CHARACTERIZATION OF DEGRADATIVE PROTEINQUALITY CONTROL MECHANISMS USING MODELSUBSTRATES DERIVED FROM TEMPERATURESENSITIVE ALLELESbySophie ComynM.Sc., The University of Alberta, 2011B.Sc., The University of British Columbia, 2007A THESIS SUBMITTED IN PARTIAL FULFILLMENTOF THE REQUIREMENTS FOR THE DEGREE OFDoctor of PhilosophyinTHE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES(Genome Science and Technology)The University of British Columbia(Vancouver)December 2016c￿ Sophie Comyn, 2016AbstractThe purpose of protein homeostasis (proteostasis) is to maintain proteome integrity,thereby promoting viability at both the cellular and organism levels. Exposure toa range of acute stresses often produces misfolded proteins, which present a chal-lenge to maintaining proteostatic balance. The accumulation of misfolded proteinscan lead to the formation of potentially toxic protein aggregates, which are charac-teristic of a number of neurodegenerative diseases such as Alzheimer’s and Parkin-son’s. Therefore, a number of protein quality control pathways exist to promoteprotein folding by molecular chaperones or target terminally misfolded proteins fordegradation via the ubiquitin proteasome system or autophagy. Within the cytosolthe mechanisms responsible for targeting substrates for proteasomal degradationremain to be fully elucidated.In this thesis, we established and employed thermosensitive model substratesto screen for factors that promote proteasomal degradation of proteins misfoldedas the result of missense mutations in Saccharomyces cerevisiae. Using a genome-wide flow cytometry based screen we identified the prefoldin chaperone subunitGim3 as well as the E3 ubiquitin ligase Ubr1. An absence of Gim3 leads to theaccumulation of model substrates in cytosolic inclusions and their delayed degra-dation. We propose that Gim3 promotes degradation by maintaining substrate sol-ubility.In the course of screening for factors involved in degradative protein qual-ity control, we identified secondary mutations in the general stress response geneWHI2 among a number of E3 ligase deletion strains. We demonstrate that an ab-sence ofWHI2 is responsible for the observed impairment in the proteolytic degra-dation of Guk1-7. We propose a link between mutations inWHI2 to a deficiency iniithe Msn2/4 transcriptional response, thereby altering the cell’s capacity to degrademisfolded cytosolic proteins.Collectively, the data in this thesis generated with the Guk1-7 model substrateunderscores how changes in the elaborate protein quality control network can per-turb proteostasis. Given that proteostasis is altered in a number of diseases rangingfrom cancer to ageing, identifying the factors that mediate protein quality controland understanding the interplay between members of the proteostatic network areimportant not only for understanding the basic biological processes but also forpotential therapeutic applications.iiiPrefaceChapter 1 is partially based on a first author publication. Comyn SA, Chan GT,Mayor T. (2014). False start: cotranslational protein ubiquitination and cytosolicprotein quality control. J Proteomics 100:92-101. doi: 10.1016/j.jprot.2013.08.005.This is a peer reviewed review article. I co-wrote most of the manuscript withThibault Mayor, whereas Gerard Chan helped with some sections. I prepared thefigure.Chapter 2 is based on a first author publication. Comyn SA, Young BP, LoewenCJ, Mayor T. (2016). Prefoldin promotes proteasomal degradation of cytosolicproteins with missense mutations by maintaining substrate solubility. PLoS Genet-ics 12(7):e1006184. doi: 10.1371/journal.pgen.1006184. I performed all of theexperiments and Barry Young prepared the barcoded yeast deletion collection usedfor the screen and all subsequent validation experiments. I and Thibault Mayor co-wrote the manuscript with input from Drs. Barry Young and Christopher Loewen.The plasmid BPM866 (pFA6a-mCherry-KanMX6) and the yeast strain YTM1919(Hsp42-mCherry), which were used in this study, were made by Mang Zhu usinga codon optimized mCherry template prepared by Dr. Patrick Chan.Chapter 3 is based on a first author publication being prepared for submission.Comyn SA, Flibotte S, Spear ED, Michaelis S, Mayor T. Recurrent background mu-tations in WHI2 alter proteostasis and impair degradation of cytosolic misfoldedproteins in Saccharomyces cerevisiae. All the experiments were designed by my-self and Thibault Mayor. I performed most of the experiments and co-wrote themanuscript with Thibault Mayor. Illumina library preparation and whole genomesequencing was conducted by Ana Kuzmin at the NextGen Sequencing facility atthe UBC Biodiversity Research Centre. Dr. Ste´phane Flibotte performed the se-ivquencing analysis which led to the identification of theWHI2mutation. Eric Spearand Susan Michaelis performed experiments that are not presented in this thesisbut form part of this publication.vTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiGlossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiiAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviii1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Protein Misfolding and Protein Homeostasis . . . . . . . . . . . . 11.2 Protein Folding and Cytosolic Molecular Chaperones . . . . . . . 31.2.1 Nascent Protein Folding . . . . . . . . . . . . . . . . . . 31.2.2 Hsp70, Hsp40, and Hsp90 . . . . . . . . . . . . . . . . . 41.2.3 TRiC/CCT Chaperonin and Prefoldin . . . . . . . . . . . 51.3 Molecular Chaperones and Protein Degradation . . . . . . . . . . 81.4 Ubiquitin Proteasome System (UPS) . . . . . . . . . . . . . . . . 81.4.1 ER Associated Degradation (ERAD) . . . . . . . . . . . . 101.4.2 Nuclear Protein Quality Control . . . . . . . . . . . . . . 121.5 Cytosolic E3 Ubiquitin Ligases Involved in Protein Quality Control 131.5.1 CHIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13vi1.5.2 Ubr1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141.5.3 Hul5 and Rsp5 . . . . . . . . . . . . . . . . . . . . . . . 151.5.4 Ltn1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151.6 Autophagy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161.7 Spatial Protein Quality Control: CytoQ, IPOD, and INQ . . . . . 181.8 Stress Responses . . . . . . . . . . . . . . . . . . . . . . . . . . 191.8.1 Heat Shock and General Stress Response . . . . . . . . . 201.9 Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211.10 Model Substrates Used to Study Proteostasis . . . . . . . . . . . . 231.11 Research Objective . . . . . . . . . . . . . . . . . . . . . . . . . 261.11.1 Specific Aims . . . . . . . . . . . . . . . . . . . . . . . . 262 Prefoldin Promotes Proteasomal Degradation of Cytosolic Proteinswith Missense Mutations by Maintaining Substrate Solubility . . . . 272.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272.2 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . 292.2.1 Yeast Strains, Plasmids, and Media . . . . . . . . . . . . 292.2.2 Stability Effect of Guk1-7 Mutations . . . . . . . . . . . 342.2.3 Cellular Thermal Shift Assay (CETSA) . . . . . . . . . . 342.2.4 Solubility Assay . . . . . . . . . . . . . . . . . . . . . . 342.2.5 Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . 352.2.6 Degradation Assay . . . . . . . . . . . . . . . . . . . . . 352.2.7 Flow Cytometry . . . . . . . . . . . . . . . . . . . . . . 362.2.8 GFP Pulldown . . . . . . . . . . . . . . . . . . . . . . . 362.2.9 Proteasome Function . . . . . . . . . . . . . . . . . . . . 362.2.10 Statistical Analysis . . . . . . . . . . . . . . . . . . . . . 372.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372.3.1 Guk1-7 is Thermally Unstable . . . . . . . . . . . . . . . 372.3.2 Fluorescence-Based Assay to Assess Protein Stability . . 382.3.3 Guk1-7 Degradation is Proteasome Dependent . . . . . . 442.3.4 FACS-Based Screen for Protein Homeostasis Factors . . . 452.3.5 Ubr1 Stabilizes Guk1 Missense Mutant . . . . . . . . . . 482.3.6 Gim3 Impairs Guk1-7-GFP Degradation . . . . . . . . . . 51vii2.3.7 Gim3 Facilitates the Clearance of Insoluble Guk1 andMain-tains Guk1-7 Solubility . . . . . . . . . . . . . . . . . . . 542.3.8 Gim3 Has a General Effect Towards Thermally Destabi-lized Proteins . . . . . . . . . . . . . . . . . . . . . . . . 592.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 622.5 Supplemental Data . . . . . . . . . . . . . . . . . . . . . . . . . 673 Recurrent Background Mutations in WHI2 Alter Proteostasis andImpair Degradation of CytosolicMisfolded Proteins in Saccharomycescerevisiae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 723.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 723.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 743.2.1 Yeast Strains, Media, and Growth Conditions . . . . . . . 743.2.2 Plasmids . . . . . . . . . . . . . . . . . . . . . . . . . . 773.2.3 Flow Cytometry . . . . . . . . . . . . . . . . . . . . . . 773.2.4 Sequencing . . . . . . . . . . . . . . . . . . . . . . . . . 803.2.5 WHI2 Plate Assay . . . . . . . . . . . . . . . . . . . . . 803.2.6 Turnover Assay . . . . . . . . . . . . . . . . . . . . . . . 813.2.7 Solubility Assay . . . . . . . . . . . . . . . . . . . . . . 813.2.8 Guk1-7-GFP Ubiquitination . . . . . . . . . . . . . . . . 813.2.9 Cellular Thermal Shift Assay (CETSA) . . . . . . . . . . 823.2.10 Statistical Analysis . . . . . . . . . . . . . . . . . . . . . 823.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 833.3.1 Multiple Strains From the Yeast Knockout Collection Dis-play Impaired Proteostasis . . . . . . . . . . . . . . . . . 833.3.2 A Secondary Mutation in WHI2 Co-Segregates with In-creased Guk1-7-GFP Stability . . . . . . . . . . . . . . . 853.3.3 Guk1-7-GFP Degradation is Impaired Owing to SecondaryMutations inWHI2 . . . . . . . . . . . . . . . . . . . . . 923.3.4 Reduced Proteostasic Capacity inWHI2Mutants is Linkedto Msn2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 953.3.5 Mutant WHI2 Impairs Guk1-7-GFP Degradation by Re-ducing Substrate Ubiquitination . . . . . . . . . . . . . . 96viii3.3.6 Essential E3 Ligase Rsp5 and Molecular Chaperones Ydj1and Ssa1 are Required for Guk1-7-GFP Degradation . . . 993.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1004 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1064.1 Chapter Summaries . . . . . . . . . . . . . . . . . . . . . . . . . 1064.2 General Discussion . . . . . . . . . . . . . . . . . . . . . . . . . 1074.2.1 Using Temperature Sensitive Alleles asModel Protein Qual-ity Control Substrates . . . . . . . . . . . . . . . . . . . . 1074.2.2 Flow Cytometry: An Ideal Method for Identifying and Char-acterizing Protein Quality Control Factors . . . . . . . . . 1094.2.3 Triage Decisions: Simply a Matter of Kinetic Partitioning? 1104.2.4 The Importance Of, and Difficulty In, Maintaining Pro-teostasis . . . . . . . . . . . . . . . . . . . . . . . . . . . 1114.3 Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . 1134.3.1 Flow Cytometry Screens for E3 Ligases Targeting HumanDisease Alleles . . . . . . . . . . . . . . . . . . . . . . . 1134.3.2 Characterizing the Role of the E3 Ligase Ubr1 in Cyto-plasmic Protein Quality Control . . . . . . . . . . . . . . 113Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116ixList of TablesTable 2.1 Yeast strains used in Chapter 2 . . . . . . . . . . . . . . . . . 30Table 2.2 Plasmids used in Chapter 2 . . . . . . . . . . . . . . . . . . . 33Table 2.3 Summary of FACS screen validation . . . . . . . . . . . . . . 48Table 3.1 Yeast strains used in Chapter 3 . . . . . . . . . . . . . . . . . 74Table 3.2 E3 ligase collection used for screening . . . . . . . . . . . . . 78Table 3.3 Plasmids used in Chapter 3 . . . . . . . . . . . . . . . . . . . 79xList of FiguresFigure 1.1 Proteostasis . . . . . . . . . . . . . . . . . . . . . . . . . . . 3Figure 1.2 Hsp70 reaction cycle . . . . . . . . . . . . . . . . . . . . . . 6Figure 1.3 E3 ubiquitin ligases . . . . . . . . . . . . . . . . . . . . . . . 10Figure 1.4 Protein quality control . . . . . . . . . . . . . . . . . . . . . 11Figure 1.5 The general stress response . . . . . . . . . . . . . . . . . . . 22Figure 2.1 Guk1-7 is thermally unstable . . . . . . . . . . . . . . . . . . 40Figure 2.2 Misfolded Guk1-7 is degraded at the non-permissive temperature 42Figure 2.3 Guk1-7 degradation is proteasome dependent . . . . . . . . . 44Figure 2.4 FACS-based screen . . . . . . . . . . . . . . . . . . . . . . . 47Figure 2.5 Ubr1 promotes Guk1-7-GFP degradation . . . . . . . . . . . 50Figure 2.6 Absence of Gim3 reduces Guk1-7 turnover . . . . . . . . . . 53Figure 2.7 Gim3 facilitates clearance of insoluble Guk1-7 . . . . . . . . 57Figure 2.8 Guk1-7-GFP puncta colocalize with Q-body markers . . . . . 58Figure 2.9 Gim3 helps maintain Guk1-7 solubility . . . . . . . . . . . . 59Figure 2.10 Thermosensitive alleles are stabilized by prefoldin subunits . . 61Figure 2.11 Model for stabilization of temperature sensitive alleles by Gim3 65Figure 2.12 Guk1-7-GFP flow cytometry . . . . . . . . . . . . . . . . . . 68Figure 2.13 Ubr1 does not act with San1 in the degradation of Guk1-7-GFP 69Figure 2.14 Guk1-7-GFP Gim3 interaction and viability assay . . . . . . . 71Figure 3.1 Flow cytometry based screen for E3 ligases targeting Guk1-7-GFPfor degradation . . . . . . . . . . . . . . . . . . . . . . . . . 85Figure 3.2 Guk1-7-GFP degradation in E3 ligase deletion strains . . . . . 87xiFigure 3.3 Guk1-7-GFP stability is not a direct effect of E3 ligase deletion 88Figure 3.4 Mutations in WHI2 segregate with the Guk1-7-GFP stabilityphenotype . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90Figure 3.5 das1∆ tetrad analysis and WHI2 addback . . . . . . . . . . . 91Figure 3.6 Absence ofWHI2 leads to Guk1-7-GFP stability . . . . . . . 94Figure 3.7 Msn2 is linked to reduced proteostatic capacity inWHI2 mutants 97Figure 3.8 whi2∆ promotes Guk1-7-GFP stability through reduced ubiq-uitination . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98Figure 3.9 A role for essential E3 ligases and molecular chaperones inGuk1-7-GFP degradation . . . . . . . . . . . . . . . . . . . . 101xiiGlossary∆∆G Free energy changeABCE1 ATP binding cassette subfamily E member 1ADP Adenosine diphosphateANOVA Analysis of varianceARS Autonomously replicating sequenceAsi1 Amino acid sensor independentAtg8 Autophagy relatedATP Adenosine triphosphateBmh2 Brain modulosignalin homologuebp Base pairBra7 Fluorocytosine resistanceBtn2 Batten diseaseCCT Chaperonin containing TCP-1Cdc48 Cell division cycleCEN Yeast centromereCETSA Cellular thermal shift assayCHX CycloheximideCHIP C-terminus of Hsc70-interacting proteinCFTR Cystic fibrosis transmembrane conductance regulatorCPY CarboxypeptidaseCytoQ Cytosolic quality control compartmentDas1 Dst1-delta6-azaurail sensitivityDBD DNA binding domainDeg1 Depressed growth rateDIC Differential interference contrastDNA Deoxynucleic acidxiiiDoa10 Suppressor of mRNA stability mutantDsRed Discosoma sp. red fluorescent proteinE1 Ubiquitin activating enzymeE2 Ubiquitin conjugating enzymeE3 Ubiquitin ligase enzymeEDTA Ethylenediamine tetraacetic acidEGFP Enhanced GFPER Endoplasmic reticulumERAD ER associated degradationERISQ Excess ribosomal protein quality controlEV Empty vectorFACS Fluorescence activated cell sortingFap1 FKBP12-associated proteinFDA Food and drug administrationFes1 Factor exchange for Ssa1pFITC Fluorescein isothiocyanateg GravityGDP Guanosine diphosphateGFP Green fluorescent proteinGim Gene involved in microtubule biogenesisGlo4 GlyoxalaseGMP Guanosine diphosphateGPD Triose-phosphate dehydrogenaseGPS Global protein stability analysisGuk1 Guanylate kinaseGus1 Glutamyl-tRNA synthetaseHbs1 Hsp70 subfamily B suppressorHCl Hydrogen chlorideHECT Homologous to the E6AP carboxyl terminus domainHEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acidHGMD Human gene mutation databaseHOP Hsp90 organizing proteinHrd1 HMG-coA reductase degradationxivHrt3 High level expression reduces Ty3 transpositionHSD Honest significant differenceHsp Heat shock proteinHul5 HECT ubiquitin ligaseINQ Intranuclear quality control compartmentIPOD Immobile protein depositIRES Internal ribosome entry siteJUNQ Juxtanuclear quality control compartmentkDa Kilo daltonLtn1 RING domain mutant killed by Rtf1 deletionµg MicrogramµL Microlitreµm MicrometermL MillilitremM MillimolarM MolarMr Molecular weightmin MinuteMAT Mating typeMPP11 DnaJ Hsp family member C2Msn Multicopy suppressor of SNF1 mutationMup3 Methionine uptaken NumberNAC Nascent chain associated complexNaCl Sodium chlorideNEDD4 Neural precursor cell expressed, developmentally downregulated 4NEF Nucleotide exchange factorNEMF Nuclear export mediator factorNES Nuclear export sequenceNLS Nuclear localization sequenceNMP Nucleoside monophosphate kinaseNP-40 Nonidet P-40Npl4 Nuclear protein localizationxvOD Optical densityORF Open reading frameP P-valuep62 Nucleoporin 62PAH Phenylalanine hydroxylasePBS Phosphate buffered salinePCR Polymerase chain reactionPDB Protein data bankPep4 Carboxy peptidase Y-deficientPgk1 3-phosphoglycerate kinasePKA Protein kinase APMSF Phenylmethane sulfonyl fluoridePolyQ PolyglutaminePrb1 Proteinase BPro3 Proline requiringPrc1 Proteinase CRAC Ribosome associated complexRBR RING-between-RINGRim15 Regulator of IME2RING Really interesting new geneRNA Ribonucleic acidRpt6 Regulatory particle triple-A proteinRQC Ribosome quality control complexRqc1 Ribosome quality control 1Rsp5 Reverses Spt-phenotypeS SolubleSan1 Sir antagonistSD Standard deviationSDG Saccharomyces genome databaseSDS Sodium dodecyl sulfateSDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresisSec61 SecretorySILAC Stable isotope labelling with amino acids in cell culturexviSir4 Silent information regulatorSis1 Slt4 suppressorSNV Single nucleotide variantSsa Stress sensitive subfamily ASsb Stress sensitive subfamily BSse1 Stress seventy subfamily ESti1 Stress inducibleSTRE Stress response elementst TimeT Total cell lysateTAD Transcriptional activating domainTae2 Translation-associated element 2TAP Tandem affinity purificationTcp Tailless complex polypeptideTom1 Trigger of mitosisTOR Target of rapamycinTPR Tetratricopeptide repeat domainTRiC TCP-1 ring complexTx-100 Triton X-100Ub UbiquitinUbc Ubiquitin conjugatingUbe2W Ubiquitin conjugating enzyme E2 WUbp Ubiquitin specific proteaseUbr1 Ubiquitin protein ligase E3 component N-recognin 1Ufd Ubiquitin fusion degradationUgp1 UDP-glucose pyrophosphorylaseUPS Ubiquitin proteasome systemUTR Untranslated regionVHL von Hippel LandauWhi2 WhiskeyX TimesYdj1 Yeast dnaJYPD Yeast extract peptone dextrosexviiAcknowledgementsI would like to thank everyone who has contributed to the work contained in thisthesis and who has helped me throughout the course of my degree. I would like toexpress my appreciation to my supervisor, Thibault Mayor, and to my supervisorycommittee members Christopher Loewen, Vivien Measday, and Michel Robergefor their advice, encouragement, guidance, and mentorship. Thank you to membersof the Mayor lab for valuable feedback and camaraderie. To Elizabeth, Allym,Aruna, and Carla, thank you for making me welcome in your lab and for yourkindness and generosity. Thank you to Megan Kofoed for all of her assistance withthe numerous yeast collections. Thank you to members of the Hansen lab, bothpast and present, for their friendship and making the office a fun place to be overthe past six years. I am especially grateful to Cheryl and Tan for their friendshipand insightful discussions. Finally, I would like to express my gratitude to myparents for their support and patience, without which none of this would have beenpossible.xviiiChapter 1Introduction1.1 Protein Misfolding and Protein HomeostasisProtein homeostasis, or proteostasis, preserves proteome integrity and thereby pro-motes viability at both the cellular and organism levels. To do so, proteostasis ismaintained by a network of interconnected pathways that influence the fate of pro-teins by directing their translation, folding, localization, and degradation [1]. Mis-folded proteins are one of many factors that challenge a cell’s ability to maintainproteostatic balance.Protein misfolding can result from a number of processes such as mutation; er-rors during transcription, RNA processing and translation; trapping of a folding in-termediate; failure to incorporate into multimeric complexes; or post-translationaldamage [2, 3]. The risk to the cell of misfolded or partially folded proteins may beattributed, at least in part, to the exposure of hydrophobic amino acid residues thatin the native state would be sequestered to the core of the protein, or at protein-protein interaction interfaces, but once exposed can engage in unspecific interac-tions with other polypeptides. These exposed hydrophobic regions of misfoldedproteins also have an inherent propensity to aggregate, forming associations notnative to the cell [4]. For example, artificial beta-sheet proteins expressed in humanHEK293T cells were found to coaggregate with proteins that have many functionalinteraction partners suggesting that the aggregates competitively bind to functionalprotein-protein interaction interfaces [5]. Moreover, the relative cytotoxicity of the1aggregates correlated with the number of interaction partners ascribed to the coag-gregating proteins. More recently, Kim et al. identified and analysed aberrant pro-tein interactions involving soluble oligomers and insoluble inclusions of the mutanthuntingtin protein [6]. Expressing a fragment of huntingtin, containing the Hunt-ington’s disease causing polyglutamine (PolyQ) repeat expansion, they found thatinsoluble inclusions predominantly interacted with members of the protein qualitycontrol machinery representing ∼85 proteins. Soluble oligomers interacted withupwards of 800 different proteins representing diverse cellular functions such astranscription, translation, and RNA-binding. Within the cytosol, macromolecularcrowding creates an environment that increases the tendency of folding interme-diates and misfolded proteins to aggregate, as aggregation is highly concentrationdependent [7]. The native conformation of a protein, however, must balance struc-tural stability with conformational flexibility that is associated with protein func-tion [8]. As such, a tightly regulated network of molecular chaperones contendwith a constant flux of protein intermediates, misfolded proteins, and aggregateformation [9].Proteostasis depends on balancing the folding capacity of chaperone networkswith the quantity of proteins in non-native conformations (Figure 1.1). When thelevel of misfolded proteins rises, the folding capacity of the cell can be temporarilyaugmented to meet the increased demand through the activation of signaling path-ways modulated by the transcription factors heat shock factor 1 (Hsf1) andMsn2/4.Stressors that precipitate protein misfolding disrupt equilibrium and if the cellularresponse is overwhelmed and insufficient to meet the increase in need, then it canlead to the accumulation and aggregation of misfolded proteins [10]. Although theexact mechanism that results in the formation of cellular aggregates has yet to befully elucidated, their presence is associated with a number of neurodegenerativeconditions such as Huntington’s, Parkinson’s, and Alzheimer’s diseases, as well asageing [11, 12]. Maintaining proteome integrity, therefore, requires an integratedprotein quality control network that monitors the proteome, mediates protein re-folding by molecular chaperones, and removes terminally misfolded proteins viathe ubiquitin proteasome system or autophagy [1].2ProteostasisStresses ResponsesPerturbations Adaptive ResponsesAgeingEnvironmental StressesGenetic MutationsMisfolded ProteinsProteasome DeficiencyUnassembled SubunitsAutophagyGeneral Stress ResponseHeat Shock ResponseMolecular ChaperonesSpatial PQCTranslation ControlUbiquitin Proteasome SystemFigure 1.1: Proteostasis. Proteostasis depends upon balancing the perturba-tions that disrupt protein folding with the network of pathways that direct thelevels, conformational state, and distribution of the proteome.1.2 Protein Folding and Cytosolic Molecular ChaperonesMolecular chaperones promote protein homeostasis by preventing protein aggrega-tion, assisting protein folding, and targeting terminally misfolded clients for degra-dation. Broadly, chaperones can be defined as any protein that recognizes andinteracts with proteins found in a non-native state for the purpose of stabilizing andpromoting folding into an active conformation without forming part of the finalstructure [10]. There are a number of distinct conserved chaperone families, oneof which is the heat shock protein (Hsp) family whose members are classified bytheir molecular weights (e.g. Hsp40, Hsp70, Hsp90, and the small Hsps). Withinthe context of protein quality control in the eukaryotic cytosol, the main chaperonemachineries involved are the: ribosome associated chaperones, Hsp70/40, Hsp90,TRiC/CCT chaperonin, and prefoldin.1.2.1 Nascent Protein FoldingA number of chaperones bind to, or associate with, the ribosome to both promoteprotein folding and prevent misfolding or aggregation of the nascent polypeptideas it emerges from the ribosome. As translation occurs at a slower rate than proteinfolding, nascent polypeptide chains emerge from the ribosome in a partially foldedaggregation prone state. Also, because the size of the ribosome exit tunnel is suchthat folding beyond the formation of alpha helical structures is prohibited, only lim-ited folding can proceed until a domain (generally 50 to 300 amino acids in length)3exits the ribosome [13]. Therefore, molecular chaperones interact co-translationalywith nascent polypeptide chains to prevent aggregation and premature non-nativefolding from occurring before the polypeptide has been fully translated. Moreover,while the ribosome exit sites are positioned in the polysome in such a way as tominimize aggregation of the nascent polypeptides, ribosome associated chaperonesare needed to further prevent aggregation of the numerous identical proteins beingtranslated from the polysome [14, 15]. The ribosome associated complex (RAC)and nascent chain associated complex (NAC) are the first chaperone complexesto interact with the nascent polypeptide as it exits the ribosome. In mammals,RAC is formed by the association of Hsp70L1 and the J-domain containing pro-tein MPP11. In yeast, RAC consists of the Hsp70 Ssz1 and the ribosome bindingHsp40 zuotin. Zuotin in turn acts in concert with the Hsp70s Ssb1 and Ssb2 that arealso associated to the ribosome [16–18]. Deleting the genes encoding the dimericNAC complex subunits and Ssb proteins in Saccharomyces cerevisiae resulted indecreased viability under conditions of protein folding stress [19]. Moreover, theabundance of ribosomal particles was altered in these mutants suggesting that ribo-some biogenesis is linked to the protein folding capacity of the ribosome associatedchaperones. Partially folded proteins are transferred by Hsp70 chaperones fromthe ribosome associated complex to further downstream folding pathways such asHsp70, prefoldin, and chaperonin.1.2.2 Hsp70, Hsp40, and Hsp90The Hsp70 family of molecular chaperones promotes protein folding through reit-erative cycles of ATP-dependent client capture and release (Figure 1.2). Hsp70s donot function independently, but as part of an Hsp70 core machinery consisting of anHsp70, an Hsp40 (J domain protein), and a nucleotide exchange factor (NEF) thatcoordinate their activities to increase the efficiency of Hsp70 client folding [20, 21].Hsp70 consists of an N-terminal ATPase domain and a C-terminal substrate bind-ing domain that binds to short 5-7 amino acid stretches of hydrophobic residueson client proteins. Dimerization is necessary for Hsp70 chaperone activity and ef-ficient Hsp40 interaction [22]. Initial client binding occurs while Hsp70 is in theATP-bound state and the Hsp70-client interaction is stabilized through ATP hy-4drolysis. The intrinsic ATPase activity of Hsp70 is relatively low however, andis enhanced by association with an Hsp40. As the number of Hsp40 proteins inthe cell greatly outnumbers that of the Hsp70s, it is thought that the repertoire ofHsp40 proteins help target Hsp70 to its clients, or bind them directly to enhancethe specificity of the system [23]. How each Hsp40 protein recognizes its cohort ofsubstrates remains to be fully understood. Once Hsp70 is bound to a client protein,Hsp40, via its conserved J domain, stimulates ATP hydrolysis and the subsequentactivity of a NEF promotes ADP dissociation and client release. Upon releasefrom Hsp70 the bound hydrophobic region is free to refold, however, it may re-associate with Hsp70 if folding is not complete. There are four non-ribosomalbinding Hsp70s (Ssa1-4) in the cytosol of S. cerevisiae and their ATP hydrolysisis enhanced by Ydj1 and Sis1 Hsp40 co-chaperones [24]. In addition to Fes1, aconfirmed NEF in the yeast cytosol, the Hsp110 Sse1 has also been reported toact as a NEF for the Hsp70s Ssa1 and Ssb1 [24]. The Hsp70 family can act bothco- and post-translationaly and are hubs to direct substrates to further downstreamchaperone networks such as Hsp90 and chaperonin [25].Hsp90 acts downstream of Hsp70 and Hsp40 to assist in the folding of nascenttranscription factors, protein kinases, and steroid hormone receptors [26]. Sub-strate transfer from the Hsp70/Hsp40 system to Hsp90 is mediated by the Hsp90organizing protein (HOP, or Sti1 in yeast) [27]. The tetratricopeptide repeat do-main (TPR) of HOP interacts with the MEEVD sequence on the C-terminal ofHsp90 bridging the two chaperone machineries and facilitating substrate transfer.Once bound to a substrate, Hsp90 ATPase activity is stimulated through an inter-action with Aha1. Chemical inhibition of Hsp90 function leads to the proteasomaldegradation of many Hsp90 substrates potentially as the result of increased inter-action with Hsp70 and Hsp70 associated factors [28].1.2.3 TRiC/CCT Chaperonin and PrefoldinChaperonin, also known as the TCP-1 ring complex (TRiC) or; the chaperonincontaining TCP-1 (CCT), functions in folding newly translated proteins and pre-venting protein aggregation in the cytosol [29, 30]. Essential in all three domains oflife, it has been estimated that upwards of 10% of the eukaryotic proteome transits5Hsp70ATPOpenOpenClosedPartially foldedproteinADPATPHsp40Hsp70Hsp40Non-native proteinNon-native proteinHsp40PiNEFATP123Figure 1.2: Hsp70 reaction cycle. 1) Hsp40 binds to misfolded substrates anddelivers them to an ATP bound Hsp70. 2) Substrates bind to Hsp70 via hy-drophobic patches (blue) and ATP hydrolysis to ADP, accelerated by Hsp40,switches Hsp70 to the closed substrate binding conformation. 3) Hsp40 dis-sociates from Hsp70 and a nucleotide exchange factor (NEF) exchanges ATPfor ADP causing Hsp70 to open and release the misfolded substrate enablingit to fold. Substrates can re-engage with Hsp40 to continue Hsp70 cycling orfold into their native conformation.through CCT, including cytoskeleton proteins and cell cycle regulators [31]. Chap-eronin is a large cylindrical 1 MDa protein complex formed through the stackingof two identical rings, each comprising eight subunits (CCT1-8) [30]. The apicaldomains of the ring subunits act as a lid to enclose the partially unfolded sub-strate in the central cavity mitigating the effects of macromolecular crowding onprotein folding. In eukaryotes, the chaperonin lid does not close entirely therebyaccommodating extended polypeptide chains and single domains of multidomainsubstrates [32]. CCT engages substrates while in the ATP-bound state with ATP6hydrolysis inducing conformational changes that lead to the closure of the chap-eronin lid and substrate encapsulation. Each ring is divided into two hemispheresbased on ATP binding affinity that leads to a cycle of asymmetric conformationalchanges [33]. CCT subunits 3, 6, 7, and 8 constitute one of the hemispheres andhave low affinity for ATP under physiological conditions and were found to be dis-pensable for chaperonin activity [33]. While all eight subunits have a conservedATP binding domain, their sequences are only ∼40% identical. The apical do-mains of each subunit, therefore, are thought to recognize different substrate mo-tifs, which is underscored by each subunit having its own specific patterns of polarand hydrophobic residues [34]. The pattern of amino acid residues in the each api-cal domain is thought to allow chaperonin to bind and fold a range of structurallydiverse proteins. Although initially identified through its requirement for fold-ing the cytoskeletal proteins actin and tubulin, the eukaryotic chaperonin has beenshown to act in the folding of a number of other substrates [35]. The human chaper-onin interactome is enriched for proteins predicted to be aggregation prone, whichcontain multiple domains and have complex topologies [36]. The importance ofCCT in human health is underscored by a mutation in CCT5 being identified asthe cause of autosomal recessive mutilating sensory neuropathology with spasticparaplegia [37]. CCT is also required for the replication of human pathogens suchas HIV and hepatitis C as well as folding a number of cancer associated proteins,such as p53 and the von Hippel Lindau tumor suppressor [38–40].Prefoldin is a hetero-oligomeric protein complex composed of six subunitsranging in size from 14–23 kDa [41]. Conserved in archaea and eukaryotes, butabsent in prokaryotes, the prefoldin hexamer forms a “jellyfish-like” structure withN- and C-terminal coiled-coil regions of each subunit forming “tentacles” that em-anate from a central region [42]. Misfolded substrates are transferred from pre-foldin to the TRiC/CCT chaperonin in an ATP-independent manner through di-rect binding of the two chaperone complexes [41]. In addition to its role in aidingnascent proteins, such as actin and tubulin, to attain their functional conforma-tions, the prefoldin chaperone complex has been shown to prevent huntingtin andalpha-synuclein aggregate formation [43, 44].71.3 Molecular Chaperones and Protein DegradationWhile molecular chaperones promote protein homeostasis by preventing proteinaggregation and promoting protein folding, they can also mediate the targetingof terminally misfolded clients for degradation; either indirectly, by maintainingmisfolded proteins in a non-aggregated degradation competent state, or directlythrough facilitating the recognition and/or transfer of substrates to degradativequality control pathways [45–49]. In an elegant study using the von Hippel Lin-dau (VHL) tumor suppressor protein as a misfolded model substrate, Frydman andcolleagues showed that a different set of chaperone proteins and co-factors mediatefolding and degradation, respectively [50]. Hsp90 and the co-chaperone Sti1 wererequired for degradation but not the folding of VHL, while the converse was truefor the chaperonin TRiC/CCT. The Ssa1/2 cytosolic Hsp70s, and the nucleotideexchange factor Sse1 were also required for degradation of VHL, as well as othercytosolic misfolded proteins [50, 51]. Although the yeast Hsp70 cofactor Ydj1 wasnot required for VHL degradation, it has been shown to mediate the degradation ofER proteins with exposed misfolded cytosolic domains and the degradation of cy-tosolic proteins after heat shock [52, 53]. Most recently Fes1, another yeast Hsp70NEF, was shown to promote proteasomal degradation of additional misfolded pro-teins [54]. In higher eukaryotes, the Bag6 Hsp70 cofactor that can bind to theproteasome via its ubiquitin-like domain is required for the efficient degradationof defective nascent polypeptides [55]. Although the degradation of distinct sub-strates has been demonstrated to require different chaperones and co-chaperones,in most cases it remains unclear how two competing systems (i.e. folding anddegradation machineries) triage misfolded proteins in the cell.1.4 Ubiquitin Proteasome System (UPS)In eukaryotes, the ubiquitin proteasome system plays a critical role in protein qual-ity control by selectively targeting intracellular proteins for degradation throughthe covalent attachment of polyubiquitin chains. Ubiquitin is a highly conserved8.5 kDa protein that is primarily conjugated onto lysine residues of target sub-strates through the activity of an enzymatic cascade involving ubiquitin activating(E1), ubiquitin conjugating (E2), and ubiquitin ligase (E3) enzymes [56]. Substrate8specificity and recruitment is mediated by an E3 ubiquitin ligase, either alone or incombination with an E2 conjugating enzyme. The inherent complexity of the ubiq-uitin system is reflected in the sheer number of putative E3 ligases (90 and 600)encoded in the genome of yeast and human, respectively [57]. Moreover, the im-portance of E3 ligases in substrate targeting is emphasized by the fact that the num-ber of putative E3 ligases greatly outnumbers that of E2 conjugating enzymes byapproximately 15:1 [58]. E3 ubiquitin ligases belong to one of three families char-acterized by their namesake domains (Figure 1.3). The really interesting new gene(RING) family are the most abundant E3 ligases with ∼600 members in humansand can be found as monomers, dimers, or as part of multisubunit complexes [59].The RING domain can be located anywhere on the protein and consists of a con-served consensus sequence of cysteine and/or histidine residues, which coordinatewith two zinc atoms to stabilize the domain structure [58]. RING E3 ligases act asscaffolds to orient the substrate with an E2-ubiquitin conjugate for efficient ubiqui-tin transfer. Approximately thirty proteins belong to the homologous to the E6APcarboxyl terminus domain (HECT) family with all known HECT domains locatedat the C-terminus of the protein. HECT E3 ligases play a direct role in substrateubiquitination through a stepwise process. First, the HECT E3 ligase, via an N-lobe E2 binding domain, receives an E2-ubiquitin conjugate to then form an E3-ubiquitin conjugate. Ubiquitin is then conjugated to a substrate from the E3 activesite via a C-lobe catalytic cysteine [58]. The final family, the RING-between-RINGor RBR E3 ligases, are the least abundant containing only 13 members in humans.RBR ligases contain an N-terminal RING domain (RING1), like the RING ligases,followed by in between RING and RING2 domains, which are unique to RBR pro-teins and do not contain the cysteine RING consensus sequence. In addition, likeHECT ligases, RBR ligases form a thioester bond with ubiquitin. Mutations inParkin, one of the most studied members of this family, are associated with earlyonset Parkinson’s disease [60].Ubiquitin has seven lysine residues, all of which can be conjugated to ubiquitinmolecules to form polyubiquitin chains. The K48 chain linkage is considered tobe the predominant signal recognized by the cell for proteasomal degradation. The26S proteasome is a large 2.5 MDa protein complex responsible for the selectiverecognition and degradation of ubiquitin conjugated proteins. It is composed of the9RING HECT RBRE2 conjugating enzyme RING domainSubstrate UbiquitinN-lobeC-lobeIBRRING1 RING2Figure 1.3: Schematic of the three E3 ubiquitin ligase families and E3 catal-ysed ubiquitin transfer.central barrel-like 20S core particle, which contains the proteolytic peptidases, andtwo 19S regulatory particles. The regulatory particle is responsible for substraterecognition, ubiquitin chain removal, and protein unfolding and translocation intothe catalytic core particle [61, 62]. In addition, the system is under constant flux asubiquitination can be reversed by deubiquitinating enzymes.There are multiple pathways that can target misfolded proteins for proteaso-mal degradation. Different types of protein damage are more prevalent in differentcellular compartments owing to the nature of the subcellular environment neces-sitating compartment-specific quality control pathways, with systems having beendescribed for the ER, nucleus, and cytoplasm [63].1.4.1 ER Associated Degradation (ERAD)Approximately one third of all eukaryotic proteins are membrane or secreted pro-teins that must pass through the ER [64]. Protein folding is monitored by ERquality control machinery and non-native or unassembled subunits are targeted fordegradation by the ER associated degradation (ERAD) pathway. If misfolded pro-teins are left to accumulate in the ER, a stress response is triggered in an attemptto rebalance the protein quality control system and to clear misfolded proteins.ERAD was initially discovered through studies of the cystic fibrosis transmem-brane conductance regulator (CFTR). While some components of the ERAD path-10way, such as the E3 ligases, are better defined in yeast, the identification of humanhomologs to several of the yeast genes involved would suggest that ERAD mayplay a role in ER proteostasis in higher eukaryotes as well. ER proteins targetedfor degradation must be retranslocated from the ER in an ATP-dependent mannerwhere they are degraded by the ubiquitin proteasome system [64]. The multispanER membrane RING E3 ligases Doa10 and Hrd1 ubiquitinate substrates on the cy-tosolic side of the ER following substrate retranslocation by a complex containingthe AAA-ATPase Cdc48, Ufd1, and Npl4 (p97, UFD1, and NPL4 in mammals)(Figure 1.4) [65]. The mechanism determining whether a substrate is targeted byDoa10 or Hrd1 is thought to be based upon the location of the degradation sig-nal. ER lumen and membrane substrates are generally recognized by Hrd1 andcytosolic substrates by Doa10 [66, 67].NucleusSan1, Asi1, Tom1Endoplasmic reticulumHrd1, Doa10CytoplasmUbr1, Hul5, Rsp5, Ltn1INQCytoQVacuoleIPODFigure 1.4: Schematic representation of the spatial distribution of proteinquality control compartments in yeast.111.4.2 Nuclear Protein Quality ControlDespite the fact that the majority of proteasomes are located within the nucleusunder non-stress conditions, our understanding of nuclear protein quality controlis relatively limited compared to that of the ER and cytoplasm [68]. In yeast, theprimary model organism used for nuclear protein quality control studies, the RINGE3 ligase San1, in conjunction with the E2 ubiquitin conjugating enzyme Ubc1,is responsible for the ubiquitination of misfolded nuclear proteins, thereby tar-geting them for degradation by the ubiquitin proteasome system in the nucleus(Figure 1.4) [69–71]. San1 recognizes exposed hydrophobic residues on misfoldedproteins that are normally buried in the native conformation [69]. These hydropho-bic stretches interact with substrate recognition sites on San1 that are interspersedbetween N- and C-terminal intrinsically disordered domains [72]. It is thought thatthese disordered regions, which lack secondary structure, provide flexibility suchthat San1 can bind to a large number of substrates with different conformations.The AAA-ATPase Cdc48/p97 has also been shown to be required for the degra-dation of some highly insoluble San1 substrates [73]. While the role of molecu-lar chaperones in nuclear protein quality control remains unclear, the chaperonesSis1 and Sse1 are required for the nuclear targeting of misfolded cytoplasmic pro-teins [74–76]. Moreover, inhibition of Hsp42 leads to the accumulation of cyto-plasmic proteins in the nucleus [77]. This would suggest a mechanism wherebysome proteins are normally retained in the cytosol and when this fails it leads tothe accumulation and aggregation of proteins within the nucleus. Why misfoldedcytoplasmic proteins are imported into the nucleus remains unknown, however, anumber of possible explanations have been proposed. The first, is that proteinsunder 40 kDa in size passively diffuse through nuclear pores into the nucleus. Thesecond, is that nuclear import is an active response in cases where cytoplasmicprotein quality control becomes overwhelmed. The third, is that it could be ad-vantageous to separate nuclear degradation from cytosolic protein folding. Morework will be required to determine whether one, or all, of these explanations iscorrect [78, 79]. It remains unclear as to which E3 ligase is required for nuclearprotein quality control in mammals.In addition to San1, two other nuclear E3 ligases have recently been character-12ized. Asi1 is a RING E3 ligase located in the inner nuclear membrane. The Asicomplex, consisting of the proteins Asi1-3, acts in parallel with Hrd1 and Doa10of the ERAD pathway to degrade soluble and integral membrane proteins [80]. Itis not yet known if Asi1 has a more general role in nuclear protein quality control,or how its substrates are recognized. Recently, the HECT E3 ligase Tom1 wasidentified in a screen, along with the E2 conjugating enzymes Ubc4 and Ubc5, tobe responsible for targeting overexpressed and unassembled ribosomal proteins fordegradation [81]. Tom1 specifically targets residues that would normally be hiddenin mature ribosome assemblies. Cells lacking TOM1 contained aggregated riboso-mal proteins. This new pathway named excess ribosomal protein quality control(ERISQ) is conserved as the human Tom1 homolog Huwe1 demonstrated a similarfunction in human cells.1.5 Cytosolic E3 Ubiquitin Ligases Involved in ProteinQuality ControlSeveral ubiquitin ligases in higher and lower eukaryotes are proposed to targetmisfolded cytosolic proteins for degradation, a number of which are described indetail below. Many of these E3 ligases act in conjunction with molecular chaper-one partners to target their misfolded clients for degradation. Their substrates arediverse ranging from nascent polypeptides that fail to attain their native confor-mation, stalled translation products, N-terminally destabilized polypeptides, andproteins which have become misfolded. How do E3 ligases recognize their sub-strates? Does each pathway target a specific subset of misfolded proteins? Doseveral ubiquitin ligases target the same proteins, potentially recognizing differentdomains or conformations? The search for the answers to these questions drivescurrent work in the cytoplasmic protein quality control field.1.5.1 CHIPC-terminus of Hsc70-interacting protein (CHIP) E3 ligase was shown over tenyears ago to be part of a major pathway targeting cytosolic misfolded proteinsfor degradation. A chaperone dependent ligase, CHIP interacts with both Hsp70and Hsp90 via its TPR domain, as well as with misfolded proteins that are then13ubiquitinated and targeted for degradation [49, 82–84]. Several co-factors werefound to influence CHIP activity. For instance, the Bag2 Hsp70 co-chaperone wasfound to interact and inhibit CHIP activity, favoring folding over degradation ofthe substrate [85, 86]. In contrast, the related Bag1 Hsp70 co-chaperone that con-tains a proteasomal interacting ubiquitin-like domain promotes CHIP activity andmay facilitate substrate delivery to the proteasome [87, 88]. In addition, CHIPauto-ubiquitination on lysine 2, mediated by the E2 conjugating enzyme Ube2W,enhances the E3 ligase activity of CHIP [89]. Preventing this self-ubiquitinationresults in a reduction of CHIP’s ability to ubiquitinate a variety of substrates. Thedeubiquitinating enzyme Ataxin 3 has been shown to regulate the ability of CHIPto ubiquitinate itself, as well as regulate the polyubiquitin chain lengths of CHIPsubstrates [90]. CHIP in turn, has been observed to ubiquitinate the polyglutamineexpanded form of Ataxin 3, targeting it for degradation [91]. It is still unclear whatcriterion determines the targeting of such proteins for degradation. Other ubiq-uitin ligases, like Parkin (for which mutations are linked to Parkinson’s disease)and Dorfin have been implicated in the targeting of cytosolic misfolded proteins,although more recent work indicates that Parkin may instead target defective mi-tochondria for macroautophagy [92–96]. Intriguingly, these ubiquitin ligases aremostly absent in lower eukaryotes like S. cerevisiae.1.5.2 Ubr1Ubr1 was first identified and characterized as being the E3 ligase of the N-end rule,a pathway whereby the half-life of a protein correlates with the identity of the N-terminal amino acid residue that is recognized by the ubiquitin ligase [97]. Ubr1recognizes N-end rule substrates through two domains: the UBR box (binds TypeI (Arg, Lys, or His) basic N-terminal amino acids) and the ClpS domain (bindsType II (Phe, Leu, Trp, Tyr, Ile) bulky hydrophobic residues) [98]. A number ofreports however, now lend support to Ubr1 playing a role in protein degradationindependent of the N-end rule [99–101]. Subsequently, it was shown that bothSan1 and Ubr1 are key E3 ligases in the cytosolic quality control machinery (Fig-ure 1.4) [74]. The cytosolic Ubr1, alone or together with the nuclear ubiquitinligase San1, was found to target a large variety of cytosolic misfolded proteins14including artificial model substrates, thermosensitive mutant alleles and unfoldedkinases [74, 101–104]. In some cases, Ubr1 ubiquitinates cytosolic substrates withthe assistance of Sse1 and Ssa1 chaperones, while the nuclear localized San1 ubiq-uitinates substrates that are delivered to it from the cytosol with the help of theSis1 Hsp40 [48, 75]. Ubr1 is highly conserved with one mammalian homologueknown to play a role in protein quality control [105]. Mutations in human Ubr1 areresponsible for the autosomal recessive Johanson-Blizzard syndrome characterizedby developmental abnormalities and pancreatic insufficiency [106].1.5.3 Hul5 and Rsp5Exposure to heat shock stress induces a conserved cytoprotective heat shock re-sponse that, in addition to transcriptional induction and repression, results in in-creased protein ubiquitination and degradation of primarily cytosolic proteins [107–109]. The HECT E3 ligases Hul5 and Rsp5 are both required for the increasedubiquitination of cytosolic proteins observed following heat shock (Figure 1.4) [52,109]. Hul5 is a proteasome associated protein, with chain elongation activity in op-position to the proteasome bound deubiquitinating enzyme Ubp6 [110, 111]. WhileHul5 is mainly nuclear in unstressed cells, it relocalizes to the cytoplasm upon heatshock and its cytosolic localization is required for the targeting of cytosolic mis-folded proteins for proteasomal degradation [109]. As an E4 ligase, Hul5 promotesthe elongation of polyubiquitin chains initiated by other E3 ligases, thereby in-creasing substrate processivity at the proteasome [110]. Rsp5 is essential for yeastviability and has a role in a number of cellular processes such as endocytosis, RNAexport, and lipid biosynthesis [112–114]. Following heat shock, Rsp5 interactswith the Hsp40 chaperone Ydj1 to promote substrate ubiquitination [52]. Rsp5’srole in ubiquitinating proteins following heat shock is conserved as the homologNEDD4 is also required for heat shock induced ubiquitination in higher eukary-otes [52].1.5.4 Ltn1Nascent polypeptides on stalled ribosomes have been shown to be ubiquitinatedand targeted for proteasomal degradation by the E3 ligase Ltn1 (Figure 1.4) [115].15Ltn1 targets non-stop proteins (derived from non-stop mRNA lacking a terminationcodon) and proteins containing polylysine stretches for ubiquitination and subse-quent degradation in yeast, but was shown not to play a role in general cytoso-lic quality control when tested against the VHL quality control substrate [115].Stalled 80S ribosomes are dissociated by the ribosome recycling factors Hbs1-Pelota-ABCE1 into 40S small subunits and 60S nascent chain tRNA complexesthat facilitate the recognition of the nascent polypeptide by Ltn1 [116]. ExposedtRNA is recognized by Rqc2 (NEMF in mammals) prohibiting 40S reassociationand promoting Ltn1 recruitment [117]. Ltn1 associates with the 60S ribosomeand functions as part of a ribosome quality control complex (RQC) comprisingCdc48, the translation-associated element 2 (Tae2), and the protein ribosome qual-ity control 1 (Rqc1) [118, 119]. Ltn1 binds to ribosomal proteins in a way suchthat its RING domain is oriented towards the exit tunnel [120]. Following ubiq-uitination by Ltn1, and as a prerequisite for proteasomal degradation, the tRNA-linked polypeptide is dissociated from the 60S ribosome through the activity of theCdc48-Ufd1-Npl4 complex [121]. Recently, Ltn1 has also been shown to mediatethe degradation of translationally stalled ER proteins [122]. This function requirescytosolic exposure of the nascent polypeptide at the ribosome-Sec61 transloca-tion channel junction [117]. Targeting proteins during cotranslational translocationprevents complete translocation into the ER, thereby eliminating the need to re-translocate the protein back into the cytosol and bypassing the ERAD network.Ltn1’s structure, determined by single-particle electron microscopy, is similar tothe cullin subunit of the cullin-RING ubiquitin ligases, but has significant confor-mational variability that could be integral for its function [123]. The importance ofLtn1 is underscored by the results of an N-ethyl-N-nitrosourea mutagenesis screenthat identified homozygous lister mouse mutants that are viable, but display pro-gressive early onset neurodegeneration [124].1.6 AutophagyAutophagy, the process whereby cytoplasmic components are degraded by thelysosome, is important for recycling amino acids during nutrient starvation and forthe clearance of aggregated proteins and damaged organelles, such as mitochondria16and ribosomes. Three types of autophagy have been described, each categorized bythe mechanisms required to deliver substrates to the lysosome [125]. Cellular com-ponents destined for degradation via macroautophagy are encapsulated through theformation of double membraned autophagosomes that fuse with the lysosome (orvacuole in fungi) delivering their contents to be degraded by enzymes. In microau-tophagy, the lysosomal membrane is remodeled to capture cellular componentsbringing them directly into the lysosome in a fashion reminiscent of phagocytosis.Finally, chaperone-mediated autophagy requires the selective import of unfoldedproteins into the lysosome through a combination of chaperone mediated substratetargeting and a set of dedicated receptors and translocation machinery. While au-tophagy was originally viewed as a non-selective process, whereby the lysosomeindiscriminately engulfed portions of the cytosol, many studies now demonstratethat macroautophagy can selectively target protein aggregates and organelles forlysosomal degradation. Moreover, the ubiquitin proteasome system and autophagyare interconnected, as a compensatory increase in autophagy is observed whenproteasome activity is impaired or inhibited [126]. Under starvation conditions,ribosomes and proteasomes undergo selective lysosomal degradation in a processcalled ribophagy and proteaphagy, respectively [127–129]. The E3 ubiquitin ligaseLtn1 protects 60S ribosomal subunits from starvation-induced selective ribophagyin a process antagonised by the deubiquitinating enzyme Ubp3 [130]. Similarly, se-lective mitochondrial degradation, or mitophagy, is important for maintaining mi-tochondrial integrity and for limiting the production of potentially harmful reactiveoxygen species [131]. Parkin, an E3 ligase of the outer mitochondrial membrane,has been associated with mitophagy suggesting that some outer mitochondrialmembrane proteins require ubiquitination in order to promote selective macroau-tophagy [132]. In cases where damaged proteins presumably can no longer beprocessed by the proteasome, protein aggregates accumulate adjacent to the vac-uole, presumably to be cleared by autophagy [125]. Such aggregates colocalizewith Atg8, the homolog to the mammalian autophagosome LC3 protein, whichacts as a receptor for ubiquitin binding proteins. For instance, p62 and Nbr1 pro-mote the turnover of polyubiquitinated protein aggregates by selectively binding toK63 ubiquitin chains, which are recognized through a ubiquitin binding domain,while also binding to LC3 to shuttle the substrates to autophagosomes [133]. While17it is clear that a level of reciprocity exists between the autophagy and UPS path-ways, a greater appreciation of the protein quality control elements will be neededbefore we can truly understand how substrates are triaged between these two com-partments.1.7 Spatial Protein Quality Control: CytoQ, IPOD, andINQProtein aggregation has traditionally been viewed as a last resort when proteinquality control is exhausted. More recently however, the perception of spatial se-questration of misfolded proteins has changed, and it is now believed to representan early event in protein quality control and to occur even under physiologicalconditions. In S. cerevisiae, there are three spatially distinct protein quality con-trol compartments that sequester misfolded or aggregated proteins into inclusionswithin the cell. These are: the cytosolic quality control compartment (CytoQ), theimmobile protein deposit (IPOD), and the intranuclear quality control compart-ment (INQ). These compartments are not unique to yeast as similar cytoplasmicinclusions have been described in mammalian cells [134, 135].CytoQ inclusions (also referred to as stress foci or Q-bodies) are found through-out the cytosol and require Hsp42 for their formation following heat stress [136,137]. Hsp42, along with Hsp26 constitute the cytosolic members of the smallheat shock family of molecular chaperones and, like all small heat shock proteins,have a conserved alpha-crystallin C-terminal domain [138, 139]. Functional understress and non-stress conditions, Hsp42 binds to misfolded proteins to prevent pro-tein aggregation. In addition to being a monomer, Hsp42 can also form barrel likeoligomeric structures from hexameric rings of dimers if present at high concentra-tions. Hsp42 is exclusive to CytoQ and is used as a marker for this compartment.A single IPOD inclusion is found adjacent to the vacuole and formation is in-dependent of stress [135]. This deposit does not co-localize with proteasomes andcontains insoluble non-ubiquitinated proteins as well as amyloid proteins [135]. Todate, most attention has been paid to the INQ quality control compartment. Orig-inally thought to associate with the nucleus while remaining in the cytosol, thejuxtanuclear quality control compartment (JUNQ) has recently been discovered to18reside within the nucleus in close proximity to the nucleolus and has, as a con-sequence, been renamed the intranuclear quality control compartment (INQ) [77,135]. Cytosolic proteins require active transport through the nuclear pore complexto reach INQ and substrate targeting and aggregation is mediated by Sis1 in thecytosol and Btn2 within the nucleus [77, 140]. Sis1 alone however, is not suffi-cient to target proteins for nuclear import suggesting that other factors remain tobe discovered [140]. Sis1 levels are relatively high under both physiological andstress conditions, while Btn2 is barely detectable and must be rapidly induced uponheat shock. Even under stress conditions Btn2 is rapidly degraded and inhibitingits degradation stabilizes INQ deposits underscoring its importance in nuclear in-clusion formation [77]. Hsp104 is an AAA-ATPase that associates with aggregatesto assist with their disassembly [141]. Hsp104 is used as a general aggregationmarker and is conserved in fungi and plants but no metazoan homolog has yetbeen identified. While not essential for viability, Hsp104 is required for inducedthermotolerance in yeast [142]. Ubiquitination was once thought to be the sortingsignal dictating protein sorting to the INQ compartment [135]. INQ’s associationwith Hsp104 however, suggests instead that sequestration of misfolded proteins oc-curs prior to, or independently from, the decision to refold or degrade a misfoldedsubstrate.1.8 Stress ResponsesExposure to a range of intrinsic or extrinsic stressors can precipitate protein mis-folding overwhelming the proteostasis capacity of the cell. Depending on the na-ture of the stress, the cell can elicit a number of cellular responses to ensure survivaland restore proteostasis. Common to most of these is the induction of molecularchaperones and other factors required to mitigate the stress as well as a decreasein the transcription, translation, and splicing of all other factors not essential to thestress response [143]. While a number of pathways have been described, the mostwidely studied are the heat shock and general stress responses.191.8.1 Heat Shock and General Stress ResponseExposure to elevated temperatures results in a highly conserved physiological heatshock response, which is characterized by induced expression of genes includingmembers of the heat shock molecular chaperone family. While the heat shockresponse is cytoprotective, many of the genes induced are not required for surviv-ing the initial stress, but are instead necessary for surviving subsequent stresses,thereby forming acquired stress resistence [144]. Genes are induced through bind-ing of the heat shock transcription factor Hsf1 to heat shock elements in promoterregions [141]. Vertebrates and plants have four Hsf proteins, with the Hsf1 iso-form primarily responsible for the heat shock response [145]. In contrast, inverte-brates and yeast have a single Hsf1 protein. Low level Hsf1 activity is essentialfor yeast viability and is required for basal expression of Hsp70 and Hsp90 chap-erones [146]. In higher eukaryotes under non-stress conditions, Hsf1 is maintainedin an inactive monomeric form in the cytoplasm through an interaction with Hsp90proteins. Exposure to stress releases Hsf1 resulting in its trimerization, which isrequired for DNA binding and gene induction [141].A broad range of environmental stresses such as heat, nutrient starvation, os-motic shock, and oxidation precipitate a transcriptional response in eukaryotes.This general stress response, resulting in the induction of approximately 200 genes,is mediated by the zinc-finger transcription factors Msn2 and Msn4 which bindto stress response elements (STRE) in the promoter regions of target genes (Fig-ure 1.5) [147]. Originally, the heat shock response was considered to be a subsetof the general stress response, but a recent report suggests that, in yeast at least,the heat shock response is largely Hsf1 independent and, instead, the heat shocktranscriptional response is predominantly driven by Msn2/4 activity [146]. Msn2and Msn4 are partially redundant transcription factors that share 41% sequenceidentity at the amino acid level. Neither gene is essential in yeast and they are notconserved from yeast to metazoans [145]. While Msn4 expression is induced bystress, Msn2 is constitutively expressed and is thought to play the dominant rolein stress response as overexpression of MSN4 can only partially suppress the phe-notype of the msn2∆ mutant [148]. Msn2 contains an N-terminal transcriptionalactivating domain (TAD), a nuclear export sequence (NES), a nuclear localization20sequence (NLS), and a C-terminal zinc finger DNA binding domain (DBD) [149].Structurally, Msn2 is predicted to be intrinsically disordered with the exceptionof two structured regions in the TAD domain. The sequences of these structuredmotifs are highly conserved in yeast and mutations result in decreased Msn2 ac-tivity and nuclear localization [149]. Msn2 activity is thought to be regulated bytwo nutrient sensing pathways: protein kinase A (PKA) and target of rapamycin(TOR) [141]. Under non-stress conditions, cAMP dependent PKA phosphorylationnegatively regulates Msn2 by phosphorylating the nuclear localization sequence,thereby retaining Msn2 in the cytoplasm [150]. Nuclear exclusion in the absenceof stress is also thought to be mediated by an interaction between Msn2 and the 14-3-3 protein homolog Bmh2, which is enhanced by TOR activity [150]. A secondPKA consensus site on Msn2 regulates nuclear export, which requires the Msn5exportin receptor that controls the nuclear localization of many transcription fac-tors (Figure 1.5) [150, 151]. Msn2 is primarily found in the nucleus under certainconditions such as when TOR activity is inhibited, in msn5∆ cells, or when PKAlevels decrease [150–152]. Interestingly, Msn2/4 display oscillatory nucleocyto-plasmic shuttling under intermediate stress conditions that is regulated by PKAlevels in the case of Msn2, but not for Msn4 [153]. How this oscillatory shuttlingrelates to transcriptional activity remains unknown. Msn2/4 bind to a five base pair(CCCCT or AGGGG) consensus binding site resulting in a transcriptional responsethat is both transient and scales with the magnitude of the stress [147]. This is inpart the product of a linear relationship between induced gene expression and theconcentration of nuclear Msn2, which is produced by low Msn2 binding affinityand a limited number of Msn2 molecules relative to the number of STRE bindingsites in the genome [154–156]. The combination of environmental sensing path-ways regulating Msn2 localization and activity and the linear relationship betweenMsn2 concentration and target gene expression means that Msn2/4 can mediate acommensurate homeostatic response to a range of extrinsic stresses.1.9 DiseasesProtein homeostasis networks maintain proteome integrity and are essential forcell viability. Perturbations that disrupt the equilibrium of this system can lead21NucleusCytoplasmSTREMsn2/4Msn2/4P Bmh2TORPKAMsn5Figure 1.5: The general stress response. Under a range of stress conditionsthe transcription factors Msn2 and Msn4 bind to stress response elements(STREs) in the promoter regions of target genes. Nuclear localization andimport is regulated by the PKA and TOR pathways. Nuclear export is medi-ated by PKA activity and the exportin Msn5.to a class of diseases known as proteopathies, which range from lysosomal stor-age diseases to cystic fibrosis and neurodegenerative disorders [157]. Protein mis-folding, which can lead to protein aggregation, is characteristic of a number ofproteopathies. Moreover, it is thought that an age related decline in the cell’s ca-pacity to respond to the presence of misfolded proteins underlies the late onset ofneurodegenerative diseases such as Alzheimer’s and Parkinson’s [12]. The effectof missense mutations on protein stability is of particular interest in the contextof disease as missense mutations represent more than half of all mutations in theHuman Gene Mutation Database (HGMD) [158]. Sahni and colleagues tested ap-proximately 3000 human disease associated missense alleles and found about onethird of the mutations altered protein stability and resulted in an increased engage-ment with components of the protein homeostasis network [159].22There is currently great interest in the potential for developing therapeutics thattarget proteostatic imbalance and components of the ubiquitin proteasome system.One such example is Bortezomib, a proteasome inhibitor used to treat relapsedmultiple myeloma [160]. The selectivity of proteasome inhibition to kill tumorcells as opposed to normal healthy cells is thought to be attributed to tumor cellsbeing more sensitive to proteasome inhibition due to higher concentrations of ab-normal proteins [161]. Recently, selective proteasome inhibition by a compoundtargeting the kinetoplastid proteasome was shown to clear mice of the parasitesresponsible for leishmaniasis, sleeping sickness, and Chagas disease, which leadto 50,000 deaths annually and affect more than 20 million people globally [162].Therapies targeting molecular chaperones are also being developed for the treat-ment of diseases ranging from cancer to neurodegeneration. For instance, a recom-binant human HSP70 therapy was shown to reduce a number of disease associatedneurological symptoms in mouse models of lysosomal storage diseases [163]. Inaddition, the drug Lumacaftor, which acts as a chaperone, was recently approved bythe food and drug administration (FDA) to treat patients with the F508∆ mutationin CFTR [164]. Together, these examples highlight the exciting potential targetingprotein homeostasis networks have for drug development and clinical applications.1.10 Model Substrates Used to Study ProteostasisProtein quality control pathways have been identified and characterized using awide range of model substrates. These substrates are essential components of ge-netic screens that have been used to probe protein quality control and will con-tinue to be vital if we are to understand what aspects within misfolded proteinsare necessary for recognition by molecular chaperones and E3 ubiquitin ligases totarget them for degradation. Model substrates used in the study of ER, nuclear, andcytoplasmic protein quality control include: VHL, CPY*, Ura3, Ubc9, and GFPfusions.VHL is an E3 ligase that acts as a tumor suppressor with mutations leadingto a disease of the same name. VHL folding and stability is coupled to its as-sembly into a complex containing elongin B and C. An absence of the elonginpartners, or mutations that disrupt binding, results in VHL being degraded by the23proteasome [165]. Folding defective mutants of VHL were used to examine howdifferent molecular chaperones contribute to the triage decision of whether to foldor degrade misfolded proteins. McClellan and colleagues demonstrated that somechaperones, such as the TRiC chaperonin, were only required for folding, whereasHsp90 was necessary for VHL degradation, and Hsp70 had a role in both foldingand degradation [50].The vacuolar carboxypeptidase (CPY) encoded by the gene PRC1, has beeninstrumental in the study of ER and cytoplasmic protein quality control. Mutantprc1-1 (or CPY*) is retained in the ER and targeted for degradation while thewild type protein is located in the vacuole. Genetic screens looking for mutantsthat are defective in CPY* degradation isolated key factors of the ERAD pathway,including the E3 ligase Hrd1 and E2 conjuating enzymes Ubc6 and Ubc7 [166].∆ssCPY*, a truncated version of the mutant CPY* protein that has had its signalsequence removed restricting its localization to the cytosol, has been used in thediscovery of cytosolic protein quality control pathways [53, 74, 103]. Primarily,∆ssCPY* substrates have been used to delineate the role of Ubr1 and San1 in thedegradation of misfolded cytoplasmic proteins [74, 103].Fused to a model substrate or short peptide, Ura3 is used as a reporter proteinin genetic screens to identify protein quality control components. It has been usedto screen for mutations in the Type I and II substrate binding sites of Ubr1 andmore recently used to generate a new panel of model substrates through fusionwith a degron library [98]. Screening this panel of substrates revealed a globalrequirement for the molecular chaperones Ssa1, 2 and Ydj1 as well as a novel rolefor Ltn1 in a mechanism distinct from ribosomal quality control [167]. The E3ligase Doa10 was also identified as the primary ligase required for these substrates.Ubc9 is essential for yeast viability and is required for cyclin degradation [168].A temperature sensitive allele of Ubc9 was identified and found to undergo con-ditional proteasomal degradation [169]. More recently, Ubc9 has been used as agreen fluorescent protein (GFP) fusion protein in the study of the INQ and CytoQpathways [135, 136]. As is the case with Ubc9, the majority of the work presentedin this thesis relies upon fusing novel model substrates to GFP to study cytosolicprotein quality control. GFP is a 27 kDa protein originally isolated from the jelly-fish Aequorea victoria that emits green light at a wavelength of 509 nm and can be24used to tag proteins at their N- or C-termini [170]. It forms a cylindrical beta barrelstructure consisting of eleven beta strands with a central alpha helix that is cova-lently bonded to the chromophore. The GFP chromophore is formed through thecyclisation and oxidation of three amino acids (Ser65, Tyr66, and Gly67), whichoccurs within two to four hours of synthesis [171, 172]. The S65T GFP mutantis more amenable to biological applications as it has a faster maturation time, ismore resistant to photobleaching, and its single excitation peak at 490 nm meansthat it can be used with fluorescein isothiocyanate (FITC) filter sets [173]. Theadvent of whole proteome GFP tagging collections has meant that it was possibleto perform high throughput studies using flow cytometry to identify factors thatinfluence protein stability or abundance and shifted the focus of flow cytometryscreens away from single substrates or a small collection of deletion strains. Twomethodologies highlight these advances: global protein stability profiling and tan-dem fluorescent protein timers. Global protein stability (GPS) analysis is a methodfor analysing protein turnover at the proteome level in mammalian cells [174]. Twofluorescent proteins, an internal control DsRed and an EGFP fusion with a proteinof interest, are translated from a single mRNA transcript containing an internalribosome entry site (IRES). The EGFP/DsRed ratio of a cell represents the sta-bility of the protein of interest as both fluorescent proteins are produced from thesame mRNA. Changes to the stability of the GFP fusion protein will therefore bereflected by a change in the EGFP/DsRed ratio. EGFP/DsRed constructs were cre-ated for the entire human ORFeome containing approximately 8000 human proteinencoding open reading frames (ORFs) and pooled transformed cells are fluores-cence activated cell sorted (FACS) into bins based on the GFP/DsRed ratio andthen microarray analysis is performed to identify the tagged ORF. This approachwas used to successfully identify substrates of the SCF ubiquitin ligase in mam-malian cells [175]. The tandem fluorescent protein timer method uses a similar dualfluorescent protein approach, however, in this case, the two proteins are fused andmature with different kinetics. The fluorescence ratio of the two proteins providesa measure of protein age and has been used to identify regulators of the N-end rulepathway in yeast [176].251.11 Research ObjectiveProtein homeostasis encompasses the network of pathways that influence the fateof proteins from synthesis to degradation for the purpose of maintaining proteomeintegrity, thereby promoting viability at both the cellular and organism levels. Mis-folded proteins challenge the cell’s capacity to maintain the proteostatic balanceand may divert resources away from essential cellular processes or result in the pro-duction of potentially toxic protein aggregates. Consequently, cells have adoptednumerous protein quality control pathways to prevent aberrant protein aggregation,promote protein folding, and to target terminally misfolded proteins for degrada-tion. Previous work from the Mayor lab identified a panel of temperature sensitivealleles of essential genes encoding for cytosolic proteins in S. cerevisiae that aredegraded in a proteasome-dependent manner once shifted to an elevated tempera-ture of 37◦C. The protein quality control pathways responsible for the degradationof a number of these alleles, which contain potentially destabilizing missense mu-tations, are unknown. Recently, there has been renewed interest in the role thatmissense mutations play in genetic disease as they can induce protein instabilitywhich leads to premature and/or increased rates of protein degradation and, asa consequence, loss of function phenotypes. My hypothesis is that a number ofquality control pathways, both known and as yet undiscovered, are present withinthe cytoplasm to aid the cell in the recognition, refolding and/or degradation ofproteins destabilized by missense mutations. This thesis is focused on identifyingand characterizing cytosolic protein quality control factors that induce proteasome-mediated degradation of thermally unstable model substrates.1.11.1 Specific Aims1. Develop a flow cytometry based assay to monitor the stability of a GFP-tagged substrate.2. Use genetic screens to identify protein quality control factors that promoteproteasomal degradation of a model substrate.3. Perform in depth characterization of the factors identified in Aim2.This work was performed using the model organism Saccharomyces cerevisiaewith a combination of cell biology, biochemical, and genetic approaches.26Chapter 2Prefoldin Promotes ProteasomalDegradation of Cytosolic Proteinswith Missense Mutations byMaintaining Substrate Solubility2.1 IntroductionThe protein homeostasis network encompasses systems required by the cell to gen-erate and maintain the correct levels, conformational state, and distribution of itsproteome [1]. Misfolded proteins threaten this balance by triggering loss of func-tion phenotypes, diverting resources away from producing essential protein prod-ucts, or precipitating the production of potentially toxic protein aggregates [4]. Thepresence of protein aggregates is characteristic of a number of neurodegenerativediseases such as Parkinson’s and Alzheimer’s disease, and a decrease in the proteinhomeostasis capacity of the cell is thought to underlie the later stages of cellularageing [11, 12, 177]. It is, therefore, not surprising that the cell has evolved a num-ber of protein quality control pathways aimed at preventing protein aggregation,promoting protein folding, and targeting terminally misfolded proteins for degra-dation [178–180]. These pathways triage misfolded proteins, which will face three27main possible fates: to be refolded back to their functional native conformation;to be targeted for degradation; or to be sequestered into spatially distinct qualitycontrol compartments.Proteins are selectively targeted to the eukaryotic ubiquitin proteasome sys-tem by the covalent attachment of polyubiquitin chains catalyzed by a cascadeof E1 (ubiquitin-activating), E2 (ubiquitin-conjugating), and E3 (ubiquitin ligase)enzymes [62, 181]. Substrate recruitment and specificity is determined by the E3ubiquitin ligases, either alone or in concert with an E2 conjugating enzyme or othersubstrate adaptors. A number of subcellular compartment-specific quality controlpathways have been identified, each associated with a particular E3 ligase or set ofligases [63, 70, 178]. In yeast, the San1 ligase is responsible for ubiquitinating nu-clear misfolded proteins [70]. Experiments have shown that San1 binds misfoldedproteins through recognition sequences located in disordered regions of its N- andC-terminal domains [72]. In contrast to the nucleus, a number of ligases have beenidentified to target cytosolic proteins for degradation in yeast. While initially char-acterized for its role as the recognin of the N-end rule pathway, Ubr1 has also beenshown to target misfolded cytoplasmic proteins for degradation [74, 76, 102–104].It does so either alone, or in conjunction with other E3 ligases such as Ubr2 in thecase of newly synthesized kinases, or with the nuclear San1 where both are requiredfor the complete degradation of the engineered ∆ssCPY*-GFP substrate [74, 104].Hul5, a nuclear protein that relocalizes to the cytoplasm upon heat shock, and Rsp5have been identified as the two ligases responsible for the marked increase in cy-toplasmic protein ubiquitination following heat shock stress [52, 182]. Finally, theribosome associated ligase Ltn1 targets non-stop polypeptides stalled during trans-lation for degradation [115].Recently, the importance of spatial organization in protein quality control hasgained recognition. Under normal physiological conditions, misfolded proteins canbe concentrated into dynamic Q-bodies where they can be refolded by chaperonesor degraded [136]. However, if the protein quality control systems become over-whelmed, misfolded proteins can be sequestered into discrete cellular inclusions.The INQ compartment acts to concentrate detergent soluble misfolded proteinscapable of being refolded, or degraded, and contains 26S proteasomes and chap-erones such as the disaggregase Hsp104 [77, 135]. The IPOD by contrast contains28insoluble non-ubiquitinated proteins; does not co-localize with proteasomes; andis the site of amyloidogenic protein sequestration, perhaps to prevent their toxicinteraction with quality control machinery [135]. The IPOD is also postulated tobe the site of yeast prion maturation [183].In this study we performed a screen to identify factors involved in degradativeprotein quality control of a model substrate that misfolds as the result of destabiliz-ing missense mutations. We show that our model substrate is thermally unstable,undergoes proteasome mediated degradation, and forms Q-body like inclusions.We then identified and characterized the prefoldin chaperone subunit Gim3 as afactor important for maintaining our substrate protein’s solubility, and thereby fa-cilitating its degradation.2.2 Materials and Methods2.2.1 Yeast Strains, Plasmids, and MediaAll yeast deletion strains used in this study are derived from BY4741 or BY4742wild type (WT) strains and are listed in Table 2.1. The temperature sensitive alleleswere generously provided by Dr. P. Hieter. The Cup1-Deg1-GFP plasmid wasa gift from T. Sommer [184]. The Hsp104-mCherry and Hsp42-mCherry strainswere constructed by homologous recombination of a PCR product amplified froma plasmid containing a yeast codon optimized mCherry ORF (BPM 866). Guk1and Guk1-7 GFP-tagged fusion plasmids (BPM 453, BPM 458) were constructedby inserting ORFs amplified from genomic DNA, with primers containing BamHIand XbaI restriction enzyme recognition sequences, into PGPD-GFP(S65T) (BPM241). Ugp1-3 (BPM 457), Pro3-1 (BPM 507), and Gus1-3 (BPM 500) GFP taggedplasmids were produced in the same fashion using: BamHI and NotI; BamHI andNotI; and NotI and XbaI, respectively. The histidine tagged fusions were producedby cloning PCR amplified inserts into PGPD (BPM 171) using BamHI and SalI(BPM 659, BPM 717). All plasmids used in this study are listed in Table 2.2. Cellswere grown in synthetic drop out media following standard procedures.29Table 2.1: Yeast strains used in Chapter 2Strain ID Alias Genotype SourceYTM 408 BY4741 ura3∆0, leu2∆0, his3∆1, met15∆0OpenBiosystemsCollectionYTM 703 ubr1∆san1∆ his3∆1, leu2∆0, met15∆0, ura3∆0,san1∆::His3MX6, ubr1∆::KanMXKhosrow-Khavaret al. 2012YTM 736 Guk1-7-13mychis3∆1, leu2∆0, LYS2,met15∆0, ura3∆0, guk1-7-13myc::KanMX6::URA3, CAN1Khosrow-Khavaret al. 2012YTM 749 Gus1-3ura3∆0, leu2∆0, his3∆1, LYS,MET, can1∆::Leu2-MFA1pr::His3,Gus1-3::UraP. HieterYTM 755 Pro3-1ura3∆0, leu2∆0, his3∆1, LYS,MET, can1∆::Leu2-MFA1pr::His3,Pro3-1::UraP. HieterYTM 758 Guk1-7ura3∆0, leu2∆0, his3∆1, LYS,MET, can1∆::Leu2-MFA1pr::His3,Guk1-7::UraP. HieterYTM 766 Ugp1-3ura3∆0, leu2∆0, his3∆1, LYS,MET, can1∆::Leu2-MFA1pr::His3,Ugp1-3::UraP. HieterYTM 938 san1∆ his3∆1, leu2∆0, met15∆0, ura3∆0,san1∆::KanMXOpenBiosystemsCollectionYTM 981 ubr1∆ his3∆1, leu2∆0, met15∆0, ura3∆0,ubr1∆::KanMXOpenBiosystemsCollectionYTM 1183 tda2∆ his3∆1, leu2∆0, met15∆0, ura3∆0,tda2∆::KanMXOpenBiosystemsCollectionYTM 1184 yak1∆ his3∆1, leu2∆0, met15∆0, ura3∆0,yak1∆::KanMXOpenBiosystemsCollectionContinued on next page30Strain ID Alias Genotype SourceYTM 1185 rim15∆ his3∆1, leu2∆0, met15∆0, ura3∆0,rim15∆::KanMXOpenBiosystemsCollectionYTM 1186 gim3∆ his3∆1, leu2∆0, met15∆0, ura3∆0,gim3∆::KanMXOpenBiosystemsCollectionYTM 1187 YOR364W∆ his3∆1, leu2∆0, met15∆0, ura3∆0,YOR364W∆::KanMXOpenBiosystemsCollectionYTM 1290 vhr1∆ his3∆1, leu2∆0, met15∆0, ura3∆0,vhr1∆::KanMXOpenBiosystemsCollectionYTM 1293 sli15∆ his3∆1, leu2∆0, met15∆0, ura3∆0,sli15∆::KanMXOpenBiosystemsCollectionYTM 1294 fau1∆ his3∆1, leu2∆0, met15∆0, ura3∆0,fau1∆::KanMXOpenBiosystemsCollectionYTM 1301 gim5∆ his3∆1, leu2∆0, met15∆0, ura3∆0,gim5∆::KanMXOpenBiosystemsCollectionYTM 1302 gim6∆ his3∆1, leu2∆0, met15∆0, ura3∆0,gim6∆::KanMXOpenBiosystemsCollectionYTM 1304 gim1∆ his3∆1, leu2∆0, met15∆0, ura3∆0,gim1∆::KanMXOpenBiosystemsCollectionYTM 1305 gim4∆ his3∆1, leu2∆0, met15∆0, ura3∆0,gim4∆::KanMXOpenBiosystemsCollectionYTM 1306 gim2∆ his3∆1, leu2∆0, met15∆0, ura3∆0,gim2∆::KanMXOpenBiosystemsCollectionYTM 1356 rpt6-20 his3∆1, leu2∆0, met15∆0, ura3∆0,RPT6::rpt6-20::KanMXBoone tscollectionContinued on next page31Strain ID Alias Genotype SourceYTM 1357 pep4∆prb1∆his3∆1, leu2∆0, met15∆0,ura3∆0, lys2∆0, PRB1::KanMX6,PEP4::His3MX6Fang et al.2015YTM 1489 Gim3-TAP his3∆1, leu2∆0, met15∆0, ura3∆0,gim3::TAP::His3MXOpenBiosystemsCollectionYTM 1677 tcp1-1 his3∆1, leu2∆0, met15∆0, ura3∆0,TCP1::tcp1-1-KanMAX6 P. HieterYTM 1678 tcp4-1 his3∆1, leu2∆0, met15∆0, ura3∆0,CCT4::cct4-1-KanMAX6 P. HieterYTM 1692 bra7∆ his3∆1, leu2∆0, met15∆0, ura3∆0,bra7∆::KanMXOpenBiosystemsCollectionYTM 1693 asp1∆ his3∆1, leu2∆0, met15∆0, ura3∆0,asp1∆::KanMXOpenBiosystemsCollectionYTM 1694 mal11∆ his3∆1, leu2∆0, met15∆0, ura3∆0,mal11∆::KanMXOpenBiosystemsCollectionYTM 1695 mup3∆ his3∆1, leu2∆0, met15∆0, ura3∆0,mup3∆::KanMXOpenBiosystemsCollectionYTM 1696 pol4∆ his3∆1, leu2∆0, met15∆0, ura3∆0,pol4∆::KanMXOpenBiosystemsCollectionYTM 1697 pph22∆ his3∆1, leu2∆0, met15∆0, ura3∆0,pph22∆::KanMXOpenBiosystemsCollectionYTM 1855 ubr1∆gim3∆ his3∆1, leu2∆0, met15∆0, ura3∆0,ubr1∆::KanMX, gim3∆::His3MX6 This thesisYTM 1901 Hsp104-mCherry his3∆1, leu2∆0, met15∆0, ura3∆0,Hsp104-mCherry::His3MX6 This thesisYTM 1902 Guk1-7-13mycgim3∆, his3∆1, leu2∆0, LYS, MET15,Guk1-7-13myc::KanMX6::Ura3,CAN1, gim3∆::His3MX6This thesisYTM 1919 Hsp42-mCherry his3∆1, leu2∆0, lys2∆0, ura3∆0,Hsp42-mCherry::KanMX6 This thesis32Table 2.2: Plasmids used in Chapter 2PlasmidID NameAuxo-trophicMarkerPlasmidType SourceBPM 42 pRS316 Ura CEN/ARS RJD CollectionBPM 45 pRS313 His CEN/ARS RJD CollectionBPM 171 PGPD His CEN/ARS F. Khosrow-KhavarBPM 241 PGPD-GFP(S65T) His CEN/ARS F. Khosrow-KhavarBPM 368 PUbr1-Ubr1 Leu CEN/ARS R. HamptonBPM 369 PUbr1-Ubr1(C1220S) Leu CEN/ARS R. HamptonBPM 453 PGPD-Guk1-GFP His CEN/ARS This thesisBPM 457 PGPD-Ugp1-3-GFP His CEN/ARS This thesisBPM 458 PGPD-Guk1-7-GFP His CEN/ARS This thesisBPM 500 PGPD-Gus1-3-GFP His CEN/ARS This thesisBPM 507 PGPD-Pro3-1-GFP His CEN/ARS This thesisBPM 509 PGPD-Guk1(E127K)-GFP His CEN/ARS This thesisBPM 510 PGPD-Guk1(T95A)-GFP His CEN/ARS This thesisBPM 511 PGPD-Guk1(F59H)-GFP His CEN/ARS This thesisBPM 513 PGPD-Guk1(A84T)-GFP His CEN/ARS This thesisBPM 551 PGPD-Gim3 Ura CEN/ARS This thesisBPM 572 PUbr1-Ubr1 Ura CEN/ARS This thesisBPM 659 PGPD-Guk1::His6 His CEN/ARS This thesisBPM 708 PCup1-Deg1-cNLS-GFP Ura CEN/ARS T. SommerBPM 717 PGPD-Guk1-7::His6 His CEN/ARS This thesisBPM 779 PGPD-GFP Ura CEN/ARS This thesisBPM 780 PGPD-Guk1-GFP Ura CEN/ARS This thesisBPM 781 PGPD-Guk1-7-GFP Ura CEN/ARS This thesisBPM 866 pFA6a-mCherry-KanMX6 KanMX CEN/ARS This thesisBPM 894 PGuk1-Guk1-GFP His CEN/ARS This thesisBPM 895 PGuk1-Guk1-7-GFP His CEN/ARS This thesis332.2.2 Stability Effect of Guk1-7 MutationsThe predicted thermodynamic stability changes of mutations in Guk1-7 were com-puted using FoldX (version 3.0). The protein structure of Guk1 was downloadedfrom the Protein Data Bank (PDB accession 1EX7) and was optimized using therepair function of FoldX. Structures corresponding to each of the single point mu-tations and all four point mutants combined were generated. The predicted effect ofmutations on protein structural stability was expressed as the predicted free energychange (∆∆G) and was obtained by subtracting the energy values of the mutantstructures from that of the wild type.2.2.3 Cellular Thermal Shift Assay (CETSA)A 50 mL yeast culture grown at 25◦C was collected at log phase and harvestedby centrifugation. Cells were then lysed with glass beads in 200 µL of nativelysis buffer (20 mM HEPES, pH 7.5, 0.5% NP-40, 200 mM NaCl, 1X proteaseinhibitor mix (Roche), 1 mM 1,10 phenanthroline, 1 mM EDTA). The solublefraction was collected by centrifugation (16,000 g, 10 min, 4◦C) and protein con-centration was determined by the DC Protein Assay (BioRad). Samples were nor-malized to 2 µg/µL and 50 µL aliquots were distributed into PCR strip tubes andrun on a PCRmachine with the following program: 25◦C, 3:00; Gradient 30–50◦C,10:00; 25◦C, 1:00. The soluble fraction was once again collected by centrifugation(16,000 g, 10 min, 4◦C). Equal volumes were resolved by SDS-PAGE. Membraneswere immunoblotted with mouse anti-HIS6 (Ablab, 1:2,500) and secondary anti-bodies (Mandel Scientific, 1:10,000) and then quantified using an Odyssey InfraredImaging System.2.2.4 Solubility AssayYeast cells were grown to log phase at 25◦C and then incubated for 20 min at ei-ther 25◦C or 37◦C. Cells were lysed with glass beads in native lysis buffer (20mM HEPES, pH 7.5, 0.5% NP-40, 200 mM NaCl, 1X protease inhibitor mix, 1mM 1,10 phenanthroline, 1 mM EDTA) and then precleared by centrifugation at2,000 g for 5 min at 4◦C. Sample protein concentrations were measured by the DCProtein Assay (BioRad) and normalized. Samples were further fractionated into34soluble and pellet fractions by centrifugation at 16,000 g for 10 min at 4◦C. Thepellet fractions were then washed twice with lysis buffer. Equal volumes of to-tal cell lysate, soluble, and pellet fractions were resolved by SDS-PAGE. Sampleswere analyzed by mouse anti-GFP (Roche, 1:2,500) and rabbit anti-Pgk1 antibod-ies (Acris Antibodies, 1:10,000) as a loading control.2.2.5 MicroscopyCells were grown in synthetic dropout media lacking histidine to log phase (OD600= 0.8–1.0) at 25◦C and then collected at the indicated time points following incu-bation at 25◦C or 37◦C with our without 100 µg/mL cycloheximide (CHX), asnoted. Samples were fixed in 3.7% formaldehyde for 15 minutes at room tempera-ture and then rinsed in 0.1 M potassium phosphate containing 1 M sorbitol beforebeing permeabilized with 0.1% Triton X-100 for ten minutes. Nuclei staining wasperformed by incubating permeabilized cells in Hoechst 33342 (25 µg/mL) for 10minutes before mounting cells on slides in mounting media (2% N-Propylgallate,80% glycerol, 0.02% sodium azide in 1X PBS). Cells were imaged with a ZeissAxio observer inverted microscope equipped with a 63x oil-immersion objectiveand a digital camera. Images were analyzed with Zeiss Axiovision software.2.2.6 Degradation AssayCells were grown to log phase in synthetic drop out media at 25◦C and cyclohex-imide was added to a final concentration of 100 µg/mL. Cells were then incubatedat either 25◦C or 37◦C, and at the indicated time points cells were collected by cen-trifugation. The cells were then resuspended in modified Laemmli buffer (50 mMTris-HCl, pH 6.8, 2% SDS, 10% glycerol), and lysed with glass beads. Protein con-centration was assessed by the DC Protein Assay (BioRad). Equal amounts of pro-tein were resolved by SDS-PAGE following the addition of 10X 2-mercaptoethanol(20%) and dye to each sample. Immunoblots were performed with a mouse anti-GFP primary antibody (Millipore, 1:2,500) and a rabbit anti-Pgk1 (1:10,000, AcrisAntibodies) as a loading control. Infrared secondary antibodies were used (MandelScientific, 1:10,000) and membranes were scanned and analyzed with an OdysseyInfrared imaging system (LI-COR).352.2.7 Flow CytometryYeast cells were grown in synthetic drop out media to log phase before the additionof 100 µg/mL cycloheximide and incubated at 25◦C or 37◦C as indicated. Sampleswere run on a BD FACSCalibur instrument (BD Biosciences) with a 488 laser andGPF was detected with a 530/30 filter. 50,000 events were collected. Analysis wasperformed with FlowJo (FlowJo Data Analysis Software, LLC). For multi-hourCHX chase experiments median GFP fluorescence values were normalized to thatof the first time point. FACS sorting was performed with a BD Influx instrumentby the UBC Flow facility.2.2.8 GFP PulldownGim3-TAP yeast cells transformed with a control empty vector (BPM 42), PGPD-GFP (BPM 779), PGPD-Guk1-GFP (BPM 780), or PGPD-Guk1-7-GFP (BPM781) were grown to log phase and then lysed with glass beads and native lysisbuffer (20 mM HEPES, pH 7.5, 0.5% NP-40, 200 mM NaCl, 1X protease inhibitormix, 1 mM 1,10 phenanthroline, 1 mM EDTA, 10 mM iodoacetamide). To pull-down GFP-tagged proteins, lysates were incubated for 2 hours at 4◦C with 20 µLGFP-Trap coupled agarose beads (Chromotek). Beads were washed three times inlysis buffer before samples were eluted with 3X SDS sample buffer. Nitrocellulosemembranes were probed with mouse anti-GFP (Roche, 1:2,500), rabbit anti-Pgk1(Acris Antibodies, 1:10,000), rabbit anti-TAP (Fisher, 1:2,500), and mouse anti-ubiquitin (Millipore, 1:2,500) primary antibodies.2.2.9 Proteasome FunctionYeast cultures were grown to saturation in synthetic drop out media overnight at30◦C and then diluted to OD600 = 0.2 and left to grow for 3 hours at 30◦C. 100 µMcopper sulphate was added to the culture and incubated at 30◦C for 4 hours. Aninitial sample was removed and then cycloheximide was added to the culture toa final concentration of 100 µg/mL. Samples were collected at the indicated timepoints. Cells were lysed with glass beads and lysis buffer (1% Tx-100, 0.1% SDS,150 mM NaCl, 5 mM EDTA, 50 mM Tris-HCl, pH 7.5, 1 mM PMSF, 1X proteaseinhibitor mix). Protein concentrations were assessed using the DC Protein Assay36(BioRad) and equal amounts were resolved by SDS-PAGE.2.2.10 Statistical AnalysisUnpaired two tailed Student’s t-tests were used to assess significance of differencesbetween wild type and gim3∆ or ubr1∆ strains. One-way ANOVA with post-hocTukey HSD (honest significant difference) was used to assess significance of dif-ferences between multiple deletion strains.2.3 Results2.3.1 Guk1-7 is Thermally UnstableOur lab previously identified a panel of temperature sensitive alleles of essentialgenes encoding for cytosolic proteins in Saccharomyces cerevisiae [101]. A largefraction of mutant proteins underwent proteasome-mediated degradation when in-cubated at the restrictive temperature of 37◦C, whereas the wild type proteins werestable. While approximately one third of the unstable alleles were found to besubstrates of the E3 ubiquitin ligase Ubr1 [101], the protein quality control path-ways responsible for the proteasomal degradation of the remaining mutant pro-teins are unknown. To screen for other proteins involved in proteasome mediateddegradation of thermosensitive mutant proteins, we sought to establish an assaybased on fluorescence intensity to facilitate the quantification of a model qual-ity control substrate fused to GFP. For this study, we selected the Guk1-7 allelethat contains four missense mutations generated by random PCR-based mutagen-esis [101, 185, 186]. Guk1 is a member of the nucleoside monophosphate kinase(NMP) family and converts GMP to GDP [187]. Similar to other members of theNMP family, Guk1’s structure contains a core, a lid, and a dynamic NMP-bindingdomain [188]. Mutants of Guk1 are defective in mannose chain elongation, havehigher cell wall porosity, and are hypersensitive to larger molecular weight an-tibiotics [189]. We first predicted the structural stability effects of the missensemutations found in Guk1-7 using FoldX (Figure 2.1A) [190]. The predicted freeenergy changes (∆∆G) between the single point mutants and the wild type proteinwere modest, whereas the combined effect of all the mutations found in Guk1-737was much larger (∼7 kcal/mol). While this value is higher than that predicted formissense mutations in transmembrane domains of disease-associated proteins suchas cystic fibrosis transmembrane conductance regulator (CFTR) and rhodopsin(1.5 and 1.9 kcal/mol, respectively), it is in line with those predicted for muta-tions in phenylalanine hydroxylase (PAH) associated with mild or severe forms ofphenylketonuria (5.7 and 14.2 kcal/mol, respectively) [191, 192]. We then com-pared the thermodynamic stability of ectopically expressed wild type Guk1 withGuk1-7 in cellular lysates by a cellular thermal shift assay (CETSA) [193]. Inagreement with its predicted lower stability, Guk1-7 was less stable than the wildtype Guk1 at incubation temperatures above 38◦C (Figure 2.1B). We further exam-ined the solubility of Guk1 and Guk1-7 proteins in cells incubated at the normalgrowth temperature of 25◦C and following a short 20 minute incubation at 37◦C.Guk1 was found predominantly in the soluble form while Guk1-7 was enriched inthe NP-40 insoluble fraction at 25◦C and 37◦C (Figure 2.1C). Together these datasuggest that Guk1-7 is much less stable than the wild type protein and misfoldsforming NP-40 insoluble aggregates.2.3.2 Fluorescence-Based Assay to Assess Protein StabilityTo determine whether the ectopically expressed mutant protein was also degradedwhen fused to GFP, we first examined fluorescence levels by microscopy. Guk1-7-GFP fluorescence was on average 58% lower than that of Guk1-GFP at 25◦C(n = 101, 108), and was nearly undetectable with an average 87% loss of fluores-cence following a two hour incubation at 37◦C in the presence of the translationinhibitor cycloheximide (n = 168) (Figure 2.2A, Figure 2.12A). By contrast, thefluorescence of the wild type Guk1-GFP only slightly decreased by 29% between37◦C and 25◦C (n = 120). To verify that the loss of fluorescence was due to pro-teolysis and not misfolding of GFP, we examined levels of Guk1 and Guk1-7 byWestern blot in a cycloheximide chase assay. While Guk1-GFP levels remainedrelatively unchanged, the level of Guk1-7-GFP decreased by 30% after a four hourincubation at 25◦C, and decreased by 70% after the same period at 37◦C (Fig-ure 2.2B). We then verified whether adding a GFP tag alters Guk1 or Guk1-7 sol-ubility. Three times as much Guk1-7 as Guk1 was found in the NP-40 insoluble38pellet fraction at 25◦C, and this rose to nine times more upon the short incubationat 37◦C (Figure 2.2C). Although Guk1-7-GFP is less insoluble than Guk1-7-His6,presumably due to stabilization conferred by the GFP moiety, the GFP tagged mu-tant was both less soluble and more degraded than the wild type protein, and couldtherefore be employed as a model substrate.39!"!"#$"%&'(%")*+,-."%/"0'&1")23)404'4&567 68 97 98 87787:77:87;77 <(=:<(=:>?! @ .;8+,6?+,;8+,6?+,<(=:<(=:>?#<)/A&01")*52BB)23)35(2%"B/"0/"-C8DEF)6GH8)=/&5I#25)*;:J-KH9!F)7G98)=/&5I#25)*L8J-!D8KF):G7D)=/&5I#25)*L8J-M:;?NF):GOH)=/&5I#25)*L8J-C8DEP)KH9!P)!D8KP)M:;?N)*<(=:>?-F)?G:)=/&5I#25)*O:J-C8DEKH9!M:;?N!D8K<(=:>?<(=:<(=:>?!"#$G)*+,- 67 6:G; 66GH 6?GO 9;G8 9OG8 9HGH 87 Q%)*=R&-6?6?Q%)*=R&-6?6?6?6?Figure 2.1: Guk1-7 is thermally unstable. (A) Ribbon structure of Guk1(PDB 1EX7). Positions of the four missense mutations and predicted ∆∆Gvalues are indicated. Loss of fluorescence measured by flow cytometry after atwo hour incubation at 37◦C with cycloheximide is indicated in brackets. (B)Cellular thermal shift assay of Guk1 and Guk1-7 fused to a six histidine tag inlysates derived from cells grown at 25◦C. One representative anti-His WesternBlot is shown. The graph represents the means and standard deviations ofGuk1 levels from three independent experiments. (C) Guk1 and Guk1-7 fusedto a six histidine tag were expressed in cells grown at 25◦C or shifted to 37◦Cfor 20 min. Total cell lysate (T), soluble (S), and pellet fractions (P) wereimmunoblotted with anti-His antibodies.40!"!"#$%&!'()* +,)*- ' . - ' ./01234 5 !'()*+,)*'()*+,)*%1#$%1#$6,+7+898-7:;<7,898$7,;$$7=898'7';<$7.898,7.;!>??>@A28B#CDE(-(-(-(-%&! F0>G/3@CH*'()*+,)*'()*+,)*%1#$6%&!%1#$6,6%&!A>2">IJ?K8@LM>8N8*FO'()* +,)*- ' . - ' .%1#$6,6%&!8N8*FO %1#$6%&!8N8*FOA28B#CDE(-(-!>2G>P@D">80Q8JPJ@JD?%1#$6%&!8'()*%1#$6%&!8+,)*%1#$6,6%&!8'()*%1#$6,6%&!8+,)*#4JR>8B/0123E- ' .$--,((-'(-Figure 2.2: Misfolded Guk1-7 is degraded at the non-permissive temperature.41Figure 2.2: (Previous page) Misfolded Guk1-7 is degraded at the non-permissive temperature. (A) Wild type cells expressing ectopic Guk1-GFPor Guk1-7-GFP were grown at 25◦C and then incubated in the presence of thetranslation inhibitor cycloheximide (CHX) at 25◦C or 37◦C for 2 hours priorto fixation and imaging. Scale bar represents 5 µm. (B) Cycloheximide chaseassay. Wild type cells expressing ectopic Guk1-GFP or Guk1-7-GFP wereincubated with CHX for 4 hours at 25◦C or 37◦C and samples were collectedat the indicated time points. Guk1-GFP and Guk1-7-GFP was immunoblottedwith anti-GFP antibodies and a representative blot is shown. GFP levels werenormalized to Pgk1 levels and shown in the graph below with results repre-senting the means and standard deviations of three independent experiments.(C) Guk1-GFP and Guk1-7-GFP were ectopically expressed in wild type cellsgrown at 25◦C or shifted to 37◦C for 20 min. Total cell lysate (T), soluble (S),and pellet fractions (P) were immunoblotted with anti-GFP antibodies. Theratio of the pellet fraction to total cell lysate is noted and represents the meanand standard deviation of three independent experiments.In order to use Guk1-7-GFP as a model substrate to screen for factors impor-tant in maintaining cytosolic protein homeostasis, we established a flow cytometryassay to monitor protein stability. Cultures were incubated at 25◦C or 37◦C inthe presence of cycloheximide for two hours and then the GFP fluorescence inten-sity from single cells was measured by flow cytometry. The relative difference inmedian intensity values between 37◦C and 25◦C was used as a measure of pro-tein stability. In a wild type strain at 25◦C Guk1-7-GFP fluorescence intensity islower than that of the wild type allele, suggesting that the model substrate is in-herently unstable even at lower temperatures. After shifting the cells to 37◦C inthe presence of CHX for two hours, GFP intensity levels remained nearly constantfor Guk1-GFP (5% loss) but decreased for Guk1-7-GFP (60% loss; Figure 2.3A).These data are consistent with our previous fluorescence microscopy and CHX-chase observations (Figure 2.2A, 2.2B). The data obtained from flow cytometrymeasurements was comparable to that acquired using traditional Western blottingtechniques (Figure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igure 2.3: Guk1-7 degradation is proteasome dependent.43Figure 2.3: (Previous page) Guk1-7 degradation is proteasome dependent.(A) Flow cytometry profiles of wild type cells expressing Guk1-GFP orGuk1-7-GFP were incubated at 25◦C or 37◦C for two hours in the pres-ence of CHX. Fluorescence in cells with the control empty vector (EV) arealso shown. Lines demark median GFP fluorescence values and i denotesthe difference in median intensity values used to measure protein stability.(B) Comparison of quantitation of Guk1-7 levels in a CHX chase assay byWestern blot or flow cytometry. (C) Wild type and rpt6-20 cells expressingGuk1-7-GFP were incubated with CHX at 25◦C or 37◦C and samples wereanalysed by flow cytometry at the indicated time points. The results rep-resent the means and standard deviations of three independent experiments.(D) Guk1-7-GFP expressing wild type or pep4∆prb1∆ cells were incubatedat 25◦C or 37◦C in the presence of CHX and samples were analyzed by flowcytometry at the indicated time points. The results represent the means andstandard deviations of three independent experiments. (E) rpt6-20 cell ex-pressing Guk1-GFP or Guk1-7-GFP were grown at 25◦C and then shiftedto 37◦C for 1 hour prior to their fixation and imaging. Scale bar represents5 µm. (F) Guk1-GFP and Guk1-7-GFP expressing cells were incubated at25◦C and cell lysates were immunoprecipitated using GFP-Trap beads andthen immunoblotted with anti-ubiquitin, anti-GFP, and anti-Pgk1 antibodies.The Guk1 and Guk1-7 constructs used in this report are ectopically expressedfrom the constitutive GPD promoter. To ensure that the overexpression from aplasmid does not influence the stability of our model substrate, we expressed bothGuk1-GFP and Guk1-7-GFP from their endogenous locus and promoters, and per-formed a cycloheximide assay. Consistent with our previous CHX chase observa-tions, Guk1-GFP levels remained relatively constant and Guk1-7-GFP levels de-creased by approximately 30% after four hours at 25◦C and by 60% when incu-bated at 37◦C (Figure 2.12B). Together, this data suggested that the flow cytometryassay was suitable for monitoring protein levels and for screening purposes.2.3.3 Guk1-7 Degradation is Proteasome DependentWe next verified that degradation of the ectopically expressed model GFP-fusionsubstrate was proteasome-dependent. When we assayed levels of Guk1-7 at 37◦Cin the temperature sensitive proteasome mutant rpt6-20 [52], degradation of the44mutant protein was largely stopped (Figure 2.3C). Conversely, no difference inGuk1-7-GFP stability was seen between the wild type strain or a double mutant ofthe two main lysosomal proteases (Figure 2.3D). These results suggest that loss ofGuk1-7-GFP fluorescence is primarily caused by proteasomal degradation. Fluo-rescence microscopy revealed that Guk1-GFP was evenly distributed with no in-clusions in rpt6-20 cells at both 25◦C and 37◦C, as was the case for Guk1-7-GFPat 25◦C (n = 126, 104, 118, respectively) (Figure 2.3E). Guk1-7-GFP inclusionswere detected in 69% of the cells incubated at 37◦C (n = 120), of which 94%of cells contained a single inclusion and 5% contained two. These data indicatethat non-degraded Guk1-7-GFP was prone to aggregation at the non-permissivetemperature. Finally, we asked whether the difference in protein stability betweenGuk1-GFP and Guk1-7-GFP was also reflected by their respective ubiquitinationlevels. We found that Guk1-7-GFP, but not Guk1-GFP, was ubiquitinated at 25◦C(Figure 2.3F). In this case, we collected cell lysates from cultures incubated at thelower growth temperature, as we encountered issues with our model substrate be-ing mostly lost to the insoluble pellet fraction when cultures were grown at highertemperatures. Together these experiments suggest that misfolded Guk1-7 is tar-geted for degradation by the ubiquitin proteasome system.2.3.4 FACS-Based Screen for Protein Homeostasis FactorsTo identify novel factors involved in targeting proteins destabilized by missensemutations for degradation, we performed a genome-wide screen based on flow cy-tometry using the Guk1-7-GFP allele. A schematic of the screen is depicted in Fig-ure 2.4A. First, we pooled and bulk transformed the yeast non-essential knockoutcollection with a low copy number plasmid containing Guk1-7-GFP (Figure 2.4Ai). Growth prior to and after transformation was limited, to avoid under repre-sentation of slow growing strains. Pooled transformants were grown in selectivemedia at 25◦C and then subjected to an initial FACS presort to obtain a narrowfluorescence range, which reduces cell-to-cell variability of GFP fusion expression(Figure 2.4A ii; compare grey and green profiles for before and after presort, re-spectively). Presorted cells were then incubated at 37◦C in the presence of CHXfor two hours (Figure 2.4A iii) and then sorted again, selecting for cells with GFP45fluorescence in the top 10% range (Figure 2.4A iv). Samples were collected overa short fifteen minute period to minimize shifting of the population during thehandling time. Cells were recovered in selective liquid media and the screen wasrepeated two more times for a total of three rounds of enrichment (Figure 2.4Av). Following the final FACS sorting, cells were collected on solid selective mediaplates (Figure 2.4A vi).We selected 170 colonies, which had been isolated using the FACS screen de-scribed above, for validation using the flow cytometry assay. In approximately twothirds of the colonies tested, Guk1-7-GFP was more stable than in the wild typecells (Figure 2.4B). The yeast knockout collection was created by replacing eachyeast open reading frame with a KanMX module (conferring resistance to the an-tibiotic geneticin) and a unique 20 base pair nucleotide sequence, referred to asa molecular barcode. Universal priming sites located upstream and downstreamof the barcodes are used for PCR amplification of the barcode region. Sequenc-ing or microarray of the resulting amplicon can be used to reveal the identity of thecorresponding yeast deletion strain. We therefore selected fifty colonies at random,spanning the range of Guk1-7-GFP stabilities, and their corresponding gene knock-outs were identified by Sanger sequencing of the unique strain-specific barcodes.From these fifty colonies, we identified fifteen different gene deletions (Table 2.3).Next, to ensure that the phenotype (i.e., stabilization of Guk1-7) was not acquiredduring the screening process, we assessed the stability of Guk1-7-GFP for each ofthe fifteen gene deletions we identified through Sanger sequencing. To do so, weindividually retransformed the Guk1-7-GFP containing plasmid into each knockoutstrain from our pre-pooled knockout collection. Results of this phenotypic valida-tion were considered positive (denoted as P in Table 2.3) if Guk1-7-GFP was atleast 15% more stable in the deletion strain than in wild type cells (Figure 2.12C).Strains that did not meet this criterion were classified as negative (denoted as Nin Table 2.3). Surprisingly, we failed to observe any stabilization of our modelsubstrate in our most frequently identified hits (e.g., tda2∆), which may have beensusceptible to the acquisition of secondary mutations (Table 2.1). Among the val-idated hits, we identified the N-end rule E3 ligase Ubr1, which was previouslyshown to target cytosolic misfolded proteins for degradation [74, 76, 101–104],and the prefoldin chaperone subunit Gim 3 (Table 2.3, Figure 2.12C).46!"#$%&'"(#)**()+(+#,)-"*."/."0 10 200 21003000310230023104$#&5$%"5(*%-$&/(/,67"-8,92:;(&/(!"#$8,92(&/(!"#$8,92:;(&/%&'($!"!"##!$%&'()* )#%$+",-"&-"./!01!!"##!$%&'()* )#%$+",-"&-"()* )#%$+",-"&-"!"##!$%&'01!2"#"-'"3<-$/*+)-6"5(="$*%(>?(.)##".%&)/@ABC(D-"*)-%E(BFGH(I-JK;LBM@ABC(*"#".%(%)D(20NO5"/%&+=(10(>?(*%-$&/*(7=(7$-.)5"(*"P,"/.&/QO*)#$%"(&/5&'&5,$#(.)#)/&"*!"D"$%(%R)(6)-"(-),/5*()+(*%"D*(&&(:(&'# ## ####$$$#8,92:;(&/%)*+,Figure 2.4: FACS-based screen. (A) Schematic of FACS-based screen. (B)Flow cytometry validation of 170 colonies isolated from the FACS screen.Relative loss of fluorescence of Guk1-GFP or Guk1-7-GFP in wild type cellsis noted as a comparison.47Table 2.3: Summary of FACS screen validationStandardNameSystematicNameNumber of TimesBarcode Identified bySanger SequencingResult ofPhenotypicValidation†TDA2 YER071C 22 NYAK1 YJL141C 10 NRIM15 YFL033C 5 PUBR1 YGR184C 2 PYOR364W YOR364W 1 PGIM3 YNL153C 1 PSLI15 YBR156C 1 NFAU1 YER183C 1 NVHR1 YIL056W 1 NBRA7 YER056C 1 PASP1 YDR321W 1 NMAL11 YGR289C 1 NMUP3 YHL036W 1 PPOL4 YCR014C 1 NPPH22 YOL188C 1 N†P: positiveN: negative2.3.5 Ubr1 Stabilizes Guk1 Missense MutantWe identified the E3 ubiquitin ligase Ubr1 in our screen for factors responsible fordegradative protein quality control of misfolded cytosolic proteins destabilized bymissense alleles. In addition to its role as the E3 ligase of the N-end rule pathway,Ubr1 has also been shown to target misfolded cytoplasmic proteins for degrada-tion [74, 101, 102]. In CHX chase experiments, Guk1-7-GFP levels were approxi-mately 10% higher in ubr1∆ cells compared to wild type (P = 0.0008) at two hours48and remained significantly higher after four hours (P = 0.00003) (Figure 2.5A).To confirm that the stability observed was directly caused by the loss of Ubr1, weperformed addback experiments whereby the wild type Ubr1 was expressed from aplasmid in ubr1∆ cells. We observed that Guk1-7-GFP levels were similar betweenthe UBR1 cells containing a control empty vector and ubr1∆ cells with the Ubr1expressing plasmid, confirming that the phenotype observed could be attributed tothe absence of Ubr1 (Figure 2.5B). To further validate our findings, we performedthe same addback experiments but this time included a mutant form of Ubr1, whichcontains a point mutation in the RING domain producing an inactive ligase [74].Guk1-7-GFP levels in the Ubr1 (C1220S) expressing cells were indistinguishablefrom those with a control plasmid, lending further support to Ubr1 having a directrole in controlling Guk1-7 stability (Figure 2.5C). We next assessed the impor-tance of Ubr1 on a second unstable allele of Guk1 (T290G, hereinafter referredto as Guk1-11) that contained a single missense mutation. This mutant was gen-erated by site directed mutagenesis and was selected based on its instability, as atwo hour incubation at 37◦C in the presence of cycloheximide typically resultedin approximately 60% loss of fluorescence. Consistent with our previous results,an absence of UBR1 led to a significantly reduced clearance of this second modelsubstrate (P = 0.00035) (Figure 2.5D). The relative fluorescence of Guk1-GFP wasnot significantly different between wild type and ubr1∆ cells (Figure 2.13A). Theseresults indicate Ubr1 participates in the clearance of these model misfolded sub-strates, although other factors are also involved.Ubr1 has been shown to act in concert with the nuclear E3 ligase San1 to targetmisfolded cytoplasmic proteins for degradation [74, 101]. To test whether Ubr1also acts with San1 in the degradation of Guk1-7-GFP, we performed flow cy-tometry experiments in the single ubr1∆ and san1∆ deletion strains along with adouble ubr1∆ san1∆ deletion. Guk1-7-GFP was not markedly more stable uponthe deletion of SAN1, although levels were slightly higher in ubr1∆ san1∆ cells incomparison to ubr1∆ cells (Figure 2.13B). These results indicate that San1 doesnot play a major role in the turnover of Guk1-7. To confirm that our assay was ca-pable of detecting an effect with San1, we ran the same assay using the previouslycharacterized Ubr1 and San1 substrate Pro3-1 [101]. In this case we were able toobserve a significant stabilization of Pro3-1 in san1∆ cells, which was even more49! "#!"#$!"#$%!"#$!&'$%%&'$!"#$%%&'$!&'$%((" )#(*((+,-*.'.+/0!0123145671!89!:4:5:6;<:=1!>?8,2@AB((#(*((" )+,-*.'.+/0!0123145671!89!:4:5:6;<:=1!>?8,2@A(%&'$(C!DE(!"#$(C!DE(!"#$(C!%&'$+,-*.'.+/0!0123145671!89!:4:5:6;(!"#$!C!DE(!"#$C!%&'$(!"#$(C!FG2*!>%*""(HA(#(*((( " )BBBBBBB BBBBBBBB$(I(*I("I(J1;65:K1!+,-*.**.+/0!9;,821@31431!:4514@:5L%&'$ !"#$<:=1!>?8,2@ABBBB!BB!BB!B B!BFigure 2.5: Ubr1 promotes Guk1-7-GFP degradation. (A) Wild type andubr1∆ cells expressing Guk1-7-GFP were incubated with CHX at 25◦C or37◦C and samples were analysed by flow cytometry at the indicated timepoints. The results represent the means and standard deviations of three inde-pendent experiments. P values were calculated with an unpaired Student’s ttest, *, ** and *** denote P< 0.05, 0.005, and 0.0005, respectively. (B)UBR1and ubr1∆ cells expressing Guk1-7-GFP along with an empty vector (EV)control or UBR1 were incubated at 37◦C and samples were collected at the in-dicated time points for flow cytometry analysis. Results represent the meansand standard deviations of three independent experiments. P values were cal-culated with a one-way ANOVA and post-hoc Tukey HSD to assess signifi-cance, ** denotes P < 0.005. (C) ubr1∆ cells coexpressing Guk1-7-GFP andan empty vector control or either UBR1 or UBR1 (C1220S) were incubatedat 37◦C with CHX and samples were collected at the indicated time points.(D) Wild type or ubr1∆ cells expressing Guk1 (T290G) fused to GFP wereincubated with CHX at 37◦C for two hours before being analyzed by flow cy-tometry. The results represent the relative fluorescence intensities from threeindependent experiments (with standard deviations). P values were calculatedwith a one-way ANOVA and post-hoc Tukey HSD to assess significance, ***denotes P < 0.0005.50pronounced in the double ubr1∆ san1∆ strain (Figure 2.13C). Together, these datasuggest that San1 does not play a role alongside Ubr1 in targeting Guk1-7-GFP fordegradation, indicating that other E3 ligases may be involved in the proteasome-mediated degradation of this substrate.2.3.6 Gim3 Impairs Guk1-7-GFP DegradationPrefoldin is a hetero-oligomeric protein complex composed of six subunits rangingin size from 14–23 kDa [41]. Conserved in archaea and eukaryotes, but absent inprokaryotes, the prefoldin hexamer forms a “jellyfish-like” structure with N- andC-terminal coiled-coil regions of each subunit forming “tentacles” that emanatefrom a central region [42]. Misfolded substrates are transferred for folding fromprefoldin to the TRiC/CCT chaperonin in an ATP-independent manner through di-rect binding of the two chaperone complexes [41]. In addition to its role in aidingnascent proteins, such as actin and tubulin, to attain their functional conformations,the prefoldin chaperone complex has been shown to prevent huntingtin and alpha-synuclein aggregate formation [43, 44]. Having identified the prefoldin subunitGim3 in our screen, we decided to further examine its potential role in degradativeprotein quality control. In CHX chase experiments, Guk1-7-GFP levels were ap-proximately 25% higher in the gim3∆ strain compared to wild type (P = 0.0057)(Figure 2.6A). To ensure that the stabilization was specifically caused by the ab-sence of Gim3, we expressed in gim3∆ cells the wild-type GIM3 from a plasmid,which rescued the degradation of the model substrate (Figure 2.6B). While degra-dation of Guk1-7-GFP is not fully inhibited in gim3∆ cells, levels are markedlyhigher than in the wild type strain, indicating that Gim3 is required for the normalturnover of our model substrate. We next wished to see if Gim3 works together withUbr1. In this case we preferred a model substrate that is misfolded as the result ofa single point mutation (Guk1-11) to eliminate or minimize potential confoundingfactors caused by multiple destabilizing mutations. The double ubr1∆ gim3∆ strainshowed increased Guk1-11 stability compared to single deletion strains, howeverthe substrate was still degraded by over 50% (Figure 2.6C). This data indicates thatpotentially other E3 ligases or chaperones are required for complete proteolysisto occur. In addition, this would suggest that Ubr1 and Gim3 work partially in51parallel or in independent pathways to target the assessed misfolded substrate fordegradation.!"#!"#$ %&'$! "! #! $! %! ! "! #! $! %!"!! "!!&' #& "& $ %#!&( $")*+",-.//+0"! # &!#''!('"!!-10",(,-.//*23*456+*7897:4:5:6;<:=*7>?812@A%&'$7#'BC%&'$7$(BC!"#$(#'BC!"#$($(BC<:=*7>=:4A)*+",-.//*23*456+*7897:4:5:6; "!!(''!#'!! #! &! %!!"#$%&'$$%<:=*7>=:4AD27>0)6AE=F5G7H*3582-./-10",(,-./!"#$%&'$#'$('!#'$('!'!#'!"'!'!"!!IIIIII:4F15-.//+0"J/K7-./LK7-./J/K7-./LK7MNO*;65:P*7-10",(,-./7.;182*@3*43*7J45*4@:5G!"#$Q7EH%&'$Q7EH%&'$)(!"#$$#"!4@I7II7I!"#$*+,-!"#$./0-%&'$*+,-%&'$./0-&#!O*;65:P*7-10","",-./7.;182*@3*43*7J45*4@:5GI7II7II7II7IIIFigure 2.6: Absence of Gim3 reduces Guk1-7 turnover.52Figure 2.6: (Previous page) Absence of Gim3 reduces Guk1-7 turnover. (A)Wild type and gim3∆ cells expressing Guk1-7-GFP were incubated with CHXat 25◦C or 37◦C and samples were analysed by flow cytometry at the indicatedtime points. The results represent the means and standard deviations of threeindependent experiments and the asterix denotes significance of P< 0.05. (B)Gim3 addback experiment. GIM3 or gim3∆ cells expressing Guk1-7-GFP andeither an empty vector (EV) control or GIM3. The results represent the meansand standard deviations of three independent experiments of the relative flu-orescence intensity after a two hour CHX incubation at 37◦C. P values werecalculated with a one-way ANOVA and post-hoc Tukey HSD to assess signifi-cance, ** denotes P< 0.005. (C) Wild type, ubr1∆, gim3∆, and ubr1∆ gim3∆cells expressing Guk1 (T290G) fused to GFP were incubated at 37◦C withCHX for two hours and then analysed by flow cytometry. P values were cal-culated with a one-way ANOVA and Holm multiple comparison to assesssignificance, * and ** denote P < 0.05 and 0.01, respectively. (D) Protea-some activity assay. Gim3 or gim3∆ cells expressing Deg1-GFP under theCup1 promoter were incubated at 30◦C in the presence of CHX and sampleswere collected at the indicated time points. Deg1-GFP was immunoblottedwith an anti-GFP antibody. The results represent the mean and standard de-viation of three independent experiments. (E) GIM3 or gim3∆ cells express-ing Guk1-7-GFP, GFP alone, or a control empty vector were grown at 25◦C.Guk1-7-GFP was immunoprecipitated with GFP-Trap beads and eluted sam-ples were immunoblotted with anti-ubiquitin and anti-GFP antibodies.The TRiC/CCT chaperonin cooperates with prefoldin in folding a number ofcellular proteins and has been shown to interact with proteasome subunits, suggest-ing that it may be involved in proteasome maturation [194]. Therefore, one possi-bility is that the stabilizing effect of Gim3 on Guk1-7 could be indirect, a result ofdecreased proteasome function. We tested for compromised proteasome functionin the gim3∆ strain using the constitutive Deg1-GFP proteasome substrate [184].We found that there was no significant difference in the degradation of Deg1-GFPin gim3∆ cells compared to the wild type strain at all time points tested, with theexception of the thirty minute sample (P = 0.77, P = 0.22, P = 0.04, P = 0.07 forthe 10, 20, 30, and 60 minute time points, respectively) (Figure 2.6D). Hence, thereduced turnover of misfolded protein observed in gim3∆ cells is unlikely causedby an impaired proteasome.53We next determined whether an absence of GIM3 could affect ubiquitinationof our misfolded model substrate. GFP tagged Guk1-7 was pulled down fromcells grown at 25◦C, where it remained mostly soluble, and then ubiquitin levelswere detected by immunoblotting. After normalizing the quantity of ubiquitin tothat of eluted Guk1-7, ubiquitination levels were essentially unchanged, with un-der 5% less ubiquitinated model substrate in gim3∆ cells compared to wild type(Figure 2.6E). This experiment indicates that the absence of Gim3 did not impairubiquitination of our model substrate, in agreement with Gim3 functioning inde-pendently of Ubr1.2.3.7 Gim3 Facilitates the Clearance of Insoluble Guk1 andMaintains Guk1-7 SolubilityWe next sought to evaluate the impact an absence of Gim3 has on Guk1-7 localiza-tion. Fluorescence microscopy performed on wild type and gim3∆ strains showedthat while there was no difference in wild type Guk1-GFP localization betweenthe two strains, Guk1-7-GFP formed cytoplasmic puncta in 93% of gim3∆ cellswhen incubated at 37◦C, and additional faint and diffuse cytoplasmic GFP wasalso visible (n = 200) (Figure 2.7A). Cells contained on average 1.5 puncta, whichwere typically located next to the nucleus. In contrast, only sixteen percent ofwild type cells contained Guk1-7-GFP puncta (n = 200). These puncta were nolonger present when we expressed Gim3 from a plasmid in gim3∆ cells and thediffuse cytoplasmic GFP signal was also not present, similar to that observed inGIM3 cells (Figure 2.7B). We decided to examine this phenomenon more closelyby performing a time course microscopy experiment incubating cells at 37◦C, butin the absence of the translation inhibitor CHX which we had been using up to thispoint and which may interfere with aggregate formation [195]. Within five minutesnumerous Guk1-7-GFP containing puncta were detected within the cytoplasm ofGIM3 cells with those in the gim3∆ strain being slightly delayed and visible af-ter 10 minutes (Figure 2.7C). While puntca remained visible in both strains after45 minutes at 37◦C, those in the gim3∆ strain appeared to coalesce slower than inGIM3 (after 15–20 minutes in GIM3 compared to 20–25 minutes in gim3∆) andremained visibly brighter. Some diffuse cytoplasmic Guk1-7-GFP signal was alsopresent in gim3∆ cells up to 25 minutes after shifting to the increased growth tem-54perature, but was only present in the first 10 minutes for GIM3 cells. While thenumber of puncta did not differ between Gim3 containing or deleted cells, the in-tensity of the gim3∆ puncta remained brighter for longer. These results suggest thatGim3may play a role in maintaining Guk1-7-GFP solubility at higher temperaturesto facilitate substrate degradation.These Guk1-7-GFP puncta observed in the time course experiment in bothGIM3 and gim3∆ cells are reminiscent of the Q-bodies described by Frydmanand colleagues [136]. To determine whether this is indeed the case, we examinedGuk1-GFP and Guk1-7-GFP colocalization in GIM3 cells with two cytosolic ag-gregate markers: Hsp104 and Hsp42. Hsp104 is an aggregate-specific chaperonethat has a diffuse cytoplasmic and nuclear localization pattern at 25◦C but formspuncta when incubated at 37◦C (Figure 2.8A, enlarged Figure 2.8C). Hsp104-mCherry and Guk1-7-GFP puncta colocalized in 100% of cells at 37◦C (n = 100).We then examined Guk1-7-GFP colocalization with the small heat shock proteinHsp42, which is required for peripheral aggregate formation during physiologicalheat stress [137]. As with Hsp104, Hsp42-mCherry formed puncta when incu-bated at 37◦C, but not at 25◦C (Figure 2.8B, enlarged Figure 2.8D). Guk1-7-GFPcolocalized in all Hsp42-mCherry puncta. However, in 34% of the cells exam-ined (n = 100), we find an average of 1.4 Guk1-7-GFP puncta per cell that do notcolocalize with Hsp42. Overall, Hsp42-free Guk1-7-GFP puncta represented 9%of all puncta observed in the one hundred cells examined. Together, the Hsp104and Hsp42 colocalization data suggest that at 37◦C Guk1-7-GFP forms cytosolicinclusions similar to Q-bodies.55! "!"#$%&'()*"+,-./ ,-./ ,-./ ,-./01./ 01./ 01./ 01./!"#$ %&'$ !"#$ %&'$*2345*"+ *234515*"+,-./ 01./ 01./ 01./!"#$ %&'$6789 6789 6789 ! !"#$#$%&'():%;<%*234515*"+=>/:%;<%#? - 4? 4- ,? ,- 0? @-!"#$*234515*"+=>/#$%&'()%&'$*234515*"+=>/#$%&'():%;<%:%;<%ABC7D)701$/Figure 2.7: Gim3 facilitates clearance of insoluble Guk1-7.56Figure 2.7: (Previous page) Gim3 facilitates clearance of insoluble Guk1-7.(A) GIM3 or gim3∆ cells expressing Guk1-7-GFP were grown at 25◦C andthen incubated at either 25◦C or 37◦C for 2 hours in the presence of CHX be-fore fixation and imaging. (B)GIM3 and gim3∆ cells expressing Guk1-7-GFPalong with an empty vector control or GIM3 were incubated at 25◦C or 37◦Cfor two hours in the presence of CHX before fixation and imaging. (C)GIM3 or gim3∆ cells expressing Guk1-7-GFP were incubated at 25◦C andthen shifted to 37◦C. Samples were collected at the indicated time points andthen fixed before imaging. For all images, the scale bar represents 5 µm anddotted lines demark cell boundaries.To verify the importance of Gim3 in maintaining Guk1-7 solubility, we exam-ined the sedimentation of the mutated protein after centrifugation. Twice as muchGuk1-7 is found in the NP-40 insoluble pellet fraction at 25◦C in gim3∆ cellscompared to wild type GIM3 cells (Figure 2.9A). There was also more Guk1-7 inthe pellet of gim3∆ than wild type cells after incubating cells at 37◦C. We thenperformed immunoprecipitation experiments to test whether Gim3 could directlyinteract with Guk1-7 as a potential mechanism for maintaining Guk1-7 solubil-ity. From cell extracts incubated at 25◦C, Guk1-7-GFP can pull down a TAP-tagged form of Gim3 whereas no interaction was detected between Gim3-TAPand Guk1-GFP (Figure 2.9B). We verified this interaction in an independent ex-periment (Figure 2.14A). Once again, a lower temperature was used for pulldownexperiments to avoid losing Guk1-7 in the insoluble pellet fraction. These resultssuggest that Gim3 could maintain Guk1-7 in a more soluble state through physicalinteraction, potentially acting as a holdase. Holdases are a type of molecular chap-erone that bind to misfolded proteins in an ATP-independent manner to preventprotein aggregation, but they do not directly refold their substrates [47]. Consis-tent with these findings, we tested the viability of the guk1-7 strain over a range oftemperatures (25◦C to 37◦C), in the presence or absence of GIM3, and found thatin both cases viability largely decreased between 32◦C and 33◦C with no growth attemperatures of 34◦C or above (Figure 2.14B). These results indicate that, whereasdegradation of poorly soluble Guk1-7 was delayed, temperature-dependent lethal-ity is not rescued in gim3∆ cells.57!!"# $%&'()* +&,-&./0$)123456#(&,,789:#;<:#89:#;<:#.=>25./0.=>25<5./0"!"# $%&'()* +&,-&./0$)14856#(&,,789:#;<:#89:#;<:#.=>25./0.=>25<5./0./0$)123456#(&,,7 +&,-&#$./0$)14856#(&,,7 +&,-&Figure 2.8: Guk1-7-GFP puncta colocalize with Q-body markers. (A) Cellswith Hsp104 endogenously tagged with mCherry and ectopically expressingGuk1-GFP or Guk1-7-GFP were grown at 25◦C and then incubated at 25◦C or37◦C for 30 minutes before fixation and imaging. Scale bar represents 5 µm.(B) Hsp42-mCherry cells ectopically expressing Guk1-GFP or Guk1-7-GFPwere grown at 25◦C prior to incubation at 25◦C or 37◦C for 30 minutes. Cellswere then fixed before imaging. The scale bar represents 5 µm. (C) Enlargedimages from cells collected as in A. Scale bar represents 2.5 µm. (D) Enlargedimages from cells collected as in B. Scale bar represents 2.5 µm.58! "!"#"$%$"#&'()#*$%$+*#"',)#+$%$!*#+'*-#*$%$!+#"'./00/12 3 .+(45+(45,*45,*45!"#$%&'$678!9*96:.678!96:.;<=1>$?/@1AB6C<,92D.E.F$6:.GF$2D.E.F$6:.GF$6:.2D.6:..H8!CI=71JB$K8LMN(-(-(-(-JB$K8LMN,*(-,*(-(-Figure 2.9: Gim3 helps maintain Guk1-7 solubility. (A) GIM3 or gim3∆Guk1-7-GFP expressing cells were grown at 25◦C or shifted to 37◦C for 20min. The ratio of the pellet fraction to total cell lysate is noted and repre-sents the mean and standard deviation of three independent experiments. (B)Guk1-7-GFP was immunoprecipitated from Gim3-TAP expressing cells incu-bated at 25◦C and then immunoblotted with anti-TAP, anti-GFP, or anti-Pgk1antibodies.2.3.8 Gim3 Has a General Effect Towards Thermally DestabilizedProteinsTo see if the effect of Gim3 on Guk1-7 solubility was specific to this prefoldin sub-unit, or common to all prefoldin subunits, we performed fluorescence microscopywith the other prefoldin mutant strains. Guk1-7-GFP puncta were visible in allof the prefoldin deletions, albeit to varying degrees, suggesting that they all playa role in Guk1-7 solubility (Figure 2.10A). Fifty-six per cent of gim1∆ cells con-tained puncta, whereas only 26% of gim4∆ did (n = 50, each). While the numberof puncta per cell only differed slightly between prefolin strains, either faint or nocytoplasmic Guk1-7-GFP was visible in gim2∆, gim4∆, gim5∆, and gim6∆ strainswhile markedly present in gim1∆ and gim3∆ cells. To better quantify the effect, wemeasured Guk1-7 levels by flow cytometry and found that only deletions of GIM1and GIM3, and to a lesser extent GIM5, retarded the degradation of the model sub-strate (Figure 2.10B). Not surprisingly, gim1∆ and gim3∆were the only strains that,in addition to puncta, also had a diffuse cytoplasmic Guk1-7-GFP signal visible byfluorescence microscopy. All together, these results suggest that while deletion of59individual members of the prefoldin complex impacted degradation of the modelsubstrate, some (i.e., Gim1 and Gim3) may play a more important role.!"!"#$!"#%!"#&!"#'!"#(!"#)!"#$%&'()$%&'(*$%+,-.&%$/0123450 1.))6.7671(.-0)80980:;<$;:#'(*$%+%'=,>0-"0?@ABCD0-"0?@AD//&%/#9+'(Figure 2.11: Model for stabilization of temperature sensitive alleles by Gim3.(A) Proposed model for how Gim3 promotes degradation of temperature sen-sitive alleles destabilized by missense mutations.As the structure of the prefoldin complex has no evidence for a nucleotidebinding site and therefore lacks ATP-regulated functionality, it is tempting to spec-ulate that prefoldin may act as a holdase [42]. Interestingly, only the mutated andnot wild type Guk1 requires Gim3 to remain soluble. In addition, mutations thataffect chaperonin components also impaired Guk1-7 degradation. These results65indicate that in addition to maintaining misfolded proteins soluble, prefoldin alsohanded them to the chaperonin for refolding. More work will be required to clearlydemonstrate whether Gim3 acts as a holdase to prevent protein aggregation to en-hance substrate accessibility for ubiquitin-proteasome mediated degradation. Sev-eral other chaperone or co-chaperone proteins have been shown to be importantfor promoting the degradation of cytosolic misfolded proteins in yeast. Sse1 wasshown to help mediate the degradation of the tumor suppressor VHL and is requiredfor the recognition of misfolded proteins by Ubr1 [50, 74]. The Ydj1 J-domaincontaining Hsp40 mediates both the degradation of ER proteins with exposed mis-folded cytosolic domains and the Rsp5 mediated degradation of cytosolic proteinsafter heat shock [52, 53]. In contrast Sis1, another J-domain containing Hsp40,was shown to be important for the relocalization of cytosolic misfolded proteinsto the nucleus [75]. Fes1, an Hsp70 nucleotide exchange factor, was also shownto be important for the degradation of cytosolic misfolded proteins and does soby interacting with the misfolded proteins bound to Hsp70 and triggering their re-lease [54, 167, 203]. By demonstrating a role for Gim3 in substrate solubility, ourwork adds to a growing body of evidence suggesting that prefoldin is important forpreventing potentially toxic protein aggregation [44, 204, 205]. In addition to ourtemperature sensitive alleles, prefoldin has been shown to inhibit human amyloid-beta fibrillation and prevents aggregation of huntingtin [44, 204]. This underscoresthe potential importance the prefoldin chaperone complex has in maintaining pro-tein homeostasis.662.5 Supplemental Data!"!"#$%&%!'()*++,)-"./!"#$%&,0. 12,3456)7 !"#$%& 0. 12 38&6)7!"#$%&,0.,,3456)7 !"#$%&,0.,,38&6)7!"#$%&'(9 4 :94559&5$99!"#$%&%!'( 456)!"#$%!'(,456)!"#$%!'(,8&6)!"#$%&%!'(,8&6)20;*,3<-"=>7(*=?*./@A*,-B,0.0/0@+!"#$%&%!'()*++,)-"./'"#()*+,$-./ &'01 #%2$#'+"-=*>?*.?*,0./*.>0/C,3@D"D7989E999F9E999G9E999!"#$456).H$9I!"#$8&6).H$49!"#$%&456).H$9$!"#$%&8&6).H$FIFigure 2:12: Guk1-7-GFP flow cytometry.67Figure 2.12: (Previous page) Guk1-7-GFP flow cytometry. (A) Box plot ofquantification for fluorescence microscopy images in Figure 2.2A. Correctedtotal cell fluorescence was calculated by subtracting the mean fluorescence ofbackground readings from the integrated density. n = 108, 120, 101, and 168for Guk1 25◦C, Guk1 37◦C, Guk1-7 25◦C, and Guk1-7 37◦C, respectively.(B) Wild type cells expressing Guk1-GFP or Guk1-7-GFP on a plasmid andexpressed from their endogenous promoters were incubated at 25◦C or 37◦Cwith CHX. Samples were collected at the indicated time points and analyzedby flow cytometry. (C) Flow cytometry validation experiments for the deletionstrains identified by barcode sequencing. Cells expressing Guk1-7-GFP wereincubated with CHX at 25◦C or 37◦C for two hours prior to flow cytometryanalysis. Note that expression of YOR364W and RIM15 from a plasmid (i.e.,add back experiments) failed to rescue the phenotype indicating that an addi-tional mutation may have caused stabilization of the model substrate. Dele-tions of UBR1 and GIM3 were further analyzed in this work but not MUP3and BRA7.68! "#!"#$ %&'$!"#$%&'"()*+,-)./(0#*12"34"54"(&5%"53&%6 789:899895353!"#$()*$%&'$()*$!"#$+,-$%&'$+,-$!"#$%&'"(#133(10()*+,-;-)./(0#*12"34"54"(,8998<989=(==(=53=(==(==(!"#$%&'"(#133(10(/21>-,-)./(0#*12"34"54"(!"#$()*$%&'$()*$!"#$+,-$%&'$+,-$,8998<989=(==(==(==(==(==(=Figure 2.13: Ubr1 does not act with San1 in the degradation of Guk1-7-GFP.(A) Guk1-GFP was expressed in wild type or ubr1∆ cells and incubated withcycloheximide for 2 hours at 25◦C or 37◦C prior to performing flow cytom-etry. The results represent the relative fluorescence intensities and standarddeviations from three independent experiments. Statistical significance wastested using an unpaired two tailed Student’s t-test. (B) Guk1-7-GFP wasexpressed in wild type, ubr1∆, san1∆, and ubr1∆ san1∆ cells and incubatedwith cycloheximide for 2 hours at 25◦C or 37◦C prior to performing flow cy-tometry. The results represent the average and standard deviations from threeindependent experiments. Statistical significance was tested using a one-wayANOVA and a Tukey HSD post-hoc test. *, **, and ns denote P < 0.05,P < 0.01, and not significant, respectively. (C) Pro3-1-GFP expressing cellswere grown and treated as in B. Samples were analysed using a one-wayANOVA followed by Tukey’s post hoc test, ** denotes P < 0.01.69!"!"#$%&'()*((+,-./012.01././012.3(45)*(645)*(3(45)*(645&'(7859,:;<6&)$,-=1=)*()$,-=)*()*(>?#%@5ABC%D8)!?0=&'(2.EF 0/EF 0-EF 02EF00EF 0GEF 0.EF 0HEF 01EF!"#$!%&'()*$!%&'!"#$!%&'+,()*$!%&'+,Figure 2.14: Guk1-7-GFP Gim3 interaction and viability assay.70Figure 2.14: (Previous page) Guk1-7-GFP Gim3 interaction and viability as-say. (A) Guk1-7-GFP was immunoprecipitated from wild type or Gim3-TAPexpressing cells incubated at 25◦C and then immunoblotted with anti-TAP,anti-GFP, or anti-Pgk1 antibodies. (B) Viability assay. Wild type, gim3∆,guk1-7, or double guk1-7, gim3∆ cells were streaked on rich media plates andincubated for two days at the indicated temperatures.71Chapter 3Recurrent BackgroundMutations inWHI2 AlterProteostasis and ImpairDegradation of CytosolicMisfolded Proteins inSaccharomyces cerevisiae3.1 IntroductionProtein homeostasis (proteostasis) is maintained by an extensive protein qualitycontrol network that promotes and mediates protein folding by molecular chap-erones and prevents the accumulation of misfolded proteins by targeting them fordegradation via the ubiquitin proteasome system or autophagy [1]. The proteostaticbalance can be challenged by exposure to a range of intrinsic or extrinsic stressors,which require the cell to mount an adequate response, most notably by regulatingthe expression of protein quality control network elements in a concerted manner.Inadequate management of misfolded proteins can have deleterious consequences,72such as aggregation, which is characteristic of some neurodegenerative diseasesthat include Alzheimer’s, Parkinson’s, and ageing [11].Proteasomal degradation of misfolded cytosolic proteins is mediated by sev-eral quality control E3 ubiquitin ligases, which typically work in concert withother chaperone proteins to recognize their substrates [178, 203]. For instance,Hsp110 Sse1, which acts as a nucleotide exchange factor, was shown to promoteubiquitination by the Ubr1 E3 ligase in yeast [74]. As well, we proposed that theYdj1 Hsp40 co-chaperone acts as a substrate adaptor for the Rsp5 E3 ligase uponacute heat stress [52]. In other cases, chaperone proteins are also required to pro-mote proteolysis. The Hsp40 co-chaperone Sis1 for example, is necessary for thetranslocation of misfolded cytosolic proteins to the nucleus where most protea-somes reside [140]. We also recently showed that the yeast prefoldin subunit Gim3is required to promote proteolysis of cytosolic proteins misfolded due to missensemutations by preventing their aggregation [206]. Therefore, although chaperoneproteins primarily promote polypeptide folding and assembly, they may also playa key role in the clearance of misfolded proteins. Understandably, the relation-ship between the folding and degradation machineries is complex. For example,the structurally related chaperone regulatory proteins Bag1 and Bag2, respectivelypromote and inhibit the degradation of cytosolic misfolded proteins by the CHIPE3 quality control ligase [85, 86, 207]. Therefore, a major challenge is to un-derstand how changes in the intricate protein quality control network can perturbproteostasis, for instance by shifting the balance between folding and proteolysis.Temperature sensitive alleles of essential genes in S. cerevisiae are invaluablemodel substrates that can be employed to characterize components of the proteinquality control machinery [70, 101, 135, 136, 146, 206]. We previously identi-fied the E3 ubiquitin ligase Ubr1 from a genetic screen for factors involved indegradative protein quality control of Guk1-7, a thermally unstable mutant alleleof the guanylate kinase Guk1 [206]. Ubr1 activity alone, however, was not suffi-cient to account for the bulk of substrate degradation. Therefore, we performeda targeted flow cytometry based screen using a panel of E3 mutant strains. Us-ing this approach, we identified a surprising number of yeast strains with impaireddegradation. However, following whole genome sequencing we identified numer-ous secondary mutations in the stress response geneWHI2, which were responsible73for the impaired proteolysis of the misfolded model substrate. We linked this phe-notype to a deficiency of the Msn2/Msn4 transcription factor response that alteredthe cell’s capacity to adeptly degrade cytosolic misfolded proteins.3.2 Methods3.2.1 Yeast Strains, Media, and Growth ConditionsThe S. cerevisiae strains used in this study are listed in Table 3.1 and Table 3.2.Yeast strains were cultured in synthetic media with 2% dextrose (lacking the appro-priate amino acids for plasmid selection) or YPD (1% yeast extract, 2% peptone,2% dextrose) and grown at 25◦C with shaking unless indicated otherwise. Whennot specified otherwise, cultures in log phase were obtained by diluting overnightsaturated cultures grown at 25◦C to an OD600 = 0.2 and grown for 4–6 hours untillog phase OD600 = 0.8–1.0 was reached.Table 3.1: Yeast strains used in Chapter 3Strain ID Alias Genotype SourceYTM 408 BY4741 his3∆1, leu2∆0, ura3∆0, met15∆0OpenBiosystemsCollectionYTM 409 BY4742 his3∆1, leu2∆0, ura3∆0, lys2∆0OpenBiosystemsCollectionYTM 445 ssa1-45 his3∆11, leu2∆3, ura3∆52, trp1∆1 T. MayorYTM 639 rsp5-1 his3∆1, leu2∆, ura3∆0, met15∆0,RSP5::rsp5-1-KanMX T. MayorYTM 660 ydj1∆ his3∆1, leu2∆, ura3∆0, ydj1∆::KanMX T. MayorYTM 1867 asi1∆ Tetrad 3a his3∆1, leu2∆0, ura3∆0, MET15,LYS2, asi1∆::KanMX4, whi2-1 This thesisYTM 1868 asi1∆ Tetrad 3b his3∆1, leu2∆0, ura3∆0, met15∆0,lys2∆0, asi1∆::KanMX4, WHI2 This thesisYTM 1869 ASI1 Tetrad 3c his3∆1, leu2∆0, ura3∆0,MET15, LYS2, whi2-1 This thesisContinued on next page74Strain ID Alias Genotype SourceYTM 1870 ASI1 Tetrad 3d his3∆1, leu2∆0, ura3∆0,met15∆0, lys2∆0, WHI2 This thesisYTM 1857 ASI1 Tetrad 1a his3∆1, leu2∆0, ura3∆0, MET15,lys2∆0, whi2-1 This thesisYTM 1856 ASI1 Tetrad 4c his3∆1, leu2∆0, ura3∆0, MET15,lys2∆0, WHI2 This thesisYTM 1871 asi1∆ Tetrad 3a/ BY4741his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0, met15∆0/MET15,LYS2/lys2∆0, asi1∆::KanMX4/ASI1,whi2-1/WHI2This thesisYTM 1872 asi1∆ Tetrad 3a/ asi1∆his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0,met15∆0/MET15, LYS2/lys2∆0,asi1∆::KanMX4/asi1∆::KanMX4,whi2-1/whi2-1This thesisYTM 1873 asi1∆ Tetrad 3a/ das1∆his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0, met15∆0/MET15,LYS2/lys2∆0, asi1∆::KanMX4/ASI1,DAS1/das1∆::KanMX4, whi2-1/whi2-2This thesisYTM 1874 asi1∆ Tetrad 3a/ fap1∆his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0, met15∆0/MET15,LYS2/lys2∆0, asi1∆::KanMX4/ASI1,FAP1/fap1∆::KanMX4, whi2-1/whi2-3This thesisYTM 1875 asi1∆ Tetrad 3a/ hrt3∆his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0, met15∆0/MET15,LYS2/lys2∆0, asi1∆::KanMX4/ASI1,HRT3/hrt3∆::KanMX4, whi2-1/whi2-4This thesisYTM 1876 asi1∆ Tetrad 3a/ hul5∆his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0, met15∆0/MET15,LYS2/lys2∆0, asi1∆::KanMX4/ASI1,HUL5/hul5∆::KanMX4, whi2-1/whi2-5This thesisYTM 1877 asi1∆ Tetrad 3a/ ufd2∆his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0, met15∆0/MET15,LYS2/lys2∆0, asi1∆::KanMX4/ASI1,UFD2/ufd2∆::KanMX4, whi2-1/whi2-6This thesisContinued on next page75Strain ID Alias Genotype SourceYTM 1878 asi1∆ Tetrad 3a/ ufd4∆his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0, met15∆0/MET15,LYS2/lys2∆0, asi1∆::KanMX4/ASI1,UFD4/ufd4∆::KanMX4, whi2/whi2-7This thesisYTM 1879 asi1∆ Tetrad 3b/ BY4741his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0, met15∆0/MET15,LYS2/lys2∆0, asi1∆::KanMX4/ASI1,WHI2/WHI2This thesisYTM 1880 asi1∆ Tetrad 3b/ asi1∆his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0,met15∆0/MET15, LYS2/lys2∆0,asi1∆::KanMX4/asi1∆::KanMX4,WHI2/whi2-1This thesisYTM 1881 asi1∆ Tetrad 3b/ das1∆is3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0, met15∆0/MET15,LYS2/lys2∆0, asi1∆::KanMX4/ASI1,DAS1/das1∆::KanMX4, WHI2/whi2-2This thesisYTM 1882 asi1∆ Tetrad 3b/ fap1∆his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0, met15∆0/MET15,LYS2/lys2∆0, asi1∆::KanMX4/ASI1,FAP1/fap1∆::KanMX4, WHI2/whi2-3This thesisYTM 1883 asi1∆ Tetrad 3b/ hrt3∆his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0, met15∆0/MET15,LYS2/lys2∆0, asi1∆::KanMX4/ASI1,HRT3/hrt3∆::KanMX4, WHI2/whi2-4This thesisYTM 1884 asi1∆ Tetrad 3b/ hul5∆his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0, met15∆0/MET15,LYS2/lys2∆0, asi1∆::KanMX4/ASI1,HUL5/hul5∆::KanMX4, WHI2/whi2-5This thesisYTM 1885 asi1∆ Tetrad 3b/ ufd2∆his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0, met15∆0/MET15,LYS2/lys2∆0, asi1∆::KanMX4/ASI1,UFD2/ufd2∆::KanMX4, WHI2/whi2-6This thesisYTM 1886 asi1∆ Tetrad 3b/ ufd4∆his3∆1/his3∆1, leu2∆0/leu2∆0,ura3∆0/ura3∆0, met15∆0/MET15,LYS2/lys2∆0, asi1∆::KanMX4/ASI1,UFD4/ufd4∆::KanMX4, WHI2/whi2-7This thesisContinued on next page76Strain ID Alias Genotype SourceYTM 1744 msn2∆ his3∆1, leu2∆0, ura3∆0, met15∆0,msn2∆::KanMX4OpenBiosystemsCollectionYTM 1745 msn4∆ his3∆1, leu2∆0, ura3∆0, met15∆0,msn4∆::KanMX4OpenBiosystemsCollectionYTM 1691 whi2∆ his3∆1, leu2∆0, ura3∆0, met15∆0,whi2∆::KanMX4OpenBiosystemsCollectionYTM 1690 glo4∆ his3∆1, leu2∆0, ura3∆0, met15∆0,glo4∆::KanMX4OpenBiosystemsCollection3.2.2 PlasmidsPlasmids used in this study are listed in Table 3.3. Guk1-GFP (BPM453), Guk1-7-GFP (BPM458), and Guk1-7-His6 (BPM717) expressed from the GPD1 promoterin pRS313 were generated in a previous study [206]; Guk1-7-GFP was subclonedwith ApaI and SacI sites in pRS315 to generate BPM609 and with XhoI and SacIIsites in pRS316 to generate BPM781. To generate the E3 ligase addback plasmids(BPM748, ASI1; BPM749, DAS1; BPM750, FAP1; BPM751, HRT3; BPM752,HUL5; BPM753, UFD2; BPM754, UFD4), the open reading frames and approx-imately 500 bp of both endogenous 5’ and 3’UTR was PCR amplified from ge-nomic DNA (BY4741) and inserted in pRS316 using XhoI and XmaI sites for allbut HUL5 where SacII and XhoI sites were used. The WHI2 (BPM863, BPM914)addback plasmids were generated as for the E3 ligases except ligated into pRS315or pRS316 using SacI and XmaI sites, respectively.3.2.3 Flow CytometryCells in log phase were treated with 100 µg/mL cycloheximide and incubated ateither 25◦C or 37◦C as indicated. GFP fluorescence was measured for 50,000 cellsusing a FACSCalibur flow cytometer. Median GFP fluorescence values were ob-tained using FlowJo software. For chase experiments, percentage remaining values77Table 3.2: E3 ligase collection used for screeningSystematic Standard Well Systematic Standard WellName Name No. Name Name No.YMR258C ROY1 A1 YHL010C ETP1 D1YOL013C HRD1 A2 YLR368W MDM30 D2YOL054W PSH1 A3 YER116C SLX8 D3YML068W ITT1 A4 YLR352W n/a D4YDR049W VMS1 A5 YAL002W VPS8 D5YDR131C n/a A6 YDR360W TFB3 D6YDR143C SAN1 A7 YLR024C UBR2 D7YHR115C DMA1 A8 YMR119W ASI1 D8YKL010C UFD4 A9 YNL008C ASI3 D9YKL034W TUL1 A10 YGL003C CDH1 D10YJL149W DAS1 A11 YMR247C RKR1 D11YLR247C IRC20 A12 YNL023C FAP1 D12YNL230C ELA1 B1 YJL157C FAR1 E1YKR017C HEL1 B2 YDL013W SLX5 E2YDR265W PEX10 B3 YBR203W COS111 E3YDR306C n/a B4 YKL059C MPE1 E4YDR313C PIB1 B5 YER068W NOT4 E5YIL001W n/a B6 YMR026C PEX12 E6YCR066W RAD18 B7 YJL210W PEX2 E7YJR036C HUL4 B8 YOR191W ULS1 E8YDL190C UFD2 B9 YDR255C RMD5 E9YBR062C n/a B10 YGL131C SNT2 E10YNL116W DMA2 B11 YLR005W SSL1 E11YBR280C SAF1 B12 YDR103W STE5 E12YIL030C DOA10 C1 YDR266C HEL2 F1YBR114W RAD16 C2 YOL138C RTC1 F2YDR457W TOM1 C3 YBR158W AMN1 F3YGL141W HUL5 C4 YJR052W RAD7 F4YGR184C UBR1 C5 YDR132C n/a F5YDL074C BRE1 C6 YLR108C n/a F6YLR224W n/a C7 YMR080C NAM7 F7YLR097C HRT3 C8 YPL046C ELC1 F8YNL311C SKP2 C9 YJR090C GRR1 F9YDR219C MFB1 C10 YJL204C RCY1 F10YLR427W MAG2 C11YOR080W DIA2 C1278Table 3.3: Plasmids used in Chapter 3PlasmidID NameAuxotrophicMarkerPlasmidType SourceBPM 42 pRS316 Ura CEN/ARS RJD CollectionBPM 45 pRS313 His CEN/ARS RJD CollectionBPM 49 pRS315 Leu CEN/ARS RJD CollectionBPM 390 PYDJ1-YDJ1 Ura CEN/ARS E. CraigBPM 453 PGPD-Guk1-GFP His CEN/ARS T. MayorBPM 458 PGPD-Guk1-7-GFP His CEN/ARS T. MayorBPM 559 PSSA1-SSA1 Ura CEN/ARS T. MayorBPM 573 PRSP5-RSP5 Ura CEN/ARS T. MayorBPM 575 PRSP5-RSP5(C777A) Ura CEN/ARS T. MayorBPM 609 PGPD-Guk1-7-GFP Leu CEN/ARS T. MayorBPM 708 PCUP1-Deg1-GFP Ura CEN/ARS T. MayorBPM 718 PGPD-Guk1-7-GFP Ura CEN/ARS T. MayorBPM 748 PASI1-ASI1 Leu CEN/ARS This thesisBPM 749 PDAS1-DAS1 Leu CEN/ARS This thesisBPM 750 PFAP1-FAP1 Leu CEN/ARS This thesisBPM 751 PHRT3-HRT3 Leu CEN/ARS This thesisBPM 752 PHUL5-HUL5 Leu CEN/ARS This thesisBPM 753 PUFD2-UFD2 Leu CEN/ARS This thesisBPM 754 PUFD4-UFD4 Leu CEN/ARS This thesisBPM 914 PWHI2-WHI2 Ura CEN/ARS This thesis79were calculated by normalizing the median GFP fluorescence intensity values foreach time point to the initial t = 0 measurement. To calculate the relative loss of flu-orescence for single time-point measurements, the difference of GFP fluorescencevalues for samples incubated at 25◦C and 37◦C was normalized to that of the 25◦Csample. To perform multiple strain comparisons, the relative loss of fluorescencevalues (as calculated above) for each strain was normalized to that of the wild typeBY4741 strain.3.2.4 SequencingWhole-genome sequencing and library preparation was performed at the NextGenSequencing facility at the Biodiversity Research Centre of the University of BritishColumbia. Yeast cells were grown overnight to saturation in YPD at 25◦C andgenomic DNA was extracted using standard protocols [208]. Barcoded librariesfor each strain were created according to Illumina protocols (Illumina 2011, allrights reserved) and 100 bp paired end fragments were sequenced by pooling allsix libraries and run on a single lane of an Illumina HiSeq2000. The short-readaligner BWAwas used to map sequence reads to the yeast reference genome S288Cversion R64 (Saccharomyces Genome Database, SDG) [209]. Single-nucleotidevariants (SNVs) were identified using the SAMtools toolbox and then each SNVwas annotated with a custom-made Perl script using gene data downloaded fromSDG on January 21, 2014 [210]. IGV viewer was used to visually inspect readalignments in the regions of candidate SNVs [211, 212].3.2.5 WHI2 Plate AssayYeast cultures were grown overnight at 25◦C in 5 mL YPD to OD600 = 1–2 andthen diluted to OD600 = 0.2 in 5 mL YPD and left to grow for 2 hours at 25◦C.1 mL was kept as an untreated control and the remaining 4 mL of culture wastreated with 200 mM acetic acid for 4 hours at 25◦C. Treated and untreated cultureswere serially diluted fivefold in 1X PBS and plated on solid media. Plates wereincubated at 30◦C for 2 days before being imaged with a Gel Doc XR+ System(Bio-Rad).803.2.6 Turnover AssayCells transformed with a Deg1-GFP containing plasmid were grown to saturationovernight at 30◦C, diluted to OD600 = 0.2 and then incubated for 3 hours at 30◦C.Deg1-GFP expression was induced for 4 hours at 30◦C with 100 µM copper sul-phate and then 100 µg/mL cycloheximide was added with samples collected at theindicated time points. Cells were lysed with glass beads in lysis buffer (50 mMTris-HCl, pH 7.5, 1% Tx-100, 0.1% SDS, 150 mM NaCl, 5 mM EDTA, 1 mMPMSF, 1X protease inhibitor mix (Roche)). Protein concentrations were measuredusing the DC Protein Assay (Bio-Rad) and normalized prior to resolving equalvolumes by SDS-PAGE. Membranes were immunoblotted with mouse anti-GFP(Roche, 1:2,500) and rabbit anti-Pgk1 (Acris Antibodies, 1:10,000) primary anti-bodies and secondary antibodies (Mandel Scientific, 1:10,000). Membranes werescanned and analyzed with an Odyssey Infrared imaging system (LI-COR).3.2.7 Solubility AssayCells expressing Guk1-7-GFP in log phase were incubated at either 25◦C or 37◦Cfor 20 minutes. Cells were lysed with glass beads in native lysis buffer (20 mMHEPES, pH 7.5, 0.5% NP-40, 200 mM NaCl, 1X protease inhibitor mix (Roche),1 mM 1,10 phenanthroline, 1 mM EDTA) and centrifuged at 2,000 g for 5 minutesat 4◦C. Protein concentrations were determined using the DC Protein Assay (Bio-Rad) and normalized to 0.5 µg/µL. Samples were then fractionated into solubleand pellet fractions by centrifuging at 16,000 g for 10 minutes at 4◦C. The pelletfraction was washed twice with native lysis buffer prior to being resuspended in 1XSDS buffer (50 mM Tris-HCl, pH 6.8, 2% SDS, 3% glycerol). Equal volumes oftotal cell lysate, soluble, and pellet fractions were resolved by SDS-PAGE. Mem-branes were immunoblotted with mouse anti-GFP (Roche, 1:2,500) and secondaryantibodies (Mandel Scientific, 1:10,000).3.2.8 Guk1-7-GFP UbiquitinationCells expressing ectopic Guk1-7-GFP, Guk1-GFP, or a control empty vector(pRS313), were grown to log phase and then lysed with glass beads in nativelysis buffer (20 mM HEPES, pH 7.5, 0.5% NP-40, 200 mM NaCl, 1X protease81inhibitor mix (Roche), 1 mM 1,10 phenanthroline, 1 mM EDTA, 10 mM iodoac-etamide). GFP-tagged Guk1-7 was pulled down with GFP-Trap coupled agarosebeads (Chromotek; 10 µL per 3 mg of lysate) for 2 hours at 4◦C. Beads werewashed three times in lysis buffer before samples were eluted with 3X SDS buffer.Equal volumes of samples were resolved by SDS-PAGE. Membranes were im-munoblotted with mouse anti-GFP (Roche, 1:2,500), rabbit anti-Pgk1 (Acris Anti-bodies, 1:10,000), and mouse anti-ubiquitin (Millipore, 1:2,500) primary antibod-ies and secondary antibodies (Mandel Scientific, 1:10,000).3.2.9 Cellular Thermal Shift Assay (CETSA)Cells expressing Guk1-7-His6 were grown to log phase and then lysed with glassbeads in 200 µL native lysis buffer. The soluble fraction was collected by spinningat 16,000 g for 10 minutes at 4◦C on a benchtop centrifuge. Protein concentrationwas determined by the DC Protein Assay (Bio-Rad) and samples were normalizedto 2 µg/mL in native lysis buffer and 50 µL aliquots were distributed into PCR striptubes. Samples were heated using a CETSA PCR Program (25◦C, 3:00; 30–50◦Cgradient, 10:00; 25◦C, 1:00) on a thermocycler. The resulting soluble fraction wascollected by centrifugation at 16,000 g for 10 minutes at 4◦C and one third volumeof 3X SDS buffer was added to samples prior to resolving equal volumes by SDS-PAGE. Membranes were immunoblotted with a mouse anti-His primary antibody(Ablab, 1:2,500) and a secondary antibody (Mandel Scientific, 1:10,000).3.2.10 Statistical AnalysisData are presented as mean± SD unless otherwise stated. Comparisons were madeusing the two-tailed Student’s t-test and differences were considered significant at ap-value of< 0.05. When indicated, multiple strains were compared with a one-wayANOVA and post-hoc Tukey HSD to assess significance.823.3 Results3.3.1 Multiple Strains From the Yeast Knockout Collection DisplayImpaired ProteostasisTo monitor the stability of the Guk1-7 mutant by flow cytometry in yeast cells,we previously generated a C-terminal GFP fusion protein ectopically expressedfrom the constitutive GPD promoter [206]. As we reported, Guk1-7-GFP lev-els are ∼50% and ∼85% lower after incubating cells at 37◦C in the presence ofthe translation inhibitor cycloheximide for two and four hours, respectively (Fig-ure 3.1A). To identify another E3 ubiquitin ligase responsible for the degradation ofthe Guk1-7-GFP model substrate, we screened a collection of 70 non-essential E3ligase deletion strains that were individually transformed with a CEN/ARS plasmidencoding the Guk1-7-GFP fusion. Cultures were grown at 25◦C and then dividedand incubated in the presence of CHX for two hours at 25◦C and 37◦C beforeperforming flow cytometry analysis (Figure 3.1B). For each deletion strain, therelative difference in median GFP fluorescence intensities from samples incubatedat 25◦C and 37◦C was normalized to that of the wild type strain, to calculate a rel-ative loss of Guk1-7-GFP fluorescence (Figure 3.1B). The collection was screenedtwice and strains that had a relative loss of Guk1-7-GFP fluorescence value of 0.75or lower in at least one of the two rounds were selected for further validation. Atotal of 20 strains met this criterion and were further analysed by flow cytometryin three independent experiments (Figure 3.1C). In agreement with our previousfindings, deleting UBR1 led to a 25% lower averaged loss of Guk1-7-GFP fluores-cence compared to that of wild type cells [206]. Surprisingly, we identified twelveE3 ligase deletion strains with a greater impairment in Guk1-7-GFP degradationthan that observed in ubr1∆ cells. Of these, seven strains had an averaged rela-tive loss of Guk1-7-GFP fluorescence value of 0.5 or lower. These results indicatethat an unusually high number of strains from our yeast knockout collection havea reduced capacity to eliminate misfolded cytosolic proteins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igure 3.2: Guk1-7-GFP degradation in E3 ligase deletion strains. Cyclohex-imide chase assay. Wild type or the corresponding E3 ligase deletion strainexpressing ectopic Guk1-7-GFP were incubated in the presence of CHX ateither 25◦C or 37◦C for four hours and samples were analysed by flow cy-tometry at the indicated time points. Results represent the mean and standarddeviation of three independent experiments. P values were calculated witha two-tailed unpaired Student’s t-test (*, **, ***, and ns denote p < 0.05,0.01, 0.005, and not significant, respectively). a) das1∆, b) fap1∆, c) hrt3∆,d) hul5∆, e) ufd2∆, f) ufd4∆.87! "!"!!!"#!$"!!!"#$%&'(%&'$%&'(%&'$%&!"#$)*++++# $!"!!!"#!$"!!(")$%&'(*&+$%&'(*&+$%&(")$)*++++,-./%&'(012/%&'(012/%&,-./!"!!!"#!$"!!)*++++% &!"!!!"#!$"!!,345%&'(0675%&'(0675%&,345)*++++3(!8%&'(6*%8%&'(6*%8%&3(!8!"!!!"#!$"!!)*++++!"!!!"#!$"!!3(!9%&'(6*%9%&'(6*%9%&3(!9)*++++'%(!)*+%,(-..,-&,/012343/56,&(0-7%.#%8#%Figure 3.3: Guk1-7-GFP stability is not a direct effect of E3 ligase deletion.Wild type and E3 ligase deletion strains expressing Guk1-7-GFP along withan empty vector (EV) control or corresponding E3 gene under its endogenouspromoter and terminator were incubated with CHX at 25◦C and 37◦C for twohours prior to flow cytometry analysis. Results represent three independentexperiments. P values were calculated with a one-way ANOVA and post-hoc Tukey HSD to assess significance (** and ns denotes p < 0.01 and notsignificant, respectively). a) das1∆ b) fap1∆ c) hrt3∆ d) hul5∆ e) ufd2∆ f)ufd4∆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igure 3.4: Mutations inWHI2 segregate with the Guk1-7-GFP stability phe-notype.89Figure 3.4: (Previous page) Mutations in WHI2 segregate with theGuk1-7-GFP stability phenotype. a) Backcross and phenotypic segrega-tion. The MATa asi1∆ strain was backcrossed with the wild type MATalphaBY4742 to produce sets of tetrads. Growth was assessed by culturing cells onYPD and 2:2 KanMX deletion marker segregation was observed by spottingonto YPD+G418 plates. This segregation pattern was compared to tetrads ex-pressing Guk1-7-GFP and analyzed by flow cytometry following incubationwith CHX at 25◦C and 37◦C for 2 hours. Results represent three indepen-dent experiments and p values were calculated with a one-way ANOVA andpost-hoc Tukey HSD to assess significance, ** denotes p < 0.01. b) CHXchase assay. Tetrad c and d, produced from the asi1∆ backcross, expressingGuk1-7-GFP were incubated with CHX at 25◦C and 37◦C four hours. Sam-ples were analysed by flow cytometry at the indicated time points. The resultsrepresent the mean and standard deviation of three independent experiments.P values were calculated with a two-tailed unpaired Student’s t-test (*, **,and ns denote p < 0.05, 0.01, and not significant, respectively). c) Comple-mentation test. MATa E3 ligase deletion strains were mated with MATalphawild type BY4742 and asi1∆ cells from tetrad a and b. The resulting diploidsexpressing Guk1-7-GFP were incubated with CHX for two hours at 25◦C and37◦C and analysed by flow cytometry. d) Whole genome sequencing of wildtype, asi1∆, and four asi1∆ backcross tetrad strains revealed a single base pairdeletion in the coding sequence ofWHI2 co-segregates with the Guk1-7-GFPstability phenotype. Arrow head denotes nucleotide base deleted in the asi1∆strain and derivatives.Next, we performed a complementation test to determine whether the sec-ondary mutations responsible for the stability phenotype observed in the sevenE3 ligase deletion strains are in the same locus, or different loci. Heterozygousdiploids were produced by mating a wild type strain (BY4742) and each of theseven E3 ligase deletions to two haploids, derived from the asi1∆ backcross shown:one, contained the secondary mutation (tetrad a) and the other, did not (tetradb). Guk1-7-GFP levels were indistinguishable between heterozygous diploids pro-duced from mating tetrad a and BY4741 and diploids produced from crossingtetrad b with any of the E3 deletion mutants, or the wild type BY4741 (Fig-ure 3.4C). Conversely, all heterozygous diploids derived from mating E3 ligasedeletions with tetrad a (harbouring the secondary mutation) demonstrated increased90!!"#$%&'"(#)**()+(,-./010,23(+#-)4"*5"65" /788798!"#$ !"#$ %&'$ %&'$"8879/78()*+:(;<,-.+:(;<,-.+: /012== ==6*!"#$%&'"(#)**()+(,-./010,23(+#-)4"*5"65"Figure 3.5: das1∆ tetrad analysis and WHI2 addback. a) Analysis of onetetrad obtained from backcrossing the MATa das1∆ strain with the wild typeMATalpha BY4742. Tetrad spores expressing Guk1-7-GFP were incubatedwith CHX at 25◦C and 37◦C for two hours prior to flow cytometry analysis.b) Wild type and glo4∆ cells co-expressing Guk1-7-GFP and an empty con-trol vector (EV) orWHI2 were incubated with CHX at 25◦C and 37◦C for twohours and then analysed by flow cytometry. Results represent three indepen-dent experiments and p values were calculated with a one-way ANOVA andpost-hoc Tukey HSD to assess significance (** and ns denote p < 0.01 andnot significant, respectively).Guk1-7-GFP stability, thereby indicating that they belong to the same complemen-tation group, and suggests that the secondary mutations present in each strain arein the same gene.To identify the locus containing the secondary mutation, we performed whole-genome sequencing on four haploid tetrads and their parental wild type and asi1∆strains. Secondary mutations in the genes GLO4 and WHI2, which are approxi-mately 3000 bp apart on chromosome fifteen, co-segregated with the strains har-bouring the increased Guk1-7-GFP stability phenotype. Interestingly, secondarymutations in the general stress response geneWHI2 have been identified previouslyin yeast knockout collections and genome evolution studies [214–216]. We identi-fied a single nucleotide deletion in the coding sequence of WHI2. This mutation,hereinafter referred to as whi2-sc1, produces a frameshift introducing a prematurestop codon and likely results in a loss ofWHI2 function (Figure 3.4D). In contrast,the coding sequence of GLO4, a mitochondrial glyoxalase, contained a single mis-sense mutation.913.3.3 Guk1-7-GFP Degradation is Impaired Owing to SecondaryMutations inWHI2To determine whether the mutation in WHI2 caused the observed stabilization, weexpressed the wild type ORF from a plasmid in cells derived from the backcross.Whereas the addition of an empty vector did not rescue the phenotype, addition ofWHI2 re-established normal Guk1-7-GFP degradation levels (Figure 3.6A). More-over, we observed a similar impairment in the degradation of the Guk1-7-GFPmodel substrate in whi2∆ cells, that could be rescued by the expression of WHI2(Figure 3.6B). Intriguingly, we found that glo4∆ cells had a similar reduction inGuk1-7-GFP degradation (Figure 3.5B). Subsequent Sanger sequencing of a PCRproduct amplified from the WHI2 locus of glo4∆ cells identified two point muta-tions that produce a premature stop codon. These results suggest that the effectobserved in glo4∆ cells is attributed to a loss of WHI2 function, not of GLO4, andthat loss of WHI2 function is sufficient to strongly impair degradation of a mis-folded cytosolic model substrate.92!"#$%&#%' ())*+,*-.%#/.*-./'01*2*3401*2*!"#$%&'$*2*34%&'$*2*!"#$()'**2*34()'**2*!"#$+()**2*34+()**2*!"#$,(-**2*34,(-**2*!"#$&./0*2*34&./0*2*!"#$&123*2*34&123*2*!"#$1,+$*2*341,+$*2*!"#$1,+4*2*341,+4*2*!"#$!"#5%6&#/7%6899*8:*;<=>?@?;AB*:6<8$%9.%%".%))CD>C)5!"#$2*345%&'$2*345%&'$2*!"#$"9EE EE5%6&#/7%*6899*8:*;<=>?@?;AB*:6<8$%9.%".% >C))CD)67#*52*3467#*5.2*3467#*5.2*!"#$67#**'2*3467#*5'2*!"#$"9 "9EE EE"9F#$&/" G<.6%8#/'%,<#&#/8"BH%"8#IJ/.F#$%"K#H B$%'/.#%'*B$8#%/"8/-* LMD1NO )CMP9:)*** Q>;N1 )CM&123 L>@/"9OR*L>Q'%6;R*L(>;N1 )CPD,(-* >P(@;N- )CD)829* >S>M;N- )CM@&./0 >>)QON- )CPQ()'* M@Q'%6O )CD>+()* DQ>'%61 )CPL1,+4 MS(ON; )CD>+;(* S)@;N1 )C@D1,+$ ((@/"91 )CD)$F(LQB0(@SFR*;(@P4PLM>PLMP@DPSLSMQ((M*TS>U(>)>QM*TPU>)(LP*TQUS)Figure 3.6: Absence of WHI2 leads to Guk1-7-GFP stability.93Figure 3.6: (Previous page) Absence ofWHI2 leads to Guk1-7-GFP stability.a) ASI1 tetrads c and d co-expressing Guk1-7-GFP and a control empty vec-tor (EV) or WHI2 were incubated with CHX at 25◦C and 37◦C for two hoursand then analysed by flow cytometry. Results represent three independentexperiments and p values were calculated with a one-way ANOVA and post-hoc Tukey HSD to assess significance (ns and ** denote not significant andp< 0.01, respectively). b) Wild type and whi2∆ cells expressing Guk1-7-GFPalong with an empty vector (EV) control or WHI2 were treated and analysedas in a. c) WHI2 function assay. Diluted overnight cultures of wild type orE3 ligase deletion strains expressing either an empty control vector or WHI2were treated with 200 mM acetic acid for four hours prior to serial dilutionand spotting onto synthetic drop out plates. Images were taken after two daysof growth at 30◦C. d) Mutations in WHI2 were identified by Sanger sequenc-ing of a PCR amplicon spanning 100 bp up and downstream of the start andstop codons. For each strain, the mutations identified are as listed and, thepredicted protein length is depicted in red. Black boxes denote the C-terminalmismatch extensions.We next sought to confirm that WHI2 was also mutated in the other E3 lig-ase mutant strains in which Guk1-7-GFP degradation was impaired. Mutations inWHI2 sensitize cells to exposure to acetic acid, which lends itself to a convenientassay for Whi2 function [214]. The whi2∆ and all seven E3 ligase deletion strainswere sensitive to acetic acid treatment (Figure 3.6C). ExpressingWHI2 from a plas-mid under its endogenous promoter restored cell viability in all strains, confirmingdata from the complementation test suggesting that all strains contain secondarymutations in the same locus (Figure 3.4C). We proceeded to sequence the entireWHI2 gene, including approximately one hundred base pairs upstream and down-stream of the start and stop codons, in all twenty of the top E3 ligase deletionstrains from our screen. Whereas the ubr1∆ and six other strains had no apparentmutations, we identifiedWHI2 mutations in a total of eleven strains (Figure 3.6D).These consist of: dma1∆, elc1∆, etp1∆, cos111∆, and all seven strains that failedin addback experiments: asi1∆, das1∆, fap1∆, hrt3∆, hul5∆, ufd2∆, and ufd4∆.In two cases, pex2∆ and tom1∆, we were unable to obtain unambiguous sequenc-ing results after two independent genomic extractions and sequencing runs. Of theWHI2 mutations identified, nine are predicted to produce truncated proteins result-94ing from the introduction of a premature stop codon. Three of the nine also containadditional C-terminal extensions (ranging from 4 to 31 amino acids in length) asthe result of frameshift mutations. The mutations are relatively evenly dispersedalong the length of the protein with the exception of a mutation free region, sev-enty amino acids in length, found approximately three quarters of the way into theprotein. Whereas the different WHI2 mutations led to varying degrees of impairedGuk1-7-GFP degradation, we did not see a clear correlation between the severityof the Guk1-7-GFP stabilisation phenotype and the predicted Whi2 length in thesestrains (Figure 3.6D).3.3.4 Reduced Proteostasic Capacity inWHI2Mutants is Linked toMsn2Exposure to stressors such as heat, oxidative or osmotic shock, and nutrient star-vation results in the transcriptional activation of approximately 200 genes in yeast[146]. Activation of this general stress response is mediated by binding of thepartially-redundant zinc finger transcription factors Msn2 and Msn4 to STREs inthe promoters of stress-response genes [146, 217]. Under non-stress conditions,Msn2 is sequestered in the cytoplasm and upon exposure to stress, Msn2 translo-cates to the nucleus [218, 219]. To determine whether Guk1-7-GFP stability inWHI2 mutants is linked to reduced Msn2/Msn4 activity, we assessed Guk1-7-GFPlevels in single deletions. An absence of MSN2, but not MSN4, led to a significantincrease in Guk1-7-GFP compared to wild type (p = 0.003) with levels similar tothose seen in whi2∆ cells (Figure 3.7A). These data would therefore suggest thatdecreased Guk1-7-GFP degradation is associated to a general impairment of stressresponse factors acting downstream of Msn2.Intriguingly, while performing WHI2 addback experiments we noticed a pro-nounced decrease in Guk1-7-GFP stability when adding a second plasmid bearingthe auxotrophic marker leucine. To further investigate this observation, we trans-formed wild type and whi2-sc1 cells with a plasmid expressing Guk1-7-GFP thatcontained one of the following selection markers: histidine, uracil, or leucine. Con-sistent with our previous data, loss of Guk1-7-GFP fluorescence was 53% lower inwhi2-sc1 cells compared to wild type when the histidine marker was used (Fig-ure 3.7B). However, when whi2-sc1 cells containing the leucine selection marker95were grown in synthetic media without additional leucine, Guk1-7-GFP degrada-tion was mostly impaired and levels were approximately two fold higher. As well,uracil selection resulted in an intermediate phenotype. Leucine was previouslyshown to activate the TORC1 kinase complex that can also inhibit Msn2/4 [220–222]. One possibility is that Whi2 is only required to maintain Msn2 active whenTORC1 is stimulated in the presence of high levels of exogenous leucine. In sup-port of this view, the addition of increasing amounts of leucine restored the im-paired Guk1-7-GFP degradation in whi2-sc1 cells (Figure 3.7C). These data sug-gest that mutations in WHI2 only impair proteostasis in conditions where Whi2 isrequired to maintain Msn2 active.3.3.5 MutantWHI2 Impairs Guk1-7-GFP Degradation by ReducingSubstrate UbiquitinationTo determine how an absence of WHI2 results in increased Guk1-7-GFP stabi-lization we first needed to clarify what aspect of protein quality control is alteredin the mutants. We first compared the thermodynamic stability of ectopically ex-pressed Guk1-7 in cellular lysates by CETSA. Solubility decreased rapidly at tem-peratures above 42◦C in extracts from both wild type and whi2-sc1 strains (Fig-ure 3.8A). While not marked, slightly more Guk1-7 remained soluble in whi2-sc1lysates compared to wild type at 46.5◦C and 48.8◦C (p = 0.04 and p = 0.032). Bycontrast, Guk1-7 was slightly, but not significantly, less soluble in whi2-sc1 cellsgrown at 25◦C or following a short twenty-minute incubation at 37◦C (p = 0.18 andp = 0.07) (Figure 3.8B). Together these data suggest that Whi2 does not markedlyinfluence Guk1-7-GFP degradation by increasing its thermal stability or inducingits aggregation.It is possible that mutations in WHI2 might generally alter the ubiquitin pro-teasome system. However, using the known proteasome substrate Deg1-GFP, wefound no significant difference in degradation in whi2-sc1 cells compared to wildtype (p = 0.31, 0.4, 0.7, 0.52 for 10, 20, 30, and 60 minute time points, respec-tively) (Figure 3.8C). We next asked whether an absence of WHI2 could affectGuk1-7-GFP ubiquitination. Ubiquitin levels were measured following pulldownof Guk1-GFP and Guk1-7-GFP from cultures grown at 25◦C. Normalizing theubiquitin signal to the amount of GFP tagged substrate eluted revealed approxi-96mately 30% less ubiquitinated Guk1-7-GFP in whi2-sc1 cells compared to wildtype (Figure 3.8D). Together these data suggest that mutated WHI2 could impairGuk1-7-GFP degradation by decreasing substrate ubiquitination.!"#!!"#$"!%& !"#$ %&'$ %&'('()*+,-(.)/00./1.234$5652781)3/9(0:((;:(;0<<<'()*+,-(.)/00./1234$565278.1)3/9(0:(;:( $"!!"#!!. !"!#. !"$. $=>>,+,/;*).)(3:,;(.*>>(>.?@ABCDE)*+$ !"#$,&-.'()*+,-(.)/00./1234$565278.1)3/9(0:(;:( $"!!"#!F,0+,>,;( D(3:,;(G9*:,))*+$ !"#$,&-.=3H/+9/IJ,:.K*94(9Figure 3.7: Msn2 is linked to reduced proteostatic capacity inWHI2mutants.a) Wild type, whi2∆, msn2∆, and msn4∆ cells expressing Guk1-7-GFP wereanalysed by flow cytometry following a two hour incubation at 25◦C and 37◦Cin the presence of CHX. P values were calculated with a one-way ANOVA andpost-hoc Tukey HSD to assess significance (*, **, and ns denote p < 0.05,0.01, and not significant, respectively). b) Guk1-7-GFP was expressed fromCEN/ARS plasmids with histidine, uracil, or leucine auxotrophic markers inwild type or whi2-sc1 cells. Cultures were incubated with CHX for two hoursat 25◦C or 37◦C before being analysed by flow cytometry. c) Wild type andwhi2-sc1 cells were co-transformed with Guk1-7-GFP and pRS315 (LEU2).Cultures were grown in synthetic drop out media containing different amountsof leucine and then incubated for two hours in the presence of CHX at 25◦Cand 37◦C before flow cytometry analysis.97! "! " # "$%&'%()*+,-.+,)*+,-.+,!"#$%&'$()*+./0123240.54.0623240657-0/232.0/5-*0)232.0*5!(89(:;<&:(2=+,>!"#$%&'$()*+-/0/ -?0) --06 -.07 4)0* 470* 4606#(:@(A<2$B2CAC!"#$%&'$()*+!C8(2=8CA0>D(E?FGH##EI?/ ?/ )/ -/ 7/ / ?/ )/ -/ 7/!"#$ %&'$()*+#(:@(A<2$B2CAC# $)*/?*/?//.**/-.CA9&< GH##EI?J#K2GH#LK2GH#J#K2GH#LK2M'N:2=ID;> %&'$()*+!"#$G&I?F.FGH#G&I?FGH#O89",+&*?"',(22'()'@1A8B",+&*?"',(22'()'@1A8B