Crosstalk between ceramide and cell signalling pathways controlling chronological lifespan in yeast

! ! 1! Vitor Hugo Freitas Teixeira Crosstalk between ceramide and cell signalling pathways controlling chronological lifespan in yeast Tese de Candidatura ao grau de Doutor em Ciências Biomédicas submetida ao Instituto de Ciências Biomédicas Abel Salazar da Universidade do Porto. Orientador – Professor Doutor Vitor Manuel Vieira da Costa Categoria – Professor Associado Afiliação – Instituto de Ciências Biomédicas Abel Salazar da Universidade do Porto Co-orientador – Professor Doutor Pedro Gaspar Moradas-Ferreira Categoria – Professor Catedrático Afiliação – Instituto de Ciências Biomédicas Abel Salazar da Universidade do Porto ! ! 2! ! ! 3! Publications De acordo com o disposto no n.º1 do artigo 34.º do Decreto-Lei n.º 74/2006, publicado em Diário da República, 1.ª série, n.º 60 de 24 de Março de 2006, e republicado pelo Decreto-Lei n.º 115/2013, publicado em Diário da República, 1.ª série, n.º 151 de 7 de Agosto de 2013, que procede à terceira alteração ao Decreto-Lei n.º 74/2006, de 24 de Março de 2006, nesta tese foram utilizados os resultados de trabalhos publicados abaixo indicados. Teixeira, V; Medeiros, T.C.; Vilaça, R.; Moradas-Ferreira, P; Costa, V. Reduced TORC1 signalling abolishes mitochondrial dysfunctions and shortened chronological lifespan of Isc1p-deficient cells, Microbial Cell 2014, 1(1): 21-36. (DOI: 10.15698/mic2014.01.121) No cumprimento do Decreto-Lei acima mencionado, o autor desta dissertação declara que interveio na concepção e execução do trabalho experimental, na interpretação e discussão dos resultados, e na sua redacção. Durante a execução deste trabalho, o autor desta tese colaborou no seguinte trabalho publicado: Vilaça, R.; Silva, E., Nadais, A.; Teixeira, V. et al., Sphingolipid signalling mediates mitochondrial dysfunctions and reduced chronological lifespan in the yeast model of Niemann-Pick type C1, Mol Microbiol 2013, 91(3): 438-451. (DOI: 10.1111/mmi.12470) ! ! 4! The work presented in this thesis was done at the Institute for Molecular and Cell Biology (IBMC, University of Porto), Porto, Portugal. This work was financially supported by FEDER funds through the Operational Competitiveness Programme – COMPETE and by National funds through FCT – Fundação para a Ciência e a Tecnologia under the projects FCOMP-01-0124-FEDER-028210 (PTDC/BBB-BQB/1850/2012) and PEst-OE/BIA/UI4050/2014 and a grant (SFRH/BD/72134/2010) from FCT. ! ! 5! Ó tocadora de harpa, se eu beijasse Ó tocadora de harpa, se eu beijasse Teu gesto, sem beijar as tuas mãos! E, beijando-o, descesse pelos desvãos Do sonho, até que enfim eu o encontrasse. Tornado Puro Gesto, gesto-face Da medalha sinistra — reis cristãos Ajoelhando, inimigos e irmãos, Quando processional o andor passasse! Teu gesto que arrepanha e se extasia. O teu gesto completo, lua fria Subindo, e em baixo, negros, os juncais. Caverna em estalactites o teu gesto. Não poder eu prendê-lo, fazer mais Que vê-lo e que perdê-lo! E o sonho é o resto. Fernando Pessoa ! ! 6! Acknowledgements ! ! 7! Acknowledgements It would not have been possible to write this doctoral thesis without the help and support of kind people around me. I won´t be able to include them all but they know they have my gratitude and my friendship for all their encouragement and patience. Above all, I would like to thank my family for their personal support and great patience at all times. In particular, my parents, Manuel and Rita Teixeira, and my sister, Carla Teixeira, have given me their unequivocal and unconditional support throughout, as always, for which my mere expression of thanks does not suffice. Indeed. They are my heroes for life and my steeping stone on my way to succeed in life. I´ve also been blessed for the birth of my twin godchildren, Ana Carolina and Jorge Gabriel, which was utterly one of the happiest moments I´ve ever experienced during my whole life. This thesis would not have been possible without the help, support and patience of my main supervisor, Prof. Vitor Costa, not to mention his advice throughout crucial parts of my work. The good advice and support of my cosupervisor, Prof. Pedro Moradas-Ferreira, have been invaluable on both an academic and a scientific level, for which I am extremely grateful. There are definitely people that were present at crucial times of my life and there are no words to express how much I thank them for being present whenever I needed and always ready to listen and give encouragement. To Márcia Monteiro, for the everlasting friendship with more than 15 years. We have shared so many and countless moments in our lives, so many journeys accomplished together, so many projects fulfilled. We have been always for each other. You are my best friend and my "life traveller" and there are no words to express my gratitude for having you in my life. To Sílvia do Carmo Lopes, you were a revelation in and for my life and you´ve taught me so much that I sometimes have the impression you don´t even realize that. I hope to fulfill my promise and allow you to wear that red dress one day, soon enough. To Cristina Neves, who was always there for me, time after time, to help me out and give me her optimistic advice, wisdom and encouragement. We have shared this journey together and you´ve already Acknowledgements ! ! 8! achieved "the goal". Thank you for letting me take part of your life the way you did and I do have some of the best memories of my life because of you. I couldn´t be more happy for you. Wait for me, I´m near. To Joana Fernandes, Flávia Santos, Ricardo Silva, Patrícia Fernandes, Helena Santos, Daniela Monteiro, Daniel, Carlos Silva and later on, Carlos Rodrigues, Rita Borges, Guilherme Silva, Ana Soares, Raquel Kritinas, Rita Silva, Teresa Santos, Pedro Rocha, Marco, Artur Costa, Sérgio Mendes, Diana Sousa, Inês Lobão, Sofia Ramalho, Rita Marques, Juliano Ribas, Lisandra Mota and so many others friends. I thank your generosity and encouragement at different times of my life, especially on my local Youth Group, for the "Taizé experience" and my book writing. I hope that the "literary lives" I´ve created in my books will be available soon. One day. I am most grateful to former and current lab colleagues, particularly to Helena Rocha, Rita Vilaça, Elísio Silva, Vanda Mendes, Andreia Pereira and Clara Pereira for their kindness, friendship and support, together with the members of the Cellular and Applied Microbiology Group and the Bioengineering and Synthetic Microbiology Group. They were important by promoting a welcoming scientific and social environment. In particular, I want to thank Tânia Catarina Medeiros, whom I´ve supervised during her Master thesis. I could not me more proud for having contributed for the buildup of a promising young scientist like you and earned a very dear friend during that journey. Hard work leads to success. I would also like to acknowledge the technical support of the IBMC and its staff, particularly to Catarina Leitão, Liliana and Alexandra, for their support and assistance since the start of my PhD work. I have to acknowledge their importance for making my work more organized and efficient. Last, but by no means least, I thank everyone else for their support and encouragement throughout, which have not been named here. I´m so grateful for being surrounded by all of you in my life. Table of Contents ! ! 9! Table of Contents Table of Contents ................................................................................................. 9 List of Figures .................................................................................................... 15 List of Tables ...................................................................................................... 18 List of Abbreviations ......................................................................................... 19 Summary ............................................................................................................. 21 Sumário ............................................................................................................... 25 Chapter I - General Introduction 1.1. Yeast as an eukaryote biological model system ...................................... 30 1.2. Aging and disease ...................................................................................... 34 1.2.1. Genomic instability and telomere shortening ........................... 35 1.2.2. Epigenetic alterations ................................................................... 37 1.2.3. Loss of protein homeostasis (proteostasis) ............................... 38 1.2.4. Impaired nutrient signalling ......................................................... 40 1.2.4.1. mTORC1 signalling ......................................................... 40 1.2.4.2. AMPK signalling .............................................................. 41 1.2.4.3. Sirtuins ............................................................................. 43 1.2.5. Stem cell exhaustion ..................................................................... 44 1.3. Oxidative stress and mitochondrial dysfunction in the aging process . 47 1.3.1. The mitochondrial free radical theory of aging .......................... 47 1.3.2. ROS scavenging systems ............................................................ 49 1.3.2.1. Non-enzymatic components of the antioxidant defense system ........................................................................................... 49 I. Glutathione ...................................................................... 49 List of Figures ! ! 16! Figure 2.S5 ........................................................................................................ 120 Figure 3.1 .......................................................................................................... 127 Figure 3.2 .......................................................................................................... 130 Figure 3.3 .......................................................................................................... 134 Figure 3.4 .......................................................................................................... 137 Figure 3.5 .......................................................................................................... 140 Figure 3.6 .......................................................................................................... 143 Figure 3.7 .......................................................................................................... 145 Figure 3.S1 ........................................................................................................ 161 Figure 3.S2 ........................................................................................................ 162 Figure 3.S3 ........................................................................................................ 163 Figure 3.S4 ........................................................................................................ 164 Figure 3.S5 ........................................................................................................ 164 Figure 3.S6 ........................................................................................................ 165 Figure 3.S7 ........................................................................................................ 166 Figure 3.S8 ........................................................................................................ 167 Figure 3.S9 ........................................................................................................ 167 Figure 3.S10 ...................................................................................................... 168 Figure 3.S11 ...................................................................................................... 169 Figure 3.S12 ...................................................................................................... 170 Figure 4.1 .......................................................................................................... 178 Figure 4.2 .......................................................................................................... 179 Figure 4.3 .......................................................................................................... 180 List of Figures ! ! 17! Figure 4.4 .......................................................................................................... 181 Figure 4.5 .......................................................................................................... 182 Figure 4.6 .......................................................................................................... 183 Figure 4.7 .......................................................................................................... 184 Figure 4.8 .......................................................................................................... 185 Figure 4.9 .......................................................................................................... 188 Figure 4.S1 ........................................................................................................ 197 Figure 5.1 .......................................................................................................... 205 List of Tables ! ! 18! List of Tables Table 2.5.1 ......................................................................................................... 116 Table 3.5.1 ......................................................................................................... 158 Table 4.5.1 ......................................................................................................... 196 List of Abbreviations ! ! 19! List of Abbreviations AbA, Aureobasidin A; ADP, Adenosine diphosphate; ALP, Alkaline phosphatase; AMPK, 5' Adenosine monophosphateactivated protein kinase; AP1, Activator protein 1 transcription factor; ATG, Autophagy-related gene; ATP, Adenosine triphosphate; CAPP, Ceramide-activated protein Phosphatase; CK2, Casein kinase 2; CL, Cardiolipin; CLS, Chronological lifespan; CMA, Chaperonemediated autophagy; COX, Cytochrome c oxidase; CPY, Carboxypeptidase Y; CR, Calorie restriction; Cvt, Cytoplasm-to-vacuole targeting; CWI, Cell Wall Integrity; DAG, Diacylglycerol; DHE, Dihydroethidium; DHS, Dihydrosphingosine; DMSO, Dimethyl sulfoxide; DNA, deoxyribonucleic acid; DNP, 2-4 dinitrophenol; DTT, Dithiothreitol; EDTA, Ethylenediamine tetraacetic Acid; ER, Endoplasmic reticulum; ERAD, Endoplasmic-Reticulum-associated protein degradation; ERMES, Endoplasmic Reticulum (ER)-mitochondria encounter structure; ETC, Electron transport chain; FACS, Fluorescence-activated cell sorting; FCCP, carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone; Fw, forward; GFP, Green fluorescent protein; GTP, Guanosine triphosphate; GSH, Glutathione; HM, hydrophobic motif; HA, Human influenza hemagglutinin; HOG, High Osmolarity Glycerol; HSC, Hematopoietic stem cells; IGF-1, Insulin-like growth factor-1; IL-13, Interleukin-13; IPC, Inositol-phosphorylceramide; ISC1, Inositol phosphosphingolipid phospholipase C; JNK, c-Jun N-terminal protein Kinase; LAMP-2A, Lysosomal associated membrane protein-2A; LCBs, Long chain bases; LCBPs, Phosphorylated long chain bases; LC-MS/MS, Liquid chromatography-tandem mass spectrometry; MAPK, Mitogen-activated protein kinases; MIPC, Mannosyl-inositol-phosphorylceramide; M(IP)2C, Mannosyldiinositol-phosphorylceramide; MNF 1,2, Mitofusins 1,2; MOPS, 3-(Nmorpholino)propanesulfonic Acid; mtUPR, mitochondrial unfolded protein response; NADH, Nicotinamide adenine dinucleotide; NADPH, Nicotinamide adenine dinucleotide phosphate; NCR, Nitrogen catabolite repression; NTCB, 2nitro-5-thiocyanatobenzoic acid; PAS, Phagophore Assembly Site; PBS, Phosphate buffer saline; PCR, Polymerase Chain Reaction; PDS, Post-Diauxic Shift; PDK1, 3-phosphoinositide dependent protein kinase-1; PG, Phosphatidylglycerol; PH, Pleckstrin homology; PHS, Phytosphingosine; PI, Propidium iodide; PI3K, Phosphoinositide 3-Kinase; PIPES, Piperazine-N,N-bis(2- List of Abbreviations ! ! 20! ethanesulfonic acid); PKA, Protein Kinase A; PKB/Akt, Protein Kinase B; PKC, Protein Kinase C; PMSF, Phenylmethanesulfonyl Fluoride; p-NPP, paraNitrophenylphosphate; PP1, Protein Phosphatase type 1; PP2A, Protein Phosphatase type 2A; PS, Phosphatidylserine; Rb, Retinoblastoma; RLS, Replicative lifespan; RNA, Ribonucleic acid; ROS, Reactive oxygen species; Rv, reverse; S1P, Sphingosine-1-phosphate; SAHF, Senescence-associated heterochromatin foci; SAP, SIT4-associated proteins; SC, Synthetic complete; SD, Standard deviation; SDS-PAGE, Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis; SNARE, Soluble NSF attachment receptor; SPH, Sphingosine; SPT, Serine palmitoyltransferase; TCA, Trichloroacetic acid; TM, turn motif; TORC1, Target Of Rapamycin Complex 1; TORC2, Target Of Rapamycin Complex 2; UPR, Unfolded Protein Response; UPRE, Unfolded Protein Response Element; UV, Ultraviolet; vps, Vacuolar protein sorting; YPD-Yeast Extract Peptone Dextrose; YPL-Yeast Extract Peptone Lactate. Summary ! ! 21! Summary !Sphingolipids regulate crucial biological processes, such as stress response and apoptosis. The budding yeast Saccharomyces cerevisiae has been pivotal to uncover the impact of sphingolipid dynamics on eukaryotic cell physiology and metabolism. The bioactive sphingolipid ceramide can be generated by the de novo biosynthetic pathway or through the hydrolysis of complex sphingolipids, the last catalysed by the inositol phosphosphingolipid phospholipase C protein, Isc1p (the yeast orthologue of the mammalian neutral sphingomyelinase 2). Cells lacking Isc1p display shortened chronological lifespan (CLS), oxidative stress sensitivity and severe mitochondrial dysfunctions, which are associated with iron overload, enhanced oxidative damage to biomolecules and increased cell death by Yca1p-mediated apoptosis. Previous results demonstrated that the PP2A-like phosphatase Sit4p and the MAPK kinase Hog1p are activated in response to increased ceramide levels displayed by isc1Δ cells during aging and contribute to mitochondrial dysfunction, hydrogen peroxide hypersensitivity and shortened chronological lifespan of isc1∆ cells. In this thesis, the role of the nutrient-sensing Target Of Rapamycin Complex 1 (TORC1) and its downstream effector, the AGC protein kinase Sch9p, in isc1Δ phenotypes was assessed. This work demonstrates that the TORC1Sch9p pathway is activated in cells lacking Isc1p. The deletion of TOR1 or SCH9 abolishes the premature aging, oxidative stress sensitivity and mitochondrial dysfunctions displayed by isc1Δ cells and this is correlated with alterations in autophagic activity and flux. The protective effect of TOR1 deletion is not associated with the attenuation of Hog1p hyperphosphorylation, which was previously implicated in isc1Δ phenotypes. Importantly, Hog1p activation is responsive to ceramide by Sch9p-dependent mechanisms. Our data support a model in which Isc1p regulates mitochondrial function and chronological lifespan in yeast through the TORC1-Sch9p pathway, although Isc1p and TORC1 also seem to act through independent pathways, as isc1Δtor1Δ phenotypes are intermediate to those displayed by isc1Δ and tor1Δ cells. We then provide evidence the TORC1-Sch9p signalling pathway acts as a central axis to integrate upstream Isc1p-driven ceramide signalling signals to downstream effectors Sch9p and Sit4p. Summary ! ! 22! Macroautophagy plays an important regulatory role in mitochondrial function, stress response and aging. Since these processes are impaired in Isc1pdeficient cells, we aimed to investigate how Isc1p-driven ceramide signalling regulates macroautophagy and other mitochondrial quality control mechanisms, namely mitochondrial dynamics and the selective degradation of mitochondria by mitophagy. The results show that Isc1p deficiency reduces autophagic flux by primarily impairing vesicular trafficking, vacuolar proteolysis, acidification and morphology. These phenotypes are suppressed by downregulation of TORC1 and its downstream effectors, Sch9p and Sit4p, which integrate nutrient and stress signals from TORC1 with ceramide signalling from Isc1p. We show that the deletion of ISC1 leads to hyperactivation of mitophagy, presumably as an adaptive response to mitochondrial dysfunction, and this was correlated with loss of cell viability of isc1Δ cells. Isc1p-deficient cells also displayed higher levels of the mitochondrial fission protein Dnm1p associated with unbalanced mitochondrial fission, contributing to mitochondrial fragmentation, oxidative stress sensitivity and shortened lifespan. The deletion of TOR1, SCH9, SIT4 and HOG1 attenuated the induction of mitophagy in respiratory conditions, and this was correlated with the suppression of mitochondrial fragmentation and improvement of mitochondrial function. We further show that Isc1p and Dnm1p physically interact in vitro, further supporting a possible regulatory role for Isc1p in mitochondrial dynamics. The deletion of SIT4 extends lifespan in yeast by boosting mitochondrial function (mitochondrial catabolic derepression) and stress response. Although Sit4p functions downstream of Isc1p, the phosphatase is downregulated by TORC1 and was recently shown to control the turnover of complex sphingolipids in a Npr1p-dependent manner. However, how TORC1 effector Sit4p integrates nutrient and sphingolipid signalling is not yet fully understood. Our results demonstrate that Sit4p regulates the sphingolipid biosynthetic pathway by a transcriptional regulatory mechanism involving the ceramide synthase LAG1 and ceramidase YPC1. An increased LCBPs/LCBs ratio is observed in sit4Δ cells and ceramide metabolism is preferably shunted towards the synthesis of αhydroxylated phytoceramide species, which have been implicated on proper mitochondrial function. Importantly, overall decreased ceramide production in sit4Δ cells is correlated with reduced activation of the TORC2-Pkh1p-Ypk1p pathway and decreased Orm phosphorylation. Moreover, the Pkh1/2p-Sch9p Summary ! ! 23! pathway is downregulated in sit4Δ cells by a Snf1p-dependent mechanism. The downregulation of this pathway is associated with improved mitochondrial fitness and extended CLS upon Sit4p deficiency. Given the regulatory role of Sit4p in nutrient signalling and the regulation of the activation of the Pkh1/2p-Sch9p pathway, our results provide evidence that Sit4p is not only an effector but also a key regulator of sphingolipid metabolism and is involved in a regulatory network of interacting pathways that integrate signals from TORC1-mediated nutrient sensing with sphingolipid dynamics to regulate cell growth and longevity in yeast. In conclusion, the work described in this thesis provides new insights on the molecular mechanisms by which cellular signalling pathways are activated in response to alterations in sphingolipid metabolism and the impact on crucial cellular functions and metabolism. This clarification is pivotal to understand sphingolipid functions in more detail and define new strategies to improve human health and extend lifespan. ! ! 24! Sumário ! ! 25! Sumário Os esfingolípidos regulam processos biológicos cruciais, tais como a resposta ao estresse e apoptose. A levedura Saccharomyces cerevisiae tem sido crucial para definir o impacto da dinâmica do metabolismo dos esfingolípidos na fisiologia e metabolismo celular dos eucariontes. O esfingolípido bioactivo ceramida pode ser gerado pela via biossintética de novo ou através da hidrólise dos esfingolípidos complexos, sendo que o último processo mencionado é catalisado pela proteína Isc1p (inositol phosphosphingolipid phospholipase C protein), que é ortóloga à proteína esfingomielinase neutra tipo 2 em mamíferos. As células deficientes na proteína Isc1p apresentam envelhecimento cronológico prematuro, sensibilidade aumentada ao estresse oxidativo e disfunção mitocondrial severa, os quais estão associados com o excessivo influxo de ferro, aumento dos danos oxidativos a biomoléculas e aumento de morte celular por apoptose mediado pela metacaspase Yca1p. Estudos anteriores demonstraram que a proteína fosfatase da família das proteínas fosfatases tipo 2A e a proteína da via das proteínas cínases activadas por mitogénios (MAPK) Hog1p são activadas em resposta ao aumento dos níveis de ceramida exibidos pelas células deficientes na proteína Isc1p durante o envelhecimento, contribuindo, desta forma, para a disfunção mitocondrial severa, sensibilidade aumentada ao estresse oxidativo e envelhecimento cronológico prematuro apresentadas por este mutante. Nesta tese, o papel desempenhado pela via de sinalização mediada por nutrientes, a via Target Of Rapamycin Complex 1 (TORC1) e a sua proteína efectora, a proteína cínase Sch9p, nos fenótipos de células deficientes na proteína Isc1p foi analisado. Este trabalho demonstra que a via sinalizadora TORC1-Sch9p está mais activa no mutante isc1Δ. A deleção dos genes TOR1 e SCH9 suprime o envelhecimento cronológico prematuro, sensibilidade aumentada ao estresse oxidativo e disfunção mitocondrial severa ostentadas pelo mutante isc1Δ e tal está associado com a supressão do defeito na autofagia exibido pelo mutante. O efeito protector promovido pela deleção do gene TOR1 não está relacionado com a activação da via sinalizadora mediada pela proteína Hog1p, a qual está associada aos fenótipos exibidos pelas células deficientes na proteína Isc1p. Por outro lado, a activação de Hog1p em resposta a ceramida ocorre por Chapter I ! ! 32! eukaryotic cellular processes (Mustacchi et al., 2006). As any model system, yeast also presents some limitations as a working biological system. One major problem is the fact that yeast is unicellular and much less complex than mammalian cells (Karathia et al., 2011, Zakrajsek et al., 2011). In order to understand the etiology of multifactorial diseases, such as diabetes or metabolic syndrome, studies in yeast are limited due to the absence of multiorganic integration signalling imposed by multicellularity during evolution. For instance, assessing the role of intercellular interactions and systemic mechanisms, such as endocrine and hormone regulation, lies beyond the scope of yeast as a biological system. In neurodegenerative disease, the study of basic aspects associated with synaptic transmission and glial-neuronal processes are hindered (Khurana & Lindquist, 2010). Another important disadvantage is the occurrence of gene duplications, which may difficult the identification of new genes due to redundancy effects (van der Klei & Veenhuis, 2006). There are also specific aspects related to yeast physiology, such as the cell wall and associated signalling pathways (eg. the CWI pathway) (Levin, 2011). For instance, the lipid and protein composition of cell surface greatly varies between yeast and mammals and therefore response to stimuli may produce distinct effects on cell physiology and metabolism. In pharmacotoxicologic studies, yeast usually require higher concentrations of the compound that is often required to produce a toxic effect in mammalian cell systems, presumably due to the barrier presented by the cell wall and various efflux pumps (Matuo et al., 2012). Importantly, mammalian cells have diversified, from an evolutionary point of view, to include cellular specialization and compartmentalization without homology in yeast. Yeast mitochondria, for example, lack a typical complex I although it possesses a structurally simpler complex named external NADH dehydrogenase 2 (Ndh2p, also known as Nde2p) (Luttik et al., 1998). In this regard, studies assessing the effect of loss-of-function of complex I, such as in Parkinson´s disease (Kosel et al., 1996, Keeney et al., 2006), are limited. In the scope of this study, the elucidation of sphingolipid metabolism and dynamics and their route of synthesis were primarily done in yeast cells and have been important to uncover new functions of sphingolipids and to understand the mechanisms for sphingolipid homeostasis in human health and disease (Dickson, 2008, Huang et al., 2013). The budding yeast has been used to identify nearly all Chapter I ! ! 33! of the genes that encode for sphingolipid metabolic enzymes and many of these were critical in identifying mammalian homologs (Rego et al., 2014). Since yeast and mammals share many similarities in sphingolipid metabolism and signalling (Dickson, 2008, Rego et al., 2014), S. cerevisiae is considered a valuable model organism to study sphingolipid functions and regulation. Chapter I ! ! 34! 1.2. Aging and disease Aging is a biological process characterized by a progressive loss of metabolic and physiological integrity of an organism, which leads to a gradual deterioration of biomolecules and impairment of stress resistance mechanisms. Aging has become the primary risk factor for the development of many pathological conditions, including cancer, neurodegenerative diseases, diabetes, cardiovascular disorders, sarcopenia and liver dysfunction (Kirkwood, 2005, Tosato et al., 2007, Newgard & Pessin, 2014). Several hypotheses have been proposed to explain the etiology of the aging process. It was firstly proposed that aging is stochastic rather than programmed due to a decrease in forces of natural selection for maintenance in the post-reproductive phase of life (reviewed in (Kirkwood, 2008)). From an evolutionary point of view however, biological processes that slow or counteract the deleterious effects of aging are more advantageous and therefore expected to be preferentially selected in individuals to prolong life and give rise to more progeny. As a result of this conception, the notion of gerontogenes (genes that increase lifespan when overexpressed or mutated) and lifespan extension mechanisms have emerged and became the basis for the theory of antagonistic pleiotropy (Williams, 2001, Williams & Day, 2003). It was postulated the existence of positive evolutionary selection of genes in early life stages that confers advantageous effects but possibly adverse effects later in life (at post-reproductive phase) (Williams, 2001, Williams & Day, 2003). Cells are constantly exposed to a harmful environment throughout life. In particular, early work on DNA damage has established it as a main cause of aging due to gradual impairment of DNA repair processes and DNA loss of integrity (Chen et al., 2007). From this, aging would be genetically determined by deleterious changes on repair and maintenance mechanisms under the genetic control (Chen et al., 2007, Kirkwood, 2008, Rodriguez-Rodero et al., 2011). The disposable soma theory predicts that stochastic accumulated damage and decline of maintenance and repair systems cause aging in somatic tissues, whereas the germ line is carefully maintained and kept immortal (Kirkwood & Holliday, 1979). However, several studies have demonstrated that aging is not fully genetically Chapter I ! ! 35! programmed and also relies on the interaction with environmental factors, raising to the general notion that aging is a multifactorial phenomenon characterized by a time-dependent decline in physiological function (Kirkwood & Melov, 2011). Over recent years, the unprecedented advance in genomics and several other ´omic´ areas has provided crucial evidence on the nature and the underlying mechanisms governing the aging process. These approaches include the discovery and identification of key genes and pathways, genetic studies of heritable diseases with common denominators with premature aging and physiological and metabolic profiling that relates cellular and organismal alterations with hallmarks of aging (Mair & Dillin, 2008, Kirkwood & Melov, 2011, Rodriguez-Rodero et al., 2011, Lopez-Otin et al., 2013, Newgard & Pessin, 2014). Importantly, advances in human genetics and cellular biology allowed to scrutinize the sources of aging-related changes, the maintenance and repair mechanisms responsible for cell homeostasis, the outcome resulting from the intersection between deleterious changes and lifespan-promoting processes and finally the ability to ameliorate and retard normal aging (hormetic regulation of aging) and hence extend healthy lifespan. On this regard, aging can be envisioned as a biological process determined by the interconnectedness between genetic and biochemical processes, usually conserved in evolution, and environmental factors. Over a lifetime, this leads to a gradual accumulation of detrimental changes in molecules, cells and tissues associated with a general decline in stress tolerance functions, therefore limiting the ability of an organism to maintain homeostasis. As a result, individuals present higher risk to develop age-related diseases (cancer, cardiovascular and neurodegenerative disorders) and premature mortality (LopezOtin et al., 2013). On the scope of this thesis, some hallmarks of aging will be described, namely genomic and telomere instability, epigenetic alterations, loss of proteostasis, impaired nutrient sensing and stem cell exhaustion. 1.2.1. Genomic instability and telomere shortening Senescence is a complex process involving genetic and environmental factors affecting most physiological pathways. Among species, the average Chapter I ! ! 36! lifespan is very wide, which indicates that maximum lifetime is partially determined by specific aspects of species genotypes and ultimately the individual´s genotype. Over a lifetime, DNA is exposed to several physical and chemical factors (e.g. phthalates and UV radiation) and biological agents (viruses) that compromise genomic stability and integrity (Rodriguez-Rodero et al., 2011). In addition, increased damage promoted by reactive oxygen species (ROS), gradual impairment of DNA repair mechanisms and the accumulation of DNA replication errors during life correlates with cell senescence and aging (Rodriguez-Rodero et al., 2011). It can encompass genomic alterations (point mutations and polymorphisms), loss (or gain) of repeated DNA sequences, telomere shortening, mitogenic signals (oncogenes), gene disruption promoted by transposons and rearrangements in chromosome number (aberrant euploidy and aneuploidy) (Burhans & Weinberger, 2007, Vijg & Suh, 2013). Importantly, the identification of genes and mutations responsible for agerelated monogenic hereditary disorders (progeroid syndromes) has been critical to uncover the function of a specific group of genes in an individual’s lifespan. Progerias are characterized by a premature aging phenotype and have become a model to study aging-associated genetic changes (Rodriguez-Rodero et al., 2011). It includes Cockayne syndrome, Werner, Fanconi anemia, Bloom, RothmundThomson and Hutchinson-Gilford syndromes, ataxia-telangiectasia and xeroderma pigmentosum, which are characterized by accelerated aging promoted by mutations in genes implicated in genetic stability and DNA repair (RodriguezRodero et al., 2011). The genomic stability also includes specific mechanisms responsible for the maintenance and functionality of telomeres. Telomeres are DNA-protein complexes that cap and stabilize the ends of linear DNA strands, thus preventing chromosome instability. A correlation between telomere shortening and somatic stem cell decline during aging has been established in the past few decades (Vijg & Suh, 2013, Townsley et al., 2014). It has been shown that repetitive DNA sequences at chromosome ends shortens with age, as observed in fibroblasts and lymphocytes, due to the lack of adequate telomerase activity (Bodnar et al., 1998), but also in hematopoietic stem cells (HSC), making cells more susceptible to mutation, apoptotic cell death and reduced ability to self-renewal (Lansdorp, 1995). In fact, telomere shortening or lengthening is associated with shortened Chapter I ! ! 37! and extended lifespan in mice, respectively (Bernardes de Jesus et al., 2012). Telomerase deficiency has also been associated with age-associated diseases, such as pulmonary fibrosis, dyskeratosis congenita and aplastic anemia, due to replicative senescence and decreased regenerative capacity in tissues (Townsley et al., 2014). 1.2.2. Epigenetic alterations The emergence of human epigenetics as a key regulator in gene expression and integrity has become the basis for the so-called epigenetic theory of aging (Berdasco & Esteller, 2012). It postulates that non-adaptive epigenetic alterations in cells and tissues contribute to aging. Epigenetic regulation includes alterations in the methylated state of genes and regulatory DNA sequences, covalent modifications of histones, chromatin remodeling and the expression of regulatory non-coding RNAs (Berdasco & Esteller, 2012). Some studies have demonstrated that epimutations accumulate throughout life, leading to activation of genes normally downregulated epigenetically. In particular, some modifications, such as histone H4K16 acetylation (Dang et al., 2009), H4K20 (Sarg et al., 2002) and H3K4 trimethylation (Greer et al., 2011) and reduced H3K9 methylation or H3K27 (Sidler et al., 2014), have been associated with increased aging. Age-related epigenetic alterations also include the formation of nuclear regions called senescence-associated heterochromatin foci (SAHF) (Narita et al., 2003). At these regions, heterochromatin proteins and Retinoblastoma (Rb) protein are recruited to E2F-dependent promoters of proliferative genes, therefore repressing E2F target genes (Narita et al., 2003). The activity of DNA methyltransferases, deacetylases and histone demethylases and protein complexes involved in chromatin remodeling is altered with aging (Berdasco & Esteller, 2012). Importantly, the activity of the NADdependent deacetylase SIRT1, which belongs to the sirtuin family, decreases throughout life (Berdasco & Esteller, 2012) and Herranz et al. have shown that the overexpression of SIRT1 improves healthy lifespan but does not extend longevity (Herranz et al., 2010). Although the mechanisms are rather complex, it affects genome stability by modifying chromatin and repressing the transcription of integrated reporter genes through its intrinsic acetylase activity (Vaquero et al., Chapter I ! ! 38! 2004). Another anti-aging factor that has come to the foreground of research area in recent years is the NAD-dependent deacetylase SIRT6. This protein contributes to genomic stability (DNA repair) and regulates H3K9 and H3K56 deacetylation in telomeric regions, which is involved in NF-kB-dependent modulation of gene expression, apoptosis, and cellular senescence (Kawahara et al., 2009, Yuan et al., 2009). SIRT6-deficient cells are characterized by enhanced genomic instability due to non-clonal chromosomal aberrations and breakage as well as hypersensitivity to γ-irradiation and age-related alterations (cachexia, kyphosis, and osteopenia) (Mostoslavsky et al., 2006), and also the hyperacetylation of the H3K9 locus that enhances NF-kB signalling, which ultimately leads to cell death and aging (Kawahara et al., 2009). On the other hand, increased expression of SIRT6 prolongs life by mechanisms involving reprogramming of aging cells and alterations in metabolism, such as reduced signalling through the serum insulinlike growth factor (IGF-1) and other members of the IGF signalling pathway (Kanfi et al., 2012), whose activities are known to be lowered in long-lived mutants (Kenyon, 2010). The DNA methylation pattern of individuals with progerias is altered and resembles the ones in aged cells. More recently, a number of disorders have been documented and associated to genetic imprinting, including cancer and syndromes involving chromosomal instabilities, namely Angelman, Prader-Willi and Beckwith-Wiedemann syndromes (Adams, 2008). These disorders are caused by abnormal activation or silencing of genes by epigenetic mechanisms, usually resulting in non-adaptive alterations of the epigenetic landscape and disease (Adams, 2008, Berdasco & Esteller, 2012). 1.2.3. Loss of protein homeostasis (proteostasis) During evolution, organisms have developed stress response mechanisms to adapt to harsh environmental conditions and intrinsic stress. In particular, perturbations on protein homeostasis or proteostasis, associated with changes in calcium concentration, glucose deprivation, redox changes or ischemia, impair the function of protein control quality mechanisms and the fidelity of the protein folding (Salminen & Kaarniranta, 2010, Taylor & Dillin, 2011). As a result, the accumulation of unfolded proteins and damaged components has severe Chapter I ! ! 39! consequences for cell physiology and metabolism and contributes to aging and development of age-related and neurodegenerative diseases, such as Alzheimer´s and Parkinson´s disease (Taylor & Dillin, 2011). This observation is the biochemical basis for a wide variety of diseases termed ER storage or conformational diseases (Rutishauser & Spiess, 2002). In order to adapt to such hostile environment, the cell has evolved an adaptive and coordinated response to limit the accumulation of unfolded proteins called unfolded protein response (UPR) (Schröder & Kaufman, 2005, Kimata et al., 2006, Mori, 2009). In this mechanism, the ER-located transmembrane protein endoribonuclease IRE1 is autophosphorylated and catalyzes the cytoplasmic splicing of the XBP1 mRNA. The spliced version of this mRNA is translated into the transcription factor XBP1 that induces various genes to alleviate the ER stress condition (Mori, 2009). At this stage, the folding demand is decreased by downregulating the transcription of genes encoding secretory proteins (Travers et al., 2000), and by promoting the clearance of slowly folding or misfolded proteins through the activation of the ER-associated degradation (ERAD), ubiquitin proteasome system and autophagy (Randolph Y, 2002, Nishikawa et al., 2005, Mori, 2009). Each of these proteostastic mechanisms are capable of eliminating damaged proteins or triggering a more global response, such as cell cycle arrest or apoptosis upon severe proteotoxic stress. To enhance the folding capacity of the ER, the expression of ER resident molecular chaperones and foldases is enhanced (Nishikawa et al., 2005) and the ER increases in size in order to dilute enhanced unfolded protein load, and alleviate the folding capacity and the accumulation of misfolded proteins (Schuck et al., 2009). The aging process is characterized by the gradual impairment of ER function and proteostasis, eventually leading to the accumulation of harmful protein modifications, misfolding and aggregation of proteins, disturbances in Ca2+ homeostasis and impairment in global protein synthesis (Squier & Bigelow, 2000, Kregel & Zhang, 2007, Lindner & Demarez, 2009, Jones, 2010). Moreover, the decline in autophagic and proteasomal degradation with aging compromises protein quality control and promotes the accumulation of misfolded and oxidized proteins. Many of the key components of the UPR such as the chaperones and enzymes display reduced expression and activity, resulting in a dysfunctional ER (Salminen & Kaarniranta, 2010). Unresolved and sustained ER stress leads to a Chapter I ! ! 40! more deleterious outcome with increased protein accumulation, loss of ER function and activation of apoptotic cascades that ultimately contribute to cell death and aging (Tabas & Ron, 2011). 1.2.4. Impaired nutrient signalling The ability to utilize substrates and nutrients to fuel energy production for the homeostatic maintenance of cell-intrinsic processes is a characteristic of all living organisms. This is dictated by coordinated nutrient-sensing systems that are able to integrate metabolic and physiological signals to maintain cell and organismal homeostasis. It is well documented that these pathways become increasingly less efficient and dysfunctional, making cells more prone to develop disease. In fact, core metabolic pathways such as glycolysis, fatty acid oxidation, amino acid oxidation, lipogenesis, and ketogenesis become less efficient with aging (Newgard & Pessin, 2014). Crucial sensing protein complexes have received attention in the past few years in the field of aging, namely the nutrient and growth factors sensing mammalian Target Of Rapamycin Complex 1 (mTORC1), the AMP-dependent protein kinase (AMPK) and sirtuins (Newgard & Pessin, 2014). 1.2.4.1. mTORC1 signalling The mTORC1 is a highly conserved serine/threonine protein kinase complex activated by growth factors, mitogens and nutrients that is involved in the temporal control of cell growth by promoting transcription, ribosome biogenesis, nutrient transport, and inhibiting autophagy (Laplante & Sabatini, 2012, Johnson et al., 2013). The downregulation of mTORC1 activity, either pharmacologically (upon treatment with rapamycin) or genetically, extends the lifespan of several organisms, ranging from yeast (Medvedik et al., 2007, Pan et al., 2011), flies (Bjedov et al., 2010), nematodes (Vellai et al., 2003) alongside with mammals (Harrison et al., 2009, Miller et al., 2011). It also partly mediates the beneficial effects of calorie restriction (CR), which is currently the classic paradigm for studying the role of metabolism in aging (Blagosklonny, 2010, Sharp & Strong, Chapter I ! ! 41! 2010). There is increasing evidence showing that mTORC1 and the protein kinase S6K, one of mTORC1 effectors, contribute to aging and age-related diseases, such as type II diabetes mellitus and cancer (Manning, 2004, Inoki et al., 2005). It has been recently shown that the role of mTORC1-S6K1 signalling in the regulation of longevity is associated with insulin and IGF-1 signalling (Manning, 2004). In mouse models of reduced IGF1 signalling, namely haploinsufficiency of IGF-1 receptor and deletion of insulin receptor substrate-1 (IRS-1), lifespan extension was correlated with reduced mTORC1–S6K1 signalling (Holzenberger et al., 2003, Selman et al., 2008). Moreover, caloric restriction (CR) prolongs lifespan in several models in part due to overlapping effects of reduced mTORC1S6K1 activity and diminished IGF-1 signalling pathway (reviewed in (Fontana et al., 2010)). Although the mechanisms involved in the anti-aging effects of downregulation of mTORC1 activity are rather complex and interconnected, some studies have provided evidence that it may promote survival by reducing overall mRNA translation rate and protein synthesis. It would then relieve ER stress, preventing the accumulation of mistranslated and misfolded proteins and loss of proteostasis (Cornu et al., 2013, Johnson et al., 2013). The inhibition of mTORC1 activity also induces macroautophagy (Johnson et al., 2013), which promotes the clearance of damaged biomolecules (DNA, proteins, lipids) and organelles that can be detrimental and induce pathological changes (Cornu et al., 2013, Johnson et al., 2013). mTORC1 inhibition also improves stem cell function and therefore the ability of tissue self-renewal (see 1.2.5.). 1.2.4.2. AMPK signalling AMP-activated protein kinase (AMPK) is a highly conserved sensor of AMP and ADP levels (Mihaylova & Shaw, 2011). In glucose sensing, AMPK stimulates energy generation from glucose and fatty acids during stress and inhibits energyconsuming processes, such as protein, cholesterol and glycogen synthesis (Hardie, 2007, Mihaylova & Shaw, 2011). Compelling evidence has demonstrated however that the function of AMPK is not restricted to the maintenance of energy metabolism but also in the regulation of several homeostatic mechanisms, such as autophagy (Alers et al., 2012) and general stress response by improving tissue Chapter I ! ! 48! can be further reduced to hydroxyl radicals by Fe2+. All together, these ROS are key contributors to oxidative damage (Vendelbo & Nair, 2011) (Figure 1.2). Numerous studies provide evidence for the impact of oxidative imbalance in aging, mostly using animal models. The involvement of mtDNA in longevity pathways has received considerable attention (Shokolenko et al., 2014). Indeed, it presents a higher mutation rate, which is attributed to its close proximity to the electron transport chain. Unlike nuclear DNA, mtDNA does not present the protective effect of histones against oxidative damage but instead forms protein-DNA complexes (nucleoids) in the mitochondrial matrix (Shokolenko et al., 2014). Although this nucleoid formation makes mtDNA more impervious to oxidative damage, it remains more susceptible to ROS-mediated damage than nuclear DNA (Shokolenko et al., 2014). During lifetime, the random accumulation of age-related somatic mtDNA mutations contributes to aging (Shokolenko et al., 2014). However, an alternative hypothesis has emerged, claiming that most of the mutations are created as replication errors during embryogenesis and then undergo clonal expansion and cause a mosaic respiratory chain dysfunction in different tissues. Mosaic respiratory chain deficiency caused by clonal expansion of mtDNA mutations is ubiquitously observed in human aging (Larsson, 2010). In an elegant work, Bonawitz et al. demonstrated that in yeast expressing a deficient mitochondrial RNA polymerase amino-terminal domain (ATD) mutant (which has efficient but imbalanced mitochondrial translation), defective coupling of transcription to translation leads to increased ROS production and inhibition of respiration in the stationary phase, ultimately limiting CLS (Bonawitz et al., 2006). Therefore, life-long exposure to ROS likely results in a preferential accumulation of mtDNA damage and accelerated aging. Nowadays, the refined mitochondrial free radical theory of aging states that the mitochondrial production of ROS, such as superoxide and H2O2, results in the accumulation of damage to macromolecules and stress resistance protein systems, which in turn overwhelm the capacity of biological systems to repair themselves, resulting in an inevitable functional decline. The cumulative increase of ROS production and agingassociated mutations and alterations in mtDNA stability and integrity can impair the function of the respiratory chain and enhance ROS production. This can Chapter I ! ! 49! subsequently leads to exponentially increasing levels of mtDNA damage and oxidative stress in the cell, which ultimately culminates in cell death and aging (Shokolenko et al., 2014). More recently, the mitochondrial free radical theory of aging has been challenged. Some recent studies have demonstrated that altered ROS signalling can either have no significant or can even be beneficial for longevity. For instance, increased ROS production can extend lifespan in yeast (Pan et al., 2011) and C. elegans (Lee et al., 2010). In yeast, enhanced ROS levels during growth produce a hormetic effect and activate adaptive stress response mechanisms, preconditioning cells to better survive the stationary phase and to extend CLS (Pan et al., 2011). 1.3.2. ROS scavenging systems To live in an oxygen-containing environment, organisms had to evolve efficient antioxidant defense systems to cope with the production of ROS. Such mechanisms comprise either non-enzymatic and enzymatic components to scavenge ROS. ROS-scavenging pathways from different cellular compartments act coordinately to avoid the deleterious effects of toxic oxygen products and oxidative damage to lipids, proteins, and nucleic acids. 1.3.2.1. Non-enzymatic components of the antioxidant defense system I. Glutathione The tripeptide glutathione (γ-glutamyl-cysteinyl-glycine, GSH) is a low molecular weight thiol with an important regulatory role against oxidative damage (Davies, 2000) (Figure 1.2). It has been detected virtually in all cellular compartments such as cytosol, ER, lysosome, and mitochondria. Due to the presence of a free thiol group that can transfer two electrons, it can effectively neutralize ROS. During ROS scavenging, two molecules of glutathione are oxidized to render glutathione disulfide (GSSG) (Davies, 2000, Circu & Aw, 2012). The balance between GSH and GSSG is crucial to maintain cell´s redox state and the activity of glutathione reductase, a NADPH-dependent enzyme, is required to Chapter I ! ! 50! preserve this balance (Davies, 2000, Circu & Aw, 2012). Due to its reducing power, GSH plays an important role in diverse biological processes, including cell growth/division, signal transduction, conjugation of metabolites, enzymatic regulation, synthesis of proteins and nucleic acids, detoxification of xenobiotics and the expression of stress-responsive genes (Aquilano et al., 2014). This relevance is supported by the fact that glutathione exist in cells at high concentrations (1-10 mM) (Circu & Aw, 2012). II. Natural Antioxidants: Vitamins and Polyphenols Diet is an important non-genetic factor for the control of some diseases. More specifically, fruits and vegetables have been reported to have a protective effect due to high content of vitamins A, C, and E and polyphenol, which present antioxidant activity (Landete, 2013). Due to their ability to scavenge ROS, they slow or prevent the oxidation of other molecules by removing free radical intermediates and inhibiting other oxidation reactions (Landete, 2013). Although some inconsistencies on the importance of exogenous antioxidants in vitro and in vivo have been reported, it is generally accepted that endogenous homeostatic repair mechanisms are unable to prevent global oxidative damage, particularly with aging, thereby making sources of dietary antioxidants especially important to ameliorate the impact of cumulative oxidative stress overtime (Landete, 2013). 1.3.2.2. Enzymatic components of the antioxidant defense system I. Superoxide dismutase (SOD) Superoxide dismutase plays a central role in defense against oxidative stress in all aerobic organisms. The enzyme SOD belongs to the family of metalloenzymes and catalyzes the dismutation of O2•− to O2 and H2O2 (Figure 1.2) (Davies, 2000, Fukai & Ushio-Fukai, 2011). It is present in most of the subcellular compartments that generate activated forms of oxygen. In eukaryotic cells, three isozymes of SOD are known, namely SOD1, a copper/zinc SOD (Cu/Zn-SOD), SOD2, a manganese SOD (Mn-SOD), and SOD3, an extracellular SOD (EC-SOD) that contains copper and zinc (Fukai & Ushio-Fukai, 2011). Mn-SOD is localized in Chapter I ! ! 51! mitochondria, whereas SOD1 is found in the cytosol, and peroxisome and mitochondria (Fukai & Ushio-Fukai, 2011). S. cerevisiae possesses a Cu/ZnSOD (Sod1p), predominantly cytosolic, and a mitochondrial form of Mn-SOD (Sod2p) (Gralla, 1997). Superoxide anions can also be eliminated by metallothioneins, which are small and ubiquitous Cys-rich proteins known to be involved in ROS scavenging and metal homeostasis (Babula et al., 2012). II. Catalase and peroxidases Among antioxidant enzymes, catalase was the first enzyme to be identified and characterized in detail. It is a ubiquitous tetrameric heme-containing enzyme that catalyzes the dismutation of two molecules of H2O2 into water and oxygen (Davies, 2000) (Figure 1.2). It has high specificity for H2O2, but weak activity against organic peroxides. Catalases break down hydrogen peroxide by a twostage mechanism in which hydrogen peroxide alternately oxidizes and reduces the haem iron at the active site. In the first step, one hydrogen peroxide molecule oxidizes the haem to an oxyferryl species. In the second step, a second hydrogen peroxide molecule is used as a reductant to regenerate the enzyme, producing water and oxygen (Gralla, 1997). Catalase is located in the cytosol and the matrix of peroxisomes and mitochondria in eukaryotic cells (Gralla, 1997). Interestingly, alterations in catalase subcellular localization have been observed in association with disease and stress conditions (Zhou & Kang, 2000). In S. cerevisiae, two forms of catalase exist, namely the cytosolic catalase Ctt1p and the peroxisomal/mitochondrial catalase Cta1p (Gralla, 1997). Glutathione peroxidase belongs to the peroxidase family whose main biological role relies on the ability to prevent oxidative stress by reducing lipid hydroperoxides to their corresponding alcohols and free hydrogen peroxide to water (Arthur, 2000, Circu & Aw, 2012, Aquilano et al., 2014). Several isozymes were identified in mammals and are encoded by different genes, which vary in cellular location and substrate specificity. Glutathione peroxidase 1 (GPx1) is the most abundant form of GPx and is mostly located in the cytosol of virtually any mammalian tissue (Figure 1.2). Chapter I ! ! 52! Figure 1.2. Cellular ROS production. Mitochondria (and ER) constitute the major source of ROS production in cells. The reduction of oxygen to water through the respiratory chain proceeds via one electron flow. In the mitochondrial respiratory chain, Complex IV (cytochrome c oxidase) retains all partially reduced intermediates until full reduction is achieved. However, the redox centers in the electron transport chain may leak electrons to oxygen, partially reducing this molecule to superoxide anion (O2•−). Dismutation of O2•− produces hydrogen peroxide (H2O2), which in turn may be fully reduced to water or partially reduced to hydroxyl radical (OH•) by the Fenton reaction. Several antioxidant defense mechanisms contribute to attenuate oxidative damage to biomolecules, namely glutathione (GSH), glutathione peroxidases (GPx), catalase and superoxide dismutase (SOD). Glutathione peroxidase 2 (GPx2) is an gastrointestinal and extracellular enzyme, while glutathione peroxidase 3 is extracellular, especially abundant in the plasma. Eight different isoforms of glutathione peroxidase (GPx1-8) have been identified in humans and they require selenium for optimal catalytic and antioxidant activity (Arthur, 2000). Chapter I ! ! 53! 1.3.3. Yeast as a model to study the role of mitochondria in aging S. cerevisiae has been crucial to unravel the mechanisms by which cells respond to mitochondrial dysfunction and oxidative stress. This involves functional and structural alterations in mitochondrial morphology and turnover in response to metabolic and environmental cues elicited by cellular signalling pathways. In this section, the importance of general cell signalling pathways and mitochondrial quality control mechanisms (mitochondrial dynamics, macroautophagy and mitophagy) will be introduced as an emerging link to the aging process. 1.3.3.1. TORC1-Sch9p and RAS/PKA pathways in the regulation of mitochondrial function, stress response and aging Aging in yeast is monitored by counting the number of daughter cells generated by an individual mother cell (replicative lifespan, RLS) or by evaluating the survival of a population of post-mitotic cells (chronological lifespan, CLS) (Fabrizio & Longo, 2003). In S. cerevisiae, glucose, aminoacids and other nutrients can activate the responsive Target of Rapamycin (TOR) and the serine threonine kinase Sch9p and the Ras/PKA signalling pathways (Fontana et al., 2010) with significant impact on longevity. An important signalling pathway involved in the regulation of cell growth and survival is the highly conserved TOR pathway (Dann & Thomas, 2006). This pathway is highly conserved among organisms, ranging from flies, nematodes, protozoa alongside with mammals. In S. cerevisiae, the TOR pathway is controlled by two Ser/Thr protein kinases, Tor1p and Tor2p, which assemble into two protein complexes with distinct subunit composition and regulatory roles (Loewith et al., 2002, Dann & Thomas, 2006, Kim & Guan, 2011, Loewith & Hall, 2011). The rapamycin-sensitive TOR complex 1 (TORC1) contains either Tor1p or Tor2p and is mostly associated with the regulation of cell growth (nutrient sensing), autophagy, ribosomal and protein turnover and cell proliferation (Evans et al., 2011, Kim & Guan, 2011). The TOR complex 2 (TORC2) contains Tor2p, but not Tor1p, and mediates the proper maintenance of the cell cytoskeleton (Cybulski & Hall, 2009) and was recently implicated in the regulation of ceramide biosynthesis Chapter I ! ! 54! by a Ypk2p-dependent mechanism (Aronova et al., 2008). Furthermore, TORC2 indirectly modulates the sphingomyelinase Isc1p activity through the phosphorylation and activation of Slm1p and Slm2p (Tabuchi et al., 2006). Some studies have linked the TORC1 pathway to the regulation of mitochondrial function and yeast CLS (Bonawitz et al., 2007, Pan et al., 2011). In fact, the deletion of TOR1 or pharmacological inhibition of TORC1 with rapamycin promote oxidative stress resistance and extends CLS in yeast and other organisms (Powers et al., 2006, Bonawitz et al., 2007, Kaeberlein & Kennedy, 2011) (Figure 1.3). TORC1 is active during early stages of growth and represses the induction of stress responses and entry into the stationary phase, in part by inhibiting the Rim15p protein kinase (Wanke et al., 2005, Wei et al., 2008). Reducing TORC1 signalling at early stages of growth extends CLS by an intrinsic mechanism involving enhanced mitochondrial membrane potential and superoxide production (Pan et al., 2011). This in turn induces an adaptive response that contributes to decrease ROS production in the stationary phase and promote longevity in yeast. Associated with improved mitochondrial function, TOR1 deletion leads to increased translation of mtDNA-encoded subunits of the oxidative phosphorylation system, which increases oxygen consumption and may limit intracellular oxygen availability to produce ROS (Bonawitz et al., 2007). As observed in yeast, administration of rapamycin leaded to the downregulation of the mTOR-S6K pathway and reduction in ROS levels and oxidative DNA damage in mammalian cells (Halicka et al., 2012). Some authors have identified downstream targets of TORC1 involved in the regulation of stress response and aging, namely the AGC protein kinase Sch9p. Like other AGC proteins, Sch9p has several conserved domains: a central kinase catalytic domain, an activation loop, a turn motif (TM) and a C-terminal regulatory domain, which contains a hydrophobic motif (HM) that is phosphorylated by TORC1 (Urban et al., 2007). At the N-terminal side of the activation loop, Sch9p has a calcium-dependent C2 domain with unknown function (Urban et al., 2007). Sch9p acts as a signalling mediator, relaying upstream signals from intracellular and extracellular cues, to downstream targets by phosphorylating them on serine and/or threonine residues (Voordeckers et al., 2011, Halicka et al., 2012, Huang et al., 2012). Chapter I ! ! 55! Figure 1.3. The yeast major regulatory signalling pathways of CLS and mitochondria function. The nutrient-sensing pathways controlled by TORC1, Sch9p and Ras/PKA converge on the protein kinase Rim15p, which in turn regulates the activation of Msn2/4p and Gis1p stress-responsive transcription factors. Pkh1/2p also regulates Sch9p activity in response to sphingolipids (long chain sphingoid bases, LCBs). The downregulation of Ras/PKA and TORC1-Sch9p pathways partly mediate the anti-aging effects of CR. Sch9p has a pivotal role in oxidative stress resistance, CLS and mitochondrial function. In fact, the deletion of SCH9 gene leads to improved mitochondrial function, which contributes to oxidative resistance and CLS extension in yeast (Fabrizio et al., 2001, Wei et al., 2008, Wei et al., 2009) (Figure 1.3). Apart from sensing nutrient and stress signals from TORC1, Sch9p also regulates CLS and mitochondria function by integrating signals from long chain sphingoid bases (LCBs), a sphingolipid species (Figure 1.3). In addition to phosphorylation in the C-terminus mediated by TORC1, Sch9p is phosphorylated in a Thr570 residue in the activation loop by Pkh1/2p protein kinases, homologs of Chapter I ! ! 56! mammalian phosphoinositide-dependent protein kinase 1 (PDK1), in response to LCBs (Voordeckers et al., 2011, Huang et al., 2012). Huang et al. have recently demonstrated that the downregulation of sphingolipid synthesis induced by myriocin (an inhibitor of the first step of de novo sphingolipid biosynthetic pathway) or the deletion of PKH2 enhances CLS and improves mitochondrial function and oxidative stress resistance by Sch9p-dependent mechanisms, which involves a decrease in the activation of the Pkh1/2p-Sch9p pathway (Huang et al., 2012). Another important signalling pathway in the regulation of mitochondrial function and CLS in yeast is the Ras/PKA pathway. S. cerevisiae contains 2 RAS genes, RAS1 and RAS2, which present significant homology to the mammalian Ras. Both Ras1p and Ras2p function to activate adenylate cyclase (Tamaki, 2007). Adenylate cyclase converts ATP into cAMP, which then binds the Bcy1p protein, a regulatory subunit of protein kinase A (PKA) (Tamaki, 2007). Three genes, TPK1, TPK2, and TPK3, encode the catalytic subunit of protein kinase A. When cAMP is hydrolyzed by phosphodiesterases encoded by the PDE1 and PDE2 gene, PKA is inactivated due to binding of Bcy1p to the catalytic protein complex (Tamaki, 2007). The TORC1-Sch9p and the RAS/cAMP/PKA pathways regulate mitochondrial function and longevity by overlapping mechanisms (Wei et al., 2008, Galdieri et al., 2010) (Figure 1.3). In fact, the downregulation of both pathways extends longevity and increases stress resistance partly through the Rim15p protein kinase (Wei et al., 2008, Wei et al., 2009). Rim15p is inhibited by Sch9p, Ras/PKA and TORC1 during growth. Mechanistically, inactivation of TORC1 results in dephosphorylation of the phospho-Thr1075 located in a 14-3-3-binding site of Rim15p, thereby causing the release of Rim15p from the cytoplasmic 14-33 anchoring protein Bmh2p (Galdieri et al., 2010). For the remaining signalling effectors (PKA and Sch9p) involved in the inactivation of Rim15p, the mechanism is not yet known. When these signalling pathways are downregulated, Rim15p becomes activated and increases the expression of a variety of genes involved in G0 entry and stress response through the translocation of transcription factors Msn2/4p and Gis1p into the nucleus (Wei et al., 2008, Galdieri et al., 2010) (Figure 1.3). The core of the post diauxic shift (PDS)-element and stress responsive element (STRE)-element genes, whose expression is induced by these transcription Chapter I ! ! 57! factors, are important in response to a variety of stresses, including nutrient limitation and oxidative stress and also in trehalose and glycogen accumulation, which are required to survive under nutrient starvation (during stationary phase) and longevity in yeast (Galdieri et al., 2010). The transcription factors Msn2p/Msn4p upregulate genes encoding for antioxidant defense mechanisms, including SOD2, CTT1, SOD2, GLR1, HSP12 and DDR2, as well as the nicotinamidase PCN1 (Jamieson, 1998, Zhang et al., 2009, Galdieri et al., 2010). Pcn1p is involved in the NAD salvage pathway and favors Sir2p activation to suppress the formation of toxic replicative ribosomal DNA (rDNA) (Lesur & Campbell, 2004). Gis1p-regulated genes include stress response (HSP26, SIP18 and GRE1), the pentose phosphate pathway (TKL2, GND2), glutamate synthesis (IDP2, GDH3) and the glyoxylate cycle (MLS1, ICL1) genes (Zhang et al., 2009). Moreover, Msn2/4p and Rim15p are required for the induction of autophagy upon inhibition of PKA and Sch9p, although they are dispensable in the induction of autophagy by rapamycin (Yorimitsu et al., 2007). It was also shown that the upregulation of Rim15p is required to improve oxidative stress response and mitochondrial function and to extend yeast lifespan in CR conditions downstream of Ras/PKA, TORC1 and Sch9p pathways (Wei et al., 2008). More recently, it was shown that aminoacids threonine and valine promote increased sensitivity to general stress and aging primarily by activating the TORC1-Sch9p pathway whereas serine stimulates Sch9p through Pkh1/2p, which in turn inhibits Rim15p activity (Mirisola et al., 2014). 1.3.3.2. Mitochondrial quality control mechanisms Mitochondrial dynamics, which is dictated by a series of fusion and fission events in the mitochondrial network, is involved in the regulation of important cellular processes, namely mitochondrial metabolism, redox dynamics, apoptosis and cell death. Together with macroautophagy and the selective mitochondrial turnover (mitophagy), mitochondrial dynamics forms an interconnected quality control system that contributes to homeostatic turnover of mitochondria and sustained ATP production and removal of damaged mitochondria and oxidized biomolecules (proteins, DNA and lipids), processes that are tightly regulated to preclude aging. It is therefore conceivable that dysfunctional quality control Chapter I ! ! 64! expansion and sealing, fusion with the lysosome (vacuole in yeast), vesicle breakdown and recycling of the resulting macromolecules (Maiuri et al., 2007) (Figure 1.5). Figure 1.5. Schematic diagram of the various stages of autophagy. Autophagy induction upon mTOR inhibition, a sensor of nutrients (aminoacids), allows the activation of the ULK1 kinase complex. In nucleation, the Class III PI(3)K complex forms PI(3)P, which is necessary for formation of the isolation membrane. The membrane expands to engulf cytosolic contents in vesicle elongation, a process that requires the two ubiquitinlike conjugation steps of ATG5–ATG12 and LC3/At8p–PE. In vesicle retrieval, the transport of ATG9 between the PAS and non-PAS sites is necessary for autophagosome formation and requires ATG18. During vesicle maturation, trafficking and fusion of the fully enclosed double-membrane autophagosome to various endosomal compartments occurs, eventually fusing to the lysosome (vacuole in yeast) to form the autolysosome. In the final stage, degradation of the contents of the autolysosome occurs by the resident lysosomal enzymes. Although the process is described for mammalian cells, similar features are also observed for yeast cells. The Figure was obtained from Maiuri et al., 2007. Chapter I ! ! 65! II. Autophagy regulation and cell signalling pathways In yeast, some signalling pathways have been characterized as playing an important role in the regulation of autophagy. The regulatory proteins of these pathways are TORC1, Sch9p, PKA, and Pho85p protein kinases (Reggiori & Klionsky, 2013). Under nutrient-rich conditions, autophagy is inhibited because TORC1 is active and hyperphosphorylates Atg13p (Kamada et al., 2010, Reggiori & Klionsky, 2013), resulting in a lower affinity for the serine-threonine protein kinase Atg1p and Atg17p to induce the process (Figure 1.5). PKA and Sch9p are also involved in the regulation of Atg13p phosphorylation and localization to the phagophore assembly site (PAS) (Stephan et al., 2009), although the mechanisms involved are yet fully understood. When TORC1 activity is inhibited, either by rapamycin or starvation, Atg13p is rapidly dephosphorylated by unknown mechanisms (to yield a hypophosphorylated form of Atg13p) and interacts with Atg1p (Kamada et al., 2010, Reggiori & Klionsky, 2013). The Atg1p-Atg13p protein complex then associates with Atg17p, which is part of a ternary complex with Atg29p and Atg31p. The Atg1p-Atg13p protein complex then recruits other Atg proteins to the phagophore assembly site and controls their assembly and dynamics (Reggiori & Klionsky, 2013). Autophagosome nucleation requires a complex containing Atg9p and the class III phosphatidylinositol 3-kinase Vps34p, the latter generating phosphatidylinositol 3-phosphate (Maiuri et al., 2007, Reggiori & Klionsky, 2013) (Figure 1.5). The expansion of autophagosomal membranes involves two ubiquitin-like proteins, Atg12p and Atg8p, an E1 ubiquitin activating enzyme (Atg7p), two analogues of ubiquitin-conjugated enzymes (Atg10p and Atg3p), an Atg8p modifying protease (Atg4p), the protein target of Atg12p attachment (Atg5p) and Atg16p (Figure 1.5). In the first ubiquitination reaction, the E1-like Atg7p and the E2-like Atg10p promote the association of Atg12p with Atg5p (Suzuki et al., 2001, Suzuki et al., 2007). This conjugate subsequently interacts with Atg16p to generate pre-autophagosomal structures (Mizushima et al., 1999). In the second ubiquitin reaction, Atg8p is cleaved by the protease Atg4p and conjugated to phosphatidylethanolamine (PE) by Atg7p (E1-like) and Atg3p (Kim et al., 1999, Chapter I ! ! 66! Kirisako et al., 1999). This lipidated form of Atg8p is essential to drive proper autophagosome biogenesis (Figure 1.5). Upon completion of autophagosome formation, the Atg12p–Atg5p–Atg16p protein complex is released into the cytosol, whereas Atg8p-PE remains stably associated with the autophagosomal membranes (Kirisako et al., 2000). At this point, the outer autophagosomal membrane fuses with the vacuolar membrane to produce an autophagic body (autolysosome in mammalian cells) (Reggiori & Klionsky, 2013) (Figure 1.5). Several proteins are known to regulate the autophagosome-to-vacuole fusion step, namely the Rab protein Ypt7p (Kim et al., 1999), its GDP-exchange factor Ccz1pMon1p (Wang et al., 2003), and several SNARE proteins such as Vam3p, Vam7p, Vti1p and Ykt6p along with the class C Vps/HOPS complex, which are essential for proper tethering and fusion (Reggiori & Klionsky, 2013). The remaining singlemembrane that envelops the cargo is lysed and the population of Atg8-PE together with the enclosed cargo are released into the vacuolar lumen and degraded by resident vacuolar hydrolases (proteases, lipases, nucleases and glucosidases). The resulting degradation products are released back into the cytosol through the activity of specific membrane permeases for recycling (Reggiori & Klionsky, 2013). III. Autophagy and lifespan The relationship between CLS and autophagy is extremely complex and not fully understood. Autophagy appears to be a common downstream process of multiple cellular pathways with well-known roles in longevity regulation and disease (Madeo et al., 2010). The upregulation of autophagy extends chronological lifespan in mice, C. elegans, yeast and other organisms. Importantly, the TORC1-Sch9p and the Ras/cAMP-dependent protein kinase proteins, which integrate the network of nutrientand energy-sensing pathways that regulate autophagy, are known to be involved in proper regulation of longevity pathways (Wei et al., 2008, Galdieri et al., 2010, Evans et al., 2011, Swinnen et al., 2013). Recently, Hansen et al. found that dietary restriction and TOR inhibition in C. elegans produce an "autophagy" phenotype and that inhibiting genes required for autophagy prevents dietary restriction and TOR inhibition from extending lifespan, corroborating with this conception (Hansen et al., 2008). Chapter I ! ! 67! Screenings performed in yeast have demonstrated that mutants for genes encoding proteins of the autophagic machinery (ATG5, ATG7, ATG8, ATG12 and ATG18) are short-lived (Matecic et al., 2010) and some genes are required to extend CLS upon rapamycin treatment (ATG1 and ATG7) (Alvers et al., 2009). Suppression of autophagy by knockdown of essential autophagy genes triggers apoptosis or necrosis in cells that would otherwise survive under stress conditions (Kourtis & Tavernarakis, 2009, Suzuki et al., 2011). Autophagy appears to serve primarily a cytoprotective function by maintaining nutrient and energy homeostasis during starvation or by degrading damaged cellular components and invasive pathogens (Reggiori & Klionsky, 2013). Although autophagy is a predominantly pro-survival mechanism, it can also play a role in cell death, which is not restricted to developmental programmed cell death, but extends to cell death that occurs in many pathological conditions (Codogno & Meijer, 2005, Tsujimoto & Shimizu, 2005). Excessive autophagy induced by extreme conditions such as toxins and necrosis-triggering insults might cause uncontrollable degradation or sequestration of cells contents resulting in undesirable cell death if not properly regulated (Codogno & Meijer, 2005). IV. Mitophagy The mitochondrial theory of aging predicts that an accumulation of oxidative damage and mtDNA mutations is associated with the onset of age-associated pathologies and cell death (Shokolenko et al., 2014). Mitophagy promotes the removal of damaged/dysfunctional or oxidized mitochondria and contributes to the homeostatic maintenance of sustainable mitochondrial function and thereby allows an efficient process for ATP production and maintenance of proper cellular energetics (Seo et al., 2010). There is some evidence in yeast studies suggesting that damaged mitochondria are eliminated by mitophagy. For example, interference with F1FoATPase biogenesis in a temperature sensitive fmc1ts mutant (Priault et al., 2005) or osmotic swelling of mitochondria caused by depletion of the mitochondrial K+/H+ exchanger Mdm38p induce mitophagy (Nowikovsky et al., 2007). It is conceivable to assume that mitophagy also mitigates the deleterious effect of mtDNA mutations in heteroplasmic cells, together with mitochondrial dynamics (Figure Chapter I ! ! 68! 1.4). Supporting this view, Mao et al. have recently reinforced the important link between mitophagy and mitochondrial dynamics (Mao et al., 2013). On this basis, both processes seem to act in a coordinate manner to assure the proper connectivity of the mitochondrial network, which is an important factor that determines cell’s fate (Figure 1.4). The core autophagic machinery used is common with other types of autophagy. The requirement of several ATG genes for mitophagy has been reported by several groups and the screening of ATG mutants in S. cerevisiae allowed the identification of some ATG proteins selectively involved in mitophagy, namely Atg32p and Atg33p (Kanki et al., 2009a, Kanki & Klionsky, 2010a). Recently, the Slt2p and Hog1p mitogen activated protein kinases (MAPK) were identified as key mediators of mitophagy by regulating the phosphorylation of Ser114 and Ser119 on Atg32p (Aoki et al., 2011, Mao et al., 2011) (Figure 1.6). Importantly, Slt2p is required for proper recruitment of mitochondria to PAS (Mao et al., 2011). Whi2p, Uth1p, and Aup1p are also required for mitophagy, although the function of these proteins in mitophagy remains to be established (Bhatia-Kiššová & Camougrand, 2010, Hirota et al., 2012) (Figure 1.6). Casein kinase 2 (CK2) phosphorylates Atg32p and activates mitophagy in yeast, but the relevance for the process is not yet understood (Kanki et al., 2013). In an elegant work, Bockler and Westermann identified a complex connecting mitochondria and the ER structure termed ERMES (ER mitochondria encounter structure), which mediated the mitochondrial-ER tethering (mitochondrial-ER contacts) and is essential for the formation of the isolation membrane destined to the formation of mitophagosomes (Bockler & Westermann, 2014). In yeast studies, there are several ways to induced mitophagy. The most common is the incubation in non-fermentable medium to induce the proliferation of mitochondria (Kissova et al., 2004, Kanki et al., 2009b) (lactate or glycerol as sole carbon sources). Cells can then be grown upon nitrogen starvation conditions to induce mitophagy (Mao et al., 2011). Rapamycin has also been reported to induce the process (Mendl et al., 2011). The importance of mitophagy is demonstrated by its association with the onset of age-associated ailments, such as neurodegenerative diseases (Mijaljica et al., 2010). Chapter I ! ! 69! Figure 1.6. Mitophagy associated signalling in yeast. The MAPK Slt2p and Hog1p regulate the phosphorylation of Atg32p at s Ser114 and Ser119 residues. This phosphorylation is important for the Atg11p-Atg32p interaction. Atg11p recruits mitochondria to the PAS to be enclosed in mitophagosomes. Proteins Atg33p, Whi2p, Uth1p, and Aup1p have been reported to be important for the induction of mitophagy. However, their regulatory role remains to be established. The Figure was modified from Hirota et al., 2012. Chapter I ! ! 70! 1.4. Sphingolipids 1.4.1. Structure Sphingolipids are important structural components of cell membranes found in essentially all animals, plants and fungi, as well as some prokaryotic organisms (Hannun & Obeid, 2008). They are mostly found on the outer leaflet of the plasma membrane, although they are also present at membranes of different organelles at variable ratio (Hannun & Obeid, 2008). Furthermore, they are major constituents of lipoproteins. Sphingolipids, including sphingosine, ceramide and sphingosine-1phosphate (S1P), have emerged as core metabolites in this metabolism as they regulate a vast number of cellular processes including cell growth, adhesion, migration, senescence, apoptosis, endocytosis, and most recently, autophagy in yeast and higher eukaryotes (Dickson, 2008, Hannun & Obeid, 2008, Young et al., 2013). Since they have the ability to modulate the activity of some proteins and the activation of some signalling pathways, variations in the relative levels of different sphingolipid species result in important changes in overall cellular functions and fate (Hannun & Obeid, 2008). From a structural point of view, sphingolipids have an amphipathic nature and are composed by a long chain sphingoid base (LCB) (sphingosine (SPH) in mammals, and phytosphingosine (PHS) and dihydrosphingosine (DHS) in yeast), with the 2-amino group amide-linked to a fatty acid, thereby forming the core unit, to which polar groups are added at C1-hydroxyl group to give rise to different types of sphingolipids (Hannun & Obeid, 2008, Rego et al., 2014) (Figure 1.7). The nature of the fatty acid (carbon length, degree of unsaturation and hydroxylation) along with other modifications of the long-chain bases and the polar head group define the vast family of sphingolipids. 1.4.2. Bioactive sphingolipids and aging Sphingolipid metabolism is highly complex and interconnected, which enable cells to orchestrate different cellular responses by regulating sphingolipid interconversions (Hannun & Obeid, 2008, Rao et al., 2013). Chapter I ! ! 71! Figure 1.7. The general structure of sphingolipids. Sphingolipid are composed of LCBs, which can be linked to a fatty acid via an amide bond and to a polar head of different chemistry. The nature of the fatty acid and the polar headgroup define the extended family of sphingolipids. The Figure was modified from Rego et al., 2014. In response to both extracellular stimuli (UV, hypoxia, toxins, heat stress, etc.) and alterations in cellular physiology, the enzymes involved in sphingolipid metabolism act in a coordinated manner to regulate not only the levels of individual bioactive lipids, but also their metabolic interconversion and the physiological response generated after proper integration of cellular signalling (Hannun & Obeid, 2008). The sphingolipids ceramide, sphingosine and sphingosine-1-phosphate are the main representatives of sphingolipid metabolism and play crucial roles in the regulation of many cellular processes (Hannun & Obeid, 2008, Rego et al., 2014). The first sphingolipid to be identified was sphingosine and it exerts pleiotropic effects on protein kinases and other targets (Hannun & Obeid, 2008). Sphingosine and its related sphingoid bases have roles in regulating the actin cytoskeleton, endocytosis, cell cycle and apoptosis. Ceramide mediates many cell-stress responses that include the promotion of pro-apoptotic events and cell senescence, by modulating the activity of ceramide-activated protein kinases (e.g. PKCζ) and phosphatases [ceramide-activated protein phosphatase (CAPP), phosphoprotein 1 (PP1) and phosphoprotein 2A (PP2A)] (Hannun & Obeid, 2008, Rao et al., 2013) (Figure 1.8). On the other hand, S1P promotes cell proliferation and survival by Chapter I ! ! 72! acting in an autocrine manner on S1P receptors (Hannun & Obeid, 2008, Rao et al., 2013) (Figure 1.8). Consequently, alterations in the relative amounts of sphingosine-1-phosphate and sphingosine/ceramide have significant effects on cell physiology and metabolism and ultimately on cell fate (Figure 1.8). This explains why sphongolipid metabolism is highly regulated (Hannun & Obeid, 2008, Rao et al., 2013). Other components of the family of sphingolipids include ceramide-1-phosphate (C1P), involved in inflammation and vesicular trafficking, glucosylceramide, mostly associated with post-Golgi trafficking and drug resistance, lyso-sphingomyelin and dihydroceramide (Hannun & Obeid, 2008, Rao et al., 2013). Figure 1.8. Sphingolipid metabolism and signalling. In response to several stimuli, sphingolipid metabolism can be adaptively activated. Whereas ceramide and sphingosine are mostly associated with apoptosis and cellular senescence by activating protein kinases and phosphatases, sphingosine-1-phosphate (S1P), conversely, promotes cell proliferation and protection from apoptosis by acting on receptors, usually in an autocrine manner. Therefore, the regulation of ceramide/S1P is crucial to determine cell fate. The Figure was adapted from Hannun & Obeid, 2008. Chapter I ! ! 73! The importance of sphingolipid signalling derives from the early recognition of their contribution in the pathobiology of human cancers and other human ailments such as diabetes and heart disease, microbial infections, neurological and immune dysfunctions (Kolter & Sandhoff, 2006, Ozbayraktar & Ulgen, 2009, Kolter, 2011, Hla & Dannenberg, 2012, Young et al., 2013). Additionally, sphingolipids have been implicated in the regulation of stress responses and longevity. In mammals, ceramide levels and SMase activity increase and induce a senescent phenotype morphologically and biochemically (inhibition of cell cycle progression, inhibition of the PKC/PLD pathway, induction of Rb dephosphorylation and inhibition of the pro-proliferative AP1 activity) (Obeid et al., 1993, Obeid & Venable, 1997, Hannun & Luberto, 2000). Senescence and growth inhibition were also correlated with the inhibition of telomerase transcription and activity by ceramide (Ogretmen et al., 2001). Yeast mutants lacking Ydc1p (dihydroceramidase) are characterized by increased CLS whereas the overexpression of YDC1 triggers mitochondria and vacuolar fragmentation, apoptosis and accelerated aging in yeast (Aerts et al., 2008). Genes involved in sphingolipid metabolism (LAG1, YPC1, YSR3, IPT1, and LCB5) show variable expression in senescent and apoptotic cells (Laun et al., 2005). More recently, it was shown that the downregulation of sphingolipid biosynthesis increases yeast CLS in part due to a reduction in long-chain bases (LCBs) mediated activation of the Pkh1p-Sch9p pathway (Huang et al., 2012). Furthermore, ceramide synthase (Lag1p) and LCB kinase (Lcb4p) activities decrease upon entry into the stationary phase, leading to a large increase in the levels of LCBs (Lester et al., 2013). Over the past few decades, many studies attempted to unveil the role of sphingolipids in signal transduction pathways involved in the regulation of these biological processes. However, the mechanisms by which sphingolipids control many aspects of cell physiology and metabolism remains to be characterized in more detail. 1.4.3. Yeast sphingolipid metabolism The sphingolipid metabolism is highly conserved between yeast and mammals and shares a similar spatial organization (Figure 1.9). Chapter I ! ! 80! interaction with the phosphate group, and as a general base in the catalytic reaction, respectively (Okamoto et al., 2003). Ideally, Isc1p presents optimal activity around pH 7.5 and very low activity in acidic and very alkaline conditions (Okamoto et al., 2003). Interestingly, Isc1p is post-translationally regulated by translocation from the ER into mitochondria upon the transition from fermentative to the respiratory metabolism during the so-called post-diauxic shift (PDS) (Vaena de Avalos et al., 2004). This appears to be associated with the regulation of mitochondrial sphingolipid metabolism and function, namely the generation of α-hydroxylatedphytoceramides, which are necessary for proper function of this organelle (Kitagaki et al., 2007, Kitagaki et al., 2009). The regulatory role of Isc1p on mitochondrial function is reinforced by the fact that Isc1p-deficient cells exhibit a growth defect upon the transition to the PDS phase associated with the inability to upregulate genes required for respiratory metabolism, despite presenting intact mitochondrial DNA (Kitagaki et al., 2007, Kitagaki et al., 2009). In fact, these mutants display defective growth on non-fermentable carbon sources and enhanced frequency of petite formation. The mutant strain also displays decreased levels of the mitochondrial ETC complex IV (cytochrome c oxidase, COX) subunits Cox3p and Cox4p and this is correlated with decreased COX activity and oxygen consumption in isc1Δ cells (Almeida et al., 2008, Kitagaki et al., 2009, Barbosa et al., 2011). II. Isc1p-driven signalling in mitochondrial function and aging Initial studies have demonstrated that Isc1p is implicated in the regulation of important cellular processes, namely responses to osmostress, heat stress and genotoxic agents. Almeida et al. have firstly reported that Isc1p is involved in the regulation of oxidative stress resistance, mitochondrial function and chronological lifespan (Almeida et al., 2008). In fact, isc1Δ cells display shortened CLS, increased hydrogen peroxide sensitivity, which appear to be associated with mitochondrial dysfunction and increased iron uptake (Almeida et al., 2008). Together with mitochondrial dysfunction, it promotes a vicious cycle of oxidative damage to biomolecules and contributes to cell death by apoptosis (Almeida et al., 2008, Kitagaki et al., 2009, Barbosa et al., 2011). Chapter I ! ! 81! In an attempt to dissect possible signalling pathways governing isc1Δ phenotypes, important downstream targets of Isc1p were identified and implicated in the regulation of mitochondrial function and CLS (Figure 1.11). Lipidomic analysis showed specific changes in sphingolipids in Isc1p-deficient cells during aging, including decreased dihydrosphingosine levels and an increase of very long chain base species, namely dihydro-C26-ceramide and phyto-C26-ceramide, the latter raising the possibility of activation of ceramide-dependent protein phosphatases. On this basis, it was shown that the deletion of SIT4 gene, which encodes for the catalytic subunit of type 2A ceramide-activated protein phosphatases and presents considerable homology to human protein phosphatase 6 (PP6) (Bastians & Ponstingl, 1996), supresses mitochondrial dysfunctions, oxidative stress sensitivity and shortened CLS of Isc1p-deficient cells (Barbosa et al., 2011) (Figure 1.11). Figure 1.11. The sphingomyelinase Isc1p modulates ceramide signalling to regulate mitochondrial function and CLS in yeast. Isc1p-mediated signalling controls the activity of the ceramide-activated PP2A-like phosphatase Sit4p and the MAPK Hog1p, by regulating ceramide content in yeast. Among other signalling pathways, the yeast high osmolarity-responsive (HOG) MAPK pathway, the HOG pathway, was also studied. It is activated in response to hyperosmotic stress via two independent osmosensing branches, the Sln1p and the Sho1p branches, which in turn control the phosphorylation of the Chapter I ! ! 82! main effector, the MAPK Hog1p (Rep et al., 2000, Alepuz et al., 2001, Proft et al., 2001, de Nadal et al., 2003, de Nadal et al., 2004, O'Rourke & Herskowitz, 2004, Proft et al., 2006). It was demonstrated that sphingolipids are also able to modulate the HOG pathway in response to the inibition of the de novo biosynthetic pathway or depletion of ergosterol (Tanigawa et al., 2012). Hog1p activation is deleterious for cells devoid of Isc1p since ceramide signalling increases Hog1p phosphorylation and the deletion of HOG1 attenuates isc1Δ phenotypes (Barbosa et al., 2012). ! ! 83! Scope of the thesis Sphingolipids are important components of membranes in eukaryotes. More recently, they are recognized as important bioactive biomolecules implicated in the regulation of signalling transduction and important biological processes. Several lines of evidence suggest an intricate interplay between sphingolipid metabolism and TOR signalling. In particular, TOR signalling is involved in the regulation of mitochondrial function, oxidative stress response and chronological lifespan (CLS) in yeast. Since isc1Δ cells exhibit mitochondrial dysfunction, hydrogen peroxide sensitivity and premature aging, we aimed to uncover the role of the nutrient-sensing Target Of Rapamycin Complex 1 (TORC1) and its downstream effector, the AGC protein kinase Sch9p, in these phenotypes. The involvement of ceramide signalling in the regulation of important biological functions, such as autophagy and related processes, and mitochondrial quality control mechanisms, has been recently addressed in mammals. On this basis, we also aimed to determine how the Isc1p sphingomyelinase-driven ceramide signalling regulates macroautophagy, mitophagy and mitochondrial dynamics and evaluate its impact on longevity in yeast. The deletion of SIT4 extends lifespan in yeast by boosting mitochondrial function (mitochondrial catabolic derepression) and stress response. Although the PP2A-like protein phosphatase Sit4p functions downstream of Isc1p, the phosphatase is downregulated by TORC1 and was recently shown to control the turnover of complex sphingolipids in a Npr1p-dependent manner. However, how Sit4p integrates nutrient and sphingolipid signalling to regulate lifespan is not yet established. On this basis, we aimed to establish a functional association between Sit4p and sphingolipid metabolism and evaluate the impact on cell survival. ! ! 84! ! ! 85! Science goes from question to question; big questions, and little, tentative answers. The questions as they age grow ever broader, the answers are seen to be more limited. George Wald CHAPTER II Chapter II ! ! 86! Reduced TORC1 signalling abolishes mitochondrial dysfunctions and shortened chronological lifespan of Isc1p-deficient cells Vitor Teixeira, Tânia C. Medeiros, Rita Vilaça, Pedro Moradas-Ferreira and Vítor Costa Keywords: Isc1p; Tor1p; Sch9p; oxidative stress; chronological aging, mitochondrial function, ROS, ceramide. Chapter II ! ! 87! Abstract The target of rapamycin (TOR) is an important signalling pathway on a hierarchical network of interacting pathways regulating central biological processes, such as cell growth, stress response and aging. Several lines of evidence suggest a functional link between TOR signalling and sphingolipid metabolism. Here, we report that the TORC1-Sch9p pathway is activated in cells lacking Isc1p, the yeast orthologue of mammalian neutral sphingomyelinase 2. The deletion of TOR1 or SCH9 abolishes the premature aging, oxidative stress sensitivity and mitochondrial dysfunctions displayed by isc1Δ cells and this is correlated with the suppression of the autophagic flux defect exhibited by the mutant strain. The protective effect of TOR1 deletion, as opposed to that of SCH9 deletion, is not associated with the attenuation of Hog1p hyperphosphorylation, which was previously implicated in isc1Δ phenotypes. Our data support a model in which Isc1p regulates mitochondrial function and chronological lifespan in yeast through the TORC1-Sch9p pathway although Isc1p and TORC1 also seem to act through independent pathways, as isc1Δtor1Δ phenotypes are intermediate to those displayed by isc1Δ and tor1Δ cells. We also provide evidence that TORC1 downstream effectors, the type 2A protein phosphatase Sit4p and the AGC protein kinase Sch9p, integrate nutrient and stress signals from TORC1 with ceramide signalling derived from Isc1p to regulate mitochondrial function and lifespan in yeast. Overall, our results show that TORC1-Sch9p axis is deregulated in Isc1pdeficient cells, contributing to mitochondrial dysfunction, enhanced oxidative stress sensitivity and premature aging of isc1Δ cells. Chapter II ! ! 88! 2.1. Introduction Sphingolipids are ubiquitous structural components of eukaryotic cell membranes, and their bioactive metabolites (sphingosine, sphingosine-1phosphate, ceramide, ceramide-1-phosphate and lyso-sphingomyelin) are known to act as second messengers in the transduction and regulation of signalling pathways (Dickson & Lester, 1999, Jenkins & Hannun, 2001, Spiegel & Milstien, 2003, Dickson, 2008, Hannun & Obeid, 2008, van Meer et al., 2008). Sphingosine (and related sphingoid bases) and ceramide are involved in the regulation of actin cytoskeleton organization, endocytosis, degradation of nutrient permeases, apoptosis, cell senescence and cell cycle arrest whereas sphingosine-1phosphate plays a key role in proliferation, mitogenesis, cell migration, cell survival and inflammation (in higher eukaryotes) (Hannun & Obeid, 2008). Thus, subtle variations on the relative amounts of sphingosine-1-phosphate and sphingosine/ceramide are expected to determine cell fate in response to environmental or metabolic stresses. The importance of sphingolipids is recognized by the fact that sphingolipid signalling is implicated in the pathobiology of human cancers and other human diseases such as diabetes and heart disease, microbial infections, neurological and immune dysfunctions (Kolter & Sandhoff, 2006, Ozbayraktar & Ulgen, 2009, Kolter, 2011, Hla & Dannenberg, 2012, Young et al., 2013). Sphingolipids metabolism and their route of synthesis are highly conserved from yeast to mammalian cells. Studies using the budding yeast Saccharomyces cerevisiae have served in many ways to foster our understanding of sphingolipid dynamics and their role in the regulation of cell cycle, cell integrity, endocytosis, cytoskeleton dynamics and protein turnover (Dickson, 2008, Hannun & Obeid, 2008). Additionally, sphingolipids have been implicated in the regulation of stress responses and longevity. For instance, yeast mutants lacking Ydc1p (dihydroceramidase) are characterized by increased chronological lifespan (CLS) whereas the overexpression of YDC1 triggers mitochondria and vacuolar fragmentation, apoptosis and accelerated aging in yeast (Aerts et al., 2008). Genes involved in sphingolipid metabolism (LAG1, YPC1, YSR3, IPT1, and LCB5) show variable expression in senescent and apoptotic cells (Laun et al., 2005). Chapter II ! ! 89! More recently, it was demonstrated that the downregulation of sphingolipid synthesis increases yeast CLS in part due to a reduction in long-chain bases (LCBs) mediated activation of Sch9p, the yeast homologue of mammalian ribosomal S6K protein kinase (Huang et al., 2012). Furthermore, ceramide synthase (Lag1p) and LCB kinase (Lcb4p) activities decrease upon entry into the stationary phase, leading to a large increase in the levels of LCBs (Lester et al., 2013). We have previously demonstrated that Isc1p, the yeast orthologue of mammalian neutral sphingomyelinase-2 (nSMase2) responsible for the hydrolysis of complex inositol phosphosphingolipids to produce ceramide, is implicated in oxidative stress resistance and chronological lifespan in yeast. Isc1p-deficient cells display shortened CLS, oxidative stress sensitivity and impaired redox and iron homeostasis (Almeida et al., 2008). Analogous to the role of ceramide and ceramide-activated protein phosphatases in regulating mammalian cell apoptosis, Isc1p acts upstream of Sit4p, the catalytic subunit of protein phosphatase related to type 2A protein phosphatases (PP2A) in yeast. Indeed, SIT4 deletion restores mitochondrial function of isc1Δ cells, increasing oxidative stress resistance and extending CLS (Barbosa et al., 2011). The activation of the HOG (High Osmolarity Glycerol) pathway is also deleterious for isc1Δ cells since ceramide signalling increases the phosphorylation of the Hog1p mitogen activated protein kinase (MAPK) and the deletion of HOG1 attenuates the phenotypes of Isc1p-deficient cells (Barbosa et al., 2012). Recent studies also link ceramide to other important signalling pathways involved in the regulation of cell growth and survival, namely the TOR (Target of Rapamycin) pathway. This pathway is highly conserved among organisms, ranging from flies, nematodes, protozoa alongside with mammals (Raught et al., 2001, Dann & Thomas, 2006, de Virgilio & Loewith, 2006, Laplante & Sabatini, 2012, Johnson et al., 2013, Markaki & Tavernarakis, 2013). In S. cerevisiae, the TOR pathway is controlled by two Ser/Thr protein kinases, Tor1p and Tor2p, which assemble into two protein complexes with distinct subunit composition and regulatory roles (Loewith et al., 2002, Kim & Guan, 2011, Loewith & Hall, 2011). The rapamycin-sensitive TOR complex 1 (TORC1) contains either Tor1p or Tor2p and is mostly associated with the regulation of cell growth (nutrient sensing), autophagy, ribosomal and protein turnover and cell proliferation (Evans et al., Chapter II ! ! 96! Figure 2.3. Increased mitochondrial coupled respiration imparted by reduced TORC1 signalling extends lifespan in isc1Δ cells. S. cerevisiae BY4741, isc1Δ, tor1Δ and isc1Δtor1Δ cells were grown in SC-medium to the PDS phase, treated with 10 µM 2,4-dinitrophenol (DNP; square) or vehicle (DMSO; circle) and kept in the medium at 26°C. The viability was determined by standard dilution plate counts and expressed as the percentage of the colony-forming units at time T in relation to T0. Values are mean ± SD of at least three independent experiments. The linear plot was used for better visual assessment. 2.2.3. Hyperpolarization and fragmentation of the mitochondrial network in isc1Δ cells are suppressed by TOR1 or SCH9 deletion To get further insights into alterations in mitochondrial function operating on isc1Δ cells, we assessed the mitochondrial membrane potential (Δψm), a parameter that has been used to monitor changes on bioenergetics as a key indicator of cell health or injury (Nicholls, 2004). Chapter II ! ! 97! For this purpose, yeast cells were labeled with a mitochondria-specific voltage-dependent dye, 3,3-dihexyloxacarbocyanine iodide [DiOC6(3)], which aggregates and preferentially accumulates into functional mitochondria, and analyzed by flow cytometry (Figure 2.4.A). At PDS phase, cells lacking Isc1p displayed enhanced Δψm when compared to parental cells, which is consistent with mitochondrial hyperpolarization. In contrast, tor1Δ and sch9Δ cells had a slightly lower Δψm, which has been associated with mild mitochondrial uncoupling (Pan et al., 2011). Both TOR1 and SCH9 disruption in isc1Δ cells reversed the mitochondrial hyperpolarization and decreased the Δψm to values similar to those observed in the respective single mutants. This probably increases cell survival in isc1Δtor1Δ and isc1Δsch9Δ cells, since isc1Δ cells die by caspase-dependent apoptosis upon oxidative stress and during cell aging (Almeida et al., 2008) and mitochondrial hyperpolarization has been associated with the activation of a mitochondrial dependent apoptotic pathway, which initially involves a transient hyperpolarization followed by depolarization of the mitochondrial membrane and release of cytochrome c from the mitochondria into the cytosol (Kroemer et al., 2007). The mitochondrial membrane potential plays a key role in the regulation of mitochondrial morphology and alterations on this parameter were demonstrated to impact on mitochondrial dynamics (Legros et al., 2002, Detmer & Chan, 2007, Berman et al., 2008, Oliveira, 2012). To assess changes in mitochondrial network dynamics, yeast cells expressing a mitochondrial-targeted DsRed protein were analyzed by fluorescence microscopy. At the exponential phase, the mitochondrial network was not yet fully developed and no significant differences were observed between the parental and isc1Δ cells (Figure 2.4.B). This was expected since cells are undergoing fermentative growth at this phase. However, in the PDS phase, loss of Isc1p led to the formation of a typical punctuate pattern contrasting with the tubular and well-organized network observed in healthy parental cells (Figure 2.4.B). This structural alteration has been associated with mitochondrial fragmentation and observed in cells undergoing apoptotic cell death. The normal tubular mitochondrial network was restored upon disruption of TOR1 or SCH9 in isc1Δ cells (Figure 2.4.B), suggesting that TORC1 and its downstream target Sch9p are also implicated in the regulation of mitochondrial dynamics by promoting network fragmentation. Chapter II ! ! 98! Figure 2.4. The mitochondrial membrane hyperpolarization and decreased autophagic flux contribute to mitochondrial dysfunction and impairment of mitochondrial dynamics in isc1Δ cells. A. S. cerevisiae BY4741, isc1Δ, tor1Δ and isc1Δtor1Δ, sch9Δ and isc1Δsch9Δ cells were grown in SC-medium to the PDS phase, stained with the potential-sensitive dye 3,3dihexyloxacarbocyanine iodide [DiOC6(3)] for 30 min and analyzed by flow cytometry. Treatment of the parental strain (BY4741) with FCCP (carbonyl cyanide 4- (trifluoromethoxy)phenylhydrazone) was used as a positive control (depolarizing event). Values are mean ± SD of at least three independent experiments. ****p<0.0001; ***p<0.001; **p<0.01. B. Yeast cells transformed with pYX222-mtDsRed were grown to Chapter II ! ! 99! the exponential and PDS phases and analyzed by fluorescence microscopy, as described in Experimental Procedures. Live cells were visualized by fluorescence microscopy. A representative experiment out of three is shown. Scale bar: 5 µm. C. S. cerevisiae BY4741, isc1Δ, tor1Δ and isc1Δtor1Δ, sch9Δ and isc1Δsch9Δ cells carrying pRS416 GFP-ATG8 were grown to the exponential phase in SC-medium and treated with either rapamycin (200 ng/mL) or DMSO (vehicle) for 3 h. Proteins were analyzed by immunoblotting, using anti-GFP antibody. D. The autophagic flux was calculated by the ratio between the free GFP signal and the sum of free GFP and GFP-Atg8p signals. Values are mean ± SD of at least three independent experiments ****p<0.0001; ***p<0.001; **p<0.01. E. S. cerevisiae BY4741 and isc1Δ cells carrying pRS416 GFP-ATG8 were grown to PDS phase, washed twice with water and then maintained in water. Proteins were analyzed by immunoblotting, using anti-GFP antibody. Previous studies have demonstrated that autophagy has an important role in maintaining proper mitochondrial function and dynamics since autophagydefective mutants present severe mitochondrial dysfunctions (Suzuki et al., 2011, Aung-Htut et al., 2013). In particular, the regulation of the mitochondrial membrane potential appears to be crucial to regulate autophagic flux. In addition, accumulating evidence show that defects in autophagy compromise mitochondrial dynamics (Lee et al., 2012). Since isc1Δ cells present similar phenotypic features and TORC1, a negative regulator of autophagy, is activated in this mutant strain, we evaluated if these cells present autophagy defects. For this purpose, we have monitored the processing of GFP-Atg8p. Once autophagy is induced, GFP-Atg8p is recruited to drive autophagosome biogenesis and then delivered to the vacuole inside the autophagic body. Whereas Atg8p is degraded by resident vacuolar hydrolases, the GFP moiety is relatively resistant to proteolysis. Therefore, the appearance of free GFP signal is indicative of autophagy induction. To induce autophagy, yeast cells were treated with rapamycin. Under these conditions, the autophagic flux was significantly lower in Isc1p-deficient cells (35%) when compared to the parental strain (60%) (Figure 2.4.C-D). Notably, a slower migrating band (above free GFP) was observed in isc1Δ cells, both under basal conditions and upon rapamycin treatment (Figure 2.S3). It probably results from an incomplete or aberrant processing of GFP-Atg8p, possibly due to vacuolar dysfunction (defective Pep4p-dependent proteolytic activity) or alterations in vacuolar morphology upon deletion of ISC1 (Seeley et al., 2002). In isc1Δtor1Δ Chapter II ! ! 100! and isc1Δsch9Δ double mutants, the autophagic flux increased to values close to those observed for the respective tor1Δ and sch9Δ single mutants (Figure 2.4.CD). It should be noted that rapamycin is predicted to have still some effect on tor1Δ cells because there is yet some functional TORC1 signalling. In fact, TORC1 complex is sensitive to rapamycin due to binding of FKBP-rapamycin complex to subunits of TORC1. Thus, TORC1 is functional (although with reduced activity) in tor1Δ cells since it still contains Tor2p (which may replace Tor1p) in its composition. These results implicate TORC1 and its downstream effector, Sch9p, in the deregulation of autophagy, possibly contributing to mitochondrial network fragmentation and impairment of oxidative stress resistance and CLS in isc1Δ cells. To provide further evidence that autophagy is impaired in isc1Δ cells, we have also monitored the autophagic flux during chronological aging. The rate of viability loss during aging of this mutant is very high when cells are grown in SCmedium (Figure 2.2.A). Thus, to avoid unspecific changes that may occur due to cell death, we assessed autophagy under conditions of calorie restriction (cells grown to PDS phase, washed and maintained in water overtime). We have previously shown that calorie restriction increases CLS in both isc1Δ and parental cells, but isc1Δ mutants still exhibit a premature aging phenotype (Barbosa et al., 2011). The results show that the autophagic flux increased in parental cells aged for 3-5 days, but it was significantly reduced in isc1Δ cells (Figure 2.4.E). Overall, the data suggests that autophagy is impaired in this mutant strain. 2.2.4. TOR1 and SCH9 deletion in isc1Δ cells decreases ROS levels, catalase A deficiency and apoptotic cell death Apoptosis and aging has been extensively associated with enhanced ROS production (Raftopoulou, 2005, Simm & Brömme, 2005, Kregel & Zhang, 2007, Marchi et al., 2012). Thus, the improvement of mitochondrial function and/or antioxidant defenses may decrease mitochondrial ROS production or increase its detoxification, leading to CLS extension. To test this hypothesis, ROS levels were measured by flow cytometry using early stationary phase cells stained with dihydroethidium (DHE), a molecular probe sensitive to superoxide radicals. The results show that ROS levels were low in parental, tor1Δ and sch9Δ cells but Chapter II ! ! 101! approximately 50% of isc1Δ cells were DHE-positive (Figure 2.5.A). In isc1Δtor1Δ and isc1Δsch9Δ mutants, ROS levels were higher than in parental cells but significantly lower when compared to Isc1p-deficient cells (by approximately onehalf), suggesting that TOR1 and SCH9 deletion increases survival in isc1Δ cells by decreasing ROS generation. Figure 2.5. The deletion of TOR1 or SCH9 decreases ROS production, attenuates catalase A deficiency and diminishes apoptotic cell death in isc1Δ cells. A. S. cerevisiae BY4741, isc1Δ, tor1Δ and isc1Δtor1Δ, sch9Δ and isc1Δsch9Δ cells were grown to early stationary phase (day 1 in the CLS assay), stained with the dihydroethidium (DHE) for 10 min and analyzed by flow cytometry. Values are mean ± SD of at least three independent experiments. ***p<0.001; **p<0.01. B. Yeast cells were grown to the PDS phase and catalase activity was detected in situ after non-denaturing polyacrylamide gel electrophoresis, using the H2O2/peroxidase system, as described in Experimental Procedures. A representative experiment out of three is shown. C. BY4741, isc1Δ, tor1Δ and isc1Δtor1Δ cells were double stained with DiOC6(3) and PI (propidium iodide) and analyzed by flow cytometry, as described in Experimental Procedures. Representative histograms are shown. Increased ROS levels have been associated with homeostatic imbalance partially dictated by impaired cellular antioxidant defences. Hence, we Chapter II ! ! 102! hypothesized that the improvement of the mitochondrial function and dynamics in isc1Δtor1Δ and isc1Δsch9Δ cells may contribute to upregulate antioxidant defence mechanisms and decrease ROS levels during the aging process. It was previously shown that isc1Δ cells fail to induce CTA1 gene expression in the PDS phase (Kitagaki et al., 2009) and display a low activity of Cta1p (Barbosa et al., 2011), the catalase A form present in mitochondria and peroxisomes. Moreover, CTA1 overexpression partially suppresses isc1Δ phenotypes (Barbosa et al., 2011). Our results show that Cta1p activity was partially restored in isc1Δtor1Δ and isc1Δsch9Δ cells (Figure 2.5.B). In sch9Δ and isc1Δsch9Δ cells, Ctt1p (cytosolic catalase) activity was not detected, which is consistent with the fact that the Sch9p kinase is directly or indirectly involved in the transcriptional control of CTT1 expression in yeast (Pascual-Ahuir & Proft, 2007). The analysis of superoxide dismutase activity did not reveal changes in Sod1p (cytosolic form) or Sod2p (mitochondrial form) activity upon the deletion of TOR1 or SCH9 in isc1Δ cells (data not shown). To evaluate if the decrease in ROS production and enhancement of antioxidant defenses (Cta1p) contribute to decrease apoptotic cell death in isc1Δtor1Δ mutants, early stationary phase cells were labeled with DiOC6(3)/propidium iodide (PI). The analysis of yeast cells by flow cytometry allows the definition of four distinct populations: healthy cells (DiOC6(3)- positive/PI-negative), early apoptosis (DiOC6(3)-negative/PI-negative), late apoptosis (DiOC6(3)-positive/PI-positive) and necrosis (DiOC6(3)-negative/PIpositive). The results (Figure 2.5.C) clearly show that the large majority of parental and tor1Δ cells remained healthy whereas 61% of the isc1Δ cell population was already undergoing early (15%) or late (46%) apoptosis. In the isc1Δtor1Δ double mutant, however, a significant decrease of apoptotic markers was observed, particularly at late stages of apoptosis, where a reduction of approximately 50% (from 46% to 24%) was detected. 2.2.5. SCH9 deletion but not reduced TORC1 signalling in isc1Δ cells attenuates Hog1p activation The hyperactivation of the HOG signalling pathway was previously implicated in the premature aging and mitochondrial dysfunction exhibited by Chapter II ! ! 103! Isc1p-deficient cells (Barbosa et al., 2012). Thus, we hypothesized that the suppression of isc1Δ phenotypes by TOR1 and SCH9 deletion could be associated with the modulation of the HOG pathway. To address this question, Hog1p phosphorylation was monitored by Western blotting using an anti-phosphop38 antibody that recognizes dually phosphorylated Hog1p, the active form of this kinase (Millar et al., 1995, Smith et al., 2004). As illustrated in Figure 2.6.A, Hog1p phosphorylation was increased in isc1Δ cells when compared to parental cells. In contrast, phosphorylated Hog1p could not be detected in tor1Δ cells. Notably, Hog1p phosphorylation levels in isc1Δtor1Δ cells were even higher to those detected in isc1Δ cells, suggesting a functional interplay between the HOG and TORC1 signalling pathways in isc1Δ cells. Figure 2.6. The deletion of TOR1 increases Hog1p phosphorylation without affecting its cytosolic localization whereas SCH9 disruption diminishes Hog1p phosphorylation in isc1Δ cells. A. Hog1p activation was monitored in BY4741, isc1Δ, tor1Δ and isc1Δtor1Δ cells by immunoblotting, using anti-phospho-p38 antibody (top panel) that detects the phosphorylated form of Hog1p, or anti-Pgk1p (loading control) as primary antibodies. A representative blot out of three is shown. B. S. cerevisiae BY4741, isc1Δ, tor1Δ and isc1Δtor1Δ cells expressing the consensus Rlm1p binding sequences fused to a LacZ reporter (2xRLM1-LacZ), were grown to the exponential phase and the β-galactosidase activity was measured as described in Experimental Procedures. Values are mean ± SD Chapter II ! ! 104! of at least three independent experiments. ****p<0.0001; ***p<0.001; *p<0.05. C. Hog1p activation was monitored in BY4741, isc1Δ, sit4Δ and isc1Δsit4Δ cells by immunoblotting, as described in A. A representative blot out of three is shown. D. Hog1p activation was monitored in BY4741, isc1Δ, sch9Δ and isc1Δsch9Δ cells by immunoblotting, as described in A (left panel). BY4741 and sch9Δ cells were grown to the exponential phase and treated with either 10 µM C2-ceramide or DMSO (vehicle) for 1 h (right panel). A representative blot out of three is shown. It was previously shown that Hog1p activation leads to its import into the nucleus where it phosphorylates the Msn2p/Msn4p, Hot1p, Sko1p and Smp1p transcription factors to promote adaptation to stress conditions (Marquez et al., 1998, Reiser et al., 1999, Rep et al., 1999, Rep et al., 2000, Alepuz et al., 2001, Proft et al., 2001, de Nadal et al., 2003, de Nadal et al., 2004, O'Rourke & Herskowitz, 2004, Proft et al., 2006). Thus, we have also monitored Hog1p cell localization by fluorescence microscopy in cells expressing HOG1-GFP (Figure 2.S4). Similarly to parental cells, Hog1p was present in the cytoplasm of isc1Δ, tor1Δ and isc1Δtor1Δ mutants, implying that TORC1 may not regulate Hog1p localization. We have also evaluated the activation of the Cell Wall Integrity (CWI) pathway, since we have previously reported a Hog1p-dependent activation of Slt2p, a MAPK of the CWI pathway in isc1Δ cells (Barbosa et al., 2012). For this purpose, we monitored the activation of Rlm1p, a transcription factor regulated by Slt2p, by measuring β-galactosidase activity in cells expressing a LacZ reporter under the control of Rlm1p promoter. Consistent with the hyperactivation of Hog1p and induction of the Rlm1p-driven LacZ reporter, β-galactosidase activity was increased by 1.7and 4.3-fold in isc1Δ and isc1Δtor1Δ cells, respectively (Figure 2.6.B). It was previously shown that the protein phosphatase Sit4p is negatively regulated by TORC1 signalling (Jacinto et al., 2001) but activated by ceramide (Nickels & Broach, 1996). Moreover, the deletion of SIT4 suppresses isc1Δ phenotypes (Barbosa et al., 2011). Notably, Hog1p phosphorylation increased in sit4Δ cells and it was exacerbated in isc1Δsit4Δ cells (Figure 2.6.C), as observed in isc1Δtor1Δ cells (Figure 2.6.A). This suggests that the phosphorylation of Hog1p is also regulated by a Sit4p-dependent mechanism. Remarkably, the deletion of SCH9 decreased Hog1p phosphorylation in isc1Δ cells (Figure 2.6.D). It was Chapter II ! ! 105! previously reported that ceramide signalling increases Hog1p phosphorylation (Barbosa et al., 2012). Thus, we tested if Sch9p regulates the HOG pathway in response to ceramide. As previously reported, C2-ceramide treatment increased Hog1p phosphorylation in parental cells (Figure 2.6.D). In contrast, the deletion of SCH9 completely abolished Hog1p phosphorylation upon treatment with ceramide (Figure 2.6.D). These results suggest that Sch9p is acting upstream of Hog1p in response to ceramide signalling. Overall, the results show that the deletion of TOR1 or SIT4 does not suppress isc1Δ phenotypes through the attenuation of Hog1p hyperactivation and Sch9p appears to regulate the activation of the HOG pathway in response to ceramide in Isc1p-deficient cells. 2.3. Discussion The TORC1 pathway is a well-established nutrient response pathway that modulates aging and age-related diseases (Johnson et al., 2013). Here we provided evidence that TORC1 signalling is deregulated in cells lacking Isc1p, the yeast orthologue of mammalian neutral sphingomyelinase. Isc1p-deficient cells exhibit increased TORC1 activity, which is detrimental for this mutant. In fact, the deletion of TOR1 alleviates the premature aging, oxidative stress sensitivity and mitochondrial dysfunctions of isc1Δ cells. However, isc1Δtor1Δ cells exhibit lower resistance to oxidative stress, shortened CLS, slightly impaired mitochondrial function, and higher levels of ROS and apoptotic cell death markers compared to the tor1Δ mutant strain, suggesting that TORC1-independent mechanisms also contribute to isc1Δ phenotypes. In agreement, the overexpression of ISC1 does not suppress the rapamycin hypersensitivity of tor1Δ mutants, suggesting that Isc1p is not acting downstream of TORC1 (Figure 2.S5). Several lines of evidence suggest an intricate interplay between sphingolipid metabolism and TOR signalling. Both TORC1 and TORC2 control the biosynthesis of sphingolipids through regulation of Orm1p and Orm2p, two evolutionarily conserved integral membrane proteins of the endoplasmic reticulum (Breslow et al., 2010, Liu et al., 2012, Shimobayashi et al., 2013). However, TORC1 and TORC2 signalling seems to function independently in the regulation of Chapter II ! ! 112! considered 100%) for the CLS assay; treated vs. untreated cells for H2O2 resistance). Values are mean ± SD of at least three independent experiments. 2.4.3. Enzymatic activities and oxygen consumption For enzyme activities, yeast cells were harvested by centrifugation for 5 min at 4,000 r.p.m. (4ºC). Cells were then ressuspended in 50 mM potassium phosphate buffer (pH 7.0) containing protease inhibitors (Complete, Mini, EDTAfree Protease Cocktail Inhibitor Tablets; Boehringer Mannhein) and phosphatase inhibitors (50 mM sodium fluoride, 5 mM sodium pyrophosphate, 1 mM sodium orthovanadate). Total protein extracts were obtained by mechanical disruption through vigorous shaking of the cell suspension in the presence of glass beads for 5 min. Short pulses of 1 min were applied followed by 1 min incubation on ice. Cell debris was removed by centrifugation at 13,000 r.p.m. for 15 min and protein content was determined by the method of Lowry, using bovine serum albumin as a standard. Catalase activity was analyzed in situ, in the presence of 3,3′- diaminobenzidine tetrahydrochloride, using the H2O2/peroxidase system (Conyers & Kidwell, 1991). Cytochrome c oxidase (COX) activity was determined by measuring cytochrome c oxidation (Poyton et al., 1995). β-galactosidase activity determination was performed as previously reported (Barbosa et al., 2012). Oxygen consumption rate was measured for 3 x 108 cells in PBS buffer (pH 7.4), using an oxygen electrode (Oxygraph, Hansatech). Data was analyzed using the Oxyg32 V2.25 software. 2.4.4. Mitochondrial membrane potential, ROS levels and cell death The mitochondrial membrane potential was measured using the potentialsensitive dye DiOC6(3). Briefly, 2x106 cells were ressuspended in sample buffer [10 mM 2-(N-morpholino) ethanesulfonic acid, 0.1 mM MgCl2 and 2% (w/v) glucose, pH 6.0]. DiOC6(3) (Molecular Probes) was added to a final concentration of 1 nM. The cell suspension was then incubated for 30 min at 26°C, collected by centrifugation and washed twice with PBS. Fluorescence was measured on the FL-1 channel of a Becton Dickinson FACS Calibur Analytic Flow cytometer with excitation and emission settings of 488 nm and 515–545 nm, respectively, without Chapter II ! ! 113! compensation. For the quantification of ROS levels, 5x106 cells were ressuspended in PBS and the superoxide anion sensitive probe dihydroethidium (DHE, Molecular Probes) was added to a final concentration of 5 µM. Cells were incubated for 10 min at 26°C, pelleted by centrifugation, washed twice with PBS and analyzed by flow cytometry with excitation and emission settings of 488 nm and ≥670 nm (FL-3 channel), without compensation. For the characterization of the cell death process, cells were dually stained with DiOC6(3) (1 nM) and propidium iodide (PI, 2 µg/mL, Molecular Probes) to evaluate mitochondrial membrane polarization and the plasma membrane integrity, respectively. Cells were incubated for 30 min at 26°C and harvested as previously described. After suitable compensation, fluorescence was measured by flow cytometry at different wavelengths: excitation/emission at 488/525 nm for DiOC6(3) (FL-1 channel), and at 536/600 nm for PI (FL-3 channel). Data treatment was performed using the FlowJo software (Tree Star). 2.4.5. Western Blot analysis To evaluate the TORC1-dependent C-terminal phosphorylation of Sch9p, cells transformed with pJU676 (expressing SCH9-5HA) were grown in SC-medium lacking uracil to the exponential phase. NTCB-chemical fragmentation analysis was done as described (Urban et al., 2007). Proteins were then analyzed by SDSPAGE using 10% polyacrylamide gels and blotted onto a nitrocellulose membrane (GE Healthcare, Buckinghamshire, United Kingdom). Immunodetection was performed using rabbit anti-HA (Sigma-Aldrich) at a 1:1,000 dilution as primary antibody, goat anti-rabbit IgG-peroxidase (Sigma-Aldrich) at a 1:5,000 dilution as secondary antibody, and the Lumigen HRP chemiluminescent substrate (GE Healthcare, RPN2109, Buckinghamshire, United Kingdom). To monitor Hog1p phosphorylation, yeast cells were grown to the exponential phase and protein extracts (50 µg) were separated by SDS-PAGE and blotted onto a nitrocellulose membrane. The membrane was incubated with the primary antibodies rabbit anti-phospho-p38 MAPK (Cell Signalling Technology, Denver, USA) at a 1:500 dilution or mouse anti-Pgk1p (Invitrogen, Carlsbad, USA) at a 1:30,000 dilution. Subsequently, the membrane was incubated with the secondary antibodies, anti-rabbit IgG-peroxidase (Sigma-Aldrich) at a 1:5,000 Chapter II ! ! 114! dilution, or anti-mouse IgG-peroxidase (Molecular Probes) at a 1:3,000 dilution. Immunodetection was performed as described above. The evaluation of LCBs mediated Phk1/2p-dependent phosphorylation of Sch9p was performed as described (Huang et al., 2012). To assess alterations in autophagic flux, cells were grown to the exponential phase in SC-medium and treated with either rapamycin (200 ng/mL, (Sigma-Aldrich) or DMSO (vehicle, Sigma-Aldrich) for 3 hours. For CLS assay in water, cells were grown to the PDS phase, washed twice with water and then maintained in water overtime. Total protein extracts (30 µg) were analyzed in similar conditions as previously described. The membrane was incubated with the primary antibodies mouse antiGFP (Roche, Basel, Switzerland) at a 1:3,000 dilution or mouse anti-Pgk1p (Invitrogen, Carlsbad, USA) at a 1:30,000 dilution. Subsequently, the membrane was incubated with the secondary antibody anti-mouse IgG-peroxidase (Molecular Probes) at a 1:3,000 dilution. Immunodetection was done as described. 2.4.6. Fluorescence microscopy For mitochondrial morphology analysis, cells transformed with pYX222mtDsRed were grown in SC-medium lacking histidine to the exponential or PDS phase. To assess Hog1p cell localization, cells expressing HOG1-GFP were grown in SC-medium lacking uracil to the exponential phase and stained with the fluorescent dye 4’,6’-diamidino-2-phenylindole, DAPI (Molecular Probes, Invitrogen, 2.5 µg/mL) to label the nucleus. Live cells were observed by fluorescence microscopy (AxioImager Z1, Carl Zeiss). Data image stacks were deconvolved by QMLE algorithm of Huygens Professional v3.0.2p1 (Scientific Volume Imaging B.V.). Maximum intensity projection was used to output final images using ImageJ 1.45v software. 2.4.7. Statistical analysis Data are expressed as mean values ± SD of at least three independent experiments. Values were compared by Student’s t-test. The 0.05 probability level was chosen as the point of statistical significance throughout. Statistical analyses were carried out using GraphPad Prism Software v5.01 (GraphPad Software). Chapter II ! ! 115! Acknowledgements We are grateful to Dr. David E. Levin (Goldman School of Dental Medicine, Boston University, Boston, USA), Dr. Robbie Loewith (University of Geneva, Switzerland), Dr. Francesc Posas (Universitat Pompeu Fabra, Barcelona, Spain), Dr. Yusuf Hannnun (Stony Brook University, Health Science Center, Stony Brook, New York, USA) and Dr. Paula Ludovico (ICVS, Universidade do Minho, Portugal) for generously providing plasmids and other reagents used in this study and fruitful discussion. We would like to thank Catarina Leitão (AFCU, IBMC) for technical support on flow cytometry and Paula Sampaio (ALM, IBMC) for technical assistance and data treatment on fluorescence microscopy. This work was supported by FEDER (Fundo Europeu de Desenvolvimento Regional) through the program “Programa Operacional Fatores de Competitividade-COMPETE”, by FCT (Fundação para a Ciência e Tecnologia) and by "Programa Operacional Regional do Norte (ON.2 – O Novo Norte)", through the projects PESTC/SAU/LA0002/2013-FCOMP-01-0124-FEDER037277 and NORTE-07-0124FEDER-000001. V.H.F.T. (SFRH/BD/72134/2010) and R.V. (SFRH/BD/48125/2008) were supported by FCT fellowships. Chapter II ! ! 116! 2.5. Supplementary Information Table 2.5.1. Saccharomyces cerevisiae strains used in this study. Strain Genotype Source BY4741 Mata his3Δ1, leu2Δ0, met15Δ0, ura3Δ0 [pJU676, pYX222-mtDsRed, pRS416-GFPATG8, pRS416-HOG1-GFP, pLGΔ3122xRLM1-LacZ, pYES2, pYES2ISC1] EUROSCARF isc1Δ BY4741 isc1Δ::KanMX4 [pJU676, pYX222mtDsRed, pRS416-GFP-ATG8, pRS416HOG1-GFP, pLGΔ312-2xRLM1-LacZ] EUROSCARF tor1Δ BY4741 tor1Δ::KanMX4 [pYX222-mtDsRed, pRS416-GFP-ATG8, pRS416-HOG1-GFP, pLGΔ312-2xRLM1-LacZ, pYES2, pYES2ISC1] This study isc1Δtor1Δ BY4741 isc1Δ::LEU2 tor1Δ::KanMX4 [pYX222-mtDsRed, pRS416-GFP-ATG8, pRS416-HOG1-GFP, pLGΔ312-2xRLM1LacZ] This study sch9Δ BY4741 sch9Δ::KanMX4 [pYX222-mtDsRed, pRS416-GFP-ATG8] EUROSCARF isc1Δsch9Δ BY4741 isc1Δ::LEU2 sch9Δ::KanMX4 [pYX222-mtDsRed, pRS416-GFP-ATG8] This study sit4Δ BY4741 sit4Δ::KanMX4 EUROSCARF isc1Δsit4Δ BY4741 isc1Δ::URA3 sit4Δ::KanMX4 Barbosa et al. (2011) Harboring plasmids are shown in square brackets Chapter II ! ! 117! BY4741 (DMSO) BY4741 (Rap) isc1Δ (DMSO) isc1Δ (Rap) 0 20 40 60 80 100 % Viable cells *** * *** Figure 2.S1. The Pkh1/2p-dependent Sch9p phosphorylation does not contribute to isc1Δ phenotypes. The Pkh1/2p-mediated phosphorylation of Sch9p at Thr570 was monitored in BY4741 and isc1Δ cells by immunoblotting, using anti-phospho-Thr570Sch9p antibody (top panel) or anti-Pgk1p (loading control) as primary antibodies. A representative blot is shown. Figure 2.S2. Inhibition of TORC1 signalling with rapamycin suppresses oxidative stress sensitivity displayed by isc1Δ cells. Yeast cells were grown in SC-medium to the early exponential phase (OD600=0.3), pre-incubated with either 200 ng/mL rapamycin (Rap) or vehicle (DMSO) for 3h and then treated with 1.5 mM H2O2 for 60 min. Cell viability was determined as described in Experimental Procedures. ***p<0.001; *p<0.05. ! Phosphorylated-Sch9p (Thr570) Pgk1p BY4741 isc1∆ Chapter II ! ! 118! ! Figure 2.S3. The autophagic flux is decreased in isc1Δ cells. S. cerevisiae BY4741 and isc1Δ cells carrying pRS416 GFP-ATG8 were grown to the exponential phase in SCmedium and treated with either 200 ng/mL rapamycin (Rap) or DMSO (vehicle) for 3 hours. Proteins were analyzed by immunoblotting, using anti-GFP antibody. The Western blot shown here is a replicate of that shown in Figure 2.4.C (only for BY4741 and isc1Δ cells) with a longer exposure time. A slower migrating band (above free GFP) was observed in isc1Δ cells, both under basal conditions and upon rapamycin treatment. A similar upper band (marked with an arrow) was observed in parental cells upon treatment with rapamycin. Notably, this upper band was more abundant in isc1Δ cells when compared to the one observed in parental cells. It probably results from impaired or aberrant GFP-Atg8p processing, possibly due to vacuolar dysfunction (defective Pep4pdependent proteolytic activity) or alterations in vacuolar morphology upon deletion of ISC1, as mentioned in the main text. Thus, we have considered only the higher migrating band as free GFP (generated from processing of GFP-Atg8p) for autophagic flux quantification (Fig. 4D). GFP BY4741 (DMSO) BY4741 (Rap) isc1Δ (DMSO) isc1Δ (Rap) GFP-Atg8p Chapter II ! ! 119! ! ! ! ! Figure 2.S4. The deletion of TOR1 does not alter Hog1p localization in isc1Δ cells. S. cerevisiae BY4741, isc1Δ, tor1Δ and isc1Δtor1Δ cells transformed with pRS416-HOG1GFP were grown to the exponential phase and live cells were visualized by fluorescence microscopy for GFP and DAPI, as described in Experimental Procedures. BY4741 cells treated with 0.7 M NaCl for 20 min were used as positive control to monitor Hog1p translocation to the nucleus. Scale bar: 5 µm. Chapter II ! ! 120! Figure 2.S5. The rapamycin sensitivity of the tor1Δ mutant was not suppressed by ISC1 overexpression. S. cerevisiae BY4741 and tor1Δ cells carrying pYES2 or pYES2ISC1 were grown in SC-medium to the exponential phase and then diluted to OD600=0.1. Fivefold dilution series were spotted on glucose or galactose media containing either rapamycin (50 ng/mL, dissolved in DMSO) or DMSO (vehicle) and cells were grown at 26°C for 5 days. ! ! ! ! ! ! ! ! ! ! ! ! ! 121! One of the major challenges for (...) who wish to work in the domain of (...) biology is becoming conversant with the daunting intricacies of existing biological knowledge. Questions about the origin, function, and structure of living systems have been pursued by nearly all cultures throughout history, and the work of the last generations has been particularly fruitful. The knowledge of living systems resulting from this research is far too detailed and complex for any one human to comprehend. Lawrence Hunter CHAPTER III Chapter III ! ! 128! Figure 3.1. Isc1p-deficient cells exhibit vacuolar fragmentation and impaired autophagic flux. A. BY4741 and isc1Δ cells expressing GFP-Atg8p were grown in SC glucose-medium to exponential phase and exposed to rapamycin (200 ng/mL) for 3h. Vacuoles were visualized by FM 4-64 staining (red) and GFP-Atg8p dynamics was observed by fluorescence microscopy. Scale bar: 5 µm. B. BY4741 and isc1Δ cells expressing Pho8Δ60 were grown in YPD to exponential phase and treated as described in A. The Pho8Δ60 activity was normalized to the activity of WT cells treated with rapamycin, which was set to 100%. Values are mean ± SD of at least three independent experiments. ****p<0.0001; ***p<0.001. C. The processing of the Cvt pathway marker protein prApeI to mApeI was analyzed by immunoblotting in BY4741, isc1Δ, vac8Δ and isc1Δvac8Δ cells grown to exponential phase and treated as described in A. Next, we corroborated these results in a different strain background (W303) by measuring the activity of a truncated form of alkaline phosphatase (Pho8Δ60), which is converted to an enzimatically active form at the vacuoles upon induction of autophagy (Yorimitsu et al., 2007). ALP activity is routinelly used as a quantitative measure of autophagy induction (Yorimitsu et al., 2007). Rapamycintreated parental cells showed an increase in Pho8Δ60 activity, which was suppressed in atg1Δ cells where autophagy is blocked. Pho8Δ60 activity was lower in isc1Δ cells (Figure 3.1.B), supporting that the induction of autophagy is impaired in the mutant strain and that this phenotype is not strain specific. The cytoplasm-to-vacuole targeting (Cvt) pathway is a vesicular transport mechanism that employs most of the macroautophagy machinery to deliver several hydrolases to the vacuole, including the percursor aminopeptidase I (prApeI) (Scott et al., 1996, Wang & Klionsky, 2003). We therefore sought to determine whether isc1Δ cells are also defective in this process by monitoring the processing of prApeI to the mature form (mApeI). We observed that prApeI accumulated in isc1Δ cells upon rapamycin treatment, a feature already observed at basal conditions when compared to parental cells (Figure 3.1.C). It was also analyzed the prApeI accumulation phenotype in a vac8Δ background. Vac8p is required for proper induction of the Cvt pathway but not for macroautophagy (Abeliovich et al., 2000). Therefore, in a vac8Δ background, prApeI can only be converted to its mature form when autophagy is induced. Non-treated vac8Δ cells Chapter III ! ! 129! accumulated prApeI and the protein was rapidly processed to mApeI after incubation with rapamycin. In contrast, the accumulation of prApeI was still observed in rapamycin-treated isc1Δvac8Δ cells (Figure 3.1.C). Together, these results suggest that Isc1p is required for proper regulation of macroautophagy and the Cvt pathway in yeast. To identify the steps at which autophagy is impaired in Isc1p-deficient cells, we have firstly examined vacuolar proteolysis. Pep4p constitutes the major vacuolar protease driving the maturation of vacuolar proteases. Accordingly, pep4Δ cells display defective GFP-Atg8p processing when autophagy is induced (Kirisako et al., 1999). To evaluate if Pep4p-dependent vacuolar proteolytic activity is impaired in isc1Δ cells, we monitored the maturation of carboxypeptidase Y (CPY), which occurs in a Pep4p-dependent manner (Figure 3.2.A). Most of the precursor CPY (prCPY) was converted to the mature form (mCPY) in parental cells, both under basal conditions and after incubation with rapamycin. In isc1Δ cells, prCPY maturation was significantly decreased, even after rapamycin treatment, suggesting that Pep4p activity is reduced in the mutant strain (Figure 3.2.A). To provide further evidence that reduced Pep4p-mediated vacuolar hydrolysis contributes to reduced autophagic flux in isc1Δ cells, prApeI processing (which is Pep4p-dependent) was monitored in cells overexpressing PEP4. Indeed, the defective maturation of prApeI in isc1Δ cells was alleviated by PEP4 overexpression (Figure 3.2.B). To determine whether Isc1p is involved in autophagosome-to-vacuole fusion, we employed a protease protection assay to evaluate the sensitivity of the prApeI to an externally added protease in rapamycin-treated isc1Δ cells (Krick et al., 2010). After removing non-lysed cells by low speed centrifugation, we separated the total lysate (Figure 3.2.C, T, lane 1) in a 13,000 x g supernatant (S13, lane 2) and pellet fraction (P13, lane 3). With this method, we also used internal controls to determine the efficiency of spheroplast lysis (Pgk1p), and to verify the integrity of intracellular compartments including the vacuole (prCPY) and, by extension, autophagosomes. The resistance of the lumenal propeptide (prCPY) to trypsin in the absence of Triton X-100 (Figure 3.2.C, compare lanes 3 and 4 versus lane 5) reflects the integrity of the vacuole (and autophagosomes). In this assay, the membrane fraction (P13) was untreated (lane 3) or treated with protease alone (lane 4) or with protease and detergent (lane 5). Chapter III ! ! 130! Figure 3.2. Defective vacuolar proteolysis and acidification contributed to reduced autophagic flux in isc1Δ cells. A. The Pep4p-dependent processing of prCPY to mCPY was analyzed by immunoblotting in BY4741 and isc1Δ cells grown to exponential phase and treated with either DMSO Chapter III ! ! 131! (vehicle) or rapamycin (200 ng/mL) for 3h. B. The Pep4p-dependent processing of prApeI to mApeI was analyzed by immunoblotting in BY4741 and isc1Δ overexpressing PEP4 treated as described in 2A. Quantification of mApeI/(prApeI+mApeI) is shown. Values are mean ± SD of at least three independent experiments. **p<0.01. C. Cells devoid of Isc1p were grown to the exponential phase and treated with rapamycin (200 ng/mL) for 2h. After removing non-lysed cells, the total lysate (T, lane 1) was centirfuged to obtain a 13,000 x g supernatant (S13, lane 2) and pellet fraction (P13, lane 3). The P13 fraction was then ressupended in PS200 buffer (lane 3) and trypsin-digested with and without detergent (Triton X-100, lanes 4-5). The efficiency of spheroplast lysis (Pgk1p) and the integrity of intracellular compartments (vacuolar prCPY) were also monitored. *mApeI-proteaseprotected intermediate form of prApeI. D. For the detection of secreted CPY, BY4741 and isc1∆ cells were grown on YPD plates and overlaid with a nitrocellulose filter. The vps4∆ mutant was used a positive control for CPY secretion. Secreted CPY was immunodetected with anti-CPY. E. Vacuolar acidification of BY4741 and isc1∆ cells grown to exponential phase was monitored using quinacrine staining and analyzed by flow cytometry. Unprobed and probed cells are depicted in red and blue, respectively. Representative histograms are shown. F. Concanamycin A-sensitive vacuolar ATPase (VATPase) activity using acridine orange was analyzed in BY4741 and isc1∆ isolated vacuoles. Reactions were started in the presence of 15 µM acridine orange. When indicated, MgATP was added. Concanamycin A was added at the same time (dashed line). At the end of the assay, FCCP (15 mM) was added to dissipate the proton gradient. If prApeI is membrane-protected (inside Cvt vesicles or within the vacuole), then prApeI is only cleaved by the exogenously added protease in the presence of detergent. Our results suggest the presence of a membrane-enclosed intermediate form of ApeI (Figure 3.2.C, compare lane 1 versus lane 3) that was fully cleaved to the mature form by trypsin in the presence of detergent (Figure 3.2.C, compare lanes 3 and 4 versus lane 5). This is consistent with our own observations that an intermediate form of ApeI accumulates in total extracts of isc1Δ cells, which may result from altered processing of ApeI, leading to its accumulation within the vacuoles (Figure 3.1.C). We also perfomed a similar assay using cells expressing GFP-Atg8p (Figure 3.S1). In parental cells, GFPAtg8p was fully degraded by trypsin, indicating it was readily accessible in the cytosol. This means GFP-Atg8p enclosed in autophagosomes is rapidly converted to free GFP in parental cells due to the rapid turnover of autophagosomes fusing Chapter III ! ! 132! to the vacuoles. In isc1Δ cells, a small pool of GFP-Atg8p remained proteaseprotected, probably enclosed by a membrane as it was degraded in the presence of the detergent (Figure 3.S1). Together, these results suggest that disturbed autophagic flux of isc1Δ cells is not attributed to alterations in the fusion of vesicles to the vacuole but rather to impaired Pep4p-mediated proteolysis. Defective processing of ApeI through the Cvt pathway may also result from alterations in vacuolar protein sorting and trafficking, which is a crucial part of autophagy regulation. To assess if isc1Δ cells exhibit vesicular trafficking defects, we have evaluated whether CPY secretion is altered in isc1∆ cells. Several vacuolar protein sorting (vps) mutants that fail to deliver CPY to the vacuole display an aberrant secretion of CPY (Rothman & Stevens, 1986, Raymond et al., 1992). Our results show that isc1∆ cells, in contrast to parental cells, exhibited CPY secretion, similarly to the vps4∆ mutant (Figure 3.2.D). Taken together, these data indicate that the accumulation of prCPY in isc1∆ cells may arise not only from defective Pep4p-mediated processing but also from impaired Golgi-to-vacuole trafficking. The mutual observation that the Cvt pathway and CPY trafficking are affected in isc1Δ cells, already under basal conditions, indicates that vesicular trafficking may contribute to reduced autophagic flux in this mutant strain. The overexpression of VAM3 (Darsow et al., 1997) and YKT6 (Kweon et al., 2003), which are involved in endocytic trafficking and vacuolar docking/fusion, had no effect on Cvt pathway and CPY secretion (Figure 3.S2.A and 3.S2.B), suggesting that the trafficking defect in isc1∆ cells does not occur at a step involving Vam3p and Ykt6p (t-SNARE/v-SNARE pairing). Proper vacuolar acidification has been shown to be pivotal for autophagy as it is required for the activation of vacuolar proteases (Nakamura et al., 1997) and the autophagosome-vacuole fusion step (Abeliovich et al., 2000). This led us to assess whether vacuolar acidification is affected in isc1∆ cells by using the fluorescent weak base quinacrine, whose accumulation preferentially relies on acidic compartments. The results show that quinacine staining in isc1Δ cells was lower to that of parental cells (Figure 3.2.E). In agreement with these results, by measuring concanamycin A-sensitive ATPase activity in isolated vacuoles, we also show that isc1Δ vacuoles exhibited lower V-ATPase activity than wild-type vacuoles (Figure 3.2.F), indicating that isc1Δ cells display defective vacuolar acidification. Chapter III ! ! 133! 3.2.2. The Sit4p protein phosphatase and the TORC1-Sch9p pathway contribute to vesicular trafficking and vacuolar defects of isc1∆ cells The ceramide-activated protein phosphatase Sit4p was previously implicated in isc1∆ phenotypes (Barbosa et al., 2011). Therefore, we analyzed whether Sit4p underlines the autophagic defects of this mutant. The deletion of SIT4 in isc1∆ cells increased GFP-Atg8p processing (Figure 3.3.A) and the maturation of prApeI (Figure 3.3.B) after rapamycin treatment. The deletion of TOR1 or SCH9, which suppresses the reduced autophagic flux of isc1∆ cells (Teixeira et al., 2014), also increased prApeI maturation in this mutant (Figure 3.S3.A). Next, we assessed if Sit4p and the TORC1-Sch9p pathway are also implicated in vesicular trafficking and vacuolar defects observed in isc1∆ cells. Lower Pep4p activity, the exacerbated CPY secretion and the defective vacuolar acidification exhibited by isc1∆ cells were abolished in isc1∆sit4∆ double mutants (Figure 3.3.C-E), as well as in isc1∆tor1∆ and isc1∆sch9∆ cells (Figure 3.S3.B-D). Isc1p-deficient cells displayed trafficking defects and vacuolar fragmentation associated with impaired autophagy, similar to vam3∆ (Darsow et al., 1997) and ypt7∆ mutants (Kim et al., 1999). Notably, the restoration of the autophagic flux was also associated with the suppression of vacuolar fragmentation in isc1∆sit4∆, isc1∆sch9∆ cells and to a lesser extent, in isc1∆tor1∆ cells (Figure 3.S4). Together, these results are consistent with the idea that the protein machinery involved in vesicular trafficking events plays an important role in the modulation of autophagy in the mutant strain. We have previously shown that isc1∆ mutants have increased levels of some ceramide species coupled with the activation of Sch9p and Sit4p and we proposed that the TORC1-Sch9p signalling pathway acts as a central axis to integrate upstream Isc1p-driven ceramide signalling signals to downstream effectors Sch9p and Sit4p (Teixeira et al., 2014). Combined with the results presented here, this suggests that ceramide plays a regulatory role in autophagy. Chapter III ! ! 134! Figure 3.3. The deletion of SIT4 suppressed Pep4p-dependent vacuolar proteolysis and the vacuolar protein sorting defects in isc1Δ cells. A. BY4741, isc1Δ, sit4Δ and isc1Δsit4Δ cells carrying GFP-Atg8p were grown in SC-medium to exponential phase and treated with either DMSO (vehicle) or rapamycin (200 ng/mL) for 3h. Proteins were analyzed by immunoblotting, using anti-GFP antibody. The quantification of autophagic flux (GFP/(GFP+GFP-Atg8p)) is shown. ****p<0.0001; A Chapter III ! ! 135! ***p<0.001; **p<0.01. B. The processing of the Cvt pathway prApeI to mApeI was analyzed by immunoblotting in BY4741, isc1Δ, sit4Δ and isc1Δsit4Δ cells treated as described in 3A. C. Specific Pep4p activity was measured in protein extracts of BY4741, isc1Δ, sit4Δ and isc1Δsit4Δ cells grown to exponential phase in SC-medium. Values were normalized to parental cells (set as 100%). Values are mean ± SD of at least three independent experiments. ***p<0.001; **p<0.01. D. CPY secretion was analyzed in BY4741, isc1Δ, sit4Δ and isc1Δsit4Δ cells as described in figure 3.2.D. E. Vacuolar acidification of isc1Δsit4Δ cells was analyzed using quinacrine staining, as described in 3.2.E. Unprobed and probed cells are depicted in red and blue, respectively. Representative histograms are shown. To explore this hypothesis, we enzymatically decreased ceramide load in isc1Δ cells by ectopically increasing YPC1 ceramidase expression (Mao et al., 2000). Overexpression of YPC1 efficiently restored Pep4p activity in isc1∆ cells to that of WT values (Figure 3.S5.A), but did not suppress the CPY vacuolar protein sorting defect (Figure 3.S5.B). This indicates that decreasing ceramide substrates of Ypc1p can overcome some but not all the mechanisms deregulated in the isc1Δ mutant. Taken together, our data suggest that ceramide species increased in isc1∆ cells signal through Sit4p and the TORC1-Sch9p pathways and impair vesicular trafficking, vacuolar function (proteolysis and acidification) and morphology, ultimately contributing to reduced autophagic flux in the mutant strain. Mechanisms other than autophagy are involved in cellular protein turnover. It is well known that autophagy is coordinated with proteasomal degradation and the unfolded protein response (UPR) signalling to govern cell fate, either by restoring protein homeostasis or inducing cell death when the folding capacity is overwhelmed (Korolchuk et al., 2010, Salminen & Kaarniranta, 2010). Isc1pdeficient cells are sensitive to tunicamycin (Kim et al., 2012), a hallmark of defects in ER stress response. We evaluated the activation of an UPRE::LacZ reporter in cells exposed to tunicamycin and DTT, two well-established ER stress inducers. The isc1Δ cells showed an UPRE-regulated induction comparable to the observed in parental cells (Figure 3.S6.A). It was previously reported that the ceramide-activated protein phosphatase Sit4 negatively regulates the UPR in response to ceramide (Mousley et al., 2008). Since very long chain ceramide species levels increased during aging in isc1Δ Chapter III ! ! 136! cells, which were previoulsy implicated in the activation of Sit4p, we hypothesized that UPR could be innapropriately downregulated in isc1Δ cells by activating Sit4p. We observed that UPR signalling was decreased both in exponential phase and during aging by a Sit4p-dependent mechanism (Figure 3.S6.B). However, the overexpression of HAC1 (encoding for the transcription factor that regulates the UPR) had minor effects on isc1Δ CLS and failed to abolish the hydrogen peroxide sensitivity of these mutants (Figure 3.S6.C-D). It is unlikely that lower UPR levels during aging of isc1Δ cells are associated with lower ER stress since the overexpression of HAC1 partially extended the CLS of isc1Δ cells. Instead, the results suggest that failure to induce UPR during aging results from inapropriate activation of Sit4p in response to deranged ceramide levels (Barbosa et al., 2011). In addition, the ER-associated degradation mechanism (ERAD) was not compromised in isc1Δ cells (Figure 3.S6.E). These results suggest that changes in the UPR do not significantly contribute to isc1Δ phenotypes. 3.2.3. Mitophagy is hyperactivated in Isc1p-deficient cells Since isc1Δ cells present mitochondrial dysfunction and fragmentation (Teixeira et al., 2014), we hypothesized that mitophagy is triggered to counteract mitochondria loss-of-function in the mutant. The induction of mitophagy was quantitatively measured by employing a modified alkaline phosphatase (mtPho8) assay (Sampaio-Marques et al., 2012). First, we have monitored mitophagy in cells incubated in lactate medium (post-log phase), as described (Kanki et al., 2009b). The growth defect of isc1∆ cells in this medium is less pronounced than that observed in acetate or glycerol medium and cells are still able to grow within this period of time (48 hours), as reported (Vaena de Avalos et al., 2004) (and our observations). In parental cells, alkaline phosphatase activity increased 2-fold after shifting cells from glucose to lactate medium. This variation was higher (3-fold) in isc1∆ cells (Figure 3.4.A). This indicates that mitophagic activity (the uptake of mitochondria through mitophagy) is enhanced in isc1∆ cells. We also observed mitochondrial fragmentation and reduced oxygen consumption in isc1∆ cells under the same conditions (Figure 3.S7). Chapter III ! ! 137! Figure 3.4. The hyperactivation of mitophagy was correlated with decreased cell viability in isc1Δ cells. A. Mitophagic activity was measured by performing the alkaline phosphatase assay. BY4741 and isc1Δ cells expressing inactive Pho8 proenzyme targeted to the mitochondrial matrix (mtPho8) were grown in SC-glucose (OD600=0.1, T0) and either maintained in the same medium or shifted to SC-lactate medium for 48 h. Alkaline phosphatase activity was measured in glucose (before the shift) and in glucose or lactate- Chapter III ! ! 144! 3.4.A. Mitophagy activity was measured by the alkaline phosphatase assay. Values are mean ± SD of at least three independent experiments. ****p<0.0001; **p<0.01; *p<0.05. C. Cells were grown in SC-medium to the exponential phase and exposed to 1.5 mM H2O2 for 60 min. Cell viability was expressed as the percentage of the colony-forming units (treated cells versus untreated cells). Values are mean ± SD of at least three independent experiments. ****p<0.0001; ***p<0.001; **p<0.01. D. Cells were grown in SCglucose and kept in the medium at 26°C. The viability was determined as the percentage of the colony-forming units at time T in relation to T0. Values are mean ± SD of at least three independent experiments. E. Cells were grown in SC-glucose to the exponential phase and then diluted to OD600=0.1. Fivefold dilution series were spotted on YPD (glucose) or YPG (glycerol) medium and grown at 26°C for 4 days. F. Oxygen consumption rate was measured in cells grown to PDS phase. Values are mean ± SD of at least three independent experiments. ****p<0.0001; **p<0.01. G. Cytochrome c oxidase (COX) activity was determined in cells grown to the PDS phase. Values are mean ± SD of at least three independent experiments. ****p<0.0001; ***p<0.001. H. Cells were grown in SC-glucose to the PDS phase, stained with the potential-sensitive dye 3,3dihexyloxacarbocyanine iodide [DiOC6(3)] for 30 min and analyzed by flow cytometry. Treatment of the parental strain (BY4741) with FCCP (carbonyl cyanide 4- (trifluoromethoxy)phenylhydrazone, 10 mM) was used as a positive control (depolarizing event). Values are mean ± SD of at least three independent experiments. ***p<0.001; **p<0.01; *p<0.05. 3.2.7. Isc1p interacts with Dnm1p Our data showing that Isc1p regulates mitochondrial dynamics and function in a Dnm1p-dependent manner led us to investigate whether these proteins interact. For this purpose, we overexpressed FLAG-tagged Isc1p (under GAL1 promoter) in yeast cells expressing endogenously C-terminally GFP-tagged Dnm1p and cells were grown under respiratory conditions (galactose, for 16-18h hours), where Isc1p is translocated from the ER to the mitochondria (Vaena de Avalos et al., 2004). In fact, FLAG-Isc1p is localized at the ER after incubation for 4-6 hours on galactose, whereas longer incubation (24 h) leads to the mitochondrial localization of Isc1p (Vaena de Avalos et al., 2004). Immunoprecipitation with a FLAG antibody co-precipitated Dnm1p-GFP in cells Chapter III ! ! 145! expressing FLAG-Isc1p but not in cells expressing untagged Isc1p (Figure 3.7), indicating that Isc1p interacts with Dnm1p in vitro. Figure 3.7. Isc1p interacts with Dnm1p in vitro. Lysates from cells expressing Dnm1p-GFP and FLAG-tagged Isc1p proteins were used for immunoprecipitation with either anti-GFP or anti-FLAG protein G beads. Cells were grown in SC-Galactose to induce FLAG-Isc1p and to monitor the interaction between these proteins under respiratory conditions, where Isc1p is known to translocate from the ER to mitochondria. Immunoprecipitated fractions were analyzed by SDS-PAGE and Western blotting. Lysate from untagged BY4741 (lane 2) and G protein beads (lane 3) were used as controls to evaluate unspecific binding of antibodies and beads to lysate proteins. 3.3. Discussion The role of sphingolipids in aging and age-associated pathologies, such as neurodegenerative diseases and cancer, has been extensively reported. In fact, it is well established that alterations on sphingolipid rheostat have significant impact on cell growth and survival and general stress response with impact on the onset of these diseases. However, the molecular mechanisms by which signalling pathways are activated in response to alterations in sphingolipid biosynthesis or turnover are not yet fully understood. This clarification is pivotal to understand sphingolipid functions in more detail and define new strategies to improve human health and extend lifespan. Chapter III ! ! 146! In this work, we report that Isc1p-driven ceramide signalling regulates macroautophagy and the Cvt pathway, mitophagy and mitochondrial dynamics. We show that isc1Δ cells present defects in autophagy flux, exacerbated mitophagy and increased mitochondrial fission, leading to mitochondrial network fragmentation. The impairment in autophagy was not associated with changes in vesicle nucleation. Instead, it resulted from primary defects observed in the mutant strain at later stages of autophagy, namely defective vesicular trafficking and Pep4p-mediated vacuolar proteolysis. The impairment of vesicular trafficking results from a general vacuolar protein sorting defect, since isc1Δ cells displayed aberrant secretion of CPY and disturbed processing of ApeI through the Cvt pathway. The lower Pep4p activity could be due to the lower vacuolar acidification and V-ATPase activity of isc1Δ cells, since Pep4p is not mislocalized in this mutant (our unpublished data). Notably, PEP4 overexpression ameliorated impaired autophagic flux in isc1Δ cells, indicating that Pep4p-mediated vacuolar proteolysis is a limiting factor in the mutant. Consistent with a role in autophagy, Isc1p has been reported to interact with Aut5p, a protein involved in the lysis of intravacuolar vesicles (Epple et al., 2003). We have observed that the induction of GFP-Atg8p (in rapamycin treated cells) and basal Atg13p phosphorylation are not significantly altered in isc1Δ cells (Figure 3.3.A and 3.S12). Nevertheless, we cannot discard the possibility that autophagic induction may be altered in isc1Δ cells due to the hyperactivation of the TORC1 signalling pathway. It is possible that hyperactive TORC1 may target the Atg1p kinase activity, which is required for efficient association of Atg1p with Atg17p and Atg13p (Atg1p kinase complex) during autophagy. Moreover, Sch9p activation, which is increased in isc1Δ cells (Teixeira et al., 2014), was shown to contribute to the inhibition of autophagy independently of phosphorylation of Atg13p (Yorimitsu et al., 2007), which possibly accounts for reduced autophagic activity in isc1Δ cells. Yamagata et al. have recently reported that MIPC or M(IP)2C complex sphingolipids are not required for autophagy but reduced IPC causes autophagic defects, as the inhibtion of IPC synthesis by aureobasidin A (AbA) causes a reduction in autophagic activity (Yamagata et al., 2011). Since Isc1p-deficient cells accumulate these sphingolipids, we may discard the possibility that complex sphingolipids may play a key role for the defects in autophagy observed in isc1Δ cells. Chapter III ! ! 147! Our results show that the TORC1-Sch9p and Sit4p-mediated signalling pathways impair autophagic flux in isc1Δ cells since the deletion of TOR1, SCH9 and SIT4 suppressed the defects on vesicular trafficking, vacuolar proteolytic capacity, vacuolar acidification and the vacuolar fragmentation phenotype, ultimately contributing to restore proper autophagic flux and the integrity of the Cvt pathway in isc1Δ cells. We therefore provide novel evidence that ceramide may constitute a quality control reporter in the process via the ceramide-activated Sit4p protein phosphatase and the activation of the TORC1-Sch9p pathway in response to Isc1p-driven ceramide signalling. Consistently, alterations in ceramide turnover promoted by ectopic overexpression of ceramidase YPC1 in isc1Δ cells improved vacuolar proteolysis. Recently, Sit4p was shown to interact with autophagy effectors, namely Atg18p (Ho et al., 2002). In this work, we also implicate ceramide as a signal that contributes to UPR silencing observed in isc1Δ cells, due to activation of Sit4p. This is supported by the restoration of the UPR in the double mutant isc1Δsit4Δ strain during aging. Together, these results demonstrate that Sit4p links ceramide homeostasis to ER stress response and autophagy in isc1Δ cells by acting as a ceramide sensor whose result is the innapropriate downregulation of these important biological processes, in part due to derranged ceramide metabolism, which then alter cell signaling through Sit4p and the TORC1-Sch9p pathway. More recently, Sit4p has been shown to be present in different protein complexes. Indeed, Sit4p activity is regulated by TORC1 in a Tip41p/Tap42p manner by phosphorylation (Jacinto et al., 2001) whereas it can be activated by ceramide in a Isc1p-dependent manner (Barbosa et al., 2011). Luke et al. (Luke et al., 1996) and Woodacre et al. (Woodacre et al., 2013) have demonstrated that Sit4p also interacts with SAP proteins (SIT4-associated proteins), indicating some degree of response specificity. We propose that ceramide signalling acts as a quality control reporter ideally suited to sense the functional status of autophagy and the ER stress response via Sit4p-mediated mechanisms. Sit4p likely integrates signalling derived from both TORC1 and Isc1p-mediated ceramide homeostasis to regulate these processes and the response specificity may be acquired by interacting with different protein subunits, depending on the stimulus. Importantly, Swinnen et al. have recently reported that ceramide levels are regulated by the TORC1-Sch9p signalling pathway at a transcriptional level Chapter III ! ! 148! (Swinnen et al., 2014). In particular, sch9Δ cells exhibit increased expression of YPC1. Therefore, the suppression of isc1Δ phenotypes imparted by SCH9 deletion can be associated with decreased ceramide levels due to upregulation of YPC1, similarly to the supression of some autophagic defects by ectopic expression of YPC1 reported in this study. We thus propose that Sit4p acts as a ceramide "sensor" and Sch9p may control the phosphatase activity by adjusting ceramide levels (through Ypc1p). The molecular mechanisms underlying the regulation of mitophagy in response to alterations in sphingolipid metabolism remain largely unknown. We provide evidence that mitophagic activity is increased and associated with mitochondrial fragmentation and impaired mitochondrial function in isc1∆ cells when respiration is induced (lactate-growing conditions). This is consistent with an adaptive response to counteract mitochondrial dysfunction. The induction of mitophagy is increased during aging of isc1∆ cells, a phenotype that was associated with shortened CLS of the mutant strain. The deletion of ATG32 aggravated the growth defect of isc1∆ cells, particularly upon the transition to PDS phase (respiratory metabolism), and further shortened its CLS, indicating that Atg32p is required for isc1∆ cell growth and that mitophagy contributes to survival of isc1∆ cells during stationary phase. Sentelle et al. have recently reported an interaction between ceramide and LC3B-II upon DNM1L/DRP1-dependent mitochondrial fission, which ultimately contributes to impaired mitochondrial function by inducing mitophagic cell death (Sentelle et al., 2012). Isc1p-deficient cells also exhibit mitochondrial fragmentation associated with Dnm1p-dependent mechanisms and with the activation of Sit4p and the TORC1-Sch9p pathways in response to derranged ceramide load, leading to mitochondrial dysfunction. Since mitophagy is hyperactivated in isc1∆ cells, could mitophagy be acting as an adaptive clearance mechanism in response to the segregation of dysfunctional mitochondria promoted by mitochondrial fragmentation? We observed that lactate-growing isc1∆ cells present mitochondrial fragmentation and mitochondrial dysfunction associated with increased induction of mitophagy. Importantly, the deletion of TOR1, SCH9, SIT4 and HOG1 lowers the induction of mitophagy when respiration is induced in isc1Δ cells (lactate) and this is associated with the restoration of mitochondrial network integrity and improvement of the organelle function (Barbosa et al., 2011, Barbosa et al., 2012, Teixeira et al., Chapter III ! ! 149! 2014). Altogether, the results point out to a molecular mechanism in which the impairment of mitochondrial dynamics leads to the segregation of dysfunctional and damaged mitochondria, in a Dnm1p-dependent manner, to be degraded by mitophagy in isc1Δ cells upon respiratory metabolism (post-log phase, lactate). This hypothesis is supported by the fact that the deletion of DNM1 reverted the mitochondrial fragmentation of isc1∆ cells, which was associated with lower activation of mitophagy in lactate-growing cells and overall improvement of mitochondrial function. Therefore, we propose that Dnm1p-mediated alterations in mitochondrial dynamics promote mitophagy, thus tipping the balance in favour of fission and thereby enforcing segregation of damaged/dysfunctional mitochondria to be removed by mitophagy in isc1Δ cells under respiratory conditions. This supports a functional link between mitophagy and mitochondrial dynamics and gives new insights on the role of ceramide in the crosstalk between these important mitochondrial quality control mechanisms. In fact, our data supports that Isc1p-driven ceramide signaling also plays a regulatory role in the regulation of mitophagy and mitochondrial dynamics in yeast by modulating the activation of the TORC1-Sch9p pathway, Sit4p and the HOG pathway, which responds to ceramide in a Sch9p-dependent manner, under respiratory conditions (post-log phase). We propose that, when Isc1p is post-translationally activated during the transition to a respiratory metabolism and relocalizes to mitochondria, the protein may adjust ceramide pools to regulate the activation of these signalling pathways and therefore contribute to proper mitochondrial function and network organization in yeast. During aging, the induction of mitophagy further increased in isc1Δtor1Δ and isc1Δhog1Δ cells, comparing with isc1Δ cells. In contrast, it was decreased in isc1Δsit4Δ and isc1Δsch9Δ cells at late stages of the aging process (in water), rendering the strongest phenotypes, which is consistent with the fact that Sch9p and Sit4p integrate ceramide signalling with TORC1 readouts. Nevertheless, these results do not allow to clearly assess the role of mitophagy in aging of isc1∆ cells. We also demonstrate, for the first time, that a protein involved in sphingolipid metabolism, namely Isc1p, interacts with the mitochondrial fission machinery protein Dnm1p under respiratory growth conditions. We speculate that Isc1p-driven ceramide signalling may regulate Dnm1p expression and activity. This is supported by the fact that ISC1 disruption leads to the fragmentation of the Chapter III ! ! 150! network by Dnm1p-dependent mechanisms, as demonstrated in this study. Kitagaki et al. have reported that a decrease in the levels of α-hydroxylated phytoceramide was associated with mitochondrial dysfunction in isc1∆ cells (Kitagaki et al., 2007). Therefore, we hypothesize that alterations in ceramide content (dependent of Isc1p activity) leading to dysfunction of mitochondria may constitute a signal that promotes the recruitment of Dnm1p, leading to the segregation and degradation of damaged mitochondria by mitophagy in yeast. In agreement with this hypothesis, Parra et al. have reported that C2-ceramide induced mitochondrial fragmentation associated with increased mitochondrial DRP1 and FIS1 content, DRP1 colocalization with FIS1, and early activation of apoptosis (Parra et al., 2008). In mammalian cells, ceramide triggers autophagy by inactivating the mTOR-signalling pathway downstream of AKT/PKB (Ravikumar et al., 2004). Furthermore, ceramide also acts by activating the c-Jun N-terminal kinase 1 (JKAMP/JNK1), an orthologue of yeast Hog1p (Galcheva-Gargova et al., 1994), which phosphorylates BCL-2 leading to its dissociation from BECN1/BECLIN1 (Pattingre et al., 2009). C2-Cer has also been shown to relieve IL-13-mediated inhibition of autophagy, whose mechanism appears to be mediated by the class I PI3K/AKT pathway (Petiot et al., 2000). Additionally, recent work has shown that amino acid deprivation caused an increase in acid sphingomyelinase activity leading to enhanced ceramide content, which in turn promoted the PP1/PP2Adependent inactivation of mTOR to induce autophagy (Taniguchi et al., 2012). Our data supports that Isc1p-driven ceramide signalling is also important for proper induction of autophagy; however, autophagy is inhibited upon increased ceramide content observed in isc1Δ cells. Importantly, ceramide acts by downregulating the mTOR signalling in mammals, whereas the TORC1 pathway is hyperactive in isc1Δ cells, contributing to impaired autophagy induction and flux, in parallel to ceramide-mediated activation of Sit4p and Sch9p. One explanation for the apparently divergent regulation of autophagy between mammalian and yeast cells may rely on the fact that specific ceramide species are altered in Isc1p-deficient cells, namely very long chain ceramide species (Barbosa et al., 2011), which in turn inappropriately activates signalling pathways responsive to these particular ceramide pools. Yeast and mammals, however, appear to retain similar molecular Chapter III ! ! 151! mechanisms regarding the regulation of mitophagy and mitochondrial dynamics in response to ceramide signalling, as suggested by our study. Overall, our results implicate Isc1p and ceramide signalling in the regulation of autophagy and support a model in which ceramide load promotes Dnm1pdriven mitochondrial fragmentation, therefore contributing to mitochondrial dysfunction and cell death in yeast. Importantly, our study provides mechanistic details and novel evidence that the modulation of ceramide metabolism by targeting these ceramide-responsive signalling pathways, which are conserved in mammals, could represent a novel strategy to improve lifespan and human health. 3.4. Experimental Procedures 3.4.1. Yeast strains and growth conditions S. cerevisiae BY4741 was the parental strain of all haploid derivatives used in this study (Table 3.5.1). Yeast cells were grown aerobically at 26°C in a gyratory shaker (at 140 r.p.m.), with a ratio of flask volume/medium volume of 5:1. The growth media used were YPD [1% (w/v) yeast extract, 2% (w/v) bactopeptone and 2% (w/v) glucose], YPL [1% (w/v) yeast extract, 2% (w/v) bactopeptone and 2% (w/v) lactate, pH 5.5], YPGlycerol [1% (w/v) yeast extract, 2% (w/v) bactopeptone, 4% (v/v) glycerol], synthetic complete (SC) drop-out medium containing 2% (w/v) glucose, 0.67% yeast nitrogen base without amino acids (Difco Laboratories) and supplemented with appropriate amino acids (80 mg histidine L-1 (Sigma-Aldrich), 400 mg leucine L-1 (Sigma-Aldrich), 80 mg tryptophan L-1 (Sigma-Aldrich) and 80 mg uracil L-1, (Sigma-Aldrich), SC drop-out medium containing 3% (v/v) glycerol and SC drop-out medium containing 2% (v/v) lactate, pH 5.5. Gene disruption was done by conventional methods and the proper integration of the cassettes was confirmed by PCR. The double mutant isc1Δatg32Δ was generated by sporulation of the diploids resultant from mating atg32Δ::KanMX4 carrying pISC1 and isc1Δ::LEU2 strains, followed by tetrad dissection on SC-URA medium to select for spores carrying pISC1, by standard techniques. The identities of the resulting strains were confirmed by growth on respective selective media. Chapter III ! ! 152! 3.4.2. Stress resistance and chronological lifespan For H2O2 resistance assay, cells were grown to the exponential phase (OD600 = 0.6) and exposed to 1.5 mM H2O2 (Merck) for 60 min. The CLS assay was performed as described (Teixeira et al., 2014). 3.4.3. Enzymatic activities and oxygen consumption For the alkaline phosphatase assay, cells were harvested and ressuspended in 200 µL assay buffer [250 mM Tris (AMRESCO), 10 mM MgSO4 (Sigma-Aldrich), 10 mM ZnSO4 (Prolabo), pH 9.0 containing protease inhibitors (Complete Mini EDTA-free protease cocktail inhibitor tablets, Boehringer Mannhein). The cells were lysed by vortexing with glass beads and cell debris was removed by centrifugation at 13,000 r.p.m. for 15 min. 10 µg of extract was added to a final volume of 500 µL (assay buffer containing 1.13 mM nitrophenylphosphate (Sigma Fast, Sigma-Aldrich), and samples were incubated for 15 min at 30°C before terminating the reaction by adding 500 µL of stop buffer [2 M glycine (Nzytech)/NaOH (Merck), pH 11.0]. Production of nitrophenol was monitored spectrophotometrically (UV Mini 1240, Shimadzu), and its concentration was calculated using ε400 = 18,000 M-1.cm-1. The activity of COX (Poyton et al., 1995), β-galactosidase (Teixeira et al., 2014) and Pep4p (Jones et al., 1982) was determined as previously described. For ER stress assays, tunicamycin (SigmaAldrich) and DTT (Nzytech) were used at indicated concentrations. For V-ATPase activity, ATP-dependent proton uptake activity was assayed by spectrophotometry using acridine orange (Sigma-Aldrich) (491-540 nm). Protein (15 µg) of isolated vacuoles were added to the reaction mixture containing 20 mM MOPS (SigmaAldrich)-Tris pH 7.0, 15 mM KCl (VWR), 135 mM NaCl (Merck) and 15 mM acridine orange followed by addition of 0.1 mM MgATP (Sigma-Aldrich). The reaction was stopped by the addition of 10 µL of 1 mM carbonyl cyanide p- (trifluoromethoxy)phenylhydrazone (FCCP, Sigma-Aldrich). Concanamycin A (0.1 µM, Enzo) was used a control to assess that V-ATPase was the main source of the vacuolar proton motive force. Protein content was determined by the method of Lowry, using bovine serum albumin (Nzytech) as a standard. Oxygen consumption rate was measured for 3 x 108 cells in PBS buffer (pH 7.4), using an Chapter III ! ! 153! oxygen electrode (Oxygraph, Hansatech). Data was analyzed using the Oxyg32 V2.25 software. 3.4.4. Western Blot analysis Cells were harvested and then ressuspended in 50 mM potassium phosphate buffer (pH 7.0) containing protease inhibitors and phosphatase inhibitors [50 mM sodium fluoride (Merck), 5 mM sodium pyrophosphate (Fluka Biochemika), 1 mM sodium orthovanadate (Sigma, Aldrich)]. Total protein extracts were lysed and centrifuged at 13,000 r.p.m. for 15 min and protein content was determined by the method of Lowry. Proteins were then analyzed by SDS-PAGE using 10% polyacrylamide (Sigma-Aldrich) gels and blotted onto a nitrocellulose membrane (GE Healthcare) and the Lumigen HRP chemiluminescent substrate (GE Healthcare). Assay for ERAD was done as reported (Ellis et al., 2004). To evaluate Dnm1p-GFP protein levels and Atg13p phosphorylation, cells were harvested and subjected to alkaline lysis. Proteins (50 µg) were resolved by SDS-PAGE using 8.5% polyacrylamide gels. The primary antibodies used were rabbit anti-Atg13p and rabbit anti-ApeI (kindly provided by Dr. Klionsky), mouse anti-GFP (Roche, Basel, 11 814 460 001), mouse anti-CPY (Molecular Probes), mouse anti-Pgk1p (Invitrogen), rabbit anti-HA (Sigma-Aldrich), rabbit anti-actin (Sigma-Aldrich), mouse anti-Por1p (Molecular Probes) and rabbit anti-Fis1p (Santa Cruz Biotechnology, Inc.). The secondary antibodies used were anti-mouse IgG-peroxidase (Molecular Probes) and goat anti-rabbit IgG-peroxidase (SigmaAldrich). 3.4.5. Protease protection assay Rapamycin-treated cells (40 OD600 units) were harvested at 4,000 r.p.m. for 5 min, washed twice with water, and incubated for 20 min at 30ºC in 4 mL of buffer A [100 mM Tris/H2SO4 (Sigma-Aldrich), pH 9.4] containing 20 mM dithiothreitol (DTT, Sigma-Aldrich). The cells were then pelleted, ressuspended in 4 mL of zymolyase buffer [1 M sorbitol (Sigma-Aldrich), 20 mM PIPES (AMRESCO), pH 6.8 containing 1.2 mg zymolyase 20T (Amsbio)] and spheroplasted for 30 min at 30°C. The spheroplasts were harvested at 2,000 r.p.m. for 5 min and hypotonically