HEPATOCELLULAR CARCINOMA By Amrita Oak A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Chemical Engineering - Doctor of Philosophy 201 9 ABSTRACT HEPATOCELLULAR CARCINOMA By Amrita Oak The endoplasmic reticu lum (ER) is the site for protein folding and maturation. ER stressors, both physiological and pharmacological , can result in activation of the unfolded protein response (UPR). Palmitate, a saturated fatty acid , is one such ER stressor and leads to inductio n of the UPR. This is primarily through the activation of Inositol Requiring Enzyme - 1 (IRE1) leading to splicing of XBP1 mRNA. However, the mechanism of this activation is unclear . With the aid of a bimolecular fluorescence complementation (BiFC) assay, we identified two crucial residues on the transmembrane domain (TM) of IRE1, S450 and W457, that are drivers of palmitate mediated activation. Previous research from our group suggested that IRE1 also has binding sites for palmitate on its cytosolic domain ( CD). However, IRE1 - CD protein expressed in E. coli was over - phosphorylated which possibly affected its binding to PA. To investigate this, we developed a protocol for expression and purification of wild type and mutant IRE1 - CD protein in insect Sf21 cells. A fluorescence polarization based binding assay was performed to determine whether palmitate binds to residues on the IRE1 - CD protein . Previously our laboratory demonstrated that palmitate induced the migration of cancer cells as well as transcription fa ctors (TF) involved in epithelial - to - mesenchymal transition (EMT) . Here, w e investigated the role of IRE1 activation on these processes . Using CRISPR gene editing to generate IRE1 knockout s in liver and breast cancer cell lines , we observed that IRE1 mediates the upregulation in EMT - TFs , a decrease in the expression of the desmoplakin (DSP) protein , and a n increase in the migration of liver and breast cancer cell s . DSP is a critical componen t of desmosomes, which function to maintain the structural integrity at adjacent cell - cell contacts . In addition to migration, t he effect of XBP1 splicing on metabolism has not be en studied. We found t he activation of IRE1 - XBP1 is accompanied by changes in the metabolic genes involved in glycolysis, fatty acid oxidation, gluconeogenesis, and ceramide metabolism , suggesting that some of the metabolic effects of palmitate are mediated through IRE1 . These results could have implications on the develop ment of c hemotherapeutic strategies. This study paves the way for further investigations into the far - reaching effects of activation of the UPR o n cell survival, metabolism , and chemo - tolerance. iv Dedicated to my family v ACKNOWLEDGEMENTS I would like to thank my advisor Dr. Christina Chan for the opportunity to work in the Cellular and Molecular Biology laboratory on many interesting projects and supporting me throughout my Ph.D. I would like to thank the members of the l aboratory for helping me in the many ways and for motivating me in my research. I had a wonderful time working, learning and growing during my Ph.D. I am deeply thankful for the advice and support offered by my Ph.D. committee: Dr. S. Patrick Walton, Dr. K aren Liby, and Dr. Michael Feig during the course of my work. I would like to mention and thank my collaborators here Dr. Amadeu Sum, Dr. Hyun Ju Cho, Dr. Aritro Nath for their input and involvement in my research objectives. Finally, I would like to tha nk my husband, Sudhanwa, for standing by me and supporting me through this journey. I am grateful for all the love, help and unconditional support from the Oak and Dewasthale families . I could not have completed my Ph.D. without their support and motivatio n. It is impossible to name everyone in this small space. All of the friends I have made throughout this time have supported and enriched my life in one way or another. My special thanks to all my colleagues in our research group for their support. Amrita vi TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ .............. x LIST OF FIGURES ................................ ................................ ................................ ........... xi LIST OF ABBREVIATI ONS ................................ ................................ ........................... xv 1 INTRODUCTION ................................ ................................ ................................ ...... 1 1.1 The Endoplasmic Reticulum ................................ ................................ ................ 1 1.2 Aggregation of misfolded proteins ................................ ................................ ....... 1 1.3 ER stress sensors ................................ ................................ ................................ .. 2 1.3.1 ATF6 ................................ ................................ ................................ ............. 2 1.3.2 PERK ................................ ................................ ................................ ............ 3 1.3.3 IRE1 ................................ ................................ ................................ .............. 6 1.4 The Unfolded Protein Response (UPR) ................................ ............................... 7 1.5 ER stress in hepatocellular carcinoma (HCC) ................................ ...................... 8 1.6 Specific aims of this work ................................ ................................ .................. 10 2 DOMAIN IN SENSING OF LIPID SATURATION ................................ ....................... 12 2.1 Introduction ................................ ................................ ................................ ........ 12 2.1.1 Endoplasmic Reticulum and the Unfolded Protein Response .................... 12 2.1.2 ................................ ....................... 12 2.1.3 ................................ ................................ . 13 2.1.4 Bimolecular Fluorescence Compl ementation Assay (BiFC) ...................... 14 2.2 Materials and Methods ................................ ................................ ....................... 16 2.2.1 Cell culture and transfection ................................ ................................ ....... 16 2.2.2 Cloning of VC - - ................................ ........ 17 2.2.3 Immunofluorescence ................................ ................................ ................... 18 2.2.4 Western Blotting ................................ ................................ ......................... 19 2.2.5 Bimo lecular fluorescence complementation (BiFC) assay ......................... 19 2.2.6 XBP1 splicing assay ................................ ................................ ................... 20 2.2.7 Tryptophan fluorescence measurements and tryptophan depth calculation 20 2.3 Results and Discussion ................................ ................................ ....................... 22 vii 2.3.1 Palmitate activates ER stress ................................ ................................ ...... 22 2.3.2 .............. 22 2.3.3 - TMD are involved in dimerization .............. 26 2.3.4 - TMD serves as a sensor for lipid membrane saturation in mammals 2 8 2.3.5 - LD in dimerization ................................ ............................. 31 2.3.6 Effect of Y161A on BiFC dimerization ................................ ...................... 34 2.4 Trp457 locates near the center of the POPC bilayer. ................................ ......... 35 3 EXPRESSION AND PURIFICATION OF A MAMMALIAN PROTEIN: .............................. 37 3.1 Introduction ................................ ................................ ................................ ........ 37 3.1.1 - CD with PA show specific residues involved in PA binding ................................ ................................ ................................ ................ 37 3.1.2 Types of heterologous protein expression systems ................................ ..... 39 3.1.3 Choosing the right expression system ................................ ......................... 40 3.2 Experimental Design ................................ ................................ .......................... 44 3.2.1 Materials ................................ ................................ ................................ ..... 44 3.3 Procedure ................................ ................................ ................................ ............ 47 3.3.1 - CD into pFastBac plasmid ................................ ............ 47 3.3.2 Preparation of recombinant bacmid ................................ ............................ 50 3.3.3 Transfection of recombinant bacmid into Sf21 cells ................................ .. 52 3.3.4 Purification of MBP fusion protein ................................ ............................. 54 3.3.5 BODIPY - PA binding assay ................................ ................................ ........ 55 3.4 Results ................................ ................................ ................................ ................ 56 3.4.1 Cloning and preparation of pFastBac - - CD ................................ ....... 56 3.4.2 Transfection and infection into Sf21 cells ................................ .................. 57 3.4.3 Optimization of protein expression and purification ................................ .. 59 3.4.4 Binding assay with BODIPY - PA ................................ ................................ 60 3.5 Conclusions ................................ ................................ ................................ ........ 65 4 - XBP1 IN HEPATOCELLULAR CARCINOMA ................................ ................................ ............ 68 4.1 Introduction ................................ ................................ ................................ ........ 68 4.1.1 ER stress ................................ ................................ ................................ ...... 68 4.1.2 IRE1: A stre ss sensor in the UPR ................................ ............................... 69 4.1.3 Desmoplakin: A component of desmosomes ................................ .............. 70 viii 4.2 Materials and Methods ................................ ................................ ....................... 71 4.2.1 Cell lines and culture conditions ................................ ................................ . 71 4.2.2 Generation of IRE1 - / - KO cell lines ................................ ............................ 72 4.2.3 Knockdown of DSP using siRNA ................................ ............................... 73 4.2.4 Wound healing assay ................................ ................................ .................. 73 4.2.5 ................................ ................................ ............. 74 4.2.6 XBP1 splicing assay ................................ ................................ ................... 74 4.2.7 qPCR for EMT transcription factors ................................ ........................... 75 4.2.8 Western Blotting for EMT markers ................................ ............................ 75 4.2.9 Luciferase assay ................................ ................................ .......................... 76 4.3 Results and Discussion ................................ ................................ ....................... 76 4.3.1 - / - hepatocellular carcinoma cell lines using CRISPR 76 4.3.2 ..................... 79 4.3.3 Desmoplakin levels decrease on palmitate treatment ................................ . 82 4.3.4 ZEB1 and ZEB2 regulate DSP transcription levels ................................ .... 86 4.3.5 XBP1 potentially regulates ZEB1/2 activity ................................ ............... 90 5 THE ROLE OF XBP1 IN CANCER CELL METABOLISM ................................ .. 92 5.1 Introduction ................................ ................................ ................................ ........ 92 5.1.1 Obesity and incidence of cancer ................................ ................................ . 92 5.1.2 ER stress, fatty acids, and cancer ................................ ................................ 94 5.1.3 Cancer cells show changes in metabolism ................................ .................. 96 5.2 Materials and Methods ................................ ................................ ....................... 97 5.2.1 Selection of genes ................................ ................................ ....................... 97 5.2.2 qP CR assay ................................ ................................ ............................... 101 5.3 Results and Discussion ................................ ................................ ..................... 103 5.3.1 LPIN1 ................................ ................................ ................................ ........ 103 5.3.2 ACAT1 ................................ ................................ ................................ ...... 108 5.3.3 IDH1 ................................ ................................ ................................ ......... 110 5.3.4 PDK1 ................................ ................................ ................................ ......... 112 5.3.5 PFKFB4 ................................ ................................ ................................ .... 113 5.3.6 CAV1 ................................ ................................ ................................ ........ 114 5.3.7 EMT - Transcriptio n Factors ................................ ................................ ...... 115 5.3.8 Effect of XBP1 levels on survival ................................ ............................ 117 5.3.9 Using the TCGA datasets to determine XBP1 targets .............................. 118 ix 6 FUTURE WORK AND CONCLUSIONS ................................ ............................. 125 6.1 Importance of junction plakoglobin (JUP) in cancer ................................ ....... 125 6.1.1 Junction plakoglobin: Desmosomal component ................................ ....... 125 6.1.2 JUP in cell signaling ................................ ................................ ................. 125 6.1.3 Regulation of JUP ................................ ................................ ..................... 127 6.2 Chemoresistance/chemo - tolerance ................................ ................................ ... 128 6.2.1 Role of XBP1 in resistance to chemotherapy ................................ ........... 128 6.2.2 Effect of IRE1 - XBP1 activation by PA on chemo - tolerance to CUDC - 101 128 6.3 Transmissible ER stress and chemo - resistance ................................ ................ 133 6.3.1 TERS confers higher survival in tumors ................................ ................... 133 6.3.2 TERS primed cells are resistant to chemotherapeutic drugs .................... 134 6.3.3 TERS may be mediated through exosomes ................................ .............. 134 6.4 Conclusions ................................ ................................ ................................ ...... 135 6.4.1 ................................ 135 6.4.2 ................................ ................................ . 136 6.4.3 stress on metabolism .............................. 136 APPENDIX ................................ ................................ ................................ ..................... 138 REFERENCES ................................ ................................ ................................ ............... 141 x LIST OF TABLE S Table 3 - - CD predicted to be involved in PA binding. .................... 38 Table 3 - 2 Characteristics of heterologous protein expression in bacterial, yeast, insect and mammalian systems ranked according to desirability. ................................ ..................... 40 Table 3 - 3 Predictions for O - glycosylation, C - mannosylation, N - glycosylation sites and glycation sites on CD - IRE1 ................................ ................................ .............................. 43 Table 3 - 4 Antibiotic concentrations and stock solutions. ................................ ................. 45 Table 3 - 5 PCR primers for insertion of IRE1 - CD (547aa - 977aa) into pFastBac plasmid . ................................ ................................ ................................ ................................ ........... 47 Table 3 - 6 Restriction digestion with SspI. ................................ ................................ ........ 48 Table 3 - 7 Mix to create pFastBac vector overhangs. ................................ ....................... 48 Table 3 - - CD insert. ................................ ........................ 49 Table 3 - 9 Working concentrations of antibiotics. ................................ ............................. 50 Table 4 - 1 Predicted binding sites for XBP1 on the ZEB1/2 promoter region relative to the transcription start site (TSS). ................................ ................................ ............................ 90 Table 5 - 1 JASPAR transcript ion factor database scores for various metabolic genes and EMT - TFs. ................................ ................................ ................................ .......................... 98 Table 5 - 2 qPCR primer sequences for selected genes and EMT - TFs . ........................... 102 xi LIST OF FIGURES Figure 1 - 1 Sequence comparison of IRE1 and PERK luminal domains 4 Figure 1 - 2 Alignment of IRE1_LD (green) and PERK_LD (magenta). ............................ 5 Figure 1 - 3 Progression of hepatocellular carcinoma. ................................ ......................... 8 Figure 2 - 1 BiFC assay for dimerization of IRE1 - TM (taken with permission from (Cho et al., 2019)). ................................ ................................ ................................ ......................... 22 Figure 2 - - - FL. ................................ .......... 23 Figure 2 - ................................ ............................. 24 Figure 2 - - ...................... 25 Figure 2 - - S450A mutant proteins. ................................ ................................ ................................ .... 25 Figure 2 - - TMD. ................................ ....... 27 Figure 2 - - ................................ ...... 29 Figure 2 - 8 BiFC assay with V437A and M440R mutations. ................................ ............ 30 Figure 2 - - M440R mutant prote ins. ................................ ................................ ................................ ... 30 Figure 2 - - TMD sense lipid aberrancy and activate - XBP1 pathway. ................................ ................................ ............................... 31 Figure 2 - - FL constructs .............................. 32 Figure 2 - - FL constructs. ................................ ....... 33 Figure 2 - 13 Effect of Y161 mutation on XBP1 splicing. ................................ ................. 34 Figure 2 - 14 Trp457 localizes in the center of the small unilamellar vesicles (SUVs). .... 35 Figure 3 - 1 Alignment of IRE1 and PKR. ................................ ................................ ......... 37 Figure 3 - 2 N - linked glycosylation patterns for insect cells and mammalian cells. .......... 42 Figure 3 - 3 Flowchart for all the steps involved in the expression and purification of proteins from Sf21 cells. ................................ ................................ ................................ ................. 46 xii Figure 3 - 4 Plasmid map for pFastBac - - CD. ................................ ........................... 56 Figure 3 - 5 Colony PCR of white E. coli transformant colonies. ................................ ...... 57 Figure 3 - 6 Signs of infection in untransfected Sf21 cells(left) and P3 treated Sf21 cells (right) for 24 hrs. Scale bar = 45µm ................................ ................................ ................. 57 Figure 3 - 7 Expression of MBP - - CD protein ................................ ......................... 59 Figure 3 - 8 Western blot showing the time course of protein expression in P0 (left) and P1 (right) infected Sf21 cells using anti - ................................ ..................... 60 Figure 3 - 9 Structure of BODIPY fluorophore (from Thermo Fisher). ............................. 61 Figure 3 - 10 Fluorescence polarizat ion (FP) assay using BODIPY - PA. ........................... 62 Figure 3 - 11 BODIPY - PA binding to BSA and lysozyme. ................................ ............... 63 Figure 3 - 12 BODIPY - - CD wild type and mutants. ............. 64 Figure 3 - - CD on PA binding. ................................ ................. 65 Figure 4 - 1 CRISPR clones obtained for Hep3B cells. ................................ ...................... 77 Figure 4 - 2 CRISPR clones obtained for HepG2 cells. ................................ ..................... 78 Figure 4 - 3 RT - PCR showing XBP1 splicing activity in wild typ e and IRE1 - / - Hep3B cells on induction with palmitate. ................................ ................................ ............................. 78 Figure 4 - 4 XBP1 splicing activity in wild type and IRE1 - / - HepG2 cells on induction with PA. ................................ ................................ ................................ ................................ .... 79 Figure 4 - 5 Wound healing assay with MEF IRE1 - / - KO cells transfected with wild type 7A - FL - - ...... 80 Figure 4 - n Hep3B(left) and HepG2 (right) cells. ................................ ................................ ................................ .......... 81 Figure 4 - 7 Boyden's chamber assay for Hep3B, HepG2, MDA MB 231. ....................... 83 Figure 4 - 8 Desmoplakin levels in Hep3B wild type vs IRE1 - / - KO treated with PA. ...... 84 Figure 4 - 9 Expression of EMT markers in DSP KD Hep3B and HepG2 cells ................ 85 Figure 4 - 10 Scatter - plots showing correlation between the mRNA expression levels of DSP and ZEB1 or ZEB2 using the TCGA PAN - CAN dataset. ................................ ................ 86 Figure 4 - 11 Heat map of gene expression profiles of DSP, ZEB1 and ZEB2 (X - axis) across all samples in the TCGA pan - cancer database grouped by cancer type (Y - axis). ............ 87 xiii Figure 4 - 12 qPCR for ZEB1 gene expression in Hep3B. ................................ ................. 88 Figure 4 - 13 Luciferase assay with Hep3B and MDA MB 231. ................................ ....... 89 Figure 5 - 1 Increase in risk of liver cancer across US populations due to obesity. ........... 93 Figure 5 - 2 Position weight matrix for XBP1 transcription factor from the JASPAR TF database. ................................ ................................ ................................ .......................... 101 Figure 5 - 3 Prediction of XBP1 binding sites on LPIN1 from the PWMScan output. .... 104 Figure 5 - 4 Heat map showing the correlation between XBP1, LPIN1, MOGAT2, AGPAT1, GPAT2, PLD1 in the TCGA BRCA dataset. ................................ ................ 105 Figure 5 - 5 Correlation of XBP1 vs LPIN1 and Kaplan - Meier survival curve for BRCA and PANCAN. ................................ ................................ ................................ ....................... 106 Figure 5 - 6 Relative gene expression for LPIN1. ................................ ............................ 107 Figure 5 - 7 ACAT1 gene expression and correlation with XBP1 and Kaplan - Meier survival curve. ................................ ................................ ................................ ............................... 109 Figure 5 - 8 IDH1 gene e xpression and correlation with XBP1. ................................ ...... 110 Figure 5 - 9 Kaplan Meier survival curve for IDH1levels in PANCAN (left) LIHC (middle) and BRC A (right). ................................ ................................ ................................ ........... 111 Figure 5 - 10 qPCR and Kaplan - Meier survival curve for gene expression levels of PDK1. ................................ ................................ ................................ ................................ ......... 112 Figure 5 - 11 qPCR and Kaplan - Meier survival curve for gene expression levels of PFKFB4. ................................ ................................ ................................ ................................ ......... 114 Figure 5 - 12 qPCR and Kaplan - Meier survival curve for gene expression levels of CAV1. ................................ ................................ ................................ ................................ ......... 115 Figure 5 - 13 qPCR for expression levels of ZEB1 and TWIST. ................................ ..... 116 Figure 5 - 14 Kaplan - Meier survival curve for BRCA, triple negative breast cancer (TNBC), LIHC and COADREAD. ................................ ................................ ................................ 117 Figure 5 - 15 Venn diagram showing the number of genes with a positive correlation with XBP1 from liver cancer (LIHC), breast cancer (BRCA) and colorectal cancer (COADREAD) databases and their overlap. ................................ ................................ .. 118 xiv Figure 5 - 16 343 genes from LIHC, BRCA and COADREAD sectioned by gene ontology (GO) biological processes using PANTHER tools. ................................ ........................ 119 Figure 5 - 17 Overrepresented genes in BRCA, LIHC and COADREAD datasets as compared to Homo sapiens reference genes. ................................ ................................ .. 119 Figure 5 - 18 343 genes from LIHC, BRCA and COADREAD sectioned into 101 gene ontology (GO) biological pathways using PANTHER tools. ................................ ......... 121 Figure 5 - 19 Pie chart showing the categorization of genes involved in metabolic processes by sub - categories and levels of processes. ................................ ................................ ...... 122 Figure 6 - 1 T - coffee expresso alignment result for human beta - catenin and JUP. .......... 126 Figure 6 - 2 Correlation between JUP and transcription factors ZEB1 and ZEB2. .......... 127 Figure 6 - 3 Structure of CUDC - 101, a multitarget inhibitor against HDACs, EGFR, and HER2 receptors. ................................ ................................ ................................ .............. 129 Figure 6 - 4 IC50 curves for treatment of Hep3B cells with CUDC 101. ........................ 130 Figure 6 - 5 EGFR kinase domain (orange) and IRE1 kinase domain (green) aligned using PDBeFold. ................................ ................................ ................................ ....................... 131 Figure 6 - 6 Alignment of IRE1 and c - Abl kinase using PDBeFold. ............................... 132 xv LIST OF ABBREVIATIONS ACAA1 Acetyl - CoA Acyltransferase 1 ACAT1 Acetyl - CoA Acetyltransferase 1 ADP Adenosine diphosphate AGPAT1 1 - Acylglycerol - 3 - Phosphate O - Acyltransferase 1 ATF6 Activating transcription factor 6 BiFC Bimolecular Fluorescence Complementation BMI Body Mass Index BRCA Breast Invasice Carcinoma Cohort BSA Bovine serum albumin bZIP basic region leucine zipper CAV1 Caveolin 1 CD Cytosolic domain CDS Coding sequence CHO Chinese hamster ovary ChREBP Carbohydrate Response Element Binding Protein CML Chronic myeloid leukemia COADREAD Colon Adinocarcinoma and Rectal Adenocarcinoma Co - IP Co - immunoprecipitation CPT2 Carnitine Palmitoyltransferase 2 CRISPR Clustered regularly interspaced short palindromic repeats CV Column volumes xvi DAG Diacylglycerol DGAT2 Diacylglycerol O - Acyltransferase 2 DMEM Dulbecco's Modified Eagle Medium DSP Desmoplakin EMT Epithelial - to - mesenchymal transition ER Endoplasmic reticulum ERAD ER associated degradation EV Extracellular vesicle FADS1 Fatty acid desaturase1 FASN Fatty acid synthase FBS Fetal bovine serum FOXC2 Forkhead box C2 FP Fluorescence polarization FRET Fluorescence resonance energy transfer GAPDH Glyceraldehyde 3 - phosphate dehydrogenase GEO Gene expression omnibus GFP Green fluorescent protein GPAT2 Glycerol - 3 - Phosphate Acyltransferase 2 HCC Hepatocellular carcinoma HEK 293 Human embryonic kidney 293 HIF1A Hypoxia Inducible Factor 1A IDH1 Isocitrate dehydrogenase IL - 6 Interleukin 6 xvii IRE1 Inositol requiring enzyme 1 JUP Junction Plakoglobin LD Luminal Domain LIC Ligation independent cloning LIHC Liver hepatocellular carcinoma LPIN1 Lipin1 MBP Maltose binding protein MD Molecular dynamics MEF Mouse embryonic fibroblast MOGAT2 Monoacylglycerol O - Acyltransferase 2 NAFLD Non - alcoholic fatty liver disease NASH Non - alcoholic steatohepatitis NCEP National Cholesterol Education Program NEFA Non - esterified fatty acid ORF Open reading frame PA Palmitate PCR Polymerase chain reaction PDK1 Pyruvate Dehydrogenase Kinase 1 PERK Protein kinase R like endoplasmic reticulum kinase PFKFB3 6 - Phosphofructo - 2 - Kinase/Fructose - 2,6 - Biphosphatase 3 PFKFB4 6 - Phosphofructo - 2 - Kinase/Fructose - 2,6 - Biphosphatase 4 PHGK1 Phosphorylase Kinase Catalytic Subunit Gamma PKM2 Pyruvate Kinase M1/2 xviii PKR Protein kinase R PLD1 Phospholipase D1 PPI Protein protein interaction PTM Post - translational modifications PWM Position Weight Matrix RIDD Regulated IRE1 dependent decay RMSD Root mean square deviation RTK Receptor tyrosine kinase S1P Site 1 protease S2P Site 2 protease SCD1 Stearyl - CoA - desaturase1 SPT1 Serine palmitoyltransferase SPTLC1 Serine Palmitoyltransferase Long Chain Base Subunit 1 SPTLC3 Serine Palmitoyltransferase Long Chain Base Subunit 3 Src3 Steroid Receptor Co - activator 3 SREBP Sterol Regulatory Element Binding Protein SUV Small unilamellar vesicles TCF T - cell factor TERS Transmissible ER stress TEV Tobacco etch virus TF Transcription factors TLR4 Toll - like receptor 4 Tm Tunicamycin xix TMD Transmembrane domain TNBC Triple negative breast cancer TSS Transcription start site UPR Unfolded protein response WHO World Health Organization XBP1 X - box binding protein 1 YFP Yellow fluorescent protein ZEB1 Zinc Finger E - Box Binding Homeobox 1 ZEB2 Zinc Finger E - Box Binding Homeobox 2 delta delta Ct 1 1 INTRODUCTION 1.1 The Endoplasmic Reticulum The endoplasmic reticulum (ER) is the site for protein folding, protein maturation, Ca 2 + storage, and redox homeostasis. However, the ER is not just a tube that proteins pass through on their way to the secretory organelles. Decades of research has established active monitoring of the folding status of the proteins passing through the ER. This monitoring transmits the information outside of the ER through a series of three ER stress sensors that activate their repertoire of downstream signaling pathways. ER stress is defined as an imbalance between the protein folding capacity of the ER and the protein tra nslation capacity. If the ratio of unfolded polypeptides to folded and post - translationally modified proteins goes up, the ER is said to be perturbed. Under these conditions, the Unfolded Protein Response (UPR) is activated to mitigate ER stress. Canonical ly, the reasons for ER stress have been a discrepancy in protein folding capacity. changes to the redox homeostasis and lipid perturbation can also cause ER stress and a ctivate the UPR (Cho, 2013b; Pineau et al., 2009; R. Volmer, van der Ploeg, & Ron, 2013; Romain Volmer & Ron, 2015; Yadav, Chae, Kim, & Chae, 2014) . Nutrient starvation as commonly seen in cancer cells and nutrient excess as seen in hepatic steatosis both can activate the UPR. 1.2 Aggregation of misf olded proteins Usually, protein folding is determined by its primary amino acid sequence. Each possible conformation after folding has a certain free energy, and the final conformation of the protein is the one with the lowest free energy (Schröder & Kaufman, 2005a) . Water is 2 then excluded and the protein is folded . This process is aided by chaperone proteins and the presence of post - translational modifications. One scenario in the aggre gation of misfolded proteins is when the hydrophobic patches on one protein collide and collapse with another protein leading to aggregation. Protein folding enzymes are broadly classified as foldases, lectins, and chaperones. How do the ER stress sensors sense misfolded proteins? Let us consider the chaperone protein BiP. It does not catalyze faster folding (unlike foldases) but works through (Schröder & Kaufman, 2005c) . BiP recognizes and binds to hydrophobic regions on the unfolded protein that should, on proper folding, be buried deep inside the core with low affinity. 1.3 ER stress sensors There are three main sensors of ER stress in the UPR: Inositol Requiring Enzyme 1(IRE1), Protein kinase RNA - like endoplasmic reticulum kinase (PERK) and Activating Transcription Factor 6 (ATF6). 1.3.1 ATF6 Activating Transcription Factor 6 is an ER stress sensor. It is a Type II transmembrane bZIP domain containing ER stress sensor. Briefly, ATF6 has two domains: a luminal domain (LD) that senses misfolded proteins and a cytosolic domain that has a bZIP motif at the N - terminal and a transcriptional activation domain at the C - terminal. In non - stressed conditions, the ATF6 - LD binds to BiP. ATF6 is also sensitive to the calnexin/calreticulin cycle, and under - glycosylated ATF6 is retained in the ER through its interaction with the lectin calreticulin. Upon ER stress, BiP is released leading to the exposure of two Golgi localization sequen ces. It migrates to the Golgi where is it cleaved 3 by S1P and S2P proteases. The cytosolic fragment translocates to the nucleus and induces transcription of UPR associated genes (Hetz & Glimcher, 2009; Malhi & Kaufman, 2011; Nagelkerke, Bussink, Sweep, & Span, 2014; Schröder & Kaufman, 2005b) . ATF6 has two homologs - ATF or repressor activity based on whether it forms homodimers or heterodimers within the homologs . ATF6 primarily is responsible for upregulating genes for responses geared towards cell survival (Vandewynckel et al., 2013) . Recent evidence has suggested that ATF6 and sXBP1 may have a combinatorial effect on expression of a separate subset of genes in the ERAD pathway (Hetz, Chevet, & Harding, 2013; Shoulders et al., 2013) . Deletion of both ATF6 homolog s is embryonic lethal. H owever develop normally (Shuda et al., 2003) . ATF6 - / - mice show a sustained activation of the - XBP1 branch of the UPR. Therefore, the ATF6 branch of the UPR is adaptive in nature - it supports the other branches in regulating proteostasis (Shoulders et al., 2013) . 1.3.2 PERK Protein kinase R - like endoplasmic reticulum kinase (PERK) is another ER stress sensor. PERK and IRE1 are similar in that both are type I transmembrane proteins with a kinase domain. BiP binds to the luminal domain of PERK and keeps it in its inactive form. On sensing of misfolded proteins by BiP and its release from the PERK luminal domain, PERK dimerizes , forms higher order oligomer s and is activated. binding of the methio nyl - tRNA f o incapable of translating proteins leading to a generalized attenuation of translation. 4 However, it is capable of translating a select subset of mRNAs with short open reading frames (ORFs). Thi s includes ATF4, a transcription factor with a bZIP motif. CHOP/GADD34 are targets of ATF4. Both PERK and IRE1 are capable of being activated by lipid sensing - preferentially in response to saturated fatty acids like palmitate. However, the luminal domain is dispensable to this sensing. Our research suggests that this originated from the sensing of membrane saturation by the transmembrane domain. Figure 1 - 1 Sequence comparison of IRE1 and PERK luminal doma ins . Alignment of luminal domains of IRE1and PERK using T - coffee Espresso (T - COFFEE, Version_11.00.d625267). Both the luminal domains are distinct from one another suggesting different modes of activation of IRE1 and PERK. 5 We compared the LD sequence of PERK and IRE1 to assess the extent of homology. Figure 1 - 1 shows a sequence alignment between them with orange showing a good degree of simi larity (score=68). A 3D homology structure prediction was done using CPHmodels 3.2 ( Figure 1 - 2 , left ) . The RMSD value (6.069Å) denotes a poor alignment. Using crystal st ructures from the RCSB PDB database improved the RMSD value of the alignment slightly (4.29Å) ( Figure 1 - 2 , right) . Overall, there is very limited homology in PERK and IRE1 LD suggesting differing mechanisms of activation (Chuan Yin Liu, Schroder, & Kaufman, 2000) . Figure 1 - 2 Alignment of IRE1_LD (green) and PERK_LD (magenta) . On the left, homology structure prediction was done using CPHmodel 3.2, and the resulting structures were aligned using PyMol. On the right, existing cr ystal PDB structures were used for the alignment of IRE1 and PERK LD. The RMSD values (depicted in the figure) do not indicate a strong alignment of the luminal domains as suggested by the Tcoffee Espresso alignment 6 1.3.3 IRE1 Inositol requiring enzyme 1 (IRE1) is one of the most conserved stress transducers. There are two homologs of IRE1 (Schröder & Kaufman, 2005b) . IRE1 is a dual function kinase and endoribonuclease protein. It is a transmembrane protein with a type I transmembrane domain connecting the luminal domain and cytosolic domain of IRE1. The luminal domain is in the ER lumen while the cytosolic domain is in the cytosolic space. Canonically, both IRE1 and PERK are activated by BiP dissociation from their luminal domains. Misfolded or unfolded proteins bind to the substrate binding domain of BiP causing BiP to dissociate from IRE1_LD (Carrara, Prischi, Nowak, Kopp, & Ali, 2015) . ADP - BiP has a high affinity for proteins and ADP - ATP cycling is require d for their release and subsequent activation of IRE1 and PERK. IRE1_LD then dimerizes in a manner reminiscent of major histocompatibility complex dimerization (Credle, Finer - moore, Papa, Stroud, & Walter, 2005) . Activation of IRE1 is not limited to dimerization. Higher order oligomers are formed that are important in activation of the RNas e - active form of IRE1 (Maurel, Chevet, Tavernier, & Gerlo, 2014) . On dimerization of the IRE1_LD, the cytosolic domain comes together and t rans - autophosphorylates at the kinase domain (Ali et al., 2011) . The phosphorylation causes a conformational change that leads to the formation of the ribonuclease pocket. The primary substrate for the IRE1 RNase domain is XBP1: a b asic region leucine zipper (b ZIP ) transcription factor in mammalian cells or Hac1 in ye ast. A 26 - bp intron in XBP1 mRNA is spliced and the spliced mRNA is joined back together with a ligase. 7 1.4 The Unfolded Protein Response (UPR) The UPR is a series of cellular responses that are activated in order to resolve unfolded/misfolded protein accumu lation through ATF6, PERK and IRE1. The activated responses include both pro - survival and pro - apoptotic arms. The UPR is activated both in normal physiological conditions as well as a response to diseases or metabolic dysfunction. In B cell differentiation , expanded production of immunoglobulins is accompanied by activation of the UPR, specifically through the IRE1 - XBP1 axis (Mitra & Ryoo, 2019) . Knockdown or absence of XBP1 is sufficient to cause an impaired UPR leading to B cell death (Iwakoshi, Lee, & Glimcher, 2003) . 8 1.5 ER stress in hepa tocellular carcinoma (HCC) Figure 1 - 3 Progression of hepatocellular carcinoma. Injuries to the liver initiate the progression from a normal liver to fatty liver to hepatic steatosis and fibrosis culminating in hepatocellular carcinoma . T his increase in rates of obesity is ac com pani e d by the symptoms of metabolic syndrome - insulin resistance , dyslipidemi a and chronic inflammation . As depicted in Figure 1 - 3 , v arious injuries to the liver can activate inflammation pathways that lead to increased production of cytokines, increased lipogenesis and retention of fat. These injuries can be viral infections like Hepatiti s B and C, toxins like alcohol and drugs (Ji & Kaplowitz, 9 2006) . The injury we will focus on in this work as our motivation, is obesity and insulin resistance. Obesity increases HCC risk by 4.5 fold in male patients (Nakagaw a, Umemura, et al., 2014) . Hepatocytes make up to 70% of liver cells. Their primary functions are lipogenesis, synthesis of cholesterol and glucose metabolism (A. - H. Lee & Glim cher, 2009; Rutkowski, 2018) . - 98% of morbidly obese patients (Lebeaupin et al., 2018) . Accumulation of lipids in hepatocytes leads to dyslipidemia and insulin resista nce. Non - clinical features of non - alcoholic disease of the liver associated with the pathological et al in 1980 (James & Day, 1998; Ludwig, Viggiano, McGill, & Oh, 1980) . Saturat ed fatty acids, specifically, contribute to progression of hepatic steatosis to non - alcoholic steatohepatitis (NASH) and hepatic cirrhosis. Studies in mice have found increased XBP1 splicing along with an increase in the levels of CHOP and caspase - 3 as a r esponse to diets high in saturated fatty acids (Dong Wang, Wei, & Pagliassotti, 2006) . The level of saturated free fatty acids in plasma increases in NASH (0.4 - 0.5mM) patients as compared to healthy controls (0.2 - 0.25 mM) (Tavares De Almeida, Cortez - Pinto, Fidalgo, Rodrigues, & Camilo, 2002) . Dyslipidemia is induced by activation of SREBP1 and PPAR by ER stress. They are both transcription factors that control expression of genes controlling lipid metabolism (J. Han & Kaufman, 2016; Luo et al., 2017; Reue, 2009) . SREBP1 is usually maintained in an inactive state in the ER. O n cleavage and activation in the Golgi, it promotes transcription of genes controlling triglyceride and cholesterol synthesis (Ju Youn Kim et 10 al., 2018) . Accumulation of fats in t he liver is accompanied by lipotoxicity and oxidative stress (Buzzetti, Pinzani, & Tsochatzis, 2016) . Mice fed a high fat diet showed an upregulation in UPR markers and activation and translocation of SREBP1 (Nakagawa, Umemura, et al., 2014) . Inhibiting ER stress with pharmacological inhibitors blocked SREBP activation and consequentially, lipid accumulation. Non - alcoholic fatty liver disease (NAFLD) progresses to HCC from h epatic cirrhosis. increasing phosphorylation of the insulin receptor substrate - 1 (IRS1). This contributes to insulin resistance and is not seen in IRE1 - / - mice (Ozcan et al., 2004) . Saturated fatty acids like palmitate have been shown to incr ease insulin resistance by increase in FFA uptake leading to an accumulation of diacylglycerol and activation of PKC which leads to phosphorylation of IRS1. The effect of palmitate can be reversed by oleate treatment presumably by reversing diacylglycerol accumulation (Coll et al., 2008) . However, this may not be the only mechanism of action. Palmitate induces activation of ER str ess sensors, specifically IRE1, leading to insulin resistance without a significant increase in triglyceride levels (J. Lee, Cho, & Kwon, 2010) . 1.6 Specific aims of this work Previous studies in our research g roup and others have shown that saturated free protein response. However, no exact mechanism of action had been elucidated. Through this work, we demonstrate that (1) the transmembrane domain of IRE1 is important in 11 dimerization and activation of IRE1 and this is mediated through the tryptophan residue at position 457 and the serine residue at 450. We developed a bimolecular fluorescence complementation (BiFC) assay to assay dimerization of IRE1 subunits in the ER membrane. The BiFC assay can be further modified for use with other protein - protein interacting partners. We also (2) investigate the hypothesis that the saturated fatty acid palmitate binds directly to the cytosolic domain of IRE1. Previous work has shown that IRE1 - CD expressed in E. coli is capable of binding PA . However, this form of the IRE1 - CD was over - phosphorylated. Therefore , we expressed IRE1 - CD produced in insect cells and compared bindin g capacity to different mutant IRE - CD. Activation of IRE1 leads to formation of the endoribonuclease pocket that splices XBP1 mRNA. The spliced form of XBP1 (sXBP1) codes for a transcription factor that activates the UPR signaling genes downstream of XB P1. In this work, we found that IRE1 activation contributed to an increase in migration ability of breast and liver cancer cells, therefore we hypothesize that (3) this occurs through loss of desmoplakin, a cell adhesion protein and investigate the transcr iptional regulators involved in this loss of expression. Given that metabolism is well known to be altered in cancer, we also hypothesize that (4) activation of IRE1 and sXBP1 regulate the transcription of genes in various metabolic pathways and comment on how th ey potentially affect the prognosis of patients battling breast , colon and liver cancer. Looking at the effects of IRE1 activation on targets beyond the unfolded protein response will help us construct a more comprehensive picture of the changes tha t could be occurring during cancer progression. 12 2 STRUCTURAL FEATURES OF THE HUMAN IRE TRANSMEMBRANE DOMAIN IN SENSING OF LIPID SATURATION 2.1 Introduction 2.1.1 Endoplasmic Reticulum and the Unfolded Protein Response The endoplasmid reticulum (ER) is the primary site for protein translation and protein folding. The ER strives to maintain homeostasis by a balance between the rates of protein translation and protein folding/maturation. In cases where this homeostasis is disrupted, the cell is said to be unde rgoing ER stress . This may be caused by different physiological, chemical, or metabolite stressors. Some common ER stressors are free fatty acids and glucose, oxidative stress leading to the accumulation of unfolded/misfolded proteins or toxic lipid specie s (S. Wang & Kaufman, 2012) . ER stress activates the un folded protein response (UPR) designed to restore the cell to homeostasis. This is mediated by three ER stress sensor proteins, IRE1 (Inositol - requiring enzyme 1), PERK (protein kinase R (PKR) - like ER kinase), and ATF6 (activating transcription factor 6) (Kaufman, 1999) . The UPR attenuates ER stress by 1) global attenuation of protein translation 2) upregulation of chaperone proteins and foldases 3) activation of the ER associated degra dation (ERAD) pathway 4) increase in the size of the ER by biogenesis . If all of these adaptive responses fail to restore ER homeostasis, apop t otic pathways are activated leading to cell death (Hetz & Papa, 2018; M. Wang & Kaufman, 2014) . 2.1.2 The IRE1 branch is the most conserved branch of the UPR. transmembrane protein with an N - terminal luminal domain (LD) that senses ER stress and 13 a bifunctional cytosolic domain containing a Ser/Thr kinase domain and an endoribonuclease domain (Korennykh & Walter, 2012) . Activation of IRE1 is through the LD . The luminal domain dimerizes and forms higher order oligomers. (Oikawa, Kimata, Kohno, & Iwawaki, 2009) . This brings together the kinase domain which trans - autophosphorylates leading to activation of the RNase domain (Ali et al., 2011) . The endoribonuclease domain of IRE1 is responsible for the splicing of X - box binding protein 1 (XBP1) mRNA. The spliced form of XBP1 is translated to form an active transcription factor that is responsible f or the upregulation of chaperone protein genes involved in enhancing ER protein - folding capacity and proteins in the ERAD pathway (A. - H. Lee, Iwakoshi, & Glimcher, 2003; Yoshida, Matsui, Yamamoto, Okada, & Mori, 2001) . 2.1.3 A study by Volmer et al ., suggested that IRE1 without its LD was still able to sense lipid perturbation in the ER membrane and show activation of its kinase and endoribonuclease domains (Romain Volm er, Ploeg, & Ron, 2013) . Membrane tethering via the transmembrane domain (TMD) was observed to be important for this activation. A subsequent study suggested that generic TMD features are sufficient for activation of (Kono, Amin - Wetzel, & Ron, 2017) . Dimerization of the LD of its transmembrane domain (TMD) in the activation process remains unclear and direct - TMD conformation is required to establish which features of the - through lipid perturbation . In this study, we can sense membrane lipid saturation. We h ypothesize - . To 14 - TMD, we developed an in vivo BiFC (Bimolecular Fluorescence Complementation) assay to confirm that lipid saturation induces th - TMD in the ER membrane. Along with MD simulations performed by our collaborators, we identified an essential residue, tryptophan 457, that - TMD. Our data reveal a novel regulatory mec hanism in which unique structural features of TMD are important in the saturated fatty - TMD through lipid perturbation by saturated free fatty acids. 2.1.4 Bimolecular Fluorescence Complementation Assay (BiFC) The primary assay used in this study to visualize the interaction of protein partners is the Bimolecular Fluorescence Complementation (BiFC) assay. Traditionally, protein - protein interactions (PPIs) have been stud ies by overexpressing a protein with an epitope tag. A co - immunoprecipitation assay (Co - IP) may be done to identify interacting partners. However, this approach has limitations because of the high expression of proteins that it requires. Overexpression may cause protein aggregation and non - specific interactions, increased cell death, alterations in the localization pattern of the protein, etc . This is especially true in the case of transmembrane or membrane - associated proteins. Fusion of fluorescent protein s to observe PPIs in techniques such as BiFC and FRET is qualitatively better since it also provides protein localization data along with information on interaction. Fluorescence resonance energy transfer (FRET) assays are performed by attaching full - lengt h fluorescence proteins to the interacting partners. FRET pairing is observed only when the FRET pair is within 10nm of each other. However, this assay is limited by the need of skill and expensive instrumentation needed to observe the FRET signal, the nee d for an exhaustive set of standards and controls as well as a sterically permissive orientation 15 of the FRET fluorophores. An advantage of BiFC over more traditional interaction analyses and FRET is the ease of visualization and the minimal perturbation of the cellular environment. Fluorescence complementation was observed by Ghosh et al. , in 2000 where they expressed two GFP peptide fragments in Escherichia coli cells that were then reconstituted to form a fluorescent protein (Ghosh, Hamilton, & Regan, 2000) . This assay was first developed by Kerppola and his colleagues in 2 002 where it was used to investigate the interaction between Fos and Jun (Hu, Chinenov, & Kerppola, 2002) . It is ba sed on the fusion of two fragments of a non - fluorescent protein to a potential interacting protein pair. The fusion proteins used in BiFC assays are unchanged in their cellular function or localization. Different fluorescent counterparts can be used as a fusion partner. Early BiFC assays had GFP fragments fused to the interaction partner. Molecular manipulation of fluorescent proteins like GFP has produced variants with increased specificity, different maturation times and temperature, increased sig nal intensity, etc . Now, yellow fluorescent protein (YFP) and its derivatives like Venus, Cyan, Cerulean, etc can be used depending on the specific application. Each protein can be fragmented at different sites to obtain fusion fragments of varying specif icity and signal to noise ratio. With BiFC fluorescent proteins, the goal is to identify a system that has maximum specificity, high signal to noise ratio and high signal intensity. The fluorescent protein fragments can be fused to the N - terminal or the C - terminal end of the interacting proteins. An ideal fusion protein pair would have the following characteristics: the fusion protein should be functionally identical to the endogenous protein, and the expression level localization pattern in the cell shoul d be the same. BiFC assays are susceptible to non - specific interactions due to overexpression of fusion 16 constructs or steric compatibility. Proper controls need to be included to draw conclusions about the interaction partners. Absence or presence of BiFC fluorescence by itself does not confirm the interaction of the protein partners without demonstrating the robustness of the assay with proper controls. It is therefore desirable to develop and test out different fluorescent protein fragments with low non - s pecific background fluorescence. A positive interaction result must be tested out with mutant proteins that disrupt the interaction. We chose the Venus BiFC system developed extensively by Kerppola and colleagues for our BiFC assay. The Venus protein is a variant of YFP with a higher fluorescence intensity and faster maturation times. In 2010, Hu et al . searched for mutations to reduce spontaneous self - association of the two non - fluorescent fragments of Venus and identified an I152L mutation that drasticall y reduced self assembly (Kodama & Hu, 2010) . Throughout the years, different Venus pairs have been used demonstrated the highest fluorescence intensity and signal to noise (S/N) ratio for the VC155 - VN 155 pair. Coupled with the I152L mutation, this makes for an efficient BiFC assay. 2.2 Materials and Methods 2.2.1 Cell culture and transfection / mouse embryonic fibroblast (MEF) cells were maintained at 37°C at 5% CO 2 MEM, Thermo cat# 11995065) containing 4500 mg/L glucose, L - glutamine, sodium pyruvate, and sodium bicarbonate supplemented with 10% fetal bovine serum. Cells were reverse - transfected with - FL ( full length ) or - dLD ( truncated luminal domain ) BiFC constructs using Lipofectamine® LTX with Plus reagent (Thermo Fisher Scientific , cat# A12621 ) for 24 h. Afterward, the media 17 was removed and replaced with complete media containing only 2% bovine serum albumin (BSA) or 0.4 mM palmitate (PA) in 2% BSA media. 2.2.2 Cloning of VC - - pBiFC - VC155 and pBiFC - VN155 (I152L) were a gift from Chang - Deng Hu (Addgene plasmid # 22011 and #27097) (Shyu et al., 2006). pcDNA - Randal J. Kaufman (University of Michigan, M - 18 aa) ( MPARRLLLLLTLLLPGLG ) was added at the N - terminal of the coding sequence upstream of the HA - and Myc - tagged VC and VN fragments. The coding sequences of the truncated lu minal domain (19 - 29aa, 408 - 443aa) followed by the transmembrane and cytosolic domains (444 - 977 aa) were inserted at the C - terminus of the VC and VN fragments (see Figure 2 - 1 - full length BiFC constructs, the HA - and Myc - tagged VC and VN fra gments were inserted between the signal peptide sequence and the C - 9 - 977aa). - TMD and the V437R, M440R, Y161A mutations were designed and performed according to guidelines from the QuikChange Lightning Site - Directed Mutagenesis kit (Agilent). The sequences of the mutations were confirmed by DNA sequencing services (Eurofins Genom ics). Cloning of VC - VN - - VC155 and pBiFC - VN155 (I152L) were a gift from Chang - Deng Hu (Addgene plasmid # 22011 and #27097) (Shyu et al., 2006). pcDNA - gift originally from Randal J. Kaufman (University of Michigan, MI, USA)) was used to - 18 aa) was added at the N - terminal of the coding sequence upstream of the HA - and Myc - tagged VC and VN fragments. The coding sequences of the truncated luminal domain (19 - 29aa, 408 - 443aa) 18 followed by the transmembrane and cytosolic domains (444 - 977 aa) were inserted at th e C - terminus of the VC and VN fragments (see Figure 2 - 1 - full length BiFC constructs, the HA - and Myc - tagged VC and VN fragments were inserted between the signal peptide sequence and the C - 9 - 977aa). Prim - TMD and the V437R, M440R, Y161A mutations (see Appendix) were designed and performed according to guidelines from the QuikChange Lightning Site - Directed Mutagenesis kit (Agilent). The sequences of the mutations were confirmed by DNA sequencing services (Eurofins Genomics). Primers used to confirm insertion were CMV - for ( - CGCAAATGGGCGGTAGGCGTG - ), seqC - f - CGTGACCGCCGCCGGGATCAC ), seqN - f - GACCACATGAAGCAGCACGACTTCTTCAAG - seq - r - CGGGTGTTTGAGCACGTGCTTCGCTG - 2.2.3 Immunofluorescence The hepatocellular carcinoma cell line Hep3B Cells were reverse transfected with pBiFC - VC155 - - - - - VC155 - - FL and pBiFC - VN155 - - FL plasmids using Lipofectamine LTX for 24 h and then fixed ( 3.7 % formaldehyde for 10 mins) , permeabilized ( 0.1 % Triton - X for 5 mins) , and co - stained with an anti - calnexin antibody ( 1:100, Santa Cruz Biotech, cat# sc - 23954), an anti - HA antibody ( 1:100, Abcam, cat# ab9110), or an anti - My c antibody ( 1:100, Abcam, cat# ab9106). Goat anti - rabbit IgG H&L secondary antibodies conjugated with Alexa 488 ( 1:1000, Abcam, cat# 150077) or goat anti - mouse IgG H&L Alexa 568 ( 1:1000, Abcam , cat# 175701) were further incubated. Representative images for cells expressing Myc - VN - 19 - - VC - - had the same colocalization pattern. 2.2.4 Western Blotting For western blotting to confirm level of expression of IRE1, eIF2A and pEIF2A, 30 - 50ug total p rotein was loaded onto an SDS - PAGE gel run in a Tris/glycine/SDS buffer system. The proteins were transferred onto a nitrocellulose membrane and blocked with 5% BSA for 3 hrs at RT. Primary antibody incubation was done overnight at 4 ° C with antibodies agai nst IRE1 (Cell Signaling Tech, cat # 3294) , eIF2A (Cell Signaling Tech, cat # 9722 ) , phospho - eIF2A (Cell Signaling Tech, cat # 9721 ) , all at a 1:1000 dilution overnight at 4 o C. Secondary antibodies used were Goat Anti - rabbit IgG (H+L) Secondary antibody, HR P (ThermoFisher, cat# 65 - 6120) at a dilution of 1:10000 in 5% milk for 1 hr at RT. 2.2.5 Bimolecular fluorescence complementation (BiFC) assay µ - plates (Ibidi , cat # 80826 ) were used for confocal microscopy to detect Venus fluorescence. / MEF cells were seeded at a density of 3×1 0 6 cells/mL in each well and co - transfected with VN155 - - - - M440R / W457A / S450A) plasmids. Cells were transfected as described above and then treated with BSA or PA media for 2 h. The Olympus FluoView FV1000 imaging system, with an UPLFLN 20X NA 0.5 objective and a PLAPON 60X O NA 1.42 objective, was ex em = 535nm. - FL was performed for the - LD constructs using the VN155 - - FL and VC155 - - FL (WT or W457A/S450A/W457A - S450A) plasmids. The corrected total cell fluorescence (CTCF) 20 was calculated for each cell (McCloy et al., 2014) using the following equation: CTCF = [Integrated Density (Area of selected cell × mean fluorescence of background readings)]. The number of cells that contain at least one foci was analyzed using ImageJ (Fiji) software with the following parameter settings, threshold 15 - 255, circularity 0 - 1, and size (pixel 2 ) 100 - 10000. 2.2.6 XBP1 splicing assay / MEF cells were reverse transfected with the plasmids as described above. Total mRNA was collected using the Qiagen RNeasy Plus kit (Qiagen, cat# 74134) and the one - step RT - PCR kit (Qiagen) was used to amplify XBP1 cDNA. PCR primers were designed to flank the 26 - bp splicing sequence of XBP1. PCR reaction was performed using the GAACCAGGAGTTAAGAACACG - AGGCAACAGTGTCAGAGTCC for 98 o C for 2 mins followed by 30 cycles of 98 o C for 30s, 61 o C for 15s, 72 o C for 30s and a final extension of 72 o C for 3 mins. The PCR products were run on a 2.5% agarose gel made in 0.5X TBE. Gels were visualized and quantified using Molecular Imager ChemiDoc XRS System (Bio Rad). The splicing efficiency of XBP1 (% XBP1) was calculated with the following equation: XBP1 splicing (%) = [Intensity (XBP1 S ) + 0.5*Intensity (XBP1 HD )] / [Intensity (XBP1 S ) + Intensity (XBP1 HD ) + Intensity (XBP1 U )]*100 2.2.7 Tryptophan fluorescence measureme nts and tryptophan depth calculation For nitroxide lipid quenching experiments, the small unilamellar vesicles (SUVs) were prepared by a modified method from a previous study (Maiorano and Davidson, 2000). The spin labels used, 5 doxyl PC (1 - palmitoyl - 2 - st earoyl - (5 - doxyl) - sn - glycero - 3 - 21 phosphocholine) and 16 doxyl PC (1 - palmitoyl - 2 - stearoyl - (16 - doxyl) - sn - glycero - 3 - phosphocholine) were obtained from Avanti Lipids. The molar ratio of POPC: doxylPC: peptide was 68:12:1 for all the assays. The lipid and peptide mixture in chloroform was dried under nitrogen, vacuum dessicated for 8 h to remove all traces of chloroform. 10 µL ethanol was further added to the lipid - peptide film. The sample was then vortexed and incubated on ice for 10 min. After diluting the sample with 1 mL of 10 mM Tris - HCl buffer (pH 7.0), bath sonication was performed for 1 h. The spectra were recorded at room temperature on a Fluoromax - ex = 280nm. Emission spectra were collected from 295nm to 450nm. The fluorescence from POPC SUVs was subt racted from the samples containing spin label and peptide. The distance of the Trp from the center of the bilayer (Z cf ) is calculated with the following equation (Ren et al., 1997): where F 0 is the fluorescence intensity in the absence of quencher, F s is the fluorescence intensity in the presence of the shallow quencher (5 - doxyl PC), F d is the fluorescence intensity in the presence of the deep quencher (16 - doxyl PC), L cd is the distan ce from the center of the bilayer to the deep quencher (2.25 Å), and L ds is the distance between the shallow (12.15 Å) and deep quenchers. 22 2.3 Results and Discussion 2.3.1 Palmitate activates ER stress S aturated fatty acid s , palmitate, are known to and this is independent of the LD (Kitai et al., 2013; R. Volmer et al., 2013) . There is no higher order clustering seen indicating oligomerization indica ting a potentially separate mechanism of (Kitai et al., 2013) through direct regulation of the conformation of - TMD. 2.3.2 To directly monitor palmitate - induced dimerization in vivo , we employed a BiFC assay that has been successfully used to quantify dimerization of both the seven transmembrane domain (7TM) recept or and the receptor tyrosine kinases (RTKs) family proteins in vivo (Kosel et al., 2010; Tao & Maruyama, 2008) . Figure 2 - 1 BiFC assay for dimeriz ation of IRE1 - TM (taken with permission from (Cho et al., 2019) ) . - - tagged Venus N - terminus fragment and HA - tagged Venus C - t erminus fragment were inserted between the signal peptide (SP) and the truncated LD. The amino acid residue numbers of the human Schematic diagram of the BiFC assay. Homodimerizat - functional Venus fluorescent protein. W - t o investigate - TMD . The N - terminal Venus 23 fragment (Myc - tagged VN ) and the C - terminal Venus fragment (HA - tagged VC) were inserted downstream of the signal peptide (1 - - Figure 2 - 1 A ) (Shyu, Liu, Deng, & Hu, 2006) . Once bo th VN - - - - - close proximity in the ER membrane, the Venus fragments will be able to reconstitute a native and fluorescent protein ( Figure 2 - 1 B ). We employed the BiFC assay in - / - MEF cells. These cells have a negligible . The BiFC assay helped us monitor t he early response of palmitate - induced dimerization in - / - MEF cells transfected with V N - - - - - / - . We first confirmed that the expression levels - and palmitate - - / - MEF cells were comparable to the parental MEF - WT cells ( Figure 2 - 2 A and B ). Figure 2 - 2 - - FL . - - / MEF. MEF or / MEF cells were seeded at a density of 3×10 6 / MEF cells were transfected with VN - - - IR - FL. After treatment with 2% BSA or 0.4 mM PA for 2 h, cell extracts were collected and subjected to western blot analysis. GAPDH served as a loading control. (B) Quantification of (A). (Figure taken with permission from (Cho et al., 2019) ) 24 We also measured changes in activation of the PERK pathway due to transfection and expression of - constructs. T he exogenous - alter the palmitate - a downstream factor of the other UPR sensor PERK ( Figure 2 - 3 ). Figure 2 - 3 (Left) The cell extracts obtained from Figure 2 - 2 were used for western blot analysis to Quantification of the blots. The phos - sample. Values are mean ± SE (n=2). (Figure taken with permission from (Cho et al., 2019) ) Figure 2 - 4 shows the results on palmitate treatment for 2 hrs. Dimerization of the VN - - - WT and VC - - - WT proteins followed by reconstitution of the Venus fluorophore results in a significant increase in fluorescence intensity. 25 Figure 2 - 4 - . - / - MEF cells co - expressing VN - - - - and S450A mutant constructs were treated with 2% BSA media, 0.4 mM palmitate (PA) or 0.4 mM oleate dissolved in 2% BSA media, or 5 µg/mL tunicamycin (Tm) for 2 h. The Figure 2 - 5 - and S450A mutant proteins. - with BSA or 0.4 mM PA for 2 h. GAPDH was used as a loading control. (B) RT - PCR of Xbp1 mRNA from cells transfected with VN - - - - W457A and S450A mutant constructs. Cells were treated with 2% BSA or 0.4 mM PA for 2 h, and total mRNA isolated from the cells was subjected to RT - PCR analysis. The 26 Figure 2 - unspliced (U) and spliced (S) products and heteroduplex species (HD) are indicated. GAPDH was used as a loading control. (G) Quantification of spliced XBP1 (XBP1S). The fraction of spliced Xbp1 mRNA in each sample was quantified as described - test). (Figure taken with permission from (Cho et al., 2019) ) Treatment with the m onounsaturated fatty acid, oleate (OA), failed to induce the - Figure 2 - 4 ), suggesting that the induced dimerization of - - treated cells is specifically due to lipid saturation. Tunicamycin, an ER stress inducer that perturbs protein folding i n the ER lumen also did not induce dimerization of - constructs ( Figure 2 - 4 ). This was consistent with the results seen by Volmer et al . (Romain Volmer et a l., 2013) . All these lines of evidence suggest that - in vivo is specific to the increased lipid saturation due to palmitate. This - - WT was further corroborated by the enhanced XBP1 splicing effic iency ( Figure 2 - 5 ). 2.3.3 - TMD are involved in dimerization W e performed alanine - - TMD to systematically - . 27 Figure 2 - 6 - TMD . - / - MEF cells co - expressing alanine mutant VN and VC plasmids were treated with BSA or 0.4 mM PA for 2 h. The fold change of Venus fluorescence between the BSA - and PA - treated cells was quantified and nor malized to WT. Values are reported as mean ± SE (n = 2) (Figure taken with permission from (Cho et al., 2019) ) All resid ues, except for two native alanines, were mutated to alanine in the VN - - - - Figure 2 - 1 ). B oth S450A and W457A were not able to reconstitute the fluorescence in the palmitate - treated cells, indicating an inability to dimerize in the ER membrane ( Figure 2 - 4 and Figure 2 - 6 ). The loss of palmitate - - the W457A and S450A mutants resulted in a loss of XBP1 splicing ability ( Figure 2 - 5 B and C ). The levels of protein expression for the mutants were similar to that of the WT and ( Figure 2 - 2 A , Figure 2 - 5 A , respectively ). We confirmed that the mutants primarily colocalize with the ER membrane marker protein, calnexin, in the BSA and palmitate - treated Hep3B cells ( Figure 2 - 7 A and B ) . The obtained Pearson's correlation coefficients for W457A and S450A were not significantly different from the 28 WT. Therefore, our in vivo data demonstrate an essential role of both Trp457 and Ser450 in modulating the palmitate - - LD. 2.3.4 - TMD serves as a sensor for lipid membrane saturation in mammals Results of a previous study with the yeast Ire1 showed that the juxta - membrane amphipathic helix, but not its TMD, plays a crucial role in sensing aberrant physical membrane propert ies (Halbleib et al., 2017) . Since the amphipathic helix is conserved on (Halbleib et al., 2017; Kono et al., 2017) , we further assessed whether - activation by membrane lipid saturation. 29 Figure 2 - 7 - cts. (A) Hep3B cells were transfected with Myc - tagged VN - - - WT and its corresponding mutants (W457A, S450A, V437R, and M440R) and then treated with 2% BSA or 0.4 mM PA (in 2% BSA) media for 2 h. The cells were fixed and stained with rabbit anti - Myc tag and mouse anti - calnexin antibodies. Images were taken on an Olympus - condition were analyzed using Image J software. Data indicate the mean ± SD (n = 10). (Figure taken with permission from (Cho et al., 2019) ) 30 Figure 2 - 8 BiFC assay with V437A and M440R mutations. (A) - / - MEF cells co - expressing VN - - - - V437R and M440R mutants were treated with 2% BSA or 0.4 mM PA for 2 h. Venus B ) Quantification of Venus fluorescence in the presence of BSA or PA. The CTCF values were calculated for each treatment and mutation and the fold changes upon PA treatment were then normalized - - WT sample. Values are reported as mean ± SD (n = 2), **p < 0.001 - test). Figure 2 - 9 - and M440R mutant proteins. - a loading control. (B) RT - - / - MEF cells co - transfected with VC - - - WT and VN - - - WT or their mutants, V437R and M440R. BSA and PA treatment conditions are the same as ( Figure 2 - 8 A). (C) Quantification of - - WT, V437R and M440R mutants in the presence of BSA or PA. Values are reported as mean ± SD (n = 3 - test). (Figure taken with permission from (Cho et al., 2019) ) 31 To further investigate this , we mutated two conserved hydrophobic residues (V437 and M440) on the amphipathic helix to arginine an d performed the BiFC assay ( Figure 2 - 8 ). In contrast to the yeast Ire1 (Halbleib et al., 2017) , the two amphipathic helix mutations did not affect the dimerization ability - ( Figure 2 - 8 A and B ). Furthermore, the amphipathic helix mu tant proteins exhibited WT - like XBP1 splicing activity ( Figure 2 - 9 B and C ). V437R and M440R helix mutant proteins had similar expression levels as the WT protein ( Figure 2 - 9 A ), and they also co - localized with calnexin in both the BSA and palmitate - treated cells ( Figure 2 - 7 A and B ). In corroboration with the study by Kono et al., our in vivo results strongly suggest than the yeast Ire1 (Kono et al., 2017) , 2.3.5 - LD in dimerization W e used full - - - FL and VC - - FL ( Figure 2 - 10 - / - MEF cells. Including the LD will help us exclude the effect of LD on dimerization of the TMD. Figure 2 - 10 - TMD sense lipid aberrancy and - XBP1 pathway. - length (FL) constructs used in the BiFC assay. The Myc - tagged Venus N - terminus frag ment and HA - tagged Venus C - terminus fragment were inserted between SP and LD ) . (B) Western blot of the WT and mutant samples with indicated antibodies. (Figure taken with permission from (Cho et al., 2019) ) 32 Full - - / - and treated with BSA or PA media. D imerization of full - was evident by the increase in the reconstituted green fluorescence ( Figure 2 - 11 A ). Ass seen with the - constructs, ( Figure 2 - 4 - FL reduced the dimerization in the palmitate - treated cells significantly (p<0.001) ( Figure 2 - 11 B and D ). Figure 2 - 11 - FL con s tructs (A ) - / - MEF cells co - expressing VN - - WT - FL and VC - - W T - FL, or their W457A and S450A mutants, were treated with 2% BSA or 0.4 mM PA for 2 hr. Venus (B) RT - PCR of Xbp1 - / - MEF cells co - transfected with VC - - FL - WT and VN - - FL - WT or their mutants W457A and S450A. Cells were treated with 0.4 mM PA for 0, 0.5, 1, 2 h. The unspliced (U) and spliced (S) products and heteroduplex species (HD) are indicated. (C) Quantification of Venus fluorescence was detected with confocal - change upon PA treatment was normalized to WT. Values are reported as mean ± SD (n = 3 - test). (D) Quantification of - FL - WT, W457A, and S450A mutants in (B). - test). (Figure taken with permission from (Cho et al., 2019) ) 33 - FL variants are similar to the endogenous - WT cell ( Figure 2 - 2 , Figure 2 - 5 , and Figure 2 - 10 ), the reduced dimerization in the W457A and S450A mutants is not likely to be due to non - - FL variants. Figure 2 - 12 - FL constructs . (A) Hep3B cells were transfected with Myc - tagged VN - - FL - WT and its corresponding mutants (W457A and S450A) and then treated with 2% BSA or 0.4 mM PA for 2 hrs. The cells were fixed and stained with rabbit anti - Myc tag and mouse anti - calnexin antibodies. Images were collected using an Olympus FV10 confocal microscope (with - FL variants and calnexin colocaliza tion. Ten cells for each condition were analyzed using Image J software. Data indicate the mean ± SD (n = 10). (Figure taken with permission from (Cho et al., 2019) ) 34 hese full - - localized with the ER membrane protein, calnexin, were very similar indicating proper localization ( Figure 2 - 12 A and B ). B - FL - - FL - S450A mutants in the palmitate - treated cells showed significantly reduced XBP1 S fractions ( Figure 2 - 11 B and D ). This suggests that both Trp457 and Ser450 are important not only t could activate XBP1 splicing. 2.3.6 Effect of Y161A on BiFC dimerization W e used an established Y161A mutation that is defective in sensing unfolded proteins to separate the effect of lipid perturbation from the effect of accumulation of unfolded proteins upon palmitate treatment (Kono et al., 2017) . Figure 2 - 13 Effect of Y161 mutation on XBP1 splicing . (A) RT - - / - MEF cells transfected with pcDNA - - FL - Y161A, pcDNA - - FL - Y161A - W457A, and pcDNA - - FL - Y161A - S450A. Cells were treated with BSA or 0.4 mM PA for 2 hrs. (B) Quantification of splicing - FL - Y1 - FL - Y161A - - FL - Y161A - S450A. - test). Error bars may not be visible due to low values. (Figure taken with permission from (Cho et al., 2019) ) 35 - Figure 2 - 5 ), the XBP1 splicing - FL - Y161A - - FL - Y161AS450A were - FL - Y161A ( Figure 2 - 13 A and B ). 2.4 Trp457 locates near the center of the POPC bilayer. Tryptophan is found abundantly in membrane proteins. Due to its amphipathic nature, tryptophan is known to stabilize the membrane proteins by anchoring the interactions in the membrane - water interface (Yau, Wimley, Gawrisch, & White, 1998) . Studies have shown that tryptophan often locates toward the center of the TMD helix (Haeger et al., 2010; Ridder, Skupjen, Unterreitmeier, & Langosch, 2005; Sal - Man, Gerber, Bloch, & Shai, 2007) . The tryptophan residue support s the aromatic interaction, which mediates t he TMD dimerization . Figure 2 - 14 Trp457 localizes in the center of the small unilamellar vesicles (SUVs). - TMD peptide in small unilamellar vesicles (SUVs). The SUVs containing TMD peptide, POPC, and spin - labeled shallow quencher (5 Doxyl 36 Figure 2 - PC - blue) or deep quencher (16 Doxyl PC - red) were prepared as descr ibed in the Methods. The tryptophan fluorescence was excited at 280 nm and the emission spectra were obtained by scanning from 295nm to 450nm. (Figure taken with permission from (Cho et al., 2019) ) To experimentally determine the location of Trp457 in the POPC lipid bilayer, we used the parallax method t o estimate the membrane penetration depth of the tryptophan residue (Chattopadhyay & London, 1987) . The par allax method allows the determination of the apparent location of tryptophan by comparing the quenching of the tryptophan fluorescence by two different spin - labeled POPC lipids, 5 - Doxyl PC (shallower quencher), and 16 - Doxyl PC (deeper quencher). Compared t o the fluorescence intensity in the absence of the quenchers ( Figure 2 - 14 , black), Trp457 was quenched dramatically by the deeply located nitroxide lipid quencher (F/F 0 = 0.099) ( Figure 2 - 14 , red), but much less efficiently by the shallow quencher (F/F 0 = 0.71) ( Figure 2 - 14 , blue). We calculated the distance of Trp457 from the center of the bilayer (Z cf ) as described in the Methods. The average depth of penetration of Trp457 from the center o f the bilayer was estimated at 0.88 ± 0.35 Å . Collectively, our in vivo results suggest that the intrinsic structural elements of the - TMD could be a key factor in sensing membrane lipid saturation in mammals. 37 3 EXPRESSION AND PURIFICATION OF A MAMMA LIAN SF21 CELLS 3.1 Introduction 3.1.1 - CD with PA show specific residues involved in PA binding Previous research from our lab investigated the effect of binding of PA to PKR (Protein kinase R) , a serine - threonine kinase protein (Cho et al., 2011) . This study showed that PA binds to PKR in the presence of ATP and prevents phosphorylation and activation of PKR . The binding site for fatty acids on PKR is at the ATP binding site. Figure 3 - 1 Alignment of IRE1 and PKR. IRE1 (depicted in blue) and PKR (depicted in red) were aligned using PyMol align. The root mean square deviation ( RMSD ) value is 1.538 Å . Alignment of IRE1 and PKR . IRE1 (depicted in blue) and PKR (depicted in red) were aligned using PyMol align. The RMSD value is 1.538 Å. 38 PKR shares a lot of structural similarities with IRE1 (K. P. K. Lee et al., 2008) . An - CD and PKR is shown in Figure 3 - 1 . - CD coordinates were obtained from RCSB PDB (PDB id: 5HGI). The PKR structure was obtained by homology modeling of the PKR protein sequence (NCBI accession no: P19525) in CPHmodels - 3.2 (Nielsen, Lundegaard, Lund, & Petersen, 2010) . The root mean square deviation ( RMSD ) - CD may also show similar binding to PA. To investigate this, in collaboration with Dr. Michael Feig, we decided to employ molecular dynamics si - CD that would bind to PA. For the simulation, the Feig group used a homology model derived from the active form of yeast Ire1(RCSB PDB: 3FBV) as well as a crystal structure of the inactive form of human IR - activation loop and as well as the RNase dimer interface ( Table 3 - 1 ) predicted through molecular dynamics ( MD ) simulations of both active and inactive IRE1 models. Table 3 - 1 - CD predicted to be involved in PA binding . %time Residue no. Active IRE1 Inactive IRE1 R611 27.61 8.54 R687 14.21 - R722 12.12 15.22 R727 4.69 23.25 R864 17.58 15.85 R887 15.11 17.95 K893 7.5 - R912 6.68 4.17 In this study, we developed a protocol to express the Inositol Requiring Enzyme - 39 protein primarily composed of three domains, the luminal domain, a transmembrane domain and cytosolic domain (CD) (Schröder & Kaufman, 2005c) . It is a bifunctional protein with kinase and endoribonuclease activity on the CD. In this chapter, we - CD binds to PA at the residues predicted by MD simulations in Table 3 - 1 . To answer this qu e stion, - CD (547 - 977aa) in insect Sf21 cells using ba E. coli expressed in E. coli and yeast systems is in a hyperphosphorylated state (Prischi, Nowak , Carrara, & Ali, 2014) . 3.1.2 Types of heterologous protein expression systems Heterologous expression of proteins is defined as the expression of proteins in an organism that does not naturally express the protein. The different systems in use for such pur poses can be broadly classified as bacterial, yeast, insect and mammalian systems. Cells in culture can be manipulated to express the protein of interest in transient or stable mode. Some factors that need to be considered while choosing the expression sys tem are the required yield of the protein, post - translational modifications, functionality, and speed of expression. Proteins of prokaryotic origin are best expressed in bacterial systems. E. coli can express and adequately process prokaryotic proteins an d is the go - to system for cheap, scalable and high yield expression. If the expression of eukaryotic proteins is required, additional factors need to be taken into consideration. The most important factor would arguably be the level of post - translational m odifications (PTM) required for a functional form of the protein. Proteins requiring extensive PTMs will not be processed correctly in 40 bacterial systems and will most likely aggregate in inclusion bodies (Rinas & Bailey, 1992; Ventura, 2005; Wingfield, 2016) . Adding a fusion protein tag to the protein of interest may sometimes help to resolubilize the proteins. However, in the interest of saving time and effort, it is advisable to switch to higher eukaryotic systems such as insect or mammalian cells . Another advantage to eukaryotic systems is the presence of extensive protein folding machinery that is vital for the function of the protein (Shi & Jarvis, 2007) . Table 3 - 2 Characteristics of heterologous protein expression in bacterial, yeast, insect and mammalian systems ranked according to desirability . Table 3 - 2 compares different systems in terms of the type of protein, post - translational modifications, and ease of large - scale production (R. Chen, 2012; X. Gao, Yo, & Harris, 2005; Gupta, Jaisani, Pandey, & Mukherjee, 1999; Huang et al., 2007; Müller, Sandal, Kamp - Hansen, & Dalbøge, 1998; Sør ensen & Mortensen, 2005) . 3.1.3 Choosing the right expression system Bacterial systems have desirable characteristics for large scale production but are not able to process and correctly fold proteins requiring extensive PTMs (Rosano & 41 Ceccarelli, 2014) . Depending on the application of the protein, Chinese hamster ovary (CHO) or Human embryonic kidney (HEK 293) cells are typically the system o f choice for eukaryotic proteins (Croset et al ., 2012) . This is true, especially for therapeutic proteins or antibodies. However, the yield is very low compared to insect or bacterial cells typically on the order of 100mg/L of expressed protein (Subedi, Johnson, Moniz, Moremen, & Barb, 2015) . The level of protein expression is higher in insect cells ranging from 100mg 1g/ L (Merlin, Gecchele, Capaldi, Pezzotti, & Avesani, 2014; Van Oers, P ijlman, & Vlak, 2015) . Insect cells and mammalian systems have a wide range of post - translational modifications. An excellent tool to determine all possible post - translational modifications of a protein of interest is dbPTM ( http://dbptm.mbc.nctu.edu.tw/ ). Insect and mammalian systems are both capable of all PTMs with one exception. Insect cells and mammalian cells have similar phosphorylation and O - linked glycosylation patterns. They can authentically proce ss phosphorylation modifications, partly due to the presence of phosphatases. However, in terms of N - glycosylation, proteins in insect cells are high - mannosylated while mammalian cells have a complex glycosylation pattern (Croset et al., 2012; Shi & Jarvis, 2007) . Figure 3 - 2 depicts the difference in N - linked glycosylation patterns in insect cells vs. mammalian cells. Varied and complex sugars such as N - acetylneuraminic acid, galactose, fucose, mannose are added in mammalian cells in a branched configuration as opposed to the addition of only mannose residues in insect cells. 42 Figure 3 - 2 N - linked glycosylation patterns for insect cells and mammalian cells . Post - translational modifications for IRE1 were predicted using algorithms developed by Center for Biological Sequence Analysis, Techni cal University of Denmark ( http://www.cbs.dtu.dk/ ). The predicted N - glycosylation, O - glycosylation, C - mannosylation and glycation sites are scored for the likelihood of modification ( Table 3 - 3 ). - linked - CD. 43 Table 3 - 3 Predictions for O - glycosylation, C - mannosylation, N - glycosylation sites and glycation sites on CD - IRE1 . The prediction algorithms used are hosted by Centre fo r Biological Sequence Analysis, Technical University of Denmark 44 3.2 Experimental Design 3.2.1 Materials All the materials are listed according to what is needed for each sub - section of the procedure. 3.2.1.1 - CD into pFastBac plasmid i. pFastBac His6 MBP N10 TEV LIC cloning vector (4C) (Addgene plasmid #30116) ii. pFastBac His 6 MBP N10 TEV LIC cloning vector (4C) (Addgene plasmid #30116) iii. SspI - HF (New England Biolabs, cat # R3132S) iv. CutSmart buffer (New England Biolabs cat# B7204S) v. QIAquick PCR Purification Kit (Qiagen, cat# 28104) vi. Deoxynucleotide (dNTP) set includes dGTP, dCTP ( New England Biolabs, cat # N0446S) vii. Bovine Serum Albumin (BSA), Molecular Grade (New England Biolabs, cat# B9000S) viii. T4 DNA Polymerase (New England Biolabs, cat# M0203S) ix. Q5 High Fidelity 2X Master Mix (New England Biolabs, cat# M0492S) x. OneShot Top10 chemicall y competent E. coli C404003) xi. Luria Broth (Sigma - Aldrich, cat# L3397) xii. Luria Agar (Sigma - Aldrich, cat# L3272) xiii. LB agar plates containing ampicillin (Sigma - Aldrich, cat # L5667 - 10EA) 3.2.1.2 Preparation of recombinant bacmid xiv. MAX Efficiency E. coli DH10Bac competent cells (Thermo Fisher cat# 10361012) 45 xv. Antibiotics ( Table 3 - 4 ) Aliquot and store at - 20 o C. Table 3 - 4 Antibiotic concentrations and stock solutions . Component Dissolve in Stock conc . Special considerations Kanamycin Water 50mg/mL Tetracycline Ethanol 10mg/mL Light sensitive Gentamicin Water 7mg/mL X - gal** Dimethylformamide 20mg/mL Light sensitive, DO NOT filter sterilize, Make solution in a glass vial or polypropylene tube IPTG Water 200mg/mL Filter sterilize ** There are other options for X - gal such as Bluo - gal. It is more expensive but more sensitive and turns a deeper, more obvious blue that aids in colony identification** xvi. SOC medium (Thermo cat# 15544034) xvii. PureLink HiPure Plasmid Miniprep Kit (Invitrogen, cat# K210002) 3.2.1.3 Transfection of recombinant bacmid into Sf21 cells i. Sf900 - III serum - free media (Thermo, cat# 12658019) ii. Cellfectin II reagent (Thermo, cat# 10362100) iii. Gelcode Blue Stain Re agent (Thermo, cat# 24590) iv. v. Xtractor cell lysis buffer (Takara, cat# 635671) vi. cOmplete , Mini, EDTA - free protease inhibitor cocktail (Sigma, cat# 11836170001) vii. MBPTrap HP column (GE, cat# 28 - 9187 - 78) viii. D - (+) - Maltose monohydrate (Sigma, cat# 63418 - 25G) ix. Phosphate buffered saline (PBS) (Sigma, cat# P7059 - 1L) x. Binding buffer for MBPTrap HP column: 20 mM Tris - HCl, 200 mM NaCl, 1mM DTT, 1 mM EDTA, pH 7.4 46 xi. Elution Buffer for MBPTrap HP column: Bindin g buffer + 10mM maltose xii. Regeneration buffer: 0.5M NaOH Figure 3 - 3 Flowchart for all the steps involved in the expression and purification of proteins from Sf21 cells . 47 3.3 Procedure Figure 3 - 3 depicts a flowchart for all the steps involved in the expression and - CD in Sf21 cells using baculoviral vectors. The proc edure is divided into 5 subunits. The time required for each subunit is indicated in brackets in the left column. 3.3.1 - CD into pFastBac plasmid Note: pFastBac His6 MBP N10 TEV LIC cloning vector (4C) was a gift from Scott Gradia (Addgene plasmi d #30116; http://n2t.net/addgene:30116 ; RRID: Addgene_30116) Another version of this plasmid (5C) is available with 2 ORFs so that two proteins can be - CD was o btained from NCBI (Gene id: 2081). Primers were designed to clone the CD fragment from 547aa - 977aa from a pCDNA - hIRE1 plasmid containing a full - length IRE1 coding sequence. Table 3 - 5 PCR primers for insertion of IRE1 - CD (547aa - 977aa) into pFastBac plasmid . CD_IRE1_forward - TACTTCCAATCCAATGCA GGCAGCAGCCCCTCCCTGGAAC - CD_IRE1_reverse - TTATCCACTTCCAATGTTATTA GAGGGCGTCTGGAGTCAC - Table 3 - 5 shows the primers designed for the insertion - CD into the empty pFastBac plasmid. The underlined tag on each primer is to facilitate ligation independent cloning (LIC) and is specific to the pFastBac backbone. The rest of the primer is specific to the insert coding sequence (CDS). This technique makes use of the endonuclease activity of T4 DNA polymerase to generate sticky overhangs for ligation between the vector plasmid and insert DNA. This technique avoids the use of ligases. 48 Protocol: 3.3.1.1 Preparing the vector pFastBac plasmid i. Linearize the plasmid with SspI - HF (NEB) ( Table 3 - 6 ). The high - fidelity versions of restriction enzymes are faster and more efficient. Incubate at 37 o C for 15mins 2hrs to completely digest all plasmid DNA. Heat inactivation is done for 20 mins at 65 o C. Table 3 - 6 Restriction digestion with SspI . Component Amount Volume pFastBac plasmid DNA (250ng/uL) 1 ug 4 uL 10X CutSmart buffer Final conc 1X 5uL SspI - HF 10 units 1uL Water Make up to 50uL 40uL ii. Use a PCR purification kit (Qiagen) to purify the large linearized fragment of plasmid DNA. iii. To create pFastBac vector overhangs, mix the following components in PCR tubes, the order of addition is as listed (T4 DNA polymerase is added last) ( Table 3 - 7 ). Incubate at 12 o C for 30 minutes followed by heat inactivation at 75 o C for 20 minutes. Table 3 - 7 Mix to create pFastBac vector overhangs . Reagent Final conc . Volume to use 10x NEB 2.1 1x 4 ul Eluted linearized pFastBac DNA 10 - 50 ng/ul 20 - 30 ul dGTP (100mM) 2.5mM 1 ul DTT (100mM) 5mM 2 ul BSA (10 ug/ul) 0.25 ug/ul 0.6 ul T4 DNA polymerase 0.075 units/ul 1 ul Sterile dH2O to 40ul 49 3.3.1.2 - CD iv. - CD insert using the primers outlined in Table 3 - 5 was set up ( Table 3 - 8 ) . Amplification was confirmed by running the PCR product on a 1.5% ag arose gel. Table 3 - 8 - CD insert . Component Volume Q5 2X Master Mix 25uL CD_IRE1_forward primer (10uM) 2 uL CD_IRE1_reverse primer (10uM) 2uL pCDNA - hIRE1 (50ng/uL) 1uL Water ( make up to 50uL) 20uL The PCR program should be run for 98 o C 1 min, followed by 35 cycles of 98 o C for 30 sec, 61 o C for 10sec, 72 o C for 90sec, and a final extension of 72 o C for 5 mins. v. Insert overhangs were created in a way similar to creating vect or overhangs in the presence of dCTP (instead of dGTP). vi. After heat inactivation, both vector and insert were mixed together (1:3 ratio). The total volume of the combined vector and insert should be less than 5 uL, at most 10 uL. Incubate for 5 mins at roo m temperature and then add 1uL of 25mM EDTA followed by another incubation for 5 mins. vii. 2 uL of this mixture was transformed into competent E. coli onto LB agar plates containing 100ug/mL ampicillin and incubated at 37 o C for 16 - 24 hrs. viii. Individual colonies were selected, and plasmid DNA was prepped. Insertion of - CD was confirmed by DNA sequencing. 50 3.3.2 Preparation of recombinant bacmid Once the insert has been cloned into the pFastBac empty plasmid, it needs to be transformed into E. coli DH10Bac cells to form the recombinant bacmid. E. coli DH10Bac competent cells are sold by Thermo Fisher. The genotype is F - mrr - hsdRMS - - rps L nup G /pMON14272 / pMON7124. DH10Bac cells have a baculovirus shuttle vector and a helper plasmid. This machinery is required for the generation of recombinant bacmid after transformation of pFastBac - - CD into the cells. The baculovirus shuttle vector (bMO N14272) also encodes kanamycin resistance, and the helper plasmid (pMON7142) use of blue/white colony screening to determine integration of the insert into the bacmid. pF astBac has the Tn7 element which includes the polyhedrin promoter, the gene of interest and gentamicin resistance. It is also possible to use regular E. coli DH10Bac cells and make them chemically competent. Protocol : i. Dissolve LB agar powder in water and autoclave and l et it cool to 55 o C ii. Add all the antibiotics and chemicals in the w orking concentrations ( Table 3 - 9 ) : Table 3 - 9 Working concentrations of antibiotics . Component Stock conc. Working conc. Dilution Kanamycin 50mg/mL 50ug/mL 1:1000 Tetracycline 10mg/mL 10ug/mL 1:1000 Gentamicin 7mg/mL 7ug/mL 1:1000 X - gal 20mg/mL 200ug/mL 1:100 IPTG 200mg/mL 40ug/mL 1:200 51 iii. Let harden, store at 4 o C in the dark. iv. Use only competent E. coli DH10Bac cells. v. Thaw cells on ice in a 15mL round bottom sterile polypropylene tube. vi. Add appropriate volume of DNA to the vial . vii. Incubate on ice for 30 mins. viii. Heat shock for 45s at 42 o C. ix. Transfer tubes back on ice and chill for 5 mins. x. Add 900uL SOC medium, xi. Incubate tubes in a shaking incubator, 37 o C at 200rpm for at least 4 hrs. *This 4 hr long outgrowth step is necessary to allow the bacteria to generate the antibiotic resistance proteins encoded on the plasmid backbone* xii. Make 3 10 - fold serial dilutions of 8 00uL cells. Plate 100uL of each dilution onto LB plates with all the selection antibiotics and X - gal/IPTG. xiii. Incubate plates at 37 o C for 48 hrs. between white and blue colonies* *Pick colonies that are large and well isolated . Avoid picking colonies that are gray or darker in the center as they may contain a mixture of empty bacmid and recombinant bacmid* xiv. Replate selected colonies for an additional round of selection on fresh LB a gar plates (with kanamycin, gentamicin, tetracycline) overnight at 37 o C. 52 xv. Once the white phenotype is confirmed, inoculate in LB media with kanamycin, gentamicin, and tetracycline and grow overnight. xvi. Isolate recombinant bacmid DNA using Purelink HiPure Plas mid Kit by Invitrogen with a modified protocol. 3.3.3 Transfection of recombinant bacmid into Sf21 cells Note: Sf21 cells are suspension cells grown at 28 o C, without the need for CO 2 incubators. They are grown in Sf - 900 III media. The media can be supplemented, if needed, with serum but bovine serum should not be used (fetal calf serum is preferred). The serum needs to be inactivated to inactivate complement fragments that can inactivate baculoviruses. Transfection reagents can be lipid based. Calcium chloride ca n be used for baculovirus transfection, but it has less efficiency than lipid - based reagents. Protocol: 3.3.3.1 Transfection of bacmid into Sf21 cells i. A day before transfection, passage Sf21 cells so that they are at a density of 3 x 10 5 cells/mL. *Actively dividing cells have a higher transfection efficiency and produce more protein* ii. Plate cells at a concentration of 8 x10 5 cells/well of a 24 - well plate. iii. Allow the cells to attach for 1 hr and replace media with fresh media (without FBS). iv. Use 2ug of bacmid DNA per well. Dilute in 75uL of SFM900 - III media. Incubate for 15 mins. v. Use 8uL of cellfectin II in 75uL media. Incubate for 15 mins. 53 *Make sure the cellfectin tube* vi. Mix together the two tube s and incubate for 15 - 30 mins. vii. Add dropwise to the well. viii. 24 hrs later, replace transfection media with fresh media containing 10% FBS. *Serum proteins in FBS act as substrates for proteases* ix. Check for signs of infection (SIF) after 24 hrs up to 96 hrs. *SIF may not be obvious in P0 infection so continue infection of 5 days and collect the supernatant and re - infect to amplify the viral stock* x. After changing media, collect supernatant after every 24 hrs for 24,48,72,96 hrs. This supernatant is the P0 vira l stock. Store at 4 o C in the dark. xi. Use the P0 viral stock to infect newly plated Sf21 cells to generate P1 stock. Add 150ul of P0 stock dropwise on top of the cells (plated in a 6 - well plate) , gently swirl a few times and incubate the plate at 28 o C. xii. Look for SIF in 24 hours post infection time. After 5 days, collect only P1 viruses with SIF and store viral stocks in the dark at 4 o C. xiii. In a similar fashion, collect P2, P3 and P4 stocks for increased baculoviral titer while an increasing number of cells infected (6 - well plate 150cm dish T25 flask). Store all at 4 o C protected from light. 3.3.3.2 Scale - up of heterologous protein expression i. Seed cells in a T75 (vented) flask as described in the transfection procedure. Make sure they are actively dividing cells i n the log phase of growth. ii. Add 150uL of the collected P1 stock. iii. Incubate at 27 o C for 5 days or until 30 - 40% of the cells have lysed. 54 iv. Collect cells as well as supernatant. This will be the P2 viral stock. v. Use P2 viral stock to infect spinner flasks with Sf 21 cells. vi. If no signs of infection are observed, infect Sf21 cells with the P2 stock of baculoviruses and collect the P3 viral stock. 3.3.4 Purification of MBP fusion protein Sf21 cells were infected with P3 stock of baculovirus for 96hrs. The cell pellet was collected and lysed with Xtractor buffer, and an MBPTrap column was used to extract MBP - - CD. Procedure: i. Prepare Xtractor buffer with protease inhibitor of choice. In this purification protocol, we used cOmplete EDTA free protease inhibitor c ocktail. ii. Pellet Sf21 cells, remove supernatant media and wash pellet once with phosphate buffered saline (PBS). Add 5mL Xtractor buffer to 1g cell pellet. iii. Mix vigorously by vortexing and shake at room temperature for 15 - 30 mins. Make sure that cell lysis is complete by observing under the microscope. iv. Centrifuge at 16,000 g for 20 - 30 mins to pellet the cell debris. v. Take the clarified supernatant and filter through a 0.22µm filter immediately before loading onto the MBPTrap HP column. vi. Prior to loading th e clarified lysate, equilibrate the MBPTrap HP column with 7 column volumes (CV) of binding buffer at a flow rate of 1mL/min vii. Load clarified lysate onto the column. The binding capacity of the 1mL MBPTrap HP column is protein dependent but can be approximated to 5mg - 7mg of MBP - tagged protein. The flow rate should be decreased to 0.5mL/min. 55 viii. Wash with 10 CV of binding buffer at a flow rate of 1mL/min. If real - time A280 readings are possible, wash until no discernible absorbance at 280nm is observed. ix. Add 5 CV of elution buffer at a reduced flow rate of 0.5mL/min, and assay elute fractions to determine fractions with the highest concentration of the protein of interest. x. Regenerate the column with 3 CV distilled water followed by 3 CV of 0.5M NaOH. Was h away the NaOH with 5 CV distilled water. The column is now ready to be used again. 3.3.5 BODIPY - PA binding assay BODIPY conjugated PA was used in a fluorescence polarization based binding assay to - CD wild type and mutant proteins. Protocol: i. Dissolve BODIPY - PA in 100% ethanol and make a concentrated stock of 10uM. ii. Use 10nM BODIPY - PA and mix w ith different concentrations (0 - 150nM) purified - CD in a PBS buffer. iii. Incubate at RT for 5 mins. iv. Use a Biotek Synergy H1 hybrid multi - mode reader. Make sure the FP filter cube is installed properly. v. Measure parallel and perpendicular intensities at ex 488nm/ em 520nm and determine fluorescence polarization value (FP) using Equation 3 - 1 . 56 3.4 Results 3.4.1 Cloning and preparation of pFastBac - - CD The pFastBa c - - CD plasmid features are shown in Figure 3 - 4 - CD is fused to the C - terminus of an MBP protein under the polyhedrin promoter. A 6X His tag is also present at the N - - CD coding sequences are separated by a TEV (tobacco etch virus) protease site. The plasmid has an ampicillin and gentam icin selection marker for selection during the preparation of recombinant bacmid. The pFastBac backbone has a Tn7att transposition element that guides the insertion of the coding sequence into the bacmid. After transformation into E. coli DH10Bac , colonies were grown for 48 hrs. Colonies show up within 24 hrs but take an additional day to develop a blue color. Large white colonies were picked and re - streaked onto a Tet - Kan - Gent - IPTG - X - gal plate. Figure 3 - 4 Plasmid map for pFastBac - - CD . 57 After this additional round of selection, a colony PCR was performed to make sure - CD was incorporated into the recombinant bacmid ( Figure 3 - 5 ). Plasmids were - CD sequence was confirmed, glycerol stocks were made for future use. Figure 3 - 5 Colony PCR of white E. coli transformant colonies . 3.4.2 Transfection and infection into Sf21 cells The recombinant bacmid was extracted from E. coli DH10Bac cells and precipitated with isopropanol to get pure bacmid samples. Sf21 cells were transfected and monitored for signs of infection for 5 days and multiple passages of viral titer. Figure 3 - 6 shows the Sf21 cells changing on infection with P3 baculoviruses. The cells become bigger in size, and the nuclei seem to occupy more of the cell. The Sf21 cells show more granularity as the in fection progresses. Figure 3 - 6 Signs of infection in untransfected Sf21 cell s (left) and P3 treated Sf21 cells (right) for 24 hrs. Scale bar = 45µm 58 - CD after transfecting by collecting cell lysate after P0 and P1 infections. Figure 3 - 7 shows a GelCode Blue - stained SDS - PAGE of cell - CD. The - CD protein expressed in P1 infected cells was more than P0 infected cells reflecting the higher concentration of baculoviral titer. The baculovirus stock was amplified to P3 to obtain high protein expressing infected Sf21 cells. The P3 viral supernatant was collected, filtered with a 0.22µm filter and stored at 4 o C in the dark for subsequent infection. Sf21 cells were passaged the day before infection so that they were in log phase at a density of 3x10 5 cells/mL. 150uL P3 supernatant was added for each mL of Sf21 cells. The final volume of Sf21cell s infected can be adjusted based on the yield of protein required. The time required for the expression of a protein depends on each protein of interest expressed. 59 Figure 3 - 7 Expression of MBP - - CD pr otein (Left) SDS - PAGE of total protein lysate for untransfected Sf21 cells . P0 and P1 infected clones 1 and 15. The 100kDa band is marked and the putative 6X His - MBP - - CD protein band is prominent in P1 infections (Right) Western blot using an anti - IRE1antibody 3.4.3 Optimization of protein expression and purification Figure 3 - 8 shows the western blot for the time course of protein expression to - CD, in P0 infect ed cells, maximum expression was seen at 96 hrs. On increasing infection titer by using P2 stocks, the overall level of protein expression increased. The optimal time point for harvesting was still after 96hrs of infection. Once the viral titer and time fo r expression was optimized, cell lysates were collected for purification of the protein - CD to aid in purification. Filtered and clarified lysate was run through the MBPTrap column to obtain 60 - CD protein with >85% p - CD is purified, it can be used for binding assays to determine binding partners for the protein (Cho et al., 2011) . Figure 3 - 8 Western blot showing the time course of protein expression in P0 (left) and P1 (right) infected Sf21 cells using anti - . 3.4.4 Binding assay with BODIPY - PA - - CD with 6 different mutations at the residues predicted by the MD simulations (R611A, R687A, R864A, R887A, K893A, R912A). To evaluate whether these residues are in fact involved in PA binding, we employed a binding assay with a fluorescently labeled PA molecule. The BODIPY fluorophore (4,4 - difluoro - 3a,4a - diaza - s - indacene) conjugated with palmitate (Molecular Probes) was used to develop a fluorescence polarization based binding assay. Fluorescence polarizatio n assays has been used to demonstrate protein - protein, protein - nucleic acid interactions (Lea & Simeonov, 2012) . In this work, we use it to - CD. The size of a fluorophore correlates inversely with the degree of polarization. Therefore, bulky fluorophore s like BODIPY are conjugated to the ligand of interest, in this case, PA to slow down its molecular rotation ( Figure 3 - 9 ). 61 Figure 3 - 9 Structure of BODIPY fluorophore (from Thermo Fisher) . BODIPY dyes stay longer in their excited states and are more hydrophilic as compared to older generation dyes. The schematic for the assay is shown i n Figure 3 - 10 . - CD on binding to PA in presence of polarized excitation light. Since PA is such a small molecule, the fluorescence polarization value is not high enough ( Figure 3 - 10 , top). However, on conjugation with a bulky BODIPY molecule, the molecular rotation slows down, and the polarization value goes up ( Figure 3 - 10 , bottom). The polarization value is calculated as follows: Equation 3 - 1 : To calculate the polarization value of the sample where I parallel is the intensity of emitted light parallel to the excitation light and I perpendicu lar is the intensity of emitted light perpendicular to excitation light. 62 Figure 3 - 10 Fluorescence polarization (FP) assay using BODIPY - PA . The FP assay is developed to assess binding of palmitate to the protein of interest. In the top two panels, the lysozyme protein (negative control) does not bind to PA or BODIPY - CD binds to PA, but still shows a low FP value. In the fourth panel, c onjugation of a bulky BODIPY molecule to the PA - CD slows down molecular tumbling leading to a high polarization value . 63 To establish that the BODIPY - PA binding assay was specific to fatty acid binding proteins, we execu ted it with BSA as a positive control and lysozyme as a negative control. Figure 3 - 11 BODIPY - PA binding to BSA and lysozyme . 10nM BODIPY - PA was mixed with different concentrations of BSA and lysozyme. FP increases with increasing BSA concentration but not with lysozyme concentration Testing a concentration of 0 - 150nM, BSA fluorescence polarization values increased and reached a plateau. Lysozyme, however, does not bind fatty acids which was reflected in the lack of increase in FP over increasing concentrations of lysozyme ( Figure 3 - 11 ) . - activation loop. K893 and R912 a re on the dimer interface. Previous work done in our lab suggested that the R611 residue was important in palmitate binding (Cho, 2013a) . - CD was expressed in an E. coli system and displayed little to no RNase activity. Therefore, it is possible that the post - translational modific ation machinery 64 in a bacterial system cannot accurately process mammalian IRE1 protein. To account for those differences, we decided to use an insect cell expression system. However, with this purified protein, BODIPY - PA binding assay did not show any di fferences in binding for R611A mutation ( Figure 3 - 12 ). We also tested other residues predicted by MD simulations to affect PA binding by mutating them to alanine. They did not display changes in binding affinity as compared to - CD ( Figure 3 - 12 and Figure 3 - 13 ). Figure 3 - 12 BODIPY - PA b - CD wild type and mutants . 0 - 150nM of each purified protein was mixed with 10nM BODIPY - PA and the fluorescence polarization values were recorded for different concentrations of purified protein. 65 Figure 3 - 13 - CD on PA binding . - CD (wild type, R611A, R687A, R864A, R887A, K893A and R912A ). There is no s ignificant change in FP values between the wild type and different mutant proteins(n=3). 3.5 Conclusions Sf21 insect cells combined with the pFastBac baculoviral system make a very robust, optimizable system for expression of eukaryotic proteins. MBP - - CD was successfully expressed and purified from Sf21 cells. This protocol can be easily modified to express any mammalian protein in insect cells by modifying the PCR primers shown 66 in Table 3 - 5 for the desired coding sequence ( CDS ) . Long term expression of the protein can be achieved by making frozen stocks of baculovirus infected Sf21 cells. The MBP tag can be cleaved off using AcTEV protease if it interferes with downstream assays and the protein can be cleaned up using a Ni - NTA column. Further characterization of the expressed protein is required to establish its structure, phosphorylation status and in vitro XBP1 splicing activity conclusively . The phosphorylation status of the protein can be identified by various methods. Feldman et al., used mass spectrometry techniques to identify the phosphorylated sites on IRE1 (Feldman e t al., 2016) . This technique in conjunction with a Phos - tag gel would lend more information on the ratio of phosphorylated to unphosphorylated species in the purified protein (Qi, Yang, & Chen, 2011; R. Volmer et al., 2013) . Using the Phos - tag lig and developed by Fujifilm Wako - Chem, an SDS - PAGE gel can be used to separate phosphorylated proteins from unphosphorylated forms. The Phos - tag ligand binds to the phosphate group and slows down migration of the phosphorylated band. This leads to a separati on of the phosphorylated and unphosphorylated species into separate bands. To confirm that the band pattern is because of distinct phosphorylation , purified protein - phosphatase enzyme with activity towards phosphorylated seri ne, threonine and tyrosine residues. On treatment and subsequent application to a Phos - tag SDS - PAGE gel, the slower migrating band should disappear. In vitro XBP1 splicing can be assayed with a hairpin - RNA cleavage assay or by a fluorescence - based assay with a FRET - paired oligonucleotide (Ali et al., 2011; Prischi et al., 2014) . If the protein structur e needs to be determined, circular dichroism or small - angle X - ray scattering techniques can be used. Circular dichroism is a very useful tool to 67 determine secondary structure as well as native folding of expressed or fusion proteins. Specifically, secondar y structure can be determined using far - UV spectra and protein folding characteristics can be determined by monitoring spectra at different temperatures or in the presence of different denaturing agents (Greenfield, 20 06) . 68 4 REGULATION OF DESMOPLAKIN MEDIATED BY IRE1 - XBP1 IN HEPATOCELLULAR CARCINOMA 4.1 Introduction 4.1.1 ER stress The endoplasmic reticulum is the site for protein synthesis, protein maturation, calcium and redox homeostasis. The quantity and type of protei ns produced varies widely between different types of cells. Some normal biological processes such as B - cell differentiation, insulin secretion are associated with transient ER stress and trigger the Unfolded Protein Response (UPR) (Schröder & Kaufman, 2005c) . The UPR is the collective term used for all the signaling pathways that are activated to bring the ER back to homeostasis and its normal physiological state. The most commonly studied cause of ER stress is when there i s a discrepancy in the rate of protein folding to the rate of protein translation. In the context of cancer nutrient deprivation, hypoxia, certain metabolites, disruptio n of calcium homeostasis. However, there are also other causes of activation of the UPR such as perturbations in the lipid metabolism, changes in redox homeostasis (Pineau et al., 2009; Romain Volmer & Ron, 2015; Yadav et al., 2014) . One such injury is increase in the level of free fatty acids such as palmitate in the plasma and increased uptake by cancer cells. P almitic acid (C16:0) can either be synthesized de novo or can be uptaken from plasma by fatty acid transporter proteins (Kinlaw, Baures, Lupien, Davis, & Kuemmerle, 2016) . Cancer cells need fatty acids for many purposes such as a source of energy, as structural blocks and for incorporation in the plasma membrane (Baenke, Peck, Miess, & 69 Schulze, 2013) . Recent research has suggested that some FAs, notably palmitate, may also be involved in altered signaling for proliferation of cancer cells and angiogenesis (Louie, Roberts, Mulvihill, Luo, & Nomura, 2013; Maloney et al., 2009) . Previous s tudies from our group and others have established that palmitate induces ER stress in mammalian and yeast cells (Cho et al., 2013; Karaskov et al., 2006a; Kono et al., 2017; Lu et al., 2012; R. Volm er et al., 2013) . PERK and ATF6 are the other two transmembrane sensors that sense ER stress. On activation by ER stress, PERK phosphorylates eIF2A that leads to a gene ral attenuation of protein translation. ATF6, on activation, translocates to the Golgi complex and is cleaved to release the cytosolic bZIP domain of ATF6, which then translocates to the nucleus and acts as a transcription factor. Treatment of HepG2 cells with palmitate shows activation of by palmitate as an inducer and the PERK branch shows minimal activation (Das et al., 2008; Das, Mondal, & Elbein, 2010) . 4.1.2 IRE1: A stress sensor in the UPR luminal domain in the E R lumen that senses misfolded proteins. The cytosolic domain that extends out into the cytoplasm is comprised of a kinase domain and an endoribonuclease transmembrane domain an d trans - autophosphorylation at its kinase domain to form the endoribonuclease pocket. The RNase domain specifically cleaves X - box binding protein 1 (sXBP1), a bZIP transcription factor and excises a 26bp intron from the unspliced XBP1 70 mRNA. sXBP1 regulates genes involved in resolving ER stress molecular chaperones, BiP/HSPA5, ERdj4/Hsp40, transcription factors such as CHOP/GADD153, ATF4 and other ER associated genes such as calnexin, protein disulphide isomerase family genes as well as genes involved in t he ER associated degradation (ERAD) pathway (Acosta - Alvear et al., 2007; A. - H. Lee et al., 2003) . XBP1 mRNA overexpression has been seen in many different types of cancers and the levels of XBP1 expression in triple negative breast cancer (TNBC) shows a significant association between high XBP1 expression and shorter relapse - free survival of patients (X. Chen et al., 2014; Jiang, Niwa, & Koong, 2015) . When XBP - / - cells were transfected into SCID mice, it was found that the XBP - / - cells did not develop into tumors after implantation (Romero - Ramirez et al., 2004) . 4.1.3 Desmoplakin: A component of desmosomes induction of ER stress with palmitate. Previous studies from our lab have shown that desmoplakin (DSP) expression levels go down in response to palmitate in HepG2 cells in a previous paper (X. Wang et al., 2011) . Therefore, we hypothesize that the migration induced by palmitate is due to the downregulation of desmoplakin . Desmosomes are intercellular junctions most co mmonly found in tissue that needs strong intercellular adhesion such as heart tissue, epidermis, gut mucosa etc. Desmosomes are composed of desmosomal cadherins (desmogleins and desmocollins), armadillo proteins (plakoglobin and plakophilin) and desmoplak in. There are two isoforms of desmoplakin that are expressed differentially in different tissues (Chidgey & Dawson, 2007; Nekrasova & Green, 2013) . 71 In this study, we demonstrated that the e pithelial - to - mesenchymal transition ( EMT ) transcription factors ZEB1 and ZEB2 are involved in downregulation of desmoplakin. EMT is a part of normal embryonic development where it is important in stem cell differentiation, gastrulation, formation of the se condary palate and other developmental processes. In adults, EMT is important in wound healing, regeneration of the ovarian epithelium. However, it is also seen in progression of cancer. EMT is characterized by loosening and loss of cell - to - cell adhesion, increased motility and loss of apico - basal polarity (Dejeans, Barroso, Fernandez - Zapico, Samali, & Chevet, 2015; Lamouille, Xu, & Derynck, 2014; Saitoh, 2018) . Cells form junction s with neighboring cells with the help of cell adhesion proteins. Different types of junctions are formed by different classes and types of cadherins, connexins, claudins and occludins (Troyanovsky, 1999) . We found that IRE1 is necessary for the ZEB media ted downregulation of desmoplakin. It is possible that spliced XBP1 plays a role in upregulati ng ZEB1 and ZEB2 transcription . We provide evidence that the ZEB1/2 transcription factor has possible XBP1 binding sites on its promoter region. Thus, XBP1 could bind ZEB1/2, which in turn negative ly regulat e DSP expression. 4.2 Materials and Methods 4.2.1 Cell lines and culture conditions Human cell lines for liver cancer ( HepG2, Hep3B , SNU - 387 ) , breast cancer ( MCF10A, MDA - MB - 231 ) and colorectal cancer ( HCT116 ) were purchased from ATCC. HepG2, Hep3B and MDA - MB - Medium (DMEM) supplemented with 10% fetal bovine serum (FBS). SNU - 387 cells were 72 cultured in RPMI1640 medium supplemented with 10% FBS. HCT116 cells w ere cultured DMEM/F12 medium supplemented with 5% horse serum, 0.5µg/ml hydrocortisone, 100ng/ml cholera toxin, 10µg/ml insulin and 20ng/ml epidermal growth factor. Experiments p erformed in 6 well plates were cultured in 2mL media. All cell lines were maintained in a humid incubator with 5% CO 2 at 37°C. 4.2.2 Generation of IRE1 - / - KO cell lines pSpCas9(BB) - 2A - GFP (PX458) and pSpCas9(BB) - 2A - Puro (PX459) V2.0 was a gift from Feng Zhang (Addgene plasmid # 48138 and # 62988) (Ran et al., 2013) . pX333 was a gift from Andrea Ventura (Addgene plasmid # 64073). Three high scoring single guide RNAs (sgRNAs) were synthesized and inserted into the plasmids. The sgRNAs wer e predicted by crispr.mit.edu . They are then verified by Cas - OFFinder, Off - Spotte r. This makes sure that there are no - off target effects within the genome. gRNA - A and gRNA - B is inserted in pX333. gRNA - C was inserted in pSpCas9(BB) - 2A - Puro (PX459) V2.0. gRNA - A and gRNA - C were inserted within Bbs1 cut sites. gRNA - B was inserted between A flII and KpnI sites. sgRNA - A TCACCGCCTCGCTGTCGTCGCGG sgRNA - B CCGTACCGCCCCCGGAGCCAGGG sgRNA - C CCAGAGGGAGGCCGCCGAAT GGG All inserted constructs were confirmed by DNA sequencing. The plasmids were transfected into Hep3B , HepG2 cells using Lipofectamine 30 00 (Thermo, cat # L3000001). After using the appropriate selection agent (4µg/mL puromycin for 2 - 5 days, or GFP based 73 cell sorting) , individual clonal populations were expanded and analyzed for loss of IRE1 expression by western blotting . Clones with no de tectable IRE1 protein were selected for further studies . 4.2.3 Knockdown of DSP using siRNA Hep3B and HepG2 cells were seeded at a density of 5x10 5 cells in 1 mL in a well of a 6 - well plate . DSP siRNA were obtained from Thermo (assay id. s4333, s433 5 ) and transfe cted with Lipofectamine RNAiMAX (Thermo Fisher , cat # 13778030) per well using reverse transfection for 24 hrs . A dose response (0, 20, 40, 80 nM) was performed to determine the concentration of siRNA required for >80% knockdown of DSP levels with no observ ed toxicity . This concentration (80nM) was used for all subsequent experiments. The transfection was carried out in 2mL of media per well of a 6 - well plate. 4.2.4 Wound healing assay Wound healing assays were performed to assay the directional migration of a cell monolayer (Rodriguez, Wu, & Guan, 2005) . A 6 - well plate was used to seed 5x10 5 cells in 1 mL - / - cells in a monolayer for the wound healing assay. We looked at the effect of IRE1 and PA on wound healing of the MEF cells. MEF cells were reverse transfected for 24 hrs with pCDNA plasmids (1.5ug) containing either full length IRE1, IRE1 - W457A or dLD - IRE1 (IRE1 without the luminal domain) . At the beginning of the experi ment, a uniform scratch is made. Cell culture media (DMEM 11965) containing 10% FBS was replaced with BSA or PA media without FBS and with mitomycin C added at 1µg/ml (Sigma Aldrich, cat# M4287). Images are taken at 0 hrs, and every 4 - 6 hrs of pre - marked a reas of the scratch after media change and addition of mitomycin C. The images 74 are analyzed using TScratch to calculate the rate and percentage of open wound closure (Gebäck, Schulz, Koumoutsakos, & Detmar, 2009) . 4.2.5 Trans - well migration assays were performed using cell culture inserts (8 µm pore size) (Sigma - Aldrich, cat# CLS3464) . Cells were counted and 5 x 10 4 c ells were seeded in 200 µ L in the upper chamber of the inserts in serum - free media supplemented with 1µg/ml mitomycin C (Sigma Aldrich, cat# M4287) to prevent proliferation. In experiments involving siRNA mediated DSP knockdown, HepG2, Hep3B and MDA - MB - 231 cells were reverse transfected as described in Section 4.2.3. After 24 hrs, the media was replaced with serum - free media supplemented with 1µg/ml mitomycin C . The inserts were placed in wells with 500 µ L regular medium containing 10% FBS as chemoattractant. For the insert was replaced with 200 µ L PA media without FBS (containing mitomycin C). The media in the outer well contains 10% FBS as a chemoattractant in BSA or PA media . A lipophilic tracer (DilC 12 ) (ThermoFisher, cat # D383) at a dilution of 1:100 was used to track migrating cells. After 24 hours of incubation at 37°C, the number of cells that migrated across the membrane were counted in 3 fields of view using a fluorescence microscope at 20X magnification. Each migration assay was repeated three times. 4.2.6 XBP1 splicing assay mRNA was collected from H ep3B and HepG2 wild type and IRE1 - / - KO cells. A one - step RT - PCR kit was used to amplify huXBP1 cDNA (Qiagen, cat # 210210). huXBP1 - - TTACGAGAGAAAACTCATGGCC - - rev - - GGGTCCAAGTTGTCCAGAATGC - on 75 sequence and the products were resolved on a 2.5% agarose gel made in 1X TBE until the bands were resolved (for 5 - 6 hrs) . 4.2.7 qPCR for EMT transcription factors mRNA was collected from Hep3B and HepG2 wt and IRE1 - / - KO cells with an RNeasy Plus mRNA extrac tion kit (Qiagen cat # 74134). cDNA was prepared (Thermo, cat # 4368814) and equal amounts (75 - 100ng) used for qPCR with the iQ SyBr Green supermix (Bio - Rad, cat # 1708882) using 95 o C for 30s followed by 58 o C for 30s, 72 o C for 45s for 40 cycles. ZEB1 was a - ATGCAGCTGACTGTGAAGGT - - GAAAATGCATCTGGTGTTCC - - CAGCCGCATCTTCTTTTGCG - - TGGAATTTGCCATGGGTGGA - 4.2.8 Western Blotting for EMT markers Hep3B and HepG2 cells were c ollected after treatment or transfection. Cell lysates were prepared using RIPA buffer containing a protease inhibitor cocktail and the protein concentration was measured using a Bradford assay. 30 - 50ug total protein was loaded onto an SDS - PAGE gel run in a Tris/glycine/SDS buffer system. The proteins were transferred onto a nitrocellulose membrane and blocked with 5% BSA for 3 hrs at RT. Primary antibody incubation was done overnight at 4 ° C with antibodies against IRE1 (Cell Signaling Tech, cat # 3294) , D SP (Abcam, cat # ab71690), E - cadherin (Cell Signaling Tech, cat # 3195), Vimentin (Cell Signaling Tech, cat # 3932) and GAPDH (GeneTex, cat # GTX100118). 76 4.2.9 Luciferase assay The LightSwitch Luciferase Assay system (Active Motif) was used to insert the DSP p romoter sequence upstream of an engineered luciferase gene (RenSP). A 96 well plate was seeded with 5000 cells/well. 150ng of reporter plasmid containing the DSP promoter was then reverse transfected into Hep3B and MDA - MB - 231 cells. siRNA designed against ZEB1 or ZEB2 or a scrambled sequence was co - transfected at a concentration of 200nM per well along with the plasmid. The siRNAs used against ZEB1 and ZEB2 are a combination of three 27 - (Origene, cat# SR321982, S R306590). The luciferase assay was carried out according to the LightSwitch Luciferase Assay Kit instructions 24 hrs later (Active Motif, cat # 32031). 4.3 Results and Discussion 4.3.1 - / - hepatocellular carcinoma cell lines using CRISPR - / - cell lines were generated. Multiple clones were grown and a ssayed for expression of protein. C3 for the Hep3B cell line showed a complete knockdown of IRE1 and so was used in further assays ( Figure 4 - 1 ) . Clones 24 and 32 for the HepG2 cell line demonstrated no expression of IRE1 and clone 24 was chosen for further experiments ( Figure 4 - 2 ) . 77 Figure 4 - 1 CRISPR clones obtained for Hep3B cells . Quantification and images of a w estern blot for IRE1 protein expression of CRISPR clones obtained for Hep3B cells normalized against control wild type Hep3B cells. Clone C3 shows a complete knockout for the IRE1 protein as detected by the western blot. To confirm the functional knockout of IRE1, Hep3B C3 cells were treated with 0.4 mM PA for 24 hrs and 48 hrs and HepG2 C24 cells were treated with 0.4mM PA for 24 hr s. Wild type Hep3B and HepG2 cells treated with PA were used as a control. As seen in Figure 4 - 3 and Figure 4 - 4 , wild type Hep3B and HepG2 cells show splicing of XBP1 on induction with PA that is abolished in the IRE1 - / - knockout cells. 78 Figure 4 - 2 CRISPR clones obtained for HepG2 cells . CRISPR clones obtained for Hep G2 cells normalized against control wild type HepG2 cells. Clones 24 and 32 show a knockout for the IRE1 protein with minimal background signal . Figure 4 - 3 RT - PCR showing XBP1 splicing activity in wild type and IRE1 - / - Hep3B cells on induction with palmitate. Cells were treated with BSA media (control) or 0.4mM PA media for 24 (24PA) or 48 (48PA) hrs of XBP1 as indicated by the presence of the XBP1 S knockout cells, there is no splicing in BSA or PA treated cells. 79 Figure 4 - 4 XBP1 splicing activity in wild type and IRE1 - / - HepG2 cells on induction with PA . Cells were treated with BSA media (control) or 0.4mM PA media for 24 hrs . Wild type HepG2 cells treated with PA show activa by the presence of the XBP1 S splicing in BSA or PA treated cells. Once the CRISPR knockouts were confirmed, they were used to further investigation on 4.3.2 determine the rate of migr ation of the cells. Wound healing assays/ scratch assays provide a qualitative measure of cell migration. offer several advantages, including the ability to control for cell proliferation effects and cell - cell interaction e designed to afford the versatility to test the effect of inhibitors or activators on cell migration (H. Chen, 2005) . To study the effect of palmitate, we added PA media (without serum) to the upper chamber where the cells are loaded. The bottom chamber separated by a porous membrane was DMEM 11965 med ia with 10% FBS. 80 Many Boyden chamber devices are available commercially. Transwell cell culture were used to assay migration . Cells were seeded on the upper compartment of the chamber at a cell density 5x 10 4 in 200µL of serum free BSA (control) or serum - free PA medium containing mitomycin C . A lipophilic tracer (DilC 12 ) at a dilution of 1:100 (2 µL in 200 µL) in th e upper chamber and used to track migrating cells after 24 hrs. Figure 4 - 5 Wound healing assay with MEF IRE1 - / - KO cells transfected with wild - FL - - On PA treatment, a higher rate (70 - 80%) of wound closure was seen in cells expressing FL - - - / - KO or FL - - axis denotes the % open wound area. Previous studies as well as publications from our research group have shown that palmitate increases the rate of migration of various cancer cell lines (Nath, Li, Roberts, & Chan, 2015) ., We performed a scratch assay/wound healing assay with mouse embryonic 0 20 40 60 80 100 120 IRE1-/- IRE-/- + PA FL-IRE1-WT +PA FL-IRE1-W457A + PA dLD-IRE1 + PA % open wound area Wound Healing Assay 0 hr 4 hr 8 hr 12 hr 81 lot slower than cells expre ssing full length - There is no effect of PA treatment on the migration of MEF IRE1 - / - as evident in Figure 4 - 5 . However, MEF cells expressing either the IRE1 full length protein (FL - IRE1 - wt) or the truncated form of the protein without the luminal domain (dLD - IRE1) migrate faster than the control in res ponse to PA. This response is abrogated when the cells express ed the W457A mutated FL - IRE1 ( Figure 4 - 5 ) . he rate of - / - KO Hep3B and HepG2 cells demonstrates that treatment with PA increases - / - KO cells (p<0.005) ( Figure 4 - 6 ) . This Figure 4 - 6 and HepG2 (right) cells. Wild type cells induced with PA migrated across the membrane much faster than BSA control cells & PA treated IRE1 - / - cells (** p - value<0.05) The y - axis denotes the average number of cells counted in 3 fields of the transwell (n=3) 82 4.3.3 Desmoplakin levels decrease on palmitate treatment In a previous study from our group, we published the effect of palmitate on the expressio n of desmoplakin (DSP) in HepG2 cells (X. Wang et al., 2011) . HepG2 cells treated with palmitate media f or 48 hrs showed a decrease in DSP expression as seen with immunofluorescence. On replacement with normal media, this expression was recovered to some extent within 24 hrs and fully recovered in 72 hrs. We demonstrated above that p almitate increases the mi gration of Hep3B and HepG2 cells , which is accompanied with changes in cell morphology and induction of EMT (Nath et al., 2015) . In this study, we investigated the effect of knockdown of DSP on the migration and invasion characteristics of the cancer cells. Hep3B , HepG2 and MDA MB 231 cells were transfected with a combination of 2 siRNA against DSP. The concentration of siRNA(80nM) to be used was determined by c omparing the level of decrease in DSP expression after 24 hrs. Additionally, u number of migrating cells w as counted ( Figure 4 - 7 ). Fetal bovine serum (FBS) was used as a chemoattractant. In both Hep3B and HepG2 cell lines, decrease in DSP levels led to a significant increase in the number of migrating cells (p< 0.05). 83 Figure 4 - 7 Boyden's chamber assay for Hep3B, HepG2, MDA MB 231 . Cells were treated with 80nM DSP siRNA as determined by the dose response to knockdown levels of DSP protein by over 80% . As compared to the cells treated with a scrambled control siRNA , DSP KD cells migrated faster through the membrane . Desmoplakin is an essential component of the desmosomal intercellular junctions. Loss of desmoplakin expression or mutations in desmoplakin lead to many disea ses such as early onset cardiomyopathy, erythrokeratodermia (Boyden et al., 2016; McGrath, 2005) . Recent evidence has suggested that desmosome deficiency may be involved in increased local invasion of tumors (Dusek & Attardi, 2011) . Nath et al . assessed the influence of DSP expression levels on patient survival rates in the TCGA pan - cancer dataset (Nath, 2015) . A comparison of survival curves of patients stratified according relative DSP expression suggests there were no differences in survival rates between patients with basal or high DSP levels. However, low DSP levels were strongly associated with reduced survival rates (p=8.7x10 - 17 ). The log rank test is the sum of the chi - square value for each e vent time in each group (Nath, 2015) . This study shows that the loss in DSP due to PA treatment may be mediated through 84 - / - e, even on palmitate treatment, there is no splicing of XBP1 mRNA ( Figure 4 - 3 and Figure 4 - 4 ) - / - KO cells with 0.3mM PA media. Figure 4 - 8 A shows the relative expression of desmoplakin in Hep3B cells on treatment with palmitate. In wild type Hep3B cells, there is a ~50% decrease in DSP levels that is - / - cells ( Figure 4 - 8 B ). This suggests that the downregulation - XBP1 pathway. Figure 4 - 8 Desmoplakin levels in Hep3B wild type vs IRE1 - / - KO treated with PA. (A ) Hep3B cells treated with media containing 2% BSA or 0.3mM PA in 2% BSA. Wild type Hep3B cells show a decrease in DSP levels on PA treatment. ( B) Quantification of blots in panel A (n=3). p - value<0.05 Palmitate leads to a loss of E - cadherin and increase in protein levels of vimentin as seen as a population effect in HepG2 cell s (Nath et al., 2015) . To investiga te whether this effect is mediated through loss of DSP, we performed an siRNA mediated knockdown of DSP and studied the effect on the protein expression levels of various epithelial - to - mesenchymal transition (EMT) genes. We found that knockdown of DSP does not lead to a decrease in expression of E - cadherin or vimentin in either Hep3B or HepG2 cells. Figure 4 - 9 shows representative western blots and quantification of levels of E cadherin and 85 vimentin. There is no loss of E - cadherin on DSP knockdown. Levels of E - cadherin typically decrease at the onset of EMT and are one of the hallmarks of cancer (Hanahan & Weinberg, 2011) . The tumor suppressor role of the cel l - adhesion protein E - cadherin has been established previously (Semb & Christofori, 1998) . Mutations in E - cadherin are seen in some diffuse gastric tumors and breast cancers . However, in the majority of cases, decrease in E - cadherin is through transcriptional repression. Loss of E - cadherin mediated cell - cell adhesion is involved in the transition to an invasive tumor (Christofori & Semb, 1999; Nakagawa, Hikiba, et al., 2014; Semb & Christofori, 1998; Singhai et al., 2011) . This indicates that induction of EMT by palmitate is independent of the loss of DSP. Figure 4 - 9 Expression of EMT markers in DSP KD Hep3B and HepG2 cells ( A) Western blot showing desmoplakin (DSP), E - cadherin (CDH), Vimentin (VIM) levels in Hep3B and HepG2 cells treated with 80nm of DSP siRNA1, DSP siRNA 2 or a combination of both siRNAs. (B) Quantification of DSP levels from panel A. On treatment with 80 nM DSP siRNAs, expression of desmoplakin goes down ~80% in Hep3B cells and 86 Figure 4 - ~50% in HepG2 cells. (C)E - cadherin and (D) Vimentin levels are unaffected by siRNA mediated knockdown of DSP 4.3.4 ZEB1 and ZEB2 regulate DSP transcription levels Previous work done in our lab demonstrated that mRNA expression of ZEB1 and ZEB2 transcription factors is elevated in PA treated HepG2 cells as compar ed to BSA control cells (Nath et al., 2015) . Figure 4 - 10 Scatter - plots showing correlation between the mRNA expression levels of DSP and ZEB1 or ZEB2 using the TCGA PAN - CAN dataset. s correlation coefficient and p - value showing significance of correlation. correlation (r= - 0.38) between DSP and ZEB1 (p<0.001) as well as ZEB2 (r= - 0.578) with a p - value<0.001. ZEB1 and ZEB2 are zinc finger homeodomain proteins. They have multiple functional domains one of which is a DNA binding transcriptional repressor domain (Vandewalle, Van Roy, & Berx, 2009) . 87 W e compared the mRNA expression profiles of DSP with ZEB1 and ZEB2 levels in the multi - cancer TCGA pan - cancer dataset ( Figure 4 - 10 ) . Correlation analysis confirmed that reduction in DSP expression was s ignificantly negatively correlated with ZEB1 (p<10 - 3 ) and ZEB2 (p<10 - 3 ) expression in the pan - cancer dataset ( Figure 4 - 10 ). DSP expression showed inverse association with both ZEB1 and ZEB2 expression across all cancer types ( Figure 4 - 11 ) . Figure 4 - 11 Heat map of gene expression profiles of DSP, ZEB1 and ZEB2 (X - axis) across all samples in the TCGA pan - cancer database grouped by cancer type (Y - axis). Red indicates high expression levels, while blue indicates low expression levels, with negative correlations observed b etween DSP a nd ZEB1 and ZEB2 in all cancer types On performing a qPCR for gene expression analysis for ZEB1 levels in Hep3B WT - / - KO cells, we found that treatment of the WT cells with PA led to an increase 88 in mRNA expression of ZEB1 ( Figure 4 - 12 ) indicating that ZEB1 transcript levels Previous experiments outlined in Chapter 2 have shown that the lev in response to PA treatment. However, the level of p hosphorylated IRE1 does change in response to PA (Cho, 2013a; Romain Volmer et al., 2013) . Phosphorylation of IRE1 is an important step in formation of the endoribonuclease domain and XBP1 splicing. Figure 4 - 12 qPCR for ZEB1 gene expression in Hep3B . ZEB1 mRNA expression levels in wild type and IRE1 - / - KO cells. mRNA levels from qPCR were normalized to GAPDH and fold change was obtained between PA and BSA treated cells. ZEB1 gene expression levels go up nearly 4 - fold in wild type cells treated with PA XBP1 has been shown to bind to EMT - transcr iption factors such as SNAI1, SNAI2, ZEB1, ZEB2, TCF3 (Cuevas et al., 2017) - XBP1 axis via overexpression of LOXL2 led to induction of EMT in MDCK - II ce lls. This was accompanied by upregulation of EMT genes such as vimentin and N - cadherin. On 89 upregulation of expression of E - cadherin and ZO - 1 (Cuevas et al., 2017) . Given the association between ZEB1/2 expression and reduction in DSP expression , we f urther performed a DSP promoter activity assay to demonstrate that the regulation of DSP expression was specifically controlled by ZEB1/2. Briefly, we inserted the DSP promoter region upstream of the CDS for RenSP luciferase. Next, we co - transfected the DS P - promoter luciferase plasmid and siRNA targeted against ZEB1 or ZEB2 into Hep3B and MDA MB 231 cells. Figure 4 - 13 Luciferase assay with Hep3B and MDA MB 231. Hep3B (left) and MDA MB231 (right) were trans fected with plasmids containing the RenSP luciferase gene downstream of the DSP promoter region and treated with siRNA against ZEB1 or ZEB2. Knockdown of ZEB1 and ZEB2 shows an increase in luciferase activity as compared to the control . On transcriptional repression of the DSP promoter region by ZEB1 and ZEB2, there was no transcription and production of the luciferase protein as seen in the control cells. 90 However, on knockdown of the ZEB transcription factors with siRNA, the transcriptional repression was lifted, and luciferase could be produced. We found that the knockdown of ZEB1 resulted in significant increase in DSP promoter activity in Hep3B cells (p<0.05) and MDA MB 231 cells (p <0.05 ) ( Figure 4 - 13 ). Similarly, ZEB2 knockdown resulted in significant increase in DSP promoter activity in Hep3B cells (p<0.05) and MDA MB 231 cells (p < 0.05). These results indicated that the repression of DSP transcription was specifically regulated by ZEB binding to the DSP promoter. 4.3.5 XBP1 potentially regulates ZEB1/2 activity From Figure 4 - 12 we have demonstrated that IRE1 - / - knockout cells do not show an upregulation in ZEB1 levels on PA treatment. This indicates that XBP1 splicing and activation of the transcription factor is important in ZEB1 upregulation. We further analyzed whether XBP1 has the potential to directly bind to the ZEB1/2 prom oter region. Table 4 - 1 Predicted binding sites for XBP1 on the ZEB1/2 promoter region relative to the transcription start site (TSS) . The JASPAR 2017 TFBS prediction algorithm was used to predict binding sites for the XBP1 consensus motif on the ZEB1 and ZEB2 promoter regions. Predicted XBP1 binding sites on ZEB promoter relative to TSS Start End Strand Predicted site sequence ZEB1 676 663 - AAGGAAACGTCTTT ZEB1 786 773 - ACTGAAACGTGACC ZEB2 1046 1033 - AGGTATACGTCTTT ZEB2 348 335 + ACAGAAGCGTCACG 91 The promoter region (up to 1200 bp upstream of the transcription start site) for ZEB1 (Gene id: 6935) and ZEB2 (Gene id: 9839) was obtained using UCSC Genome Browser and used to run a JASPAR TFBS prediction. Table 4 - 1 shows the predicted binding sites for XBP1 on the ZEB1/2 promoter region. This needs to be further investigated with the help of ChIP assays and luciferase assays to establish the transcriptional regulation of ZEB1/2 by XBP1. In this chapter, we have investigated the link between loss of DSP expression and activation of IRE1 by PA . DSP , a cell adhesion protein , is an important protein in the progression of the EMT cascade. Loss of DSP due to treatment with PA is accompanied by inc reased cell migration. factors act as transcriptional repressors for DSP. This is through an IRE1 - XBP1 mediated mechanism. Further studies are necessary to conclusively establish the XBP1 mediate d upregulation of ZEB1. 92 5 THE ROLE OF XBP1 IN CANCER CELL METABOLISM 5.1 Introduction 5.1.1 Obesity and incidence of cancer Obesity is an important factor in the development of multiple types of cancer. Various cohort studies over the years have looked at the association of BMI and cancer risk (Campbell et al., 2016; Y. Chen, Wang, Wang, Yan, & Luo, 2012; Y. Ma et al., 2013; Renehan, Tyson, Egger, Heller, & Zwahlen, 2008; Roberts , Dive, & Renehan, 2010) . The increase in risk is not uniform or even present across all types of cancers indicating that there is some specificity associated with this increase in risk (Eheman et al., 2012; Renehan et al., 2008) . Breast cancer is the most commonly diagnosed cancer in wom en with an estimated 268,000 cases in 2019. Similar tends are seen with the incidence and recurrence of breast cancer in women (Brown & Simpson, 2010; Neuhouser et al., 2015; Picon - Ruiz, Morata - Tarifa, Valle - Goffin, Friedman, & Slingerland, 2017; Simone et al., 201 6) . Obesity is present as a risk factor in up to 50% of breast cancer affected patients (Renehan et al., 2008) . Production of inflammatory cytokines, insulin resistance and systemic inflammation a re responsible for development of breast cancer and have been identified as molecular mechanisms triggering tumorigenesis in patients. Colorectal cancer is the third most commonly diagnosed cancer accounting for almost 700,000 deaths in 2012 and the morta lity has now increased to 880,000 in 2018 (Ferlay et al., 2018) . Obesity is a significant risk factor in the development of colore ctal cancer. Accumulation of visceral abdominal fat coupled with lack of physical activity lead 93 to development of hyperinsulinemia and changes in the mucosal lining of the colon (Frezza, Wachtel, & Chiriva - Internati, 2006) . Primary liver cancer ranks as the sixth most common cancer worldwide. Although the injuries that lead to hepatocellula r carcinoma (HCC) differ, it accounts for 85 - 90% or primary liver cancer cases (Y. Chen et al., 2012) . HCC is caused by viral hepatitis in most parts of the world except America. In the United States, the rates of HCC have increased proportionally with the rate of increase in obesity (Campbell et al., 2016) . The difference in risk can also be attributed partly to the presence of other confounding factors across different genders, races, and socio - economic levels. For example, a meta - analysis study in 2008 looked at the effect of body mass index (BMI) on lung cancer in men and women and found a reduced risk ratio (0.76 - 0.80) across six studies (Renehan et al., 2008) . But, the smoking status of the patients was not taken into consideration during this analysis. I n some distinct c ancers, the increase in risk was attributable to obesity and increased BMI or waist circumference ( Figure 5 - 1 ) . Figure 5 - 1 Increase in risk of liver cancer across US populations due to obesity. The values were adjusted for age, sex, alcohol, cigarette smoking, race and diabetes status ( data from (Campbell et al., 2016) ) 0 20 40 60 80 100 120 140 160 Overweight Class I Class II Class III Percentage Increase in risk of liver cancer by level of obesity Increase in risk of liver cancer 94 A study in 2016 on the risk of liver cancer in US adults attributed the 3 - fold increase in liver cancer to the increasing incidence of ob esity in the US population (Campbell et al., 2016) . Figure 5 - 1 outlines the overall conclusions of this study , which are backed by other studies that also looked at the relationship of BMI and risk of liver cancer (Y. Chen et al., 2012; Renehan et al., 2008) . Another strong predictor for liver cancer is Type II diabetes mellitus (Campbell et al., 2016; Lai, Park, Hartge, Hollenbeck, & Freedman, 2013; Yang et al., 2011) . A distinction needs to be made here , to underscore that BMI may not be the best indicator for high levels of fatty acids. Other EMT markers need to be considered as was discussed in Nath et al in 2015 (Nath et al., 2015) . The unreliability of BMI as a n indicator of fatty acid levels has been compensated for by adding other indicators to epidemiological studies. Examples of such indicators are waist circumference, insulin level s, estrogen levels , etc . (Frezza et al., 2006; Neuhouser et al., 2015) . Plasma free fatty acid levels are elevated in obese individuals (Opie & Walfish, 1963) . Thus other than BMI, plasma fatty acid levels could be better indicators of cancer risk. Along those lines, saturated fatty acid levels have been found elevated in various cancers, e.g. prostate (Chavarro et al., 2013) , colorectal (May - Wilson et al., 2017) , breast (Saadatian - Elahi et al., 2002) , and liver (Nath, 2015) c ancers. 5.1.2 ER stress , fatty acids, and cancer O besity has been shown to activat e i nflammatory pathways that lead to changes in the cellular stress response (Y. Z. Liu, Wang, & Jiang, 2017; Monteiro & Azevedo, 2010) . Levels of fatty acids are upregulated in the plasma of obese individuals (Arner & Rydén, 2015; Boden, 2008) . Elevated saturated fatty acids, e.g. PA, can also come from diet, and 95 likewise have been shown to induce inflammation (Bujisic & Martinon, 2017; Coll et al., 2008; Ralston, Lyons, Kennedy, Kirwan, & Roche, 2017) and ER stress (Das et al., 2010; Karaskov et al., 2006b) . Genetic mouse models (ob/ob) of obesity fed high fat or normal diets were analyzed for ER stress indicators . It was found that both the PERK and IRE1 arms of the UPR were activated in high - fat diet fed mice. This effect was seen exclusively in adipocytes a nd hepatocytes and not in muscle cells (Ozcan et al., 2004) . The link betwee n ER stress and metabolism has been actively investigat ed and ER stress is thought to play a role in obesity - associated inflammation, type 2 diabetes, insulin resistance as well as different types of cancer (Alasiri et al., 2018; Iizuka, 2017; H. Kim, Bhattacharya, & Qi, 2015; Sha, He, Yang, & Qi, 2011; Wellen & Hotamisligil, 2005) . Studies have used XBP1 knockout animals to tease out the effects of activation of UPR , on metabolism. However, these studies come with a caveat overactivation of IRE1 is seen as a consequence of the GRP78 feedback loop making it difficult to completely understand the role of IRE1 activation (A. - H. Lee, Scapa, Cohen, & Glimcher, 2008; Wu , Zhang, Lu, Ren, & Yi, 2015; K. Zhang & Kaufman, 2008) . Hence, using IRE1 knockouts, instead of XBP1 knockouts remove s the feedback over activation through XBP1 entirely leading to a clearer picture on the role of IRE1 on metabolism . XBP1u expression ha s been ed mice and the presence of XBP1u could affect hepatic lipid metabolism through an uncharacterized mechanism (Wu et al., 2015) . Therefore, in this chapter, we will investigate IRE1 KO cells to obtain further insights into the changes in lipid metabolism. 96 5.1.3 Cancer cells show changes in metabolism Otto Warburg described the Warburg effect as the switch seen in cancer cells where instead of relying on oxidative phosphorylation to satisfy the energy needs of the cell, most cancer cells u se aerobic glycolysis to generate energy. Even though this process is inefficient in terms of energy generation, it seems to b e the preferred pathway for most cancer cells (Vander Heiden, Cantley, & Thompson, 2009) . The metabolism of glucose to lactate generates 2 ATP per molecule of glucose as opposed to 36 ATPs generated by oxidative phosphorylation. This preference can be due to accumulation of mutations in cancer cells that lead to constitutive activation of met abolic pathways, overexpression of transporter proteins that ensure a constant flux and overabundance of glucose, amino acids into the cell (Pavlova & Thompson, 2016; Vander Heiden et al., 2009) . Another hypothesis for the switch is that tumor hypoxia selects for cancer cells that exhibit the Warburg effect and are dependent on anaerobic metabolism (Vander Heiden et al., 2009) . However, hypoxia is seen predominantly in late - st age tumors, so it would not satisfactorily explain the onset of the Warburg effect (Ackerman & Simon, 2014; Gatenby & Gillies, 2004) . In addition to glucose, cancer cells also take u p amino acids, nucleic acids and fatty acids from their surroundings. This is due to overexpression or deregulation of transporter proteins (Luengo, Gui, & Vander Heiden, 2017; Pavlova & Thompson, 2016; Vander Heiden et al., 2009) . In this chapter, we look ed at different methodologies to investigate global changes in metabolism related to t he activation of ER stress in cancer . Therefore, we analyzed target genes that are potentially regulated by XBP1 by identifying consensus binding sequences in their regulatory regions. We then developed 97 qPCR assays to qu antify gene expression in IRE1 expressing and knockout liver cancer cell lines. Genes that showed differential expression on treatment with PA in IRE1 WT cells as compared to IRE1 KO cells were also analyzed for their importance in survival using patient d ata from the TCGA databases. We also analyzed 3 individual cancers, namely, liver, breast and colon cancer. Given that the primary site for metastasis of colon cancer is the liver, while breast cancer metastasizes to several locations, mainly the bones, the brain, as well as the liver, we aimed to identify genes that could be dysregulated in all three cancer typ es. We proposed the genes that are identified in this intersection potentially could be candidates that could be evaluated for their potential in metastasizing breast or colorectal cancer to the liver. Therefore, we used the publicly available TCGA data to identify an intersection of enriched metabolic genes that correlate with XBP1 on the liver (LIHC), breast (BRCA), colorectal (COADREAD) cancer datasets. 5.2 Materials and Methods 5.2.1 Selection of genes In this study, we gathered a list of genes that were predicted to be regulated by XBP1. This was achieved with the help of data obtained from genome - wide ChIP to map (Acosta - Alvear et al., 2007) . In addition to that, an extensive curation of the literature was performed to identify putative tar gets of XBP1. Out of this list of over 100 genes, we narrowed down to 30 metabolic genes and transcription factors to test with a qPCR assay based on gene function and relevance to metabolic pathways . The criteria for choosing the 30 gene s was 98 a high relat ive score for XBP1 binding in JASPAR transcription factor binding database ( Table 5 - 1 ) (Khan et al., 2018; Sandelin, Alkema, Engstrom, Wasserman, & Lenhard, 2004) . The JASPAR vertebrate core was used to scan the promoter region of each gene with the XBP1 - hand side was used to input a FASTA formatted promoter sequence for each of the 30 genes. The promoter sequence was obtained using the UCSC genome browser to obtain the region upstream of the TSS. Th selected profile. On scanning, the output contains a relative score as computed by the JASPAR algorithm (Khan et al., 2018) . The relative score threshold was set at 0.7. Table 5 - 1 JASPAR transcript ion factor database scores for various metabolic genes and EMT - TFs . Relative score Start End Strand Predicted sequence L PIN 1 0.78 777 790 + TAGGTCAAGTCAGA 0.76 378 391 - TGAGCCACATAACG 0.74 133 146 + GAAGAAACATCAGG 0.73 900 913 + AGTGCCTCTTAACC 0.73 456 469 + CTTCCCACCTCAGG FADS1 0.73 257 270 + TTTCACACGTATCA 0.73 259 272 - AGTGATACGTGTGA 0.72 172 185 - TGCCATACTTCACC 0.72 550 563 - GAGATCACGCCACT 0.70 195 208 + AAGGCAAAGTCCTT FASN 0.79 25 38 - GCTGCCAGGTCAGG SCD1 0.85 662 675 + GCTGCCACGTCTCC 0.77 326 339 + AAACACATGTCAGT SPTLC1 0.75 477 490 + TACTAAACGTAAAC SPTLC3 0.78 364 377 + TGTTCTACCTCAGC 0.75 604 617 + ATGTATACATCACA ZEB1 0.76 928 941 - TTTACGACATCACC ZEB2 0.77 152 165 + ACAGAAGCGTCACG 0.76 338 351 - TATTAAACGGCATG HIF1A 0.75 839 852 + ACGAGCACGTGAGC 99 Table 5 - 1 (cont d) Relative score Start End Strand Predicted sequence 0.86 380 393 + GCGGCCACGTCTTC TWIST 0.79 588 601 + GATGAGACATCACC 0.77 949 962 + CTCCTCACGTCAGG 0.77 39 52 + ATTCCCACTTCACT SNAI1 0.77 443 456 - AATGCCACGGCCTT 0.77 105 118 + GTTGCCACTTCTTC SLUG 0.72 684 697 - GGTGCAGCGCCAGC 0.70 357 370 - GTACTCATGTCACC ACAA1 0.85 110 123 - AGGGCCACGGCATG 0.80 230 243 - AGGGCCATGTCTTC 0.80 137 150 + TTGGGGACGTCATG 0.80 261 274 - AAGGCCACGTGCCC 0.75 605 618 + CCAACCACGTCGGA 0.75 12 25 + GGGAACACGCAATC ACAT1 0.85 620 633 + GTGGTCACGTCACT 0.77 855 868 + CCATCCACGTCCTT 0.75 866 879 + CTTCACTCGTCAAC CPT2 0.75 199 212 + AATCCGAGGTCACC 0.73 778 791 - GCCACCACTTCAGT 0.73 982 995 + CATAACATGTTATC 0.70 891 904 - ATTGACATATAATG 0.70 982 995 - GATAACATGTTATG 0.75 117 130 - AAGGACATGTTACT CAV1 0.73 564 577 - ATGGAAACTTAAGC 0.73 117 130 + AGTAACATGTCCTT 0.72 800 813 + TAGGACAGGGCAGG 0.70 861 874 + GCAGGCGCGTCGGC PH KG 1 0.83 426 439 + ACAGACACGTCCTG 0.81 912 925 - TCTTCCACCTCAGC 0.76 1362 1375 + ATTCAGACATCAGC PKM2 0.78 379 392 + TTCGCCACGTTGGC 0.76 1170 1183 - AATATCAAGTCAGA 0.76 1635 1648 + AAAGACACCCCATC 0.75 324 337 - TGGTGCACGTCTGT DGAT2 0.77 984 997 + CCCGCCGCGTCGGC 0.72 532 545 - TAAGAAACCTCACA ERDJ5 0.81 982 995 + GTTTCCCCGTCACC Calnexin 0.79 661 674 + TGGGCCATGTCGGC CD36 0.77 34 47 + TCAGGCACGTGAGC OCT1 0.78 260 273 + AATCACACGACACT 100 Table 5 - 1 (cont d) Relative score Start End Strand Predicted sequence FOXC2 0.79 239 252 + AGAGTCACTTCACC GRP78 0.79 885 898 - CAATAAACGTCACT PDK1 0.79 371 384 + TATTCAAAGTCATC PFKFB3 0.78 802 815 + AGTTCCGCGTCTGG 0.77 118 131 - CCTTCCACGTCCCG 0.75 971 984 - CCTGGCGCGTCACG PFKFB4 0.79 220 233 - TACCCCAGGTCAGC 0.77 35 48 - AAAAACACTTCAGA IDH1 0.75 328 341 - CCTCCCACCTCAGC 0.74 707 720 - GTTTCCAGGTCGGC 0.73 928 941 - TGGGCGGAGTCAGC 0.73 667 680 - TGGGCGGCGGCAGC 0.72 96 109 + AGTCCAACTTCATT 0.71 969 982 - CCCGACACGCCTCC 0.71 911 924 + GTGGCCACGCCCCT ASMase 0.85 765 778 + AAGGAGACGTCTTC 0.81 13 26 - AATGACACCTAAGC NSMase 0.84 599 612 + CAGGACACGCCAGC 0.81 406 419 - TGTTCTACGTCAAT To corroborate the predictions obtained in Table 5 - 1 , we further used a positional weight matrix (PWM) for XBP1 (length 14) obtained from the JASPAR TF database ( Figure 5 - 2 ) to scan the Homo sapiens genome (Dec 2013 GRCh38/hg38) and compared the results to a preloaded XBP1 matrix (length 11) from HOCOMOCO v11 Human TF collection. The JASPAR TF PWM returned 14156 hits as opposed to the HOCOMOCOv11 PWM which returned 67326 hits suggesting more selectiv ity in transcription factor binding predictions . PWMScan (from PWMTools; https://ccg.epfl.ch/pwmscan/ ) was used for both the analys es . All the genes selected showed predicted binding through both JASPAR and PWMS can predictions. 101 Figure 5 - 2 Position weight matrix for XBP1 transcription factor from the JASPAR TF database. 5.2.2 qPCR assay Hep3B wild type and IRE1 - / - cells were treated with 2% BSA or 0.3mM PA in 2% BSA media for 24 and 48 hrs. mRNA extraction was performed with a kit (Qiagen, cat# 74134) and transcribed into cDNA (ThermoFisher, cat# 4368814). qPCR primers were designed for use with Biorad iCycler and are listed in Table 5 - 2 . The PCR conditions used were: 95C for 3 mins, 40 cycles of 95C for 15s, 58C for 30s, 58C for 1min followed by a final extension of 72C for 3 mins. 102 Table 5 - 2 qPCR primer sequences for selected genes and EMT - TFs . Primer set Species Gene name forward reverse hu SNAIL CACTATGCCGCGCTCTTTC GCTGGAAGGTAAACTCTGGATTAGA hu SLUG GGACACATTAGAACTCACACGGG GCAGTGAGGGCAAGAAAAAGG hu ZEB1 ATGCAGCTGACTGTGAAGGT GAAAATGCATCTGGTGTTCC hu ZEB2 TATGGCCTACACCTACCCAAC AGGCCTGACATGTAGTCTTGTG hu TWIST GCAGGGCCGGAGACCTAG TGTCCATTTTCTCCTTCTCTGGA hu FOXC2 GCAGGGCTGGCAGAACAG CGCGGCACTTTCACGAA hu GAPDH CAGCCGCATCTTCTTTTGCG TGGAATTTGCCATGGGTGGA hu OCT1 GTGTGTAGACCCCCTGGCTA GTGTAGCCAGCCATCCAGTT hu PDK1 CGGATCAGAAACCGACACA ACTGAACATTCTGGCTGGTGA hu HIF1A GACAGAGCCGGCGTTTAG AGAACTCATCTTTTTCTTCTCGTTC hu PFKFB4 GGGATGGCGTCCCCACGGG CGCTCTCCGTTCTCGGGTG hu PFKFB3 CAGTTGTGGCCTCCAATATC GGCTTCATAGCAACTGATCC hu PKM2 ATCGTCCTCACCAAGTCTGG GAAGATGCCACGGTACAGGT hu ACAA1 TCTGTTAACTCCGCGGTCAG TCATGACTGCCGAGAGAAGC hu CD36 GAGAACTGTTATGGGGCTAT TTCAACTGGAGAGGCAAAGG hu CAV1 CAGGCTTGTAACCTTTACAGGAC CATAGATGCTTAGTCCCTCATGC hu PHKG1 GTACGCCGTGAAGGTCATC CTTCCCATAGGTCAAACACCAA hu LPIN1 GGAAGAAACACCACAATCAAGGA TGGAGGATGATTCATGCTTTACC hu DGAT2 AGTGGCAATGCTATCATCAT GAGGCCTCGACCATGGAAGAT hu SCD 1 TTTCACTTGGAGCTGTGGGTGAG GGAACCTGAGGGACCCCAAA hu GRP78 TCATCAACGAGCCTACGGCA CTTTTCTACCTCGCGCCGGA hu C A NX CGCGGGGCAAGATCATGGAA TGGCCCGAGACATCAACACAAG hu ERDJ5 TTTCGGTCTGGAATGGCCCC TGAGTTGAGAAACAACATGCCACT hu ACAT1 CGCTAGGGGTGCGGGGTT CACTTCTTCTTTTGGAATCCCTGCC hu CPT2 CGCTAGGGGTGCGGGGTTG GGAATCCCTGCCTTTTCAATGGCTC hu FAS N CCAAAGAAGCAGCAATGGGCCAG GCCATCAGAGACCACAGGGG hu FADS1 AGGAGCGGTGGCTAGTGAT GGCTGCTCTGGAGACAGTTC hu ASMASE CCTCAGAATTGGGGGGTTCTATGC CACACGGTAACCAGGATTAAGG hu NSMASE CATGGTGACTGGTTCAGTGG TAGAGCTGGGGTTCTGCTGT hu IDH1 ACCAATCCCATTGCTTCCATTTT TCAAGTTTTCTCCAAGTTTATCCA hu SPT1 CTTGTTCCTCCTGTCCCAAA CCCCACGCCATACTTCTTT hu SPT3 CGACTCTCAGGTGCAACCAT GGCCCCAATACTGTGAGCTT 103 5.3 Results and Discussion Each of the genes described in this section were selected because they displayed putative regulation by XBP1 through increase in gene expression in IRE1 WT cells treated with PA and not KO cells coupled with TCGA database analysis. The following results depict the qPCR analysis of gene expression levels in IRE1 WT and KO cells that showed a difference in response to PA treatment. We have also analyzed the importance of each gene as seen in differential survival because of high or low expression l evels using patient data obtained from TCGA databases. Kaplan - Meier survival was analyzed from the BRCA, LIHC and COADREAD databases as well as PANCAN databases. Representative images are shown here although the trends seen in all the databases were simila r . 5.3.1 LPIN1 Lipin1 is an essential enzyme in triglyceride metabolism and glycerolipid synthesis. It catalyzes the formation of diacylglycerol (DAG) from phosphatidic acid through its (Holthuis & Menon, 2014) . It is a transcriptional activator of PPARA in adipogenesis and differentiation. LPIN1 is important in lipodystrophy syndromes (Gómez de Cedrón & Ramírez de Molina, 2015) . High LPIN1 expression correlates with poor prognosis in multiple cancers including, breast (TNBC), colorectal, prosta te cancer etc. (Brohée et al., 2015; He et al., 2017a; Jin Young Kim, Kim, Lim, & Ch oi, 2016; Meana et al., 2018; S. Zhao et al., 2019) . Figure 5 - 3 shows the output from PWMScan for the LPIN1 region on the genome (Ambrosini, Groux, & Bucher, 2018) . The top vertical line s in the green box indicate predicted sit es for XBP1 binding upstream of the LPIN1 transcription start site (TSS) under the heading PWMScan motif matches (in 104 red box) . This is followed by the position of the transcript set for LPIN1 (highlighted in green box) . Figure 5 - 3 Prediction of XBP1 binding sites on LPIN1 from the PWMScan output . The CpG island track shows the presence of CpG islands, commonly found near the TSS. Genes with CpG islands in their promoter regions have distinctive patte rns of transcription initiation and methylated CpG islands are usually associated with stable loss of gene expression (Deaton & Bir d, 2011) . The size of the CpG islands is indicated (85 and 29 bases). - shows levels of each transcript as assayed by RNA - seq in 9 cell lines. The layered H3K4Me1, H3K4Me3 and H 3K27AC track s shows the enrichment level of the histone mark s at each position in th e genomic region of LPIN1 . These modifications suggest regulatory sites for the transcription of that protein. 105 Figure 5 - 4 Heat map showing the correlation between XBP1, LPIN1, MOGAT2, AGPAT1, GPAT2, PLD1 in the TCGA BRCA dataset . Yellow denotes low levels of expression and red denotes high levels of gene expression. A heat map showing the correlation between gene expression of XBP1 and LPIN1 in the TCGA BRCA dataset showed a significant negative correlation (p<0.05) ( Figure 5 - 4 ). Rstudio was used to develop this heat map (Rstudio Team, 2015) . However, XBP1 expression levels did not seem to affect lev els of other genes in the triglyceride synthesis and glycerolipid synthesis pathway s such as MOGAT2, AGPAT1, GPAT2, and PLD1. Correlation analysis of LPIN1 vs XBP1 mRNA Z - scores shows a significant negative - 0.58 (p - value<0.001) ( Figure 5 - 5 , left). All these lines of evidence ( Figure 5 - 3 , Figure 5 - 4 ) taken together with the JASPAR predictions ( Table 5 - 1 ) predict XBP1 regulation of transc ription of LPIN1. 106 A Kaplan - Meier survival analysis comparing low and high LPIN1 expression in the BRCA database shows a significant difference in survival (p - value=0.004) in groups with high levels of LPIN1 ( Figure 5 - 5 , right). The LIHC and COADREAD database s show a similar trend, however, due to the small number of patient samples, it did not show statistical significance. Figure 5 - 5 Correlation of XBP1 vs LPIN1 and Kaplan - Meier survival curve for BRCA and PANCAN . ( Top l eft) Correlation plot of LPIN1 vs XBP1 mRNA Z - scores using data from TCGA BRCA database . The Pearson's coefficient (r = - 0.585 ) is displa yed along with the p - value. 107 Figure 5 - (Right) A Kaplan - Meier survival curve showing log - rank test statistics for survival of low vs high LPIN1 expression in the TCGA BRCA database with a p - value of 0.004 ( Bottom l eft) Correlation plot of LPIN1 vs XBP1 mRNA Z - scores from the TCGA PANCAN database . The Pearson's coefficient (r = - 0.19 ) shows a significant negative correlation ( p - value <0.0001) . ( Bottom r ight) A Kaplan - Meier survival curve showing log - rank test statistics for survival of low vs high LPIN 1 expression in the TCGA PANCAN database with a p - value <0.001 Figure 5 - 6 shows the relative expression levels of Hep3B cells wild type or IRE1 - / - KO treated with PA for 24 or 48 hrs normalized to a BSA control by calculating the value . As is evident from the figure, LPIN1 expression is reduced at 24 hrs of PA treatment (relative to BSA control) in Hep3B wild type cells. Figure 5 - 6 Relative gene expression for LPIN1 . Hep 3B wild type and IRE1 - / - KO treated with palmitate for 24 or 48 hrs and primers specific for LPIN1 were used to calculate the fold change in LPIN1 gene expression (* p<0.05) (n=3 ) 108 This downregulation is not observed in Hep3B IRE1 - / - KO cells on PA treatment ( Figure 5 - 6 ) . Given the strong negative correlation of LPIN1 with XBP1 ( Figure 5 - 5 ) , this - XBP1 may be involved in the downregulation of LPIN1 on PA treatment. As seen in Figure 5 - 5 , high levels of LPIN1 correlate with poor prognosis in BRCA and PANCAN databases. This is corroborated by independent observations in breast, prostrate, lung and other cancers (Brohée et al., 2015; X. Fan et al., 2018; He et al., 2017b; Me ana et al., 2018) . LPIN1 mRNA expression is reduced in obesity (Miranda et al., 2010) . Our data suggest a mechanism by which this may be taking place. Increased PA levels as se (J. Han & Ka ufman, 2016; Pfaffenbach et al., 2010; Dong Wang et al., 2006) . LPIN1 may be transcriptionally repressed through recruitment of a repressor protein like Mist1 by XBP1 (Acosta - Alvear et al., 2007; Lemercier, To, Carrasco, & Konieczny, 1998) (RIDD ) may also be responsible for the lower levels of LPIN1 in Hep3B wild type cells treated with PA pointing to a protective effect of IRE1 activation . The method of regulation of LPIN1 by XBP1 needs to be studied in further detail to elucidate the involvement of IRE1, Mist1 and RIDD. 5.3.2 ACAT1 ACAT1, also known as stearoyl - O - acetyltransferase 1 (SOAT1) catalyzes the conversion of cholesterol to cholesterol esters. This eventually leads to the storage of fat in the form of lipid droplets. ACAT1 knockdown led to a decrease in tumor size in mouse xenograft experiments (J. Fan et al., 2014) . In concert with SIRT3, ACAT1 promoted glycolysis in cancer cells and promot es the Warburg effect. ACAT1 and ACAT2 109 transcription increases on treatment with free fatty acids (FFA) (Antalis et al., 2010; Seo et al., 2001) . Figure 5 - 7 ACAT1 gene expression and correlation with XBP1 and Kaplan - Meier survival curve . ( Top l eft) Correlation of XBP1 and ACAT1 expression analysis (Pearson's coefficient r=0. 022, p<0.00 5 ). ( Top r ight) Relative gene expression for ACAT1 in 3B wild type and IRE1 - / - KO treated with palmitate for 24 or 48 hrs . (p<0.05) (n=3) (Bottom) Kaplan Meier survival curve for ACAT1levels in TCGA - PANCAN. The samples are stratified by low and high ACAT1 levels in the pan - cancer (PANC AN) database. Low levels are correlated with poor prognosis. and formation of spliced XBP1. ACAT1 levels increase with exposure to PA after 24 hrs 110 but return to normal level s at the 48 hr time point ( Figure 5 - 7 ). This increase is not seen in IRE1 - / - KO cells pointing to an IRE1 - XBP1 dependent mechanism of action. TCGA PANCAN survival data shows a reversal of trend for ACAT1 levels in certain cancer types eg. bladder cancer, liver hepatocellular carcinoma making it difficult to conclusively analyze the PANCAN data ( Figure 5 - 7 , bottom) . 5.3.3 IDH1 Isocitrate dehydrogenase 1 (IDH1) is responsible for the conversion of isocitrate to 2 - oxyglutarate. IDHs are known to be important in lipogenesis and controlled by sterol regulatory element SREBP1 (Shechter, D ai, Huo, & Guan, 2003) . It has been found to be highly mutated in breast, liver, colon cancer as well as gliomas, sarcomas and other subtypes of cancer (Dang, Yen, & Attar, 2016; Fathi et al., 2014; X. Ma, Wang, Zhang, & Gazdar, 2013) . Figure 5 - 8 IDH1 gene expression and correlation with XBP1. (Left) Correlation of XBP1 and IDH 1 expression analysis (Pearson's coefficient r= - 0. 515 , p<0.0001). (Right) Relative gene (PA/BSA) expression for IDH 1 in 3B wild type and IRE1 - / - KO treated with palmitate for 24 or 48 hrs (p<0.05) (n=3) 111 Figure 5 - 9 Kaplan Meier survival curve for IDH1levels in PANCAN (left) LIHC ( middle ) and BRCA (right). The samples are stratified by low and high IDH1 levels in the pan - cancer (PANCAN) , liver cancer (LIHC) and breast cancer (BRCA). High IDH1 levels are correlated with poor prognosis in both liver and breast cancers . Figure 5 - 8 (left) shows the correlat ion between IDH1 and XBP1 in the pan - cancer database with a very robust (p<0.001) - 0.515. On analysis of gene expression levels, we found that , in Hep3B cells, IDH1 levels increase in response to s and then goes down in expression after 48 hrs ( Figure 5 - 8 , right). A Kaplan Meier survival analysis shows that high IDH1 expression leads to poorer prognosis in liver (L IHC) and breast (BRCA) cancer with a hazard ratio of ~1.4 ( Figure 5 - 9 ) . The hazard ratio is a measure of the difference between the high and low expression curves in the KM plot (Spruance, Reid, Grace, & Samore, 2004) . A value of 1 indicates that there is no difference. A value of 1.4, as seen in Figure 5 - 9 indicates that patient survival for high expressing IDH1is lower than for low g ene expression signatures of IDH1 in both liver and breast cancer . Similar to ACAT1, there was a discrepancy in the survival trends for different types of cancer. Additional in - depth investigation is required to understand the role of this gene as a functi on of different types of cancer. 112 5.3.4 PDK1 Pyruvate dehydrogenase kinase 1 is responsible for phosphorylation of the pyruvate dehydrogenase subunits PDHA1 and PDHA2 which inhibits pyruvate dehydrogenase activity thereby switching tumor metabolism towards glyco lysis (L. Yu, Chen, Sun, Wang, & Chen, 2017) . Recent evidence has implicated PDK1 inhibition in the induction of cell death in breast cancer, glioblasto ma multiforme and metastasis of liver cancer to breast and colorectal sites (Dupuy et al., 2015; Emmanouilidi & Falasca, 2017; J. E. Han et al., 201 7; Qin et al., 2019; Raimondi & Falasca, 2011) . Induction of PDK1 was seen under hypoxic condition accompanied by activation of the HIF1A factor (Dupuy et al., 2015; J. E. Han et al., 2017) . As seen in Figure 5 - 10 , we found that PDK1 activation by PA within 48 hrs. This increase is in an IRE1 dependent manner. A study on colorectal cancer, breast cancer and glioblasto ma multiforme cell lines suggested that increase in PDK1 levels was through the HIF1A factor , which has been associate d with cancer (Imamura et al., 2009; Semenza, 2013) . Figure 5 - 10 qPCR and Kaplan - Meier survival curve for gene expression levels of PDK1 . (Left) Hep3B WT and IRE1 - / - KO cells were treated with PA or BSA (control) and the gene expression levels for PA as compared to BSA were analyzed for 24 and 48 hrs. A 113 Figure 5 - significant increase (* p<0.05) is seen in gene expression of PDK1 at 24 hrs and 48 hrs in wild type Hep3B cells but not in IRE1 - / - KO cells treated with PA . qPCR expression significantly decreased at 48 hrs between Hep3B WT and IRE1 - / - KO cells (*p<0.05) (n=3) (Right) Kaplan - Meier survival analysis for PDK1 using samples from the TCGA - PANCAN databa se. The samples are stratified by low (dark blue), mid - range (light blue) and high (red) PDK1 levels. High PDK1 levels are correlated with poor prognosis (p - value<0.0005). A study on triple negative breast cancer cells suggested that XBP1 regulates HIF1A targets (X. Chen et al., 2014) . PDK1 was also seen to be enriched in breast cancer stem cells and non - small cell lung cancer cells, whereas inhibition or depletion of PDK1 decreased tumor growth and decreased migration (Dupuy et al., 2015; T. Liu & Yin, 2017; Peng et al., 2018) . 5.3.5 PFKFB4 The enzyme 6 - phosphofructo - 2 - kinase or fructose - 2,6 - bisphosphatase 4 (PFKFB4) is responsible for synthesis and degradation of fructose 2,6 - bisphosphate. It is a bifunctional enzyme possessing both kinase and phosphatase activities. Fructose 2,6 - bisphosphate allosterically activates glycolysis (Chesney, Clark, Lanceta, Trent, & Telang, 2015) . PFKFB4 has been shown to be essential for tumor survival and progression for breast, liver , colon cancer as well as glioblastomas, prostate cancer (Chesney et al., 2014; R. Gao et al., 2018; Goidts et al., 2012; O. Minchenko, Opentanova, & Caro, 2003; Ros et al., 2012) . The kinase activity of PFKF B4 has recently been shown to activate the steroid receptor co - activator ( Src - 3 ) suggesting a role in transcriptional reprogramming (Dasgupta et al., 2018) . Src3 is an important oncogene both in hormone depe ndent and independent cancers such as breast, liver, colorectal cancer (Yan, Tsai, & Tsai, 2006) . Inhibitors against PFKFB4 have be en shown to inhibit xenograft growth and are postulated to be viable chemotherapeutic targets (Chesney et al., 2015; Telang, Trent, Chesney, & Clark, 2015) . 114 Figure 5 - 11 qPCR and Kaplan - Meier survival curve for gene expression levels of PFKFB4. (Left) Hep3B WT and IRE1 - / - KO cells were treated with PA or BSA (control) and the gene expression levels for PA as compared to BSA were analyzed for 24 and 48 hrs. A significant increase (* p<0.05) is seen in gene expression of PFKFB4 at 48 hrs in wild type Hep3B cells but not in I RE1 - / - KO cells treated with PA (n=3) (Right) Kaplan - Meier survival analysis for PFKFB4 using samples from the TCGA - PANCAN database. The samples are stratified by low (dark blue), mid - range ( white ) and high (red) PFKFB4 levels. High PFKFB4 levels are correl ated with poor prognosis (p - value = 1.152 e - 1 1 ) . Our experiments found that PFKFB4 is upregulated in Hep3B wild type cells in response to PA and this increase is abrogated on knockout of IRE1 ( Figure 5 - 11 ). Based on our findings and previous studies, PFKFB4 may be regulated in a XBP1 - HIF1A dependent manner (O. H. Minchenko et al., 2005; H. Zhang et al., 2016) 5.3.6 CAV1 Studies from our research group and others have shown that CAV1 is amplified in metastatic tumors (Burgermeister, Liscovitch, Röcken, Schmid, & Ebert, 2 008; Chatterjee et al., 2015; Nath & Chan, 2016; Qian et al., 2019; Sugie et al., 2015; T. Yu et al., 2012) . CAV1 is the major structural protein found in caveolae and also has roles in lipid homeostasis and signal transduction via the tyrosine kinase Fyn (Chatterjee et al., 2015; Rudick & And erson, 2002) . 115 Figure 5 - 12 qPCR and Kaplan - Meier survival curve for gene expression levels of CAV1. (Left) Hep3B WT and IRE1 - / - KO cells were treated with PA or BSA (control) and the gene expression le vels for PA as compared to BSA were analyzed for 24 and 48 hrs. A significant increase (* p<0.05) is seen in gene expression of CAV1 at 24 and 48 hrs in wild type Hep3B cells but not in IRE1 - / - KO cells treated with PA (n=3) (Right) Kaplan - Meier survival an alysis for CAV1 using samples from the TCGA - PANCAN database. The samples are stratified by low (dark blue), mid - range (white) and high (red) CAV1 levels. High CAV1 levels are correlated with poor prognosis (p - value= 2.798e - 8 ). CAV1 levels increase in an IRE1 dependent manner on treatment with PA. Figure 5 - 12 depicts the qPCR results (left) and Kaplan - Meier survival analysis for CAV1 (right) . Hi gh levels of CAV1 correlate with a poor prognosis as seen in the TCGA PANCAN database. 5.3.7 EMT - Transcription Factors Epithelial - to - mesenchymal transition (EMT) is mediated by three major groups of transcription factors (TFs): Snail, TWIST and ZEB (Danhier et al., 2017; Sánchez - Tilló et al., 2012) . Snail levels were increased in response to PA treatment both in IRE1 WT cells as well as IRE1 - / - KO cells. This indi cates that Snail levels are responsive to treatment with saturated fatty acids, although this may not be through activation of XBP1 or IRE1. We 116 found during our qPCR experiments that TWIST and ZEB1 expression levels increased in ated by PA ( Figure 5 - 13 ) . Figure 5 - 13 qPCR for expression levels of ZEB1 and TWIST. Hep3B WT and IRE1 - / - KO cells were treated with PA or BSA (control) and the gene expre ssion levels for PA as compared to BSA were analyzed for 24 and 48 hrs. (Left) qPCR for ZEB1 levels in Hep3B WT and IRE1 - / - KO treated with PA for 24 and 48 hrs. Levels of ZEB1 (PA/BSA) increase nearly 2 fold at 24 hrs in wild type cells and not in IRE1 - / - KO cells (*p<0.05) (n=3) . (Right) qPCR for TWIST levels in Hep3B WT and IRE1 - / - KO treated with PA for 24 and 48 hrs. Leve ls of TWIST (PA/BSA) increase nearly 2.7 fold at 48 hrs in wild type cells and not in IRE1 - / - KO cells (*p<0.05) (n=3) ZEB1 and ZEB2 are zinc finger domain transcription factors responsible for repression of certain genes, e.g. E - cadherin in EMT. TWIST transcription factors have a basic helix - loop - - terminal end (Sánchez - Tilló et al., 2012) . TWIST is bo th an activator and repressor depending on which other regulating proteins it binds to. TWIST has been reported to repress E - cadherin and activate N - cadherin expression during EMT. Our results are corroborated by other studies that investigated upregulation of TWIST and ZEB TFs by XBP1s (Cuevas et al., 2017) . 117 5.3.8 Effect of XBP1 levels on survival We looked at the effect of XBP1 levels on survival in different types of cancers. Contrary to what has been reported in some studies (X. Chen et al., 2014) , we found that a lower XBP1 expression level correlated with a worse survival probability for breast cancer samples ( Figure 5 - 14 ) . When we classified samples accordin g to ER, PR and HER2 status and considered a cohort of triple negative breast cancer (TNBC) samples , the results were the same. Figure 5 - 14 Kaplan - Meier survival curve for BRCA, triple negative breast cancer (TNBC) , LIHC and COADREAD. (Left) All samples from the TCGA - BRCA datasets were stratified according to high (red) and low (blue) XBP1 levels . (Right) Triple negative breast cancer samples were chosen from the TCGA - BRCA dataset and stratified by high (red) and low (blue) XBP1 levels. 118 5.3.9 Using the TCGA datasets to determine XBP1 targets We also decided to use another approach to find other unknown targets that could be regulated by XBP1. In this approach we used TCGA datasets for liver cancer (L IHC), breast cancer (BRCA) and colorectal cancer (COADREAD). The datasets were downloaded on 04.01.2019 from the GDC portal. mRNA gene expression data was used to determine the correlation of XBP1 levels with all the genes in the TCGA dataset. Only signifi cant correlations with a p - value <0.05 were used for further analysis. From the pool, , LIHC and COADREAD datasets respectively. Figure 5 - 15 Venn diagram showing the number of genes with a positive correlation with XBP1 from liver cancer ( LIHC ) , breast cancer ( BRCA ) and colorectal cancer ( COADREAD ) databases and their overlap . 119 Figure 5 - 16 3 43 genes from LIHC, BRCA and COADREAD sectioned by gene ontology (GO) biological processes using PANTHER tools. Figure 5 - 17 Overrepresented genes in BRCA, LIHC and COADREA D datasets as compared to Homo sapiens reference genes. (Top) Bar chart showing the percentage of overrepresented genes from the input dataset (blue) as compared to the reference dataset (red). (Bottom) A pie - chart representation of over - represented genes with each color representing different Gene Ontology sets. 120 The Venn diagram in Figure 5 - 15 shows 3 45 genes that are significantly correlated with XBP1 levels in all three cancer types - LIHC, BRCA, COADREAD. We further analyzed those 3 45 genes using Gene Ontology tools. Out of 3 45 gen es, 3 43 genes were mapped to known human refseqs and divided into classes according to gene ontology for biological processes ( Figure 5 - 16 ) and also according to biologic al pathways ( Figure 5 - 18 ). 121 Figure 5 - 18 3 43 genes from LIHC, BRCA and COADREAD sectioned into 10 1 gene ontology (GO) biological pathways using PANTHER tools. 122 Figure 5 - 19 Pie chart showing the categorization of genes involved in metabolic processe s by sub - categories and levels of processes . 123 Figure 5 - 17 compares the input gene list with the Homo sapiens reference gene dataset and classifies the genes by their categories and over - representation. The y - axis in Figure 5 - 17 (top) represents the percentage of genes in the gene list with blue denoting the input gene percentage and red denoting the reference gene percentage. As is evident from Figure 5 - 17 (bottom) genes from the macromolecule processes and metabolic processes are significantly overrepresented in the input dataset. We were able to further categorize these genes by the ir individual pathways as shown in Figure 5 - 18 . Each section of the pie chart represents gene ontology curated pathways for the overrepresented genes. Figure 5 - 19 depicts the steps of each selection with an example of selecting genes for fatty acid and carbohydrate metabolic processes. We select the metabolic processes section and pick th e primary metabolic processes slice. The levels in that slice include carbohydrate metabolic processes which when selected is divided into three categories: cellular carbohydrate processes, monosaccharide and polysaccharide metabolic processes. Opening the se levels of processes will display a list of genes that are overrepresented for the biological process. We also visualized this process for fatty acid metabolic processes as displayed in Figure 5 - 19 (bottom). This analysis can be used to highlight the fold change in gene expression . The datasets used can be microarray datasets from Gene Expression Omnibus (GEO) or RNA - seq data acquired on treatment with PA in Hep3B wild type or IRE1 - / - KO cells. Overrepresented genes can then be quantified with a fold change value and a p - value to denote significance. This strategy will aid us in analyzing global changes to metabolism due to activation of I RE1 - XBP1. Using gene expression values from samples treated with BSA or PA in wild type or IRE - / - KO cells , we can separate the effects of PA from the 124 effects of XBP1 splicing on metabolic changes. This will lead to a deeper understanding of how tumors dev elop and proliferate. 125 6 FUTURE WORK AND CONCLUSIONS 6.1 Importance of junction plakoglobin ( JUP ) in cancer 6.1.1 Junction plakoglobin: Desmosomal component In Chapter 5, we looked at the effect of PA on DSP expression and its implications on survival in the TCGA cancer database. Plakoglobin, a member of the armadillo family of proteins, is also a component of desmosomes. Plakoglobin (JUP) and plakophilins associate with the desmosomal cadherins and DSP to form the desmosome. Experiments have shown that JUP is also loosely associated with E - cadherin in the adherens junction although further research is needed to quantify and elucidate the association (Aberle, Schwartz, & Kemler, 1996; Aktary & Pasdar, 2012) . 6.1.2 JUP in cell signaling - - catenin, both belonging to the armadillo family of proteins. Both the catenins are highly homologous to each other ( Figure 6 - 1 - catenin, JUP can take its place in adherens junctions. However, there is still some disagreement on - catenin targets (Aktary & Pasdar, 201 2; Maeda et al., 2004; Shimizu, Fukunaga, Ikenouchi, & Nagafuchi, 2007) . Previous research in our group has shown that on knockdown of desmoplakin, JUP is released from desmosomes and it translocates to the nucleus (Nath, 2015) . It will be interesting to see whether this translocatio n leads to a non - canonical activation of Wnt signaling pathways. 126 Figure 6 - 1 T - coffee expresso alignment result for human beta - catenin and JUP. The scale for the quality of the alignment is given at the top of the figure with pink indicating a good alignment. Both armadillo proteins are highly homologous except for the N - termina l and C - terminal ends. . 127 This can be done with the help of TOP/FOP luciferase reporter plasmids (Qu et al., 2016; C. Zhao, 2014) - catenin binding sites from T - cell factor (TCF) while FOP Flash plasmids have mutated TCF binding sites that can - catenin. 6.1.3 Regulation of JUP Figure 6 - 2 shows the correlation between JUP and levels of the transcriptional repressors ZEB1 and ZEB2. The correlation is depicted as a heat map ( Figure 6 - 2 , left) with mRNA normalized Z - scores from the TARGET and TCGA PANCAN database. Figure 6 - 2 Correlation between JUP and transcription factors ZEB1 and ZEB2 . (Left) Heat map showing correlation between JUP and transcription factors ZEB1 and ZEB2 in the TARGET and TCGA PANCAN database. (Right) Correlation between gene expression values of JUP vs ZEB1 (top) and JUP vs ZEB2 (bottom) showing a strong negative co rrelation at a p - value <0.0001 128 - 0.5 (p - value<0.0001) and similarly for ZEB2 was - 0.598 (p - value<0.0001) ( Figure 6 - 2 , right). In Chapter 4, we have pres ented our data showing regulation of DSP through ZEB1 and ZEB2. It is possible that JUP is regulated in a similar manner. 6.2 Chemoresistance/chemo - tolerance 6.2.1 Role of XBP1 in resistance to chemotherapy Splicing of XBP1 has been implicated in the progression of various cancers like multiple myeloma, hepatocellular carcinoma, breast cancer etc. (Carrasco et al., 2007 ; A. - H. Lee et al., 2003; K. Lee et al., 2002) . Tumor survival under hypoxic conditions needs - XBP1 branch of the UPR (Romero - Ramirez et al., 2004) . ER stress pathways were shown to be activated constitutively in drug - resistant U266 cells as opposed to control drug - sensitive cell lines and correlated with a worse prognosis in multiple myeloma patients (Nakamura et al., 2006) . The IRE1 - XBP1 branch has also been shown to be invol of tumor - associated dendritic cells. This is important for long - term effectiveness of chemotherapeutic drugs (Rubio - Patiño, Bossowski, Chevet, & Ricci, 2018) . 6.2.2 Effect of IRE1 - XBP1 activation by PA on chemo - tolerance to CUDC - 101 Our experiments using Hep3B WT and IRE1 - / - KO cell lines treated with the drug CUDC - 101 gave interesting results. Briefly, 7000 cells in a 96 well plate were seeded (both WT or IRE1 - / - KO cells ) and were treated with BSA or 0.4mM PA media for 24 hrs to induce ER stress. They were then treated with CUDC - 101 for 24 hrs , a multitarget inhibitor of HDAC, EGFR and HER2. CUDC - 101 is currently in clinical trials and indicated for head and neck squamous cell carcinoma ( Figure 6 - 3 ) (Galloway et al., 2015) . Different 129 concentrations of CUDC - 101 were used (0.025 , 0.05, 0.1, 0.15, 0.3 and 0.5 µM) and the viability of the cells was measured with an Alamar blue assay (Ex 530 - 560 nm / Em 590 nm ). The relative fluorescence units (RFU) are a measure of cell viability after treatment. Figure 6 - 3 Structure of CUDC - 101, a multitarget inhibitor against HDACs, EGFR, and HER2 receptors. (Image taken from https://www.invivochem.com/cudc - 101/ ) GraphPad Prism was used to fit the RFU values to the Hill equation to cal culate the IC50 values for each treatment. The Hill equation is where Y is the RFU for different concentrations and X is the concentration of drug used. As seen in Figure 6 - 4 , the IC50 values of the WT cells treated with PA differed s ignificantly from the IC50 value of IRE1 - / - KO cells treated with PA. This difference was also seen in the IC50 value of BSA treated WT cells indicating that the increase in chemo - tolerance (and therefore, an increase in IC50 value) was due to the presence and activation 130 Figure 6 - 4 IC50 curves for treatment of Hep3B cells with CUDC 101 . IC50 curves for Hep3B wild type (left) and IRE1 - / - KO (right) cells treated with CUDC - 101 and normal media, BSA media or PA media. The IC50 values (in µM) are indicated in the tables below the graphs We would like to make the distinction here between chemo - tolerance and chemo - resistance. Tolerance is defi ned as a diminished response to a drug whereas resistance is the ability of cancer cells to withstand the drug. Tolerance can be offset by treating with an increased dose of the drug. Resistance, on the other hand, occurs because of a mutation in the prote in t hat the drug target s . This can give rise to a sub - population of cells that all carry the mutation and are not responsive to the drug. Chemoresistance is generally not reversible. Drug tolerance on the other hand, is reversible within a few doublings of the cells (Sharma et al., 2010) . In our case, PA increases drug tolerance which is mediated through activation of ER stress and the IRE1 - XBP1 axis. It is r eversible once the ER stressor PA is removed. However, the mechanism that leads to increased tolerance needs to be elucidated. We hypothesize that it could be because of enhanced survival through activation of the pro - survival XBP1s activity by PA . Anothe r interesting point to note here is that CUDC - 101 is a reported EGFR Crystallography structure co - ordinates for EGFR (PDB ID: 3W2S) and IRE1 kinase (PDB 131 ID: 3P23 - 1 ) were obtained f rom the RCSB PCD database. Using PDBeFold to align the EGFR and IRE1 kinase domains, we obtained a root mean square deviation (RMSD) value of 2.441 Å. The aligned domains are shown in Figure 6 - 5 with EGFR in orange and IRE1 in green. This similarity between the domains suggests that we may be observing an off - target effect of CUDC - 101. Analyzing this would lead to a more comprehensive understanding of how chemo - tolerance develops in cancer cells. Then there is the question of whether this effect is limited to EGFR inhibitors or if it applies to multiple classes of inhibitors. This would depend on whether the che mo - tolerance is mediated through a direct CUDC - 101 interaction with IRE1. Figure 6 - 5 EGFR kinase domain (orange) and IRE1 kinase domain (green) aligned using PDBeFold. PyMol was used to visualize the proteins. The RMSD value is calculated to be 2.441Å 132 IRE1 kinase is also very similar to the c - Abl kinase domain (PDB ID: 6NPE) ( Figure 6 - 6 ) . The c - Abl protein is implicated in the progression of chronic myeloid leukemia (CML). Imatinib, a drug designed against c - Abl binds to its kinase domain. It will be i nteresting to perform a chemo - tolerance assay with PA treated wild type and IRE1 - / - KO cells to see whether a similar increase in IC50 values is observed. Does this increase in cell viability stem from an increase in EMT markers? Throughout the preceding chapters, we have shown how IRE1 - XBP1 affects cell migration, increases expression of EMT - transcription factors in the presence of PA and possibly affects t he overall metabolism of the cell. Western blot and qPCR analysis of EMT genes will shed more light on whether drug tolerance due to PA leads to activation and progression of EMT. Figure 6 - 6 Alignment of I RE1 and c - Abl kinase using PDBeFold. PyMol was used to visualize the proteins. Green : c - Abl kinase domain, Magenta : IRE1 kinase domain . RMSD value: 4.24. 133 We have highlighted the difference between chemo - tolerance and chemo - resistance. However, there m ay be a possibility that continuous treatment of PA and the drug may lead to the emergence of a chemo - resistance population of cells. It will be interesting to see how this chemo - resistant sub - population differs from the original population in terms of sen sitivity and response to ER stressors. 6.3 Transmissible ER stress and chemo - resistance Transmissible ER stress (TERS) is demonstrated in communication of cancer cells with each other, other cells in the tumor micro - environment , as well as cells of the immun e system. Mahadevan et. al showed that macrophage cells treated with conditioned media from ER - stressed cells showed an increase in UPR genes (Mahadevan et al., 2011) independent of interleukin 6 (IL - 6) signaling. Upregulation of BiP, CHOP and XBP - 1 was observed (Mahadevan et al., 2011) . ER stress is transmissible regardless of whether the stressor is physiological or pharmacological and is potentiated by toll - like receptor 4 (TLR4). 6.3.1 TERS confers higher survival in tumors TERS is med iated through secretion of soluble factors that can transmit the ER similar to that seen in chronic adaptive UPR responses (Rodvold et al., 2018) . Further stresses on TER primed cells led to initiation of a pro - survi val response in heterologously and homologously treated cancer cells. This may lead to increased cell fitness seen in tumor cells with an increased ability to cope with nutrient deprivation and hypoxia. 134 6.3.2 TERS primed cells are resistant to chemotherapeutic drugs Recently, it has been demonstrated through our preliminary work and through other studies, that ER stress may be responsible for induction of drug - tolerance or resistance to chemotherapeutic drugs (Rubio - Patiño et al., 2018; Sharma et al., 2010) . However, the mechanism is unknown. There is recent evidence that it could be mediated through TERS (Doron et al., 2018; Kanemoto et al., 2016; Rodvold e t al., 2018) . 6.3.3 TERS may be mediated through exosomes The soluble factors that transmit ER stress to other cells may do so through packaging of stress - related proteins or molecules in extracellular vesicles (EVs). EVs are formed by s of cells, both prokaryotic and eukaryotic. The quantity, frequency and contents of EVs may change depending on the cell type, nutrient availability and signaling needs. They can deliver, DNA, mRNA or proteins to neighboring cells and are therefore primer candidates for TERS. Hepatocytes are particularly prone to ER stress and a recent study found that treatment with PA stimulated release of EVs in hepatocytes (Kakazu, Mauer, Yin, & Malhi, 2016) - XBP1, are involved in mediating EV formation and release, possibly through SPT1. EVs enhance angiogen esis and promote tumor metastases and are instrumental in evading apoptosis (Costa - . EVs are released after stimulation of neurons and are important in neuronal communication. With the evidence that Spt1 and ceramide is packaged in exosomes, EVs ma y be a propagating factor for neurodegenerative diseases (Budnik, Ruiz - Cañada, & Wendler, 2016; Kakazu et al., 2016) . 135 6.4 C onclusions 6.4.1 , Endoplasmic Reticulum stress and the UPR in the dimerization and activation of protein. Multiple studies including ours have shown the lipid sensing properties of certain transmembrane proteins (IRE1, MGA2) (Cho et al., 2019; Covino et al., 2016; Covino, Hummer, & Ernst, 2018; Romain Volmer & Ron, 2015) . Our work focused on developing an assay to measure protein - protein interaction for transmembrane protein IRE1 . With the help of an alanine scanning assay, we were able to investigate the effect of individual residues on the dimerization ability . We identified two residues S450 and W457 on the transmembrane domain that are important drivers for dimerization. S450A and W457A mutations lead to loss of XBP 1 splicing activity. The current study suggests a novel role of the transmembrane domain in the regulation and activation of IRE1 through sensing of membrane fluidity and saturation. We also in vestigated the hypothesis of direct binding and activation of IRE1 by PA. - CD from insect cells to obtain a hypo - phosphorylated form of the protein. This protocol can be adapted for use with other mammalia n proteins requiring post - translational modifications that are too complex for a - CD has the potential to bind PA directly. Mutational studies and ligand binding assays will help furt her determine which amino acid residues are important in binding of PA. This may be mitigate the deleterious effects of ER stress in disease pathologies. 136 With obesity becom ing a global epidemic, we need to understand the effects of rising levels of fatty acids in the plasma on various human diseases. There is a wealth of evidence on the relationship between obesity, cardiovascular disease, cancer, neurodegeneration and many more (Y. Ma et al., 2013; Mazon, de Mello, Ferreira, & Rezin, 2017; Dingzhi Wang & Dubois, 2012) . The effect of obesity on ER stress is an active field of research and will prove crucial in understanding the pathology behind progression of multiple diseases . 6.4.2 Cell adhesion has long been recognized to be important in migration and metastases of cancer cells. We determined that loss of DSP , a cell adhesion protein was enough to promote migration of cancer cells. The ZEB family of EM T - TF is involved in the regulation of DSP levels. We also determined that ZEB levels are responsive to the s ZEB1 and ZEB2 levels by transcriptional activation , however, f urther studies need to be performed to identify the binding regions for XBP1. 6.4.3 The Warburg effect has been document ed for some decades now but the effect of ER stress and chronic inflammation on incidence and progression of cancer is actively being investigated (Hanahan & Weinberg, 2011; So et al., 2012; Vander Heiden et al., 2009; L. Yu et al., 2017) and XBP1 splicing may be importa nt in global metabolic changes observed in certain cancers. Increased incidence of obesity and increased plasma levels of saturated free fatty acids contribute to the activat ion of and splicing of XBP1 . The XBP1 transcription 137 factor has binding sites on various metabolic genes including those found in glycolytic, gluconeogenesis, fatty acid oxidation, purine and pyrimidine synthesis pathways. We studies on how met abolism is altered especially due to fatty acids that can activat e 138 APPENDIX 139 APPENDIX 140 REFERENCES 141 REFERENCES 1. Aberle, H., Schwartz, H., & Kemler, R. (1996). Cadherin - catenin complex: Protein interactions and their implications for cadherin function. Journal of Cellular Biochemistry , 61 (4), 514 523. https://doi.org/10.10 02/(SICI)1097 - 4644(19960616)61:4<514::AID - JCB4>3.0.CO;2 - R 2. Ackerman, D., & Simon, M. C. (2014). Hypoxia, lipids, and cancer: Surviving the harsh tumor microenvironment. Trends in Cell Biology , 24 (8), 472 478. https://doi.org/10.1016/j.tcb.2014.06.001 3. Acosta - Dynlacht, B. D. (2007). XBP1 Controls Diverse Cell Type - and Condition - Specific Transcriptional Regulatory Networks. Molecular Cell , 27 (1), 53 66. https://doi.org/10.1016/j.molcel.2 007.06.011 4. Aktary, Z., & Pasdar, M. (2012). Plakoglobin: Role in Tumorigenesis and Metastasis. International Journal of Cell Biology , 2012 , 1 14. https://doi.org/10.1155/2012/189521 5. Alasiri, G., Fan, L. Y. N., Zona, S., Goldsbrough, I. G., Ke, H. L., Auner , H. W., & Lam, E. W. F. (2018). ER stress and cancer: The FOXO forkhead transcription factor link. Molecular and Cellular Endocrinology , 462 , 67 81. https://doi.org/10.1016/j.mce.2017.05.027 6. Ali, M. M. U., Bagratuni, T., Davenport, E. L., Nowak, P. R., Si lva - Santisteban, M. C., complex and implications for the unfolded protein response. The EMBO Journal , 30 (5), 894 905. https://doi.org/10.1038/emboj.2011.18 7. Ambrosini, G., Grou x, R., & Bucher, P. (2018). PWMScan: a fast tool for scanning entire genomes with a position - specific weight matrix. Bioinformatics (Oxford, England) , 34 (14), 2483 2484. https://doi.org/10.1093/bioinformatics/bty127 8. Antalis, C. J., Arnold, T., Rasool, T., Lee, B., Buhman, K. K., & Siddiqui, R. A. (2010). High ACAT1 expression in estrogen receptor negative basal - like breast cancer cells is associated with LDL - induced proliferation. Breast Cancer Research and Treatment , 122 (3), 661 670. https://doi.org/10.1007/s10549 - 009 - 0594 - 8 142 9. Arner, P., & Rydén, M. (2015). Fatty acids, obesity and insulin resistance. Obesity Facts , 8 (2), 147 155. https://doi.org/10.1159/000381224 10. Baenke, F., Peck, B., Miess, H., & Schulze, A. (2013). H ooked on fat: the role of lipid synthesis in cancer metabolism and tumour development. Disease Models & Mechanisms , 6 (6), 1353 1363. https://doi.org/10.1242/dmm.011338 11. Boden, G. (2008). Obesity and Free Fatty Acids (FFA). Endocrinology and Metabolism Clini cs of North America , 37 (3), 635. https://doi.org/10.1016/j.ecl.2008.06.007.Obesity 12. Boyden, L. M., Kam, C. Y., Hernández - Martín, A., Zhou, J., Craiglow, B. G., Sidbury, erythrokeratodermia - car diomyopathy syndrome. Human Molecular Genetics , 25 (2), 348 357. https://doi.org/10.1093/hmg/ddv481 13. Brohée, L., Demine, S., Willems, J., Arnould, T., Colige, A. C., & Deroanne, C. F. (2015). Lipin - 1 regulates cancer cell phenotype and is a potential target to potentiate rapamycin treatment. Oncotarget , 6 (13), 11264 11280. https://doi.org/10.18632/oncotarget.3595 14. Brown, K. A., & Simpson, E. R. (2010). Obesity and breast cancer: Progress to understanding the relationship. Cancer Research , 70 (1), 4 7. https://d oi.org/10.1158/0008 - 5472.CAN - 09 - 2257 15. Budnik, V., Ruiz - Cañada, C., & Wendler, F. (2016). Extracellular vesicles round off communication in the nervous system. Nature Reviews Neuroscience , 17 (3), 160 172. https://doi.org/10.1038/nrn.2015.29 16. Bujisic, B., & Ma rtinon, F. (2017). IRE1 gives weight to obesity - associated inflammation. Nature Immunology , 18 (5), 479 480. https://doi.org/10.1038/ni.3725 17. Burgermeister, E., Liscovitch, M., Röcken, C., Schmid, R. M., & Ebert, M. P. A. (2008). Caveats of caveolin - 1 in can cer progression. Cancer Letters , 268 (2), 187 201. https://doi.org/https://doi.org/10.1016/j.canlet.2008.03.055 18. Buzzetti, E., Pinzani, M., & Tsochatzis, E. A. (2016). The multiple - hit pathogenesis of non - alcoholic fatty liver disease (NAFLD). Metabolism: Cl inical and Experimental , 65 (8), 1038 1048. https://doi.org/10.1016/j.metabol.2015.12.012 143 19. Campbell, P. T., Newton, C. C., Freedman, N. D., Koshiol, J., Alavanja, M. C., diabet es, and risk of liver cancer for U.S. adults. Cancer Research , 76 (20), 6076 6083. https://doi.org/10.1158/0008 - 5472.CAN - 16 - 0787 20. Carrara, M., Prischi, F., Nowak, P. R., Kopp, M. C., & Ali, M. M. u. (2015). Noncanonical binding of BiP ATPase domain to Ire1 a nd Perk is dissociated by unfolded protein CH1 to initiate ER stress signaling. ELife , 4 , 1 16. https://doi.org/10.7554/eLife.03522 21. Carrasco, D. R., Sukhdeo, K., Protopopova, M., Sinha, R., Enos, M., Carrasco, D. E., ation and Stress Response Factor XBP - 1 Drives Multiple Myeloma Pathogenesis. Cancer Cell , 11 (4), 349 360. https://doi.org/10.1016/j.ccr.2007.02.015 22. Chatterjee, M., Ben - Josef, E., Thomas, D. G., Morgan, M. A., Zalupski, M. M., Khan, 015). Caveolin - 1 is associated with tumor progression and confers a multi - modality resistance phenotype in pancreatic cancer. Scientific Reports , 5 (December 2014), 1 15. https://doi.org/10.1038/srep10867 23. Chattopadhyay, A., & London, E. (1987). Parallax method for direct measurement of membrane penetration depth utilizing fluorescence quenching by spin - labeled phospholipids. Biochemistry , 26 (1), 39 45. https://doi.org/10.1021/bi00375a006 24. Chavarro, J. E., Ke nfield, S. A., Stampfer, M. J., Loda, M., Campos, H., Sesso, H. D., & Ma, J. (2013). Blood levels of saturated and monounsaturated fatty acids as markers of de novo lipogenesis and risk of prostate cancer. American Journal of Epidemiology , 178 (8), 1246 125 5. https://doi.org/10.1093/aje/kwt136 25. Chen, H. (2005). Boyden Chamber Assay. In J. Guan (Ed.), Methods in Molecular Biology: (Vol. 294, pp. 15 23). Totowa, NJ: Humana Press Inc. 26. Chen, R. (2012). Bacterial expression systems for recombinant protein producti on: E. coli and beyond. Biotechnology Advances , 30 (5), 1102 1107. https://doi.org/10.1016/j.biotechadv.2011.09.013 27. Chen, X., Iliopoulos, D., Zhang, Q., Tang, Q., Greenblatt, M. B., Hatziapostolou, M., - negative breast cancer by controlling Nature , 508 (7494), 103 107. https://doi.org/10.1038/nature13119 144 28. Ch en, Y., Wang, X., Wang, J., Yan, Z., & Luo, J. (2012). Excess body weight and the risk of primary liver cancer: An updated meta - analysis of prospective studies. European Journal of Cancer , 48 (14), 2137 2145. https://doi.org/10.1016/j.ejca.2012.02.063 29. Chesn ey, J., Clark, J., Klarer, A. C., Imbert - Fernandez, Y., Lane, A. N., & Telang, S. (2014). Fructose - 2,6 - bisphosphate synthesis by 6 - phosphofructo - 2 - kinase/fructose - 2,6 - bisphosphatase 4 (PFKFB4) is required for the glycolytic response to hypoxia and tumor gr owth. Oncotarget , 5 (16), 6670 6686. https://doi.org/10.18632/oncotarget.2213 30. Chesney, J., Clark, J., Lanceta, L., Trent, J. O., & Telang, S. (2015). Targeting the sugar metabolism of tumors with a first - in - class 6 - phosphofructo - 2 - kinase (PFKFB4) inhibitor. Oncotarget , 6 (20). https://doi.org/10.18632/oncotarget.4534 31. Chidgey, M., & Dawson, C. (2007). Desmosomes: a role in cancer? British Journal of Cancer , 96 (12), 1783 1787. https://doi.org/10.1038/sj.bjc.6603808 32. Cho, H. (2013a). Role of Palmitic Acid on the Kinase Proteins . Michigan State University. 33. Cho, H. (2013b). Roles of palmitic acid on the kinase proteins . ProQuest Dissertations and Theses . Michigan State University, Ann Arbor. Retrieved from http://ezproxy.msu.edu/login?url=http://search.proquest.com/ docview/1468949553?a ccountid=12598 34. C. (2011). Molecular mechanism by which palmitate inhibits PKR autophosphorylation. Biochemistry , 50 (6), 1110 1119. https://doi.org/10 .1021/bi101923r 35. Membrane Lipid Saturation. Cell Reports , 27 (1), 307 - 320.e5. https://doi.org/10. 1016/j.celrep.2019.03.017 36. Cho, H., Wu, M., Zhang, L., Thompson, R., Nath, A., & Chan, C. (2013). Signaling dynamics of palmitate - induced ER stress responses mediated by ATF4 in HepG2 cells. BMC Systems Biology , 7 (1), 9. https://doi.org/10.1186/1752 - 0509 - 7 - 9 145 37. Christofori, G., & Semb, H. (1999). The role of the cell - adhesion molecule E - cadherin as a tumour - suppressor gene. Trends in Biochemical Sciences , 24 (2), 73 76. https://doi.org/10.1016/S0968 - 0004(98)01343 - 7 38. Chuan Yin Liu, Schroder, M., & Kaufman, R. J. (2000). Ligand - independent dimerization activates the stress response kinases IRE1 and PERK in the lumen of the endoplasmic reticulum. Journal of Biological Chemistry , 275 (32), 24881 24885. https://doi.org/10.1 074/jbc.M004454200 39. Coll, T., Eyre, E., Rodríguez - Vázquez - Carrera, M. (2008). Oleate reverses palmitate - induced insulin resistance and inflammation in skeletal muscle cells. Journal of Biological Chemist ry , 283 (17), 11107 11116. https://doi.org/10.1074/jbc.M708700200 40. Costa - Lyden, D. (2018). Pancreatic cancer exosomes initiate pre - metastatic niche formation in the liver. Nat Cel l Biol , 17 (6), 816 826. https://doi.org/10.1038/ncb3169.Pancreatic 41. R. (2016). A Eukaryotic Sensor for Membrane Lipid Saturation. Molecular Cell , 63 (1), 49 59. https://d oi.org/10.1016/j.molcel.2016.05.015 42. Covino, R., Hummer, G., & Ernst, R. (2018). Integrated Functions of Membrane Property Sensors and a Hidden Side of the Unfolded Protein Response. Molecular Cell , 71 (3), 458 467. https://doi.org/10.1016/j.molcel.2018.07.0 19 43. Credle, J. J., Finer - moore, J. S., Papa, F. R., Stroud, R. M., & Walter, P. (2005). On the mechanism of sensing unfolded protein in the endoplasmic reticulum. Proceedings of the National Academy of Sciences of the United States of America , 102 (52), 1877 3 18784. 44. Antonsson, B. (2012). Differences in the glycosylation of recombinant proteins expressed in HEK and CHO cells. Journal of Biotechnology , 161 (3), 336 348. https://doi.or g/10.1016/j.jbiotec.2012.06.038 45. Cuevas, E. P., Eraso, P., Mazón, M. J., Santos, V., Moreno - bueno, G., Cano, A., & Portillo, F. (2017). LOXL2 drives epithelial - mesenchymal transition via activation of IRE1 - XBP1 signalling pathway. Scientific Reports , 7 (449 88), 1 11. https://doi.org/10.1038/srep44988 146 46. Dang, L., Yen, K., & Attar, E. C. (2016). IDH mutations in cancer and progress toward development of targeted therapeutics. Annals of Oncology , 27 (4), 599 608. https://doi.org/10.1093/annonc/mdw013 47. Danhier, P., Porporato, P. E. (2017). Cancer metabolism in space and time: Beyond the Warburg effect. Biochimica et Biophysica Acta - Bioenergetics , 1858 (8), 556 572. https://doi.org/10.1016/j.bbabio.2 017.02.001 48. Das, S. K., Chu, W. S., Mondal, A. K., Sharma, N. K., Kern, P. A., Rasouli, N., & Elbein, S. C. (2008). Effect of pioglitazone treatment on endoplasmic reticulum stress response in human adipose and in palmitate - induced stress in human liver and adipose cell lines. American Journal of Physiology. Endocrinology and Metabolism , 295 , 393 400. https://doi.org/10.1152/ajpendo.90355.2008. 49. Das, S. K., Mondal, A. K., & Elbein, S. C. (2010). Distinct gene expression profiles characterize cellular response s to palmitate and oleate. Journal of Lipid Research , 51 (8), 2121 2131. https://doi.org/10.1194/jlr.M004275 50. B. W. (2018). Metabolic enzyme PFKFB4 activates transcriptio nal coactivator SRC - 3 to drive breast cancer. Nature , 556 (7700), 249 254. https://doi.org/10.1038/s41586 - 018 - 0018 - 1 51. Deaton, A. M., & Bird, A. (2011). CGIs and the regulation of transcription. Genes Dev , 25 , 1010 1022. https://doi.org/10.1101/gad.2037511 52. De jeans, N., Barroso, K., Fernandez - Zapico, M. E., Samali, A., & Chevet, E. (2015). Novel roles of the unfolded protein response in the control of tumor development and aggressiveness. Seminars in Cancer Biology , 33 , 67 73. https://doi.org/10.1016/j.semcancer.2015.04.007 53. Doron, B., Abdelhamed, S., Butler, J. T., Hashmi, S. K., Horton, T. M., & Kurre, P. (2018). Transmissible ER stress reconfigures the AML bone marrow compartment. Leukemia . https://doi.org/10.1038/s41375 - 018 - 0254 - 2 54. Dupuy, F., Tabariès, S., Andrzejewski, S., P. M. (2015). PDK1 - dependent metabolic reprogramming dictates metastatic potential in breast cancer. Cell Metabolism , 22 (4), 577 589. https://doi.org/10.1016/j.cmet.2015.08.007 147 55. Dusek, R., & Attardi, L. (2011). Desmosomes: new perpetrators in tumour suppression. Nature Reviews Cancer , 11 , 317 323. 56. Eheman, C., Henley, S. J., Ballard - Barbash, R., Jacobs, E. J., Schymura, M. J., Noone, cer, 1975 - 2008, featuring cancers associated with excess weight and lack of sufficient physical activity. Cancer , 118 (9), 2338 2366. https://doi.org/10.1002/cncr.27514 57. Emmanouilidi, A., & Falasca, M. (2017). Targeting PDK1 for chemosensitization of cancer cells. Cancers , 9 (10), 1 25. https://doi.org/10.3390/cancers9100140 58. Phosphorylation of PDP1 Toggles Recruitment between ACAT1 and SIRT3 to Regulate the Pyruvate Dehydrogen ase Complex. Molecular Cell , 53 (4), 534 548. https://doi.org/10.1016/j.molcel.2013.12.026 59. Fan, X., Weng, Y., Bai, Y., Wang, Z., Wang, S., Zhu, J., & Zhang, F. (2018). Lipin - 1 determines lung cancer cell survival and chemotherapy sensitivity by regulation o f endoplasmic reticulum homeostasis and autophagy. Cancer Medicine , 7 (6), 2541 2554. https://doi.org/10.1002/cam4.1483 60. Borger, D. R. (2014). Isocitrate Dehydrogenase 1 ( IDH1) Mutation in Breast Adenocarcinoma Is Associated With Elevated Levels of Serum and Urine 2 - Hydroxyglutarate. The Oncologist , 19 (6), 602 607. https://doi.org/10.1634/theoncologist.2013 - 0417 61. Feldman, H. C., Tong, M., Wang, L., Meza - acevedo, R., Theodore Maly, D. J. (2016). Structural and Functional Analysis of the Allosteric Inhibition of - Competitive Ligands. ACS Chem Biol , 11 (8), 2195 2205. https://doi.org/10.1002/jmri.25711.PET/MRI 62. Ferlay, J., Ervik, M., Lam, F., Colo Global Cancer Observatory: Cancer Today.Lyon, France: International Agency for Research on Cancer. Retrieved from https://gco.iarc.fr/today 63. Frezza, E. E., Wachtel, M. S., & Chiriva - Internati, M. (2006). I nfluence of obesity on the risk of developing colon cancer. Gut , 55 (2), 285 291. https://doi.org/10.1136/gut.2005.073163 148 64. Jimeno, A. (2015). A phase I study of CUDC - 10 1, a multitarget inhibitor of HDACs, EGFR, and HER2, in combination with chemoradiation in patients with head and neck squamous cell carcinoma. Clinical Cancer Research , 21 (7), 1566 1573. https://doi.org/10.1158/1078 - 0432.CCR - 14 - 2820 65. Gao, R., Liu, Y., Li, Promotes breast cancer metastasis via induction of hyaluronan production in a p38 - dependent manner. Cellular Physiology and Biochemistry , 50 (6), 2108 2123. https://doi.org/10.1159/000495055 66. Gao, X., Yo, P., & Harris, T. K. (2005). Improved yields for baculovirus - mediated expression of human His 6 - PDK1 and His6 - phospho - specific isoforms for design of inhibitors that stabilize inactive conformations. Protein Expression and Purification , 43 (1), 44 56. https://doi.org/10.1016/j.pep.2005.03.017 67. Gatenby, R. A., & Gillies, R. J. (2004). Why do cancers have high aerobic glycolysis? Nature Reviews Cancer , 4 (11), 891 899. https://doi.org/10.1038/nrc1478 68. Gebäck, T., Schulz, M. M . P., Koumoutsakos, P., & Detmar, M. (2009). TScratch: A novel and simple software tool for automated analysis of monolayer wound healing assays. BioTechniques , 46 (4), 265 274. https://doi.org/10.2144/000113083 69. Ghosh, I., Hamilton, A. D., & Regan, L. (2000). Antiparallel leucine zipper - directed protein reassembly: Application to the green fluorescent protein [12]. Journal of the American Chemical Society , 122 (23), 5658 5659. https://doi.org/10.1021/ja994421w 70. Goid Radlwimmer, B. (2012). RNAi screening in glioma stem - like cells identifies PFKFB4 as a key molecule important for cancer cell survival. Oncogene , 31 (27), 3235 3243. https://doi.org/10 .1038/onc.2011.490 71. Gómez de Cedrón, M., & Ramírez de Molina, A. (2015). Microtargeting cancer metabolism: opening new therapeutic windows based on lipid metabolism. Journal of Lipid Research , 57 , jlr.R061812. https://doi.org/10.1194/jlr.R061812 72. Greenfield, N. J. (2006). Using circular dichroism spectra to estimate protein secondary structure. Nat. Protocols , 1 (6), 2876 2890. https://doi.org/10.1038/nprot.2006.202.Using 149 73. Gupta, J. C., Jaisani, M., Pandey, G., & Mukherjee, K. J. (1999). Enhancing recombinant p rotein yields in Escherichia coli using the T7 system under the control of Journal of Biotechnology , 68 (2 3), 125 134. https://doi.org/10.1016/S0168 - 1656(98)00193 - X 74. Haeger, S., Kuzmin, D., Detro - Dassen, S., Lang, N., Kilb, M. Schmalzing, G. (2010). An intramembrane aromatic network determines pentameric assembly of Cys - loop receptors. Nature Structural and Molecular Biology , 17 (1), 90 99. https://doi.org/10.1038/nsmb.1721 75. Halbleib, K., Pesek, K., Covino, R., Ho R. (2017). Activation of the Unfolded Protein Response by Lipid Bilayer Stress. Molecular Cell , 67 (4), 673 - 684.e8. https://doi.org/10.1016/j.molcel.2017.06.012 76. Han, J. E., Lim, P. W., Na, C. M., Choi, Y. S. Experimental Neurobiology , 26 (5), 295. https://doi.org/10.5607/en.2017.26.5.295 77. Han, J., & Kaufman, R. J. (2016). The role of ER stress in lipid metabolism and lipotoxicity. Journal of Lipid Research , 57 (8), 1329 1338. https://doi.org/10.1194/jlr.R067595 78. Hanahan, D., & Weinberg, R. A. (2011). Hallmarks of cancer: The next generation. Cell , 144 (5), 646 674. https://doi.org/10.1016/j.cell.20 11.02.013 79. (2017a). Lipin - 1 regulation of phospholipid synthesis maintains endoplasmic reticulum homeostasis and is critical for triple - negative breast cancer cell survival. FASEB Journ al , 31 (7), 2893 2904. https://doi.org/10.1096/fj.201601353R 80. (2017b). Lipin - 1 regulation of phospholipid synthesis maintains endoplasmic reticulum homeostasis and is critical for triple - negative breast cancer cell survival. FASEB Journal , 31 (7), 2893 2904. https://doi.org/10.1096/fj.201601353R 81. Hetz, C., Chevet, E., & Harding, H. P. (2013). Targeting the unfolded protein response in disease. Nature Publishing Group , 12 (9), 703 7 19. https://doi.org/10.1038/nrd3976 82. Hetz, C., & Glimcher, L. H. (2009). Fine - tuning of the unfolded protein response: Assembling the IRE1alpha interactome. Molecular Cell , 35 (5), 551 561. 150 https://doi.org/10.1016/j.molcel.2009.08.021 83. Hetz, C., & Papa, F. R. (2018). The Unfolded Protein Response and Cell Fate Control. Molecular Cell , 69 (2), 169 181. https://doi.org/10.1016/j.molcel.2017.06.017 84. Holthuis, J. C. M., & Menon, A. K. (2014). Lipid landscapes and pipelines in membrane homeosta sis. Nature , 510 (7503), 48 57. https://doi.org/10.1038/nature13474 85. Hu, C. D., Chinenov, Y., & Kerppola, T. K. (2002). Visualization of interactions among bZIP and Rel family proteins in living cells using bimolecular fluorescence complementation. Molecular Cell , 9 (4), 789 798. https://doi.org/10.1016/S1097 - 2765(02)00496 - 3 86. Huang, Y., Li, Y., Wang, Y. G., Gu, X., Wang, Y., & Shen, B. F. (2007). An efficient and targeted gene integration system for high - level antibody expression. Journal of Immunological Metho ds , 322 (1 2), 28 39. https://doi.org/10.1016/j.jim.2007.01.022 87. Iizuka, K. (2017). The transcription factor carbohydrate - response element - binding protein (ChREBP): A possible link between metabolic disease and cancer. Biochimica et Biophysica Acta - Molecul ar Basis of Disease , 1863 (2), 474 485. https://doi.org/10.1016/j.bbadis.2016.11.029 88. Imamura, T., Kikuchi, H., Herraiz, M. T., Park, D. Y., Mizukami, Y., Mino - Kenduson, - 1alpha and HIF - 2alpha have divergent roles in colon cance r. International Journal of Cancer , 124 (4), 763 771. https://doi.org/10.1002/ijc.24032.HIF - 1 89. Iwakoshi, N. N., Lee, A. - H., & Glimcher, L. H. (2003). The X - box binding protein - 1 transcription factor is required for plasma cell differentiation and the unfolde d protein response. Immunological Reviews , 194 , 29 38. Retrieved from http://ovidsp.ovid.com/ovidweb.cgi?T=JS&PAGE=reference&D=emed6&NEWS=N &AN=2003297477 90. James, O. F. W., & Day, C. P. (1998). Non - alcoholic steatohepatitis (NASH): A disease of emerging iden tity and importance. Journal of Hepatology , 29 (3), 495 501. https://doi.org/10.1016/S0168 - 8278(98)80073 - 1 91. Ji, C., & Kaplowitz, N. (2006). ER stress: Can the liver cope? Journal of Hepatology , 45 (2), 321 333. https://doi.org/10.1016/j.jhep.2006.06.004 151 92. Jiang - XBP1 branch of the unfolded protein response in human diseases. Seminars in Cancer Biology , 33 , 48 56. https://doi.org/10.1016/j.semcancer.2015.04.010 93. Kakazu, E., Mauer, A. S., Yin, M., & Malhi, H. (2016). Hepatocytes release ceramide - enriched pro - - dependent manner. Journal of Lipid Research , 57 (2), 233 245. https://doi.org/10.1194/jlr.M063412 94. Kanemoto, S., Nitani, R., Murakami, T., Kaneko, M., Asada, R Imaizumi, K. (2016). Multivesicular body formation enhancement and exosome release during endoplasmic reticulum stress. Biochemical and Biophysical Research Communications , 480 (2), 166 172. https://doi.org/10.1016/j.bbrc.2016.10.019 95. Kar askov, E., Scott, C., Zhang, L., Teodoro, T., Ravazzola, M., & Volchuk, A. (2006a). Chronic palmitate but not oleate exposure induces endoplasmic reticulum stress, which may contribute to INS - 1 pancreatic beta - cell apoptosis. Endocrinology , 147 (7), 3398 34 07. https://doi.org/10.1210/en.2005 - 1494 96. Karaskov, E., Scott, C., Zhang, L., Teodoro, T., Ravazzola, M., & Volchuk, A. (2006b). Chronic palmitate but not oleate exposure induces endoplasmic reticulum stress, which may contribute to INS - - cell apoptosis. Endocrinology , 147 (7), 3398 3407. https://doi.org/10.1210/en.2005 - 1494 97. coordination of gene transcriptional and translational controls. Genes & Development , 13 , 1211 1233. 98. Khan, A., Fornes, O., Stigliani, A., Gheorghe, M., Castro - Mondragon, J. A., Van Der - access database of transcription factor binding profiles and its web framework. Nucleic Acids Research , 46 , D260 D266. https://doi.org/10.1093/nar/gkx1126 99. Kim, H., Bhattacharya, A., & Qi, L. (2015). Endoplasmic reticulum quality control in cancer: Friend or foe. Seminars in Cancer Biology , 33 , 25 33. https://doi.org/10.1016/j.semcancer.2015.02.003 100. Kim, Jin Young, Kim, G., Lim, S. C., & Choi, H. S. (2016). LPIN1 promotes epithelial cell transformation and mammary tumourigenesis via enhancing insulin receptor substrate 1 stability. Carcinogenesis , 37 (12), 1199 1209. https://doi.org/10.1093/carcin/bgw1 04 152 101. Kim, Ju Youn, Garcia - Carbonell, R., Yamachika, S., Zhao, P., Dhar, D., Loomba, - 2 Activation of S1P. Cell , 175 (1), 133 - 145.e15. https://doi.org/10.1016/j.cell.2018.08.02 0 102. Kinlaw, W. B., Baures, P. W., Lupien, L., Davis, W. L., & Kuemmerle, N. B. (2016). Fatty acids and breast cancer: Make them on site or have them delivered. Journal of Cellular Physiology , 231 (10), 2128 2141. https://doi.org/10.1038/jid.2013.21.Propioniba cterium 103. Kitai, Y., Ariyama, H., Kono, N., Oikawa, D., Iwawaki, T., & Arai, H. (2013). Membrane lipid saturation activates IRE1?? without inducing clustering. Genes to Cells , 18 (9), 798 809. https://doi.org/10.1111/gtc.12074 104. Kodama, Y., & Hu, C. D. (2010). An improved bimolecular fluorescence complementation assay with a high signal - to - noise ratio. BioTechniques , 49 (5), 793 803. https://doi.org/10.2144/000113519 105. Kono, N., Amin - Wetzel, N., & Ron, D. (2017). Generic membrane - spanning Molecular Biology of the Cell , 28 (17), 2318 2332. https://doi.org/10.1091/mbc.E17 - 03 - 0144 106. Korennykh, A., & Walter, P. (20 12). Structural basis of the unfolded protein response. Annual Review of Cell and Developmental Biology , 28 , 251 277. https://doi.org/10.1146/annurev - cellbio - 101011 - 155826 107. Kosel, D., Heiker, J. T., Juhl, C., Wottawah, C. M., Bluher, M., Morl, K., & Beck - Si ckinger, A. G. (2010). Dimerization of adiponectin receptor 1 is inhibited by adiponectin. Journal of Cell Science , 123 (8), 1320 1328. https://doi.org/10.1242/jcs.057919 108. Lai, G. Y., Park, Y., Hartge, P., Hollenbeck, A. R., & Freedman, N. D. (2013). The ass ociation between self - reported diabetes and cancer incidence in the NIH - AARP diet and health study. Journal of Clinical Endocrinology and Metabolism , 98 (3), 497 502. https://doi.org/10.1210/jc.2012 - 3335 109. Lamouille, S., Xu, J., & Derynck, R. (2014). Molecula r mechanisms of epithelial - mesenchymal transition. National Review Molecular Cell Biology , 15 (3), 178 196. https://doi.org/10.1038/nrm3758.Molecular 153 110. Lea, W., & Simeonov, A. (2012). Fluorescence Polarisation Assay in Small Molecule Screening. Expert Opin Dr ug Discov. , 6 (1), 17 32. https://doi.org/10.1517/17460441.2011.537322.Fluorescence 111. Lebeaupin, C., Vallée, D., Hazari, Y., Hetz, C., Chevet, E., & Bailly - Maitre, B. (2018). Endoplasmic reticulum stress signalling and the pathogenesis of non - alcoholic fatty liver disease. Journal of Hepatology , 69 (4), 927 947. https://doi.org/10.1016/j.jhep.2018.06.008 112. Lee, A. - H., & Glimcher, L. H. (2009). Intersection of the unfolded protein response and hepatic lipid metabolism. , 66 (17), 2835 2850. https://doi.org/10.1007/s00018 - 009 - 0049 - 8 113. Lee, A. - H., Iwakoshi, N. N., & Glimcher, L. H. (2003). XBP - 1 Regulates a Subset of Endoplasmic Reticulum Resident Chaperone Genes in the Unfolded Protein Response. Molecular and Cellular Biology , 23 (21), 7448 7459. https://doi.org/10.1128/MCB.23.21.7448 114. Lee, A. - H., Scapa, E. F., Cohen, D. E., & Glimcher, L. H. (2008). Regulation of hepatic lipogenesis by the transcription factor XBP1. Science , 320 (5882), 1492 1496. https://doi.org/10.1038/jid.2014 .371 115. Lee, J., Cho, H. - K., & Kwon, Y. H. (2010). Palmitate induces insulin resistance without significant intracellular triglyceride accumulation in HepG2 cells. Metabolism: Clinical and Experimental , 59 (7), 927 934. https://doi.org/10.1016/j.metabol.2009.1 0.012 116. Lee, K. P. K., Dey, M., Neculai, D., Cao, C., Dever, T. E., & Sicheri, F. (2008). Structure of the dual enzyme Ire1 reveals the basis for catalysis and regulation in nonconventional RNA splicing. Cell , 132 (1), 89 100. https://doi.org/10.1016/j.cell.2 007.10.057 117. Kaufman, R. J. (2002). IRE1 - mediated unconventional mRNA splicing and S2P - mediated ATF6 cleavage merge to regulate XBP1 in signaling the unfolded protein response. Genes & Development , 16 (4), 452 466. https://doi.org/10.1101/gad.964702 118. Lemercier, C., To, R. Q., Carrasco, R. A., & Konieczny, S. F. (1998). The basic helix - loop - helix transcription factor Mist1 functions as a transcriptional repressor of MyoD. EMBO Journal , 17 ( 5), 1412 1422. https://doi.org/10.1093/emboj/17.5.1412 154 119. Liu, T., & Yin, H. (2017). PDK1 promotes tumor cell proliferation and migration by enhancing the Warburg effect in non - small cell lung cancer. Oncology Reports , 37 (1), 193 200. https://doi.org/10.3892/ or.2016.5253 120. Liu, Y. Z., Wang, Y. X., & Jiang, C. L. (2017). Inflammation: The common pathway of stress - related diseases. Frontiers in Human Neuroscience , 11 (June), 1 11. https://doi.org/10.3389/fnhum.2017.00316 121. Louie, S. M., Roberts, L. S., Mulvihill, M. M., Luo, K., & Nomura, D. K. (2013). Cancer cells incorporate and remodel exogenous palmitate into structural and oncogenic signaling lipids. Biochimica et Biophysica Acta - Molecular and Cell Biology of Lipids , 1831 (10), 1566 1572. https://doi.org/10.1016 /j.bbalip.2013.07.008 122. breast cancer. Clinical Chemistry , 61 (12), 1457 1465. https://doi.org/10.1373/clinchem.2015.240028 123. Lu, J., Wang, Q., Huang, L., Dong, H., Lin, L., Lin, N causes endoplasmic reticulum stress and apoptosis in human mesenchymal stem cells: prevention by AMPK activator. Endocrinology , 153 (11), 5275 5284. https://doi.org/10.1210/en.2012 - 1418 124. Ludwig, J., Viggiano, T. R., McGill, D. B., & Oh, B. J. (1980). Nonalcoholic steatohepatitis: Mayo Clinic experiences with a hitherto unnamed disease. Mayo Clinic Proceedings , 55 (7), 434 438. Retrieved from http://europepmc.org/abstract/MED/7382552 125. Luengo, A., Gui, D. Y., & Vander Heiden, M. G. (2017). Targeting Metabolism for Cancer Therapy. Cell Chemical Biology , 24 (9), 1161 1180. https://doi.org/10.1016/j.chembiol.2017.08.028 126. Luo, X., Cheng, C., Tan, Z., Li, N., Tang, M., Yang, L., & Cao, Y. (2017). Emerging roles of lipid metabolism in cancer metastasis. Molecular Cancer , 16 (1), 1 10. https://doi.org/10.1186/s12943 - 017 - 0646 - 3 127. Ma, X., Wang, Y. W., Zhang, M. Q., & Gazdar, A. F. (2013). DNA methylation data analysis and its application to cancer research. Epigenomics , 5 (3), 301 316. https://doi.org/10.2217/epi.13.26 128. Ma, Y., Yang, Y., Wang, F., Zhang, P., Shi, C., Zou, Y., & Qin, H. (2013). Obesity 155 and Risk of Colorectal Cancer: A Systematic Review of Prospective Studies. PLoS ONE , 8 (1), e53916. https://doi.org/10.1371/journal.pone.005 3916 129. - catenin) has TCF/LEF family - dependent - catenin - deficient cell line. Oncogene , 23 (4), 964 972. https://doi.org/10. 1038/sj.onc.1207254 130. Mahadevan, N. R., Rodvold, J., Sepulveda, H., Rossi, S., Drew, A. F., & Zanetti, M. (2011). Transmission of endoplasmic reticulum stress and pro - inflammation from tumor cells to myeloid cells. Proceedings of the National Academy of Scie nces , 108 (16), 6561 6566. https://doi.org/10.1073/pnas.1008942108 131. Malhi, H., & Kaufman, R. J. (2011). Endoplasmic reticulum stress in liver disease. Journal of Hepatology , 54 (4), 795 809. https://doi.org/10.1016/j.jhep.2010.11.005 132. Maloney, E., Sweet, I. R., Hockenbery, D. M., Pham, M., Rizzo, N. O., Tateya, S., - for NADPH oxidase - derived superoxide in respons e to TLR4 activation. Arteriosclerosis, Thrombosis, and Vascular Biology , 29 (9), 1370 1375. https://doi.org/10.1161/ATVBAHA.109.188813 133. Maurel, M., Chevet, E., Tavernier, J., & Gerlo, S. (2014). Getting RIDD of RNA: IRE1 in cell fate regulation. Trends in B iochemical Sciences , 39 (5), 245 254. https://doi.org/10.1016/j.tibs.2014.02.008 134. May - R. S. (2017). Pro - inflammatory fatty acid profile and colorectal cancer risk: A Mendelian r andomisation analysis. European Journal of Cancer , 84 , 228 238. https://doi.org/10.1016/j.ejca.2017.07.034 135. Mazon, J. N., de Mello, A. H., Ferreira, G. K., & Rezin, G. T. (2017). The impact of obesity on neurodegenerative diseases. Life Sciences , 182 , 22 28 . https://doi.org/10.1016/j.lfs.2017.06.002 136. McGrath, J. A. (2005). Inherited disorders of desmosomes. Australasian Journal of Dermatology , 46 (4), 221 229. https://doi.org/10.1111/j.1440 - 0960.2005.00188.x 137. Meana, C., García - rostán, G., Peña, L., Lordén, G., Balboa, M. (2018). The phosphatidic acid phosphatase lipin - 1 facilitates inflammation - 156 lipin - 1 facilitates inflammation - driven colon carcinogen esis. JCI Insight , 3 (18), e97506. 138. Merlin, M., Gecchele, E., Capaldi, S., Pezzotti, M., & Avesani, L. (2014). Comparative evaluation of recombinant protein production in different biofactories: The green perspective. BioMed Research International , 2014 . htt ps://doi.org/10.1155/2014/136419 139. Minchenko, O. H., Opentanova, I. L., Ogura, T., Dmytro, O., Komisarenko, S. V, Caro, J., & Esumi, H. (2005). Expression and hypoxia - responsiveness of 6 - phosphofructo - 2 - kinase/fructose - 2,6 - bisphosphatase 4 in mammary gland m alignant cell lines. Acta Biochimica Polonica , 52 (4), 881 888. 140. Minchenko, O., Opentanova, I., & Caro, J. (2003). Hypoxic regulation of the 6 - phosphofructo - 2 - kinase/fructose - 2,6 - bisphosphatase gene family (PFKFB - 1 - 4) expression in vivo. FEBS Letters , 554 (3 ), 264 270. https://doi.org/10.1016/S0014 - 5793(03)01179 - 7 141. Miranda, M., Escoté, X., Ceperuelo - Vendrell, J. (2010). Relation between human LPIN1, hypoxia and endoplasmic reticulum stress genes in subcutaneous a nd visceral adipose tissue. International Journal of Obesity , 34 (4), 679 686. https://doi.org/10.1038/ijo.2009.290 142. Mitra, S., & Ryoo, H. D. (2019). The unfolded protein response in metazoan development. Journal of Cell Science , 132 (5), jcs217216. https://doi.org/10.1242/jcs.217216 143. Monteiro, R., & Azevedo, I. (2010). Chronic inflammation in obesity and the metabolic syn drome. Mediators of Inflammation , 2010 (Atp Iii). https://doi.org/10.1155/2010/289645 144. Müller, S., Sandal, T., Kamp - Hansen, P., & Dalbøge, H. (1998). Comparison of expression systems in the yeasts Saccharomyces cerevisiae, Hansenula polymorpha, Klyveromyces lactis, Schizosaccharomyces pombe and Yarrowia lipolytica. Cloning of two novel promoters from Yarrowia lipolytica. Yeast , 14 (14), 1267 1283. https://doi.org/10.1002/(SICI)1097 - 0061(1998100)14:14<1267::AID - YEA327>3.0.CO;2 - 2 145. Nagelkerke, A., Bussink, J., Swe ep, F. C. G. J., & Span, P. N. (2014). The unfolded protein response as a target for cancer therapy. Biochimica et Biophysica Acta - 157 Reviews on Cancer , 1846 (2), 277 284. https://doi.org/10.1016/j.bbcan.2014.07.006 146. Nakagawa, H., Hikiba, Y., Hirata, Y., Font - Burgada, J., Sakamoto, K., Hayakawa, - cadherin induces sclerosing cholangitis and promotes carcinogenesis. Proceedings of the National Academy of Sciences of the United States of America , 111 (3), 1090 1095. https://d oi.org/10.1073/pnas.1322731111 147. Nakagawa, H., Umemura, A., Taniguchi, K., Font - Burgada, J., Dhar, D., Ogata, H., - Dependent Spontaneous HCC Development. Cancer Cell , 26 (3), 331 343. https://doi.org/10.1016/j.ccr.2014.07.001 148. H. (2006). Activation of the endoplasmic reticulum stress pathway is associated with survival of myeloma cells. Leukemia and Lymphoma , 47 (3), 531 539. https://doi.org/10.1080/10428190500312196 149. Nath, A. (2015). Investigating the Role of Elevated Free Fatty Acids in Epithelial - Mesenchymal Transition of Hepatocellular Carcinoma . https://doi.org/10.25335/M5PX3Q 150. Nath, A., & Chan, C. (2016). G enetic alterations in fatty acid transport and metabolism genes are associated with metastatic progression and poor prognosis of human cancers. Scientific Reports , 6 (August 2015), 1 13. https://doi.org/10.1038/srep18669 151. Nath, A., Li, I., Roberts, L. R., & Chan, C. (2015). Elevated free fatty acid uptake via CD36 promotes epithelial - mesenchymal transition in hepatocellular carcinoma. Scientific Reports , 5 (April), 14752. https://doi.org/10.1038/srep14752 152. Nekrasova, O., & Green, K. J. (2013). Desmosome assembl y and dynamics. Trends in Cell Biology , 23 (11), 537 546. https://doi.org/10.1016/j.tcb.2013.06.004 153. Neuhouser, M. L., Aragaki, A. K., Prentice, R. L., Manson, J. A. E., Chlebowski, opausal randomized clinical trials. JAMA Oncology , 1 (5), 611 621. https://doi.org/10.1001/jamaoncol.2015.1546 158 154. Nielsen, M., Lundegaard, C., Lund, O., & Petersen, T. N. (2010) . CPHmodels - 3.0 - remote homology modeling using structure - guided sequence profiles. Nucleic Acids Research , 38 (SUPPL. 2), 576 581. https://doi.org/10.1093/nar/gkq535 155. Oikawa, D., Kimata, Y., Kohno, K., & Iwawaki, T. (2009). Activation of ER stress depends on dissociation of BiP rather than on direct interaction with unfolded proteins. Experimental Cell Research , 315 (15), 2496 2504. https://doi.org/10.1016/j.yexcr.2009.06.009 156. Opie, L. H., & Walfish, P. G. (1963). Plasma Free Fatty Acid Con centrations in Obesity. New England Journal of Medicine , 268 (14), 757 760. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7993409 157. Ozcan, U., Cao, Q., Yilmaz, E., Lee, A. - Hotamisigil, G. (2004). Endoplasmic Reticulum Stress Links Obesity, Insulin Action, and Type 2 Diabetes. Science , 306 (Oct), 457 461. https://doi.org/10.1002/9780470114735.hawley00624 158. Pavlova, N. N., & Thompson, C. B. (2016). The Emerging Hallmarks of Cancer Metabolism. Cell Metabolism , 23 (1), 27 47. https://doi.org/10.1016/j.cmet.2015.12.006 159. (2018). Glycoly sis gatekeeper PDK1 reprograms breast cancer stem cells under hypoxia. Oncogene , 37 (8), 1062 1074. https://doi.org/10.1038/onc.2017.368 160. Pfaffenbach, K. T., Gentile, C. L., Nivala, A. M., Wang, D., Wei, Y., & Pagliassotti, M. J. (2010). Linking endoplasmic reticulum stress to cell death in hepatocytes: roles of C/EBP homologous protein and chemical chaperones in palmitate - mediated cell death. American Journal of Physiology. Endocrinology and Metabolism , 298 (5), E1027 - 35. https://doi.org/10.1152/ajpendo.00642 .2009 161. Picon - Ruiz, M., Morata - Tarifa, C., Valle - Goffin, J. J., Friedman, E. R., & Slingerland, J. M. (2017). Obesity and adverse breast cancer risk and outcome: Mechanistic insights and strategies for intervention. CA: A Cancer Journal for Clinicians , 67 (5) , 378 397. https://doi.org/10.3322/caac.21405 162. Pineau, L., Colas, J., Dupont, S., Beney, L., Fleurat - Lessard, P., Berjeaud, J. - Ferreira, T. (2009). Lipid - induced ER stress: synergistic effects of sterols and saturated fatty acids. Traffic (Copenhagen, Denmark) , 10 (6), 673 690. https://doi.org/10.1111/j.1600 - 0854.2009.00903.x 159 163. Prischi, F., Nowak, P. R., Carrara, M., & Ali, M. M. U. (2014). Phosphoregulation of Ire1 RNase splicing activity. Nature Communications , 5 , 1 10. https://doi.org/10.1038/ncomms455 4 164. Qi, L., Yang, L., & Chen, H. (2011). Detecting and quantitating physiological endoplasmic reticulum stress. Methods in enzymology (1st ed., Vol. 490). Elsevier Inc. https://doi.org/10.1016/B978 - 0 - 12 - 385114 - 7.00008 - 8 165. Qian, X. L., Pan, Y. H., Huang, Q. Y., Shi, Y. B., Huang, Q. Y., Hu, Z. Z., & Xiong, L. X. (2019). Caveolin - 1: A multifaceted driver of breast cancer progression and its application in clinical treatment. OncoTargets and Therapy , 12 , 1539 1552. https://doi.org/10.2147/OTT.S191317 166. Qin, W., Tian double inhibition of PDK1 and STAT3 - Y705 prevents liver metastasis in colorectal cancer. Scientific Reports , 9 (12973), 1 12. https://doi.org/10.1038/s41598 - 019 - 49480 - 8 167. Qu, Y., Gharbi, N., Yu - catenin signaling and directs asymmetric cell division in cancer. Proceedings of the National Academy of Sciences , 113 (33), 9339 LP 9344. https://doi.org/10.1073/pnas.1 604520113 168. Raimondi, C., & Falasca, M. (2011). Targeting PDK1 in cancer. Current Medicinal Chemistry , 18 (18), 2763 2769. 169. Ralston, J. C., Lyons, C. L., Kennedy, E. B., Kirwan, A. M., & Roche, H. M. (2017). Fatty Acids and NLRP3 Inflammasome Mediated Inflam mation in Metabolic Tissues. Annual Review of Nutrition , 37 , 77 102. https://doi.org/10.1146/annurev - nutr - 071816 - 064836 170. Ran, F. A., Hsu, P. D., Wright, J., Agarwala, V., Scott, D. A., & Zhang, F. (2013). Genome engineering using the CRISPR - Cas9 system. Nat . Protocols , 8 (11), 2281 2308. Retrieved from http://dx.doi.org/10.1038/nprot.2013.143 171. Renehan, A. G., Tyson, M., Egger, M., Heller, R., & Zwahlen, M. (2008). Body - - analysis of prospective ob servational studies. The Lancet , 371 (9612), 569 578. 172. Reue, K. (2009). The Lipin Family: Mutations and Metabolism. Current Opinion in 160 Lipidology , 20 (3), 165 170. https://doi.org/10.1097/MOL.0b013e32832adee5.The 173. Ridder, A., Skupjen, P., Unterreitmeier, S., & Langosch, D. (2005). Tryptophan supports interaction of transmembrane helices. Journal of Molecular Biology , 354 (4), 894 902. https://doi.org/10.1016/j.jmb.2005 .09.084 174. Rinas, U., & Bailey, J. E. (1992). Protein compositional analysis of inclusion bodies produced in recombinant Escherichia coli. Applied Microbiology and Biotechnology , 37 (5), 609 614. https://doi.org/10.1007/BF00240735 175. Roberts, D. L., Dive, C., & R enehan, A. G. (2010). Biological Mechanisms Linking Obesity and Cancer Risk: New Perspectives. Annual Review of Medicine , 61 (1), 301 316. https://doi.org/10.1146/annurev.med.080708.082713 176. Rodriguez, L. G., Wu, X., & Guan, J. - L. (2005). Wound - healing assay. In Methods (Vol. 294, pp. 23 29). Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15576902 177. Rodvold, J. J., Chiu, K. T., Hiramatsu, N., Nussbacher, J. K., Mahadevan, N. R., (2018). Intercellular transmission of the unfolded protein response promotes survival and drug resistance in cancer cells. Science Signaling , 10 (482). https://doi.org/10.1126/scisignal.aah7177.Intercellular 178. Romero - Ramirez, L., Cao, H., Nelson, D., Hammond, E., Lee, A. - H., Yoshida, H., is required for tumor growth. Cancer Research , 64 (17), 5943 5947. https://doi.org/10.1158/0008 - 5472.CAN - 04 - 1606 179. Ros, S., Santos, C. R., Moco, S (2012). Functional metabolic screen identifies 6 - phosphofructo - 2 - kinase/fructose - 2, 6 - biphosphatase 4 as an important regulator of prostate cancer cell survival. Cancer Discovery , 2 (4), 328 343. https://d oi.org/10.1158/2159 - 8290.CD - 11 - 0234 180. Rosano, G. L., & Ceccarelli, E. A. (2014). Recombinant protein expression in Escherichia coli: Advances and challenges. Frontiers in Microbiology , 5 (APR), 1 17. https://doi.org/10.3389/fmicb.2014.00172 181. Rstudio Team. (201 5). RStudio: Integrated Development for R. 161 182. Rubio - Patiño, C., Bossowski, J. P., Chevet, E., & Ricci, J. E. (2018). Reshaping the Immune Tumor Microenvironment Through IRE1 Signaling. Trends in Molecular Medicine , 24 (7), 607 614. https://doi.org/10.1016/j.mo lmed.2018.05.005 183. Rudick, M., & Anderson, R. G. W. (2002). Multiple functions of caveolin - 1. Journal of Biological Chemistry , 277 (44), 41295 41298. https://doi.org/10.1074/jbc.R200020200 184. Rutkowski, D. T. (2018). Liver function and dysfunction - a unique window into the physiological reach of ER stress and the unfolded protein response. FEBS Journal , 286 , 356 378. https://doi.org/10.1111/febs.14389 185. Saadatian - Elahi, M., Toniolo, P., Ferrari, P., Goudable, J., Akhmedkhanov, A., Zeleniuch - Jacquotte, A., & Riboli, E. (2002). Serum fatty acids and risk of breast cancer in a nested case - Study. Cancer Epidemiology Biomarkers and Prevention , 11 (11), 1353 1360. 186. Saitoh, M. (2018). Involvement of partial EMT in cancer progression. Journal of Biochemistry , 0 (June), 1 8. https://doi.org/10.1093/jb/mvy047 187. Sal - Man, N., Gerber, D., Bloch, I., & Shai, Y. (2007). Specificity in transmembrane helix - helix interactions mediated by aromatic residues. Journal of Biological Chemistry , 282 (27), 19753 19761. https://doi.org/10.1074/jbc.M610368200 188. Sánchez - Tilló, E., Liu Postigo, A. (2012). EMT - activating transcription factors in cancer: Beyond EMT and tumor invasiveness. Cellular and Molecular Life Sciences , 69 (20), 3429 3456. https://doi.org/10.1007/s00018 - 01 2 - 1122 - 2 189. Sandelin, A., Alkema, W., Engstrom, P., Wasserman, W., & Lenhard, B. (2004). JASPAR: an open - access database for eukaryotic transcription factor binding profiles. Nucleic Acids Research , 32 , D91 D94. https://doi.org/10.1093/nar/gkh012 190. Schröder, M. , & Kaufman, R. J. (2005a). ER stress and the unfolded protein response. Mutation Research , 569 (1 2), 29 63. https://doi.org/10.1016/j.mrfmmm.2004.06.056 191. Schröder, M., & Kaufman, R. J. (2005b). ER stress and the unfolded protein response. Mutation Research /Fundamental and Molecular Mechanisms of Mutagenesis , 569 (1 2), 29 63. https://doi.org/10.1016/j.mrfmmm.2004.06.056 162 192. Schröder, M., & Kaufman, R. J. (2005c). The Mammalian Unfolded Protein Response. Annual Review of Biochemistry , 74 , 739 789. https://doi.org /10.1146/annurev.biochem.73.011303.074134 193. Semb, H., & Christofori, G. (1998). The Tumor - Suppressor Function of E - Cadherin. American Journal of Human Genetics , 63 , 1588 1593. 194. Semenza, G. L. (2013). HIF - 1 mediates metabolic responses to intratumoral hypoxia and oncogenic mutations. Journal of Clinical Investigation , 123 (9), 3664 3671. https://doi.org/10.1172/JCI67230.3664 195. Seo, T., Oelkers, P. M., Giattina, M. R., Worgall, T. S., Sturley, S. L., & Deckelbaum, R. J. (2001). Differential modulation of ACAT1 and ACAT2 transcription and activity by long chain free fatty acids in cultured cells. Biochemistry , 40 (15), 4756 4762. https://doi.org/10.1021/bi0022947 196. Sha, H., He, Y., Yang, L., & Qi, L. (2011). Stressed out about obesity: IRE1?? - XBP1 in metabolic disorders . Trends in Endocrinology and Metabolism , 22 (9), 374 381. https://doi.org/10.1016/j.tem.2011.05.002 197. Sharma, S. V, Lee, D. Y., Li, B., Quinlan, M. P., Maheswaran, S., Mcdermott, U., - mediated reversible drug tolerant state in cancer cell subpopulations. Cell , 141 (1), 69 80. https://doi.org/10.1016/j.cell.2010.02.027.A 198. Shechter, I., Dai, P., Huo, L., & Guan, G. (2003). IDH1 gene transcription is sterol regulated and activated by SREBP - 1a and SREBP - 2 in human hepatoma HepG2 c ells. Journal of Lipid Research , 44 (11), 2169 2180. https://doi.org/10.1194/jlr.m300285 - jlr200 199. Shi, X., & Jarvis, D. (2007). Protein N - Glycosylation in the Baculovirus - Insect Cell System. Current Drug Targets , 8 (10), 1116 1125. https://doi.org/10.2174/1389 45007782151360 200. Shimizu, M., Fukunaga, Y., Ikenouchi, J., & Nagafuchi, A. (2007). Defining the Roles of - Catenin and Plakoglobin in LEF/T - Cell Factor - Dependent Transcription Using - Catenin/Plakoglobin - Null F9 Cells. Molecular and Cellular Biology , 28 (2), 82 5 835. https://doi.org/10.1128/mcb.02375 - 06 201. Shoulders, M. D., Ryno, L. M., Genereux, J. C., Moresco, J. J., Tu, P. G., Wu, C., - Independent Activation of XBP1s and/or ATF6 163 Reveals Three Functionally Diverse ER Proteostasis En vironments. Cell Reports , 3 (4), 1279 1292. https://doi.org/10.1016/j.celrep.2013.03.024 202. Yamamoto, M. (2003). Activation of the ATF6, XBP1 and grp78 genes in human hepatocellular carcinoma: a possible involvement of the ER stress pathway in hepatocarcinogenesis. Journal of Hepatology , 38 , 605 614. https://doi.org/10.1016/S0 203. Shyu, J., Liu, H., Deng, X., & Hu, C. - D. (2006). Identification of new fluorescent protein fragments for bimolecular fluorescence complementation analysis under physiological conditions. BioTechniques , 40 (1), 61 66. https://doi.org/10.2144/000112036 204. & Silvestris, F. (2016). Obesity and Breast Cancer: Molecular Interconnections and Potential Clinical Applications. The Oncolo gist , 21 (4), 404 417. https://doi.org/10.1634/theoncologist.2015 - 0351 205. Singhai, R., Patil, V. W., Jaiswal, S. R., Patil, S. D., Tayade, M. B., & Patil, A. V. (2011). E - Cadherin as a diagnostic biomarker in breast cancer. North American Journal of Medical Sc iences , 3 (5), 227 233. https://doi.org/10.4297/najms.2011.3227 206. So, J. S., Hur, K. Y., Tarrio, M., Ruda, V., Frank - Kamenetsky, M., Fitzgerald, K., - mediated mRNA decay lowers plasma lipid s in mice. Cell Metabolism , 16 (4), 487 499. https://doi.org/10.1016/j.cmet.2012.09.004 207. Sørensen, H. P., & Mortensen, K. K. (2005). Advanced genetic strategies for recombinant protein expression in Escherichia coli. Journal of Biotechnology , 115 (2), 113 128 . https://doi.org/10.1016/j.jbiotec.2004.08.004 208. Spruance, S. L., Reid, J. E., Grace, M., & Samore, M. (2004). Hazard Ratio in Clinical Trials. Antimicrobial Agents and Chemotherapy , 48 (8), 2787 2792. https://doi.org/10.1128/AAC.48.8.2787 2792.2004 209. Subedi, G. P., Johnson, R. W., Moniz, H. A., Moremen, K. W., & Barb, A. (2015). High Yield Expression of Recombinant Human Proteins with the Transient Transfection of HEK293 Cells in Suspension. Journal of Visualized Experiments , 106 , e53568. https://doi.o rg/10.3791/53568 164 210. Sugie, S., Mukai, S., Yamasaki, K., Kamibeppu, T., Tsukino, H., & Kamoto, T. (2015). Significant association of caveolin - 1 and caveolin - 2 with prostate cancer progression. Cancer Genomics and Proteomics , 12 (6), 391 396. 211. Tao, R. - H., & Maruy ama, I. N. (2008). All EGF(ErbB) receptors have preformed homo - and heterodimeric structures in living cells. Journal of Cell Science , 121 (19), 3207 3217. https://doi.org/10.1242/jcs.033399 212. Tavares De Almeida, I., Cortez - Pinto, H., Fidalgo, G., Rodrigues, D., & Camilo, M. E. (2002). Plasma total and free fatty acids composition in human non - alcoholic steatohepatitis. Clinical Nutrition , 21 (3), 219 223. https://doi.org/10.1054/clnu.2001.0529 213. Telang, S., Trent, J., Chesney, J., & Clark, J. (2015). Targeting 6 - Phosphofructo - 2 - kinase/Fructose - 2,6 - Bisphosphatase - 4 (PFKFB4) in Cancer. The FASEB Journal , 29 (1_supplement), 725.29. https://doi.org/10.1096/fasebj.29.1_supplement.725.29 214. Troyanovsky, S. (1999). Mechanism of cell - cell adhesion complex assembly. Current Opinion in Cell Biology , 11 , 561 566. 215. Van Oers, M. M., Pijlman, G. P., & Vlak, J. M. (2015). Thirty years of baculovirus - insect cell protein expression: From dark horse to mainstream technology. Journal of General Virology , 96 (1), 6 23. https://doi.org/10. 1099/vir.0.067108 - 0 216. Vander Heiden, M., Cantley, L., & Thompson, C. (2009). Understanding the Warburg effect: The metabolic Requirements of cell proliferation. Science , 324 (5930), 1029 1033. https://doi.org/10.1126/science.1160809.Understanding 217. Vandewalle, C., Van Roy, F., & Berx, G. (2009). The role of the ZEB family of transcription factors in development and disease. Cellular and Molecular Life Sciences , 66 (5), 773 787. https://doi.org/10.1 007/s00018 - 008 - 8465 - 8 218. Vandewynckel, Y., Laukens, D., Geerts, A., Bogaerts, E., Paridaens, A., Verhelst, Cancer. Anticancer Research , 33 (11), 4683 4694. Retrieved from http:// ar.iiarjournals.org/content/33/11/4683.abstract 219. Ventura, S. (2005). Sequence determinants of protein aggregation: Tools to increase protein solubility. Microbial Cell Factories , 4 (11), 1 8. https://doi.org/10.1186/1475 - 2859 - 4 - 11 165 220. Volmer, R., van der Ploeg, K., & Ron, D. (2013). Membrane lipid saturation activates endoplasmic reticulum unfolded protein response transducers through their transmembrane domains. Proceedings of the National Academy of Sciences , 110 (12), 4628 4633. https://doi.org/10.1073/pnas.121 7611110 221. Volmer, Romain, Ploeg, K. Van Der, & Ron, D. (2013). Membrane lipid saturation activates endoplasmic reticulum unfolded protein response transducers through their transmembrane domains. Proceedings of the National Academy of Sciences of the United States of America , 110 (12), 4628 4633. https://doi.org/10.1073/pnas.1217611110/ - /DCSupplemental.www.pnas.org/cgi/doi/10.1073/pnas.1217611110 222. Volmer, Romain, & Ron, D. (2015). Lipid - dependent regulation of the unfolded protein response. Current Opinion in C ell Biology , 33 , 67 73. https://doi.org/10.1016/j.ceb.2014.12.002 223. Wang, Dingzhi, & Dubois, R. N. (2012). Associations between obesity and cancer: The role of fatty acid synthase. Journal of the National Cancer Institute , 104 (5), 343 345. https://doi.org/10 .1093/jnci/djs010 224. Wang, Dong, Wei, Y., & Pagliassotti, M. J. (2006). Saturated fatty acids promote endoplasmic reticulum stress and liver injury in rats with hepatic steatosis. Endocrinology , 147 (2), 943 951. https://doi.org/10.1210/en.2005 - 0570 225. Wang, M., & Kaufman, R. J. (2014). The Impact of the Endoplasmic Reticulum Protein Folding Environment on Cancer Development. Nature Reviews Cancer , 14 (Sept 2014), 581 597. 226. Wang, S., & Kaufman, R. J. (2012). The impact of the unfolded protein response on human disea se. The Journal of Cell Biology , 197 (7), 857 867. https://doi.org/10.1083/jcb.201110131 227. Wang, X., Nath, A., Yang, X., Portis, A., Walton, S. P., & Chan, C. (2011). Synergy analysis reveals association between insulin signaling and desmoplakin expression in palmitate treated HepG2 cells. PLoS ONE , 6 (11). https://doi.org/10.1371/journal.pone.0028138 228. Wellen, K. E., & Hotamisligil, G. S. (2005). Inflammation , stress , and diabetes, 115 (5), 1111 1119. https://doi.org/10.1172/JCI200525102.The 166 229. Wingfield, P. T. (2 016). Overview of the Purification of Recombinant Proteins. Current Protocols in Protein Science , 80 , 6.1.1 - 6.1.35. https://doi.org/10.1002/0471140864.ps0601s80 230. Wu, R., Zhang, Q. - H., Lu, Y. - J., Ren, K., & Yi, G. - H. (2015). Involvement of the - XBP1 Pat hway and XBP1s - Dependent Transcriptional Reprogramming in Metabolic Diseases. DNA and Cell Biology , 34 (1), 6 18. https://doi.org/10.1089/dna.2014.2552 231. Yadav, R. K., Chae, S. - W., Kim, H. - R., & Chae, H. J. (2014). Endoplasmic Reticulum Stress and Cancer. Jou rnal of Cancer Prevention , 19 (2), 75 88. https://doi.org/10.15430/JCP.2014.19.2.75 232. Yan, J., Tsai, S. Y., & Tsai, M. J. (2006). SRC - 3/AIB1: Transcriptional coactivator in oncogenesis. Acta Pharmacologica Sinica , 27 (4), 387 394. https://doi.org/10.1111/j.1745 - 7254.2006.00315.x 233. Yang, W. S., Va, P., Bray, F., Gao, S., Gao, J., Li, H. L., & X iang, Y. B. (2011). The role of pre - existing diabetes mellitus on hepatocellular carcinoma occurrence and prognosis: A meta - analysis of prospective cohort studies. PLoS ONE , 6 (12). https://doi.org/10.1371/journal.pone.0027326 234. Yau, W. M., Wimley, W. C., Gaw risch, K., & White, S. H. (1998). The preference of tryptophan for membrane interfaces. Biochemistry , 37 (42), 14713 14718. https://doi.org/10.1021/bi980809c 235. Yoshida, H., Matsui, T., Yamamoto, a, Okada, T., & Mori, K. (2001). XBP1 mRNA is induced by ATF6 and spliced by IRE1 in response to ER stress to produce a highly active transcription factor. Cell , 107 (7), 881 891. Retrieved from http://www.ncbi.nlm.nih.go v/pubmed/11779464 236. Yu, L., Chen, X., Sun, X., Wang, L., & Chen, S. (2017). The glycolytic switch in tumors: How many players are involved? Journal of Cancer , 8 (17), 3430 3440. https://doi.org/10.7150/jca.21125 237. Yu, T., Zeng, X., He, F., Liao, Y., Qian, N., & Toi, M. (2012). Caveolin - 1 is related to invasion, survival, and poor prognosis in hepatocellular cancer. Medical Oncology , 29 (2977 984). 238. - 167 activates hypoxia - induced PFKFB4 ex pression in human bladder cancer cells. Biochemical and Biophysical Research Communications , 476 (3), 146 152. https://doi.org/https://doi.org/10.1016/j.bbrc.2016.05.026 239. Zhang, K., & Kaufman, R. J. (2008). From endoplasmic - reticulum stress to the inflammato ry response. Nature , 454 (7203), 1 20. https://doi.org/10.1038/jid.2014.371 240. Zhao, C. (2014). Wnt Reporter Activity Assay. Bio - Protocol , 4 (14), e1183. https://doi.org/10.21769/BioProtoc.1183 241. . (2019). Exosomal miR - 451a Functions as a Tumor Suppressor in Hepatocellular Carcinoma by Targeting LPIN1. , 53 (1), 19 35. https://doi.org/10.33594/000000118