, 2006. 5(12): p. 993-6. 12. Hediger, M.A., et al., The ABCs of membrane transporters in health and disease (SLC series): introduction. Mol Aspects Med, 2013. 34(2-3): p. 95-107. 13. Borbas, E., et al., Investigation and Mathematical Description of the Real Driving Force of Passive Transport of Drug Molecules from Supersaturated Solutions. Mol Pharm, 2016. 13(11): p. 3816-3826. 14. Wolfersberger, M.G., Uniporters, symporters and antiporters. J Exp Biol, 1994. 196: p. 5-6.

68 15. Bean, B.P., The action potential in mammalian central neurons. Nat Rev Neurosci, 2007. 8(6): p. 451-65. 16. Mueckler, M. and B. Thorens, The SLC2 (GLUT) family of membrane transporters. Mol Aspects Med, 2013. 34(2-3): p. 121- 38. 17. Wright, E.M., Glucose transport families SLC5 and SLC50. Mol Aspects Med, 2013. 34(2-3): p. 183-96. 18. Young, J.D., et al., The human concentrative and equilibrative nucleoside transporter families, SLC28 and SLC29. Mol Aspects Med, 2013. 34(2-3): p. 529-47. 19. da Fonseca, C.A.R., et al., Na(+)/K(+)-ATPase, acetylcholinesterase and glutathione S-transferase activities as new markers of postmortem interval in Swiss mice. Leg Med (Tokyo), 2019. 36: p. 67-72. 20. Jonckheere, A.I., J.A. Smeitink, and R.J. Rodenburg, Mitochondrial ATP synthase: architecture, function and pathology. J Inherit Metab Dis, 2012. 35(2): p. 211-25. 21. Jones, P.M. and A.M. George, The ABC transporter structure and mechanism: perspectives on recent research. Cell Mol Life Sci, 2004. 61(6): p. 682-99. 22. Yu, S.P. and D.W. Choi, Na(+)-Ca2+ exchange currents in cortical neurons: concomitant forward and reverse operation and effect of glutamate. Eur J Neurosci, 1997. 9(6): p. 1273-81. 23. Wright, E.M., Renal Na(+)-glucose cotransporters. Am J Physiol Renal Physiol, 2001. 280(1): p. F10-8. 24. Wright, E.M. and E. Turk, The sodium/glucose cotransport family SLC5. Pflugers Arch, 2004. 447(5): p. 510-8. 25. Alberts B, J.A., Lewis J, et al., Carrier Proteins and Active Membrane Transport. 2002: Molecular Biology of the Cell. 4th edition. New York: Garland Science;. 26. Cordat, E. and R.A. Reithmeier, Structure, function, and trafficking of SLC4 and SLC26 anion transporters. Curr Top Membr, 2014. 73: p. 1-67. 27. Donowitz, M., C. Ming Tse, and D. Fuster, SLC9/NHE gene family, a plasma membrane and organellar family of Na(+)/H(+) exchangers. Mol Aspects Med, 2013. 34(2-3): p. 236-51. 28. Khananshvili, D., The SLC8 gene family of sodium-calcium exchangers (NCX) - structure, function, and regulation in health and disease. Mol Aspects Med, 2013. 34(2-3): p. 220-35. 29. Hediger, M.A., et al., The ABCs of solute carriers: physiological, pathological and therapeutic implications of human membrane transport proteinsIntroduction. Pflugers Arch, 2004. 447(5): p. 465- 8.

69 30. Palmieri, F., The mitochondrial transporter family SLC25: identification, properties and physiopathology. Mol Aspects Med, 2013. 34(2-3): p. 465-84. 31. Csala, M., et al., Transport and transporters in the endoplasmic reticulum. Biochim Biophys Acta, 2007. 1768(6): p. 1325-41. 32. Ishida, N., et al., Identification and characterization of human Golgi nucleotide sugar transporter SLC35D2, a novel member of the SLC35 nucleotide sugar transporter family. Genomics, 2005. 85(1): p. 106-16. 33. Bissa, B., A.M. Beedle, and R. Govindarajan, Lysosomal solute carrier transporters gain momentum in research. Clin Pharmacol Ther, 2016. 100(5): p. 431-436. 34. Lawal, H.O. and D.E. Krantz, SLC18: Vesicular neurotransmitter transporters for monoamines and acetylcholine. Mol Aspects Med, 2013. 34(2-3): p. 360-72. 35. McIntire, S.L., et al., Identification and characterization of the vesicular GABA transporter. Nature, 1997. 389(6653): p. 870-6. 36. Reimer, R.J., SLC17: a functionally diverse family of organic anion transporters. Mol Aspects Med, 2013. 34(2-3): p. 350-9. 37. Hoglund, P.J., et al., The solute carrier families have a remarkably long evolutionary history with the majority of the human families present before divergence of Bilaterian species. Mol Biol Evol, 2011. 28(4): p. 1531-41. 38. Yu, F.H. and W.A. Catterall, The VGL-chanome: a protein superfamily specialized for electrical signaling and ionic homeostasis. Sci STKE, 2004. 2004(253): p. re15. 39. Fredriksson, R., et al., The G-protein-coupled receptors in the human genome form five main families. Phylogenetic analysis, paralogon groups, and fingerprints. Mol Pharmacol, 2003. 63(6): p. 1256-72. 40. Fotiadis, D., Y. Kanai, and M. Palacin, The SLC3 and SLC7 families of amino acid transporters. Mol Aspects Med, 2013. 34(2-3): p. 139-58. 41. Jung, J., H.M. Genau, and C. Behrends, Amino Acid-Dependent mTORC1 Regulation by the Lysosomal Membrane Protein SLC38A9. Mol Cell Biol, 2015. 35(14): p. 2479-94. 42. Rebsamen, M., et al., SLC38A9 is a component of the lysosomal amino acid sensing machinery that controls mTORC1. Nature, 2015. 519(7544): p. 477-81. 43. Wang, S., et al., Metabolism. Lysosomal amino acid transporter SLC38A9 signals arginine sufficiency to mTORC1. Science, 2015. 347(6218): p. 188-94. 44. Gray, K.A., et al., A review of the new HGNC gene family resource. Hum Genomics, 2016. 10: p. 6.

70 45. Yates, B., et al., Genenames.org: the HGNC and VGNC resources in 2017. Nucleic Acids Res, 2017. 45(D1): p. D619-D625. 46. Saier, M.H., Jr., et al., The Transporter Classification Database (TCDB): recent advances. Nucleic Acids Res, 2016. 44(D1): p. D372-9. 47. Saier, M.H., Jr., C.V. Tran, and R.D. Barabote, TCDB: the Transporter Classification Database for membrane transport protein analyses and information. Nucleic Acids Res, 2006. 34(Database issue): p. D181-6. 48. Finn, R.D., et al., The Pfam protein families database: towards a more sustainable future. Nucleic Acids Res, 2016. 44(D1): p. D279- 85. 49. Saier, M.H., Jr., et al., The major facilitator superfamily. J Mol Microbiol Biotechnol, 1999. 1(2): p. 257-79. 50. Pao, S.S., I.T. Paulsen, and M.H. Saier, Jr., Major facilitator superfamily. Microbiol Mol Biol Rev, 1998. 62(1): p. 1-34. 51. Schioth, H.B., et al., Evolutionary origin of amino acid transporter families SLC32, SLC36 and SLC38 and physiological, pathological and therapeutic aspects. Mol Aspects Med, 2013. 34(2-3): p. 571- 85. 52. Perland, E., et al., Characteristics of 29 novel atypical solute carriers of major facilitator superfamily type: evolutionary conservation, predicted structure and neuronal co-expression. Open Biology, 2017. 7(9). 53. Perland, E. and R. Fredriksson, Classification Systems of Secondary Active Transporters. Trends Pharmacol Sci, 2017. 38(3): p. 305-315. 54. Marger, M.D. and M.H. Saier, Jr., A major superfamily of transmembrane facilitators that catalyse uniport, symport and antiport. Trends Biochem Sci, 1993. 18(1): p. 13-20. 55. Madej, M.G., et al., Functional architecture of MFS D-glucose transporters. Proc Natl Acad Sci U S A, 2014. 111(7): p. E719-27. 56. Madej, M.G., et al., Evolutionary mix-and-match with MFS transporters. Proc Natl Acad Sci U S A, 2013. 110(15): p. 5870-4. 57. Yan, N., Structural Biology of the Major Facilitator Superfamily Transporters. Annu Rev Biophys, 2015. 44: p. 257-83. 58. Reddy, V.S., et al., The major facilitator superfamily (MFS) revisited. FEBS J, 2012. 279(11): p. 2022-35. 59. Sreedharan, S., et al., Long evolutionary conservation and considerable tissue specificity of several atypical solute carrier transporters. Gene, 2011. 478(1-2): p. 11-8. 60. Lekholm, E., et al., Putative Membrane-Bound Transporters MFSD14A and MFSD14B Are Neuronal and Affected by Nutrient Availability. Front Mol Neurosci, 2017. 10: p. 11.

71 61. Massa Lopez, D., et al., The lysosomal transporter MFSD1 is essential for liver homeostasis and critically depends on its accessory subunit GLMP. Elife, 2019. 8. 62. Perland, E., et al., The Novel Membrane-Bound Proteins MFSD1 and MFSD3 are Putative SLC Transporters Affected by Altered Nutrient Intake. J Mol Neurosci, 2017. 61(2): p. 199-214. 63. Perland, E., et al., Structural prediction of two novel human atypical SLC transporters, MFSD4A and MFSD9, and their neuroanatomical distribution in mice. PLoS One, 2017. 12(10): p. e0186325. 64. Perland, E., et al., The Putative SLC Transporters Mfsd5 and Mfsd11 Are Abundantly Expressed in the Mouse Brain and Have a Potential Role in Energy Homeostasis. PLoS One, 2016. 11(6): p. e0156912. 65. Ceder, M.M., et al., The Neuronal and Peripheral Expressed Membrane-Bound UNC93A Respond to Nutrient Availability in Mice. Front Mol Neurosci, 2017. 10: p. 351. 66. Zhang, Y., et al., The SLC transporter in nutrient and metabolic sensing, regulation, and drug development. J Mol Cell Biol, 2019. 11(1): p. 1-13. 67. Kilberg, M.S., et al., The transcription factor network associated with the amino acid response in mammalian cells. Adv Nutr, 2012. 3(3): p. 295-306. 68. Efeyan, A., W.C. Comb, and D.M. Sabatini, Nutrient-sensing mechanisms and pathways. Nature, 2015. 517(7534): p. 302-10. 69. Brasse-Lagnel, C., A. Lavoinne, and A. Husson, Control of mammalian gene expression by amino acids, especially glutamine. FEBS J, 2009. 276(7): p. 1826-44. 70. Kilberg, M.S., J. Shan, and N. Su, ATF4-dependent transcription mediates signaling of amino acid limitation. Trends Endocrinol Metab, 2009. 20(9): p. 436-43. 71. Dey, S., et al., Transcriptional repression of ATF4 gene by CCAAT/enhancer-binding protein beta (C/EBPbeta) differentially regulates integrated stress response. J Biol Chem, 2012. 287(26): p. 21936-49. 72. Battu, S., et al., Amino Acid Sensing via General Control Nonderepressible-2 Kinase and Immunological Programming. Frontiers in Immunology, 2017. 8(1719). 73. Gururajan, A. and M. van den Buuse, Is the mTOR-signalling cascade disrupted in Schizophrenia? J Neurochem, 2014. 129(3): p. 377-87. 74. Laplante, M. and D.M. Sabatini, mTOR Signaling. Cold Spring Harb Perspect Biol, 2012. 4(2). 75. Laplante, M. and D.M. Sabatini, mTOR signaling at a glance. J Cell Sci, 2009. 122(Pt 20): p. 3589-94.

72 76. Gran, P. and D. Cameron-Smith, The actions of exogenous leucine on mTOR signalling and amino acid transporters in human myotubes. BMC Physiol, 2011. 11: p. 10. 77. Lynch, C.J., Role of leucine in the regulation of mTOR by amino acids: revelations from structure-activity studies. J Nutr, 2001. 131(3): p. 861S-865S. 78. Mergenthaler, P., et al., Sugar for the brain: the role of glucose in physiological and pathological brain function. Trends Neurosci, 2013. 36(10): p. 587-97. 79. Bélanger, M., I. Allaman, and Pierre J. Magistretti, Brain Energy Metabolism: Focus on Astrocyte-Neuron Metabolic Cooperation. Cell Metabolism, 2011. 14(6): p. 724-738. 80. Thorens, B., GLUT2, glucose sensing and glucose homeostasis. Diabetologia, 2015. 58(2): p. 221-232. 81. Gao, H.F., et al., SLC5A1 promotes growth and proliferation of pancreatic carcinoma via glucose-dependent AMPK/mTOR signaling. Cancer Manag Res, 2019. 11: p. 3171-3185. 82. Havula, E. and V. Hietakangas, Glucose sensing by ChREBP/MondoA-Mlx transcription factors. Semin Cell Dev Biol, 2012. 23(6): p. 640-7. 83. Ma, L., L.N. Robinson, and H.C. Towle, ChREBP*Mlx is the principal mediator of glucose-induced gene expression in the liver. J Biol Chem, 2006. 281(39): p. 28721-30. 84. Havula, E., et al., Mondo/ChREBP-Mlx-regulated transcriptional network is essential for dietary sugar tolerance in Drosophila. PLoS Genet, 2013. 9(4): p. e1003438. 85. Hunt, L.C., et al., The glucose-sensing transcription factor MLX promotes myogenesis via myokine signaling. Genes Dev, 2015. 29(23): p. 2475-89. 86. Stoltzman, C.A., et al., Glucose sensing by MondoA:Mlx complexes: a role for hexokinases and direct regulation of thioredoxin- interacting protein expression. Proc Natl Acad Sci U S A, 2008. 105(19): p. 6912-7. 87. Donowitz, M., C. Ming Tse, and D. Fuster, SLC9/NHE gene family, a plasma membrane and organellar family of Na⁺⁺/H exchangers. Mol Aspects Med, 2013. 34(2-3): p. 236-51. 88. Wang, Y.Y., et al., SLC9A3 Protein Is Critical for Acrosomal Formation in Postmeiotic Male Germ Cells. Int J Mol Sci, 2017. 19(1). 89. Bergersen, L.H., Lactate transport and signaling in the brain: potential therapeutic targets and roles in body-brain interaction. J Cereb Blood Flow Metab, 2015. 35(2): p. 176-85. 90. Anderson, C.M. and A. Stahl, SLC27 fatty acid transport proteins. Mol Aspects Med, 2013. 34(2-3): p. 516-28.

73 91. Nguyen, L.N., et al., Mfsd2a is a transporter for the essential omega-3 fatty acid docosahexaenoic acid. Nature, 2014. 509(7501): p. 503-6. 92. Jeong, J. and D.J. Eide, The SLC39 family of zinc transporters. Mol Aspects Med, 2013. 34(2-3): p. 612-9. 93. Song, Z., Roles of the nucleotide sugar transporters (SLC35 family) in health and disease. Mol Aspects Med, 2013. 34(2-3): p. 590-600. 94. Cesar-Razquin, A., et al., A Call for Systematic Research on Solute Carriers. Cell, 2015. 162(3): p. 478-87. 95. Rask-Andersen, M., et al., Solute carriers as drug targets: current use, clinical trials and prospective. Mol Aspects Med, 2013. 34(2- 3): p. 702-10. 96. Hauser, A.S., et al., Trends in GPCR drug discovery: new agents, targets and indications. Nature Reviews Drug Discovery, 2017. 16(12): p. 829-842. 97. Wang, W.W., et al., The Druggability of Solute Carriers. J Med Chem, 2020. 63(8): p. 3834-3867. 98. Sriram, K. and P.A. Insel, G Protein-Coupled Receptors as Targets for Approved Drugs: How Many Targets and How Many Drugs? Mol Pharmacol, 2018. 93(4): p. 251-258. 99. Rives, M.L., J.A. Javitch, and A.D. Wickenden, Potentiating SLC transporter activity: Emerging drug discovery opportunities. Biochem Pharmacol, 2017. 135: p. 1-11. 100. Garibsingh, R.A. and A. Schlessinger, Advances and Challenges in Rational Drug Design for SLCs. Trends Pharmacol Sci, 2019. 40(10): p. 790-800. 101. Eddy, S.R., Accelerated Profile HMM Searches. PLoS Comput Biol, 2011. 7(10): p. e1002195. 102. Eddy, S.R., What is a hidden Markov model? Nat Biotechnol, 2004. 22(10): p. 1315-6. 103. Katoh, K., et al., MAFFT: a novel method for rapid multiple sequence alignment based on fast Fourier transform. Nucleic Acids Res, 2002. 30(14): p. 3059-66. 104. Notredame, C., D.G. Higgins, and J. Heringa, T-coffee: a novel method for fast and accurate multiple sequence alignment11Edited by J. Thornton. Journal of Molecular Biology, 2000. 302(1): p. 205- 217. 105. Di Tommaso, P., et al., T-Coffee: a web server for the multiple sequence alignment of protein and RNA sequences using structural information and homology extension. Nucleic Acids Res, 2011. 39(Web Server issue): p. W13-7. 106. Kemena, C. and C. Notredame, Upcoming challenges for multiple sequence alignment methods in the high-throughput era. Bioinformatics, 2009. 25(19): p. 2455-65.

74 107. Gregory, T.R., Understanding Evolutionary Trees. Evolution: Education and Outreach, 2008. 1(2): p. 121-137. 108. Yang, Z. and B. Rannala, Molecular phylogenetics: principles and practice. Nature Reviews Genetics, 2012. 13(5): p. 303-314. 109. Myung, I.J., Tutorial on maximum likelihood estimation. Journal of Mathematical Psychology, 2003. 47(1): p. 90-100. 110. Fienberg, S.E., When did Bayesian inference become "Bayesian"? Bayesian Anal., 2006. 1(1): p. 1-40. 111. Douady, C.J., et al., Comparison of Bayesian and Maximum Likelihood Bootstrap Measures of Phylogenetic Reliability. Molecular Biology and Evolution, 2003. 20(2): p. 248-254. 112. Suzuki, Y., G.V. Glazko, and M. Nei, Overcredibility of molecular phylogenies obtained by Bayesian phylogenetics. Proc Natl Acad Sci U S A, 2002. 99(25): p. 16138-43. 113. Erixon, P., et al., Reliability of Bayesian posterior probabilities and bootstrap frequencies in phylogenetics. Syst Biol, 2003. 52(5): p. 665-73. 114. García-Sandoval, R., Why some clades have low bootstrap frequencies and high Bayesian posterior probabilities. Israel Journal of Ecology & Evolution, 2014. 60(1): p. 41-44. 115. Efron, B., E. Halloran, and S. Holmes, Bootstrap confidence levels for phylogenetic trees. Proceedings of the National Academy of Sciences, 1996. 93(23): p. 13429-13429. 116. Stamatakis, A., RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics, 2006. 22(21): p. 2688-2690. 117. Stamatakis, A., RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics, 2014. 30(9): p. 1312-3. 118. Berman, H.M., et al., The Protein Data Bank. Nucleic Acids Res, 2000. 28(1): p. 235-42. 119. Bai, X., T.F. Moraes, and R.A.F. Reithmeier, Structural biology of solute carrier (SLC) membrane transport proteins. Molecular Membrane Biology, 2017. 34(1-2): p. 1-32. 120. Wang, H., J. Elferich, and E. Gouaux, Structures of LeuT in bicelles define conformation and substrate binding in a membrane-like context. Nat Struct Mol Biol, 2012. 19(2): p. 212-9. 121. Solcan, N., et al., Alternating access mechanism in the POT family of oligopeptide transporters. EMBO J, 2012. 31(16): p. 3411-21. 122. Jensen, S., et al., Crystal structure of a substrate-free aspartate transporter. Nat Struct Mol Biol, 2013. 20(10): p. 1224-6. 123. Verdon, G., et al., Coupled ion binding and structural transitions along the transport cycle of glutamate transporters. Elife, 2014. 3: p. e02283.

75 124. Deng, D., et al., Crystal structure of the human glucose transporter GLUT1. Nature, 2014. 510(7503): p. 121-5. 125. Deng, D., et al., Molecular basis of ligand recognition and transport by glucose transporters. Nature, 2015. 526(7573): p. 391-6. 126. Gruswitz, F., et al., Function of human Rh based on structure of RhCG at 2.1 A. Proc Natl Acad Sci U S A, 2010. 107(21): p. 9638- 43. 127. Omasits, U., et al., Protter: interactive protein feature visualization and integration with experimental proteomic data. Bioinformatics, 2014. 30(6): p. 884-6. 128. Kelley, L.A., et al., The Phyre2 web portal for protein modeling, prediction and analysis. Nat Protoc, 2015. 10(6): p. 845-58. 129. Omasits, U., et al., Protter: interactive protein feature visualization and integration with experimental proteomic data. Bioinformatics, 2013. 30(6): p. 884-886. 130. Zhang, Y., Protein structure prediction: when is it useful? Curr Opin Struct Biol, 2009. 19(2): p. 145-55. 131. Gupta, C.L., S. Akhtar, and P. Bajpai, In silico protein modeling: possibilities and limitations. EXCLI J, 2014. 13: p. 513-5. 132. Yi, S., et al., Endoplasmic Reticulum Stress PERK-ATF4-CHOP Pathway Is Associated with Hypothalamic Neuronal Injury in Different Durations of Stress in Rats. Front Neurosci, 2017. 11: p. 152. 133. Zhang, Q., et al., Central activating transcription factor 4 (ATF4) regulates hepatic insulin resistance in mice via S6K1 signaling and the vagus nerve. Diabetes, 2013. 62(7): p. 2230-9. 134. Cavin Perier, R., T. Junier, and P. Bucher, The Eukaryotic Promoter Database EPD. Nucleic Acids Res, 1998. 26(1): p. 353-7. 135. Aerts, S., Chapter five - Computational Strategies for the Genome- Wide Identification of cis-Regulatory Elements and Transcriptional Targets, in Current Topics in Developmental Biology, S. Plaza and F. Payre, Editors. 2012, Academic Press. p. 121-145. 136. Sandelin, A., et al., JASPAR: an open access database for eukaryotic transcription factor binding profiles. Nucleic Acids Research, 2004. 32(suppl_1): p. D91-D94. 137. Khan, A., et al., JASPAR 2018: update of the open-access database of transcription factor binding profiles and its web framework. Nucleic Acids Research, 2017. 46(D1): p. D260-D266. 138. Hu, J., B. Li, and D. Kihara, Limitations and potentials of current motif discovery algorithms. Nucleic Acids Res, 2005. 33(15): p. 4899-913. 139. Gearing, L.J., et al., CiiiDER: A tool for predicting and analysing transcription factor binding sites. PLoS One, 2019. 14(9): p. e0215495.

76 140. Kaur, G. and J.M. Dufour, Cell lines: Valuable tools or useless artifacts. Spermatogenesis, 2012. 2(1): p. 1-5. 141. Katt, M.E., et al., In Vitro Tumor Models: Advantages, Disadvantages, Variables, and Selecting the Right Platform. Front Bioeng Biotechnol, 2016. 4: p. 12. 142. Perlman, R.L., Mouse models of human disease: An evolutionary perspective. Evol Med Public Health, 2016. 2016(1): p. 170-6. 143. Koboziev, I., et al., Pharmacological intervention studies using mouse models of the inflammatory bowel diseases: translating preclinical data into new drug therapies. Inflamm Bowel Dis, 2011. 17(5): p. 1229-45. 144. Pendleton, R.G., et al., Effects of pharmacological agents upon a transgenic model of Parkinson's disease in Drosophila melanogaster. J Pharmacol Exp Ther, 2002. 300(1): p. 91-6. 145. McBride, S.M., S.L. Holloway, and T.A. Jongens, Using Drosophila as a tool to identify pharmacological therapies for fragile X syndrome. Drug Discov Today Technol, 2013. 10(1): p. e129-36. 146. Adams, M.D., et al., The genome sequence of Drosophila melanogaster. Science, 2000. 287(5461): p. 2185-95. 147. Roote, J. and A. Prokop, How to Design a Genetic Mating Scheme: A Basic Training Package for <em>Drosophila</em> Genetics. G3: Genes|Genomes|Genetics, 2013. 3(2): p. 353. 148. Featherstone, D.E., Glial solute carrier transporters in Drosophila and mice. Glia, 2011. 59(9): p. 1351-63. 149. Michael Ashburner, K.G.G., and R Scott Hawley, Drosophila: A Laboratory Handbook. Second Edition. The Quarterly Review of Biology, 2005. 80(2): p. 249-249. 150. Hales, K.G., et al., Genetics on the Fly: A Primer on the Drosophila Model System. Genetics, 2015. 201(3): p. 815-42. 151. Klar, A.J. and H.O. Halvorson, Studies on the positive regulatory gene, GAL4, in regulation of galactose catabolic enzymes in Saccharomyces cerevisiae. Mol Gen Genet, 1974. 135(3): p. 203-12. 152. Duffy, J.B., GAL4 system in Drosophila: a fly geneticist's Swiss army knife. Genesis, 2002. 34: p. 1-15. 153. Saurabh, S., A.S. Vidyarthi, and D. Prasad, RNA interference: concept to reality in crop improvement. Planta, 2014. 239(3): p. 543- 64. 154. Liu, Y. and M. Lehmann, A genomic response to the yeast transcription factor GAL4 in Drosophila. Fly (Austin), 2008. 2(2): p. 92-8. 155. Jonchere, V. and D. Bennett, Validating RNAi phenotypes in Drosophila using a synthetic RNAi-resistant transgene. PLoS One, 2013. 8(8): p. e70489.

77 156. Dietzl, G., et al., A genome-wide transgenic RNAi library for conditional gene inactivation in Drosophila. Nature, 2007. 448(7150): p. 151-6. 157. Green, E.W., et al., A Drosophila RNAi collection is subject to dominant phenotypic effects. Nat Methods, 2014. 11(3): p. 222-3. 158. Vissers, J.H., et al., A Drosophila RNAi library modulates Hippo pathway-dependent tissue growth. Nat Commun, 2016. 7: p. 10368. 159. Pandey, U.B. and C.D. Nichols, Human disease models in Drosophila melanogaster and the role of the fly in therapeutic drug discovery. Pharmacol Rev, 2011. 63(2): p. 411-36. 160. Reiter, L.T., et al., A systematic analysis of human disease- associated gene sequences in Drosophila melanogaster. Genome Res, 2001. 11(6): p. 1114-25. 161. Beyenbach, K.W., H. Skaer, and J.A. Dow, The developmental, molecular, and transport biology of Malpighian tubules. Annu Rev Entomol, 2010. 55: p. 351-74. 162. Maddrell, S.H. and M.J. O'Donnell, INSECT MALPIGHIAN TUBULES: V-ATPase ACTION IN ION AND FLUID TRANSPORT. J Exp Biol, 1992. 172(Pt 1): p. 417-429. 163. Dow, J.A. and M.F. Romero, Drosophila provides rapid modeling of renal development, function, and disease. Am J Physiol Renal Physiol, 2010. 299(6): p. F1237-44. 164. Weavers, H., et al., The insect nephrocyte is a podocyte-like cell with a filtration slit diaphragm. Nature, 2009. 457(7227): p. 322-6. 165. Wang, Y., et al., The fruit fly Drosophila melanogaster as an innovative preclinical ADME model for solute carrier membrane transporters, with consequences for pharmacology and drug therapy. Drug Discov Today, 2018. 23(10): p. 1746-1760. 166. Jeibmann, A. and W. Paulus, Drosophila melanogaster as a model organism of brain diseases. Int J Mol Sci, 2009. 10(2): p. 407-40. 167. Nichols, C.D., Drosophila melanogaster neurobiology, neuropharmacology, and how the fly can inform central nervous system drug discovery. Pharmacology & Therapeutics, 2006. 112(3): p. 677-700. 168. Bharucha, K.N., The Epicurean Fly: Using Drosophila Melanogaster to Study Metabolism. Pediatric Research, 2009. 65(2): p. 132-137. 169. Musselman, L.P. and R.P. Kuhnlein, Drosophila as a model to study obesity and metabolic disease. J Exp Biol, 2018. 221(Pt Suppl 1). 170. Pfeiffenberger, C., et al., Locomotor activity level monitoring using the Drosophila Activity Monitoring (DAM) System. Cold Spring Harb Protoc, 2010. 2010(11): p. pdb prot5518. 171. Itskov, P.M., et al., Automated monitoring and quantitative analysis of feeding behaviour in Drosophila. Nature Communications, 2014. 5(1): p. 4560.

78 172. Schellinger, J.N. and A.R. Rodan, Use of the Ramsay Assay to Measure Fluid Secretion and Ion Flux Rates in the Drosophila melanogaster Malpighian Tubule. J Vis Exp, 2015(105). 173. Williams, M.J., et al., Recurrent Sleep Fragmentation Induces Insulin and Neuroprotective Mechanisms in Middle-Aged Flies. Front Aging Neurosci, 2016. 8: p. 180. 174. Julienne, H., et al., Drosophila PINK1 and parkin loss-of-function mutants display a range of non-motor Parkinson's disease phenotypes. Neurobiol Dis, 2017. 104: p. 15-23. 175. De Lazzari, F., et al., Circadian Rhythm Abnormalities in Parkinson's Disease from Humans to Flies and Back. Int J Mol Sci, 2018. 19(12). 176. Limmer, S., et al., The Drosophila blood-brain barrier: development and function of a glial endothelium. Front Neurosci, 2014. 8: p. 365. 177. Wheeler, T.J. and S.R. Eddy, nhmmer: DNA homology search with profile HMMs. Bioinformatics, 2013. 29(19): p. 2487-2489. 178. Mattila, J. and V. Hietakangas, Regulation of Carbohydrate Energy Metabolism in Drosophila melanogaster. Genetics, 2017. 207(4): p. 1231-1253. 179. McCue, M.D., Starvation physiology: reviewing the different strategies animals use to survive a common challenge. Comp Biochem Physiol A Mol Integr Physiol, 2010. 156(1): p. 1-18. 180. Chen, R., et al., The general amino acid control pathway regulates mTOR and autophagy during serum/glutamine starvation. J Cell Biol, 2014. 206(2): p. 173-82. 181. Sunwoldt, J., et al., Neuronal Culture Microenvironments Determine Preferences in Bioenergetic Pathway Use. Front Mol Neurosci, 2017. 10: p. 305. 182. Rebouças, E.d.L., et al., Real time PCR and importance of housekeepings genes for normalization and quantification of mRNA expression in different tissues. Brazilian Archives of Biology and Technology, 2013. 56: p. 143-154. 183. Dowling, C.M., et al., The importance of selecting the appropriate reference genes for quantitative real time PCR as illustrated using colon cancer cells and tissue. F1000Res, 2016. 5: p. 99. 184. Adeola, F., Normalization of Gene Expression by Quantitative RT- PCR in Human Cell Line: comparison of 12 Endogenous Reference Genes. Ethiop J Health Sci, 2018. 28(6): p. 741-748. 185. Panina, Y., et al., Validation of Common Housekeeping Genes as Reference for qPCR Gene Expression Analysis During iPS Reprogramming Process. Scientific Reports, 2018. 8(1): p. 8716. 186. Barber, R.D., et al., GAPDH as a housekeeping gene: analysis of GAPDH mRNA expression in a panel of 72 human tissues. Physiol Genomics, 2005. 21(3): p. 389-95.

79 187. Vandesompele, J., et al., Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol, 2002. 3(7): p. RESEARCH0034. 188. de la Cruz, I.P., et al., sup-9, sup-10, and unc-93 may encode components of a two-pore K+ channel that coordinates muscle contraction in Caenorhabditis elegans. J Neurosci, 2003. 23(27): p. 9133-45. 189. Levin, J.Z. and H.R. Horvitz, The Caenorhabditis elegans unc-93 gene encodes a putative transmembrane protein that regulates muscle contraction. J Cell Biol, 1992. 117(1): p. 143-55. 190. Lee, B.L., et al., UNC93B1 mediates differential trafficking of endosomal TLRs. Elife, 2013. 2: p. e00291. 191. Huh, J.W., et al., UNC93B1 is essential for the plasma membrane localization and signaling of Toll-like receptor 5. Proc Natl Acad Sci U S A, 2014. 111(19): p. 7072-7. 192. Liu, Y., et al., The human homologue of unc-93 maps to chromosome 6q27 - characterisation and analysis in sporadic epithelial ovarian cancer. BMC Genet, 2002. 3: p. 20. 193. Son, H.Y., et al., Family-Based Association Study of Pulmonary Function in a Population in Northeast Asia. PLoS One, 2015. 10(10): p. e0139716. 194. Schlosser, P., et al., Genetic studies of urinary metabolites illuminate mechanisms of detoxification and excretion in humans. Nature Genetics, 2020. 52(2): p. 167-176. 195. Campbell, C.L., et al., A role for endosomal proteins in alphavirus dissemination in mosquitoes. Insect Mol Biol, 2011. 20(4): p. 429- 36. 196. Mohr, S.E., et al., Zinc Detoxification: A Functional Genomics and Transcriptomics Analysis in Drosophila melanogaster Cultured Cells. G3 (Bethesda), 2018. 8(2): p. 631-641. 197. Stofanko, M., S.Y. Kwon, and P. Badenhorst, A misexpression screen to identify regulators of Drosophila larval hemocyte development. Genetics, 2008. 180(1): p. 253-67. 198. Wang, J., et al., Function-informed transcriptome analysis of Drosophila renal tubule. Genome Biol, 2004. 5(9): p. R69. 199. Duprat, F., et al., The TASK background K2P channels: chemo- and nutrient sensors. Trends Neurosci, 2007. 30(11): p. 573-80. 200. Karschin, C., et al., Expression pattern in brain of TASK-1, TASK-3, and a tandem pore domain K(+) channel subunit, TASK-5, associated with the central auditory nervous system. Mol Cell Neurosci, 2001. 18(6): p. 632-48. 201. Marinc, C., et al., Immunocytochemical localization of TASK-3 protein (K2P9.1) in the rat brain. Cell Mol Neurobiol, 2014. 34(1): p. 61-70.

80 202. O'Donnell, M.J. and S.H. Maddrell, Fluid reabsorption and ion transport by the lower Malpighian tubules of adult female Drosophila. J Exp Biol, 1995. 198(Pt 8): p. 1647-53. 203. Gautam, N.K., P. Verma, and M.G. Tapadia, Drosophila Malpighian Tubules: A Model for Understanding Kidney Development, Function, and Disease. Results Probl Cell Differ, 2017. 60: p. 3-25. 204. Soderberg, J.A., R.T. Birse, and D.R. Nassel, Insulin production and signaling in renal tubules of Drosophila is under control of tachykinin-related peptide and regulates stress resistance. PLoS One, 2011. 6(5): p. e19866. 205. Wu, Y., et al., Two inwardly rectifying potassium channels, Irk1 and Irk2, play redundant roles in Drosophila renal tubule function. Am J Physiol Regul Integr Comp Physiol, 2015. 309(7): p. R747-56. 206. Dow, J.A.T., S.A. Davies, and M.A. Sozen, Fluid secretion by the Drosophila Malpighian tubule. American Zoologist, 1998. 38(3): p. 450-460. 207. Wishart, D.S., et al., DrugBank 5.0: a major update to the DrugBank database for 2018. Nucleic Acids Res, 2018. 46(D1): p. D1074-D1082. 208. Zelle, K.M., et al., The genetic architecture of degenerin/epithelial sodium channels in Drosophila. G3 (Bethesda), 2013. 3(3): p. 441- 50. 209. Giannakou, M.E. and J.A. Dow, Characterization of the Drosophila melanogaster alkali-metal/proton exchanger (NHE) gene family. J Exp Biol, 2001. 204(Pt 21): p. 3703-16. 210. Rodan, A.R., M. Baum, and C.L. Huang, The Drosophila NKCC Ncc69 is required for normal renal tubule function. Am J Physiol Cell Physiol, 2012. 303(8): p. C883-94. 211. Stadler, C., et al., Immunofluorescence and fluorescent-protein tagging show high correlation for protein localization in mammalian cells. Nat Methods, 2013. 10(4): p. 315-23. 212. Hagglund, M.G., et al., Identification of SLC38A7 (SNAT7) protein as a glutamine transporter expressed in neurons. J Biol Chem, 2011. 286(23): p. 20500-11. 213. Hagglund, M.G.A., et al., Transport of L-glutamine, L-alanine, L- arginine and L-histidine by the neuron-specific Slc38a8 (SNAT8) in CNS. J Mol Biol, 2015. 427(6 Pt B): p. 1495-1512. 214. Hellsten, S.V., et al., The neuronal and astrocytic protein SLC38A10 transports glutamine, glutamate, and aspartate, suggesting a role in neurotransmission. FEBS Open Bio, 2017. 7(6): p. 730-746. 215. Bazzone, A., M. Barthmes, and K. Fendler, Chapter Two - SSM- Based Electrophysiology for Transporter Research, in Methods in Enzymology, C. Ziegler, Editor. 2017, Academic Press. p. 31-83.

81 Acta Universitatis Upsaliensis Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Pharmacy 287 Editor: The Dean of the Faculty of Pharmacy

A doctoral dissertation from the Faculty of Pharmacy, Uppsala University, is usually a summary of a number of papers. A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Pharmacy. (Prior to January, 2005, the series was published under the title “Comprehensive Summaries of Uppsala Dissertations from the Faculty of Pharmacy”.)

ACTA UNIVERSITATIS UPSALIENSIS Distribution: publications.uu.se UPPSALA urn:nbn:se:uu:diva-416506 2020