Gαq is a novel modulator of autophagy

IS A NOVEL MODULATOR Gαq

OF AUTOPHAGY

Sofía Cabezudo Violero

Madrid, 2018

Este trabajo ha sido realizado en el Departamento de Biología Molecular.

Centro de Biología Molecular ’Severo Ochoa’ (CSIC-UAM)

DOCTORA EN BIOCIENCIAS MOLECULARES (BIOQUÍMICA, BIOLOGÍA MOLECULAR, BIOMEDICINA

Y BIOTECNOLOGÍA)

DIRECTORES DE TESIS:

Dra. CATALINA RIBAS NÚÑEZ

Dr. FEDERICO MAYOR MENÉNDEZ

La realización de esta Tesis ha sido posible gracias a una Ayuda para la Formación del

Profesorado Universitario (FPU) del Ministerio de Educación, Cultura y Deporte y a fondos

de la Fundación Severo Ochoa. La autora también ha recibido una ayuda de movilidad para

estancias breves de EMBO.

Abstract ________________________________________________________________________29

Resumen ________________________________________________________________________31

Abbreviations ________________________________________________________________________33

INTRODUCTION

Autophagy ________________________________________________________________________39 1. Macroautophagy ____________________________________________________________________40 1.1 Phases and molecular machinery involved in autophagy ________________________________40 1.2 Autophagy substrates, receptors and adaptors ________________________________________42 1.3 The autophagy receptor p62/SQSTM1 _______________________________________________42 2. Chaperone-mediated autophagy (CMA) ___________________________________________________44 3. Physiological functions and regulation of autophagy ________________________________________45 3.1 Metabolic regulation and autophagy _________________________________________________46 3.1.1 Nutrient signaling pathways __________________________________________________47 G protein-coupled receptor signaling _______________________________________________________51 4. Functional and structural features of GPCR and its regulation _______________________________51 5. Heterotrimeric G proteins _____________________________________________________________52 6. The Gαq/11 family ___________________________________________________________________52 6.1 Structural aspects in Gαq signaling _________________________________________________53 6.2 Gαq interactome _________________________________________________________________54 6.3 Cellular microenvironments in Gαq signaling ________________________________________57 7. Cellular and physiological functions of Gαq/11 ____________________________________________58 7.1 GPCR and Gαq/11-mediated signaling in nutrient sensing ______________________________58

OBJECTIVES ________________________________________________________________________63

MATERIALS AND METHODS

1. Materials ________________________________________________________________________67 1.1 Buffers and solutions _____________________________________________________________ 67

1.2 Oligonucleotides _________________________________________________________________70 1.3 Primary antibodies _______________________________________________________________71 1.4 Secondary antibodies _____________________________________________________________72 1.5 Plasmids ________________________________________________________________________73 1.6 Treatments ______________________________________________________________________75 1.7 Reagents ________________________________________________________________________76 1.8 Cell lines _______________________________________________________________________76

2. Methods ________________________________________________________________________78 2.1 DNA Manipulations _______________________________________________________________78 2.1.1 DNA sequencing ____________________________________________________________78 2.1.2 DNA mutagenesis ___________________________________________________________78 2.1.3 DNA quantification __________________________________________________________78 2.1.4 DNA agarose gel electrophoresis _______________________________________________79 2.2 Culture and Manipulation of Mammalian Cells ________________________________________79 2.2.1 Maintaining and subculturing cells ____________________________________________79 2.2.2 Freezing and thawing cells ___________________________________________________80 2.3 Transient transfection ____________________________________________________________80 2.3.1 Calcium phosphate method ___________________________________________________80 2.3.2 Lipofectamine method _______________________________________________________80 2.3.3 Metafectene method _________________________________________________________80 2.4 Generation of stable cell lines ______________________________________________________80 2.4.1 Generation of Gαq interaction mutants stable MEF cells lines _______________________80 2.4.2 Generation of the Dendra-KFERQ and mCherry-GFP-LC3 stable cells lines ____________81 2.5 Cell treatments __________________________________________________________________82 2.6 Preparation of cell lysates _________________________________________________________84 2.6.1 Cell lysis ___________________________________________________________________84 2.6.2 Determination of protein concentration _________________________________________84 2.6.3 Immunoprecipitation ________________________________________________________84 2.7 Immunoblotting _________________________________________________________________ 85 2.8 Fluorescence microscopy and imaging _______________________________________________85

2.8.1 Immunofluorescence ______________________________________________________85 2.8.2 Image acquisition _________________________________________________________86 2.8.3 Image quantification ______________________________________________________86 2.9 Analysis of lysosome distribution ___________________________________________________86 2.10 Cell proliferation assays __________________________________________________________86 2.11 Cell viability assays _____________________________________________________________88 2.12 Methods for studying macroautophagy ______________________________________________88 2.12.1 Monitoring autophagic flux _________________________________________________88 2.12.2 Transmission electron microscopy ___________________________________________ 92 2.12.3 Autophagosome and autolysosome isolation from rat liver ________________________94 2.13 Methods for studying chaperone-mediated autophagy __________________________________95 2.13.1 Methods to identify if a protein is a CMA substrate _______________________________95 2.13.2 Monitoring CMA ___________________________________________________________98 2.14 Animal models and study approval _________________________________________________99 2.15 Statistics ______________________________________________________________________99

RESULTS

1. Cells lacking Gαq/11 display higher autophagic flux _____________________________________ 103 1.1 Gαq/11 knockout mouse embryonic fibroblasts display higher levels of basal and

low serum-induced autophagy _____________________________________________________ 103 1.2 Cells lacking Gαq/11 display a higher amount of autophagic vacuoles ____________________103 1.3 Gαq re-expression in Gαq/11 knockout MEFs restores the basal patterns of

autophagy markers ______________________________________________________________ 105 1.4. Despite displaying higher levels of autophagy flux viability is not altered in

Gαq/11 KO MEFs _______________________________________________________________106 2. Cells lacking Gαq/11 display an earlier autophagic response compared to wild type cells

under different nutrient stress conditions ______________________________________________ 107 2.1 Low serum conditions ____________________________________________________________107 2.2 Absence of amino acids __________________________________________________________108 2.3 Absence of glucose ______________________________________________________________ 109 3. Gαq/11 is localized in organelles related to the autophagic process and may play a role as

autophagy adaptor __________________________________________________________________ 110 3.1 Gαq is dynamically localized in lysosomal and autophagic compartments __________________111

3.2. Gαq/11 is not a substrate of autophagy: Potential role as autophagy adaptor _______________112 4. The altered autophagy in Gαq/11 KO cells correlates with the different activation status of

the mTORC1 pathway and is independent of the AMPK cascade _____________________________ 115 4.1 Gαq/11 KO MEFs display lower basal activation levels of the mTORC1 pathway _____________ 115 4.2 Gαq/11 KO MEFs display an earlier and prolonged inactivation of the mTORC1 pathway

in response to different nutrient stress conditions ____________________________________ 116 5. The specific activation of Gαq/11 leads to the modulation of mTORC1/autophagy pathways _______ 118 5.1. Characterization of the DREADD-Gq system __________________________________________119 5.2. The specific activation of a Gq-coupled GPCR reverts the autophagic phenotype

promoted by nutrient stress conditions by a mechanism that correlates with the activation of the mTORC1 pathway __________________________________________________________ 120 5.3 Constitutively active mutants of Gαq/11 maintain the mTORC1 pathway stimulated under low serum conditions ____________________________________________________________122 6. Gαq/11 is required for the reactivation of mTORC1 in response to nutrients and modulates

lysosomal positioning _______________________________________________________________ 123 6.1 Cells lacking Gαq/11 are unable to reactivate the mTORC1 pathway and therefore inactivate autophagy in response to nutrient recovery __________________________________________123 6.2 Gαq/11 modulates lysosomal positioning ____________________________________________124 7. Gαq/11 participates in the activation mechanism of mTORC1 via non-canonical effectors

involving its PB1-binding region ______________________________________________________ 126 7.1. Gαq/11 is part of the mTORC1 complex _____________________________________________ 126 7.2 The modulation of the mTORC1 pathway exerted by Gαq/11 is only partially dependent on its canonical effector PLCβ _______________________________________________________127 7.3 The PB1-binding region of Gαq is implicated in the modulation of mTORC1 ________________ 128 8. p62 displays features of a Gαq/11 effector _______________________________________________130 8.1. Endogenous Gαq/p62 association in different cell types _______________________________130 8.2. PKC is not required for Gαq/p62 association ________________________________________130 8.3. Gαq/p62 association is potentiated in response to the activation of Gαq __________________132 8.4 The Gαq interactor GRK2 is able to compete the Gαq/p62 association_____________________133 9. Identification of the interaction surfaces involved in Gαq/p62 complex formation _______________134 9.1. The classical effector binding region of Gαq is not key for p62 interaction _________________134 9.2 Gαq interacts with p62 through its PB1-binding region ________________________________136 9.3. p62 interacts with Gαq through its PB1 domain ______________________________________ 138 10. Modulation of the Gαq/p62 complex dynamics by nutrient availability _______________________ 139

10.1 Gαq and p62 dissociates in response to nutrient stress ________________________________ 139 10.2 Recovery of nutrients after starvation restores Gαq/p62/mTORC1 association _____________140 11. The modulation of mTORC1/autophagy pathways by Gαq/11 is dependent on its interaction

with p62 _______________________________________________________________________141 11.1. Basal mTORC1 activation and autophagy levels correlate with the ability of Gαq mutants to interact with p62 _____________________________________________________________141 11.2 The Gαq/p62 interaction is involved in the activation of mTORC1 and the consequent

reduction of autophagy in response to nutrient recovery after starvation __________________ 141 12. Gαq/11 as a potential modulator of chaperone-mediated autophagy (CMA) ____________________ 145 12.1 Gαq/11 is localized in lysosomes positive for CMA ___________________________________ 145 12.2 Gαq/11 displays CMA-targeting motifs _____________________________________________ 147 12.3 Gαq/11 is not a substrate of CMA _________________________________________________148 12.4. Gαq/11 positively modulates CMA ________________________________________________ 150

DISCUSION

Gαq/11 modulates macroautophagy _______________________________________________________ 155 Gαq/11 affects autophagic flux via the modulation of the mTORC1 pathway ______________________158 Potential signaling molecules upstream the Gαq/mTORC1/autophagy axis _______________________ 160 Molecular mechanisms underlying the modulation of the mTORC1/autophagy cascade by Gαq/11 ____162 p62 is a novel effector linking Gαq/11 to mTORC1 pathway activation ___________________________162 Gαq/11 modulates lysosomal positioning __________________________________________________ 165 Potential additional roles of the Gαq/p62 signaling module ___________________________________ 168 Gαq/11 positively modulates chaperone-mediated autophagy __________________________________ 169 Potential pathophysiological implications of the Gαq/autophagy axis ___________________________ 172

CONCLUSIONS _______________________________________________________________________177

CONCLUSIONES ______________________________________________________________________181

REFERENCES _______________________________________________________________________183

APPENDIX: SCIENTIFIC PUBLICATIONS ___________________________________________________206

Abstract

The highly conserved autophagy processes play a central role in cellular homeostasis by allowing the lysosomal degradation of cellular components either in basal conditions or in response to fluctuations in the internal or external microenvironment, thus helping to adapt to different stress situations. Autophagy is regulated in a cell-autonomous fashion mainly through AMP-activated protein kinase (AMPK) and mam- malian target or rapamycin complex 1 (mTORC1) modulators. Recent studies suggest that nutrients may also modulate autophagy in a systemic manner through different nutrient-sensing receptors, and emerging evidence suggest that members of the G protein-coupled receptors (GPCRs) membrane receptor family could play such a role. These GPCR nutrient receptors act via different G proteins, including Gαq/11. In this work, we unveil an unanticipated role of Gαq/11 as a key component of the cellular nutrient-sensing machinery. Cells lacking Gαq/11 display higher basal autophagy and an earlier and prolonged induction of this process upon serum removal or in the absence of amino acids or glucose, suggesting that Gαq/11 acts as a general sensor of nutrient availability. Moreover, we describe the presence of Gαq in lysosomal and autophagic compartments and a dynamic redistribution of Gαq between autophagic vacuoles upon nutrient deprivation. However, Gαq is not a substrate of autophagy under nutrient stress conditions, consistent with a role as an autophagy modulator. Importantly, we demonstrate that Gαq affects autophagy processes via the modulation of the key mTORC1 signaling hub, and identify p62/SQSTM1 as the effector linking Gαq/11 to the activation of mTORC1 via the formation of Gαq/p62/mTORC1 multi-molecular complexes. We find that p62 displays most of the features of a bona fide Gαq effector and that these proteins associate through a PB1-like interaction, involving the PB1 domain of p62 and the acidic PB1-binding region of Gαq. The dynamic formation of Gαq/p62 complexes in the presence of nutrients contributes to the activation of mTORC1, thus allowing the subsequent inactivation of macroautophagy in these metabolic conditions. Interestingly, we show that Gαq/11 is also required for the proper lysosomal positioning upon nutrient starvation or recovery, suggesting a role for Gαq/11 in the coordination of mTORC1 activation and lysosomal trafficking processes.

On the other hand, Gαq localizes in lysosomes enriched in chaperone-mediated autophagy (CMA) and appears to positively modulate this process in both basal and prolonged starvation conditions. Our results postulate Gαq as a central molecular switch between different autophagic pathways, thus helping to maintain cellular homeostasis upon nutrient fluctuations.

Resumen

La autofagia es un proceso altamente conservado que posee un papel central en la homeostasis celular, mediando la degradación lisosomal de componentes celulares, tanto en condiciones basales como en respuesta a variaciones en el microambiente intra- o extracelular. Esto permite la adaptación de la célula a distintas situaciones de estrés. Se ha descrito una regulación celular autónoma del proceso de autofagia mediada principalmente por la proteína quinasa activada por AMP (AMPK) y el complejo 1 diana de la rapamicina en mamíferos (mTORC1). Sin embargo, estudios recientes sugieren que diversos nutrientes podrían modular la autofagia de manera sistémica a través de la activación de distintos receptores, entre los que se encuentran miembros de la superfamilia de los receptores acoplados a proteínas G (GPCR). Estos GPCR sensores de nu- trientes actúan a través de distintas proteínas G, entre las que se incluye Gαq/11. En este trabajo desvelamos que Gαq/11 es un componente esencial de la maquinaria celular implicada en la detección de nutrientes. Las células deficientes en Gαq/11 presentan niveles elevados de autofagia en condiciones basales e inducen este proceso de forma temprana y prolongada en respuesta a la ausencia de suero, aminoácidos o glucosa, lo que indica que Gαq/11 actúa como un sensor general de la disponibilidad de nutrientes. Además, describimos la localización de Gαq en vesículas autofágicas y en lisosomas y una dinámica redistribución de Gαq entre las distintas vesículas autofágicas en respuesta a la privación de nutrientes. Estos resultados, junto con el hecho de que Gαq no es un sustrato autofágico, son coherentes con un papel como modulador del proceso.

Por otra parte, demostramos que Gαq controla los procesos de autofagia a través de la activación del nodo de señalización mTORC1, e identificamos a la proteína p62/SQSTM1 como un nuevo efector de Gαq/11 implicado en esta función. En presencia de nutrientes, la asociación entre Gαq y p62 a través de una interacción de tipo PB1 contribuye a la activación de mTORC1 y a la consecuente inactivación de macroautofagia mediante la formación de complejos multimoleculares Gαq/p62/mTORC1. Por otra parte, Gαq/11 es esencial para una correcta localización subcelular de los lisosomas tanto en condiciones de privación como de recuperación de nutrientes, lo que sugiere que Gαq/11 participa en la coordinación de los procesos de activación de mTORC1 y de redistribución de lisosomas. Finalmente, hemos determinado que Gαq/11 está localizado preferente- mente en lisosomas implicados en autofagia mediada por chaperonas (CMA) y que modula positivamente este proceso tanto en condiciones basales como tras una privación de nutrientes prolongada. En conjunto, nuestros resultados sugieren que Gαq puede actuar ejerciendo un balance entre distintos tipos de autofagia, ayudando a mantener la homeostasis celular en respuesta a fluctuaciones de nutrientes.

Abbreviations

3’UTR: 3’ Untranslated Region 7TMR: 7 Transmembrane Receptor aa: Amino acids

AC: Adenyl cyclase AKT: Protein Kinase B AL: Autolysosome

ALR: Autophagic lysosome reformation AMP: Adenosine monophosphate AMPK: AMP-activated protein kinase AngII: Angiotensin II

APG: Autophagosome AR: Adrenergic Receptor AT1R: Angiotensin Receptor 1 ATP: Adenosine Tri-Phosphate AV: Autophagic vacuoles BafA1: Bafilomycin A1 BSA: Bovine Serum Albumin BTK: Bruton Tyrosine Kinase

cAMP: Adenosine 3’5’ Cyclic Monophosphate CaSR: Calcium-sensing receptor

CHO: Chinese hamster ovary CK2: Casein Kinase 2

CMA: Chaperone-mediated autophagy CNO: Clozapine-N-Oxyde

CaSR: Calcium sensing receptor C-ter: Carboxyl terminus Cyt: Cytosol

DAG: Diacyl glycerol

DAPI: 4’,6-Diamidino-2-phenylindole dH2O: Distilled water

DMEM: Dulbecco’s Modified Eagle Medium DMSO: Dimethyl Sulfoxide

DNA: Deoxyribonucleic acid

DREADD: Designer Receptor Exclusively Activated by a Designer Drug DTT: Dithiothreitol

EBSS: Earl’s Balance Salt Solution EE: Protein tag with two glutamic acids EDTA: Ethylenediaminetetraacetic Acid EGFR: Epidermal Growth Factor Receptor EGTA: Ethylene glycol tetraacetic acid e-MI: Endosomal microautophagy ER: Endoplasmic reticulum

ERK: Extracellular signal-regulated kinases FBS: Fetal Bovine Serum

GAP: GTPase activating protein GAPDH: Glyceraldehyde 3-phosphate

dehydrogenase

GDI: GDP dissociation inhibitor GDP: Guanosine Diphosphate

GEF: Guanine nucleotide exchange factor GFP: Green Fluorescent Protein GLUT4: Glucose transporter 4 GPCR: G protein-coupled Receptor GRK: G protein-coupled Receptor Kinase GTP: Guanosine Triphosphate

HDAC: Histone Deacetylase

HEK293 cells: Human Embryonic Kidney 293 cells

HEPES: 4-(2-hydroxyethyl)-1-piperazineethane- sulfonic acid

HIF1α: Hypoxia-inducing factor 1α HRP: Horseradish peroxidase Hsc70: Heat shock 70 kDa protein

HUVEC: Human umbilical vein endothelial cells IF: Immunofluorescence

IGF-I: Insuline-like Growth Factor 1 IGFR: Insulin-like Growth Factor Receptor IKK: Inhibitor of Kappa B Kinase IP: Immunoprecipitation

IP3: Inositol 1,4,5 triphosphate (Ins(1,4,5)P3);

Inositol 3-phosphate

IP3R: Inositol 1,4,5 triphosphate receptor JNK: Jun N-terminal Kinase

KI: Knock-in KO: Knock out L: Leupeptin

LAMP1: Lysosomal-associated membrane protein 1 LAMP2A: Lysosomal-associated membrane

protein 2 type A LB: Luria-Bertani broth

LC3: Microtubule-associated protein 1A/1B-light chain 3

Leu: Leupeptin

LIR: LC3-interacting region Lys: Lysosome

M/L: Mitochondria-lysosomal fraction MA: Macroautophagy

MAPK: Mitogen-activated Protein Kinase

MEF: Mouse Embryonic Fibroblasts MEF2: Myocyte enhancer factor

MEK: MAPK and extracellular signal-regulated kinase (ERK) kinase

MEKK: MAP kinase kinase and extracellular signal-regulated kinase (ERK) kinase MEM: Modified Eagle Medium

mTORC1: Mammalian target of Rapamycin Complex 1

mTORC2: Mammalian target of Rapamycin Complex 2

Mtz: Metrizamide

MVBs: Multivesicular bodies M3: Muscarinic 3 receptor N: Ammonium Chloride NBR1: Neighbor of Brca1 gene NFAT: Nuclear Factor of Activated T cell NF-κB: Nuclear Factor κB

NOX: Nicotinamide adenine dinucleotide phos- phate reduced (NADPH) oxidase

NP40: Nonidet P-40 N-term: Amino terminus OPTN: Optineurin p: p value

Par: partitioning protein

PAGE: Polyacrilamide gel electrophoresis PBS: Phosphate Buffered Saline

PB1: Phox/Bem1

PCNA: Proliferating Cell Nuclear Antigen PCR: Polymerase Chain Reaction

PDGFR: Platelet-derived Growth Factor Receptor PH domain: Pleckstrin Homology domain PI3K: Phosphoinositide 3-kinase

PIP2: Phosphatidylinositol (4,5)-biphosphate (PI(4,5)P2)

PIP3: Phosphatidylinositol (3,4,5)-triphosphate PKA: Protein Kinase A

PKCζ: Protein Kinase C zeta PLC: Phospholipase C

PMSF: phenylmethylsulfonyl fluoride PMT: Pasteurella multocida toxin

PTEN: Phosphatidylinositol (3,4,5)-trisphosphate 3-phosphatase

PTX: Pertussis toxin

PVDF: Polyvinylidene fluoride RGS: Regulator of G protein Signaling RH domain: Regulator of G protein signaling

Homology domain RNA: Ribonucleic acid ROS: Reactive Oxygen Species ROCK: Rho-kinase

rpm: Revolutions per minute RTK: Receptor Tyrosine Kinase SEM: Standard error of the mean SD: Standard deviation

SDS: Sodium dodecyl sulfate Ser: Serine

SQMST1: Sequestosome 1 shRNA: short hairpin RNA Sirt1: Sirtuin 1

SOC: Super Optimal broth with catabiolite repression

STAT: Signal Transducer and Activator of Transcription

STI: Soybean trypsin inhibitor t: Time

T1R1/3: Taste Receptor type 1 member 1 or 3 TAE: Tris base, acetic acid and EDTA TE: Tris-EDTA

TEM: Transmission electron microscopy TEMED: Tetramethylethylenediamine TNF: Tumor necrosis factor

TFs: Transcription Factors Thr: Threonine

TNF-α: Tumor Necrosis Factor α TSC1/2: Tuberous sclerosis complex 1/2 Tyr: Tyrosine

Ub: Ubiquitin

ULK1: Unc-51 like autophagy activating kinase 1 UV: Ultraviolet

WB: Western Blot WT: Wild type

Int ro

Autophagy constitutes a major protective mechanism that allows cells to survive in response to multiple stressors and helps defend organisms against degenerative, inflammatory, infectious, and neoplastic diseases.

As a core homeostatic process, autophagy contributes to proper cell function through the degradation and recycling of cytoplasmic constituents in response to prolonged nutrient deprivation (Kaur & Debnath 2015).

Although autophagy has been classically perceived as a cell-autonomous process, emerging evidence sug- gests that nutrients can also modulate this process in a “systemic” manner via different types of cell surface receptors. In particular, several members of the largest family of mammalian receptors, G protein-coupled receptors (GPCRs), have been proposed as important nutrient sensors, hence emerging as a new avenue for the regulation of autophagy (Husted et al. 2017). Investigating the links between GPCR signaling cascades and autophagic processes may identify new targets for the treatment of autophagy-related disorders.

Autophagy

Autophagy is a prominent mechanism through which eukaryotic cells preserve homeostasis in basal con- ditions and in response to perturbations of the intracellular or extracellular microenvironment. In 1963, the biochemist Christian De Duve first coined the term autophagy (“self-eating” from the Ancient Greek) to describe the presence of single- or double-membraned intracellular vesicles containing part of the cyto- plasm and organelles in various states of degradation (Duve & Russel L Deter 1967). Afterwards, molecular regulators known as Atg genes were discovered in yeast, which allowed the initial understanding of this process (Cheong et al. 2007).

Two main features characterize autophagic responses: the involvement of cytoplasmic material and the culmination with lysosomal degradation. Thus, although autophagy substrates can be endogenous, such as damaged mitochondria or nuclear fragments, or exogenous, such as viruses or bacteria, autophagy operates on entities that are freely accessible to cytoplasmic components. This feature is important to distinguish between autophagy and other processes involving vesicular trafficking from the plasma membrane, such as phagocytosis (which also ends up in lysosomal degradation), receptor-mediated endocytosis or pinocytosis (Foot et al. 2017). In addition, the strict dependency on lysosomal activity is important to discriminate autophagy from other proteolytic pathways, such as the ubiquitin-proteasome system, that also participates in the homeostatic balance through the degradation of proteins (Bhattacharyya et al. 2014).

Depending on how the products to be degraded are delivered to the lysosomes, the autophagic pathways are divided into macroautophagy (MA), chaperone-mediated autophagy (CMA) and microautophagy (Fig. I1).

Macroautophagy is the best characterized variant so far. During this process, cellular components to be de- graded are sequestered in a double-membrane vesicle called autophagosome that finally fuses with lysosomes (Ktistakis & Tooze 2016). CMA has only been described in mammals. It is a very selective process in which a specific subset of soluble proteins (not organelles or other macromolecules) are directly delivered to the lyso- some for degradation previous unfolding (Kaushik & Cuervo 2012)(Kaushik & Ana Maria Cuervo 2018), with no involvement of vesicles or membrane invaginations. Finally, microautophagy is a less understood process by which cytoplasmic entities destined for degradation are directly taken up via direct membrane invagination (Mayer 2008). In endosomal microautophagy (eMI), pro- teins containing a KFERQ-motif are internalized into lysosomes upon sequestration by invaginations of the lysosomal membrane that then pinch off in the shape of small vesicles into the lysosomal lumen for degra- dation (Sahu et al. 2011)(Tekirdag & Cuervo 2017). All autophagy processes finish at lysosomes, which display more than 60 acidic hydrolases able to fully catabolize lipids, carbohydrates, nucleic acids and polypeptides, so the degradation products eventually become available for metabolic reactions or repair processes (Boya 2012) (Settembre et al. 2013).

1. Macroautophagy

1.1 Phases and molecular machinery involved in autophagy

Macroautophagy, hereafter referred to as autophagy, is a highly complex process involving many sequential steps (Fig. I2). In mammalian cells, the process begins with a nucleation step, where membranes of varied origins form phagophores, which then expand and fuse to form a complete double-membraned vesicle called autophagosome (Ravikumar et al. 2010). Autophagosomes are transient organelles that mediate cargo seques- tration and delivery to lysosomes. The isolation membranes may be generated from multiple sources that include endoplasmic reticulum (ER), Golgi complex, outer mitochondrial membrane or plasma membrane (Lamb et al. 2013). The process is initiated by the assembly and activation of the ULK1 multiprotein kinase Figure I1. Types of mammalian autophagy. Scheme

of the basic steps of the three different types of mammalian au- tophagy: macroautophagy (MA), endosomal microautophagy (e- MI) and chaperone-mediated autophagy (CMA) (Patel&Cuervo, 2015).

complex and the subsequent phosphorylation of Atg9 to trigger the formation of the isolation membrane from phospholipids of different membrane sources (Orsi et al. 2012)(Papinski et al. 2014)(Stanley et al. 2014) (Karanasios et al. 2016). This step allows the recruitment of class III phosphatidylinositol-3-kinase (PI3K)/

Beclin complexes, which contain the kinase Vps34 and the proteins Vps15, Atg14 and AMBRA1 (Kihara et al. 2001)(Matsunaga et al. 2009) (see “Induction” and “Phagophore formation” steps in Fig. I2). Vps34 produces phosphatidylinositol 3-phosphate (PI3P), which supports the expansion of autophagosomal mem- branes (Proikas-Cezanne et al. 2015). This process also involves two ubiquitin-like conjugation systems: the Atg12-Atg5-Atg16L1 molecular complex, which together with Atg4 and Atg7 conjugates phosphatidylethanol- amine to microtubule-associated protein 1 light chain 3 beta (MAP1LC3B; best known as LC3B). Lipidated LC3B (often referred as LC3-II) is generated onto forming autophagosomes and allows for substrate uptake upon binding to several receptors of autophagy, including p62/Sequestosome1 (p62/SQSTM1), through its LC3-interacting region (LIR). During autophagy, both LC3-II and the cargo receptors are finally degraded, and are widely used in experimentation as autophagic markers (Klionsky et al. 2016) (see “Autophagosome formation” inset in Fig. I2).

Autophagosomes are formed at random sites in the cytoplasm. They move along microtubules in a dynein-de- pendent fashion towards the microtubule-organizing center (MTOC), where they encounter lysosomes (Mizushima & Klionsky 2007)(Yang & Klionsky 2010). Autophagosomes can fuse with late endosomes (forming amphisomes, single- or double-membraned organelles) or with lysosomes (forming autolysosomes, single-membraned organelles), where the

engulfed substrates are degraded or re- cycled by lysosomal hydrolytic enzymes, finishing with the release of the result- ing molecules to the cytosol (Boya 2012) (“Fusion” and “Degradation” steps in Fig.

I2). Finally, autolysosomes contribute to the regeneration of the lysosomal pool via autophagic lysosome reformation (ALR), maintaining lysosome homeostasis and avoiding over-acidification of autolyso- somes (Yu et al. 2010)(Shen & Mizushima

2014)(Chen & Yu 2017). Figure I2. Molecular machinery involved in the autophagic process. See main text for details. (Autophagy: Mechanisms and functions.

Patricia Boya and Patrice Codogno).

1.2 Autophagy substrates, receptors and adaptors

Besides the autophagy machinery described above, it is important to discriminate among autophagy sub- strates, receptors and adaptors. The term autophagy substrate, or cargo, is used to describe a large set of cytoplasmic entities of endogenous or exogenous origin, ranging from proteins and other macromolecules to organelles or bacteria and viruses, that are targeted to lysosomal degradation by autophagy (Sica et al. 2015)(Galluzzi et al. 2017). Autophagy substrates should be differentiated from autophagy receptors because, while both are subjected to lysosomal degradation, only the latter are part of the autophagy apparatus, binding autophagy substrates and allowing for their recognition by the autophagy machinery (Boya et al. 2013)(Zaffagnini & Martens 2016)(Stolz et al. 2014). Most receptors for macroautophagy share an evolutionarily conserved LC3-interacting region (LIR), which allows bringing substrates in the proximity of LC3 forming autophagosomes (LC3+). This applies to p62, NBR1, OPTN, NDP52, BNIP3, BNIP3L, TRIM5, Atg19 or Atg32, among others (reviewed in (Lamark et al. 2017)). Many macroautophagy receptors also contain ubiquitin-binding domains (UBA), allowing them to recruit ubiquitinated substrates (Khaminets et al. 2016). Importantly, autophagy receptors can provide selectivity to the autophagic process. Starvation- induced autophagy is a non-selective bulk degradation of cytosolic material that rapidly fuel bioenergetics metabolism or repair processes to recover homeostasis. However, other types of stresses such as damaged organelles or aggregated proteins require selective sequestration into autophagosomal membranes through autophagy receptors. In this sense, specific autophagy can clear mitochondria, peroxisomes, lysosomes, ER, lipid droplets, polyribosomes, proteasomes, and the nucleus (Levine & Kroemer 2008)(Anding & Baehrecke 2017). This selective removal is critical for the regulation of cellular homeostasis in organisms from yeast to humans (Mizushima et al. 2008).

Finally, autophagy adaptors could be defined as proteins that can interact with LC3 family members but are not involved in cargo recognition and hence not degraded during macroautophagy, but serve as anchor points for the autophagy machinery (Stolz et al. 2014). Two examples of autophagy adaptors outside the ATG protein family are FYVE and coiled-coil domain containing 1 (FYCO1) and sorting nexin 18 (SNX18) (Olsvik et al. 2015)(Knævelsrud et al. 2013).

1.3 The autophagy receptor p62/SQSTM1

The best-known mammalian autophagy receptor is p62/SQSTM1, here-after referred to as p62, that was initially described as a partner of the atypical protein kinase C family members PKCζ and PKCλ (Sanchez et al. 1998). p62 is a multidomain protein able to interact through its different regions and motifs with a wide variety of signaling proteins to regulate multiple cellular functions and processes, including autophagy (Moscat & Diaz-Meco 2009) (see Box1). The fact that p62 accumulates in autophagy-deficient mice was the

first observed link between autophagy and p62 (Komatsu et al. 2007). Regarding its role as an autophagy receptor, its PB1 (N-terminal Phox and Bem1) domain plays a crucial role in self-oligomerization of p62 pro- tein, that is required for its recruitment to the autophagosome formation site on the endoplasmic reticulum allowing to cluster multiple substrates simultaneously (Itakura & Mizushima 2011). Moreover, p62 contains a LIR motif that allows its own autophagic degradation through LC3+ forming autophagosomes (Pankiv et al.

2007)(Ichimura & Komatsu 2010)(Hirano et al. 2016), and also a UBA domain that permits the recruitment and the autophagosomal degradation of ubiquitinated proteins. Besides ubiquitinated proteins, the UBA domain of p62 recognizes aggregates, damaged mitochondria, ubiquitin-tagged peroxisomes, midbody rings, microbes, ribosomal proteins, and virus capsid proteins, in order to deliver them to autophagosomes (Stolz et al. 2014). In sum, all three domains described above are essential for the role of p62 as an autophagic receptor (Lin et al. 2013).

The activity of autophagy receptors is tightly regulated by changes in expression, cellular localization or post-translational modifications (Stolz et al. 2014). For instance, p62 is phosphorylated by PKA in Ser24 (PB1 domain) thus increasing its interaction with the cyclic AMP-degrading phosphodiesterase PDE4A4, what impairs its homo-oligomerization, and the first steps of autophagy (Christian et al. 2014). p62 phosphoryla- tion in Ser351 by mTORC1 increases its

affinity for Keap1 and favors antioxidant responses via Nrf2 (Ichimura et al. 2013), while phosphorylation in Ser403 by ca- sein kinase 2 (CK2) enhances its affinity to ubiquitinated proteins (Matsumoto et al. 2011). Interestingly, p62 phosphoryla- tion in Thr269/Ser272 by p38δ promotes the activation of mTORC1 in response to amino acids (Linares et al. 2015), where- as AMPK-mediated phosphorylation in Ser293/294 promotes its translocation to the mitochondria and the induction of autophagic cell death (Ha et al. 2017), and phosphorylation in Ser349 by PKCδ fos- ters association to Vps34 and contributes to tumor development (Jiang et al. 2017).

Box1. p62/SQSTM1: More than an autophagy receptor. The sig- naling adaptor p62 modulates several cellular processes by its ability to complex with a variety of other regulatory proteins through its different interaction domains. Through its PB1 domain, besides self-oligomerization, p62 binds the PB1 regions of atypical PKCs, MEK5 or NBR1 and negatively regulates ERK1.

Through its ZZ-type zinc finger domain p62 binds RIP1. Moreover, p62 displays TRAF6 (TB), LC3 (LIR), Keap1 (KIR) and UBA binding domains. Through all these interactions, p62 modulates important cellular processes and is involved in different pathophysiological roles (reviewed in (Nezis & Stenmark 2012)(Lin et al. 2013) (Long et al. 2017).

The correlation between each interaction and the process or diseases modulated

2. Chaperone-mediated autophagy (CMA)

Chaperone-mediated autophagy (CMA) is a very selective form of autophagy in which different cytosolic pro- teins are targeted to the lysosome for degradation (Kaushik & Ana Maria Cuervo 2018). This pool of proteins is characterized by harboring a CMA targeting motif – a penta-peptide amino-acid sequence biochemically similar to KFERQ -. About 40% of soluble cytosolic proteins contain this motif, and posttranslational mod- ifications may generate additional motifs increasing the number of possible substrates (Kaushik & Cuervo 2012). This motif is recognized by the constitutively expressed intracellular chaperone Hsc70 (Cuervo 2010).

Upon binding of Hsc70, substrates are targeted one-by-one to the lysosomal membrane for docking at the cytosolic tail of the monomeric CMA receptor LAMP2A (Cuervo & Dice 1996). Substrate binding initiates multimerization of the LAMP2A monomers at the lysosomal membrane, which constitutes the basis of the CMA translocation multimeric complex (Bandyopadhyay et al. 2008), a process stabilized by a lysosomal resident form of Hsp90. Finally, unfolding of the substrate proteins is followed by their translocation into lysosomes assisted by a lysosome resident Hsc70 (lys-Hsc70) leading to its complete degradation (Tekirdag

& Cuervo 2017)(see Fig.I3).

Regarding CMA regulation, lysosomal membrane LAMP2A levels directly determine CMA activity because substrate binding to its cytosolic region is the limiting factor of the process (Cuervo et al. 1995). In addition, CMA activity is also limited by lys-Hsc70 since it is necessary to achieve a complete translocation of the substrates within the lysosomes (Agarraberes et al. 1997). Besides the existence of a basal CMA activity, CMA is activated in response to prolonged nutrient deprivation, oxidative stress, ER stress, hypoxia or damaged protein accumulation, being nutrient scarcity the utmost activator of this process (Orenstein & Cuervo 2010) (Kaushik & Ana Maria Cuervo 2018). After the transient upregulation of macroautophagy that typically occurs after a short time of starvation, CMA is activated from 10h of starvation, reaching its maximum at 16-24h, depending on the system. In tissues such as liver, CMA activity plateaus at approximately 36h of starvation and activation persists for 3 days (Cuervo et al. 1995), providing a source of recycled amino acids for protein synthesis (Backer & Dice 1986). Recent studies support that CMA may modulate cellular energetics through mechanisms other than mere amino acid recycling. Mice with selective knock out of LAMP2A in hepatocytes display defective handling of glucose and lipids due to loss of their ability to regulate levels of key metabolic enzymes that are usually turned over by CMA when their activity needs to be suppressed (Schneider et al. 2014). Moreover, reactive oxygen species (ROS) and changes in acetylation increase the degradation of substrates through CMA by fostering the accessibility of the KFERQ-motif to Hsc70 and the efficiency of internalization through LAMP2A (Kiffin Roberta et al. 2004)(Bonhoure et al. 2017). The molecular mechanisms involved in the regulation of CMA include the TORC2-Akt-PHLPP1 axis (Arias et al.

2015), a membrane metalloprotease and cathepsine A (Kaushik et al. 2006) as well as the phosphorylation of LAMP2A by the PERK/MKK4/p38 MAPK axis (Li et al. 2017). Interestingly, a decrease in CMA activity is associated with aging, lysosomal storage disorders, neurodegeneration, cancer and metabolic diseases (Orenstein & Cuervo 2010)(Cuervo & Wong 2014)(Tasset & Cuervo 2016)(Kaushik & Ana Maria Cuervo 2018).

3. Physiological functions and regulation of autophagy

Autophagy is an essential process for cell survival, differentiation, development, and homeostasis by helping to adapt to different stress situations (reviewed in (Rybstein et al. 2018)(Galluzzi et al. 2017)). The physiological relevance of autophagy stems from its role in the integrated stress response of the cell (Murrow & Debnath 2013).

Indeed, autophagy is induced by energy stress, ER stress, pathogen-associated molecular patterns (PAMPs) and danger-associated molecular patterns (DAMPs) or hypoxia, being nutrient stress the most potent known physiological inducer of autophagy (Kroemer et al. 2010). The housekeeping functions performed by basal and induced autophagy include the elimination of defective proteins and organelles, the prevention of abnormal protein aggregates, or the removal of intracellular pathogens, thus playing an essential role in cellular quality Figure I3. Phases and modulators of CMA. Steps that mediate degradation of cytosolic proteins via chap- erone-mediated autophagy (CMA): (1) substrate recognition by Hsc70, (2) binding to the cytosolic tail of LAMP2A and unfolding of the substrate protein by the chaperone complex, (3) multimerization of LAMP2A to form the translocation complex, (4) substrate translocation with the assistance of the lysosome resident Hsc70 (lys-Hsc70), (5) substrate degradation, (6) disassembly of LAMP2A from the translocation complex to initiate a new round of substrate binding and translocation, and (7) changes in the turnover of LAMP2A at the lysosomal membrane that contribute to modulate CMA activity (Tekirdag & Cuervo 2017).

control (Mizushima et al. 2008). Consistently, conditional knock out mice for Atg5 or Atg7 in different tissues display accumulation of ubiquitin-positive protein aggregates that are associated with cellular degeneration and cardiac hypertrophy (Komatsu et al. 2006)(Kuma et al. 2004). Moreover, the phenotype of Atg gene-deficient immortalized epithelial cells shows that autophagy can limit DNA damage and chromosomal instability and thus may prevent tumorigenesis (Mathew & White 2007). Conversely, once cancer is established, increased autophagic flux often enables tumor cell survival and growth (Galluzzi et al. 2015). Furthermore, autophagy imbalance is also related with metabolic, inflammatory, cardiovascular diseases or age-related disorders (Levine

& Kroemer 2008)(Madeo et al. 2010)(Kroemer 2015)(Rybstein et al. 2018)(Galluzzi et al. 2017).

Autophagy processes integrate with other cellular stress responses to ensure a proper function of the organism.

Interactions amid autophagy and inflammatory pathways (Mathew et al. 2009)(Fujishima et al. 2011), cell cycle regulation (Mathiassen et al. 2017) and a strong correlation between mitogen signaling and autophagy inhibition (Levine & Kroemer 2008) has been reported. Furthermore, there is a continuous coordination between the different autophagic pathways (Sridhar et al. 2012). Although CMA and macroautophagy are not redundant, cells respond to blockage of CMA by upregulating macroautophagy and vice versa (Kaushik et al. 2008)(Wu et al. 2015). These compensatory mechanisms preserve cellular homeostasis under basal conditions, but the deficiency of the altered pathway becomes manifest as cells are exposed to stress. However, the molecular mechanisms involved in coordinating such macroautophagy-CMA network are still unknown.

On the other hand, autophagy is usually a self-limited process that protects cells from death, being the induction of apoptosis coupled with the inactivation of autophagy (Wirawan et al. 2010)(Kroemer et al. 2010). However, autophagy mediates cell death in some cancer cells (Shimizu et al. 2004) and also in vivo during D. Melanogaster development (Denton et al. 2009), suggesting a complex cross talk between autophagy and apoptosis processes via not well-understood mechanisms (Maiuri et al. 2007)(Mariño et al. 2014).

3.1 Metabolic regulation and autophagy

The ability of cells to respond to nutrient and energy fluctuations is essential for the maintenance of metabolic homeostasis and viability. For instance, high levels of autophagy are observed in mouse tissues immediately following birth (Kuma et al. 2004), and mice lacking either Atg5 or Atg7 die after birth, presumably due to their inability to adapt to the neonatal starvation period (Levine & Kroemer 2008). The complex molecular mechanisms involved in the regulation of mammalian autophagy in response to changes in nutrient avail- ability are being thoroughly investigated (summary in Fig. I4).

3.1.1 Nutrient signaling pathways

As detailed in the following sections, signaling from the nutrient-sensitive kinases mTORC1, mTORC2 and AMPK is key for the modulation of autophagy, although other phosphorylation cascades involving JNK1 (Wei et al. 2008), p38α MAPK (Webber & Tooze 2010), eIF2a (Kouroku et al. 2007) or IKK (Criollo et al. 2010) may be involved in starvation-induced autophagy. In addition, starvation-induced autophagy seems to require the action of sirtuins deacetylases (Morselli et al. 2010)(Lee et al. 2008). Moreover, in response to a prolonged stress, autophagy is regulated at a transcriptional level by the transcription factor EB (TFEB) or by FoxO3 (Füllgrabe et al. 2014). In the presence of nutrients, TFEB is retained in the cytosol by its phosphorylation by ERK2 or mTORC1 and, in response to starvation translocates to the nucleus regulating the genetic expression of Atg and lysosomal biogenesis genes (Sardiello et al. 2009).

• mTORC1

mTORC1 is a highly conserved serine/threonine kinase that is capable of integrating signals from many stimuli, including amino acids, energy levels, oxygen, growth factors, and stress, to coordinate cell growth and maintain metabolic homeostasis. The essential core components of mTORC1 are Raptor (regulatory-as- sociated protein of mTOR) and mLST8 (mammalian lethal with SEC thirteen 8) (Laplante & Sabatini 2012) (Saxton & Sabatini 2017).

• Nutrient-sensing mechanism of mTORC1

In response to growth factors such as insulin or insulin-like growth factor 1 (IGF-1), there is a stimulation of PI3K and Ras pathways. The effector kinases of these pathways, protein kinase B (Akt/PKB) and extra- cellular-signal-regulated kinase 1/2 (ERK1/2), directly phosphorylate and inactivate the tuberous sclerosis complex 1/2 (TSC1/TSC2). Upon inhibition of the TSC complex, lysosomal GTP-loaded Rheb binds the mTORC1 catalytic domain to activate it (Inoki et al. 2003)(Yoshida et al. 2011), (reviewed in (Laplante & Sabatini 2012)).

Additionally, Akt also phosphorylates Raptor promoting mTORC1 activation (Frey et al. 2014). On the other hand, amino acids signal to mTORC1 is mainly transduced through the Ras-related GTP-binding protein (Rag) family of small GTPases (Eunjung kim et al. 2008). An active Rag heterodimer (GTP-loaded Rag-A/B and GDP-loaded Rag-C/D), tethered by the Ragulator complex and vacuolar H+-ATPase (v-ATPase), mediates the translocation of mTORC1 from the cytoplasm to the surface of lysosomes, where mTORC1 encounters and is activated by Rheb (Sancak et al. 2008)(Sancak et al. 2011)(Zoncu et al. 2010).

Adding other layers of complexity, priming amino acids firstly can sensitize mTORC1 for subsequent stimu- lation by activating amino acids (Dyachok et al. 2016). In addition, leucine, glutamine and arginine are key mediators of mTORC1 activation. Leucine and arginine activate mTORC1 through the intracellular sensors

Sestrin2 or Castor1, respectively (Wolfson et al. 2016), while glutamine acts by supplying energy through glutaminolysis (Durán et al. 2012), a process essential for GTP-loading of RagB, and together with leucine to inhibit the activation of AMPK (Gleason et al. 2007). For its part, intra-lysosomal arginine activates mTORC1 through the SLC38A9 transporter (Wang Shuyu et al. 2015)(Rebsamen et al. 2015), and amino acid uptake by dynamin-dependent endocytosis plays also a critical role in mTORC1 activation (Shibutani et al. 2017).

Therefore, lysosomal amino acids (AA) sensors (PAT1, SLC38A9), cytoplasmic AA sensors (LRS, Sestrin2, CASTOR1) and the Golgi AA sensor PAT4 can participate in regulating mTORC1 activation together with the Rag GTPases-mediated general mechanism (reviewed in (Zheng et al. 2016)). Because the lysosomal docking of mTORC1 results essential for its activation, its full activation is only achieved in the presence of both amino acids and growth factors (reviewed in (Laplante & Sabatini 2012))(Saxton & Sabatini 2017).

Interestingly, p62/SQSTM1 is also an essential component of the mTORC1 activation mechanism (Duran et al. 2008)(Duran et al. 2011). p62 interacts with Raptor and orchestrates a PB1 domain-driven multimolecular complex including MEKK3/MEK3-6/p38 and TRAF6 to favor the formation of the active Rag heterodimer and the translocation of mTORC1 to lysosomes for its full activation in response to amino acids (Duran et al. 2011)(Linares et al. 2015b).

• mTORC1 activation promotes anabolic metabolism

Active mTORC1 directly phosphorylates the translational regulators eukaryotic translation initiation factor 4E (eIF4E)-binding protein 1 (4E-BP1) and S6 kinase 1 (S6K1), which, in turn, promote protein synthesis and ribosome biogenesis, that is accompanied by a de novo synthesis of nucleotides (Ben-Sahra et al. 2013).

Moreover, mTORC1 also promotes lipogenesis through the activation of the transcription factors SREBPs in a S6K-dependent manner (reviewed in (Shimobayashi & Hall 2014) (Laplante & Sabatini 2012)) .

Importantly, mTORC1 is a conserved and critical repressor of autophagy. mTORC1 represses the ULK1 kinase complex by directly phosphorylating the Atg13L and ULK1(Ser757) subunits. This ULK1 phosphorylation inhibits its interaction with AMPK (Kim et al. 2011). Moreover, mTORC1 directly phosphorylates and inhibits Atg14-containing Vps34 complexes (Yuan et al. 2013). The mTORC1-mediated repression of both ULK1 and Vps34 complexes inhibits autophagosome formation and autophagy. Finally, mTORC1 also indirectly regulates autophagy by controlling lysosome biogenesis through the direct phosphorylation of the transcription factor EB (TFEB)(Settembre et al. 2012).

Conversely, in response to nutrient deprivation, mTORC1 is inactivated at different levels, thus allowing autophagy. In such conditions, activation of the TSC complex inactivates Rheb (Long et al. 2005), inactivation

of Rags impairs the location of mTORC1 in lysosomes and dephosphorylated TFEB translocates to the nucleus favoring autophagic genes transcription (Settembre et al. 2012). Interestingly, a p62-mTORC1-autophagy feedforward loop also negatively regulates mTORC1 during chronic nutrient deprivation. A prolonged lack of nutrients produces chronic activation of autophagy and long-term reduction of p62 (Duran et al. 2011) (Moscat & Diaz-Meco 2011), thus hampering p62-mediated mTORC1 reactivation upon nutrient recovery.

Paradoxically, under prolonged starvation the role of mTORC1 in autophagy flips from a repressor to a pro- moter. Autophagy provides nutrients to the cell, which in turn reactivates mTORC1 to terminate autophagy.

mTORC1 terminates autophagy by promoting the release of tubular proto-lysosomes from autolysosomes, a process known as autophagic lysosome reformation (ALR) (Yu et al. 2010)(Chen & Yu 2017). The molecular mechanism remains to be fully elucidated but implication of microtubules, clathrin (Rong et al. 2012), the small GTPase Rab7 (Yu et al. 2010), Spinster (Rong et al. 2011) and Vps34 phosphorylation by mTORC1 (Munson et al. 2015) has been suggested. Remarkably, it has been recently described that the generation of alpha ketoglutarate (α-KG) by glutaminolysis under prolonged nutrient starvation in tumor contexts reacti- vates mTORC1 leading to the inhibition of autophagy and the induction of cell death through a mechanism dependent on p62 and caspase 8. This process, known as glutamoptosis, point out p62 as an integrator of signals to detect nutrient imbalance (Villar et al. 2017).

• Relevance of lysosomal localization for mTORC1 controlled processes

As detailed above, the lysosomal membrane is the site where mTORC1 complex regulation by amino ac- ids and growth factors converge (Menon et al. 2014), thus defining the lysosomal surface as a metabolic signaling center that integrates nutritional information from the lysosome’s interior and the cytoplasm to promote mTORC1-mediated signals (Perera & Zoncu 2016)(Rabanal-Ruiz & Korolchuk 2018). Furthermore, lysosomal positioning coordinates anabolic and catabolic responses with changes in nutrient availability by orchestrating early plasma-membrane signaling events, mTORC1 signaling and autophagy (Korolchuk et al. 2011). However, the precise molecular mechanisms of lysosomal trafficking depending of nutrient availability are still unknown.

• mTORC2

The essential core components of mTORC2 are Rictor (rapamycin-insensitive companion of mTOR), Sin1 (SAPK- interacting 1) and mLST8 (Cybulski & Hall 2009)(Saxton & Sabatini 2017). mTORC2 signaling is insensitive to nutrients but does respond to growth factors such as insulin through a poorly defined mechanism that requires PI3K and ribosomes (Zinzalla et al. 2011). Differentially with mTORC1, mTORC2 is not modulated by p62 because of the absence of Raptor in the complex (Duran et al. 2011). Through the phosphorylation

of different members of the AGC kinase subfamily, mTORC2 regulates cytoskeletal organization (through PKCα) (Jacinto et al. 2004), cell survival and metabolism (through SGK1) (García-Martínez & Alessi 2008) and the inhibition of TSC2 (through Akt activation at Ser273) (Sarbassov et al. 2004). Interestingly, it has been described that mTORC2 exerts a negative modulation over CMA activation through its phosphorylation of Akt (Ser473) in the membrane of CMA-active lysosomes (Arias et al. 2015).

• AMPK

Besides nutrient fluctuations, the cell must strictly match the generation and consumption of ATP. The intracellular ratio of ATP:ADP:AMP is an important indicator of the energy status of the cell, and these nu- cleotides are directly sensed by the trimeric serine/threonine kinase AMP-activated kinase (AMPK) (Hardie et al. 2015). Decrease in ATP levels due to glucose withdrawal or mitochondrial dysfunction activates AMPK to generate energy by increasing glucose uptake and glycolysis and stimulating lipid catabolism (Hardie 2007). Several upstream kinases, including liver kinase B1 (LKB1, which is activated by energy depletion), calcium/calmodulin kinase kinase-β (CaMKKβ, which is activated by cytosolic Ca2+), and TGFβ-activated kinase-1 (TAK-1, which is also involved in IKK activation), can activate AMPK by phosphorylating a threonine residue on its catalytic α subunit (Ruderman et al. 2010). Regarding its role in autophagy, activated AMPK phosphorylates TSC2 in a distinct residue from the one phosphorylated by Akt in response to nutrients, leading to its activation, and also phosphorylates Raptor, impairing its essential interaction with mTORC1 (Laplante

& Sabatini 2012). Moreover, AMPK phosphorylates ULK1 in Ser555 promoting the activation of the Class III PI3K initiating complex leading to the activation of autophagy (reviewed in (Laplante & Sabatini 2012) and (Shimobayashi & Hall 2014)). Finally, Sirt1 and AMPK can engage in a co- ordinated positive ampli- fication loop favoring the activation of autophagy (Ruderman et al. 2010).

Figure I4. Nutrient signaling pathways that modulate autophagy. See main text for details.

G protein-coupled receptor signaling

As detailed in previous sections, intracellular nutrient fluctuations and the energy status converge in mTORC and AMPK as central metabolic regulators to control autophagy and to ensure metabolic homeostasis.

Interestingly, it has been recently suggested that nutrients can also be sensed by membrane receptors and transmit the signal into the cell in a certain way (Wauson et al. 2014). G protein-coupled receptors (GPCR), the largest, most versatile and most ubiquitous membrane receptor family (Pierce et al. 2002), mediate the actions of many neurotransmitters, hormones, chemokines, and many other stimuli such as calcium ions, odorants, bitter and sweet taste, and even photons of light (Gurevich & Gurevich 2017). GPCR regulate most physiological processes and are the targets for circa 25% of all approved drugs (Sriram & Insel 2018)(Hauser et al. 2017). Interestingly, recent studies suggest that GPCRs can also act as nutrient sensors (Blad et al. 2012) (E M Wauson et al. 2013)(Milligan et al. 2014)(Ichimura et al. 2014)(Husted et al. 2017). Such “extracellular”

nutrient information would facilitate context-specific decision making to conserve organismal homeostasis and cooperate with intracellular sensors to achieve a multilayered response to nutrient fluctuations.

4. Functional and structural features of GPCR and its regulation

The GPCR superfamily can be grouped into three different families (A, B and C) on the bases of sequence similarity. Family A is the largest group and includes rhodopsin or adrenergic receptors as best-studied members. Family B comprises only 25 members with higher homology to secretin/glucagon receptors.

Family C is also relatively small, and comprises the metabotropic glutamate receptors, the GABAB receptor, calcium-sensing receptor, as well as several taste receptors.

GPCR trigger the activation of intracellular G proteins (Hepler & Gilman 1992)(Stevens et al. 2013). Agonist binding to GPCRs promotes an important rearrangement of intracellular helices 6 and 3 (Standfuss et al.

2011)(Kobilka 2007)(Hilger et al. 2018) leading to the activation of Gα subunits by sequentially promoting GDP dissociation and GTP binding (Fig. I5). This leads to the dissociation of the heterotrimeric G protein (Gαβγ) into a GTP-bound α subunit and a βγ dimer, both of which have an independent capacity to regulate an ever-expanding list of different effectors. GTP hydrolysis in the nucleotide-binding pocket of Gα sub- units, a process modulated by regulator of G-protein signaling proteins (RGS), causes re-association of the heterotrimer and signal termination (Magalhaes et al. 2012).

Additionally, GPCRs are desen- sitized via phosphorylation of active receptors by G protein-cou- pled receptor kinases (GRKs) and subsequent high affinity binding of arrestins to the phosphorylated GPCR, thus precluding G protein binding while acting as endocytic adaptors that facilitate the target- ing of GPCRs for clathrin-medi- ated endocytosis (Wilden 1995).

This process targets receptors for degradation or, alternatively, contributes to re-sensitize and recycle them back to the plasma membrane, in a mechanism dependent on several small GTP-binding proteins of the Ras superfamily (Rab proteins) (Fig.

I5) (Ribas et al. 2007). Moreover, β-arrestins act as multifunctional adaptor proteins that couple GPCRs to numerous pathways independently, and often spatially and temporarily segregated, from G protein-dependent signaling (DeWire et al. 2007)(Peterson & Luttrell 2017)(Jean-Charles et al. 2017).

5. Heterotrimeric G proteins

A relatively small number of heterotrimeric G proteins mediate intracellular signaling cascades triggered by the huge GPCR superfamily. In the human genome 35 genes encode G proteins (16 correspond to α-subunits, five to β and 14 to γ) (Milligan & Kostenis 2006). On the basis of sequence similarity, the Gα subunits have been divided into four major families (Gs, Gi, Gq, G12 (Strathmann & Simon 1991) ), and this classification has served to define both receptor and effector coupling. While most GPCRs are able to activate more than one Gα subtype, often they also show a clear preference for one subtype over another (Gudermann et al.

1997). In addition, Gβγ subunits are also capable of binding to and activate a number of cellular proteins, with particular relevance in Gαi/o-coupled GPCR signaling (Milligan & Kostenis 2006).

6. The Gαq/11 family

The Gαq/11 family comprises four members (Hubbard & Hepler 2006). Gαq and Gα11 are ubiquitously expressed, Gα14 is found in liver, lung and kidney, and Gα15/16 (mouse/human orthologues, respectively) is only expressed in hematopoietic cells. Gαq and Gα11 are highly homologous proteins (90% amino acid Figure I5. GPCR activation and deactivation cycle. See main text for details.

(Sánchez-Fernández et al. 2014)

identity), indistinguishably activating the same cellular effectors (Hubbard & Hepler 2006) (reviewed in (Sánchez-Fernández et al. 2014)). To date, only two differential features have been specifically described for Gαq and not for Gα11: the susceptibility to be activated by Pasteurella multocida toxin (PMT) (Orth et al. 2009) and the ability to associate to the cold-activated channel TRMP8 (Li & Zhang 2013). Therefore, an assumed functional redundancy has led to a generalized use of experimental tools that do not differentiate between these two proteins (e.g. anti-Gαq/11 antibodies). Indeed, only double Gαq/Gα11 knock out mice models show drastic cardiac abnormalities whereas mice with single deletions are seemingly normal. However, there are some differential roles in which only Gαq, but not Gα11, is involved, such as certain platelet functions (only Gαq is expressed in platelets) and craniofacial defects that are also specifically displayed by Gαq^-/- mice (Hubbard & Hepler 2006). Additionally, it has been shown that Gαq, but not Gα11, is involved in glutamate receptor-dependent long-term depression in the hippocampus (Kleppisch et al. 2001). Thus, we have assumed the functional redundancy between Gαq and Gα11 and we have mainly focused on Gαq, without distinguishing from Gα11, during the development of this thesis.

6.1 Structural aspects in Gαq signaling

The Gα subunit structures reveal a conserved protein fold comprising a GTPase domain and α helical domain (Oldham & Hamm 2008) (Fig. I6). The GTPase domain, similar to that of other GTP-binding proteins, par- ticipates in the hydrolysis of GTP to GDP and is involved in binding GPCRs, GDP/GTP, the Gβγ subunit, and effectors. This domain containing three highly conserved flexible loops (termed switches I, II and III) that undergo important conformational changes between the GDP/GTP-bound forms of the protein. The helical domain, forms a lid over the nucleotide-binding site, is involved in increasing the affinity of Gα for guanine nucleotides and in increasing intrinsic GTP hydrolysis activity (Echeverría et al. 2000). The N- and C-terminus are essential for receptor specificity, interaction with the Gβγ subunit and for membrane targeting of the G protein. All Gα subunits are lipid-modified at the N-terminus. Gαq is dually palmitoylated at cysteines C9 and C10 (Linder et al. 1993) what, together with its interaction with the isoprenylated Gβγ dimer, promote the peripheral membrane localization of the G protein (Crouthamel et al. 2008) (Crouthamel et al. 2010).

In all crystallized structures of G protein–effector complexes, a general binding domain with the effectors has been identified (Sprang et al. 2007). Since the structure of these regions is well conserved amongst the different Gα families, it is assumed that effector specificity is mostly determined by the amino acid sequence (Oldham & Hamm 2006). Although effectors are thought to be mutually exclusive and to compete for binding to the effector-binding site (Lutz et al. 2005)(Carman et al. 1999), the specific determinants of these interactions are different (see Fig. I6). Specifically, available crystal structures of Gαq in complex with PLCβ3 (Waldo et al. 2010), p63RhoGEF (Williams et al. 2007) and GRK2 (Tesmer 2005) shed light on

the classical effector–Gαq binding surfaces. Comparison of the structures for PLCβ and p63RhoGEF reveals that both effectors interact with essentially identical residues in the Gαq-binding interface. GRK2 forms an effector-like interaction with Gαq in a very similar fashion to PLCβ3 and p63RhoGEF, and different to the modulatory RGS proteins (Tesmer 2005), which have been shown to bind into high order G protein–effector complexes without competing for the effector binding site, but instead allosterically inhibiting effector asso- ciation (Shankaranarayanan et al. 2008). Recently, our group described a novel acidic effector-binding region in Gαq, different from the classical effector-binding region, responsible for its interaction with the protein kinase C ζ (PKCζ), where glutamic acids 234 and 245 are essential for the association (Sánchez-Fernández et al. 2016). Interestingly, residues 221-245 of Gαq show striking sequence similarity with PB1-type I domains (also present in MEK5, p40phox or type I/II PB1-containing proteins such as p62, reviewed in (Moscat et al. 2006), more details about PB1 domains in Box2). Since PKCζ interacts with Gαq through its PB1-type II domain, we have termed this module as a PB1-binding region, potentially able to engage Gαq in additional signaling and cellular functions.

6.2 Gαq interactome

A variety of cellular proteins have been described to interact with Gαq, resulting in either propagation or deacti- vation of Gαq signaling (see a scheme of the Gαq interactome in Fig. I7). Generally, effectors are proteins whose activity depends on its interaction with GTP-bound Gαq after receptor activation, whereas inhibitory interactors limit Gαq signaling by enhancing the GTPase activity of the G protein or by sequestering it away from effectors.

Figure I6. Gαq structural and binding regions. Linear representation of the Gαq sequence depicting the corresponding three-dimensional domains, secondary structure and effector binding regions. Crucial contacts between Gαq and different effectors are shown. Modified from (Sánchez-Fernández et al. 2014).