An Investigation of the Mitochondrial Sirtuin Enzyme SIRT3 and Its Influence on the Response of Skeletal Muscle and Liver to Lipid Oversupply

An investigation of the mitochondrial sirtuin enzyme SIRT3 and its influence on the response of skeletal muscle and liver to lipid oversupply

Brenna Osborne

A thesis in fulfillment of the requirements for the degree of Doctor of Philosophy

Garvan Institute & St Vincent’s Clinical School Faculty of Medicine UNSW Australia

February 2015

i PLEASE TYPE THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet

Surname or Family name: OSBORNE

First name: BRENNA Other name/s:

Abbreviation for degree as given in the University calendar: PhD

School: St Vincent’s Clinical School Faculty: Faculty of Medicine

Title: An investigation of the mitochondrial sirtuin enzyme SIRT3 and its influence on the response of skeletal muscle and liver to lipid oversupply

Abstract 350 words maximum: (PLEASE TYPE)

SIRT3, a member of the sirtuin family of NAD+-dependent deacetylases, has been shown to directly regulate a range of mitochondrial proteins, suggesting a key role for this enzyme in energy metabolism. Some studies in SIRT3 knockout mice and high fat diet (HFD) fed mice support a link between low SIRT3 protein and detrimental metabolic outcomes. The aim of this thesis was to investigate the effects of acute tissue-specific overexpression of SIRT3 in liver and skeletal muscle under conditions of lipid excess, to see if increasing levels of this deacetylase enzyme could have beneficial metabolic effects. Eleven-fold overexpression of SIRT3 in hind-limb muscles in rats was achieved via intramuscular injection of SIRT3 adenoassociated virus (AAV). Rats were then fed a chow or HFD for 4 weeks after which they were assessed for either mitochondrial substrate oxidation using a Clark electrode, or insulin sensitivity via a hyperinsulinemic-euglycemic clamp. SIRT3 overexpression in muscle had no impact on body composition or serum profiles, however in isolated mitochondria SIRT3 overexpression caused an increase in oxygen consumption. Although HFD-feeding for 4 weeks induced significant impairments in skeletal muscle metabolism, including increased intramuscular triglyceride and a ~30% reduction in glucose uptake into muscle during the clamp, there was no significant effect of SIRT3 overexpression on these parameters. Approximately 2-fold SIRT3 overexpression in mouse liver was induced using the hydrodynamic tail vein injection (HTVI) technique. Isolated primary hepatocytes from SIRT3 overexpressing mice were found to have increased oxygen consumption and reduced triglyceride accumulation following fatty acid incubation compared to control hepatocytes. However, in vivo assessment of SIRT3 overexpression in liver tissue showed that even though there were changes induced in the amount of acetylation detected by mass spectrometry over the timeframe studied, SIRT3 overexpression had no effect on glucose tolerance, body composition or liver triglyceride accumulation in response to HFD. These results suggest that despite beneficial effects ex vivo, overexpression of SIRT3 in both skeletal muscle and liver in the whole body setting is not protective against the detrimental metabolic outcomes associated with feeding a HFD to rodents .

Declaration relating to disposition of project thesis/dissertation I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).

…………………………………………… ……………………………………..……………… ……….…………… Signature Witness Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

THIS SHEET IS TO BE GLUED TO THE INSIDE FRONT COVER OF THE THESIS

ORIGINALITY STATEMENT

‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.’

Signed ……………………………………………...... Brenna Osborne

Date ……………………………………………......

‘I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.'

Signed ……………………………………………...... Date ……………………………………………......

AUTHENTICITY STATEMENT

‘I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.’

Signed ……………………………………………...... Date ……………………………………………......

iii

ACKNOWLEDGEMENTS

Financial support via my Dora Lush Postgraduate Scholarship is gratefully acknowledged from the National Health and Medical Research Council (NHMRC Grant # 1017534). Many thanks also go to the staff of the Biomedical Testing Facility at the Garvan Institute for their excellent care of the animals used in this thesis. Mass spectrometry was conducted at the Bioanalytical Mass Spectrometry Facilities within the Analytical Centre of UNSW Australia with the assistance of Dr Valerie Wasinger. Subsidised access to this facility is acknowledged.

Tremendous thanks go to my supervisors Nigel Turner and Greg Cooney for being everything that good supervisors should be, knowledgeable, supportive, approachable, and for always having my best interests at heart. As supervisors they have shaped my work ethic and approach to scientific questions and have both been generous with their time, their animal skills, and their networks in the field of diabetes. Particular thanks go to Amanda Brandon, and her team of research assisstants Jen, Ella and Eurwin, for their help with animal experiments and technical expertise in the clamp technique.

On a more personal note, many people have contributed to the journey over the last few years, I’d like to thank all those in the Diabetes program at Garvan who have been great friends and lab buddies: Amanda, Ella, Magda, Lewin, Vi, Nicole, Jen, Lauren, Jane, Grace, Jayne, Nancy, as well as those from the Centenary Institute, particularly Fiona Keane and Mark Gorrell.

Special thanks go to my friends and family, especially to Ryan, for the ongoing support, and efforts to keep me happy, sane and well-caffeinated over the last few years.

TABLE OF CONTENTS

Table of Contents ...... v

List of Figures ...... x

List of Tables ...... xii

Abstract ...... xiii

Papers and Presentations Arising from this Thesis ...... xiv

Papers Arising in Conjunction with this Thesis ...... xv

Abbreviations ...... xvi

CHAPTER 1 General Introduction ...... 1 1.1 The Health Burden of Diabetes and Obesity ...... 1 1.2 Metabolism ...... 3 1.2.1 Obesity is a failure of energy homeostasis ...... 3 1.2.2 The importance of skeletal muscle and liver for metabolism ...... 3 1.2.3 Canonical insulin signalling ...... 5 1.3 Fat and Glucose Metabolism ...... 7 1.3.1 Glucose metabolism: Glycolysis ...... 7 1.3.2 β-Oxidation in detail ...... 8 1.4 Mitochondrial Metabolism ...... 9 1.4.1 The mitochondrion: A cellular powerhouse ...... 9 1.4.2 The tricarboxylic acid (TCA) cycle ...... 11 1.4.3 Oxidative phosphorylation ...... 12 1.4.4 Mitochondrial dysfunction and insulin resistance ...... 12 1.4.4.1 ER stress and oxidative stress models of insulin resistance ...... 13 1.4.4.2 Mitochondrial dysfunction model of insulin resistance ...... 14 1.5 Post Translational Acyl-modifications in Mitochondria ...... 15 1.5.1 Post translational modifications and regulation of mitochondrial metabolism ...... 15 1.5.2 Reversible lysine acetylation ...... 16 1.5.3 Acetylation is highly prevalent in the mitochondrion ...... 17 1.5.4 Sirtuin deacylase enzymes: An introduction ...... 18 1.6 Sirtuin Enzymes in Metabolism ...... 20 1.6.1 SIRT1: A deacetylase and master regulator of mitochondrial metabolism ...... 21 1.6.2 Metabolic role of SIRT4 ...... 22 1.6.3 Novel deacylase functions of SIRT5 ...... 23 1.6.4 SIRT2, SIRT6 and SIRT7 ...... 24 1.7 SIRT3: The Major Mitochondrial Deacetylase ...... 25 1.7.1 Sirtuins as sensors of nutritional flux ...... 25 1.7.2 The mitochondrial deacetylase SIRT3 ...... 25 1.7.3 Protein targets of SIRT3 ...... 27 1.7.4 Metabolic pathways regulated by SIRT3 ...... 28 1.7.5 SIRT3 and disease ...... 32

1.7.6 Timing and site of action may be important for SIRT3 function ...... 34 1.7.7 Evidence for non-enzymatic acetylation ...... 34 1.8 Main Aims: Modulating SIRT3 to Improve Metabolism ...... 36 1.8.1 Very few gain of function models for SIRT3 ...... 36 1.8.2 Central hypothesis of this project ...... 37 1.8.3 Specific aims ...... 37

CHAPTER 2 Research Design and Methods ...... 38 2.1 Animals and Diet Composition ...... 38 2.2 Blood Analysis ...... 38 2.2.1 Plasma non-esterified fatty acids (NEFA) ...... 39 2.2.2 Plasma triglyceride (TAG) ...... 39 2.2.3 Blood ketones ...... 39 2.3 Tissue Measurements ...... 40 2.3.1 Dismembration of tissue ...... 40 2.3.2 Isolation of proteins and determination of total protein content ...... 40 2.3.3 Western blot analysis ...... 41 2.3.4 Tissue triglyceride determination ...... 43 2.3.5 Tissue glycogen determination ...... 43 2.4 Mitochondrial Measurements ...... 44 2.4.1 Mitochondrial isolation ...... 44 2.4.2 Mitochondrial respiration using Clark-type electrode ...... 45 2.5 Statistics ...... 46

CHAPTER 3 Investigation of SIRT3 Overexpression in Skeletal Muscle of Chow and High Fat Diet-Fed Rats ...... 47 3.1 Introduction ...... 47 3.2 Methods ...... 50 3.2.1 Animals and diet composition ...... 50 3.2.2 Generation and propagation of SIRT3-FLAG-AAV ...... 50 3.2.3 SIRT3 overexpression in skeletal muscle using SIRT3-AAV ...... 51 3.2.4 Assessment of oxidative capacity in isolated mitochondria ...... 51 3.2.5 Analysis of insulin sensitivity by hyperinsulinemic-euglycemic clamp ...... 52 3.2.5.1 Dual jugular cannulation ...... 52 3.2.5.2 Hyperinsulinemic-euglycemic clamp ...... 52 3.2.5.3 Blood analysis ...... 53 3.2.5.4 Tracer dose ...... 54 3.2.5.5 Rate of tracer disappearance from plasma (Rd) ...... 54 3.2.5.6 Rate of glucose uptake (Rg’) in skeletal muscle ...... 54 3.2.5.7 Incorporation of tracer into glycogen ...... 55 3.2.6 Statistical analysis ...... 56 3.3 Results ...... 57 3.3.1 Overexpression of SIRT3 protein in rat skeletal muscle ...... 57 3.3.2 Effect of SIRT3 overexpression on total pan-lysine acetylation ...... 59 3.3.3 Effect of SIRT3 overexpression in rat skeletal muscle on oxidative capacity in isolated mitochondria ...... 59 3.3.4 The effect of 4 weeks of HFD on whole body insulin action in rats...... 61 3.3.5 The effect of 4 weeks overexpression of SIRT3 on insulin action in muscle of chow and high fat diet-fed rats ...... 63

3.3.6 Effect of SIRT3 overexpression on skeletal muscle glycogen and triglyceride content after 4 weeks HFD ...... 64 3.3.7 Insulin, but not SIRT3 or HFD, impacts on insulin signalling pathways ...... 66 3.4 Discussion ...... 68

CHAPTER 4 Investigation of SIRT3 Overexpression in the Liver ...... 75 4.1 Introduction ...... 75 4.2 Methods ...... 77 4.2.1 Animals and diet composition ...... 77 4.2.2 Generation and propagation of SIRT3-FLAG-pLIVE plasmid DNA ...... 77 4.2.2.1 Provenance of constructs ...... 77 4.2.2.2 Subcloning of SIRT3-FLAG into pLIVE ...... 78 4.2.2.3 Phosphatase treatment, ligation and transformation of SIRT3-FLAG-pLIVE ...... 78 4.2.2.4 Small scale preparation of plasmid DNA and sequencing ...... 79 4.2.2.5 Large-scale preparation of plasmid DNA ...... 79 4.2.3 Overexpression of SIRT3 in liver ...... 80 4.2.3.1 Preparation of buffer for hydrodynamic tail vein injection (HTVI) ...... 80 4.2.3.2 Preparation and injection of animal ...... 81 4.2.3.3 In vivo imaging of mice post-HTVI ...... 81 4.2.4 Gene expression analysis using quantitative PCR ...... 81 4.2.4.1 RNA preparation ...... 81 4.2.4.2 Quantitative PCR ...... 82 4.2.5 Isolation of primary hepatocytes from mice following HTVI ...... 83 4.2.6 Cell culture conditions ...... 83 4.2.7 Preparation of fatty acid or vehicle treatment media ...... 84 4.2.8 Oxygen consumption in hepatocytes ...... 84 4.2.9 Triglyceride in hepatocytes ...... 85 4.2.10 Substrate utilisation assays: ...... 85 4.2.10.1 Glucose oxidation in primary hepatocytes ...... 85 4.2.10.2 Palmitate oxidation in primary hepatocytes ...... 85 4.2.11 Assessment of body weight and body composition ...... 86 4.2.12 Assessment of metabolic parameters: GTT and serum measures ...... 86 4.2.13 Substrate oxidation in liver tissue homogenates ...... 87 4.2.14 Statistical analysis ...... 87 4.3 Results ...... 88 4.3.1 Verification of HTVI technique and SIRT3 overexpression ...... 88 4.3.1.1 SIRT3 mRNA is increased following HTVI ...... 88 4.3.1.2 Timecourse of SIRT3 overexpression in liver ...... 89 4.3.1.3 SIRT3 protein overexpression in liver, isolated mitochondria, and primary hepatocytes ...... 89 4.3.1.4 Verification of HTVI using in vivo imaging ...... 91 4.3.2 SIRT3 overexpression in an ex vivo model of isolated hepatocytes ...... 92 4.3.2.1 Viability and purity of murine hepatocyte isolations ...... 93 4.3.2.2 Oxygen consumption in primary hepatocytes after SIRT3 HTVI ...... 93 4.3.2.3 Oxygen consumption in SIRT3KO hepatocytes ...... 95 4.3.2.4 Triglyceride accumulation in cultured hepatocytes with SIRT3 overexpression ... 97 4.3.2.5 Substrate utilisation in primary hepatocytes after SIRT3 HTVI ...... 97 4.3.3 Acute SIRT3 overexpression in liver of chow and HFD-fed mice ...... 99 4.3.3.1 SIRT3 overexpression in vivo in liver at 3 weeks by western blot ...... 100 4.3.3.2 Metabolic characteristics and weight gain in acute cohort ...... 100 4.3.3.3 SIRT3 overexpression and glucose tolerance ...... 103

vii

4.3.3.4 SIRT3 overexpression in liver for 3 weeks increases substrate utilisation in ex vivo liver homogenates...... 103 4.3.4 48 hour fasting in an acute SIRT3 overexpression model ...... 105 4.3.4.1 Effect of 48 h fasting on body weight, tissue weights and serum parameters ... 105 4.3.4.2 Oxygen consumption in isolated mitochondria from SIRT3 overexpressing liver following 48 h fast ...... 107 4.3.4.3 Ketone bodies and SIRT3 overexpression during fasting ...... 108 4.3.5 Long-term overexpression of SIRT3 in vivo ...... 109 4.3.5.1 SIRT3 overexpression by western blot in 13 week cohort ...... 109 4.3.5.2 Effect of long-term overexpression of SIRT3 on body and tissue weights, fat distribution and liver triglyceride ...... 110 4.3.5.3 Long-term SIRT3 overexpression in mice and effects on glucose tolerance ...... 112 4.3.5.4 Mitochondrial substrate oxidation in isolated mitochondria in long-term SIRT3 overexpression ...... 114 4.4 Discussion ...... 115

CHAPTER 5 Mass Spectrometry Analysis of Lysine Acetylation in a SIRT3 Overexpression Model in Liver ...... 122 5.1 Introduction ...... 122 5.2 Method ...... 125 5.2.1 Workflow of liver MS/MS sample preparation and analysis ...... 125 5.2.2 Animals and mitochondria preparation ...... 126 5.2.3 Proteomics sample preparation ...... 126 5.2.4 LC-MS/MS analysis ...... 127 5.2.5 Data processing ...... 128 5.2.6 Bioinformatics analysis ...... 128 5.3 Results and Discussion ...... 129 5.3.1 Metrics of lysine acetylation sites in liver mitochondria ...... 129 5.3.2 Reduced abundance of acetylated proteins with SIRT3 overexpression ...... 129 5.3.3 Method development ...... 130 5.3.4 Quantitation of proteins in SIRT3KO and SIRT3 overexpression datasets ...... 131 5.3.5 Overlap between the pathways regulated by both SIRT3 deletion and overexpression 136 5.3.6 Mitochondrial proteins are almost wholly represented in cell compartment analysis 137 5.3.7 Results of Reactome pathway enrichment analysis ...... 138 5.3.8 Pathway enrichment analysis identifies mitochondrial metabolism and fat oxidation as key pathways in SIRT3 overexpression dataset ...... 141 5.3.9 SIRT3 overexpression in liver causes decreased acetylation of known SIRT3 targets at the peptide level ...... 141 5.3.9.1 Acetylation at K179 of SDH is decreased with SIRT3 overexpression, increased with SIRT3 deletion ...... 142 5.3.9.2 Known SIRT3 targets OTC and LCAD show variable effects of SIRT3 overexpression and deletion ...... 144 5.3.9.3 Known SIRT3 target HMGCS2 show variable effects of SIRT3 overexpression and deletion 144 5.3.9.4 Known SIRT3 target GDH shows decreased acetylation with SIRT3 overexpression 145 5.3.10 MS analysis of muscle SIRT3 overexpression ...... 145 5.3.11 Limitations of non- MS techniques ...... 146

viii

5.3.12 Limitations of MS techniques (I): Functional annotation of acetylation sites is needed for more SIRT3 targets ...... 147 5.3.13 Limitations of MS techniques (II): The problem of stoichiometry of PTMs ...... 148 5.4 Summary and Conclusions ...... 150

CHAPTER 6 General Discussion ...... 152

References ...... 164

LIST OF FIGURES

CHAPTER 1 General Introduction ...... 1 Figure 1.1 Interplay between two key metabolic tissues, liver and skeletal muscle: ... 4 Figure 1.2 Insulin signalling via the insulin receptor: ...... 6 Figure 1.3 Summary of the main reactions of glucose and fat metabolism, glycolysis and β-oxidation: ...... 8 Figure 1.4 Schematic of main mitochondrial processes of nutrient oxidation: ...... 10 Figure 1.5 The TCA cycle: ...... 11 Figure 1.6 The electron transport chain of OXPHOS: ...... 13 Figure 1.7 Subcellular localisation and activity of the sirtuin family of deacylases: ... 20 Figure 1.8 SIRT3 as a sensor of nutrient status: ...... 26

CHAPTER 3 Investigation of SIRT3 Overexpression in Skeletal Muscle of Chow and High Fat Diet-Fed Rats ...... 47 Figure 3.1 SIRT3 overexpression in rat skeletal muscle using SIRT3 AAV: ...... 58 Figure 3.2 SIRT3 overexpression and effect on global lysine acetylation: ...... 59 Figure 3.3 Percentage change in ADP-stimulated oxygen consumption in isolated mitochondria with SIRT3 AAV treatment in rats on chow and HFD: ...... 60 Figure 3.4 Reduced GIR and Rd with HFD indicate systemic insulin resistance in clamped rats: ...... 63 Figure 3.5 HFD reduces glucose uptake into muscle (Rg’) under clamp conditions which is not ameliorated by SIRT3 overexpression: ...... 64 Figure 3.6 SIRT3 overexpression has no effect on muscle triglyceride, glycogen content, or glycogen synthesis in chow and HFD-fed rat muscle: ...... 65 Figure 3.7 Effect of clamp and diet on SIRT3 protein, and insulin signaling: ...... 67

CHAPTER 4 Investigation of SIRT3 Overexpression in the Liver ...... 75 Figure 4.1 SIRT3 mRNA overexpression 3 weeks after HTVI using three different qPCR primer sets: ...... 88 Figure 4.2 Western blots of SIRT3 overexpression in liver: ...... 90 Figure 4.3 Verification of HTVI delivery method by in vivo imaging: ...... 92 Figure 4.4 Oxygen consumption in primary hepatocytes overexpressing SIRT3, over time with various inhibitors: ...... 94 Figure 4.5 Mitochondrial respiration is increased in SIRT3 overexpressing hepatocytes: ...... 95 Figure 4.6 Mitochondrial respiration in SIRT3KO hepatocytes compared to WT: ...... 96 Figure 4.7 Triglyceride accumulation in vehicle and fatty acid treated hepatocytes from SIRT3 overexpressing liver: ...... 98 Figure 4.8 Substrate utilisation in SIRT3 overexpressing hepatocytes with and without fatty acid treatment: ...... 99 Figure 4.9 SIRT3 overexpression 3 weeks after HTVI: ...... 100 Figure 4.10 Effect of 3 weeks of overexpression of SIRT3 and HFD on body weight, fat pad weight and liver triglycerides: ...... 102

Figure 4.11 Effect of 3 weeks overexpression of SIRT3 in liver on glucose tolerance in chow and HFD-fed mice: ...... 103 Figure 4.12 Substrate utilisation in liver homogenates of animals overexpressing SIRT3 for 3 weeks: ...... 104 Figure 4.13 The effect of SIRT3 overexpression on oxygen consumption in isolated liver mitochondria from 48 h fasted mice: ...... 107 Figure 4.14 Levels of beta hydroxybutyrate (β-OHB) in SIRT3 overexpressing mice following a 48 h fast: ...... 108 Figure 4.15 SIRT3 overexpression at protein level 13 weeks post-HTVI in chow and HFD-fed mice: ...... 109 Figure 4.16 Effect of SIRT3 overexpression and HFD on body weight and liver triglyceride in long-term cohort: ...... 112 Figure 4.17 Effect of SIRT3 overexpression and HFD on glucose tolerance after 12 weeks of HFD: ...... 113 Figure 4.18 Substrate oxidation in isolated liver mitochondria after 13 weeks SIRT3 overexpression in both chow and HFD: ...... 114

CHAPTER 5 Mass Spectrometry Analysis of Lysine Acetylation in a SIRT3 Overexpression Model in Liver ...... 122 Figure 5.1 Workflow and sample comparisons of LC-MS/MS experiments: ...... 125 Figure 5.2 Overlap between increased proteins in SIRT3KO and decreased proteins in SIRT3 overexpression: ...... 137 Figure 5.3 Cell compartment pathway analysis in acetylated proteins differentially acetylated in SIRT3KO and SIRT3 overexpressing liver mitochondria: ...... 138 Figure 5.4 Acetylation of individual peptides in a selection of known SIRT3 targets from SIRT3KO and SIRT3OE liver mitochondria: ...... 143

CHAPTER 6 General Discussion ...... 152 Figure 6.1 Illustration of the importance of stoichiometry for functional effects: ... 158

LIST OF TABLES

CHAPTER 1 General Introduction ...... 1 Table 1.1 Reported Substrates of SIRT3 ...... 29

CHAPTER 2 Research Design and Methods ...... 38 Table 2.1 Antibodies used for immunoblotting ...... 42

CHAPTER 3 Investigation of SIRT3 Overexpression in Skeletal Muscle of Chow and High Fat Diet-Fed Rats ...... 47 Table 3.1 Mean overexpression of SIRT3 protein with intramuscular injection of SIRT3 AAV in tibialis cranialis (TC) and extensor digitorum longus (EDL) assessed by western blot ...... 57 Table 3.3 Characteristics of rats undergoing hyperinsulinaemic-euglycaemic clamp . 62

CHAPTER 4 Investigation of SIRT3 Overexpression in the Liver ...... 75 Table 4.1 Hepatocyte yield and viability upon isolation ...... 93 Table 4.2 Physiological effects of 3 weeks overexpression of SIRT3 in liver in vivo in mice ...... 101 Table 4.3 The effects of 3 week SIRT3 overexpression on metabolic parameters in fed versus 48hr fasted mice ...... 106 Table 4.4 Effects of SIRT3 overexpression on physiological parameters after 13 weeks in chow and HFD-fed mice ...... 111 Table 4.5 Summary of liver SIRT3 overexpression experiments ...... 116

CHAPTER 5 Mass Spectrometry Analysis of Lysine Acetylation in a SIRT3 Overexpression Model in Liver ...... 122 Table 5.1 Metrics from MS/MS analysis ...... 130 Table 5.2 50 most abundant proteins detected in dataset with increased abundance in SIRTKO mouse liver compared to WT ...... 132 Table 5.3 The 50 most abundant proteins detected in dataset with decreased abundance in SIRT3 overexpressing mouse liver compared to control ...... 134 Table 5.4 Heirarchical pathway analysis from Reactome associated with SIRT3 overexpression in mouse liver…………………………………………………………………………...…139

xii

ABSTRACT

xiii

PAPERS AND PRESENTATIONS ARISING FROM THIS THESIS

Papers:

Osborne, B., Cooney, G.J., Turner, N., (2014) Are sirtuin deacylase enzymes important modulators of mitochondrial energy metabolism? (Review) Biochim Biophys Acta. Apr;1840(4):1295-302.

Oral Presentations:

Osborne, B, Brandon, AE, Stuart, E, Wright, LE, Cooney GJ, Turner N (2014) In vivo overexpression of the mitochondrial deacylase SIRT3 in muscle has no effect on glucose metabolism and peripheral insulin resistance in the high fat fed rat, Australian Diabetes Society Annual Meeting, August 2014, Melbourne, Australia.

Osborne, B, Cooney GJ, Turner N (2013) Overexpression of a mitochondrial deacetylase, SIRT3, in liver and its effects on metabolic parameters during excess lipid availability, IDF World Diabetes Congress, 2-6 December 2013, Melbourne, Australia.

Osborne B, Montgomery M, Cooney GJ, Turner N (2012) Acute Overexpression of SIRT3 in Liver has no Effect on Fat Mass or Glucose Tolerance in Mice Fed a High Fat Diet for Three Weeks, Pincus Taft Young Investigators Session, Australian Diabetes Society Annual Meeting, Gold Coast, Australia.

Poster Presentations:

Osborne B, Montgomery M, Cooney GJ, Turner N (2013) Overexpression of the mitochondrial deacetylase SIRT3 in liver has no effect on metabolic parameters in mice fed a high fat diet, despite increased oxygen consumption in isolated hepatocytes, Mitochondrial Physiology Society (MiP) Summer School, Mitochondrial Physiology: Theory & Praxis, University of Copenhagen, Copenhagen, Denmark.

Osborne B, Montgomery M, Cooney GJ, Turner N (2013) In vivo and ex vivo models of SIRT3 overexpression in liver show differing effects on metabolic parameters in mice in response to excess lipid, FASEB Science Research Conferences: Histone Deacetylases, Sirtuins and Reversible Acetylation in Signaling and Disease, Barga, Lucca, Italy.

Osborne B, Montgomery M, Cooney GJ, Turner N (2012) Effect of acute liver-specific overexpression of SIRT3 on metabolic parameters in mice fed a high fat diet for three weeks, AussieMit Meeting, Monash University, Melbourne, Australia.

Osborne B, Montgomery M, Reznick J, Cooney GJ, Turner N (2012) Effect of acute hepatic overexpression of SIRT3 on metabolic parameters in short-term high fat fed mice, 48th Annual European Association for the Study of Diabetes, Berlin, Germany.

xiv

PAPERS ARISING IN CONJUNCTION WITH THIS THESIS

Liu, M., Montgomery, M.K., Fiveash, C.E., Osborne, B., Cooney, G.J., Bell-Anderson, K., Turner, N. (2014) PPARα-independent actions of omega-3 PUFAs contribute to their beneficial effects on adiposity and glucose homeostasis. Scientific Reports. Jul 2;4:5538.

Montgomery, M.K., Osborne, B., Small, L., Cooney, G.J., Turner, N., (2013) Contrasting effects of medium- versus long-chain fatty acids in skeletal muscle, Journal of Lipid Research Dec;54(12):3322-33.

ABBREVIATIONS

2DG: 2-deoxy-D-glucose HK: hexokinase 3H-2DG: 3H-2-deoxy-D-[2,6]-glucose HMGCS2: 3-hydroxy-3-methylglutaryl AAV: adenoassociated virus CoA synthase 2 AcK: acetyl-lysine HRP: horseradish peroxidase ACS: acyl-CoA synthetase HTVI: hydrodynamic tail vein injection ADP: adenosine diphosphate KO: knockout AMPK: AMP-activated protein kinase LCACoA: long chain acyl-CoA ATP: adenosine triphosphate LCAD: long chain acyl-CoA BAT: brown adipose tissue dehydrogenase BSA: bovine serum albumin LC-MS/MS: liquid chromatography CS: citrate synthase mass spectrometry DAG: diacylglycerol MS: mass spectrometry DAVID: Database for Annotation, NAD: nicotinamide adenine Visualization, and Integrated Discovery dinucleotide EDL: extensor digitorum longus NEFA: non-esterified fatty acids ERRα: estrogen-related receptor alpha OE: overexpression ETC: electron transport chain OTC: ornithine transcarbamylase FA: fatty acid OXPHOS: oxidative phosphorylation FCCP: carbonylcyanide-p- PCR: polymerase chain reaction trifluoromethoxyphenylhydrazone PDH: pyruvate dehydrogenase G6P: glucose-6-phosphate PGC-1: peroxisome proliferator- GAPDH: glyceraldehyde-3-phosphate activated receptor gamma coactivator 1 dehydrogenase PPAR: peroxisome proliferator- GDH: glutamate dehydrogenase activated receptor GIR: glucose infusion rate PTM: post-translational modification GLUT: glucose transporter qPCR: quantitative PCR

H2O2: hydrogen peroxide RIPA: radioimmunoprecipitation assay HDAC: histone deacetylase RNA: ribose nucleic acid HFD: high fat diet ROS: reactive oxygen species

xvi

SDH: succinate dehydrogenase SIRT: sirtuin SOD: superoxide dismutase T2D: type 2 diabetes TAG: triglyceride TC: tibialis cranialis TCA: tricarboxylic acid VEH: vehicle treatment WAT: white adipose tissue WT: wild-type

xvii CHAPTER 1: Introduction

CHAPTER 1 GENERAL INTRODUCTION

1.1 The Health Burden of Diabetes and Obesity

As part of today’s modern lifestyle, decreasing daily activity and the around-the-clock availability of energy-dense fast food has led to an increase in the incidence of overweight and obese populations worldwide. Once considered a problem only in developed economies, overweight and obesity are now also on the rise in developing countries such as China, India and Brazil. The World Health Organisation (WHO) records that the worldwide prevalence of obesity has nearly doubled between 1980 and 2008 (WHO 2011). A widely adopted measure of overweight and obesity is the Body Mass Index, or BMI, a person’s weight (in kilograms) divided by the square of his or her height in metres. A person with a BMI of 30 kg/m2 or more is generally considered obese. A person with a BMI between 25 kg/m2 and 30 kg/m2 is considered overweight. In 2008, 35% of adults worldwide aged 20 and over had a BMI above 25 kg/m2, although in some WHO regions such as the Americas this figure was closer to 60% (WHO 2011). Overweight and obesity rates are shown to be increasing rapidly across Australia. In 1989, 44% of adults were overweight or obese, rising to 63% in 2011–12 (NHPA 2013). The obese also face health care costs that are 30% higher than those with more healthy body weights (Withrow and Alter 2011).

Overweight and obesity are major risk factors for diseases including type 2 diabetes (T2D), cardiovascular disease, osteoarthritis, and some forms of cancer (Kopelman 2000). Closely associated co-morbidities include insulin resistance, hypertension, dislipidemia, and inflammation. Because these markers often coexist in the one individual, this state became recognised as a distinct syndrome (Reaven 1988) now known as the Metabolic Syndrome, where insulin resistance is almost always an underlying characteristic (Ferrannini, Haffner et al. 1991, Alberti, Zimmet et al. 2005, Ferrannini 2006).

1 CHAPTER 1: Introduction

In response to nutrient absorption after a meal, there is a post-prandial rise in blood glucose concentrations, leading to the release of insulin from the beta cells of the pancreatic islets. Insulin resistance is defined as a condition where tissues of the body have a significantly reduced response to insulin. Under normal conditions, insulin responsive tissues such as skeletal muscle and white adipose tissue (WAT), are able to sense insulin levels rising and increase glucose uptake from the blood (Biddinger and Kahn 2006, Haas and Biddinger 2009). Simultaneously, insulin action inhibits gluconeogenesis or glucose production and release by the liver. In insulin resistance these responses are blunted, which results in increased production of insulin by the pancreatic beta cell in order to overcome the resistance and maintain normal blood glucose levels (euglycaemia). The overproduction of insulin can eventually lead to failure of the pancreatic beta cells, particularly in genetically susceptible individuals, leading to loss of blood glucose control and T2D with all its associated pathologies (Reaven 2005, Biddinger and Kahn 2006).

The International Diabetes Federation (IDF) currently estimates that 387 million people worldwide have diabetes (IDF 2014), and suggests a significant number (46%) of people remain undiagnosed. Diabetes can have dire consequences for health, including complications such as cardiovascular disease, retinopathy, nephropathy and neuropathy, as well as causing complications during pregnancy. In addition, the number of people living with diabetes is projected to increase to more than 500 million people by 2035 (IDF 2014). As such, dealing with the global healthcare burden due to this disease is often described as an epidemic (Zimmet, Magliano et al. 2014) and has been depicted as one of the great public health challenges of our time (Alberti and Zimmet 2014). In Australia alone, over 850,000 people have diabetes (IDF 2014). In 2005, Diabetes Australia estimated the total healthcare cost of diabetes to Australia, including lost productivity, to be AU$10.3 billion (Access Economics (The economic cost of obesity 2006).

2 CHAPTER 1: Introduction

1.2 Metabolism

1.2.1 Obesity is a failure of energy homeostasis

The common feature underpinning the development of overweight and obesity is failure to maintain energy balance (Kopelman 2000). Energy homeostasis is defined as the correct balance between calorie intake from the diet and energy expenditure. A diet of energy-dense foods combined with sedentary lifestyles is typical of many developed countries such as Australia. This creates an environment where many people are in a constant state of positive energy balance, ultimately leading to excess adiposity, obesity and the metabolic syndrome.

During periods of overnutrition, excess nutrients are stored as fat in adipose tissue, and adipose tissue has a remarkable ability to expand in order to continue to store fat as triglyceride. However, when overnutrition is essentially continuous and obesity develops, other tissues including skeletal muscle, heart and liver also exhibit increased storage of fat, a condition known as ectopic fat deposition (Perseghin, Scifo et al. 1999, Ravussin and Smith 2002). Excess lipid stored in tissues such as liver and muscle is associated with insulin resistance in these tissues, although the exact mechanisms under which ectopic fat distribution leads to insulin resistance is still not fully elucidated (Lettner and Roden 2008).

1.2.2 The importance of skeletal muscle and liver for metabolism

While oxidative metabolism of glucose and fatty acids occurs in nearly all cells, the capacity for some metabolic processes differs in certain organs as opposed to others. Tissues such as brain, adipose tissue, heart, liver and skeletal muscle all have distinct preferred fuels for oxidation and fuel storage capacities (Berg, Tymoczko et al. 2002). Of the many tissues in the body that are important for energy balance and glucose metabolism, skeletal muscle and liver are of particular interest.

Skeletal muscle accounts for up to 40% of body mass and can account for more than 30% of the resting energy expenditure in the body (Smith and Muscat 2005). Because

3 CHAPTER 1: Introduction

of the large contribution of muscle mass to total tissue mass and its insulin responsiveness, it is considered the most important tissue for insulin-stimulated glucose uptake, disposal, and storage in the body (DeFronzo and Tripathy 2009). Skeletal muscle is particularly relevant for substrate utilisation during times of high energy demand, such as during intense exercise when muscle blood flow can increase up to 20-fold (Andersen and Saltin 1985, Rose and Richter 2005). Muscle is able to use glucose, fatty acids and ketone bodies from the circulation as fuel sources, but also has reserves of glycogen that can be converted to glucose-6-phosphate (G6P) for use within the muscle cell.

The liver also plays a central role in whole body glucose and fatty acid homeostasis as the main site of gluconeogenesis, glycogenolysis, lipogenesis and ketogenesis. As can be seen in Figure 1.1, there is significant interplay between the liver and skeletal muscle as products of metabolism in muscle such as lactate must be carried to the liver for processing back into useable fuel such as glucose.

Figure 1.1 Interplay between two key metabolic tissues, liver and skeletal muscle: Adapted from Biochemistry 5th ed (Berg, Tymoczko et al. 2002)

4 CHAPTER 1: Introduction

Due to the anatomical location of the liver it is the major organ initially exposed to most of the compounds and nutrients that are absorbed by the intestine into the bloodstream, and hence plays a vital role in the generation of many metabolites that are used around the body. For example the liver is involved in the formation of glycogen, fatty acids, cholesterol, bile salts, the generation of lipoproteins, ketone bodies, as well as amino acid metabolism including protein synthesis and the urea cycle. The liver is an organ primarily of glucose storage and gluconeogenesis, producing and exporting glucose to the other tissues of the body. The hepatic glucose output of the liver is significantly reduced, and glycogen synthesis increased, following insulin stimulation (Berg, Tymoczko et al. 2002, Dzugaj 2006).

1.2.3 Canonical insulin signalling

Insulin is a major endocrine hormone that has a multitude of effects. Its main role is to regulate blood glucose levels by increasing glucose uptake into tissues, stimulating glycogen and lipid synthesis, and reducing gluconeogenesis and lipolysis. Upon binding to its receptor, insulin stimulates a well-described signalling cascade illustrated in Figure 1.2.

5 CHAPTER 1: Introduction

Figure 1.2 Insulin signalling via the insulin receptor: Classical insulin signalling occurs via a phosphorylation activated signalling cascade. Insulin binding causes autophosphorylation and activation of the insulin receptor, receptor activation leads to the phosphorylation of key tyrosine residues on insulin receptor substrate (IRS) proteins, activating phosphatidyl inositol kinase (PI3K), which in turn phosphorylates phosphatidylinositol 3, 4 (PIP2) to phosphatidylinositol-3,4,5 (PIP3). A key downstream effector of this action is Akt, activated via the phosphoinositide-dependent protein kinase-1 (PDK) and mammalian target of rapamycin complex 2 (mTORC2). Once active, Akt has effects on a many cell functions including to increase glycogen synthesis via phosphorylation and inactivation of glycogen synthase kinase 3 (GSK3), gene transcription via mTOR complex 1 (mTORC1), sterol regulatory element-binding proteins (SREBP), and forkhead box protein 01 (FOXO1), and on GLUT4 translocation via phosphorylation and activation of the 160 kDa-Akt Substrate (AS160) protein.

Briefly, the membrane-bound insulin receptor is autophosphorylated upon insulin binding, which catalyses a phosphorylation dependent signalling cascade of cellular proteins such as insulin receptor substrates (IRS), which once phosphorylated in turn recruit phosphoinositide 3-kinase (PI3K) and associated subunits which convert phosphatidylinositol 4,5-bisdiphosphate (PIP2) to phosphatidylinositol 3,4,5- tridiphosphate (PIP3), activating a phosphorylation cascade involving phosphoinositide-dependent protein kinase 1 (PDK1) which phosphorylates the

6 CHAPTER 1: Introduction

threonine 308 residue of Akt and mammalian target of rapamycin complex 2 (mTORC2) phosphorylation at serine residue 473 to give full activation of Akt. The activation of Akt ultimately affects a variety of cellular targets including inhibition of the 160 kDa-Akt Substrate (AS160), leading to GLUT4 translocation, inhibition of glycogen synthase kinase 3 (GSK3β) leading to increased glycogen synthesis and activation of mammalian target of rapamycin complex 1 (mTORC1)(Laplante and Sabatini 2009) (Saltiel and Kahn 2001). Lipid-induced defects in insulin signalling have been widely hypothesised to be an important mechanism of insulin resistance in liver and muscle (Shulman 2000, Samuel, Petersen et al. 2010), although this finding remains controversial as insulin signalling defects are not always observed with insulin resistance (Storgaard, Jensen et al. 2004)(reviewed in (Turner, Cooney et al. 2014)).

1.3 Fat and Glucose Metabolism

Energy metabolism can broadly be defined as the processes that convert nutrients into chemical energy and the building blocks necessary for the maintenance of cell function. Two key pathways for substrate oxidation and energy conservation are glycolysis (the breakdown of glucose) and β-oxidation (the breakdown of fatty acids) illustrated in Figure 1.3.

1.3.1 Glucose metabolism: Glycolysis

Once glucose has entered the cell, it is broken down into pyruvate via glycolysis, in a stepwise process. The first step is the phosphorylation of glucose by hexokinase, to create glucose-6-phosphate (G6P). G6P acts as a branch point for a number of metabolic pathways, thus in addition to being an intermediate in glycolysis, G6P can also be shunted into other pathways such as the synthesis and storage of glycogen. Secondly, G6P is isomerised to fructose-6-phosphate (F6P), and then further phosphorylated to fructose-1, 6-bisphosphate (F16BP) by phosphofructokinase. Both phosphorylation steps consume ATP. This second step of the reaction commits the glucose molecule to the glycolytic pathway, rather than conversion into another sugar

7 CHAPTER 1: Introduction

or storage as glycogen. The subsequent steps of the glycolytic pathway convert F16BP into pyruvate. Pyruvate can be converted to lactate or enter the mitochondria to be converted to acetyl-CoA by pyruvate dehydrogenase (PDH) complex and undergo further metabolism in the TCA cycle.

1.3.2 β-Oxidation in detail

Fatty acids are a major form of energy for many organisms. Short chain fatty acids are able to diffuse through the plasma membrane, while long-chain fatty acids require transporters to cross membrane barriers (Kamp and Hamilton 2006, Hagberg, Mehlem et al. 2013). Fatty acids are activated for metabolism by the addition of CoA by acyl-

Figure 1.3 Summary of the main reactions of glucose and fat metabolism, glycolysis and β- oxidation: Glycolysis occurs in the cytosol following entry of glucose to the cell via glucose transporters (GLUTs), and consists of a 9-step enzymatic process whereby the 6-carbon molecule glucose is broken down to the 3-carbon pyruvate molecule. β-oxidation of fatty-acyl-CoAs occurs once the fatty acid has entered the mitochondria, and is a repeating cycle of 4 reactions whereby fatty acyl-CoAs are broken down to acetyl-CoA and a fatty acyl-CoA 2 carbons shorter, the cycle continues depending on the chain length of the original fatty acid until all the carbons have been converted to acetyl-CoA.

8 CHAPTER 1: Introduction

CoA synthetases in the cytoplasm, and the resultant fatty acyl-CoA can be converted to triglyceride for storage or transported into the mitochondria via carnitine palmitoyltransferase (CPT) and oxidised. The β-oxidation pathway for saturated fatty acids is a repeating cycle of four enzymatic reactions (Figure 1.3), which act to produce one unit of acetyl-CoA and a shorter fatty acyl-CoA. Unsaturated fatty acids are also broken down in the β-oxidation pathway but require additional steps to rearrange any cis- double bonds. The key end products of β-oxidation are acetyl-CoA, a fuel source for the TCA cycle, and FADH2, and NADH, which can be used for ATP generation via the electron transport chain (ETC).

1.4 Mitochondrial Metabolism

1.4.1 The mitochondrion: A cellular powerhouse

At the subcellular level, energy balance and energy transfer converge at the mitochondrion. Mitochondria are crucial, double membrane-bound organelles found in most eukaryotic cells. They have a specialised structure comprising an outer membrane enclosing the intermembrane space, and a highly folded inner membrane that encloses the mitochondrial matrix.

Mitochondria have a critical role in energy transduction, cellular metabolism, signalling, and apoptotic pathways. Often referred to as the “powerhouse of the cell”, mitochondria are where energy is conserved by oxidising the nutrients that are consumed and transforming their energy into the most abundant form of chemical energy that is used by cells, ATP (adenosine triphosphate) (Figure 1.4). Mitochondria are not static organelles, but exist as a large dynamic network, with both metabolic plasticity and the ability to change morphology via fission and fusion to allow for adjustment in response to cellular stresses and metabolic requirements. For example in response to physiological changes in nutrient availability, cold exposure or disease states, mitochondria can change their number, shape, activity and preferred fuel substrates to appropriately sustain the bioenergetic needs of the cell. The critical role of mitochondria in regulating cellular homeostasis is highlighted by the fact that

9 CHAPTER 1: Introduction

mitochondrial dysfunction has been implicated in many diseases including neurodegeneration, cancer and diabetes, as well as the ageing process (Turner and Heilbronn 2008, Mammucari and Rizzuto 2010, Ren, Pulakat et al. 2010, Su, Wang et al. 2010).

Figure 1.4 Schematic of the main mitochondrial processes of nutrient oxidation: Nutrients from food we eat such as carbohydrates, fats and protein go through catabolic processes that ultimately lead to the TCA cycle in the mitochondria. The TCA cycle is linked through the production of electron donors NADH and FADH2 to the generation of ATP via the electron transport chain (ETC- shown here in green), the site of oxidative phosphorylation. Abbreviations: IMM, intermembrane space; H+, protons; I, II, III, IV, V, complexes of the ETC; Q, ubiquinone; CytC, cytochrome C.

10 CHAPTER 1: Introduction

1.4.2 The tricarboxylic acid (TCA) cycle

Fatty acid, glucose and amino acid metabolism all converge to a central pathway in the mitochondrial matrix, the TCA cycle. In this pathway shown in Figure 1.5, acetyl-CoA is completely oxidized to CO2 with the production of coenzymes NADH and FADH2. The TCA cycle comprises eight reactions, beginning with the condensing of acetyl-CoA and oxaloacetate to give citrate. Two of the next seven reactions release CO2 and conserve reducing equivalents as NADH, while the other reactions serve to regenerate oxaloacetate for further cycling and at the same time produce more reduced coenzymes NADH and FADH2, used to generate ATP via the ETC during oxidative phosphorylation (OXPHOS).

Figure 1.5 The TCA cycle: The tricarboxylic acid cycle (TCA), also known as the Krebs cycle and the Citric Acid cycle is a central metabolic pathway that takes place in the mitochondria of cells. It is composed of 8 reactions, nominally beginning with the entry of acetyl-CoA from the breakdown of carbohydrates and fatty acids. Acetyl CoA is completely oxidized to CO2 in the TCA cycle, with concomitant reduction of electron transporting coenzymes NADH and FADH2 which feed in to the electron transport chain (ETC). Adapted from Biochemistry 5th ed (Berg, Tymoczko et al. 2002).

11 CHAPTER 1: Introduction

Other TCA cycle intermediates such α-ketoglutarate are also used as the precursor material for many other biosynthetic processes such as purine nucleotide synthesis. Oxidative metabolism of other molecules, notably several amino acids, is also made possible by conversion at different reaction steps of the TCA cycle.

1.4.3 Oxidative phosphorylation

OXPHOS is a process that utilises an electro-chemical gradient across the inner mitochondrial membrane to produce ATP, via a co-ordinated group of protein complexes known as the electron transport chain (ETC). The OXPHOS process is intrinsically linked with both the TCA cycle, and β-oxidation which take place largely in the mitochondrial matrix. For every molecule of glucose oxidized to CO2 and H2O in the mitochondria, about 36 molecules of ATP can be generated. Electron transport through the ETC after the oxidation of NADH and FADH2 is tightly coupled to phosphorylation of ADP to ATP. This coupling comes about because electron transport produces a proton gradient across the inner mitochondrial membrane. As illustrated in Figure 1.6, electron flow through the respiratory chain results in the pumping of protons out of the mitochondrial matrix and the generation of membrane potential. ATP is synthesised when protons flow back to the matrix through complex V, also known as ATP synthase. Each complex of the ETC is actually a large multi-subunit assembly of proteins.

1.4.4 Mitochondrial dysfunction and insulin resistance

Insulin resistance has been shown to be an early metabolic defect in rodents during high fat feeding (Oakes, Cooney et al. 1997, Turner, Kowalski et al. 2013). Development of insulin resistance has also been shown to be an early event in the aetiology of T2D in humans (Martin, Warram et al. 1992, Vaag, Henriksen et al. 1995) . Several theories have been proposed to explain the development of insulin resistance in peripheral tissues for which some evidence exists in rodent and human models, however the precise molecular basis of insulin resistance remains unclear.

12 CHAPTER 1: Introduction

Figure 1.6 The electron transport chain of OXPHOS: CI to CV represent the 5 complexes of the ETC located on the inner mitochondrial membrane. Red arrows indicate proton flow across the membrane, while blue arrows represent electron flow. Abbreviations: IMM, intermembrane space; H+, protons; e-, electrons; Q, ubiquinone; CytC, cytochrome C.

1.4.4.1 ER stress and oxidative stress models of insulin resistance One of these models is endoplasmic reticulum (ER) stress, which is the perturbation of the protein synthesis, folding and quality control functions of the ER. ER function has been shown to be affected in both human and rodent models of insulin resistance, and can be alleviated by treatments that ameliorate this ER stress (Nakatani, Kaneto et al. 2005, Ozcan, Yilmaz et al. 2006, Boden, Duan et al. 2008). Oxidative stress via elevated reactive oxygen species (ROS) generation has also been suggested as an underlying mechanism for insulin resistance, primarily investigated in muscle, although there is evidence both for and against its role in insulin resistance in the literature (Bonnard, Durand et al. 2008, Hoeks, Briede et al. 2008, Anderson, Lustig et al. 2009, Hoehn, Salmon et al. 2009, Loh, Deng et al. 2009, Samocha-Bonet, Campbell et al. 2012). Obesity-induced chronic inflammation has also been linked to the development of insulin resistance in a variety of tissues including adipose tissue and liver, however a

13 CHAPTER 1: Introduction

clear causative role has not been fully established (Hotamisligil, Arner et al. 1995, Uysal, Wiesbrock et al. 1997, Gregor and Hotamisligil 2011).

1.4.4.2 Mitochondrial dysfunction model of insulin resistance Another popular theory for how insulin resistance might develop in peripheral tissues is in association with mitochondrial dysfunction. The mitochondrial dysfunction theory of insulin resistance posits that defects in mitochondrial oxidative metabolism leads to intracellular lipid accumulation, and through effects on signalling (Biddinger and Kahn 2006) and possibly other mechanisms, this excess lipid deposition may reduce insulin action (Kim, Hickner et al. 2000, Kelley, He et al. 2002, Lowell and Shulman 2005). Insulin resistance in skeletal muscle has indeed been correlated with increased ectopic lipid accumulation (Savage, Petersen et al. 2007, Turner, Cooney et al. 2014), particularly metabolically active long chain acyl-CoAs (LCACoAs) (Thompson, Lim-Fraser et al. 2000) and other bioactive lipid metabolites such as ceramides and diacylglycerol (DAG), which are thought to directly impact insulin signalling pathways (Samuel and Shulman 2012). Insulin resistance in liver is also closely associated with liver lipid deposition, as seen in conditions such as non alcoholic fatty liver disease (NAFLD) where liver steatosis is severe. In NAFLD, the liver appears to be able to upregulate mitochondrial oxidation to adapt to increased fatty acid levels (Begriche, Massart et al. 2013), although other defects such as increased reactive oxygen species (ROS) may instead explain the insulin resistance seen in liver (Begriche, Igoudjil et al. 2006).

With respect to mitochondrial dysfunction, early studies reported reduced activity of oxidative enzymes and decreased lipid oxidation in skeletal muscle of insulin-resistant and T2D humans (Kelley and Mandarino 2000, Kim, Hickner et al. 2000). Since then, many studies have found correlations between abnormal muscle mitochondrial metabolism and various insulin resistant states including obesity, type 2 diabetes, and ageing (Kelley, He et al. 2002, Petersen, Befroy et al. 2003, Turner and Heilbronn 2008) , however not all studies agree (Turner, Bruce et al. 2007, Hancock, Han et al. 2008). There is still confusion in the field and the exact molecular mechanisms remain unclear. In respect to lipid-induced insulin resistance in skeletal muscle, studies have

14 CHAPTER 1: Introduction

shown that increased fat can actually lead to increased fat oxidative capacity however coupled with an increase in fat-oxidation intermediates and ROS (Koves, Ussher et al. 2008, Boyle, Canham et al. 2011, Muoio and Neufer 2012). Back to back papers in Diabetes in 2013 highlighted the current controversy over the involvement of mitochondrial dysfunction in insulin resistance arguing both for and against its causative role. In this debate, Goodpaster suggests that derangements in mitochondria more broadly are central to the development of insulin resistance, while Holloszy refutes this with arguments that mitochondrial oxidation actually increases in conditions associated with insulin resistance including in obese and diabetic humans (Goodpaster 2013, Holloszy 2013). Some controversy also exists as to whether changes in mitochondrial biogenesis and mitochondrial content per cell is driving these effects, or if there is a defect in intrinsic mitochondrial metabolism (per mitochondria) that could be to responsible (Boushel, Gnaiger et al. 2007, Phielix, Schrauwen-Hinderling et al. 2008). Theories for what may underlie the development of mitochondrial dysfunction in insulin resistant tissues are many, but include altered mitochondrial fission and fusion dynamics (Jheng, Tsai et al. 2012, Sebastian, Hernandez-Alvarez et al. 2012), oxidative stress (Houstis, Rosen et al. 2006, Anderson, Lustig et al. 2009, Boden, Brandon et al. 2012), and epigenetic regulation of important mitochondrial control genes such as PGC1α at the level of DNA (Sookoian, Rosselli et al. 2010). One mechanism of mitochondrial regulation that may be involved in the development of insulin resistance will be discussed in detail below, the post-translational acyl- modification of mitochondrial proteins.

1.5 Post Translational Acyl-modifications in Mitochondria

1.5.1 Post translational modifications and regulation of mitochondrial metabolism

Mitochondrial metabolism is tightly regulated to maintain normal cellular function, and recently it has been shown that post-translational modification (PTM) of lysine residues on mitochondrial proteins is a key feature of this regulation. Lysines are

15 CHAPTER 1: Introduction

amino acid residues within proteins that are susceptible to a wide range of post- translational modifications. One important form of PTM is lysine acylation, which is the addition of specific acyl groups to the lysine residue. There are a large number of reported acyl modifications, such as acetylation, malonylation, succinylation, propionylation, butyrylation, and crotonylation (Du, Zhou et al. 2011, Peng, Lu et al. 2011, Denu and Gottesfeld 2012). Of these acyl modifications, reversible lysine acetylation has been the most extensively studied and a number of reports in the last decade have shown that acetylation is a highly prevalent and functionally relevant post-translational modification in mitochondria (Kim, Sprung et al. 2006, Choudhary, Kumar et al. 2009, Zhao, Xu et al. 2010, Anderson and Hirschey 2012, Rardin, He et al. 2013).

1.5.2 Reversible lysine acetylation

Acetylation was first described in histones in the nucleus approximately fifty years ago (Allfrey, Faulkner et al. 1964), and has since been shown to be a major regulator of gene expression and chromatin structure (Grunstein 1997). Acetylation is the covalent addition of an acetyl group from acetyl-CoA to the ε-amino group of lysine residues, which neutralises the positively charged lysine, changing the way it interacts with other nearby proteins and molecules (Glozak, Sengupta et al. 2005). As a result of this change, reversible lysine acetylation is known to affect enzymatic activity, protein stability, protein interactions and subcellular localisation of target proteins (Glozak, Sengupta et al. 2005, Xiong and Guan 2012).

With advances in mass spectrometry, proteomic studies in the last decade have shown that acetylation is a post-translational modification that extends well beyond the nucleus, being common to many non-histone proteins across the cell (Kim, Sprung et al. 2006, Choudhary, Kumar et al. 2009, Wang, Zhang et al. 2010, Zhao, Xu et al. 2010, Henriksen, Wagner et al. 2012, Lundby, Lage et al. 2012). In combination, these studies have shown that more than 4000 mammalian proteins are acetylated, pointing to reversible lysine acetylation as a major post-translational modification, that has a

16 CHAPTER 1: Introduction

regulatory scope comparable to that of other major protein modifications such as phosphorylation or ubiquitination (Kouzarides 2000). Consistent with this wide spectrum of target proteins, acetylation has been shown to influence a multitude of cellular processes, including apoptosis and the cell cycle, ageing, antioxidant defences, cancer, circadian rhythms, gene expression, and metabolism (Saunders and Verdin 2007, Choudhary, Kumar et al. 2009, Guan, Yu et al. 2010, Verdin, Hirschey et al. 2010, Finley and Haigis 2012).

1.5.3 Acetylation is highly prevalent in the mitochondrion

While it is wide-spread across the cell, acetylation is particularly prominent in mitochondria. In 2006, Kim et al used a combination of studies in HeLa cells and mouse liver to show that acetylation was abundant in mitochondria (Kim, Sprung et al. 2006). Subsequent proteomic studies examining the global acetylome of whole cells and tissues revealed thousands of potential acetylation sites, and mitochondrial proteins were highly represented in these studies (Choudhary, Kumar et al. 2009, Zhao, Xu et al. 2010). In addition to identifying individual proteins that may be acetylated, these reports also highlighted that proteins may be acetylated at multiple lysine sites. Integrated analysis of several recent mammalian proteomic screens estimates that approximately 35% of all mitochondrial proteins have at least one lysine that is able to be acetylated (Anderson and Hirschey 2012), while another recent report puts this number as high as 65% (Hebert, Dittenhafer-Reed et al. 2013). The majority of these proteins have only one or two acetylation sites, however, just over 10% of identified proteins have greater than 10 unique acetylation sites (Anderson and Hirschey 2012).

Additional proteomic studies have been published recently expanding the scope of specific tissues, species and conditions under which acetylation has been assessed. Some of these recent reports investigating the acetylome include examination of tissue specific changes across multiple rat tissues (Lundby, Lage et al. 2012), liver acetylation changes in mice during calorie restriction (Hebert, Dittenhafer-Reed et al. 2013), an alcoholic liver disease model in mice (Fritz, Galligan et al. 2012), and SIRT3 dependent changes in murine embryonic fibroblasts and mouse liver (Sol, Wagner et al. 2012,

17 CHAPTER 1: Introduction

Rardin, Newman et al. 2013). Collectively these studies have further highlighted the diversity and potential scope of acetylation for impacting upon mitochondrial metabolism.

Acetylated residues are observed in all major metabolic pathways in the mitochondria, including enzymes of the TCA cycle, the urea cycle and fatty acid β-oxidation (Zhao, Xu et al. 2010). With respect to the functional impact of acetylation, Zhao et al. showed that the enzyme malate dehydrogenase (MDH) in the TCA cycle was able to be acetylated at four lysine sites, and that this acetylation was dependent on the glucose concentration in the cells and caused an increase the activity of the enzyme (Zhao, Xu et al. 2010). Increasing the complexity of these systems, acetylation appears to both inhibit and activate different metabolic enzymes, such that while MDH is activated by acetylation, other mitochondrial enzymes, such as long chain acyl-CoA dehydrogenase (LCAD) and ornithine transcarbamylase (OTC) are inhibited by acetylation (Hirschey, Shimazu et al. 2010, Hallows, Yu et al. 2011). In addition, because some acetylated proteins have multiple lysine acetylation sites, validation studies are necessary to delineate which of the modified lysines are responsible for changes in activity of the target protein. In the case of LCAD, an enzyme involved in fatty acid oxidation, while there are 8 acetyl-lysine sites on the protein, only one has been shown to be associated with modified enzyme activity in the models studied thus far (Hirschey, Shimazu et al. 2010).

1.5.4 Sirtuin deacylase enzymes: An introduction

Reversible acetylation is controlled by the actions of acetyltransferase and deacetylase enzymes, which catalyse the addition and the removal of acetyl groups on lysine residues of target protein. In contrast to the many hundreds of enzymes that control protein phosphorylation and ubiquitination, there are only a limited number of regulatory enzymes for acetylation, with 22 acetyltransferases and 18 deacetylases identified in humans (Guan, Yu et al. 2010, Choudhary, Weinert et al. 2014). Amongst the enzymes that regulate deacetylation is the sirtuin family of deacetylase enzymes. Sirtuins are highly conserved nicotinamide adenine dinucleotide (NAD)+-dependant

18 CHAPTER 1: Introduction

deacylases and mono-ADP-ribosyl transferases (Michan and Sinclair 2007, He, Newman et al. 2012).

Sirtuins are categorised as Class III deacetylases and unlike classic deacetylases, which hydrolyse the acetyl group, sirtuins deacetylate lysine residues in an unusual chemical reaction that consumes NAD+ and releases nicotinamide, O-acetyl ADP ribose, and the deacetylated substrate. The name sirtuin is derived from its founding member, yeast Sir2 (silent information regulator 2). The sirtuins are highly conserved across species from bacteria to humans (Frye 2000) and are generally associated with increased lifespan in yeast, nematodes and fruit fly (Sinclair and Guarente 1997, Tissenbaum and Guarente 2001, Rogina and Helfand 2004) and improved healthspan1 in mammals (Howitz, Bitterman et al. 2003, Lagouge, Argmann et al. 2006, Minor, Baur et al. 2011) although not all studies are in agreement (Burnett, Valentini et al. 2011). In mammals there are seven sirtuin proteins (SIRT1-7), which display diverse subcellular localizations (Figure 1.7). SIRT1, SIRT6 and SIRT7 are chiefly nuclear, SIRT2 is cytoplasmic and SIRT3, SIRT4 and SIRT5 reside predominantly in the mitochondria. In addition to being present in disparate parts of the cell, it has come to light that mammalian sirtuins also catalyse a range of different enzymatic reactions other than deacetylation (Figure 1.7). These additional enzymatic roles include desuccinylation, demalonylation, demyristolation, debutyrylation and ADP-ribosylation (Haigis, Mostoslavsky et al. 2006, Du, Zhou et al. 2011, Fritz, Green et al. 2013, Jiang, Khan et al. 2013), and consistent with phylogenetic analyses of mammalian sirtuins (Frye 2000), it appears that different mammalian sirtuins have evolved to have distinct roles within the cell.

1 Healthspan is a term that defines the healthy lifespan of an animal, i.e. the period of life remaining free of illness or chronic injury Roth, G. S., D. K. Ingram and M. A. Lane (1999). "Calorie restriction in primates: Will it work and how will we know?" J. Am. Geriatr. Soc. 47(7): 896-903, Gerstbrein, B., G. Stamatas, N. Kollias and M. Driscoll (2005). "In vivo spectrofluorimetry reveals endogenous biomarkers that report healthspan and dietary restriction in caenorhabditis elegans." Aging cell 4(3): 127-137..

19 CHAPTER 1: Introduction

1.6 Sirtuin Enzymes in Metabolism

The mitochondrial sirtuin SIRT3, which is the primary focus of this thesis, will be described in detail in the following pages. However a brief introduction to the other members of the sirtuin family and their roles in metabolism will first be described (for an exhaustive review of the sirtuin family and their functions see (Morris 2013)). Other than SIRT1 and SIRT3 with their robust deacetylase activity providing a clear avenue of study, the exact functions of the other members of the sirtuin family are still not fully resolved, although this is slowly changing as new deacylation activities are reported in the literature.

Figure 1.7 Subcellular localisation and activity of the sirtuin family of deacylases: Representation of the localisation and deacyl-activity of sirtuins as currently understood. SIRT1 has been reported to have both nuclear and cytoplasmic localisation, SIRT2 also has targets in both the cytoplasm and the nucleus. SIRT3, SIRT4, and SIRT5 have primarily mitochondrial localisation. SIRT6 has nuclear localisation, but has also been reported in ER, while SIRT7 is localised in the nucleus and nucleolus. Black dotted arrows indicate ability to translocate to the nucleus.

20 CHAPTER 1: Introduction

1.6.1 SIRT1: A deacetylase and master regulator of mitochondrial metabolism

SIRT1 is the most widely studied of the mammalian sirtuins, and is expressed in the majority of tissues in the body. It deacetylates key histone residues involved in the regulation of transcription, and multiple non-histone protein targets including p53, forkhead box protein O1/3 (FOXO1/3), peroxisome proliferator-activated receptor gamma coactivator 1α (PGC-1α), and nuclear factor (NF)-κB (Rodgers, Lerin et al. 2005, Hubbard and Sinclair 2014). By targeting these proteins, SIRT1 is able to co-ordinately regulate numerous pathways, including muscle and fat differentiation, mitochondrial biogenesis, glucose and insulin homeostasis, hormone secretion, cell stress responses, genome integrity, and circadian rhythm (Chang and Guarente 2014). The myriad roles of SIRT1 in cellular metabolism are too numerous to cover in full here, but a link with longevity and healthy ageing persists with specific SIRT1 overexpression in the brain recently shown to enhance lifespan in mice (Satoh, Brace et al. 2013). SIRT1 activating compounds (STACs) have been developed and show promise for both longevity and improving healthspan (Minor, Baur et al. 2011, Mitchell, Martin-Montalvo et al. 2014), despite much initial controversy in the literature as to their precise mode of action (Dai, Kustigian et al. 2010, Pacholec, Bleasdale et al. 2010). Several animal models of SIRT1 manipulation have been reported in the literature, however studies have been held back by the finding that global germline deletion of SIRT1 is largely lethal or causes developmental defects (Cheng, Mostoslavsky et al. 2003, McBurney, Yang et al. 2003). Furthermore, there are conflicting reports of the effects of SIRT1 overexpression and deletion on metabolism in different models that requires more comprehensive work to be resolved (Sun, Zhang et al. 2007, Herranz and Serrano 2010, White, McCurdy et al. 2013, Boutant and Canto 2014).

Of interest to metabolic function, several independent SIRT1 transgenic mouse lines have shown beneficial metabolic phenotypes including a lean phenotype, glucose tolerance and resistance to HFD-induced metabolic perturbations compared to wildtype mice (Bordone, Cohen et al. 2007, Banks, Kon et al. 2008, Pfluger, Herranz et al. 2008), however an inducible-SIRT1 deletion model in adult mice showed no overt metabolic phenotype (Price, Gomes et al. 2012). A mutation in SIRT1 has also been

21 CHAPTER 1: Introduction

linked to a monogenic form of type 1 diabetes in a family (Biason-Lauber, Boni- Schnetzler et al. 2013). Some studies have attempted to delineate which tissues are important for SIRT1 function following these differing findings in whole-body models have also thrown up disparate findings, with a muscle-specific SIRT1 overexpression mouse model showing no effect on metabolic parameters (White, McCurdy et al. 2013) even though studies in muscle cell culture models showed there were effects on insulin action in muscle (Sun, Zhang et al. 2007).

1.6.2 Metabolic role of SIRT4

Relatively little is known about the substrate specificity and physiological relevance of SIRT4, or its precise role in metabolism. SIRT4 deletion does not cause any major phenotype in chow-fed mice under standard laboratory conditions (Haigis, Mostoslavsky et al. 2006). SIRT4KO mice do not display hyperacetylation of mitochondrial proteins, suggesting that SIRT4 is not involved in the global deacetylation of lysine residues in mitochondrial proteins (Lombard, Alt et al. 2007). Concomitant with this, SIRT4 was initially reported to exhibit little deacetylase activity, but did exhibit ADP-ribosylase activity (Haigis, Mostoslavsky et al. 2006). The ADP- ribosyltransferase activity of all mitochondrial sirtuins is low, and has even been described as an inefficient side-reaction (Du, Jiang et al. 2009), however it remains likely that this activity of SIRT4 is physiologically relevant, especially in the case of its regulation of a key enzyme, glutamate dehydrogenase (GDH) (Haigis, Mostoslavsky et al. 2006). Several groups have shown that SIRT4 is expressed abundantly in pancreatic beta cells and has the ability to regulate GDH and impact insulin secretion (Ahuja, Schwer et al. 2007, Nasrin, Wu et al. 2010). The regulation of GDH by SIRT4 has been shown to also be involved in pathways of cell survival and apoptosis (Verma, Shulga et al. 2013). SIRT4 has also been proposed to be a potent tumour suppressor in human cancer and mouse models, acting via fuel switching to support proliferation and tumour progression (Jeong, Xiao et al. 2013). In terms of lipid metabolism, SIRT4 knockdown was reported to enhance fatty acid oxidation (Haigis, Mostoslavsky et al. 2006, Nasrin, Wu et al. 2010). SIRT4 was also implicated in regulating lipid metabolism via action on malonyl-CoA decarboxylase (MCD), increasing malonyl-CoA levels and

22 CHAPTER 1: Introduction

affecting lipogenesis and fatty acid oxidation (Laurent, German et al. 2013). Consistent with this newly described function, SIRT4 deficient mice are partially protected from the effects of high fat feeding (Laurent, German et al. 2013). Interestingly, unlike levels of SIRT1 and SIRT3, which increase during calorie restriction (CR), SIRT4 levels have been shown to decrease in response to fasting and CR (Schwer, Eckersdorff et al. 2009). SIRT4 may also have a newly-described function as a lipoamidase to delipoylate lysine residues in the PDH enzyme complex in the liver (Mathias, Greco et al. 2014), suggesting that the functional repertoire of these less-studied sirtuin enzymes is still emerging.

1.6.3 Novel deacylase functions of SIRT5

The enzymatic role of SIRT5 in the mitochondria has until recently been as mysterious as the role of SIRT4. SIRT5 null mouse strains were shown to have unremarkable acetylation profiles, and SIRT5 was described as a weak deacetylase enzyme (Lombard, Alt et al. 2007). One of the first substrates reported to be deacetylated (and activated) by SIRT5 was carbamoyl phosphate synthetase (CPS1), an important enzyme of the urea cycle (Nakagawa, Lomb et al. 2009, Ogura, Nakamura et al. 2010). An exciting development in 2011 was the discovery of novel enzymatic activities for SIRT5, namely, lysine demalonylation and desuccinylation (Du, Zhou et al. 2011, Peng, Lu et al. 2011). Using a combination of in vitro studies and experiments in SIRT5 KO tissues, it was shown that amongst all protein deacetylases, SIRT5 has the unique ability to remove malonyl and succinyl groups from lysine residues. The significance of these post- translational modifications is still unclear, however many metabolic enzymes, including isocitrate dehydrogenase 2 (IDH2), GDH, citrate synthase, CPS1 and 3-hydroxy-3- methylglutaryl CoA synthase 2 (HMGCS2) have been identified to be malonylated or succinylated (Du, Zhou et al. 2011, Peng, Lu et al. 2011, Zhang, Tan et al. 2011). A detailed validation of the lysine modifications removed by SIRT5 will be necessary to fully appreciate the importance of this enzyme in mitochondrial metabolism.

23 CHAPTER 1: Introduction

1.6.4 SIRT2, SIRT6 and SIRT7

Ascribing functions to these lesser-known sirtuin proteins is a rapidly moving field. SIRT2 is localised to the cytoplasm, yet has been shown to bind to many transcriptional regulators leading to a wide range of downstream metabolic effects including increasing gluconeogenesis and promoting lipolysis via FOXO1 and FOXO3 (Wang, Nguyen et al. 2007, Wang and Tong 2009), and cell survival pathways via the transcription factor NF-κB (Rothgiesser, Erener et al. 2010). SIRT2 has also been postulated to have roles in cell cycle progression (Inoue, Hiratsuka et al. 2007)

SIRT6 is thought to exert its influence as a histone deacetylase in the nucleus (Figure 1.7). Mice deficient in the nuclear-located SIRT6 were originally shown to have severe defects associated with DNA repair, and only lived to one month of age (Mostoslavsky, Chua et al. 2006). Importantly, male transgenic mice overexpressing SIRT6 have been shown to have an increased lifespan of ~15% (Kanfi, Naiman et al. 2012). SIRT6 is a tumour suppressor that regulates glycolysis in cancer cells (Sebastian, Zwaans et al. 2012), and also has a role in maintaining circadian rhythm (Masri and Sassone-Corsi 2014). In respect to metabolism, SIRT6 has been reported to have the opposite effect as SIRT1 and increases the acetylation of PGC1α in hepatocytes, leading to a reduction in the gluconeogenic genes PEPCK and G6Pase (Dominy, Lee et al. 2012). Recently the role of SIRT6 in hydrolysing long-chain fatty acids was reported, followed closely by a report that fatty acids could enhance the deacetylation activity of this and other sirtuins (Feldman, Baeza et al. 2013, Jiang, Khan et al. 2013).

The function of the nucleolar SIRT7 remains elusive, although very recently new work has shown a role in metabolism, however not without controversy. Shin et al. report that SIRT7 null mice show hepatic steatosis morphology and increased total liver triglyceride compared to WT controls, while its re-introduction rescued the phenotype (Shin, He et al. 2013). In line with this, Ryu et al. report that SIRT7 functions as a nuclear transcriptional regulator of mitochondrial homeostasis, with its deletion in mice causing mitochondrial dysfunction and reduced OXPHOS function (Ryu, Jo et al. 2014). Conversely, Yoshizawa et al describe SIRT7 as a regulator of hepatic fatty acid

24 CHAPTER 1: Introduction

metabolism, with SIRT7 null mice shown to be resistant to HFD-induced fatty liver, obesity and glucose intolerance, and to show reductions in genes associated with lipogenesis (Yoshizawa, Karim et al. 2014). Further work needs to be done to delineate the roles of these less studied sirtuin proteins. As with any family of closely related proteins that have overlapping functions within the cell, the concept of redundancy must be acknowledged as a potential source of complexity in the analysis of the roles of individual sirtuins within metabolism. However, as can be seen in Figure 1.7, there does appear to be both subcellular compartmentation, and substrate specificity amongst the mammalian sirtuins in terms of their roles in cellular metabolism, and indeed the effects of sirtuin knockout models as described in this section do not support a large compensatory role for the other sirtuins in the case of the well described SIRT1 models at least, although this must be more thoroughly examined by the field.

1.7 SIRT3: The Major Mitochondrial Deacetylase

1.7.1 Sirtuins as sensors of nutritional flux

From a metabolic perspective, the NAD+ dependency of the sirtuin deacylation reaction implies that sirtuins are perfectly positioned at the crossroads of metabolic flux, with the ability to sense and respond to both the redox and metabolic state of the cell (Figure 1.8). Furthermore, the fact that acyl donor molecules for acylation such as acetyl-CoA, malonyl-CoA and succinyl-CoA are themselves important metabolites that fluctuate with changes in cellular energy flux, highlights the potential role of sirtuins in the metabolic control system.

1.7.2 The mitochondrial deacetylase SIRT3

SIRT3 is the most well characterised member of the mitochondrial sirtuins and is a soluble protein found in the mitochondrial matrix (Schwer, North et al. 2002). SIRT3 is nuclear encoded and expressed as a 45 kDa protein containing an N-terminal mitochondrial targeting sequence, that is cleaved off after import into the mitochondria, leaving an enzymatically active 28 kDa protein (Schwer, North et al.

25 CHAPTER 1: Introduction

2002). Reports persist that SIRT3 may also be found in the nucleus, (Scher, Vaquero et al. 2007, Iwahara, Bonasio et al. 2012), however most publications now agree that SIRT3 is predominantly mitochondrial in localisation (Cooper and Spelbrink 2008, Hallows, Albaugh et al. 2008, Gurd, Holloway et al. 2012).

Figure 1.8 SIRT3 as a sensor of nutrient status: This schematic outlines the relationship between nutrient pathways in the mitochondria with energy generation and acetylation/deacetylation. Sirtuins are NAD+ dependent, thus mitochondrial sirtuins such as SIRT3 are well placed to act as sensors of nutrient flux and respond to cellular energy changes such as the NAD/NADH ratio. Acetylation is also dependent on levels of its donor molecule, Acetyl CoA, which also fluctuates in the cell in response to nutrient intake.

SIRT3 has the most robust deacetylase activity of the three mitochondrial sirtuins, and SIRT3 deletion in mice leads to marked upregulation of global acetylation in mitochondria (Lombard, Alt et al. 2007, Peng, Lu et al. 2011). The levels of SIRT3 are highly responsive to the prevailing nutrient availability of the cell. Calorie restriction, fasting and exercise training have all been shown to increase SIRT3 levels in different tissues (Palacios, Carmona et al. 2009, Schwer, Eckersdorff et al. 2009, Hallows, Yu et

26 CHAPTER 1: Introduction

al. 2011, Hirschey, Shimazu et al. 2011), although there is some controversy with recent studies clearly showing a SIRT3 reduction in fasted skeletal muscle (Jing, O'Neill et al. 2013). In contrast, the expression and/or activity of SIRT3 has been shown to be lower in high-fat fed rodents (Bao, Scott et al. 2010, Hirschey, Shimazu et al. 2011, Kendrick, Choudhury et al. 2011), in mouse models of type 2 diabetes (Jing, Emanuelli et al. 2011) and in human subjects with the Metabolic Syndrome (Hirschey, Shimazu et al. 2011).

1.7.3 Protein targets of SIRT3

The first reported target of SIRT3 described in the literature was Acetyl-CoA synthetase 2 (AceCS2) (Hallows, Lee et al. 2006, Schwer, Bunkenborg et al. 2006). AceCS2 was shown to be activated upon the deacetylation of lysine residue Lys642 by SIRT3. Since then, SIRT3 has been reported to directly affect the function of diverse mitochondrial enzymes and components of mitochondrial respiration complexes which are discussed in detail below. A comprehensive list of reported SIRT3 substrates is presented in Table 1.1, highlighting the tissues or cell model in which these substrates were validated. Interestingly, most SIRT3 targets were confirmed in vitro in immortalised cell lines, or in vivo in mouse liver, with most studies taking advantage of liver tissues from SIRT3 knockout (SIRT3 KO) mouse models. Because the interaction of SIRT3 with its substrates may be tissue and condition specific, more studies on other tissues and in conditions other than global SIRT3 deletion are required.

More recently, detailed proteomic studies have sought to investigate the full acetylome regulated by SIRT3. Quantitative proteomics of SIRT3 KO murine embryonic fibroblasts (MEF) cells found that of all the acetylation sites identified in the screen, SIRT3 knockdown was found to modulate the acetylation of more than a quarter of them (Sol, Wagner et al. 2012). Further investigation of the effect of calorie restriction or fasting in combination with SIRT3 deletion using SIRT3 KO liver cells and high resolution mass spectrometry have further established SIRT3 as a major regulator of mitochondrial acetylation state (Hebert, Dittenhafer-Reed et al. 2013, Rardin, Newman et al. 2013). Collectively, these studies have identified a number of new putative SIRT3

27 CHAPTER 1: Introduction

targets that require validation, and as this information accumulates further understanding of the dynamic regulatory roles this enzyme plays in mitochondrial metabolism will emerge.

1.7.4 Metabolic pathways regulated by SIRT3

SIRT3 KO mice have been utilised extensively in the field to elucidate the role of SIRT3 in metabolism. Lombard et al. first published that a global SIRT3 KO mouse model, despite exhibiting marked hyperacetylation of mitochondrial proteins were similar to wildtype (WT) controls in their response to fasting or cold exposure, and exhibited normal weight, food intake and oxygen consumption (Lombard, Alt et al. 2007). An independently generated SIRT3 KO mouse model also showed marked hyperacetylation of mitochondrial extracts, and although accompanied by functional changes including a 50% reduction in ATP levels in liver, kidney and heart tissues, combined with a reduction in Complex I activity, there was no major pathology observed in these knockout animals (Ahn, Kim et al. 2008).

Since the initial investigation of these knockout mice, SIRT3 has been ascribed a number of metabolic functions. Hirschey et al. reported an important role for SIRT3 in regulating hepatic fatty acid oxidation (Hirschey, Shimazu et al. 2010). In this study SIRT3 KO mice displayed reduced markers of fatty acid oxidation compared to control mice under conditions of fasting and cold exposure. This was linked with regulation of the enzyme LCAD, which is deacetylated by SIRT3 in normal mice, activating LCAD and stimulating fatty acid oxidation when dietary energy intake is low (Hirschey, Shimazu et al. 2010).

Hallows et al. also employed metabolomics screening and peptide spot arrays in tissues and plasma from SIRT3 KO mice to show that SIRT3 is involved in regulating fatty acid oxidation, likely by affecting multiple β-oxidation enzymes in addition to LCAD (Hallows, Yu et al. 2011). In addition to directly regulating fatty acid oxidation, SIRT3 also regulates other aspects of lipid metabolism. The activity of HMGCS2, the rate-limiting step in ketone body formation, is regulated by SIRT3-mediated

28 CHAPTER 1: Introduction

deacetylation, with SIRT3 KO mice displaying an impaired production of ketone bodies during prolonged fasting (Shimazu, Hirschey et al. 2010).

Table 1.1 Reported Substrates of SIRT3 VALIDATION TISSUE TYPE / MODEL SUBSTRATE DEACET- REF STUDIED YLASE MS ASSAY ACAT1 Mouse liver mitochondria !" ! (Still, Floyd et al. 2013) (Hallows, Lee et al. 2006, In vitro & cell culture AceCS2 !" Schwer, Bunkenborg et models; E.coli al. 2006) Human aortic endothelial ALDH2 cells (HAECs) (Xue, Xu et al. 2012) overexpressing SIRT3

HepG2 cells; mouse

ATPase α / CV MEFs; mouse primary (Bao, Scott et al. 2010)

hepatocytes

Mouse heart and heart (Hafner, Dai et al. 2010,

CyclophillinD mitochondria from SIRT3 ! Shulga, Wilson-Smith et

KO mice; human cell lines al. 2010)

Mouse cardiomyocytes (Sundaresan, Gupta et al. FOXO3a ! overexpressing SIRT3 2009) (Lombard, Alt et al. 2007, SIRT3 KO mouse liver GDH ! Schlicker, Gertz et al. mitochondria 2008) (Shimazu, Hirschey et al. SIRT3 KO mouse liver HMGCS2 ! ! 2010, Hebert, Dittenhafer- mitochondria Reed et al. 2013) " CR model in brain, inner !" (Schlicker, Gertz et al. ! IDH2 ear & liver of SIRT3 KO 2008, Someya, Yu et al. mice; human cell lines 2010) Mouse cardiomyocytes (Sundaresan, Samant et Ku70 * and cell lines ! al. 2008) overexpressing SIRT3 " SIRT3 KO mouse liver: (Hirschey, Shimazu et al. LCAD !" ! high fat fed / fasted 2010) Mouse heart, SIRT3 KO (Pillai, Sundaresan et al. LKB1 / STK11 ! mouse heart 2010)

SIRT3 KO mouse liver (Hebert, Dittenhafer-Reed MDH2 mitochondria et al. 2013) Bovine liver mitochondrial ribosomes; human cell MRPL10 ! (Yang, Cimen et al. 2010) lines; SIRT3 KO mouse liver; C2C12 cells Table continued over page

29 CHAPTER 1: Introduction

Table 1.1 continued… (Ahn, Kim et al. 2008, NDUFA9/ CI SIRT3 KO mouse liver ! Rardin, Newman et al. 2013) Cell lines, (Samant, Zhang et al. OPA1 cardiomyocytes, cardiac ! 2014) fibroblasts, mouse heart Mouse liver & blood; (Hallows, Yu et al. 2011, OTC ! ! SIRT3 KO mice and CR Lundby, Lage et al. 2012)

P300/ Histone (Vempati, Jayani et al. HeLa cells H3* 2010)

Muscle SIRT3 KO mice; PDH E1α ! (Jing, O'Neill et al. 2013) C2C12 cells BAT, liver and liver mitochondria of SIRT3 KO (Cimen, Han et al. 2010, SDHa ! ! mice; brown preadipocyte Finley, Haas et al. 2011) cell line Cell lines and MEFs overexpressing SIRT3; (Qiu, Brown et al. 2010, SOD2 CR model in liver of ! Tao, Coleman et al. 2010) SIRT3 KO mice; SIRT3 KO MEFS *Nuclear localisation- requires further validation Abbreviations: CR = calorie restriction, KO = knockout mouse model, MEFs = mouse embryonic fibroblasts, BAT = brown adipose tissue, MS = mass spectrometric validation. Targets not mentioned elsewhere in text: ACAT1, acetyl-CoA acetyltransferase 1; ALDH2, aldehyde dehydrogenase 2; Ku70, Lupus Ku autoantigen protein p70; LKB1 / STK11, liver kinase B1/serine- threonine kinase 11; MRPL10, mitochondrial ribosomal protein L10; NDUFA9, NADH dehydrogenase [ubiquinone] 1 alpha subcomplex subunit 9.

Additionally, as mentioned above, SIRT3 also deacetylates and activates AceCS2, which is an enzyme that converts acetate (largely derived from lipid in the liver) to acetyl-CoA in extrahepatic tissues (Hallows, Lee et al. 2006, Schwer, Bunkenborg et al. 2006). These latter two functions of SIRT3 promote the coordinated use of lipid-derived carbons under conditions when lipid catabolism is high, such as during fasting.

A recent study in muscle of SIRT3KO mice showed that the deletion of SIRT3 in this tissue led to a decrease in pyruvate dehydrogenase activity (PDH), associated with the increased acetylation of the E1α subunit of this important enzyme in carbohydrate metabolism (Jing, O'Neill et al. 2013). This decrease in PDH activity in muscle switches cells towards fatty acid utilisation rather than glucose metabolism, even when nutrient

30 CHAPTER 1: Introduction

levels are high, a switch that is usually associated with fasting. This effect on PDH of SIRT3 deletion leads to decreased metabolic flexibility,

Another important pathway regulated by SIRT3 involves the acceleration of amino acid utilisation and subsequent disposal of nitrogen waste via the urea cycle. SIRT3 deacetylates and activates GDH, a key enzyme involved in the catabolic processing of most amino acids (Lombard, Alt et al. 2007). The reaction catalysed by GDH releases nitrogen to the urea cycle as ammonia, and SIRT3 enhances the urea cycle by activating the key mitochondrial enzyme involved, OTC (Hallows, Yu et al. 2011).

Several protein subunits of the mitochondrial respiratory chain have been reported to be regulated by SIRT3. These include NDUFA9 (complex I), succinate dehydrogenase subunit a (SDHa) (a subunit of complex II) and ATP synthase α subunit (complex V) (Ahn, Kim et al. 2008, Bao, Scott et al. 2010, Cimen, Han et al. 2010, Finley, Haas et al. 2011). Accordingly, tissues and cells with deletion or knockdown of SIRT3 show reduced activity of specific mitochondrial complexes, as well as reduced oxygen consumption. These changes in oxidative phosphorylation may, in part, explain the reduced ATP levels observed in tissues such as heart, liver and kidney in mice lacking SIRT3 (Ahn, Kim et al. 2008).

In addition to regulating enzymes involved in the metabolism of specific nutrients, SIRT3 also modulates a number of stress-related pathways in mitochondria. Firstly, SIRT3 enhances the capacity of mitochondria to detoxify reactive oxygen species (ROS), which are a by-product of mitochondrial substrate metabolism. SIRT3 deacetylates and activates isocitrate dehydrogenase 2 (IDH2) (Someya, Yu et al. 2010), which is a TCA cycle enzyme that plays a critical role in maintaining the mitochondrial pool of NADPH, which is in turn used by glutathione reductase to maintain glutathione in its reduced antioxidant form. Secondly SIRT3-mediated deacetylation of the superoxide scavenger Mn superoxide dismutase (SOD2) activates this enzyme and reduces mitochondrial ROS production (Qiu, Brown et al. 2010, Tao, Coleman et al. 2010). Thirdly, SIRT3 has also been shown to act via interaction with FOXO3a to increase the transcription of antioxidants in cardiomyocytes (Sundaresan, Samant et al. 2008). As well as its role in

31 CHAPTER 1: Introduction

regulating oxidative stress levels, SIRT3 also modifies the activity of another key stress- sensitive pathway in mitochondria, the mitochondrial permeability transition pore (mPTP). SIRT3 deacetylates one of the core regulatory components of the mPTP, cyclophilin D, leading to reduced opening of this mitochondrial pore (Hafner, Dai et al. 2010). This idea that SIRT3 may function as a mitochondrial fidelity protein that maintains the “health” of the mitochondria under stress is further supported by the finding that mitochondrial fusion protein optic atrophy protein 1 (OPA1) is a target of SIRT3, as mitochondrial fission and fusion processes are important for quality control in mitochondria (Samant, Zhang et al. 2014).

Collectively many of the metabolic pathways regulated by SIRT3 indicate that this enzyme may function as a master switch that mediates the change from glucose- mediated metabolism towards a fasting metabolism, promoting the utilisation of lipids and amino acids under conditions where nutrient availability is low. Furthermore, in addition to its effects on intermediary metabolism, SIRT3 also appears to play a key role in the resistance to various mitochondrial stresses.

1.7.5 SIRT3 and disease

Due to their ability to regulate longevity in lower organisms (Sinclair and Guarente 1997, Tissenbaum and Guarente 2001), mammalian sirtuin proteins have received much attention for their role in modifying age-related diseases. SIRT3 has been reported to mediate the benefits of calorie restriction (CR) on hearing loss triggered by increased oxidative damage in the cochlear in ageing mice (Someya, Yu et al. 2010). SIRT3 also protects against palmitate induced lipotoxicity and its attendant increase in ROS in kidney cells and hepatocytes (Bao, Scott et al. 2010, Koyama, Kume et al. 2011). Via its interaction with the mPTP, SIRT3 has also been shown to suppress age-related cardiac hypertrophy (Hafner, Dai et al. 2010). Recently, much work has focussed on the role of SIRT3 in cardiac disorders, with one report showing SIRT3 deficiency is linked with pulmonary arterial hypertension in rodents and humans (Paulin, Dromparis et al. 2014).

32 CHAPTER 1: Introduction

Cancer can be described as a prime example of aberrant metabolism, where the metabolism of a cell is largely shifted towards a state of increased proliferation and anabolic processes. It has been reported that loss of SIRT3 can trigger metabolic reprogramming, supporting tumorigenesis, and providing a selective advantage that drives tumour growth via increased glycolytic flux (Kim, Patel et al. 2010, Finley, Carracedo et al. 2011, Finley and Haigis 2012). Reduced SIRT3 activity has also been proposed to drive tumourigenesis through excess ROS production and stabilisation of HIF-1α (Bell, Emerling et al. 2011). Many transformed cell lines and human tumours have reduced SIRT3 expression (Shulga, Wilson-Smith et al. 2010) and there is a growing body of evidence supporting a role for SIRT3 as a tumour suppressor (Finley, Carracedo et al. 2011, Finley and Haigis 2012), including the observation that SIRT3 KO mice have a high incidence of spontaneous tumours (Kim, Patel et al. 2010). However, it should be noted that SIRT3 levels have also been shown to be increased in certain types of cancers (e.g. oral cancer) (Alhazzazi, Kamarajan et al. 2011), and a recent report suggests that SIRT3 favours carcinogenesis, by providing resistance to stress and apoptotic stimuli through its interaction with cyclophilin D (Verma, Shulga et al. 2013). Thus the precise role of SIRT3 in regulating tumourigenesis may be cell-type and condition-specific, and additional studies in this area are required.

With regards to metabolic disease, SIRT3 expression and/or activity has been consistently shown to be reduced in rodent models of obesity and T2D (Hirschey, Shimazu et al. 2011, Jing, Emanuelli et al. 2011). SIRT3 knockout mice display mild glucose intolerance and knockdown of SIRT3 in muscle cells reduces insulin action, likely as a result of oxidative stress (Jing, Emanuelli et al. 2011). SIRT3 has also recently been shown to be reduced in pancreatic islets from humans with T2D, with its knockdown in islet cell lines leading to β-cell dysfunction and increased ROS (Caton, Richardson et al. 2013). Verdin and colleagues have also shown that SIRT3 KO mice display accelerated development of the metabolic syndrome, but only when exposed to a very long-term (8-12 month) high fat diet (Hirschey, Shimazu et al. 2011). With respect to humans, a polymorphism in the human SIRT3 gene encoding a protein with reduced enzymatic efficiency is correlated with the development of the Metabolic Syndrome (Hirschey, Shimazu et al. 2011).

33 CHAPTER 1: Introduction

1.7.6 Timing and site of action may be important for SIRT3 function

One caveat of the abovementioned studies is that they have relied largely on global SIRT3 knockout mice to define the physiological roles of SIRT3. Deletion of SIRT3 in all tissues from birth is likely to elicit a number of developmental adaptations that may have secondary effects on metabolism. Furthermore, there has been little investigation of the effect of SIRT3 gain-of-function, and it is also unclear if particular tissues play a more dominant role in mediating the whole body effects of SIRT3. In this regard, Auwerx and colleagues recently generated liver- and muscle-specific SIRT3 knockouts and reported that these animals have no detectable changes in metabolic phenotype in response to high fat feeding (Fernandez-Marcos, Jeninga et al. 2012), throwing some doubt on the idea that changes in fatty acid oxidation in liver are responsible for the detrimental phenotype observed in the global SIRT3 KO mice (Hirschey, Shimazu et al. 2011). Of note is that in the tissue specific model, mice were only maintained on the high fat diet for 8 to 16 weeks, which is shorter than studies in the global SIRT3 KO. Also, as the authors point out, the genetic background of the animals, and the developmental onset of the SIRT3 deletion differ between models, which may additionally have a bearing on these disparate findings (Fernandez-Marcos, Jeninga et al. 2012). Overall these studies highlight that much work is required to fully elucidate the metabolic role of SIRT3.

1.7.7 Evidence for non-enzymatic acetylation

As noted above, SIRT3, SIRT4 and SIRT5 are all present in the mitochondria, and these enzymes have been reported to catalyse different reactions, with SIRT3 being the bona fide mitochondrial deacetylase, SIRT4 acting as a ADP-ribosylase and SIRT5 being a desuccinylase and demalonylase enzyme. Interestingly, despite the mitochondrial sirtuins being quite well studied, so far no mitochondrial acyltransferases have been described, although GCN5L1 has been reported in the literature as a possible mitochondrial acetyltransferase (Scott, Webster et al. 2012). Since the majority of the mitochondrial proteins are nuclear encoded, it is plausible that nuclear encoded mitochondrial proteins may be acylated prior to their trafficking to the mitochondria,

34 CHAPTER 1: Introduction

however, since proteins encoded by the mitochondrial genome are also found to be acetylated, it is likely that unidentified protein acetyltransferases are present (Kim, Sprung et al. 2006). One alternate possibility is that mitochondrial proteins may in some cases be directly acetylated due to high intra-mitochondrial concentrations of acetyl-CoA. Non-protein mediated addition of an acetyl group has been demonstrated in vitro for histones (Paik, Pearson et al. 1970) and more recently for mitochondrial proteins (Wagner and Hirschey 2014). Work by Wagner & Payne and others has led to more support for the idea that acetylation, and acylation, is driven by non-enzymatic processes (Wagner and Payne 2013, Weinert, Lesmantavicius et al. 2014). Wagner and Payne show that in vitro acetylation and succinylation of lysine residues can occur under the conditions likely to be present in the mitochondrial matrix, those of high concentrations of donor molecules acetyl-CoA and succinyl-CoA with an alkaline pH, and that this is likely to be independent of enzyme activity (Wagner and Payne 2013). Similar results have also been shown linking acetyl-CoA concentrations with acetylation in yeast mitochondria (Weinert, Lesmantavicius et al. 2014).

A follow-up review took this idea one step further, delineating the model that sirtuins may have evolved as a protein quality control mechanism to remove acyl groups that accumulate non-enzymatically on proteins, thus restoring proper metabolic control (Wagner and Hirschey 2014). In a system analogous to the way antioxidant systems have evolved to remove damaging ROS molecules under conditions of stress, sirtuins may have evolved to manage the huge amount of spontaneous acyl modifications that occur under conditions of metabolic stress, i.e. when levels of acyl-CoAs are elevated in the mitochondrial matrix following caloric restriction, fasting and caloric excess (Wagner and Hirschey 2014). This model displays a shift in thinking from the majority of literature that views sirtuins as a stand-alone regulatory system, but instead presents it as a putative quality control mechanism. This model with an emphasis on stress, also nicely accounts for the evidence that opposing stresses such as calorie excess and fasting have the same effect on levels of SIRT3 and mitochondrial acetylation in tissues such as muscle (Jing, Emanuelli et al. 2011, Jing, O'Neill et al. 2013).

35 CHAPTER 1: Introduction

The extent of acetylation changes and their functional significance has come under renewed scrutiny with the advent of new mass spectrometry techniques to assess the stoichiometry of acetylation changes across the proteome. Weinert et al. have shown in yeast that while relative changes in abundance of acetylation on a specific peptide between two sample sets can be quite high, the actual total fraction of the protein population which carries the acetyl modification is quite low (Weinert, Lesmantavicius et al. 2014). Estimated stoichiometries of mitochondrial acetylation are that 95% of acetylation sites occurred with <1% stoichiometry, which is substantially lower than the stoichiometry seen in phosphorylation modifications (Choudhary, Weinert et al. 2014). This low stoichiometry is also lower than that seen for acetylation in the nucleus, where the regulatory role of acetyltransferases has been confirmed, lending further support to the idea that mitochondrial acetylation may be non-enzymatic (Choudhary, Weinert et al. 2014).

1.8 Main Aims: Modulating SIRT3 to Improve Metabolism

1.8.1 Very few gain of function models for SIRT3

It is now clear that dynamic changes in post-translational modifications of mitochondrial proteins can influence energy metabolism. Due to their NAD+ dependency, the activity of mitochondrial sirtuins is intimately tuned to the metabolic and redox state inside mitochondria and alterations in specific acyl modifications appears to allow for rapid adjustments in mitochondrial metabolism and signalling in response to changes in nutrient flux.

Most studies to date in the literature on the role of SIRT3 have relied on global knockout mouse models, which only manifest a significant metabolic phenotype under specific conditions, and on in vitro studies. It remains to be determined if enhancing mitochondrial SIRT3 activity, by overexpressing this enzyme, can have beneficial metabolic effects that are therapeutically relevant for treating metabolic diseases such as obesity and type 2 diabetes.

36 CHAPTER 1: Introduction

1.8.2 Central hypothesis of this project

Global deletion of SIRT3 in mice is reported to be associated with a worsened metabolic outlook, including an accelerated development of the metabolic syndrome and impaired fatty acid oxidation. SIRT3 levels also appear to be reduced in obesity and T2D in humans, and in models of metabolic abnormality such as the ob/ob mouse. The central hypothesis for this project was that SIRT3 overexpression in metabolically relevant tissues such as liver and skeletal muscle could yield beneficial metabolic outcomes in vivo, via its action on fatty acid oxidation and glucose metabolism, and hence protect against the metabolic defects induced by high fat feeding.

1.8.3 Specific aims

• To investigate the effect of SIRT3 overexpression in vivo on mitochondrial metabolism and insulin action in skeletal muscle of chow and HFD-fed rats.

• To investigate the effect of SIRT3 overexpression in the liver on mitochondrial function, lipid metabolism and glucose homeostasis using two main models: o Ex vivo assessment of SIRT3 overexpression in isolated murine hepatocytes. o In vivo assessment of SIRT3 overexpression in liver tissue in mice.

• To assess the effect of varying levels of SIRT3, both overexpressed and deleted, on mitochondrial protein acetylation in the liver using mass spectrometry techniques.

37 CHAPTER 2: Research Design and Methods

CHAPTER 2 RESEARCH DESIGN AND METHODS

Detailed methods specific to each model can be found in the method section for that particular chapter. General biochemical techniques common to all chapters are detailed below.

2.1 Animals and Diet Composition

Mice and rats used in this thesis were sourced from Animal Resource Centre (Perth, Australia) or from Australian BioResources (Moss Vale, Australia). All animals were housed under standard laboratory conditions at 22 ± 0.5°C on a 12:12-hr light-dark cycle, with free access to water. Mice were fed ad libitum either a standard chow (8% calories from fat, 21% calories from protein, 71% calories from carbohydrate; Gordon’s Specialty Stock Feeds, Yanderra NSW, Australia), or high-fat diet made in house (HFD; 45% calories from fat (lard), 20% calories from protein, 35% calories from carbohydrates, 4.7kcal/g; based on Rodent Diet #D12451 Research Diets Inc., New Brunswick, NJ, USA). All experimental procedures were approved by the Garvan Institute/St. Vincent’s Hospital Animal Experimentation Ethics Committee and were in accordance with the National Health and Medical Research Council of Australia Guidelines on Animal Experimentation.

2.2 Blood Analysis

Blood for assessment of NEFA’s, triglycerides or ketones from plasma was collected either from tail-tip for in vivo sampling in the case of pre-fasting samples and ketones (mice), or from heart puncture at time of sacrifice. Blood from cardiac puncture was collected into 1.5ml tubes containing trace amounts of EDTA. Samples were centrifuged at 1000 g for 10 minutes and the top layer of plasma was carefully collected into a fresh tube and stored at -80 °C.

38 CHAPTER 2: Research Design and Methods

2.2.1 Plasma non-esterified fatty acids (NEFA)

NEFA were assessed by an enzymatic colorimetric method (NEFA-C; Wako Pure Chemical Industries, Osaka, Japan) immediately upon the first thaw cycle of plasma samples. The assay is based on the acyl-CoA oxidase method, which involves the acylation of CoA by the fatty acids in the serum, when in the presence of added acyl- CoA synthase enzyme. The resultant acyl-CoA is oxidized by the addition of acyl-CoA oxidase. This reaction also produces H2O2 which, in the presence of peroxidase, allows the oxidative condensation of 3-methy-N-ethyl-N (β-hydroxyethyl)-aniline with 4- aminoantipyrine. The resultant pigment is quantitated by measuring the absorbance at 550 nm on a microplate reader (SpectraMax 384Plus, Molecular Devices, CA, USA), and FFA concentration calculated from a standard curve (0-10 mM).

2.2.2 Plasma triglyceride (TAG)

TAG content was determined using an enzymatic colorimetric technique (Triglycerides GPO-PAP; Roche Diagnostics, IN, USA). Briefly, the assay uses lipoprotein lipase to break down plasma triglycerides into glycerol and FFA. In the presence of ATP and glycerol kinase (GK), the glycerol is converted to glycerol-3-phosphate, which then is oxidized by glycerol phosphate oxidase (GPO) to yield hydrogen peroxide. The oxidative condensation of 4-Chlorophenol and 4-aminophenazone in the presence of peroxidase and hydrogen peroxide produces a rose coloured dye (quinoneimine) which is measured at 490 nm on a microplate reader (SpectraMax 384Plus, Molecular Devices, CA, USA). The intensity of the colour formed is directly proportional to the triglyceride concentration in the sample. TAG content is calculated against a glycerol standard curve (Precimat Glycerol 21 mg/dL, Roche: 0-42 mg/dL).

2.2.3 Blood ketones

Blood ketones were measured from tail-tip or plasma samples using the Optium FreeStyle Ketone Blood β-Ketone Test Strips and the Abbott Optium FreeStyle Xceed Glucose Meter (Abbott Laboratories, Abbott Park, IL, USA). The detection range of this method is reported by the manufacturer to be 0.1 – 6 mmol/l. The meter measures

39 CHAPTER 2: Research Design and Methods

the change in electrochemical current across the capillaries of the strip generated by the conversion of the ketone 3-beta-hydroxybutyrate (beta-OHB) into acetoacetate in the presence of hydroxybutyrate dehydrogenase.

2.3 Tissue Measurements

2.3.1 Dismembration of tissue

All freeze-clamped liver and muscle tissue was powdered prior to analysis, using either a Mikro-Dismembrator (B.Braun Instruments, Melsungen, Germany), or a Cell Crusher device (Cellcrusher, Cork, Ireland), both of which were maintained at below freezing temperatures using liquid nitrogen, to prevent thawing of samples. Powdering of tissues is essential in tissues where overexpression may not be uniform across the entire tissue, to ensure samples used for analysis represent the organ as a whole.

2.3.2 Isolation of proteins and determination of total protein content

Powdered tissue (~40 mg) was homogenised in ice-cold radioimmunoprecipitation (RIPA) buffer (65 mM Tris [pH 7.4], 150 mM NaCl, 1% nonidet NP-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulphate (SDS), 10% glycerol) supplemented with deacetylase, protease and phosphatase inhibitors (10 mM nicotinamide, 10 mM sodium fluoride, 1 mM sodium orthovanadate (Na3VO4), 1 μg/ml leupeptin, 100 mg/ml phenylmethylsulfonyl fluoride (PMSF), and 2 mg/ml aprotinin) using a Polytron homogeniser (Kinematica AG, Lucerne, Switzerland) and solubilized with gentle shaking for 1 h at 4 °C. Insoluble proteins and connective tissue were removed by centrifugation (10000 g, 10 min, 4 °C). Total protein content of the supernatant was determined using the Bradford protein assay (Bradford 1976) provided by Bio-Rad as per the manufacturer’s instructions (Bio-Rad Laboratories, CA, USA), and relative absorbance measured at 595 nm on a microplate reader (SpectraMax 384Plus, Molecular Devices, CA, USA). Total protein concentration was calculated from a Bovine Serum Albumin (BSA; 0-1 mg/ml) standard curve.

40 CHAPTER 2: Research Design and Methods

2.3.3 Western blot analysis

Lysates were diluted to appropriate concentrations (1-5 mg/ml) and prepared in Laemmli buffer (62.5 mM Tris-HCl, 2% SDS, 10% glycerol, 5% 2-mercaptoethanol, 0.0025% bromophenol blue (Laemmli 1970)), and heated (65 °C, 10 min) to completely denature all proteins. Samples were resolved on 10% Tris-based SDS-polyacrylamide gel electrophoresis (SDS-PAGE) prepared using Biorad Electrophoresis Gel Purity reagents (Biorad Laboratories) alongside a molecular weight marker (Kaleidoscope Plus prestained standards, Bio-Rad Laboratories) to confirm molecular weight of proteins of interest. Gels were run at 187 V for approximately 1 hour, and proteins were transferred to a PVDF membrane (Hybond-P, Amersham Biosciences, NJ, USA) at 80 V for 80 minutes or 23 V for 16 hours. Membranes were blocked in 1% BSA-Tris buffered saline with 0.1% Tween-20 (TTBS) for 1.5 hr at room temperature before probing with primary and secondary antibodies as detailed below. PVDF membrane sections were incubated with primary antibody for 2 hours at room temperature or overnight at 4 °C. Following incubation, excess primary antibody was removed from the membranes by repeated washes in TTBS, followed by incubation (2 hr, room temperature) in the appropriate secondary antibody. Immunolabelled bands were visualised by chemiluminescence (Western Lightning, Perkin Elmer Life Sciences, MA, USA) with exposure to film (Super RX, Fuji Photo Film Co., Tokyo, Japan). Films were developed using an automatic film developer (Fuji RGII X-ray Film Processor, Tokyo, Japan), and protein bands were quantitated by performing densitometry on scanned films (Epson Perfection V700) using the PlotProfile functions in Image J (Version 1.38x, NIH, Bethesda, MD).

41 CHAPTER 2: Research Design and Methods

Table 2.1 Antibodies used for immunoblotting Antibody Source Secondary Primary Cat # antibody Antibody conc. Human/mouse sirtuin 3 Cell Signaling Donkey anti-rabbit 1:1000 5490S (h/mSIRT3) (D22A3), Technology Inc., (DAR) -HRP recognises human, mouse Danvers, MA, USA (Jackson and rat ImmunoResearch, PA, USA) Human sirtuin 3 (hSIRT3) Cell Signaling DAR 1:1000 2627S (C73E3), recognises human Technology Inc. only Pan-acetyl-lysine (AcK) Cell Signaling DAR 1:1000 9441S Technology Inc. ImmuneChem DAR 1:1000 ICP0380 Pharmaceutical Inc, Burnaby, BC, Canada Mitomix OXPHOS Cocktail Mitosciences, Sheep anti-mouse 1:1000 MS601 recognising several Eugene, OR, USA (SAM) - HRP (GE subunits of the ETC: Healthcare, Buckinhamshire, • Complex I - NDUFB6, UK) • Complex II - FeS subunit • Complex III - Core2 subunit • Complex IV - subunit II • Complex V- subunit α Akt Cell Signalling DAR 1:1000 9272 Technology Inc. Phospho-Akt (Ser273) Cell Signalling DAR 1:1000 9271 Technology Inc. AS-160 (C6947) Cell Signalling DAR 1:1000 2670 Technology Inc. Phospho-AS-160 Cell Signalling DAR 1:1000 4288 Technology Inc. Glyceraldehyde-3- Cell Signaling DAR 1:1000 2118 phosphate dehydrogenase Technology Inc. (GAPDH) Porin/ voltage-dependant Cell Signaling DAR 1:1000 4866 anion channel 1 (VDAC1) Technology Inc. Skeletal α-actin (clone 5C5) Sigma, St Louis, SAM 1:1000 A2172 MO, USA

42 CHAPTER 2: Research Design and Methods

2.3.4 Tissue triglyceride determination

Lipids were extracted from tissues by standard methods (Folch, Lees et al. 1957). 30 mg powdered liver or 50 mg powdered muscle was weighed out with liquid nitrogen to prevent thawing. Frozen tissue was homogenised in 2 ml chloroform:methanol (2:1 vol/vol) in a glass homogeniser and dispensed into a 10 ml tube. The homogeniser was rinsed with an additional 2 ml of chloroform:methanol to ensure all triglyceride was collected. Samples were extracted overnight at room temperature on an orbital mixer. 2 ml of 1 M sulphuric acid was added to the extract and phases were separated by centrifugation at 2000 g for 10 mins. The lipid-containing lower phase was collected using a pasteur pipette into 6 ml glass scintillation vials. Extracts were dried down under nitrogen gas at 45 °C. Lipids were redissolved in 300 µL absolute ethanol. Triglycerides were determined using the enzymatic colorimetric technique used in Section 2.2.2 (Triglycerides GPO-PAP; Roche Diagnostics, IN, USA).

2.3.5 Tissue glycogen determination

~30 – 40 mg of tissue was digested in 200 µl of 1 M potassium hydroxide at 70 °C. The glycogen was precipitated with the addition of 1.75 ml 95% ethanol with 75 µl saturated sodium sulphate. The samples were centrifuged at 13000 g at 4 °C for 15 min with the resulting supernatant discarded. The pellet was dissolved in 200 µl of distilled water at 70 °C and glycogen precipitated with the addition of 1.8 ml of 95% ethanol, which was centrifuged again. Digestion of glycogen into glucose was performed by the addition of 1 ml of a 3.5 U/ml amyloglucosidase (Sigma, St Louis, MO, USA) solution prepared in 0.25 M acetate buffer, pH 4.75. Samples were incubated at 37 °C overnight with constant agitation.

The glucose of the digest was assayed using a glucose assay buffer which contained 0.5 mg/ml 4-aminoantipyrinem pH 7 containing 1% phenol. 50 µl of standards were pipetted into the 96-well microplate with the addition of 250 µl of glucose assay buffer. The microplates were incubated at 37 °C for 25 mins and the absorbance was

43 CHAPTER 2: Research Design and Methods

determined at 490 nm on a microplate reader (SpectraMax 384Plus, molecular Devices, Sunnyvale, CA, USA). The glucose content was determined by comparing the absorbance of the sample with a standard curve made from glucose (0-2 mM).

2.4 Mitochondrial Measurements

2.4.1 Mitochondrial isolation

Mitochondria were isolated from fresh muscle and liver by the commonly used differential centrifugation technique (Chappell and Perry 1954, Turner, Li et al. 2008). Liver was dissected and rapidly placed in ice-cold isolation medium (250 mM sucrose, 10 mM Tris-HCl, 1 mM EGTA, 1% fatty-acid free BSA, pH 7.4) supplemented with deacetylases and protease inhibitors (10 mM nicotinamide, 20 mM sodium fluoride, 0.8 mM sodium orthovanadate (Na3VO4), 1 µg/ml leupeptin, 100 mg/ml phenylmethylsulfonyl fluoride (PMSF), and 2 µg/ml aprotinin) and diced finely with scissors, before homogenisation in a glass/teflon dounce homogeniser. Muscle was dissected rapidly, weighed, and placed in ice-cold CP-1 isolation medium (100 mM KCl, 50 mM Tris/HCl, pH 7.4, and 2 mM EGTA, 0.2% fatty-acid free BSA), supplemented with inhibitors as for liver, digested on ice for 3 min in CP-2 medium (CP-1, to which was added 5 mM MgCl2, 1 mM ATP) and the addition of 1.5 mg/g tissue Nagarse bacterial proteinase (Type XXIV, Sigma P8038) and homogenised gently using an ultraturrax homogeniser (IKA-Works, Staufen im Breisgau, Germany). Liver and muscle homogenates were cleared of cell debris, nuclei, and any unbroken cells by centrifugation at low speed (800 g, 5 min, 4 °C). The resultant supernatant was spun at high speed (10600 g, 10 min, 4 °C) to pellet mitochondria. Pellets were washed twice by resuspension in isolation medium followed by centrifugation (10600 g, 10 min, 4 °C). For downstream processing of mitochondria pellets such as western blot, final wash was in isolation medium without BSA, the supernatant discarded, and the mitochondrial pellet was snap frozen in liquid nitrogen before storage at -80 °C. Mitochondria to be assessed for respiration were resuspended in ice-cold respiration medium (225 mM mannitol, 75 mM sucrose, 10 mM Tris-HCl, 10 mM KH2PO4, 0.8 mM

44 CHAPTER 2: Research Design and Methods

MgCl2, 10 mM KCl, 0.1 mM EDTA, pH 7.0). Mitochondrial protein content was measured using the Bradford method as described earlier.

2.4.2 Mitochondrial respiration using Clark-type electrode

Oxygen consumption was measured polarographically at 30 °C using a Clark-type electrode (Strathkelvin Instruments, Motherwell, Scotland). The principle of the clark- type electrode is the measurement of oxygen consumption using a silver anode and a platinum cathode connected by a salt bridge (KCl) (Clark, Wolf et al. 1953). Oxygen reduction and electron flow are coupled by the following reactions:

4Ag + 4Cl- 4 AgCl + 4e- e- 4H+ + 4 + O2 2H2O

Oxygen concentration is proportional to current (electron flow) at the cathode, measured as voltage by the system. ~100-300 µg of muscle mitochondria or ~200- 500 ug liver mitochondria were added to the 1 ml closed cell respiration chamber with air- saturated pre-warmed respiration buffer as described above, with the addition of 0.3% BSA. Respiration was measured in four separate substrate solutions, substrate combinations used were 5 mM pyruvate + 2 mM malate, 10 mM succinate + 4 μM rotenone, 10 mM glutamate + 2 mM malate, and 20 μM palmitoyl carnitine + 2 mM malate. State II respiration (basal) was measured for several minutes in the presence of substrate alone. State III respiration (oxidative phosphorylation) was initiated by adding 0.2 mM ADP. State IV respiration was measured via the addition of 5 μg oligomycin to block mitochondrial ATP synthesis. 2 μM carbonylcyanide-p- trifluoromethoxyphenylhydrazone (FCCP) was added to stimulate maximal (uncoupled) respiration. Air-saturated media was calculated to contain 223 nmol O2 / ml and measurements were calculated in nmol oxygen/min/mg of mitochondrial protein. Alternate methods for oxygen consumption measurements using the Seahorse XF analyser are described in Chapter 4.

45 CHAPTER 2: Research Design and Methods

2.5 Statistics

All data are reported as means ± SEM. Statistical analysis relevant to individual chapters are detailed therein. In general, data were analysed with unpaired or paired t- tests and two-way ANOVA where appropriate. Statistical significance was accepted at P<0.05. Statistical analysis was performed in GraphPad Prism software (Prism 6, Version 6.0b, Oct 3, 2012).

46 CHAPTER 3: Muscle Overexpression

CHAPTER 3 INVESTIGATION OF SIRT3 OVEREXPRESSION IN SKELETAL MUSCLE OF CHOW AND HIGH FAT DIET-FED RATS

3.1 Introduction

Obesity and T2D are characterised by underlying peripheral insulin resistance at the level of skeletal muscle. The precise molecular basis of muscle insulin resistance has not been fully elucidated, although it has been shown to be correlated with increased lipid accumulation in skeletal muscle (Savage, Petersen et al. 2007, Turner, Cooney et al. 2014). Mitochondrial dysfunction has been proposed as one of the factors that may underlie this increased accumulation of intramuscular lipid (Lowell and Shulman 2005). There is some debate in the literature as to the causative role of mitochondrial dysfunction in insulin resistance (Goodpaster 2013, Holloszy 2013). Despite this controversy, mitochondrial function is clearly important for healthy and insulin- responsive skeletal muscle, and unravelling the role of SIRT3 in muscle mitochondrial metabolism is of clear importance.

One process that has been shown to regulate mitochondrial metabolism is post- translational modification of proteins by acetylation (Kim, Sprung et al. 2006, Choudhary, Kumar et al. 2009, Zhao, Xu et al. 2010). Lysine acetylation of proteins in the mitochondria has been shown to be regulated by the mitochondrial deacylase, SIRT3 (Rardin, Newman et al. 2013).

Although much work has been done in recent years identifying the targets of SIRT3, most of these targets were confirmed in vitro in immortalised cell lines, or in vivo in mouse liver, with many studies taking advantage of liver tissue from the global SIRT3 knockout (SIRT3 KO) mouse model (summary of validation tissues can be seen in Table 1.1). Surprisingly, of the validated SIRT3 targets in the literature, only relatively few have been demonstrated in skeletal muscle models, and of these studies the majority were in immortalised muscle cells such as C2C12 cells (Osborne, Cooney et al. 2014). Because the interaction of SIRT3 with its substrates may be tissue and condition

47 CHAPTER 3: Muscle Overexpression

specific, there is an obvious need for more studies on the role of SIRT3 specifically in skeletal muscle.

Despite the overall lack of focus on muscle, there are some published studies that examined SIRT3 function in this tissue. SIRT3 as a mitochondrial protein, has been shown to be more highly expressed in oxidative muscle such as heart and red muscle compared to white muscle fibres (Gurd, Holloway et al. 2012). SIRT3 was also shown to be located in both the subsarcolemmal and intermyofibrillar mitochondrial subpopulations, and is increased following exercise stimulation (Palacios, Carmona et al. 2009, Hokari, Kawasaki et al. 2010).

Previous work in our laboratory has studied the ob/ob mouse model of leptin deficiency, which is associated with uncontrolled feeding and a grossly obese phenotype associated with hyperglycaemia. These unpublished studies looked at SIRT3 levels and lysine acetylation in skeletal muscle from ob/ob mice. The leptin deficient mice were found to have significantly lower SIRT3 mRNA levels in skeletal muscle compared to wildtype littermates, and marked hyperacetylation of muscle mitochondrial proteins was observed by western blot (Lauren E. Wright, PhD thesis, 2011). Similarly, in the diabetic mouse model of leptin receptor deficiency, the db/db mouse, SIRT3 mRNA levels were also found to be significantly decreased (unpublished). In human skeletal muscle, SIRT3 mRNA levels have also been reported to decrease with obesity in a study in pregnant women (Boyle, Newsom et al. 2013), and decrease in the elderly (Joseph, Adhihetty et al. 2012).

Another relevant body of work includes studies investigating the role of SIRT3 in cardiomyocytes. Gupta and colleagues have shown that SIRT3 is involved in mitochondrial fusion via deacetylation of OPA1 (Samant, Zhang et al. 2014), in protection from stress-mediated cell death (Sundaresan, Samant et al. 2008), and in cardiac hypertrophy via antioxidant defence mechanisms that protect from ROS (Sundaresan, Gupta et al. 2009). SIRT3 has also been implicated in ageing and cardiac hypertrophy in cardiomyocytes via deacetylation of cyclophillin D and action on the mitochondrial permeability transition pore (mPTP) (Hafner, Dai et al. 2010).

48 CHAPTER 3: Muscle Overexpression

A recent study in muscle of SIRT3KO mice showed that the deletion of SIRT3 in this tissue leads to metabolic inflexibility (Jing, O'Neill et al. 2013). Metabolic flexibility is the ability of muscle to change fuel oxidation to match fuel availability (Muoio, Noland et al. 2012), with metabolic inflexibility first described in obese and diabetic humans who fail to shift from fatty acid to glucose oxidation when glucose again becomes available following a meal (Kelley and Mandarino 2000). SIRT3 deletion was found to lead to increased acetylation of the E1α subunit of pyruvate dehydrogenase (PDH), ultimately decreasing activity of the entire multi-complex enzyme (Jing, O'Neill et al. 2013). This decrease in PDH activity in muscle reduces glucose oxidation and results in a switch to fatty acid oxidation, even when glucose is available.

While the exact role that mitochondria play in the development of insulin resistance is still unclear, it is possible that strategies designed to enhance mitochondrial metabolism by increasing SIRT3 may have beneficial effects on insulin action, and may mitigate some of the detrimental effects of nutrient excess. In this Chapter the aim was to (1) overexpress SIRT3 in the skeletal muscle of rats and investigate the effect on mitochondrial metabolism and (2) determine if this overexpression has a beneficial effect on insulin action in both chow and high fat fed rats.

49 CHAPTER 3: Muscle Overexpression

3.2 Methods

3.2.1 Animals and diet composition

Animal experiments in this chapter use male Wistar rats obtained at 100 – 150 g from Animal Resource Centre (Perth, Australia). All animals were housed under standard laboratory conditions at 22 ± 0.5 °C on a 12:12-hr light-dark cycle, with free access to water. Rats were fed ad libitum for 3-4 weeks prior to the experiment day on either a standard chow (8% calories from fat, 21% calories from protein, 71% calories from carbohydrate; Gordon’s Specialty Stock Feeds, NSW, Australia), or high-fat diet (HFD; 45% calories from fat (lard), 20% calories from protein, 35% calories from carbohydrates, 4.7 kcal/g; based on Rodent Diet #D12451 Research Diets, Inc., NJ, USA). Rats were euthanased with an overdose of pentobarbital sodium (Nembutal, Abbott Laboratories, Sydney, Australia). All experimental procedures were approved by the Garvan Institute/St. Vincent’s Hospital Animal Experimentation Ethics Committee and were in accordance with the National Health and Medical Research Council of Australia Guidelines on Animal Experimentation.

3.2.2 Generation and propagation of SIRT3-FLAG-AAV

Human SIRT3-FLAG.pcDNA3.1+ (North, Marshall et al. 2003) was obtained from Eric Verdin via Addgene (Addgene plasmid 13814, Cambridge, MA). SIRT3-FLAG was cloned into the pENN.AAV.tMCK.eGFP.wpre.bgh plasmid by the University of Pennsylvania Vector Core Facility (Philadelphia, PA, USA). This plasmid allows the expression of the construct under the control of the muscle-specific muscle creatine kinase (MCK) promoter. 3.621 x1013 GC (genome copies) of AAV2/9.tMCK.PI.hSIRT3-Flag.SV40 was prepared by the University of Pennsylvania Vector Core. GFP expressing AAV2/9.tMCK.PI.eGFP.WPRE.bGH was purchased from University of Pennsylvania Vector Core as a serotype and promotor-specific control (Catalog no.V1969). Viruses will hereafter be referred to as SIRT3-AAV and GFP-AAV in this document.

50 CHAPTER 3: Muscle Overexpression

3.2.3 SIRT3 overexpression in skeletal muscle using SIRT3-AAV

SIRT3 overexpression was induced in rat skeletal muscle via intramuscular injection. Rats were anaesthetized using 3% isoflourane in oxygen, and hindlimbs were shaved and swabbed with alcohol wipes. Both left and right-hand-side tibialis cranialis (TC) and extensor digitorum longus (EDL) muscles received a total of 70 units of hyaluronidase as a pre-treatment to break down hyaluronic acid and increase muscle fibre permeability (Favre, Cherel et al. 2000, McMahon, Signori et al. 2001). One hour later, the right-hand-side TC received 1x1011 GC of SIRT3-AAV, while the right EDL received 0.5x1011 GC SIRT3-AAV, diluted in 0.9% sterile saline. The contralateral leg was used as an internal control and received saline or GFP AAV (same dose as for SIRT3-AAV). Rats were then placed on either a chow or HFD for 3 – 4 weeks. For studies in isolated mitochondria and immunoblotting, rats were sacrificed and TC and EDL excised. EDL and a portion of the TC were snap-frozen in liquid nitrogen, while the remainder of the TC was used to isolate mitochondria. For animals undergoing hyperinsulinaemic-euglycaemic clamp, rats were injected with AAV, then placed on chow or HFD for 3-4 weeks, and underwent surgery one week prior to the clamp experiment.

Muscle-specific in vivo overexpression has a number of advantages over the generation of transgenic and knockout animal models: i) the gene of interest is able to be overexpressed acutely in the adult animal, eliminating the impact of developmental effects, ii) the construct is selectively introduced into one hindlimb of the animal, allowing for targeting of specific muscle groups, and iii) the contralateral limb provides an in vivo control that is exposed to the same circulating factors and conditions as the test muscles (Cleasby, Davey et al. 2005).

3.2.4 Assessment of oxidative capacity in isolated mitochondria

Mitochondria were prepared from fresh skeletal muscle using differential centrifugation following the procedure outlined in detail in Chapter 2. Oxygen

51 CHAPTER 3: Muscle Overexpression

consumption under different substrate conditions was measured using a Clark electrode as described in Section 2.4.2. Data are presented as percentage increase in oxygen consumption between the right leg and the left control leg in each individual rat, measured as nanomoles of oxygen consumed per minute, per mg of mitochondrial protein.

3.2.5 Analysis of insulin sensitivity by hyperinsulinemic-euglycemic clamp

All hyperinsulinemic-euglycemic clamp procedures were conducted with the help of Dr. Amanda Brandon at the Garvan Institute of Medical Research according to our published methods (Cleasby, Lau et al. 2011, Wright, Brandon et al. 2011, Boden, Brandon et al. 2012).

3.2.5.1 Dual jugular cannulation All surgery procedures were conducted on heating pads using sterile surgical practices. Rats were anaesthetised with a mixture of 80 mg/kg ketamine/20 mg/kg xylazine injected intraperitoneally. Once anaesthetised, the surgical site was shaved, sterilised with alcohol/chloramphenicol and a small incision made with a scalpel. Superficial tissue was blunt dissected to expose the jugular veins, into which silastic cannula (internal diameter 0.64 mm, filled with 10 U/ml heparinised saline) were inserted and sutured in place. Cannulae were tunnelled subcutaneously and exteriorised at the back of the neck and were sealed with 1.2 g/ml polyvinylpyrrolidine in heparinised saline. Animals received the analgesic/anti-inflammatory ketuprofen (5 mg/kg) and local injection of bupivicaine (4 mg/kg) at the suture site at the end of the surgery. Animals were used in the hyperinsulinemic-euglycemic clamp one week after surgery. All rats used in clamp procedures had regained any weight lost after surgery and in most cases had gained extra weight.

3.2.5.2 Hyperinsulinemic-euglycemic clamp On the day of experimentation, conscious rats with dual jugular cannulae were fasted for approximately 5 hr, and fitted with extension cannulae to enable infusion and

52 CHAPTER 3: Muscle Overexpression

sampling. Rats were infused with insulin at a rate of 0.25 U/kg/hr, and glucose (30%) was also infused at a variable rate to maintain animals at euglycemia. Once a steady rate of glucose infusion was reached, a bolus injection of 50 μCi 3H-2-deoxy-D-[2,6]- glucose (3H-2DG) (GE Healthcare Life Sciences, Buckinghamshire, UK) and 22.5 μCi 14C- Glucose was administered, and blood samples were taken at 2, 5, 10, 15, 20, 30, and 45 min after the tracer bolus entered the bloodstream. At the conclusion of the tracer period, rats were euthanased and the TC and EDL muscles rapidly dissected and freeze clamped. A small number of rats were injected with tracer but not submitted to the clamp procedure, providing a control group of ‘tracer basal’ tissues for insulin signalling assessment by western blot.

3.2.5.3 Blood analysis Plasma glucose was determined using an automated glucose analyser (YSI, Yellow Springs, OH, USA). Blood was transferred into tubes containing EDTA, centrifuged and the plasma stored at -80 °C for later analysis. Plasma NEFA and triglyceride during clamp were measured as described in Chapter 2. Insulin levels were assessed using a sensitive rat/mouse insulin enzyme-linked radioimmunoassay (Linco, MO, USA) according to the manufacturers instructions with minor modification. The kit uses 125I- labelled insulin and a competitive binding principle to determine the level of insulin in the serum. Briefly, serum samples are diluted in assay buffer (0.05 M Phospho-saline, 0.025 M EDTA, 0.08% sodium azide, 1% RIA grade BSA, pH 7.4) and incubated with 50 μl 125I-Insulin, and 50μl anti-rat Insulin serum in assay buffer for 20-24hr at 4 °C. In this assay, the labelled tracer and the unlabelled serum antigen compete for the limited number of antibody binding sites. The antibody-bound insulin is isolated by precipitation and centrifugation (40 min, 3000 g, 4°C), and radioactivity counted using an automatic gamma counter (Perkin Elmer, CT, USA). The concentration of serum insulin is calculated from the sample counts and compared to the provided insulin standard curve (0-10 ng/ml).

53 CHAPTER 3: Muscle Overexpression

3.2.5.4 Tracer dose The dose of tracer injected into each animal was accurately determined from the before and after weight of the syringe. Total tracer activity was determined by counting 100 μl of a 1:10000 dilution of tracer, after the addition of scintillation fluid (Ultima Gold XR, Packard Biosciences, Groningen, The Netherlands) and measuring radioactivity in each sample using a liquid scintillation counter (Beckman LS6000, Beckman Instruments, CA, USA). All calculations involving tracers were completed using Microsoft Excel (Microsoft Corporation, USA).

3.2.5.5 Rate of tracer disappearance from plasma (Rd) During the tracer period, at each timepoint 50 μl of plasma was taken and deproteinised by mixing with 1 ml 2.75% ZnSO4 and 350μl saturated BaOH. To determine Rd the deproteinised sample was centrifuged (3000 g, 10 min, room temperature) and the amount of 3H and 14C present in 500 μl of the supernatant was determined by the addition of scintillation fluid and counting the activity using a liquid scintillation counter (Beckman LS6000, Beckman Instruments, CA, USA). To calculate the Rd (in mg/kg/min) disappearance curves for plasma 3H and 14C, the measured reduction in plasma radioactivity was fitted to a double exponential equation, which was integrated to determine the area under the curve to the final time point (45 min) and to estimate the area to infinity:

!".!"#$ Rd = ! ! !"∗ ! !"

In this equation, Cp is the plasma glucose concentration (mM), dose is the dpm of ! Cp ∗ t dt tracer administered, ! is the area under the tracer disappearance curve to infinity.

3.2.5.6 Rate of glucose uptake (Rg’) in skeletal muscle Glucose uptake into TC and EDL muscles of both left and right leg was calculated by measuring the accumulation of 3H-2DG -6-phosphate in skeletal muscle over the tracer

54 CHAPTER 3: Muscle Overexpression

period. 2DG is taken into cells and phosphorylated in a similar way as glucose, however 2DG-6-phosphate cannot be metabolised substantially (whereas glucose-6-phosphate can) and therefore the tracer accumulates in the tissue. Calculating Rg’ (in μmol/min/100g) utilises the accumulation of 3H-2DG-6-phosphate, Rd’ and plasma glucose levels in the equation:

! Cp. Cm* Rg = !" Cp ! *(t)dt

In this equation, Cp is the plasma glucose concentration (mM), Cm* is the accumulation of 3H-2DG-6-phosphate (dpm/100g tissue), Cp*(t) is the plasma 3H-2DG concentration (dpm/ml), and t = 0 represents the start of the tracer period.

To determine Rg’ from clamped animals and tracer basal animals, ~40 mg of powdered tibialis, or ~30 mg of EDL was weighed and homogenised in 1.5 ml dH2O using a polytron (Kinematica AG, Lucerne, Switzerland). The homogenate was centrifuged (3000 g, 10 min, 4 °C) to pellet insoluble debris, and total 3H counts within a sample of supernatant was determined using a liquid scintillation counter (Beckman LS6000, Beckman Instruments). Phosphorylated 3H-2DG (radiation within muscle fibres) was separated from unphosphorylated 3H-2DG (extracellular) by running 400 μl of supernatant through an ion exchange column (AG 1-X8 Resin, Bio-Rad Laboratories, CA, USA), 6 ml of dH2O was used to wash the column, and 2 ml of the flow through was mixed with scintillation fluid to measure the amount of unphosphorylated glucose counts using a liquid scintillation counter (Beckman LS6000, Beckman Instruments). Rg’ is calculated by subtracting the unphosphorylated 3H-2DG counts (flow through) from the total counts.

3.2.5.7 Incorporation of tracer into glycogen Glycogen was extracted from frozen muscle tissue and the final pellet resuspended in 1ml of amyloglucosidase buffer as described in Chapter 2. The incorporation of tracer into glycogen was assessed by adding 500 µl of glycogen sample to 5 ml of scintillation fluid and 14C content of glycogen assessed using a liquid scintillation counter (Beckman

55 CHAPTER 3: Muscle Overexpression

LS6000, Beckman Instruments). Incorporation into glycogen during the clamp was calculated from the glucose tracer disappearance curve and counts of 14C in the extracted muscle glycogen.

3.2.6 Statistical analysis

All results are presented as mean ± SEM. Results between chow and HFD animals were compared using unpaired t-test. Results for effect of SIRT3 on an individual muscle were compared using two-way ANOVA for effects of SIRT3 overexpression and diet, with Tukey post-hoc test. Where appropriate a paired t-test was also used to compare contralateral legs. Statistical analysis was performed in GraphPad Prism software (Prism 6, Version 6.0b, Oct 3, 2012).

56 CHAPTER 3: Muscle Overexpression

3.3 Results

3.3.1 Overexpression of SIRT3 protein in rat skeletal muscle

SIRT3 protein expression was increased ~11-fold in tibialis (TC) and ~3-fold in EDL muscles at 3-4 weeks (Table 3.1). The 45kDa SIRT3-FLAG was processed correctly into the catalytically active 28 kDa mitochondrial form of the protein, with the exogenous SIRT3 visible in Figure 3.1 clearly seen due to the size shift induced by the addition of a FLAG tag. SIRT3 AAV did not significantly affect total mitochondrial content as assessed by the mitochondrial marker porin, or by amounts of some subunits of mitochondrial OXPHOS complexes as assessed by western blot (Figure 3.1A).

Over-expression was assessed using the cross-species antibody to SIRT3 which detects both the introduced human, and endogenous mouse and rat species with high affinity (hereafter referred to as h/mSIRT3; Cell Signalling mAb D22A3). The human specific SIRT3 antibody (referred to in text as hSIRT3;Cell Signalling mAb C73E3) is reported by the manufacturer to also detect the rat protein. However, as can be seen in Figure 3.1, endogenous SIRT3 has a stronger band using the h/mSIRT3 antibody and therefore this antibody was used for analysis of protein overexpression by densitometry.

Table 3.1 Mean overexpression of SIRT3 protein with intramuscular injection of SIRT3 AAV in tibialis cranialis (TC) and extensor digitorum longus (EDL) assessed by western blot

Mean Increase in SIRT3 Protein n TC 10.9 ± 1.3 23 EDL 3.4 ± 0.3 10 Data are ± SEM

57 CHAPTER 3: Muscle Overexpression

A Tibialis L R L R L R L R SIRT3-FLAG ~30 kDa Ab: h/mSIRT3 Endogenous 28 kDa

hSIRT3 ~28 kDa CV 55 kDa CIII 48 kDa CII 30 kDa CI 20 kDa

Porin/VDAC 30 kDa

60000 5 15 LTC 4 RTC 40000 3 10 (A.U) (A.U) 2 20000 5

SIRT3 SIRT3 Protein 1 0 0 0 LTC RTC Mito. Complex Protein CI CII CIII CV Porin

B EDL L R L R L R L R

h/mSIRT3 ~28 kDa

hSIRT3 ~28 kDa Skeletal Actin 45 kDa

25000 20000 15000

(A.U) 10000

SIRT3 SIRT3 Protein 5000 0 LEDL REDL

Figure 3.1 SIRT3 overexpression in rat skeletal muscle using SIRT3 AAV: A. Representative blot of whole muscle lysates from control (left leg-L) and SIRT3 overexpression (right leg-R) tibialis muscle showing both the ~30 kDa SIRT3-FLAG and endogenous SIRT3 at 28 kDa. Lower panel shows densitometric analysis for SIRT3 protein and OXPHOS complexes in left (blue bars) and right (red bars) tibialis cranialis muscle (LTC/RTC). B. Representative blot of SIRT3 overexpression targeting the EDL muscle in rats from control (left leg-L) and SIRT3 overexpression (right leg-R) muscle. Lower panel shows densitometric analysis for SIRT3 protein in left (blue bars) and right (red bars) EDL muscle (LEDL/REDL). Antibodies used were h/mSIRT3 and hSIRT3, OXPHOS antibody cocktail to complexes I-III, Complex V and the mitochondrial marker, porin/VDAC. Data are mean ± SEM, n=23 for TC, n= 10 for EDL.

58 CHAPTER 3: Muscle Overexpression

3.3.2 Effect of SIRT3 overexpression on total pan-lysine acetylation

Section 3.3.1 shows effective overexpression of SIRT3 protein following intramuscular injection of SIRT3 AAV, so next we investigated if this overexpression had a functional effect on lysine acetylation in skeletal muscle. SIRT3 overexpression would be predicted to lower the level of acetylation and there was a trend towards a reduction in global acetylation in whole muscle lysates overexpressing SIRT3 compared to control lysates, although this difference did not reach statistical significance (Figure 3.2).

A Leg L R L R L R L R L R L R B hSIRT3 28 kDa 100000 150 kDa 80000

60000 50 kDa 40000

20000 25 kDa Pan Acetylation Pan (A.U) Acetylation Pan-Acetylysine Ab Pan-Acetylysine 0 20 kDa LTC RTC

Figure 3.2 SIRT3 overexpression and effect on global lysine acetylation: A: Western Blot of whole muscle lysates from control (left leg-L) and SIRT3 overexpression (right leg-R) tibialis muscle probed with hSIRT3 to show overexpression, and with a pan- acetyllysine (AcK) antibody. B. Densitometry across entire individual lanes shows a reduction in lysine acetylation in TC muscles that are overexpressing SIRT3 compared to control leg. LTC & RTC denote left (control) and right (SIRT3 overexpressing) tibialis cranialis muscle. Data are mean ± SEM, n=6, p=0.08 for paired t-test.

3.3.3 Effect of SIRT3 overexpression in rat skeletal muscle on oxidative capacity in isolated mitochondria

To investigate the effect of overexpression of SIRT3 on the oxidative capacity of muscle, the ability of isolated mitochondria to oxidise a variety of substrates was examined. The functional capacity of isolated mitochondria to oxidise succinate, glutamate, and palmitoyl carnitine was increased following acute SIRT3 overexpression

59 CHAPTER 3: Muscle Overexpression

in muscle for 3 weeks, in both chow and HFD-fed rats (Figure 3.3). Although ADP- stimulated or state III respiration showed a large variation from animal to animal in oxygen consumption measures, the use of mitochondria isolated from the contralateral leg as an internal control enables percentage increase as a paired comparison to assess the effect of SIRT3 overexpression. Figure 3.3 shows that the percentage increase in oxygen consumption in isolated mitochondria using succinate, glutamate and palmitoyl carnitine showed a significant increase with SIRT3 overexpression by 2-way ANOVA for both diets, with the magnitude of change approximately 25% for all substrates. Oxygen consumption was also assessed using the substrate pyruvate however there was no significant change with SIRT3 overexpression.

A B CON (Left leg) +SIRT3 (Right leg) ‡ ‡ 150 150

100 100

50 50 CONSUMPTION CONSUMPTION 2 2 (SIRT3 AAV/CON) (SIRT3 AAV/CON) O 0 O 0 CHOW HFD CHOW HFD Succinate Glutamate

C D

150 150 ‡

100 100

50 50 CONSUMPTION CONSUMPTION 2 2 (SIRT3 AAV/CON) O (SIRT3 AAV/CON) 0 O 0 CHOW HFD CHOW HFD Pyruvate Palmitoyl Carnitine

Figure 3.3 Percentage change in ADP-stimulated oxygen consumption in isolated mitochondria with SIRT3 AAV treatment in rats on chow and HFD: ADP-Stimulated respiration shown as percentage increase in oxygen consumption from CON leg (blue bars) compared to the SIRT3 overexpressing right leg (red bars) under both chow and HFD-fed conditions with A: Succinate, B: Glutamate, C: Pyruvate, and D: Palmitoyl carnitine as the substrate. Data are mean ± SEM in nmol of oxygen per minute, per mg of protein. ‡ p<0.05, effect of SIRT3 only by 2-way ANOVA. n = 3-4 animals per group.

60 CHAPTER 3: Muscle Overexpression

3.3.4 The effect of 4 weeks of HFD on whole body insulin action in rats.

To assess a potential role for SIRT3 in regulating substrate metabolism and insulin action in muscle, rats fed a normal chow or a HFD with overexpression of SIRT3 in the right tibialis muscle were subjected to a hyperinsulinemic-euglycemic clamp.

The effect of HFD on the whole body-parameters of the rats that underwent hyperinsulinaemic-euglycamic clamp can be seen in Table 3.3. Due to the excess fat deposits in HFD-fed rats, there is additional difficulty in both inserting and maintaining the patency of cannulae in these animals following surgery, hence as can be seen in Table 3.3, we successfully clamped 10 chow-fed animals, but only 5 HFD-fed animals from this cohort. Although 4 weeks of high fat feeding did not produce a significant change in body weight of the animals, there was an increase in the mass of the epididymal and inguinal fat pads with HFD (Table 3.3). There was also an effect of HFD on fasting insulin levels prior to clamp, and blood glucose at basal was also slightly elevated in HFD, providing evidence that insulin resistance was present following 4 weeks of HFD. During the clamp procedure insulin concentration was elevated to a similar level in both groups, and plasma glucose was clamped within the 7-8mM range. While there was suppression of NEFAs and plasma triglyceride under clamp conditions, only plasma triglyceride during the clamp showed a significant difference following HFD.

61 CHAPTER 3: Muscle Overexpression

Table 3.2 Characteristics of rats undergoing hyperinsulinaemic-euglycaemic clamp CHOW HFD P VALUES

Number of animals 10 5 -

Body mass (g) 335±13.6 351±4.6 0.4

Epididymal fat (g) 2.6±0.4 4.2±0.3 0.01*

Inguinal fat (g) 4.4±0.5 6.8±0.4 0.01*

Fasting plasma insulin (mU/L) 40.1 ± 3.5 67.8 ± 6.3 0.001* p<0.0001† Insulin during Clamp (mU/L) 143.6 ± 12.2 167.6 ± 5.5 0.3

Plasma glucose during Clamp (mM) 7.0±0.2 7.8±0.4 0.05

Plasma Glucose at Basal (mM) 7.7±0.3 8.5 ±0.5 0.09

Basal NEFA (mM) 0.67±0.1 0.64±0.1 0.9 p=0.0003† NEFA during Clamp (mM) 0.22±0.04 0.19±0.04 0.7

Basal Plasma TAG (mM) 1.45±0.08 1.48±0.14 0.9 p=0.001† Clamp Plasma TAG (mM) 1.06±0.04 1.25±0.08 0.04* TAG=triglyceride, NEFA=non-esterified fatty acids. Data are ±SEM *Significant p<0.05 effect of diet by unpaired t-test of chow vs. HFD animals. †Significant effect of insulin during clamp by 2-way ANOVA.

During the hyperinsulinemic-euglycemic clamp, HFD-fed rats had a reduced glucose infusion rate (GIR) and a lower disappearance of glucose from the blood over time (Rd) (Figure 3.4). These results clearly show that whole body insulin resistance occurred in response to high fat feeding for 4 weeks.

62 CHAPTER 3: Muscle Overexpression

40 40 * 30 * 30

20 20

10 10 Rd (mg/kg/min) GIR (mg/kg/min) GIR

0 0 CHOW HFD CHOW HFD

Figure 3.4 Reduced GIR and Rd with HFD indicate systemic insulin resistance in clamped rats: 4 week HFD-fed rats (black bars) show inability to respond to insulin during the clamp to the same extent as chow fed rats (grey bars). Data are mean ±SEM *p<0.01, unpaired t-test. Chow n = 10, HFD n = 5.

3.3.5 The effect of 4 weeks overexpression of SIRT3 on insulin action in muscle of chow and high fat diet-fed rats

To determine if SIRT3 overexpression in muscle of chow and HFD-fed rats altered insulin action, we examined in vivo glucose uptake into both TC and EDL muscle. Under clamp conditions, glucose uptake into control (LTC) muscle of HFD rats was 30% lower than observed in control muscle of chow-fed rats (Figure 3.5). Despite the overexpression of SIRT3 in the right TC muscle of HFD-fed rats, glucose uptake into this muscle was also 28% lower than insulin * p< 0.01, unpaired-stimulated glucose uptake in the c ttest omparative muscle of chow-fed rats. Glucose uptake into the EDL muscle was 41% lower in the LTC and 22% lower in the RTC of HFD-fed rats compared to chow-fed rats, with no significant effect observed for SIRT3 overexpression (Figure 3.5).

63 CHAPTER 3: Muscle Overexpression

CON (Left leg) +SIRT3 (Right leg)

† † 25 15

20 10 15

10 5 (umol/100g.min) 5 (umol/100g.min) Rg' / Glucose Uptake Rg' / Glucose Uptake 0 0 CHOW HFD CHOW HFD TC EDL

Figure 3.5 HFD reduces glucose uptake into muscle (Rg’) under clamp conditions which is not ameliorated by SIRT3 overexpression: Effect of HFD on labeled glucose * p< 0.02, sidak uptake into both the TC post test 2-way anova, (left panel) effect and EDL of HFD,(right panel) muscle. Control muscles from left leg (blue bars) are compared to SIRT3 AAV treated muscle † p <0.005 2 way anova diet effect from right leg (red bars). # p = Data are m0.03 diet onlyean ±SEM. †p<0.005 effect of diet only by 2 2-way ANOVA sidak post test -way ANOVA. Effect of SIRT3 was not significant. Chow n = 10 and HFD n= 5 rats per group.

3.3.6 Effect of SIRT3 overexpression on skeletal muscle glycogen and triglyceride content after 4 weeks HFD

SIRT3 has been shown to have effects on fatty acid oxidation and fuel oxidation (Osborne, Cooney et al. 2014), therefore we also investigated the effect of overexpression of SIRT3 on intramuscular triglyceride accumulation and glycogen stores during high-fat feeding in rats that underwent the clamp. HFD significantly increased intramuscular triglyceride in the TC from 3.4 ± 0.08 in chow fed animals to 4.8 ± 0.4 in HFD-fed rats (Figure 3.6A). However, targeted SIRT3 overexpression in TC had no effect on the accumulation of triglyceride content with both legs equally impaired. Total glycogen content of the muscle was unchanged by dietary intervention or SIRT3 overexpression (Figure 3.6B). Glucose incorporation into glycogen in TC muscle during the tracer period is a measure of glycogen synthesis in response to insulin infusion. Glucose incorporation into glycogen was significantly reduced in muscles of HFD-fed rats (Figure 3.6C) indicating insulin resistance, however there was no affect of SIRT3 overexpression. These results indicate that acute overexpression of

64 CHAPTER 3: Muscle Overexpression

SIRT3 in rodent muscle in vivo does not protect against lipid accumulation or glucose disposal defects induced by HFD.

A B CON (Left leg) +SIRT3 (Right leg) † 6 80

60 4 40 (nmol/mg) (nmol/mg) (nmol/mg) 2 Triglyceride 20 Glycogen Content Glycogen

0 0 CHOW HFD CHOW HFD

10 # 8

(umol/100g.min) 2

0 Glucose incorp. into Glycogen into incorp. Glucose CHOW HFD

Figure 3.6 SIRT3 overexpression has no effect on muscle triglyceride, glycogen content, or glycogen synthesis in chow and HFD-fed rat muscle: A. Muscle triglyceride increases equally with HFD in tibialis muscle from control (blue bar) and SIRT3 overexpressing (red bar) TC muscle compared to chow-fed rats. B. Glycogen content is unchanged with SIRT3 overexpression in the right TC (red bars) compared to the control TC (blue bars), and with HFD-feeding. C. Glucose incorporation into glycogen in TC muscle during clamp is reduced with HFD compared to chow-fed rats, however there is no effect of SIRT3 overexpression in the right TC (red bars) compared to the control TC (blue bars). Data are mean ±SEM. †p=0.002, #p=0.02, main effect of diet by 2-way ANOVA. Effect of SIRT3 was not significant. Chow n = 10 and HFD n= 5 rats per group.

65 CHAPTER 3: Muscle Overexpression

3.3.7 Insulin, but not SIRT3 or HFD, impacts on insulin signalling pathways

In addition to the measures of tracer uptake and metabolism, we also investigated whether the insulin resistance seen in vivo in rats was associated with changes in muscle insulin signalling in rats that underwent hyperinsulinaemic-euglycaemic clamp. SIRT3 overexpression was confirmed in these animals by western blot, and similar to previous cohorts infected with SIRT3 AAV, the animals used in the clamp procedure showed a 9-10 fold increase in SIRT3 protein in the right TC compared to the control TC (Figure 3.7A & B). There was no significant effect of HFD on the level of endogenous or overexpressed SIRT3 using t-test or 2-way ANOVA.

Insulin resistance in muscle of HFD–fed rats is frequently associated with defects in insulin-stimulated Akt phosphorylation and its downstream effects such as GLUT4 trafficking (Kim, Nikoulina et al. 1999, Frangioudakis, Ye et al. 2005). TC muscle from animals that underwent the clamp procedure and basal controls (animals that received tracer but not insulin or glucose infusion) were assessed for phosphorylation of Ser473 of Akt and Thr642 of AS160 (Figure 3.7). Western blot analysis showed a significant increase in phosphorylation of these residues of Akt and AS160 in insulin stimulated (clamped) TC muscle compared to basal TC muscle. Interestingly in post-clamp tissues, HFD for 4 weeks had no effect on Akt and AS160 phosphorylation in muscle despite the evidence of whole body and muscle-specific insulin resistance. SIRT3 overexpression in the right TC had no effect on the degree of Akt or AS160 phosphorylation under clamp conditions in either chow-fed or HFD-fed rats.

Collectively, the studies in this chapter show that while muscle specific SIRT3 overexpression is sufficient to increase some parameters of mitochondrial metabolism ex vivo, SIRT3 overepxression did not lead to measurable changes in insulin action in skeletal muscle in vivo, assessed using the hyperinsulinemic/euglycemic clamp in control and HFD-fed insulin resistant rats.

66 CHAPTER 3: Muscle Overexpression

CHOW CHOW HFD HFD A BASAL CLAMP BASAL CLAMP

L R L R L R L R L R L R L R L R L R L R L R SIRT3 skActin

pAkt

tAkt

pAS160

tAS160

CON (Left leg) +SIRT3 (Right leg) B 50000 *** ***

25000 (A.U.)

SIRT3 SIRT3 Protein 0 CHOW HFD

C D ‡ # 1.5 1.0

1.0 0.5 0.5 (p/tAkt A.U.) (p/tAS160 (p/tAS160 A.U.)

Akt Phosphorylation 0.0 0.0 CHOW HFD CHOW HFD AS160 Phosphorylation CHOW HFD CHOW HFD BASAL CLAMP BASAL CLAMP

Figure 3.7 Effect of clamp and diet on SIRT3 protein, and insulin signaling: A. Representative blot from whole muscle lysate from clamped animals shows SIRT3 protein levels, Akt phosphorylation, and AS160 phosphorylation. B. Densitometry of SIRT3 protein by western blot in chow and HFD-fed rats following clamp, n = 11-12 animals per group. C & D. Densitometry of phosphorylated Ser473 of Akt and Thr642 of AS160 from western blot in both basal controls and animals that underwent the clamp, with and without SIRT3 overexpression, n= 4 – 8 per group. Data are mean ± SEM. ***p<0.0001 for left TC vs. right TC by 2-way ANOVA with Tukey post-hoc test. #p<0.0001 main effect of clamp procedure cf. basal by 2-way ANOVA. ‡p=0.0001 main effect of clamp procedure cf. basal by 2-way ANOVA.

67 CHAPTER 3: Muscle Overexpression

3.4 Discussion

SIRT3 has been reported to play an important role in the regulation of mitochondrial metabolism. Reduced mitochondrial capacity and/or activity has been proposed as a mechanism underlying the development of obesity related insulin resistance in muscle (Lowell and Shulman 2005, Turner and Heilbronn 2008). Despite this, there are few studies investigating a possible beneficial role of SIRT3 in the prevention of mitochondrial dysfunction and insulin resistance in skeletal muscle. This study aimed to be the first to assess insulin action using a hyperinsulinaemic-euglycaemic clamp with muscle specific overexpression of the mitochondrial deacetylase SIRT3. The results herein demonstrate successful overexpression of SIRT3 in the skeletal muscle of rats using a muscle-specific AAV vector, and that this overexpression was unaffected by HFD. The overexpression of SIRT3 in skeletal muscle resulted in a small decrease in global lysine acetylation, and was sufficient to increase mitochondrial respiration in isolated mitochondria under a variety of conditions. However, in vivo data shows that in the context of insulin action at the level of the skeletal muscle, SIRT3 overexpression does not alter the insulin resistance in muscle caused by high fat feeding.

The data presented in this chapter shows that injection of the SIRT3 AAV induced significant expression of human SIRT3-FLAG at the protein level in both the tibialis cranialis (TC) and extensor digitorum longus (EDL) muscle. In our model, SIRT3 overexpression did not affect mitochondrial content as assessed by porin/VDAC levels or protein subunits of the electron transport chain (ETC) (Figure 3.1), however more detailed examination of mitochondrial content using other methods such as high- resolution imaging was not undertaken. SIRT3 overexpression in vitro in brown adipose cells and in cortical neuron cultures has been reported to increase mitochondrial biogenesis (Shi, Wang et al. 2005, Dai, Chen et al. 2014), however no evidence exists for the effects of increasing SIRT3 protein on mitochondrial content in vivo in muscle. While changes to other subunits, or the assembly of the complexes of the OXPHOS system remain possible, this data suggests that changes in the proteins of the ETC are unlikely to be responsible for any changes in mitochondrial respiration observed after 4 weeks of increased SIRT3 expression. Conversely, recent studies suggest that SIRT3

68 CHAPTER 3: Muscle Overexpression

may also have an inhibitory effect on mitochondrial biogenesis at least in skeletal muscle via its action on mitochondrial ribosomal protein L10 (MRPL10) to suppress mitochondrial protein synthesis, indeed SIRT3KO muscle cells were shown to have increased protein synthesis (Yang, Cimen et al. 2010, White and Schenk 2012), however no corresponding decrease in mitochondrial content was observed with SIRT3 overexpression in muscle in this study.

SIRT3 was significantly overexpressed in both the TC and the EDL, by 10-fold and 3-fold respectively (Table 3.1). This potentially puts the level of SIRT3 expression seen in this model in both muscles into the supra-physiological range. For example, exercise training has been shown to increase SIRT3 levels in muscle by approximately 50% (Lanza, Short et al. 2008). Other studies showing upregulation of SIRT3 protein in liver and muscle during fasting (Palacios, Carmona et al. 2009, Hirschey, Shimazu et al. 2010), exercise training (Palacios, Carmona et al. 2009) and in muscle during fructose- feeding (Warren, Lou et al. 2014) have not quantified protein change. In one early paper, tissue distribution of SIRT3 showed expression to be relatively low in skeletal muscle compared to other tissues, although the exact muscle used was not noted (Lombard, Alt et al. 2007). Although the increase in SIRT3 protein is large relative to any documented changes observed so far in physiological situations, the usefulness of examining protein function by overexpression is no less valid than examining protein function in gene knockout models, whereby the protein is deleted completely and metabolic differences assessed.

SIRT3 overexpression appeared to reduce total lysine acetylation as measured by western blot, although the variability of measuring total acetylation by this methodology did not reach statistical significance (Figure 3.2). Measures of total acetylation can be problematic, as it is possible that other deacetylases, or acetyltransferases in the system are also modulating acetylation levels. Global SIRT3KO mice show marked hyperacetylation when assessed by western blot (Lombard, Alt et al. 2007), however the major mitochondrial deacetylase has been deleted from birth in these animals. One caveat of assessing acetylation by western blot is that (as discussed in Section 1.5.3) not all acetylated lysines on a given target protein are regulated by

69 CHAPTER 3: Muscle Overexpression

SIRT3 (Hirschey, Shimazu et al. 2010), and indeed may not have functional relevance at all. Unfortunately, commercial availability of functionally validated site-specific lysine acetylation antibodies, at least those that can detect changes in tissue lysates (as opposed to transfected cell culture lysates), is so far lacking. In this context, analysis by mass spectrometry techniques that can differentiate precise acetylation sites known to have a functional role is a superior technique, and will be discussed in detail in Chapter 5.

In the present study mitochondria from SIRT3-overexpressing rat TC muscles had higher rates of ADP-stimulated mitochondrial respiration than control muscle, supporting a role for SIRT3-mediated control of mitochondrial metabolism (Figure 3.3). SIRT3 has a growing list of validated targets, that include most important mitochondrial metabolic pathways, including the TCA cycle, urea cycle and fat oxidation (described in detail in Chapter 1). In addition SIRT3 directly deacetylates components of Complex I (Ahn, Kim et al. 2008), II (Finley, Haas et al. 2011), and III (Jing, Emanuelli et al. 2011) of the ETC. In this study, respiration was significantly increased compared to mitochondria from the control leg in chow and HFD-fed rats using succinate, glutamate and palmitoyl carnitine as substrates, with only pyruvate not showing a significant increase (Figure 3.3), despite the deacetylation of PDH E1α subunit reported as being a key component of SIRT3’s role in muscle (Jing, O'Neill et al. 2013).

This increase in fuel oxidation is in agreement with previous unpublished work in this laboratory showing increased respiration in muscles overexpressing SIRT3 via the in vivo electroporation method. Other in vitro studies also show an increase in mitochondrial respiration with SIRT3 overexpression, including in brown adipose cultures (Shi, Wang et al. 2005) and in HEK293T cells (Barbi de Moura, Uppala et al. 2014). The current report is the first report of SIRT3 overexpression in vivo in muscle increasing respiration in mitochondria isolated from that muscle.

The major finding of this chapter was that SIRT3 overexpression had no effect on insulin action in the skeletal muscle of normal or HFD insulin resistance rats. To date,

70 CHAPTER 3: Muscle Overexpression

no other animal studies investigating SIRT3 have looked at insulin action using the gold-standard technique, the hyperinsulinemic-euglycaemic clamp. Studies in both global SIRT3KO and muscle-specific SIRT3KO mice have used GTT and ITT to look at glucose metabolism and insulin sensitivity (Hirschey, Shimazu et al. 2011, Fernandez- Marcos, Jeninga et al. 2012). Schenk et al used the clamp technique to show that muscle specific deletion of SIRT1 abolished the increase in insulin sensitivity (increased GIR and increased uptake into muscle) seen in calorie restricted mice (Schenk, McCurdy et al. 2011), however no similar experiments have been done in SIRT3KO models.

Data reported in this Chapter shows that 4 weeks of high fat feeding caused an increase in the mass of the epididymal and inguinal fat pads and fasting hyperinsulinemia. As expected under insulin stimulation, serum NEFAs and triglycerides were suppressed during the clamp, although triglycerides were less suppressed in the HFD group (Table 3.3). Systemic insulin sensitivity as measured by GIR and Rd clearly showed that HFD for 4 weeks produced whole body insulin resistance (Figure 3.4), in line with previous studies in our laboratory (Cleasby, Davey et al. 2005, Bruce, Hoy et al. 2009, Kanzleiter, Preston et al. 2010, Wright, Brandon et al. 2011, Boden, Brandon et al. 2012).

The main purpose of these clamp studies was to look at peripheral insulin sensitivity specifically in the muscle overexpressing SIRT3, to see if SIRT3 (presumably by increasing mitochondrial oxidation) can mitigate any of the insulin resistance induced by HFD. Figures 3.5 and 3.6 clearly show that defects in HFD-fed animals such as decreased glucose uptake, increased intramuscular triglyceride, and reduction in glycogen synthesis was impaired equally in muscles over-expressing SIRT3 and contralateral control muscles. An enhanced capacity for substrate oxidation in skeletal muscle has been linked in many studies with improved insulin action (Bruce, Hoy et al. 2009, Benton, Holloway et al. 2010). In contrast to the ex vivo experiments performed for this thesis where enhanced substrate oxidation is seen under conditions of unlimited substrate availability and disruption of potential regulatory mechanisms,

71 CHAPTER 3: Muscle Overexpression

there appeared to be no impact of SIRT3 overexpression on insulin action in muscle in vivo under the clamp conditions.

Defects in the insulin signalling pathway are thought to be a key feature of insulin resistance in peripheral tissues. Analysis of phosphorylation of two key components of the insulin signalling pathway shows a significant increase in activation of Akt and its downstream target AS160 in insulin stimulated (clamped) TC muscle compared to basal TC muscle as expected. However this level of phosphorylation was not modulated in response to HFD or SIRT3 overexpression (Figure 3.7). The inability to detect changes in phosphorylation despite marked insulin resistance has been reported previously (Hoy, Bruce et al. 2007). It has been speculated that the insulin resistance seen in those animals may involve the inhibitory effects of replete glycogen stores in muscle (Nielsen, Derave et al. 2001), although other factors including GLUT4 and hexokinase activity may also play a role (Wasserman and Ayala 2005). Akt phosphorylation is also dynamic and occurs within minutes in response to insulin stimulation (Humphrey, Yang et al. 2013). In a clamp experiment where insulin has been infused into the animal over a period of 2 – 3 hours, it may be that insulin resistance at the level of phosphorylation is not detectable, and that specific insulin stimulation experiments in this model are required to clearly see any defect induced by HFD-feeding (Frangioudakis, Ye et al. 2005). Without a detectable defect in insulin signalling pathways in these animals following HFD, we are unable to delineate if SIRT3 may be playing a beneficial role under these conditions. While the clamp technique is the “gold-standard” method for assessing insulin sensitivity in vivo due to its ability to assess muscle insulin action, it is important to note that the hyperinsulinaemia induced during the clamp technique is a supraphysiological state and other methods may also be helpful in assessing more subtle changes in insulin sensitivity (Kowalski and Bruce 2014).

There are few published studies looking at the role of SIRT3 in glucose metabolism in vivo. Global SIRT3 KO studies have used techniques such as the glucose and insulin tolerance tests, and although they did find evidence of insulin resistance at 3-months of HFD, and impaired glucose tolerance after 12-months, the liver was the main organ

72 CHAPTER 3: Muscle Overexpression

investigated (Hirschey, Shimazu et al. 2011). Furthermore, many studies have shown that insulin resistance in muscle occurs within a few weeks of starting the high fat feeding regimes (Kraegen, Clark et al. 1991, Turner, Kowalski et al. 2013). The extended time-frame required to observe the effects of the SIRT3KO on glucose homeostasis are not consistent with SIRT3 playing a role in the early aetiology of insulin resistance. Supporting this, muscle specific SIRT3KOs were reported to have no differences in glucose metabolism when assessed by GTT and ITT after 8 weeks of HFD (Fernandez-Marcos, Jeninga et al. 2012).

One other gain of function model looking at SIRT3 in skeletal muscle was recently published. Lin et al have generated a muscle transgenic mouse model that, unusually, overexpresses the short M3 isoform of SIRT3 selectively in skeletal muscle (Lin, Chen et al. 2014). This isoform has previously been reported to be of very low expression level, and to have a short half life and uncertain localisation (Cooper, Huang et al. 2009). Despite this, Lin et al. reported changes in energy expenditure, exercise ability, reduction in muscle mass, and a preponderance towards more type I muscle fibres compared to WT controls (Lin, Chen et al. 2014). Glucose tolerance by GTT, and respiration in isolated mitochondria were measured in this study with both showing no alteration in the transgenic animal (Lin, Chen et al. 2014). The variability of findings associated with the SIRT3KO model, muscle-specific SIRT3KO model, and SIRT3 muscle transgenic model, in combination with the results presented in this thesis clearly point out that there is still much to learn about the precise role of SIRT3 in skeletal muscle.

However, mitochondrial metabolism is only one factor that can regulate insulin action in muscle. It is important to note is that while some studies show that enhancing mitochondrial substrate oxidation can improve insulin action (Bruce, Hoy et al. 2009, Benton, Holloway et al. 2010), this is not the case in all models where the oxidation of specific substrates is increased (Hoehn, Turner et al. 2010). It may be that under the conditions experienced in vivo in rats the enhanced mitochondrial respiration seen even with supra-physiological SIRT3 overexpression is not able to change energy balance in the whole animal, with all the compensatory mechanisms that make-up metabolic homeostasis.

73 CHAPTER 3: Muscle Overexpression

In conclusion, despite in vitro results and SIRT3KO models suggesting a role for SIRT3 in diet-induced obesity, muscle specific overexpression of SIRT3 has no effect on the acute development of skeletal muscle insulin resistance in the high fat fed rat. It is clear that more work needs to be done to fully delineate the role of SIRT3 in skeletal muscle.

74 CHAPTER 4: Liver Overexpression

CHAPTER 4 INVESTIGATION OF SIRT3 OVEREXPRESSION IN THE LIVER

4.1 Introduction

Chapter 3 investigated the effect of overexpression of SIRT3 on mitochondrial metabolism and insulin action in skeletal muscle. Although in muscle SIRT3 overexpression was not able to improve the insulin resistance seen in response to HFD feeding, there was an increase in substrate respiration in isolated mitochondria. The liver plays a central role in glucose and lipid homeostasis and is of great importance for maintaining energy homeostasis. In circumstances of excess lipid availability and increased lipid accumulation, the liver develops insulin resistance and is unable to adequately suppress endogenous hepatic glucose output or increase glycogen synthesis in response to insulin (Biddinger, Hernandez-Ono et al. 2008, Samuel and Shulman 2012). Mitochondrial dysfunction has also been reported in the liver of patients with insulin resistance associated with type 2 diabetes and non-alcoholic fatty liver disease (NAFLD) (Perez-Carreras, Del Hoyo et al. 2003, Szendroedi, Chmelik et al. 2009, Begriche, Massart et al. 2013), highlighting the importance of mitochondrial metabolism homeostasis in maintaining liver health. SIRT3KO mice have been used to define a role for SIRT3 in liver mitochondrial metabolism, including SIRT3 deletion in the development of fatty liver and the metabolic syndrome (Hirschey, Shimazu et al. 2011, Kendrick, Choudhury et al. 2011). However, the potential benefit of overexpressing SIRT3 in the liver in the context of metabolic syndrome remains unexplored.

Mass spectrometry analyses of the SIRT3KO liver acetylome have shown a role for SIRT3 in all major mitochondrial processes, including fatty acid metabolism, electron transport chain function, acetyl-CoA metabolism, and amino acid catabolism (Hebert, Dittenhafer-Reed et al. 2013, Rardin, Newman et al. 2013). Indeed much of the current knowledge of pathways that are regulated by SIRT3 deacetylation has come from studies in liver utilising the global SIRT3KO mouse model. Hirschey et al provided evidence of a role for SIRT3 in fatty acid and glucose metabolism showing SIRT3

75 CHAPTER 4: Liver Overexpression

deletion in liver caused defects in glucose tolerance, hepatic steatosis, and an accelerated development of the metabolic syndrome (Hirschey, Shimazu et al. 2011). Comprehensive investigation of overexpression of SIRT3 in the liver using an in vivo model or transgenic mouse has not been reported to date.

Because SIRT3KO animals have been shown to have a dysregulated metabolic phenotype that is associated with changes to acetylation status of liver enzymes, the hypothesis of this chapter was that overexpression of SIRT3 in the liver could counteract the effects of HFD in the in vivo setting. The aims were to overexpress SIRT3 in the liver of mice, and in both (1) an ex vivo model of isolated hepatocytes, and (2) in vivo in mice, to investigate the effects on mitochondrial metabolism and to determine if this overexpression is protective against metabolic defects induced by excess lipid. In vivo assessments were investigated under three main conditions; an acute study of 3 weeks SIRT3 overexpression and HFD-feeding, a fasting study of 3 weeks SIRT3 overexpression combined with a 48 hour fast prior to tissue collection, and a long-term overexpression study of 13 weeks SIRT3 overexpression and HFD- feeding regime.

76 CHAPTER 4: Liver Overexpression

4.2 Methods

General methods for western blotting, mitochondrial isolation, triglyceride determination and serum parameters are described in Chapter 2. Methods specific to the liver overexpression model are described below.

4.2.1 Animals and diet composition

Animal experiments in this chapter use C57BL6 mice from Animal Resource Centre (Perth, Australia) or from Australian BioResources (Moss Vale, Australia). 129- Sirt3tm1.1Fwa/J SIRT3KO mice and WT littermates (Lombard, Alt et al. 2007) were originally sourced from Jackson Laboratories (Bar Harbor, MA, USA), and were obtained from our own colony at Australian Bio Resources (Moss Vale, NSW, Australia). All animals were housed under standard laboratory conditions at 22 ± 0.5°C on a 12:12-hr light-dark cycle, with free access to water. Mice were fed ad libitum either a standard chow (8% calories from fat, 21% calories from protein, 71% calories from carbohydrate; Gordon’s Specialty Stock Feeds, NSW, Australia), or high-fat diet (HFD; 45% calories from fat (lard), 20% calories from protein, 35% calories from carbohydrates, 4.7 kcal/g; based on Rodent Diet #D12451 Research Diets, Inc., NJ, USA). Fasting cohorts were fasted for 48 hours prior to tissue collection. All experimental procedures were approved by the Garvan Institute/St. Vincent’s Hospital Animal Experimentation Ethics Committee and were in accordance with the National Health and Medical Research Council of Australia Guidelines on Animal Experimentation.

4.2.2 Generation and propagation of SIRT3-FLAG-pLIVE plasmid DNA

4.2.2.1 Provenance of constructs Human SIRT3-FLAG.pcDNA3.1+ (North, Marshall et al. 2003) was obtained from Eric Verdin via Addgene (Addgene plasmid 13814, Cambridge, MA). This construct corresponds to the SIRT3 transcript variant 2 (Gene ID: 23410), which encodes the

77 CHAPTER 4: Liver Overexpression

slightly shorter isoform b of human SIRT3 (NCBI Reference Sequence: NM_001017524.2) that has been published on extensively (Schwer, North et al. 2002, North, Marshall et al. 2003, Schwer, Bunkenborg et al. 2006, Hirschey, Shimazu et al.

2010). SIRT3-FLAG was cloned into the pLIVE™ (Liver In Vivo Expression) Vector (Mirus Bio, Madison, WI, USA) using traditional subcloning techniques as detailed below. The luciferase reporter Luc2 was obtained from pGL4.30-luc2P/NFAT-RE/Hygro vector (Promega, Madison, WI, USA), and cloned into pLIVE using similar techniques to those described below.

4.2.2.2 Subcloning of SIRT3-FLAG into pLIVE SIRT3-FLAG was excised from pcDNA3.1 using sequential digestion with NheI and XbaI (Promega). 2 µg of DNA was digested for 2 hours each at 37 °C in Buffer B, supplemented with 100 mM NaCl with addition of XbaI. Empty pLIVE plasmid was digested with NheI in Buffer B for 2 hours at 37 °C. Sizes were checked on a 1% agarose gel run at 100 V for ~1 hour with ethidium bromide, and imaged on a GelDoc (BioRad, Hercules, CA, USA). Following identification of correct bands, digested SIRT3-FLAG DNA was further purified by agarose gel electrophoresis with 0.5% crystal violet in 50% methanol added to agarose gel to allow identification of the SIRT3 band without the use of damaging UV transillumination. After running for ~1 hour at 100 V, the gel was transferred to a lightbox for easier visualisation and the desired band was cut out of the gel using a scalpel blade. The DNA was purified from the gel slice using the Promega Wizard Gel extraction kit (Promega) according to the manufacturer’s instructions.

4.2.2.3 Phosphatase treatment, ligation and transformation of SIRT3-FLAG-pLIVE To prevent re-circularisation of plasmid DNA during the sub-cloning procedure, 1 μg of linearised plasmid was treated with shrimp alkaline phosphatase (SAP) (Promega) according to manufacturers instructions. The reaction was incubated at 37° C for 15 minutes and the enzyme was heat inactivated at 65 °C. A 10 µl ligation reaction was set up with a 1:3 molar ratio of vector:insert DNA. 60 ng of vector DNA was used and the appropriate amount of insert was calculated and ligated using the T4 Rapid ligation

78 CHAPTER 4: Liver Overexpression

system (Promega) and incubated at 16 °C for 15 minutes. Ligation mix was transformed into chemically competent TOPO 10 E. coli cells (Life Technologies, Carlsbad, CA, USA) as per the manufacturer’s instructions. Following initial 1 hour incubation, cells were plated on an agar plate with 50 µg/ml Kanamycin (Sigma, St Louis, MO, USA) to select for transformed cells. Colonies were grown overnight at 37 °C.

4.2.2.4 Small scale preparation of plasmid DNA and sequencing Single colonies were selected and used to inoculate 6 ml of Lysogeny Broth media (LB; 1% Bacto tryptone, 0.5% yeast extract, 0.5% NaCl, 4 mM NaOH) supplemented with 50 µg/ml kanamycin with a single bacterial colony and culturing overnight at 37 °C with shaking. Plasmid DNA was purified from 4 ml of bacterial culture using Promega Wizard Plus SV miniprep system (Promega) according to the manufacturer’s protocol. Colonies were initially screened using a restriction digest with the enzyme KpnI (Promega) at 37 °C for 2 hours, and run on a 1% agarose gel at 80 V for 2 hours. Colonies that had the correct size bands of 4.3k b and 300 bp were selected for sequencing. Constructs were verified by sequencing performed by Garvan Molecular Genetics (GMG, Garvan Institute Core Facility, Sydney, Australia). 50 ng of SIRT3-FLAG- pLIVE was sequenced using 3.2 pmol of the following primers: SIRT3 F2, TTT CTG TGG GTG CTT CAA GTG; pLIVE F1, GAA GAG TCT AAC AGC CAG; pLIVE REV, GCT CTT GGA AAT GGT CAA TC; and pLIVE REV2, GAA GTC TGG AAT GCC ACT GG.

4.2.2.5 Large-scale preparation of plasmid DNA Plasmids were propagated for injection into animals using commercially available endotoxin-free gigaprep kits (Qiagen, Melbourne, Victoria, Australia) as per the manufacturer’s instructions. Frozen stocks of transformed E.coli were inoculated in a 5ml starter culture of LB medium with added antibiotic for 6 hours in a 37 °C shaking incubator. The starter culture was then transferred into ~3 L of LB medium with antibiotic and incubated in a 37 °C shaking incubator overnight. Bacteria were lysed and the plasmid was purified following manufacturers instructions. The final purified plasmid pellet was resuspended in sterile 0.9% saline to a final concentration of 1

79 CHAPTER 4: Liver Overexpression

mg/ml. 1-2 µl of DNA was quantified using the Nanodrop Spectrophotometer (ND- 1000 3.3.1, Thermo Scientific, Waltham, MA, USA) to determine the concentration. The 260 nm/280 nm ratio was measured to determine the purity of the DNA preparation. The ratio was routinely between 1.8 and 1.9 indicating the preparation was free of contamination.

4.2.3 Overexpression of SIRT3 in liver

SIRT3 overexpression or expression of a control gene was achieved in mouse liver using a hydrodynamic tail vein injection (HTVI) technique in anaesthetised mice as detailed below. Empty pLIVE plasmid or pLuc-pLIVE encoding luciferase was used for injection of control mice. Mice were allowed to recover from the injection for 1 -2 days, and were then placed on either a chow or high fat diet for 1-16 weeks depending on the study.

4.2.3.1 Preparation of buffer for hydrodynamic tail vein injection (HTVI) This procedure facilitates gene transfer to the liver. The method involves direct tail vein injection of naked plasmid DNA into a mouse. The plasmid is diluted in a large volume equivalent to 10% of the body weight of the animal and the rapid injection (~5 seconds) of this solution causes expansion of the liver vasculature, allowing the entry of the plasmid into the hepatocytes (Wolff and Budker 2005, Budker, Subbotin et al. 2006) (Liu, Song et al. 1999). KRB injection buffer (68 mM NaCl, 2.35 mM KCl, 0.6 mM

KH2PO4, 2.5 mM NaHCO3 and 0.6 mM MgSO4.7H20, 10 mM Hepes, 100 mM CaCl2, and 2.8 mM glucose) was buffered by bubbling through an air mixture for 15 mins, adjusted to pH 7.4, and filtered through a 0.2 µm filter (Millipore, Billarica, MA, USA) and kept at room temperature. 40 µg of SIRT3-FLAG, 40 μg empty-pLIVE, or 10 μg of pLuc-pLIVE DNA was added to the appropriate volume (10% animal’s body weight) of buffer immediately before injection. The injection solution was loaded into a 3 ml syringe with a 27½-gauge needle and any air bubbles were removed.

80 CHAPTER 4: Liver Overexpression

4.2.3.2 Preparation and injection of animal Anaesthesia in animals was induced initially with 3% isoflourane in oxygen, and animals maintained on a nose cone at 2% isoflurane and placed on a pre-warmed heating pad. To faciliate tail vein visualisation and optimal injection the tail vessels were dilated immediately prior to injection using a heat lamp (120 W bulb) for 3-5 minutes. The injection was made to the dilated tail vein on the ventral side of the tail starting from the distal end. The area to be injected was swabbed with alcohol. The syringe needle was inserted into the vein and the entire liquid was injected in 4-7 seconds at a constant rate. Animals were immediately removed from anaesthesia and injected with 5 mg/kg ketoprofen intraperitoneally (i.p). Animals were monitored closely upon recovery before being returned to home cage. Cages of animals receiving HTVI were placed on heat pads for overnight recovery, and weighed daily until pre- injection weight was reached. HTVI causes minimal long-term damage, with any inflammation reported to be resolved within 24 hrs (Zhang, Gao et al. 2004).

4.2.3.3 In vivo imaging of mice post-HTVI In vivo imaging of animals co-injected with pLuc-pLIVE was conducted 1 week post injection using the IVIS Spectrum (Perkin Elmer, Waltham, MA). Mice were injected i.p. with 150 mg/kg D-luciferin (Gold Bio, St-Louis, MO, USA) made up at 15 mg/ml in sterile PBS and sterilised using a 0.2 μm filter (Millipore). 10 minutes after injection, animals were anesthetised with 2% isoflurane and placed into the IVIS Spectrum. 4 mice at a time were imaged using standard bioluminescent settings. IVIS images were analysed using Living Image 4.0 software (Caliper Life Sciences, Hopkinton, MA).

4.2.4 Gene expression analysis using quantitative PCR

4.2.4.1 RNA preparation Total RNA was isolated from frozen liver tissue using TRIzol reagent (Life Technologies). Tissue was homogenised in TRIzol using a Polytron homogeniser (Kinematica AG, Lucerne, Switzerland) to disrupt all cell membranes and release RNA, DNA, and protein. Chloroform was used to separate the RNA phase from DNA and protein. Total RNA was precipitated with isopropanol and resuspended in 30-50 μl of DEPC-treated

81 CHAPTER 4: Liver Overexpression

water. RNA was further cleaned-up by addition of 1/10th the volume of 3M sodium acetate pH 5.5 and 3 volumes of 100 percent ethanol and precipitated at -80 °C for 2 hours. Precipitate was recovered by centrifugation, washed in 70% ethanol, dried and resuspended in 40 μL of RNase-free water. Following quantitation by nanodrop, 4 μg of total RNA was incubated with DNase enzyme (Promega) to digest any remaining genomic DNA. 625 ng of DNase-treated RNA was then reverse transcribed with Superscript III (Life Technologies) employing a mixture of 2.5 μM oligo d(T) and 10 μM random 9-mer primers (both New England Biolabs, Ipswich, MA, USA) to enable more sensitive detection and the use of a ribosomal endogenous control gene.

4.2.4.2 Quantitative PCR SIRT3 gene expression was analysed with Universal Probe Library (UPL) probes and standard oligo primers using the LightCycler480 system and software (Roche, Basel, Switzerland). Relative quantification was assessed using the standard curve method of gene expression determination. 3 primer sets were designed to differentiate between endogenous mouse SIRT3, human SIRT3 and to pick-up both human and mouse SIRT3 together. Relative gene expression was normalized to the ribosomal gene RPS13 as an endogenous control. UPL Probe and primer sequences used were: humSIRT3 (Probe 39, Forward: 5’-GATCTGCTGCTCATCCTTGG-3, Reverse: 5’-TCGTTCCCCGACTGCTC-3’), musSIRT3 (Probe 25, Forward: 5’-AGGCCCAATGTCACTCACTAC-3’, Reverse: 5’- GAGCATCTGGGATCCCTG-3’), hum/musSIRT3 (Probe 10, Forward: 5’- CCGACATTGTGTTCTTTGGGGAGC-3’, Reverse: 5’-CCCCGACTGCTCATCAA-3’) and RPS

(Probe 110, Forward: 5’-TGCTCCCACCTAATTGGAAA-3’, Reverse: 5’- TGCTGTTGTGTGCACAAGC-3’). LightCycler480 was programmed according to the following protocol: Initial activation step (95 °C, 10 min), followed by 40 repetitions of the cycling stage (95 °C, 20 sec template denaturation; 60 °C, 45 sec primer annealing; 72 °C, 20 sec sequence extension). The reaction was concluded with a final cooling step (40 °C, 30 sec). Relative gene expression in arbitrary units was extrapolated from a standard curve of pooled sample diluted in a 1:4 serial dilution and normalised to the level of RPS13.

82 CHAPTER 4: Liver Overexpression

4.2.5 Isolation of primary hepatocytes from mice following HTVI

Primary hepatocytes were isolated from mice 3 weeks after HTVI using a two-step perfusion method (Berry and Friend 1969, Klaunig, Goldblatt et al. 1981). Mice were injected i.p. with 125 mg/kg ketamine and 25 mg/kg xylazine to induce deep anaesthesia. The liver was exposed and perfused in-situ using a 21-gauge blood collection needle (Vacutainer Safety-Lok, BD Biosciences, Franklin Lake, NJ, USA) connected to a perfusion pump set-up. Initial perfusion occurred via the inferior vena cava, with outflow via the hepatic portal vein, with pre-warmed 1x HBSS (Hanks Buffered Salt Solution) with 0.5 mM EDTA (138 mM NaCl, 50 mM HEPES, 5.6 mM

Glucose, 5.4 mM KCl, 0.34 mM Na2HPO4, 0.44 mM KH2PO4, 4.17 mM NaHCO3, pH 7.4, at 37 °C). After 15 mins perfusion at a flow rate of 5 ml/min, liver was perfused with 50 ml of 1 mg/ml collagenase H (Roche) in pre-warmed 1xHBSS supplemented with 2 mM

CaCl2. The digested liver was removed intact and placed in HBSS-CaCl2. Hepatocytes were released by gentle teasing of the softened liver and then filtered through a 100 μm cell strainer (BD Biosciences). Cells were then washed three times with cold HBSS- CaCl2 at 50 g for 3 min and the supernatant discarded. The final cell pellet was resuspended in serum-free M199 + Earles Salts (Life Technologies) for cell counting.

4.2.6 Cell culture conditions

Hepatocytes were cultured on plates coated with 0.5 μg/cm2 rat-tail collagen (BD Biosciences) in M199 media with 5.5 mM glucose containing the following supplements: Pen/Strep (Life Technologies, 100 U/ml), 0.1% BSA (Life Technologies), Ultroser G (Pall Corp, Port Washington, NY, USA), 100 nM dexamethasone (Sigma), 100 nM insulin (Actrapid, Novo-Nodisk, Copenhagen, Denmark). After 4 hours attachment, media was removed and fresh media added as described below. All cells were used in subsequent experiments 16 - 24 hours after plating. After 4 hours media was replaced with basal M199 media supplemented with Pen/Strep (Life Technologies, 100 U/ml) and 100 nM dexamethasone. For fatty acid treatments basal M199 media was supplemented with either fatty acid or vehicle media prepared as described below.

83 CHAPTER 4: Liver Overexpression

4.2.7 Preparation of fatty acid or vehicle treatment media

Fatty acid (FA) media was prepared by the addition of 250 μM each of oleic acid, linoleic acid and palmitic acid (750 μM total) to basal M199 media containing 2% BSA to simulate the effects of high physiological concentrations of long-chain fatty acids as recommended for cell culture experiments (see discussion in (Watt, Hoy et al. 2012)). Vehicle (VEH) control media was prepared by addition of ethanol to basal M199 with 2% BSA. Both media were allowed to conjugate for 2 hours at 55 °C. Media was sterilised by passing through a 0.2 μm syringe filter and allowed to cool to 37 °C before being placed on cells for overnight incubation for 14-24 hours.

4.2.8 Oxygen consumption in hepatocytes

For oxygen consumption experiments, cells were seeded in collagen-coated Seahorse XF24 cell culture plates at 20,000 cells per well. Following overnight FA treatment, media was removed, and cells washed in seahorse media (Dulbecco’s Modified Eagle’s Media, unbuffered, with 5.5 mM glucose (Life Technologies) supplemented with 10 mM sodium pyruvate, pH 7.4). Cells were allowed to equilibrate in 500 μL seahorse media for 30 mins at 37 °C without CO2. Oxygen consumption was measured using a Seahorse XF24 Extracellular Flux Analyzer (Seahorse Bioscience, North Billerica, MA, USA). Hepatocytes from SIRT3 and CON animals under different treatment conditions were measured simultaneously with minimum triplicate wells per condition. Bioenergetic profiling was performed by monitoring basal oxygen consumption for 30 minutes followed by the sequential injection of the following inhibitors, 1 μg/ml oligomycin, 1 μM Carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (FCCP), 4 μM rotenone and 1 μM antimycin A. Basal oxygen consumption, ATP-linked respiration, and proton leak, were calculated from the primary data. The data are shown as percentage change from control vehicle-treated cells (CON), or from WT vehicle- treated cells in the SIRT3KO experiments.

84 CHAPTER 4: Liver Overexpression

4.2.9 Triglyceride in hepatocytes

To assess triglyceride accumulation in hepatocyte cultures, cells were seeded in 6-well culture dishes at 4 x105 cells/well and cultured overnight as described earlier. After 16 hours, spent media was removed and cells were washed twice with PBS and collected into a 1.5 ml tube in 200 μl of ice-cold PBS using a cell scraper (Corning Inc., Corning, NY, USA). The well was rinsed with an additional 100 μl of PBS to ensure all cell material was collected. Cell lysate was sonicated for 5 secs and 10 μl taken for protein determination using Bradford assay. Lipid was extracted from the cell lysate in 2 ml of chloroform:methanol and processed as for tissue triglycerides described in Chapter 2.

4.2.10 Substrate utilisation assays:

4.2.10.1 Glucose oxidation in primary hepatocytes To assess substrate utilisation in hepatocyte cultures, cells were seeded in 6-well culture dishes at 4 x105 cells/well and cultured overnight with or without FA as described earlier. After 16 hours, spent media was removed and cells were washed with PBS and incubated for 1 hour in 1 ml of FA-conjugated glucose oxidation media (fatty acid or vehicle treatment media (prepared as described in Section 4.2.7) supplemented with 2 μCi U-14C-glucose) with and without 100 nM insulin. After 1 hour media is removed and placed in a sealed glass vial with 400 μl 1M perchloric acid and a

1.5 ml tube containing 100 μl 1 M NaOH to collect CO2 at room temperatue for 2 hours. After 2 hours the NaOH was removed from the tube and counts were determined by the addition of scintillation fluid (Ultima Gold XR, Packard Biosciences, Groningen, Netherlands) and measurement using a liquid scintillation counter (Beckman LS6000, Beckman Coulter, Pasadena, CA, USA). Cells were washed twice with PBS and collected in 300 μl 1 M KOH using a cell scraper and cell lysate assessed for protein concentration.

4.2.10.2 Palmitate oxidation in primary hepatocytes Cells were seeded as described for glucose oxidation and cultured overnight with and without FA treatment. After 16 hours, spent media was removed and cells were

85 CHAPTER 4: Liver Overexpression

washed twice with PBS and incubated for 1 hour in 1 ml of palmitate oxidation media (Basal M199 media with 2% BSA, conjugated to 200 μM palmitate (prepared as described in Section 4.2.7) and supplemented with 0.5 μCi 1-14C-palmitic acid). After 1 hour media was removed and placed in a sealed glass vial with 400 μl 1 M perchloric acid and a 1.5 ml tube containing 100 μl 1 M NaOH to collect CO2 at room temperature. After 2 hours the NaOH was removed from the tube and counts were determined by the addition of scintillation fluid (Ultima Gold XR, Packard Biosciences) and measurement using a liquid scintillation counter (Beckman LS6000, Beckman). Acidified media was collected into a 1.5 ml tube to determine the acid-soluble metabolites. Acidified media was centrifuged at maximum speed to pellet insoluble material, and supernatant (acid soluble metabolites) were collected and 100 μL added to 5 ml scintillation fluid for counting as above. Cells were washed twice with PBS and collected in 300 μl 1 M KOH using a cell scraper and cell lysate assessed for protein concentration.

4.2.11 Assessment of body weight and body composition

Body weight was assessed weekly or twice-weekly, which also accustomed mice to handling. Lean mass and fat mass were measured using a Piximus2 Dual Energy X-ray Absorptiometry (DEXA) scanner according to the manufacturer’s instructions (GE Lunar, Madison, WI, USA).

4.2.12 Assessment of metabolic parameters: GTT and serum measures

Glucose tolerance tests (GTT) were carried out following a 5 hour fast from 8am. Mice were injected i.p. with glucose (2 g/kg), and blood glucose levels were monitored from the tail-tip using an Accucheck II glucometer (Roche) for 90 min following glucose injection. Insulin levels during GTT were measured using 10 μl of whole blood in the Ultra Sensitive Insulin ELISA Kit (Crystal Chem, Downers Grove, IL, USA). Blood ketones were measured from tail-tip using the Optium Ketone Blood β-Ketone Test Strips (Abbott).

86 CHAPTER 4: Liver Overexpression

4.2.13 Substrate oxidation in liver tissue homogenates

To examine substrate utilisation in control and SIRT3 overexpressing liver, a 1:19 homogenate of fresh liver tissue was prepared in ice-cold homogenising buffer (250 mM sucrose, 10 mM Tris-HCl, 1 mM EDTA, pH 7.4) using a Polytron homogeniser (Kinematica AG). 50 μl of homogenate was incubated in 450 μl pre-warmed (30°C) oxidation medium (100 mM sucrose, 10 mM Tris-HCl, 5 mM KH2PO4, 1 mM MgCl2, 80 mM KCl, 0.2 mM EDTA, 1 mM DTT, 2 mM ATP, 0.3% fatty-acid free BSA, pH 7.4 with 10 mM nicotinamide). For pyruvate oxidation, oxidation medium was supplemented with 2 mM malate, 5 mM pyruvate, and 0.2 μCi 2-14C pyruvic acid, while for glutamate oxidation measurements the medium was supplemented with 2.2 mM malate, 5.5 mM glutamate, and 0.2 μCi 1-14C glutamate (Turner, Bruce et al. 2007). Homogenates were incubated in glass vials for 90 min in a 30 °C waterbath, with gentle shaking. All vials contained an open 1.5 ml tube with 100 μl 1 M NaOH to capture CO2 produced. The reaction was terminated by acidification with 100 μl 1 M perchloric acid, and vials were 14 left at room temperature for 2 hr to capture all CO2 produced. The NaOH was removed from the 1.5 ml tube and counts were determined by the addition of scintillation fluid (Ultima Gold XR, Packard Biosciences) and measurement using a liquid scintillation counter (Beckman LS6000, Beckman).

4.2.14 Statistical analysis

All results are presented as mean ± SEM. Results were compared using 2-way ANOVA for effects of overexpression and diet, or overexpression and fasting as appropriate, with Sidek’s post-hoc test to compare between groups as appropriate. GTT results were compared using repeated measures 2-way ANOVA and Tukey’s post-hoc test to compare between groups. Where appropriate unpaired t-tests were used. Statistical analysis was performed in GraphPad Prism software (Prism 6, Version 6.0b, Oct 3, 2012).

87 CHAPTER 4: Liver Overexpression

4.3 Results

4.3.1 Verification of HTVI technique and SIRT3 overexpression

4.3.1.1 SIRT3 mRNA is increased following HTVI To verify the success of the SIRT3 hydrodynamic tail vein injection (HTVI) technique, mice that received SIRT3-FLAG by HTVI for 3 weeks were assessed for SIRT3 mRNA in liver. Primers were designed to differentiate between the endogenous mouse SIRT3 transcript (musSIRT3), the introduced human SIRT3-FLAG (humSIRT3), and one common primer pair that detected both forms of SIRT3 transcript (hum/musSIRT3). Real-time quantitative PCR clearly shows that while there was no difference in the level of endogenous SIRT3 between the mice that underwent the SIRT3-FLAG transfection or control HTVI, the human specific primer pair showed a high level of SIRT3 mRNA detectable only in those mice that received the SIRT3-FLAG construct, demonstrating specificity of this primer pair for human SIRT3. The common primer pair showed an approximate 10-fold increase in total SIRT3 transcript by qPCR (Figure 4.1C).

A B C musSIRT3 primers humSIRT3 primers hum/musSIRT3 primers

80 80 80

60 60 60

40 40 40

mRNA A.U. 20 20 20

0 0 0 CON SIRT3 FLAG CON SIRT3 FLAG CON SIRT3 FLAG

Figure 4.1 SIRT3 mRNA overexpression 3 weeks after HTVI using three different qPCR primer sets: A. Liver mRNA assessed with mouse specific SIRT3 primers (musSIRT3) show endogenous SIRT3 mRNA 3 weeks after HTVI. B. Human specific SIRT3 primers (humSIRT3) show exogenous SIRT3 mRNA in liver following HTVI. C. SIRT3 mRNA using primers designed to pick up both endogenous and exogenous SIRT3 (hum/musSIRT3) with equal efficiency 3 weeks after HTVI. n=4 animals per group, data are mean of triplicates in arbitrary units (A.U.) extrapolated from standard curve using the Roche UPL qPCR system.

88 CHAPTER 4: Liver Overexpression

4.3.1.2 Timecourse of SIRT3 overexpression in liver

To verify the timeframe of SIRT3 overexpression following HTVI, a cohort of mice underwent a timecourse of SIRT3 overexpression from 1 – 16 weeks. Significant overexpression of SIRT3-FLAG protein was detected by western blot from 1 -10 weeks post-HTVI as shown in Figure 4.2A using the h/mSIRT3 antibody (Cell Signalling Technologies) that detects both endogenous and exogenous SIRT3. Endogenous SIRT3 protein shown in the lower band does not change over the timecourse, while the slightly larger FLAG-tagged human SIRT3 protein is overexpressed at all timepoints from 1 week through to 10 weeks. SIRT3 overexpression at 13 and 16 weeks was also observed in independent experiments (data not shown).

4.3.1.3 SIRT3 protein overexpression in liver, isolated mitochondria, and primary hepatocytes SIRT3 is expressed as a 45 kDa protein that has a mitochondrial localisation sequence at the N-terminus which is cleaved upon trafficking to the mitochondria, leaving the enzymatically active 28 kDa mitochondrial form. Figure 4.2C shows both the uncleaved and cleaved forms of the protein in both whole liver lysates and isolated mitochondria using the human-specific SIRT3 antibody. This demonstrates that SIRT3-FLAG was successfully overexpressed at the protein level, and processed correctly by the cell to its mature ~28 kDa form, as well as being enriched in the mitochondrial fraction suggesting correct subcellular targeting.

The liver is made up primarily of hepatocytes, although other cell types including kupffer cells, endothelial cells and fibroblasts are also present. The pLIVE plasmid contains a modified albumin promoter to drive expression. To confirm that the SIRT3 overexpression seen in whole liver lysates is occurring in the appropriate liver cells, primary hepatocytes were isolated and subjected to western blot for SIRT3. Figure 4.2D shows lysates from hepatocyte isolations from 2 control and 5 SIRT3 overexpressing animals 2 – 3 weeks after HTVI. SIRT3 overexpression is confirmed to be occurring in the main liver parenchymal cell, the hepatocyte.

89 CHAPTER 4: Liver Overexpression

A LIVER LYSATES Weeks CON 1 2 3 4 5 6 8 10

hSIRT3-FLAG 30 kDa endogenous SIRT3 28 kDa

B 5 *** 4 3 2 CON (A.U.)

SIRT3 SIRT3 Protein/ 1 0 1 2 3 4 5 6 8 10 CON Weeks post-HTVI

C Mitochon- Whole Liver drial Lysate Isolation

45 kDa hSIRT3-FLAG uncleaved

hSIRT3-FLAG cleaved 28 kDa

GAPDH 37 kDa

D ISOLATED HEPATOCYTES CON + SIRT3-FLAG

hSIRT3-FLAG 30 kDa 28 kDa endogenous SIRT3

Figure 4.2 Western blots of SIRT3 overexpression in liver: A: Representative timecourse of overexpression from 1 – 10 weeks post-HTVI in liver lysates using h/mSIRT3 antibody which detects both the 28kDa endogenous SIRT3 protein, and the ~30kDa hSIRT3-FLAG. B. Average fold-increase in SIRT3 over control for timecourse by densitometry. Data are mean± SEM, n=1-3, average of triplicate gels. ***p=0.0002 by one-way ANOVA compared to control. C. Enrichment of SIRT3-FLAG in mitochondrial fraction compared to whole liver lysate using hSIRT3 antibody which detects only the SIRT3-FLAG. D. SIRT3 overexpression in isolated primary hepatocytes from animals that underwent ether control or SIRT3-FLAG HTVI using the h/mSIRT3 antibody that detects both endogenous and human SIRT3.

90 CHAPTER 4: Liver Overexpression

4.3.1.4 Verification of HTVI using in vivo imaging Using fluorescent protein constructs and the HTVI technique we have previously confirmed that the liver specific pLIVE plasmid induces overexpression specifically in the liver and not in other organs including heart or kidney (Jane Reznick, PhD Thesis, 2011). Due to the technical difficulty of the HTVI technique however (poor overexpression despite a seemingly successful injection), a method of confirming the success of the HTVI technique prior to the commencement of extended diet studies was developed. To this end, a luciferase expressing pLIVE plasmid was made that could be co-injected with other constructs to identify successful injections at 1 week post- injection (pLuc-pLIVE). Figure 4.3 shows in vivo luminescence in the livers of mice that underwent the HTVI procedure versus a control mouse at 1 week post-injection. Co- HTVI with pLuc-pLIVE is therefore a good method to confirm HTVI success prior placing cohorts of animals on diets for a considerable amount of time. Data in Figure 4.2 and Figure 4.3 clearly establishes the utility of using HTVI to specifically overexpress SIRT3 in liver for at least 10 weeks to enable investigation of the effects of increased SIRT3 protein on liver metabolism in vitro and in vivo.

91 CHAPTER 4: Liver Overexpression

CON +SIRT3 / pLuc

Figure 4.3 Verification of HTVI delivery method by in vivo imaging: SIRT3-FLAG co-injected with pLuc-pLIVE at 1 week post-injection shows successful luminescence following injection of luciferin. CON denotes control animal that did not receive HTVI with 3 animals that were co-injected with 40 µg SIRT3-FLAG-pLIVE and 10 µg pLuc-pLIVE. Scale is shown on right in Radiance, a measure of photon emission from the subject (photons/second/cm2/steradian).

4.3.2 SIRT3 overexpression in an ex vivo model of isolated hepatocytes

The aim of this chapter is to investigate the effect of SIRT3 overexpression in the liver, specifically in terms of glucose and fatty acid metabolism. Isolated primary hepatocytes are an ideal model for assessing liver metabolism as they retain many of the cellular metabolic networks and subcellular compartmentalisation of liver tissue, in comparison to assays in homogenised tissues (Hewitt, Lechon et al. 2007). Isolated hepatocytes are superior to hepatoma cell lines such as HepG2 which do not replicate all the metabolic functions of true liver cells. For example HepG2 lack enzymes of the urea cycle, and can exhibit cancer-like metabolic properties depending on the culture conditions (Hewitt, Lechon et al. 2007, Iyer, Yang et al. 2010).

92 CHAPTER 4: Liver Overexpression

4.3.2.1 Viability and purity of murine hepatocyte isolations Primary hepatocytes were isolated from mice that had undergone the HTVI technique to induce overexpression of SIRT3 specifically in the liver (see Figure 4.2C). SIRT3KO mice and their wildtype (WT) littermates were also used to investigate the effect of SIRT3 deletion. Table 4.1 details average viability and cell yield for all isolations performed in both mouse lines across all experiments. Hepatocytes isolated showed high viability of close to 90% and yields on average of 40 million cells per mouse liver, illustrating a robust experimental protocol in line with published methods (Hatano, Bradham et al. 2000, Gonçalves, Vigário et al. 2007). Viability was not different between mice that received HTVI or control plasmid, SIRT3-FLAG, or no hydroporation at all (data not shown).

Table 4.1 Hepatocyte yield and viability upon isolation BL6 MICE SIRT3KO/WT n 16 25

Cell viability at Plating (%) 85 ± 3 90 ± 1

Cell Yield per mouse (million cells) 40 ± 5 43 ± 4

4.3.2.2 Oxygen consumption in primary hepatocytes after SIRT3 HTVI SIRT3 deacetylates a variety of metabolic enzymes, and an increase in its activity is expected to increase flux through mitochondrial metabolic pathways. To ascertain whether increasing SIRT3 protein could increase mitochondrial metabolism, oxygen consumption was measured in cultured hepatocytes using the XF24 Seahorse analyser (Seahorse Bioscience). Hepatocytes were treated overnight with fatty acid (FA) or vehicle (VEH) control to simulate the environment of excess lipid availability seen by the liver of a high fat fed animal. Figure 4.4 shows primary mouse hepatocyte respiration over time, measured under both basal conditions in media alone, and with the addition of inhibitors such as the Complex V inhibitor oligomycin to inhibit ATP production, maximal respiration with the addition of the uncoupler FCCP, and rotenone and antimycin A to knock-down all mitochondrial respiration. As shown in Figure 4.5, basal and maximal respiration were significantly increased with SIRT3 overexpression. However there was no significant increase in ATP-linked respiration or

93 CHAPTER 4: Liver Overexpression

the rate of oligomycin-insensitive oxygen consumption (which reflects proton leakage across the inner mitochondrial membrane) with SIRT3 overexpression, although both tended upwards. As can be seen in Figure 4.4, the spare respiratory capacity of isolated hepatocytes was not increased above basal with 1 μM FCCP, although previous optimisation experiments showed this dose to be optimal in hepatocyte cultures for eliciting maximal respiration (data not shown). The addition of FA treatment to SIRT3 overexpressing hepatocytes slightly blunted the increase in respiration with SIRT3 overexpression, with only the vehicle treated cells showing a significant increase with SIRT3 overexpression when assessed using unpaired t-test (Figure 4.5). The addition of fatty acid had no effect on respiration of control cells.

Basal +Oligo +FCCP +Rot/AA 250 +CON VEH +CON FA 200 +SIRT3 VEH +SIRT3 FA 150

100 (% (% Change/CON)

Oxygen Consumption Oxygen 50

0 0 9 35 43 61 70 87 96 TIME (mins)

Figure 4.4 Oxygen consumption in primary hepatocytes overexpressing SIRT3, over time with various inhibitors: Primary hepatocytes were assessed using the Seahorse XF24 analyser for basal respiration, with sequential injection of 1 ug/ml oligomycin, 1 μM FCCP, 4 μM rotenone with 1 uM antimycin A. Blue solid circles, and open circles denote vehicle-treated control cells and FA- treated control cells respectively. Red solid triangles and open triangles denote vehicle-treated SIRT3 overexpressing hepatocytes and FA-treated SIRT3 hepatocytes respectively. Data are mean± SEM from n= 7 animals, over 5 independent experiments. Data presented as percentage change from basal control vehicle treated cells in pmol O2/min/mg protein.

94 CHAPTER 4: Liver Overexpression

A B +CON 250 † 250 ‡ * +SIRT3 200 200

150 150 Consumption 2 Consumption 100

2 100

50 50 (% (% Change/CON) (% (% Change/CON)

Basal O 0 0 VEH FA O Maximal VEH FA

C D 150 80

60 100 40 50

Proton Leak Proton 20 (% (% Change/CON) (% (% Change/CON)

ATP-linked Respiration 0 0 VEH FA VEH FA

Figure 4.5 Mitochondrial respiration is increased in SIRT3 overexpressing hepatocytes: Primary hepatocytes were assessed using the Seahorse XF24 analyser to measure oxygen consumption. A. Basal cellular respiration. B. Maximal uncoupled respiration rate following injection of 1 μM FCCP. C. ATP-linked respiration is basal respiration minus respiration with the addition of 1 μg/ml oligomycin. D. Proton Leak extrapolated from respiration with oligomycin minus respiration with the addition of 4 uM rotenone and 1 uM antimycin A. Blue bars denote control cells and red bars denote SIRT3 overexpressing hepatocytes. Data are mean± SEM presented as percentage change from basal control vehicle treated cells in pmol O2/min/mg protein. n= 7 animals, over 5 independent experiments. †p=0.05, ‡p=0.03, main effect of SIRT3 by 2-way ANOVA; *p<0.03 unpaired t-test CON cf. SIRT3.

4.3.2.3 Oxygen consumption in SIRT3KO hepatocytes To investigate the effect of SIRT3 deletion in this model, oxygen consumption was also measured in cultured hepatocytes from SIRT3KO and WT animals using the XF24 Seahorse analyser (Seahorse Bioscience). SIRT3 deletion had a less robust effect on oxygen consumption than that seen in the SIRT3 overexpression model above. As shown in Figure 4.6, basal and maximal respiration was slightly decreased with SIRT3

95 CHAPTER 4: Liver Overexpression

overexpression, however not significantly. The addition of FA treatment to SIRT3KO and WT hepatocytes also had no significant effect on cellular respiration.

Basal +Oligo +FCCP +Rot/AA A 150 WT+VEH WT+FA SIRT3KO+VEH 100 SIRT3KO+FA

50 (% (% Change/CON) Oxygen Consumption Oxygen

0 0 9 35 43 61 70 87 96 TIME (mins)

B C

150 150 WT SIRT3KO 100 100 Consumption 2 Consumption 2 50 50 (% (% Change/CON) (% (% Change/CON)

Basal O 0 0

VEH FA O Maximal VEH FA

Figure 4.6 Mitochondrial respiration in SIRT3KO hepatocytes compared to WT: Primary hepatocytes from SIRT3KO and WT mice were assessed using the Seahorse XF24 analyser to measure oxygen consumption. A. Basal respiration followed by sequential injection of 1 μg/ml oligomycin, 1 μM FCCP, 4 μM rotenone with 1 μM antimycin A. Blue solid and open circles denote WT cells with vehicle or FA-treatment respectively. Green solid and open triangles denote SIRT3KO cells with vehicle or FA-treatment respectively. B. Basal cellular respiration. C. Maximal uncoupled respiration rate following injection of 1 μM FCCP. Data are mean± SEM from n= 10-11 animals, over 7 independent experiments. Data presented as percentage change from basal control vehicle treated cells in pmol O2/min/mg protein.

96 CHAPTER 4: Liver Overexpression

4.3.2.4 Triglyceride accumulation in cultured hepatocytes with SIRT3 overexpression

SIRT3 deletion or knockdown in liver has been associated with increased liver triglycerides and lipotoxicity (Bao, Scott et al. 2010, Hirschey, Shimazu et al. 2011). In this overexpression study we investigated if SIRT3 overexpression ex vivo would reduce liver triglyceride accumulation through its reported action on fatty acid oxidation pathway enzymes such as LCAD. Figure 4.7 shows changes in liver triglyceride accumulation with SIRT3 overexpression and SIRT3KO following overnight treatment with either vehicle or fatty acid. As expected, fatty acid treatment caused more than a 2-fold increase in triglyceride accumulation in primary hepatocytes in both SIRT3 overexpression and SIRT3KO models. SIRT3 overexpression in vehicle treated cells resulted in a significant 50% reduction in triglyceride (Figure 4.7A). In hepatocytes from SIRT3KO animals, SIRT3 deletion caused a significant increase in triglyceride by 2-way ANOVA (Figure 4.7B), corroborating the published finding that SIRT3KO mice have increased liver triglyceride (Hirschey, Shimazu et al. 2010, Hirschey, Shimazu et al. 2011).

4.3.2.5 Substrate utilisation in primary hepatocytes after SIRT3 HTVI Given the increase in the oxygen consumption rate in SIRT3 overexpressing hepatocytes, and the decrease in liver triglyceride accumulation, we performed substrate utilisation assays for both glucose and palmitate to determine if oxidation of specific substrates might be driving these outcomes (Figure 4.8). There were no significant differences in insulin-stimulated glucose oxidation with SIRT3 overexpression in isolated hepatocytes. Palmitate oxidation in primary hepatocytes increased significantly in response to overnight fatty acid treatment, but SIRT3 overexpression had no effect on palmitate utilisation under these cell culture conditions.

97 CHAPTER 4: Liver Overexpression

A 400 +CON † +SIRT3 300

200 * Triglyceride 100 (% (% change / CON)

0 VEH FA

B 300 WT † ‡ SIRT3KO

200

100 Triglyceride (% (% change / CON)

0 VEH FA

Figure 4.7 Triglyceride accumulation in vehicle and fatty acid treated hepatocytes from SIRT3 overexpressing liver: A. Triglyceride accumulation in CON (blue bars) and SIRT3 overexpressing (red bars) in primary hepatocytes following overnight vehicle (VEH) or fatty acid treatment (FA) expressed as percentage change. Data are mean± SEM percentage increase in nmol/mg cells over CON vehicle treated cells from n= 5 - 6 animals, all treatments done in duplicate wells over 4 independent experiments. B. Triglyceride accumulation in WT (blue shaded) and SIRT3KO (red shaded) in primary hepatocytes following overnight vehicle (VEH) or fatty acid treatment (FA) expressed as percentage change. Data are mean± SEM percentage increase in nmol/mg cells over WT vehicle treated cells from n= 3 - 4 animals, all treatments done in duplicate wells over 2 independent experiments. †p<0.005 main effect of FA by 2-way ANOVA; ‡p<0.001 by 2-way ANOVA main effect of SIRT3 deletion; *p=0.002 CON cf. SIRT3 unpaired t-test.

98 CHAPTER 4: Liver Overexpression

A B 2.5 2.5 † +CON +SIRT3 2.0 2.0

1.5 1.5

1.0 1.0

0.5 cells nmol/h/mg 0.5 Palmitate OxidtionPalmitate nmol/hr/mg cells nmol/hr/mg Glucose Oxidation Glucose

0.0 0.0 VEH FA VEH FA

Figure 4.8 Substrate utilisation in SIRT3 overexpressing hepatocytes with and without fatty acid treatment: 14 A. Insulin-stimulated C-glucose oxidation to CO2 in control and SIRT3 overexpressing hepatocytes. Data are mean± SEM, n=4-5. B. 14C-palmitate oxidation in control and SIRT3 overexpressing hepatocytes. Data are mean ± SEM, n=6-7. All stimulations and treatments prepared in duplicate wells. †p<0.0001 main effect of FA treatment by 2-way ANOVA.

In summary, hepatocytes isolated from mice overexpressing SIRT3 in liver have increased oxygen consumption rates and accumulate less lipid than control animals when exposed to a high fat environment, however there was no detectable difference in glucose or fatty acid oxidation detectable in these cells. Interestingly, while SIRT3KO hepatocytes showed increased lipid accumulation compared to WT cells, no major defect was seen in oxygen consumption in SIRT3KO cells.

4.3.3 Acute SIRT3 overexpression in liver of chow and HFD-fed mice

The previous sections have shown that SIRT3 overexpression in liver has effects on triglyceride accumulation and oxygen consumption when hepatocytes are isolated and assessed in vitro. To ascertain if these effects observed in hepatocytes in culture are indicative of changes in metabolism in vivo we carried out a series of in vivo SIRT3 overexpression studies. The first cohort underwent an acute 3 week overexpression of SIRT3 in the liver with either chow or HFD-feeding to induce the metabolic defects associated with excess lipid availability.

99 CHAPTER 4: Liver Overexpression

4.3.3.1 SIRT3 overexpression in vivo in liver at 3 weeks by western blot To look at the initial effect of SIRT3 overexpression during high fat feeding, a cohort of animals was followed for 3-4 weeks post-HTVI. The level of SIRT3 at 3 weeks was increased on average 2.6 ± 0.4 -fold compared to control animals (Figure 4.9).

CHOW HFD CON +SIRT3 CON +SIRT3 hSIRT3-FLAG 30 kDa endogenous SIRT3 28 kDa GAPDH 37 kDa

CON 2.5 SIRT3 2.0 1.5 1.0

CON (A.U.) 0.5 SIRT3 SIRT3 Protein/ 0.0 CHOW HFD

Figure 4.9 SIRT3 overexpression 3 weeks after HTVI: Upper panel shows western blot of whole liver lysates from animals that underwent either empty vector (CON) or SIRT3-FLAG (+SIRT3) HTVI and then were placed on chow HFD for 3 weeks, using an antibody that picks up both mouse and human SIRT3. Lower panel shows average increase in densitometry as SIRT3 liver over control liver from 3 replicate blots. Data are mean ±SEM, n=5-6 per group.

4.3.3.2 Metabolic characteristics and weight gain in acute cohort Parameters for the effects of 3 week overexpression of SIRT3 in liver of mice can be seen in Table 4.2. HFD feeding for 3 weeks caused a significant increase in total body weight, and a significant decrease in liver weight. There was no significant effect of HFD on plasma triglycerides or plasma NEFAs by 2-way ANOVA. SIRT3 overexpression did not significantly influence any of these parameters except for liver weight which did not decrease with HFD in the SIRT3 group to the same extent as in control HFD-fed group.

100 CHAPTER 4: Liver Overexpression

Figure 4.10 displays the data for the gross metabolic phenotype of mice overexpressing SIRT3 in the liver for 3 weeks. There was no significant divergence in body weight over the three weeks of the study, although importantly, the HTVI technique itself did not affect the increase in body weight of the mice (Figure 4.10A). Both visceral and subcutaneous white adipose tissue (WAT) depots showed significant increases in size following 3 weeks of HFD in agreement with earlier studies, however despite the ability of SIRT3 to increase substrate oxidation in vitro, there was no effect on the size of the WAT. Similarly, liver triglyceride was significantly increased with HFD-feeding, however no improvement in liver triglyceride was seen with SIRT3 overexpression.

Table 4.2 Physiological effects of 3 weeks overexpression of SIRT3 in liver in vivo in mice p- CON SIRT3-FLAG value CHOW HFD CHOW HFD n 5 7 5 5 -

Body Weight (g) 25.6 ± 0.7 26.8 ± 0.8 25.4 ± 0.6 26.7 ± 0.7 0.03*

Liver (% BW) 4.8 ± 0.2 3.9 ± 0.1 4.8 ± 0.1 4.8 ± 0.2 0.02#

Liver Weight (g) 1.24± 0.09 1.06± 0.06 1.23± 0.04 1.13± 0.05 0.03*

Plasma TG (mM) 0.79 ± 0.06 1.58 ± 0.4 1.57 ± 0.4 1.70 ± 0.4 ns

Plasma NEFA 0.83 ± 0.06 1.04 ± 0.2 0.86 ± 0.05 0.69 ± 0.05 ns (mM)

BW: body weight. TG: triglyceride. NEFA: non-esterified fatty acids. Data are ±SEM *Diet effect 2-way ANOVA, # SIRT3 effect by 2-way ANOVA, effect of diet and SIRT3 p = 0.02.

101 CHAPTER 4: Liver Overexpression

A 50 B +CON 3 † +SIRT3 40

30 2 ns 20

visWAT 1

Body Weight (g) 10

0 (% of Body Weight) 0 1 2 3 0 CHOW HFD Weeks post-HTVI CON SIRT3

C D 2.0 25 † † 20 1.5 ns 15 1.0 10 SubQ WAT

0.5 g) per (µmol 5 (% (% Body Weight ) Liver Triglyceride Liver 0.0 0 CHOW HFD CHOW HFD

Figure 4.10 Effect of 3 weeks of overexpression of SIRT3 and HFD on body weight, fat pad weight and liver triglycerides: A: Body weight over time: Mice were weighed at least weekly following administration of HTVI to ensure no ill effects of procedure and initiation of HFD. Blue solid lines denote chow fed animals receiving a control injection, while blue dotted lines denote HFD-fed control injected animals. Red solid line and red dotted lines denote SIRT3-FLAG injected animals from both chow and HFD groups respectively. B-D. Control HTVI in blue and SIRT3 HTVI in red following 3 weeks of either chow or HFD. B. Visceral WAT weight as a percentage of body weight from epididymal depot. C. Subcutaneous WAT weight as a percentage of body weight from inguinal depot. D. Liver triglyceride. †p<0.005, effect of diet by 2-way ANOVA, ns = non-significant. Data are mean ±SEM, n = 5 - 6 animals per group.

102 CHAPTER 4: Liver Overexpression

4.3.3.3 SIRT3 overexpression and glucose tolerance Global SIRT3 deletion has been shown previously to have effects on glucose tolerance (Hirschey, Shimazu et al. 2011). Figure 4.11 shows the glucose tolerance curves and incremental area under the curve for animals after overexpression of SIRT3 in the liver for 3 weeks. While 3 weeks of HFD had a highly deleterious effect on glucose tolerance in these mice, there was no effect of SIRT3 overexpression on this measure.

A B 40 1500 *** HFD +CON 30 +SIRT3 1000

20 CHOW iAUC (mmol/L) 500 10 Blood Glucose during ipGTT 0 0 0 20 40 60 80 100 CHOW HFD Time (min) CON SIRT3 Figure 4.11 Effect of 3 weeks overexpression of SIRT3 in liver on glucose tolerance in chow and HFD-fed mice: A. Glucose excursion during ipGTT following glucose administration. Blue solid lines denote chow fed control (CON) animals, while blue dotted lines denote HFD-fed CON animals. Red solid line and red dotted lines denote SIRT3 overexpressing animals from both chow and HFD groups respectively. B. Incremental area under the curve from the ipGTT shown in (A). Data are mean ±SEM, n = 5 - 6 animals per group. ***p<0.0001 by 2-way ANOVA effect of diet only.

4.3.3.4 SIRT3 overexpression in liver for 3 weeks increases substrate utilisation in ex vivo liver homogenates. SIRT3 overexpression increased mitochondrial respiration in isolated hepatocytes. Another in vitro method to examine mitochondrial oxidative capacity is to measure substrate oxidation in freshly prepared tissue homogenates. Liver tissue from the acute 3 week overexpression study was assessed for both pyruvate and glutamate oxidation (Figure 4.12). In agreement with the general trends seen in isolated hepatocytes, pyruvate oxidation was higher in liver homogenates overexpressing SIRT3

103 CHAPTER 4: Liver Overexpression

compared to control animals, in both chow and HFD-fed mice, although only the chow group approached significance using a post-hoc test. In contrast, glutamate oxidation in liver homogenates was significantly lower in HFD mice compared to chow-fed, however there was no significant effect of SIRT3 overexpression on glutamate oxidation.

p=0.8 † 5 ‡ 10 +CON +SIRT3 4 8 3 6 2 4 (umol/g/hr) (umol/g/hr) 1 2 Pyruvate Oxidation Pyruvate Glutamate Oxidation Glutamate 0 0 CHOW HFD CHOW HFD

Figure 4.12 Substrate utilisation in liver homogenates of animals overexpressing SIRT3 for 3 weeks: Blue panel denotes empty vector (CON) HTVI animals and red bars denote SIRT3 overexpressing animals. Left panel shows pyruvate oxidation rate in liver homogenates from chow and HFD fed mice. Right panel shows glutamate oxidation rate in liver homogenates from chow and HFD-fed mice. Data are mean ±SEM, n = 5 – 6 animals per group, ‡p=0.015 effect of SIRT3 by 2-way ANOVA,. †p=0.03 effect of diet by 2-way ANOVA, no effect of SIRT3.

In summary, SIRT3 overexpression in the liver over an acute timeframe of 3 weeks appeared to show little in vivo effect on any of the metabolic parameters measured. Although substrate utilisation was increased in liver homogenates in line with the decrease in triglyceride accumulation seen in the primary hepatocyte model reported earlier, there was no effect on body weight, fat pad weights or glucose tolerance at this acute timeframe, or effects in the context of lipid oversupply.

104 CHAPTER 4: Liver Overexpression

4.3.4 48 hour fasting in an acute SIRT3 overexpression model

As an NAD+ dependent deacetylase, it is thought that SIRT3 activity can respond dynamically to changes in cellular nutrient status. NAD+/NADH ratio and acetylation have both been shown to increase in response to fasting and caloric restriction (Yang, Vaitheesvaran et al. 2011, Hebert, Dittenhafer-Reed et al. 2013), with NAD+ levels in the fasted liver reported to be 50% higher than control levels (Rodgers, Lerin et al. 2005). SIRT3KO mice were also shown to have defects in fatty acid oxidation in the liver, however this was only evident during significant stress such as during fasting and cold exposure (Hirschey, Shimazu et al. 2010). Here, the effects of acute liver SIRT3 overexpression in the context of 48 hour fasting in mice was investigated.

4.3.4.1 Effect of 48 h fasting on body weight, tissue weights and serum parameters Basic characteristics of this fasted cohort are given in Table 4.3. As expected, long-term fasting of 48 hours caused weight loss, and a reduction in serum markers of nutrient status such as blood glucose and plasma insulin. Subcutaneous fat pad weights showed a significant reduction with fasting, although visceral fat pad weight was not affected. SIRT3 overexpression in the liver did not modulate this decrease in white adipose tissue mass. Although serum triglycerides and NEFAs are often reported to increase with fasting as fat stores are mobilized for fuel (Dole 1956, Rodgers and Puigserver 2007), at this long-term fasting level we observed a reduction in serum triglycerides, with little change in NEFA levels. SIRT3 overexpression had no significant effect on these parameters in the context of long-term fasting.

105 CHAPTER 4: Liver Overexpression

Table 4.3 The effects of 3 week SIRT3 overexpression on metabolic parameters in fed versus 48hr fasted mice CON SIRT3-FLAG p-value

FED FAST FED FAST

n 4 6 5 5 -

Initial BW (g) 30 ± 0.7 28 ± 0.4 28 ± 0.3 30 ± 0.6 ns

Final BW (g) 30 ± 0.8 23 ± 0.6 28 ± 0.5 24 ± 0.7 <0.001*

Liver (% BW) 4.8 ± 0.3 3.8 ± 0.2 4.9 ± 0.4 3.4 ± 0.3 0.0003*

VisWAT (% BW) 0.99 ± 0.07 0.92 ± 0.02 1.1 ± 0.2 0.70 ± 0.2 ns

SubWAT (% BW) 0.39 ± 0.02 0.29 ± 0.03 0.41 ± 0.06 0.24 ± 0.05 0.009*

Initial BG (mM) 8.7 ± 0.8 9.4 ± 0.7 9.2 ± 0.4 10.0 ± 0.6 ns

Fasted BG (mM) 8.5 ± 0.5 7.1 ± 0.3 9.3 ± 0.4 5.8 ± 0.4 <0.001*

Final Plasma 0.76 ± 0.1 0.4 ± 0.04 0.63 ± 0.3 0.3 ± 0.07 0.03* Insulin (ng/ml) Final Plasma TG 1.27 ± 0.10 0.86 ± 0.06 1.34 ± 0.2 0.96 ± 0.04 <0.001* (mM) Final Plasma NEFA 0.73 ± 0.1 0.58 ± 0.1 0.66 ± 0.1 0.71 ± 0.1 ns (mM) Final and Initial refer to pre-fasting measurements, and measurements taken following either a 48 hour fast or fed controls. BW: body weight. BG: blood glucose. VisWAT: visceral white adipose tissue from epididymal depot. SubWAT: subcutaneous white adipose tissue from inguinal depot. TG: triglyceride. NEFA: non- esterified fatty acids. Data are Mean ± SEM *Significant: 2-way ANOVA fasting effect only

106 CHAPTER 4: Liver Overexpression

4.3.4.2 Oxygen consumption in isolated mitochondria from SIRT3 overexpressing liver following 48 h fast To assess whether fasting had an effect on the ability of SIRT3 overexpression to increase mitochondrial metabolism, mitochondria were isolated from liver and oxygen consumption was measured using a Clark-type electrode. Figure 4.13 shows ADP- stimulated respiration in mitochondria using both succinate and palmitoyl carnitine as substrates. There was no detectable difference between SIRT3 overexpressing and CON animals in succinate respiration, nor was the increase in substrate oxidation with fasting significant. However, using the fatty acid substrate palmitoyl carnitine, there was a significant increase in oxygen consumption with SIRT3 overexpression compared to control mitochondria in the fasting state.

+CON +SIRT3 300 150 ‡ * 200 100 CONSUMPTION 2 CONSUMPTION CONSUMPTION 2 100 50 nmol O/min/mg protein O/min/mg nmol nmol O/min/mg protein O/min/mg nmol 0 0 FED FASTED FED FASTED SUCCINATE O SUCCINATE PALMITOYL C. O

Figure 4.13 The effect of SIRT3 overexpression on oxygen consumption in isolated liver mitochondria from 48 h fasted mice: Left panel shows ADP-stimulated respiration rate with succinate as substrate. Right panel shows ADP-stimulated respiration rate with palmitoyl carnitine as substrate. Data are mean ± SEM, n = 4-6 animals per group , ‡p=0.013, main effect of SIRT3 cf. CON by 2 way-ANOVA; *p=0.03 SIRT3 cf. CON Tukey post-hoc test.

107 CHAPTER 4: Liver Overexpression

4.3.4.3 Ketone bodies and SIRT3 overexpression during fasting Ketone bodies are made in the liver to provide fuel for extra-hepatic tissues during periods of fasting. SIRT3KO mice have previously been shown to have reduced levels of the ketone β-hydroxybutyrate (β-OHB) compared to WT controls (Shimazu, Hirschey et al. 2010). Mice overexpressing SIRT3 for 3 weeks were assessed for levels of β-OHB in the blood following fasting (Figure 4.14). While β-OHB was sharply elevated with 48 hours fasting in all animals, there was a significant elevation in this ketone in SIRT3 overexpressing animals compared to similarly fasted controls.

1.5 +CON * +SIRT3 1.0

0.5 -OHB -OHB (mmol/L) β

0.0 FED FASTED

Figure 4.14 Levels of beta hydroxybutyrate (β-OHB) in SIRT3 overexpressing mice following a 48 h fast: Levels of β-OHB in 3 week SIRT3 liver overexpressing mice that were either fed or fasted for 48 hours prior to blood collection, n = 9 animals per group. All data are mean ±SEM, *p<0.05 SIRT3 cf. CON unpaired t-test.

In summary, acute SIRT3 overexpression in the liver under the conditions of long-term fasting did not affect the majority of the metabolic parameters measured such as tissue weights or serum measures. However, SIRT3 overexpression in this fasted cohort did increase mitochondrial respiration in the presence of fatty acid as a substrate. SIRT3 overexpression also increased the production of the ketone β-OHB following a long-term fast.

108 CHAPTER 4: Liver Overexpression

4.3.5 Long-term overexpression of SIRT3 in vivo

Despite promising results in isolated hepatocytes, SIRT3 overexpression in the liver for 3 weeks did not influence the excess weight gain and liver triglyceride accumulation associated with feeding a HFD to mice. Hirschey et al demonstrated that SIRT3 protein levels decreased with more prolonged high-fat feeding (Hirschey, Shimazu et al. 2011) and therefore the timeframe of HFD intake was increased to 13 weeks to determine whether SIRT3 overexpression in liver resulted in beneficial effects at this timepoint.

4.3.5.1 SIRT3 overexpression by western blot in 13 week cohort Figure 4.15 shows a representative blot of SIRT3 overexpression in mice 13 weeks after chow they had received the SIRT3 construct via HTVI. At this more extended timepoint the hfd SIRT3 overexpression was not equally robust in all animals, as can be seen in lane 1.

CHOW HFD +SIRT3 CON +SIRT3 CON hSIRT3-FLAG 30 kDa endogenous SIRT3 28 kDa

2.0 +CON 1.5 +SIRT3

1.0

CON (A.U.) 0.5 SIRT3 SIRT3 Protein/

0.0 CHOW HFD

Figure 4.15 SIRT3 overexpression at protein level 13 weeks post-HTVI in chow and HFD-fed mice: Upper panel shows representative western blot of whole liver lysates from chow and HFD-fed mice that underwent HTVI with either empty vector (CON) or SIRT3-FLAG (+SIRT3) for 13 weeks. Lower panel shows densitometry from all animals included in cohort normalised to densitometry from CON chow-fed animals (adjusted to 1). Data are mean ±SEM, n = 5 - 6 animals per group.

109 CHAPTER 4: Liver Overexpression

Successful transduction of the liver was shown at 1 week using pLuc-pLIVE in this same cohort. Therefore the lack of SIRT3 protein in all liver samples is likely due to a loss of SIRT3 overexpression over the 13 weeks, rather than a failure of the original HTVI. Only animals that showed at least a 40% increase in SIRT3 protein were used in subsequent analysis, with the range of protein overexpression ranging from 40% to a 108% increase in SIRT3 protein level.

4.3.5.2 Effect of long-term overexpression of SIRT3 on body and tissue weights, fat distribution and liver triglyceride Table 4.4 shows characteristics of the mice that underwent 13 weeks of control or SIRT3 overexpression, excluding those that did not show sufficient overexpression by western blot. Percentage fat by dual energy X-ray absorptiometry (DEXA) was assessed after 10 weeks of HFD-feeding and showed a significant increase. Similarly, both visceral and subcutaneous fat pad weights were significantly increased with HFD- feeding compared to chow controls. Interestingly, of all these measures only liver weight showed a significant difference between CON and SIRT3-FLAG, although the magnitude of this difference was small.

As shown in Figure 4.16, SIRT3 overexpression and HFD-feeding had no significant effect on body weight gain over time in this cohort of mice, with HFD-fed mice only ~1g heavier than their chow fed controls after 13 weeks. We have shown that SIRT3 overexpression can reduce triglyceride content in isolated hepatocytes. In this long- term study the effect of SIRT3 on liver triglyceride was assessed. While there was an approximate 2-fold increase in liver triglyceride with HFD-feeding, there was no significant effect on liver triglyceride with SIRT3 overexpression (Figure 4.16B).

110 CHAPTER 4: Liver Overexpression

Table 4.4 Effects of SIRT3 overexpression on physiological parameters after 13 weeks in chow and HFD-fed mice CON SIRT3-FLAG p-value CHOW HFD CHOW HFD n 6 6 5 6 -

Body Weight (g) 28 ± 1.04 29.5 ± 0.8 28.4 ± 0.6 29.4 ± 0.9 ns

0.001* / Liver Weight (g) 1.2± 0.05 1.0± 0.05 1.4± 0.05 1.1± 0.05 0.02#

0.0004*/ Liver (% BW) 4.2 ± 0.2 3.5 ± 0.2 4.8 ± 0.2 3.8 ± 0.2 0.04#

VisWAT (% BW) 1.0 ± 0.1 2.3 ± 0.2 1.3 ± 0.4 2.5 ± 0.5 0.0016*

SubWAT (% BW) 0.4 ± 0.04 0.8 ± 0.07 0.5 ± 0.04 0.9 ± 0.1 <0.0001*

%Fat by DEXA at 13.1 ± 0.8 21.0 ± 1.2 13.1 ± 1.1 22.9 ± 2.5 <0.0001* 10wk BW: body weight; VisWAT: visceral white adipose tissue from epididymal depot; SubWAT: subcutaneous white adipose tissue from inguinal depot; ns: non significant. Data are mean ±SEM. *Significant effect of diet and # significant effect of SIRT3, by 2-way ANOVA.

111 CHAPTER 4: Liver Overexpression

A 35

Body weight (g) 20

15 2 3 4 5 6 7 8 9 10 11 13 Weeks post-HTVI CON SIRT3

B 20 † CON ns SIRT3

10 (nmol/mg) 5 Liver Triglyceride

0 CHOW HFD

Figure 4.16 Effect of SIRT3 overexpression and HFD on body weight and liver triglyceride in long-term cohort: A: Body weight over time: Mice were weighed at least weekly following HTVI with control or SIRT3 constructs. Control (blue) and SIRT3 overexpressing (red) following 13 weeks of either chow (solid line) or HFD (dotted line). B. Liver triglyceride after 13 weeks liver overexpression of control (blue bars) SIRT3 (red bars). Data are mean ±SEM, n = 5 - 6 animals per group. †p<0.0001, main effect of diet by 2-way ANOVA, ns = non-significant.

4.3.5.3 Long-term SIRT3 overexpression in mice and effects on glucose tolerance Figure 4.17 shows glucose tolerance and insulin response during the ipGTT in the longer-term overexpression cohort at 12 weeks. Although HFD significantly impacted on glucose tolerance, even after a much longer exposure to SIRT3 overexpression, there remained no difference between SIRT3 overexpressing and CON liver animals in terms of glucose excursion during the GTT, as indicated by the area under the curve (Figure 4.17B). Control HFD mice showed significant hyperinsulinemia during the GTT compared to control chow fed mice. Interestingly, despite HFD-fed SIRT3

112 CHAPTER 4: Liver Overexpression

overexpressing animals showing equally impaired glucose tolerance, SIRT3 overexpressing animals showed lower insulin at basal, and during the GTT compared to control HFD mice, reaching significance at the 30 minute timepoint (Figure 4.17C).

+CON 40 1500 A B † +SIRT3 * * 30 * 1000 ‡ ns 20

ipGTT AUC 500 10 Blood Glucose (mmol/L)

0 0 0 20 40 60 80 100 CHOW HFD Time (min) CON SIRT3 FLAG

C 2.0 ‡ #

1.5

1.0

0.5 ipGTT Insulin (ng/ml)

0.0 0 20 40 60 80 100 Time (min) CON SIRT3 FLAG

Figure 4.17 Effect of SIRT3 overexpression and HFD on glucose tolerance after 12 weeks of HFD: A. Response to i.p. glucose injection over time (ipGTT) at 12 weeks post-HTVI: Control animals (Blue lines) and SIRT3 overexpressing animals (red lines) under both chow (solid line) and HFD- fed (dotted line) conditions. B. Area under the curve for glucose during ipGTT . C. Insulin levels during ipGTT. Data are mean ±SEM, n = 5-6 animals per group. *p<0.005, †p=0.006 , main effect of diet by 2-way ANOVA. ‡p<0.0001 effect of CHOW cf. HFD in SIRT3 animals by 2-way ANOVA (Tukey’s post-hoc test). #p<0.0001 effect of SIRT3-HFD cf. CON-HFD by 2 way ANOVA (Tukey’s post-hoc test). ns= not significant.

113 CHAPTER 4: Liver Overexpression

4.3.5.4 Mitochondrial substrate oxidation in isolated mitochondria in long-term SIRT3 overexpression Oxidation of succinate, glutamate and palmitoyl carnitine substrates in isolated liver mitochondria from mice that received either empty plasmid or SIRT3 HTVI are displayed in Figure 4.18. Other experiments in this thesis have shown that SIRT3 overexpression increased substrate oxidation both in isolated hepatocytes, in liver homogenates in the acute 3 week cohort, and in mitochondria isolated from 48 hour fasted liver tissue. In this long-term overexpression study, 13 weeks of HFD had no effect on oxygen consumption except in the case of the fatty acid substrate, palmitoyl carnitine, where HFD caused a small but significant decrease in oxidation of this substrate (Figure 4.18C). SIRT3 overexpression, however, had no effect on mitochondrial oxygen consumption under any of the conditions tested.

A B +CON +SIRT3 200 200 CONSUMPTION CONSUMPTION 2 2 100 100 nmol O/min/mg protein nmol O/min/mg protein 0 0 SUCCINATE O SUCCINATE CHOW HFD O GLUTAMATE CHOW HFD

C † 100 CONSUMPTION 2 50 nmol O/min/mg protein 0

PALMITOYL C. O CHOW HFD

Figure 4.18 Substrate oxidation in isolated liver mitochondria after 13 weeks SIRT3 overexpression in both chow and HFD: A. ADP-stimulated respiration rate with succinate as substrate. B. ADP-stimulated respiration rate with glutamate as substrate. C. ADP-stimulated respiration rate with palmitoyl-carnitine as substrate. Data are mean ±SEM, n = 5-6 animals per group , †p-0.04 effect of diet by 2 way- ANOVA.

114 CHAPTER 4: Liver Overexpression

In summary, longer-term overexpression of SIRT3 in vivo in the livers of mice was unable to reduce the deleterious effects of HFD-feeding in terms of body and WAT weights, liver triglyceride levels, or glucose tolerance. However, SIRT3 overexpressing mice showed lower insulin levels after HFD than their control littermates, despite showing similar glucose intolerance.

4.4 Discussion

The aim of the studies reported in this Chapter was to overexpress SIRT3 in the liver of mice, and investigate the effects on mitochondrial metabolism. Because global SIRT3 deletion is reported to be associated with a generally worsened metabolic phenotype (Hirschey, Shimazu et al. 2011, Paulin, Dromparis et al. 2014, Winnik, Gaul et al. 2014), particularly under the added stress of high fat diet, we also aimed to determine if this overexpression in liver could be protective against metabolic defects induced by excess lipid.

In addition to an ex vivo model in isolated hepatocytes, we also employed in vivo models to investigate the effects of SIRT3 overexpression in liver in: i) an acute 3 week timeframe, ii) under fasting conditions, and iii) in a long-term 13 week overexpression model. A summary of the main findings of the liver overexpression model is presented in Table 4.5.

The ex vivo model in isolated hepatocytes showed 2 main effects of SIRT3 overexpression: An increase in oxygen consumption and a decrease in triglyceride accumulation. Similarly, in liver homogenates from 3 week SIRT3 overexpressing animals an increase in pyruvate oxidation was also shown. A similar ex vivo system, isolated mitochondria from SIRT3 overexpressing liver, was assessed for oxygen consumption and also showed an increase in mitochondrial respiration using a fatty acid substrate, palmitoyl carnitine, however this effect was only seen in fasted animals.

115 CHAPTER 4: Liver Overexpression

Table 4.5 Summary of liver SIRT3 overexpression experiments ! SIRT3& OE& in& Acute& SIRT3& Fasted& SIRT3& Long;term& Isolated& OE OE SIRT3&OE Hepatocytes

Overexpression! ✓! ✓! ✓! ✓!

ΔBody!Wt.! NA! No!change!

ΔFat!Pad!Wt.! NA! No!change!

ΔLiver!Wt.! NA! !! No!change! !!

ΔLiver! "TG! No!change! triglyceride! No!change!in! Δ! Glucose! No!change!in! GTT! NA! @! Tolerance! GTT! Lower!Insulin! during!GTT! ! Δ!Ketones! NA! @! !!β@OHB! @! ! No!effect!on! No!effect!on! Oxygen! ! O ! succinate! succinate,! 2 @! Consumption! consumption! ! FA! glutamate!or! respiration FA! No!effect!on!!CO2! ! Pyruvate! Substrate! production!with! oxidation!in! @! @! Oxidation! glucose!!or!FA! liver! oxidation! homogenates NA!=!not!applicable,!TG!=!triglyceride,!FA=!fatty!acid! !

SIRT3 overexpression caused an increase in both basal and maximal oxygen consumption in isolated hepatocytes, supporting the proposed role of SIRT3 as a deacetylase that can increase the activity of a variety of mitochondrial enzymes including those of the oxidative phosphorylation (for a detailed list see Table 1.1, Chapter 1). The Seahorse flux analyser has recently emerged as a method of choice to assess mitochondrial function in both isolated mitochondria and intact cells (Rogers, Brand et al. 2011, de Moura and Van Houten 2014). Measurements of oligomycin- sensitive and insensitive respiration suggest that the increase in oxygen consumption in SIRT3 overexpressing cells is due to increases in ATP-linked respiration although

116 CHAPTER 4: Liver Overexpression

effects for these measures were not large enough to be significant. Very similar results have been reported in the literature in HEK293 cells stably transfected with SIRT3 (Barbi de Moura, Uppala et al. 2014), where an increase in basal respiration and proton leak was reported, although no mechanism was suggested. One caveat of these studies in intact cells is that normalisation methods only take account of protein content and/or cell number, hence it remains possible that changes in mitochondrial density may be underlying changes in oxygen consumption (Brand and Nicholls 2011). Although in this study SIRT3KO mice did not show a significant decrease in oxygen consumption compared to WT, the fact that there is a trend towards this opposing effect occurring with SIRT3 deletion provides additional support for this model of SIRT3 overexpression.

In a similar opposing pattern, SIRT3 overexpressing hepatocytes showed a decrease in triglyceride accumulation, while SIRT3KO hepatocytes showed an increase. A similar decrease with SIRT3 overexpression has been reported in the literature using a retrovirus to overexpress SIRT3 in HepG2 cells (Shi, Fan et al. 2010). HepG2 cells overexpressing SIRT3 showed no reduction in triglyceride levels under normal media conditions, but in the presence of 2 mM oleate for 24 hr they showed significantly lower fat (Shi, Fan et al. 2010). The results in this chapter showed SIRT3 effects on triglyceride under both vehicle and fatty acid conditions, in this physiologically relevant primary cell type.

Changing liver lipid levels are commonly associated with changes in fatty acid oxidation, i.e hepatic steatosis is associated with defects in fatty acid oxidation (Kurtz, Rinaldo et al. 1998). However in our isolated hepatocytes the changes in liver triglyceride were not correlated with an increase in fatty acid oxidation in isolated hepatocytes. Nevertheless an increase in palmitate-supported respiration was seen in liver mitochondria isolated from SIRT3 overexpressing mice following long-term fasting, a condition in which we expect to see maximal activation of SIRT3 activity (Yu and Auwerx 2009), suggesting that under the ideal conditions SIRT3 overexpression may be able to increase fatty acid oxidation. It should be noted in the context of these results that while oxygen consumption and CO2 production are often assumed to be

117 CHAPTER 4: Liver Overexpression

well coupled in mitochondrial oxidation experiments, they are different measurements. The finding that oxygen consumption is changed with SIRT3 overexpression in isolated hepatocytes without any effect on substrate oxidation to

CO2 could also be affected by differences in the flux in internal fuel storage and oxidiation rather than just the oxidation of the labelled fuel that was supplied in the media (Dyck and Bonen 1998). Although FA treatment did reduce oxygen consumption on its own compared to vehicle treated cells, it did not ablate the effect of SIRT3 overexpression. The mechanism for this reduced respiration with FA treatment was not specifically investigated, but fatty acids have previously been shown to increase mitochondrial respiration at lower doses than used in this study (Nobes, Hay et al. 1990). Conversely, high levels of palmitate have been shown to cause lipotoxicity, ER stress and cell death (Listenberger, Ory et al. 2001, Egnatchik, Leamy et al. 2014), however the addition of oleate has been shown to reverse this (Noguchi, Young et al. 2009). Results are normalised to protein content, however cell viability assays were not performed following the seahorse analysis hence palmitate-induced lipotoxicity and cell death is a possible mechanism for the reduction in respiration in these cells compared to vehicle treatment, although why this was enhanced in SIRT3 overexpressing cells remains to be determined.

Remarkably, the effects of SIRT3 overexpression in vivo yielded very few findings. In the complex regulatory environment that is whole-body metabolism, increasing protein levels of SIRT3 in the liver had no effect on whole body measures such as body weight, fat distribution or serum markers of glucose and fat metabolism. The metabolic health of the liver, particularly liver lipid, is a key determinant of whole-body glucose tolerance assessed by techniques such as the GTT (Fabbrini, Magkos et al. 2009). Glucose tolerance was not affected by SIRT3 overexpression in liver, although it did appear that SIRT3 overexpression may have had the effect of reducing insulin levels in HFD-fed mice following a glucose challenge.

SIRT3KO mice have been reported to become more obese on a HFD compared to WT mice, a difference ascribed to induction of liver lipogenic pathways (Hirschey, Shimazu et al. 2011). However, a significant difference in body weight in that study did not

118 CHAPTER 4: Liver Overexpression

occur until mice had been on HFD for 33 weeks, proving the effect on body weight to be a relatively mild phenotype (Hirschey, Shimazu et al. 2011). We are not the first to report that SIRT3 changes in liver do not affect metabolic phenotype, as liver specific SIRT3KO mice showed no difference in body weight, or the many other metabolic parameters assessed in that study (Fernandez-Marcos, Jeninga et al. 2012).

Liver triglyceride more than doubled with HFD-feeding in both short and long-term overexpression studies, however this increase was not ameliorated by increased SIRT3 protein. This goes against previous studies where Hirschey et al showed a 50% reduction in liver triglyceride levels in liver homogenates from WT mice that overexpressed SIRT3 from adenovirus for 1 week, although the extent of overexpression was never quantified (Hirschey, Shimazu et al. 2011). Importantly, only a modest increase in palmitate oxidation was seen in WT mice following adenovirus administration, which does not explain the large reduction in liver triglyceride seen in Hirschey and colleagues study (Hirschey, Shimazu et al. 2010). Intriguingly, in the present study SIRT3 overexpression appeared to protect mice from the decrease in liver weight seen in control animals in response to both short-term and longer-term HFD (Tables 4.2 and 4.4), but had no effect on the changes in liver weight associated with fasting (Table 4.3). The basis for these changes in liver weight seen in this model are not clear, although it occurs independent of the HTVI procedure itself (as control animals also undergo HTVI). We have reported previously that the diet used in our study has been reported to cause reductions in liver weight (Liu, Montgomery et al. 2014). In addition, while steatosis associated with HFD-diets can cause increases in liver weight (Ye, Tid-Ang et al. 2011,) this change in weight occurred despite no detectable difference in liver triglyceride between SIRT3 overexpressing and control animals. These changes in liver weight could be associated with changes in glycogen stores as glycogen accounts for up to 10% of liver weight (Berg, Tymoczko et al. 2002) and glycogen synthesis is impaired in response to high-fat stimulus in muscle (Oakes, Cooney et al. 1997), however this was not tested experimentally in this model. It remains possible that the small but significant effect of SIRT3 overexpression on changes in liver weight induced by HFD, but not by fasting, is the result of other

119 CHAPTER 4: Liver Overexpression

unforseen effects on liver proliferation, inflammation or water retention that may be modulated by SIRT3.

The only circulating measure that showed a difference with SIRT3 overexpression was the level of the ketone β-hydroxybutyrate (β-OHB). This is the first report of SIRT3 overexpression increasing blood ketone levels. β-OHB is one of the three major ketone bodies, the others being acetoacetate and acetone. However β-OHB has a relatively higher level in the blood than the others and accounts for approximately 70% of circulating ketones (Dedkova and Blatter 2014). Ketones are made in the liver mitochondria from fatty acid derived acetyl-coA, and increase in the blood during starvation, providing a blood borne source of fuel for other tissues such as the brain. A role for SIRT3 in the response to fasting has previously been reported. Specifically, SIRT3 has been reported to regulate the activity of the rate-limiting step of ketone body synthesis, 3-hydroxy-3-methylglutaryl CoA synthase 2 (HMGCS2), which is associated with reduced β-OHB levels in SIRT3KO mice (Shimazu, Hirschey et al. 2010), hence this result in SIRT3 overexpression validates this finding.

As is clear from Table 4.5, SIRT3 overexpression exerts its major effects when tested in an isolated or ex vivo system. Under these conditions, substrates and metabolites such as NAD+ are available to the mitochondria in amounts that are potentially much higher than that found under regulated cellular conditions. The challenges of teasing out the role of SIRT3 in the in vivo setting are important to point out. While SIRT3 may have the potential to have large effects on many different pathways due to its huge number of substrates, it is important to remember that whole body metabolism is in highly regulated homeostasis.

It remains possible that the level of SIRT3 overexpression achieved in these studies is not sufficient to effect changes in in vivo metabolism, particularly in the long-term overexpression cohort where exogenous protein levels were found to drop off considerably compared to the 3 week study. The HTVI method has been reported to maintain overexpression for up to 8 months protein (MirusBio), however this is

120 CHAPTER 4: Liver Overexpression

dependent on the exact model being investigated, and we found that the success of overexpression of SIRT3 in the longer-term cohorts was quite variable (even after we developed the bioluminescence method to confirm initial injection success). Other methods for longer-term overexpression studies such as adeno-associated viral vectors or inducible transgenic mice may need to be developed to more adequately attain significant overexpression at long-term timepoints. Another possible limitation of the longer-term study is that the weight gain in this particular cohort was quite minimal for 13-week high-fat feeding (Figure 4.16), with fat pad weights and liver triglyceride closely resembling what was seen after only 3 weeks of high fat feeding in the acute cohort. Variation between different cohorts of mice is a common problem and repeating this longer-term study in a larger cohort of mice to ensure a greater metabolic defect may be warranted for future experiments.

Another caveat of these studies remains the inability to adequately assess levels of SIRT3 activity in these overexpression studies, as the increase in SIRT3 protein levels does not solely account for deacetylase activity. Chapter 5 will investigate the effect of SIRT3 overexpression in liver using mass spectrometry techniques to see if this increase in protein level is causing acetylation changes, as we did not have success applying commercially available sirtuin activity kits to measurements in vivo in liver tissue.

In conclusion, despite the effects seen in isolated hepatocytes, in vivo SIRT3 overexpressing mice displayed similar impairments as control animals in glucose tolerance and triglyceride accumulation in response to HFD and fasting. These results suggest that overexpression of SIRT3 in mouse liver in the whole body setting is not protective against hepatic lipid accumulation and glucose intolerance induced by HFD.

121 CHAPTER 5: MS Analysis

CHAPTER 5 MASS SPECTROMETRY ANALYSIS OF LYSINE ACETYLATION IN A SIRT3 OVEREXPRESSION MODEL IN LIVER

5.1 Introduction

Chapter 4 investigated the effect of overexpression of SIRT3 on mitochondrial metabolism in liver. SIRT3 overexpression enhanced respiration and reduced TAG accumulation, however in vivo this overexpression did not attenuate the negative events associated with HFD feeding. One possible explanation for these disparate effects is that the increase in SIRT3 levels was not sufficient to alter acetylation of target proteins in an in vivo context. To investigate this possibility we used mass spectrometry (MS) to study the acetylation of mitochondrial proteins in the liver SIRT3 overexpression model.

MS studies have shown that acetylation is a post-translational modification that has particular relevance in mitochondrial metabolism (Kim, Sprung et al. 2006, Choudhary, Kumar et al. 2009, Zhao, Xu et al. 2010). Advances in MS techniques including the use of antibody enrichment strategies, have demonstrated that more than 4000 mammalian proteins are acetylated, with a high proportion of mitochondrial proteins presenting with this PTM (Kim, Sprung et al. 2006, Choudhary, Kumar et al. 2009, Wang, Zhang et al. 2010, Zhao, Xu et al. 2010, Henriksen, Wagner et al. 2012, Lundby, Lage et al. 2012). The importance of this PTM is reflected by the fact that acetylation has been shown to regulate a variety of different cellular processes including apoptosis and the cell cycle, ageing, antioxidant defenses, tumour progression, circadian rhythms, gene expression, and intermediary metabolism (Saunders and Verdin 2007, Choudhary, Kumar et al. 2009, Guan, Yu et al. 2010, Verdin, Hirschey et al. 2010, Finley and Haigis 2012).

Effective liquid chromatography mass spectrometry (LC-MS) methods have been developed to assess the relative quantification of acetylation. Regardless of whether

122 CHAPTER 5: MS Analysis

isobaric-tag labelling (Hebert, Dittenhafer-Reed et al. 2013, Still, Floyd et al. 2013) or label-free MS techniques (Lundby, Lage et al. 2012, Rardin, Newman et al. 2013) have been used, most lysine acetylation studies make use of antibody enrichment strategies to enrich for the acetyl modification prior to analysing samples. Kim et al. were the first to employ immunoaffinity enrichment for the study of acetylation with other greater coverage acetylomes quickly employing the same strategy (Kim, Sprung et al. 2006, Choudhary, Kumar et al. 2009, Zhao, Xu et al. 2010). In this method, trypsin digested peptides are enriched in the sample via immunoprecipitation with the AcK antibody, and the acetyl-containing peptides eluted before running the MS. Since MS detection is dependent on abundance, the enrichment strategy allows greater coverage of lower-abundance acetylation modifications against a background of all the non-acetylated peptides in the sample. Choudhary et al report that a replicate experiment without affinity enrichment showed the number of acetylation sites to be 60-fold lower (Choudhary, Kumar et al. 2009). These MS methods are crucial for systems biology approaches and for proteome analyses where the goal is to find new substrates or new central nodes of interacting pathways. However, due to the current limitations of western blot and the lack of site-specific acetyl-lysine antibodies for SIRT3 targets, MS analysis is also critical for determining the functional relevance of SIRT3 overexpression or deletion.

In direct investigation of SIRT3 influence on the acetylome, several studies have made use of SIRT3KO mouse tissues for proteomic analyses. Hebert et al. conducted studies in liver looking at the effect of calorie restriction (CR) in both WT and SIRT3KO liver tissue, finding that SIRT3 was a major regulator of mitochondrial metabolic proteins in response to CR and detailing over 3000 acetylation sites in the process (Hebert, Dittenhafer-Reed et al. 2013). Fritz et al. used SIRT3KO mice to discover the regulatory role of SIRT3-dependent acetylation in a model of alcoholic liver disease where hyperacetylation was known to occur (Fritz, Galligan et al. 2012). Similarly, other studies employing murine embryonic fibroblasts (MEFs) and mouse liver tissue have also been carried out using SIRT3KO mice that also confirm the importance of SIRT3- dependent changes to the acetylome, specifically across several metabolic pathways

123 CHAPTER 5: MS Analysis

including fatty acid oxidation, ketogenesis, amino acid catabolism, and the urea and TCA cycles (Sol, Wagner et al. 2012, Rardin, Newman et al. 2013).

In this Chapter the aim was to examine how SIRT3 overexpression for 3 weeks in mouse liver affected lysine acetylation of mitochondrial proteins. Chow fed SIRT3 overexpressing liver mitochondria were compared to CON (empty vector) liver mitochondria for acetylation abundance by LC-MS/MS, and SIRT3KO and WT liver mitochondria were included in the analysis both to validate the methodology and to provide an important counterpoint to the effects of SIRT3 overexpression, that of SIRT3 deletion.

124 CHAPTER 5: MS Analysis

5.2 Method

General methods for protein extraction and western blotting are described in Chapter 2. Methods relating to the generation of the liver overexpression model are described in Chapter 4. Methods specific to this chapter are described below.

5.2.1 Workflow of liver MS/MS sample preparation and analysis

Figure 5.1 displays the workflow and sample comparisons for the MS/MS analysis of the 3-week liver SIRT3 overexpression mice. Pooled liver mitochondrial samples were trypsin digested and acetylated-lysine sites were enriched using an immunoaffinity approach. We performed label-free LC-MS/MS and performed downstream data processing including semi-quantitative abundance of peptides and pathway analysis.

Figure 5.1 Workflow and sample comparisons of LC-MS/MS experiments: Liver mitochondria were isolated from 4 individual chow-fed mice from SIRT3 knockout and WT colony, and from the 3 week cohort of Control and liver SIRT3 overexpressing mice. Alignment, normalisation and relative abundance measurements were processed in the Progenesis LC-MS analysis package. AcK = acetylated lysine; SIRT3 OE = SIRT3 overexpression.

125 CHAPTER 5: MS Analysis

5.2.2 Animals and mitochondria preparation

The SIRT3 liver overexpression model was detailed in Chapter 4. C57BL6 mice were transduced to overexpress SIRT3 using the HTVI technique and livers were collected at 3 weeks post overexpression. Samples used in MS experiments were from chow fed animals only. SIRT3KO and WT mice were from our colony at Australian Bio Resources (Moss Vale, Australia). Isolated liver mitochondria were prepared as detailed in Chapter 2. Briefly mitochondria were freshly prepared using differential centrifugation, in the presence of nicotinamide (to inhibit SIRT-mediated deacetylation), the final pellet was washed in isolation buffer without BSA, wash buffer removed and the pellet snap frozen in liquid nitrogen and stored prior to MS analysis.

5.2.3 Proteomics sample preparation

Isolated liver mitochondria pellets were homogenised in urea buffer (8 M Urea, 10 mM HEPES, pH 8.0) with the addition of complete EDTA-free protease inhibitor cocktail (Roche, Basel, Switzerland) and 10 mM nicotinamide. Mitochondrial fractions from 4 animals in each group were pooled. After sonication on ice for 10 seconds, lysates were centrifuged at 10000 g. 10 mg of protein was reduced with 5 mM DTT for 30 minutes at room temperature and alkylated with 15 mM iodoacetamide for 30 minutes protected from light at room temperature. Extracts were acetone precipitated to halt alkylation, and resuspended in urea buffer and assayed for protein content using the Bradford method (BioRad). 2 mg of protein from pooled samples was diluted 5-fold with 25 mM ammonium bicarbonate pH 8.0 and digested overnight with trypsin (Promega, Madison, WI) at 37 °C. The resulting peptides were acidified to pH 2 with trifluoroacetic acid (TFA) and desalted using a C18 SepPak cartridge (Waters, Milford, MA) and dried by vacuum centrifugation. For enrichment of acetylated peptides by immunoprecipitation, dried peptides were resuspended in MOPS IP buffer (50 mM

MOPS, 10 mM Na2HPO4, 50 mM NaCl, pH 7.2) with addition of 10 mM nicotinamide and incubated overnight at 4 °C with two pan-acetyl-lysine antibodies as recommended by Schilling and colleagues (Schilling, Rardin et al. 2012, Rardin, Newman et al. 2013) (Cell Signalling Technologies #9441S and Immunechem

126 CHAPTER 5: MS Analysis

Pharmaceutical #ICP0380) conjugated to protein A sepharose. Immunoprecipitates were washed five times and eluted with 0.1% TFA, 1% acetonitrile (CH3CN). Eluted peptides were desalted with STAGE tips (Thermo-Fisher, Waltham, MA), and dried by vacuum centrifugation.

5.2.4 LC-MS/MS analysis

Mass spectrometry was conducted at the Bioanalytical Mass Spectrometry Facilities within the Analytical Centre of the University of New South Wales. Peptides were resuspended in 0.1% formic acid and analyzed by LC-MS on an LTQ OrbitrapVelos MS system (Thermo Fisher, USA) coupled to an UltiMate 3000 nano-LC system (Dionex, Sunnyvale, CA, USA). A volume of 0.1 µL was loaded onto a micro C18 precolumn (500

µm×2 mm, MichromBioresources, Auburn, CA, USA) with Buffer A (98% H2O, 2%

-1 CH3CN, 0.1% TFA) at 10 µL min . After a 4 min wash the pre-column was switched (Valco 10 port valve, Dionex) into line with a fritless nano column (75 µm id×10 cm) containing reverse phase C18 media (3 µm, 200 Å Magic, MichromBioresources).

Peptides were eluted using a linear gradient of Buffer A (98% H2O, 0.1% TFA) to Buffer -1 B (98% CH3CN, 2% H2O, 0.1% formic acid) at 250 nL min over 60 min. High voltage (2000 V) was applied to a low volume tee (Upchurch Scientific, Oak Harbor, WA, USA) and the column tip positioned ~0.5 cm from the heated capillary (T=280 °C) of an OrbitrapVelos mass spectrometer. Positive ions were generated by electrospray and the Orbitrap operated in data-dependent acquisition mode. A survey scan m/z 350– 1750 was acquired in the Orbitrap (Resolution=30,000 at m/z 400, with an accumulation target value of 1,000,000 ions). Up to the 10 most abundant ions (>5000 counts) with charge states ≥ 2 were sequentially isolated and fragmented within the linear ion trap using collisionally induced dissociation with an activation q=0.25 and activation time of 30 ms at a target value of 30,000 ions. Ions selected for MS/MS were dynamically excluded for 30 s. Before each experiment, chromatographic and Orbitrap performance were confirmed by injecting GluFibrino peptide standard (5 fmol) using the above conditions and monitoring peak shape and mass calibration.

127 CHAPTER 5: MS Analysis

5.2.5 Data processing

Raw MS files were processed with Progenesis LC-MS software v4 for label-free analysis (Non-linear Dynamics, Newcastle upon Tyne, UK). Progenesis uses an alignment, ion abundance quantification, and normalisation strategy to allow accurate quantification between samples. Ion intensity maps from each run were aligned to a reference sample and ion feature matching was achieved by aligning consistent ion m/z and retention times. The peptide intensities were normalised against total intensity and compared between groups. Normalised MS/MS spectra were searched against the Mouse Swiss-Prot database (Sprot_Mus_25_10_13) and a contaminant database (Contamin_29_10_10) compiled by the Max Planck Institute of Biochemistry (Martinfreid, Germany) using database search program MASCOT (Matrix Science, London, UK), with enzyme selected as trypsin and allowing for up to 3 missed cleavages. Peptide mass tolerance was ± 6 ppm and fragment ion tolerance was ± 0.6 Da. Peptide charge states were set at +2 and +3. Oxidized methionine and acetylated lysine were searched as variable modifications. Following identification, a filter was applied to select proteins containing at least 2 unique peptides and those that contained at least one acetylation modification. Protein quantification is presented as average normalised abundance, with protein abundance calculated by Progenesis using the sum of all unique normalised peptide ion abundances for a specific protein on each run.

5.2.6 Bioinformatics analysis

All proteins submitted to further analysis were from the filtered dataset containing more than 1 unique peptide and at least one acetyl-lysine modification. Proteins that were upregulated in SIRT3KO and downregulated with SIRT3 overexpression were submitted to the Database for Annotation, Visualization and Integrated Discovery (DAVID) Bioinformatic Resources online (http://david.abcc.ncifcrf.gov/, version 6.7) for gene ontology (GO) and pathway enrichment analysis (Huang, Sherman et al. 2008, Huang da, Sherman et al. 2009). Area informative venn diagrams were produced using BioVenn, a web application for the comparison and visualization of biological lists

128 CHAPTER 5: MS Analysis

(Hulsen, de Vlieg et al. 2008). A list of acetylated proteins that were reduced in abundance with SIRT3 overexpression were submitted to Reactome’s Pathway Database platform (http://www.reactome.org/ (1_12_14)) (Croft, O'Kelly et al. 2011) and pathways mapped with the use of Cytoscape (version 3.2.0) and the Reactome Cytoscape FI Plugin (version 4.1.1beta).

5.3 Results and Discussion

5.3.1 Metrics of lysine acetylation sites in liver mitochondria

Comparisons were made between the SIRT3KO line and its WT counterpart, and between the SIRT3 overexpressing mice used in the HTVI overexpression studies relative to control mice transduced with an empty vector (described in Chapter 4). SIRT3KO mice have been extensively studied in the literature by proteomic methods and provide a control dataset to compare the effects of SIRT3 deletion and overexpression concurrently. As can be seen in Table 5.1, both datasets revealed reasonable coverage of the mitochondrial acetylome with ~1500 – 1900 peptides identified, and >400 of these containing acetylated lysines. Using a cut-off of >1 unique peptides per protein identification and a false discovery rate of 0.05, more than 200 proteins were identified in both datasets, with more than 50% of these containing at least one acetyl-lysine modification. Of these, 55% were shown to have increased abundance in the SIRT3KO livers compared to WT counterparts.

5.3.2 Reduced abundance of acetylated proteins with SIRT3 overexpression

Because the overexpression of the mitochondrial deacetylase SIRT3 was expected to decrease acetylation of SIRT3 regulated proteins, we compared the relative abundance of proteins in the SIRT3 overexpression (SIRT3OE) livers compared to control animals (Table 5.1). In line with this expectation, ~80% of proteins containing at least 1 acetyl modification site were decreased with SIRT3 overexpression.

129 CHAPTER 5: MS Analysis

Table 5.1 Metrics from MS/MS analysis # # Proteins # Protein # Peptides # Proteins Acetyl with at least abundance Detected Detected Peptides one AcK Δ SIRT3 KO>WT KO & WT 1887 573 224 143 79 / 143 dataset

SIRT3OE

Calculated in Progenesis (Non-linear Dynamics). #Proteins detected with >1 unique peptide. Relative protein abundances are calculated from abundance of all unique peptides identified for a particular protein.

5.3.3 Method development

Method development plays a key role in the establishment of new techniques. As detailed in Figure 5.1, this analysis was a pilot experiment to trial using the immunoaffinity enrichment strategies that have been used in the literature to assess lysine acetylation in SIRT3KO models and in mice under conditions such as calorie restriction or fasting (Lundby, Lage et al. 2012, Rardin, Newman et al. 2013, Still, Floyd et al. 2013). Liver mitochondria preparations from four animals were pooled (with equal contribution from each mouse) to provide a large amount of starting material for this experiment. Given the very good coverage of the acetylome that was achieved in this analysis, future experiments could be run with each sample in parallel without pooling, as the enrichment process allowed high recovery of acetylated peptides. Additionally, greater statistical power can be achieved for intensity-based quantitation with software such as Progenesis when samples are pooled during the downstream analysis stage rather than during sample preparation. Methods used in this chapter were adapted from Lundby et al. and Li et al (Lundby, Lage et al. 2012, Li, Silva et al. 2013). The use of tryptic digestion prior to the immunoprecipitation exposes a greater amount of residues to the acetylated lysine antibody, and significantly helps to enrich the acetylated peptides (Yang, Vaitheesvaran et al. 2011). In this chapter, we identified more than 200 individual proteins in each dataset, from more than a thousand unique peptides (Table 5.1). This is much more than was reported in early proteomics studies such as Kim et al. where only 233 acetylation sites were identified in 133 proteins (Kim,

130 CHAPTER 5: MS Analysis

Sprung et al. 2006), but less than recent experiments designed to provide a comprehensive survey of all acetylation sites such as Rardin and colleagues in 2013 who identified more than 2000 unique acetylation sites across more than 400 proteins using similar label-free methods (Rardin, Newman et al. 2013).

5.3.4 Quantitation of proteins in SIRT3KO and SIRT3 overexpression datasets

A list of the top 50 most-abundant proteins containing at least 1 acetyl-site in each dataset are listed in Tables 5.2 & 5.3. The most acetylated and most abundant protein in both datasets was the urea cycle enzyme carbamoyl phosphate synthase (CPS1) with 28 of its 66 detected peptides acetylated in the SIRT3 overexpression dataset, which is comparable to another published study (Rardin, Newman et al. 2013). Although very abundant in mitochondria and showing a large amount of acetylation, CPS1 is not known to be regulated by SIRT3 (Rardin, Newman et al. 2013), which is reflected by the relatively small fold-change seen in this experiment. Although this particular dataset does not have enough technical replicates to be truly quantitative, in general, the increase in the abundance of acetylated proteins in this immunoaffinity-enriched analysis broadly indicates that SIRT3 overexpression is affecting acetylation in the liver as expected.

131 CHAPTER 5: MS Analysis

Table 5.2 50 most abundant proteins detected in dataset with increased abundance in SIRTKO mouse liver compared to WT Confid- Normalised Access- Peptides Max Protein Name ence abundance Ion # Detected fold ∆ Score SIRT3KO WT Carbamoyl-phosphate CPS1 97 6116 1.53E+07 1.33E+07 1.1 synthase 1 [ammonia] GDH/ Glutamate dehydrogenase 1 34 2224 4.83E+06 4.59E+06 1.1 DHE3 ATPA ATP synthase subunit alpha 23 1617 1.50E+06 1.18E+06 1.3 HMGCS Hydroxymethylglutaryl-CoA 26 1547 1.62E+06 1.27E+06 1.3 2 synthase Non-specific lipid-transfer NLTP 22 1276 4.09E+05 3.12E+05 1.3 protein CH60 60 kDa heat shock protein 21 1227 8.47E+05 6.06E+05 1.4 THIL Acetyl-CoA acetyltransferase 17 1069 5.45E+05 5.03E+05 1.1 Trifunctional enzyme subunit ECHA 20 1027 5.28E+05 2.79E+05 1.9 alpha Peroxisomal acyl-coenzyme A ACOX1 16 994 1.70E+05 1.49E+05 1.1 oxidase 1 M2GD Dimethylglycine 16 862 2.67E+05 2.31E+05 1.2 dehydrogenase SARDH Sarcosine dehydrogenase 14 818 2.04E+05 1.62E+05 1.3 ATP synthase F(0) complex AT5F1 13 818 2.31E+05 2.31E+05 1.0 subunit B1 DLDH Dihydrolipoyl dehydrogenase 14 802 2.72E+05 1.89E+05 1.4 Long-chain-fatty-acid--CoA ACSL1 13 780 2.46E+05 2.24E+05 1.1 ligase 1 ECI1 Enoyl-CoA delta isomerase 1 13 738 2.94E+05 1.66E+05 1.8 Peroxisomal bifunctional ECHP 14 720 1.50E+05 1.28E+05 1.2 enzyme ECHM Enoyl-CoA hydratase 11 683 1.22E+05 8.88E+04 1.4 Delta-1-pyrroline-5- AL4A1 11 634 2.21E+05 1.97E+05 1.1 carboxylate dehydrogenase MDHM Malate dehydrogenase 10 592 1.49E+05 1.42E+05 1.0 BPHL Valacyclovir hydrolase 11 591 2.04E+05 1.79E+05 1.1 ATP5J ATP synthase-coupling factor 6 10 549 8.13E+05 6.01E+05 1.4 Hydroxyacyl-coenzyme A HCDH 10 532 3.82E+05 1.59E+05 2.4 dehydrogenase PDIA1 Protein disulfide-isomerase 9 518 4.39E+05 1.69E+05 2.6 Electron transfer flavoprotein ETFA 7 509 1.73E+05 1.52E+05 1.1 subunit alpha Medium-chain specific acyl- ACADM 10 508 4.57E+05 2.66E+05 1.7 CoA dehydrogenase Alpha-aminoadipic 1.6 AASS 7 493 5.22E+05 3.36E+05 semialdehyde synthase GLYAT Glycine N-acyltransferase 7 481 4.82E+05 2.25E+05 2.1

SODM Superoxide dismutase [Mn] 6 477 1.33E+06 1.13E+06 1.2

132 CHAPTER 5: MS Analysis

Table 5.2 continued… Confid- Normalised Max Access- Peptides Protein Name ence abundance fold ion # Detected Score SIRT3KO WT ∆ Propionyl-CoA carboxylase PCCB 7 464 1.65E+05 3.49E+04 4.7 beta chain Very long-chain specific acyl- ACADV 9 426 4.26E+04 3.77E+04 1.1 CoA dehydrogenase Acyl-CoA synthetase family ACSF2 8 406 6.34E+04 5.12E+04 1.2 member 2 Histidine triad nucleotide- HINT2 4 398 1.37E+06 1.16E+06 1.2 binding protein 2 3-hydroxyisobutyryl-CoA HIBCH 7 375 7.39E+04 3.35E+04 2.2 hydrolase Calcium-binding mitochondrial CMC2 9 362 1.74E+05 4.04E+04 4.3 carrier protein Aralar2 GTP:AMP phosphotransferase KAD3 8 359 5.99E+05 3.15E+05 1.9 AK3 Acyl-coenzyme A synthetase ACSM5 7 355 5.85E+04 3.60E+04 1.6 ACSM5 ACON Aconitate hydratase 7 338 9.00E+04 6.03E+04 1.5 Succinyl-CoA ligase [ADP- SUCB1 5 329 3.06E+04 1.55E+04 2.0 forming] subunit beta Glycine N-acyltransferase-like GLYAL 6 321 1.40E+05 1.11E+05 1.3 protein Glutathione S-transferase GSTK1 5 320 2.46E+05 1.97E+05 1.2 kappa 1 MPCP Phosphate carrier protein 7 312 5.80E+04 5.54E+04 1.0 NIPS1 Protein NipSnap homolog 1 7 309 7.38E+04 4.56E+04 1.6 2-amino-3-ketobutyrate KBL 4 304 9.28E+04 7.05E+04 1.3 coenzyme A ligase CISY Citrate synthase 7 303 7.00E+04 5.91E+04 1.2 ATP5I ATP synthase subunit e 5 285 1.89E+05 1.52E+05 1.2 Citrate lyase subunit beta-like CLYBL 4 280 9.36E+04 8.14E+04 1.2 protein D-beta-hydroxybutyrate BDH 7 278 1.93E+05 1.17E+05 1.7 dehydrogenase ATP-binding cassette sub- ABCD3 5 275 6.69E+04 3.89E+04 1.7 family D member 3 Lipoamide acyltransferase ODB2 4 267 1.44E+05 1.03E+05 1.4 component of BCKDHA Acyl-coenzyme A synthetase ACSM1 7 262 8.30E+04 5.12E+04 1.6 ACSM1 Protein list generated using Progenesis LC-MS Software (Non-linear Dynamics). ‘Accession#’ is UniProt/SwissProt entry name. ‘Confidence score’ is the data type confidence score calculated from the sum of the score of all peptides in the run calculated by the peak-picking software MASCOT, and calculated by Progenesis LC-MS. ‘Normalised abundance’ is relative protein abundances calculated from sum of all unique peptides identified for a particular protein and normalised using ion intensity. ‘Max fold change’ is calculated from normalised abundance SIRT3KO/WT.

133 CHAPTER 5: MS Analysis

Table 5.3 The 50 most abundant proteins detected in dataset with decreased abundance in SIRT3 overexpressing mouse liver compared to control Confid- Normalised Access- Peptides Max Protein Name ence abundance ion # Detected fold ∆ score CON SIRT3OE Carbamoyl-phosphate CPS1 66 3888 1.85E+06 1.02E+06 1.8 synthase [ammonia] THIM 3-ketoacyl-CoA thiolase 20 1478 4.33E+05 1.97E+05 2.2 Glutamate dehydrogenase 1 DHE3 23 1468 1.03E+06 7.06E+05 1.5

Non-specific lipid-transfer NLTP 23 1361 1.29E+05 1.00E+05 1.3 protein ATPA ATP synthase subunit alpha 19 1344 2.41E+05 1.78E+05 1.4 AATM Aspartate aminotransferase 17 999 1.81E+05 8.76E+04 2.1 ALDH2 Aldehyde dehydrogenase 17 923 1.33E+05 7.95E+04 1.7 ATPB ATP synthase subunit beta 13 849 8.07E+04 5.79E+04 1.4 Peroxisomal multifunctional DHB4 13 840 3.36E+04 2.56E+04 1.3 enzyme type 2 Peroxisomal bifunctional ECHP 17 828 9.80E+04 8.10E+04 1.2 enzyme CH60 60 kDa heat shock protein 14 792 5.27E+04 4.48E+04 1.2 HMGCS Hydroxymethylglutaryl-CoA 13 791 1.19E+05 6.68E+04 1.8 2 synthase URIC Uricase 15 638 5.00E+04 1.67E+04 3.0 Alpha-aminoadipic AASS 11 614 5.56E+04 4.89E+04 1.1 semialdehyde synthase MUP2 Major urinary protein 2 8 591 1.18E+05 8.55E+04 1.4 PDIA1 Protein disulfide-isomerase 12 590 1.96E+05 1.68E+05 1.2 Ornithine OTC 9 539 1.16E+05 7.32E+04 1.6 carbamoyltransferase ATP5H ATP synthase subunit d 10 516 1.26E+05 6.43E+04 2.0 Succinate dehydrogenase SDHa/ [ubiquinone] flavoprotein 7 490 1.98E+04 1.90E+04 1.0 DHSA subunit GRP75 Stress-70 protein 8 481 1.46E+04 1.35E+04 1.1 MDHM Malate dehydrogenase 10 473 1.69E+04 1.20E+04 1.4 H4 Histone H4 7 467 1.80E+05 2.04E+04 8.8 Acetyl-CoA acetyltransferase THIL 6 459 2.32E+04 1.88E+04 1.2

EF1A1 Elongation factor 1-alpha 1 6 445 8.85E+04 4.96E+04 1.8 ADT2 ADP/ATP translocase 2 7 433 1.50E+05 4.66E+04 3.2 Methylmalonate- MMSA semialdehyde 5 408 4.68E+04 2.10E+04 2.2 dehydrogenase [acylating] ATP-binding cassette sub- ABCD3 8 404 5.84E+04 3.54E+04 1.7 family D member 3 SARDH Sarcosine dehydrogenase 6 376 2.19E+04 2.06E+04 1.1

PYC Pyruvate carboxylase 7 376 2.28E+04 1.98E+04 1.2

134 CHAPTER 5: MS Analysis

Table 5.3 continued… Confid- Normalised Access- Peptides Max Protein Name ence abundance ion # Detected fold ∆ score CON SIRT3OE DLDH Dihydrolipoyl dehydrogenase 6 359 2.05E+04 1.56E+04 1.3 Trifunctional enzyme subunit ECHA 7 344 1.54E+04 9.06E+03 1.7 alpha ATPG ATP synthase subunit gamma 6 339 4.04E+04 2.58E+04 1.6 CALR Calreticulin 5 326 1.09E+05 6.20E+04 1.8 78 kDa glucose-regulated GRP78 5 325 2.00E+04 9.07E+03 2.2 protein ATP synthase-coupling factor ATP5J 6 307 5.92E+04 2.56E+04 2.3 6 ATPO ATP synthase subunit O 5 306 1.14E+04 7.75E+03 1.5

PYGL Glycogen phosphorylase 6 297 1.48E+04 1.26E+04 1.2

CES1D Carboxylesterase 1D 6 297 2.41E+04 2.03E+04 1.2 Isovaleryl-CoA IVD 4 296 4.17E+04 2.40E+04 1.7 dehydrogenase ECHM Enoyl-CoA hydratase 4 295 7.22E+03 4.92E+03 1.5 ACADL/ Long-chain specific acyl-CoA 4 291 1.25E+04 9.49E+03 1.3 LCAD dehydrogenase Dimethylglycine M2GD 5 285 3.66E+04 2.53E+04 1.4 dehydrogenase Trifunctional enzyme subunit ECHB 7 280 2.90E+04 1.60E+04 1.8 beta Acyl-CoA synthetase family ACSF2 6 277 7.67E+03 6.03E+03 1.3 member 2 Microsomal triglyceride MTP 5 270 5.27E+03 5.13E+03 1.0 transfer protein large subunit BPHL Valacyclovir hydrolase 4 266 3.40E+04 1.63E+04 2.1 Cytochrome b-c1 complex QCR2 3 259 1.05E+04 5.52E+03 1.9 subunit 2 Short-chain specific acyl-CoA ACADS 4 254 1.11E+04 5.60E+03 2.0 dehydrogenase 3-hydroxyisobutyrate 3HIDH 4 250 1.87E+04 8.05E+03 2.3 dehydrogenase Medium-chain specific acyl- ACADM 5 239 9.14E+04 4.91E+04 1.9 CoA dehydrogenase Protein list generated using Progenesis LC-MS Software (Non-linear Dynamics). ‘Accession#’ is UniProt/SwissProt entry name. ‘Confidence score’ is the data type confidence score calculated from the sum of the score of all peptides in the run calculated by the peak-picking software MASCOT, and calculated by Progenesis LC-MS. ‘Normalised abundance’ is relative protein abundances calculated from sum of all unique peptides identified for a particular protein and normalised using ion intensity. ‘Max fold change’ is calculated from normalised abundance CON/SIRT3OE.

135 CHAPTER 5: MS Analysis

5.3.5 Overlap between the pathways regulated by both SIRT3 deletion and overexpression

We sought to illustrate which proteins were equally regulated by both SIRT3 deletion and SIRT3 overexpression in this analysis. As can be seen from the diagram in Figure 5.2, 41-46% of proteins from each dataset overlapped and appeared to be increased in SIRT3KO compared to WT, and decreased in SIRT3OE compared to control.

The SIRT3KO and WT mice were analysed in a separate, parallel experiment to avoid false positives due to the different strains of mice, as the SIRT3KO line is on a 129 background, and C57BL6 mice were used for the overexpression studies. However, there was still overlap in the most abundant proteins detected in the SIRT3KO analysis and the SIRT3 overexpression analysis.

Another indication of the robustness of the data generated from this dataset comes from the fact that many of the acetyl sites detected (Tables 5.2 & 5.3) have been previously reported in other studies. Weinert et al. made the results of their mouse liver acetylation screen accessible as supplementary data accompanying their recent publication (Weinert, Schölz et al. 2013). Many of the acetyl sites identified in this chapter were present in this published acetylation screen. Additionally, because trypsin cleavage usually cleaves after a lysine residue in a protein, yet acetylation tends to protect residues from protease digestion, the identification of missed cleavages associated with lysine acetylation sites provides good evidence that an acetylation site has been correctly assigned by the MASCOT program.

136 CHAPTER 5: MS Analysis

Figure 5.2 Overlap between increased proteins in SIRT3KO and decreased proteins in SIRT3 overexpression: Numbers represent individual proteins in dataset that appear to be increased with SIRT3 deletion alone in blue, with SIRT3 overexpression alone in red, and and that overlap in both datasets in purple. Area proportional venn diagrams generated by BioVenn (Hulsen, de Vlieg et al. 2008).

5.3.6 Mitochondrial proteins are almost wholly represented in cell compartment analysis

Gene-set enrichment analysis was undertaken on protein lists from Table 5.2 & 5.3 using cell compartment gene ontology (GO) terms to further validate the mass spectrometry results using the DAVID database tool. Figure 5.3 shows that the majority of these acetylated proteins were considered to have mitochondrial localisation. A small proportion (9%) of proteins were found to have peroxisomal descriptions, which is perhaps not surprising given the similar cellular characteristics of peroxisomes and mitochondria and the method of isolation used. The preponderance of mitochondrial proteins identified in this analysis confirms the quality of the sample preparation and bioinformatics methods used.

137 CHAPTER 5: MS Analysis

% Proteins

9 86

Mitochondria (141/164) Peroxisome (14/164) Other

Figure 5.3 Cell compartment pathway analysis in acetylated proteins differentially acetylated in SIRT3KO and SIRT3 overexpressing liver mitochondria: Proteins that map to gene ontology (GO) terms associated with the mitochondrial cell compartment (in blue) comprise 86% of combined datasets, proteins that map to peroxisomal cell compartment GO terms (in red) comprise 9% of combined dataset. Other includes proteins that map to other compartments, and unmapped/unknown proteins. Pathway analysis calculated using Database for Annotation, Visualization and Integrated Discovery (DAVID) tool. Data includes some proteins that were not identified by DAVID tool, those that map to more than 1 compartment, and accession numbers that correspond to multiple isoforms. Numbers in brackets are number of proteins in list that were mapped to each GO term.

5.3.7 Results of Reactome pathway enrichment analysis

Pathway analysis is a powerful tool to see how the proteins identified in a certain experiment cluster together in terms of known networks and pathways in the cell, giving biological significance to a set of results and highlighting pathways that may be of particular importance. Pathway enrichment analysis using Reactome was also carried out on this dataset to further understand the biological significance of the proteins that were differentially acetylated. Reactome is a highly curated pathway database and network-modeling tool (Haw, Hermjakob et al. 2011, Croft, Mundo et al.

138 CHAPTER 5: MS Analysis

2014). Table 5.5 shows the top pathways identified by Reactome as being overrepresented/enriched in the SIRT3 overexpression proteomics dataset. The advantage of Reactome analysis is that information is further enhanced by the addition of hierarchy organisation to highlight exact pathways and sub-pathways these acetylated proteins are involved in.

Of the 85 proteins submitted to Reactome, only 77 were processed in the pathway analysis due to non-conforming accession numbers. Reactome analysis found that all highly significant pathways (p<0.05) were from the metabolism pathway, with 45 proteins out of the 77 in the dataset enriched in the metabolism GO term. In line with the understanding of the role of SIRT3 in mitochondrial metabolism, and in agreement with the effects on oxygen consumption and fat oxidation presented earlier in this thesis, the most represented Reactome pathways in the SIRT3 overexpression dataset were lipid metabolism, the TCA cycle and electron transport, and amino acid metabolism. One of the most significant hits was mitochondrial fatty acid beta- oxidation, with 8 of the 15 proteins associated with this pathway displaying altered acetylation in the SIRT3 overexpression analysis. As can be seen in Table 5.4, several interesting sub-pathways are highlighted, including the association with specifically saturated fatty acid metabolism, where 70 – 100% of the proteins ascribed to these pathways are affected by SIRT3 overexpression. Similarly, as a subset of fatty acid metabolism, ketone body synthesis is revealed as a significant pathway represented in this dataset.

139 CHAPTER 5: MS Analysis

Table 5.4 Heirarchical pathway analysis from Reactome associated with SIRT3 overexpression in mouse liver &pValue 2.84E)04 1.44E)03 3.31E)10 7.75E)09 2.19E)05 2.08E)06 2.08E)06 2.08E)06 2.08E)06 1.29E)03 1.29E)03 2.19E)05 3.50E)03 2.27E)03 2.48E)02 8.67E)03 4.99E)03 2.51E)02 1.18E)02 8.67E)03 1.23E)05 8.03E)03 3.60E)04 8.74E)07 6.27E)04 2.09E)07 1.31E)05 6.29E)07 1.09E)02 3 5 8 8 9 7 8 53 75 80 80 80 80 67 67 40 50 14 25 33 11 25 21 33 50 35 22 100 100 %&in&Data 8 6 3 4 4 4 4 2 2 3 2 2 2 2 2 3 3 2 4 4 6 8 4 6 2 45 10 12 16 Data SIRT3OE& Matching& Proteins&in& 8 3 5 5 5 5 3 3 3 5 4 8 6 8 8 9 15 14 36 27 45 19 18 17 184 149 108 189 1588 in&Pathway Total&Proteins& Beta/oxidation/of/myristoyl)CoA/to/lauroyl)CoA Beta/oxidation/of/decanoyl)CoA/to/octanoyl)CoA)CoA Beta/oxidation/of/hexanoyl)CoA/to/butanoyl)CoA Beta/oxidation/of/lauroyl)CoA/to/decanoyl)CoA)CoA Beta/oxidation/of/octanoyl)CoA/to/hexanoyl)CoA Beta/oxidation/of/butanoyl)CoA/to/acetyl)CoA Beta/oxidation/of/palmitoyl)CoA/to/myristoyl)CoA ✚ ✚ ✚ ✚ ✚ ✚ ✚ Mitochondrial/fatty/acid/beta)oxidation/of/saturated/fatty/acids Mitochondrial/fatty/acid/beta)oxidation/of/unsaturated/fatty/acids Synthesis/of/Ketone/Bodies ✚ ✚ ✚ Pathway&name&and&heirarchy Mitochondrial/Fatty/Acid/Beta)Oxidation Ketone/body/metabolism Chylomicron)mediated/lipid/transport Vitamin/C/(ascorbate)/metabolism Acyl/chain/remodeling/of/CL Synthesis/of/bile/acids/and/bile/salts Beta)oxidation/of/pristanoyl)CoA Citric/acid/cycle/(TCA/cycle) ✚ ✚ ✚ ✚ ✚ ✚ ✚ ✚ Fatty/acid,/triacylglycerol,/and/ketone/body/metabolism Lipid/digestion,/mobilisation/and/transport Metabolism/of/vitamins/and/co)factors Phospholipid/metabolism Bile/acid/and/bile/salt/metabolism Peroxisomal/lipid/metabolism Pyruvate/metabolism/and/Citric/Acid/(TCA)/cycle Formation/of/ATP/by/chemiosmotic/coupling Respiratory/electron/transport,/ATP/synthesis/by/chemiosmotic/coupling,/and/heat/production/by/uncoupling/proteins. Lysine/catabolism Branched)chain/amino/acid/catabolism Urea/cycle ✚ ✚ ✚ ✚ ✚ ✚ ✚ ✚ ✚ ✚ ✚ ✚ Metabolism&of&lipids&and&lipoproteins The&citric&acid&(TCA)&cycle&and&respiratory&electron&transport Metabolism&of&amino&acids&and&derivatives ✜ ✜ ✜ METABOLISM Data/prepared/using/Reactome/Pathway/Analysis/Tool,/

140 CHAPTER 5: MS Analysis

5.3.8 Pathway enrichment analysis identifies mitochondrial metabolism and fat oxidation as key pathways in SIRT3 overexpression dataset

The importance of SIRT3 for mitochondrial fatty-acid oxidation is consistent with findings in SIRT3KO mouse liver where it has been shown that SIRT3 deletion leads to reduced fatty acid oxidation due to hyperacetylation of certain residues on LCAD (Hirschey, Shimazu et al. 2010). Considering that in isolated hepatocytes we saw effects on liver triglyceride accumulation, the members of this pathway that have been detected by MS may require further investigation as possible SIRT3 targets. Ketone synthesis was another pathway that was highlighted in the Reactome analysis. This is of particular interest given that SIRT3 overexpression in liver was shown to increase levels of the ketone β-hydroxybutyrate (β-OHB) in vivo in mice (Chapter 4). Other previously reported pathways associated with SIRT3KO such as that of the urea cycle and branched-chain amino acid catabolism (Hebert, Dittenhafer-Reed et al. 2013), were also found to be enriched with SIRT3 overexpression using the Reactome tool. Future experiments allowing integration of mass spectrometry guided pathway analysis with metabolomics data in packages such as Reactome could highlight how these protein changes affect metabolite flux, and will provide a fuller picture of how SIRT3 overexpression affects mitochondrial metabolism networks.

5.3.9 SIRT3 overexpression in liver causes decreased acetylation of known SIRT3 targets at the peptide level

Although normalized abundance of acetylated proteins as shown in Tables 5.2 and 5.3 is informative, the changes that are occurring at the peptide level within these proteins is critical. As discussed at the beginning of this chapter, while there are often many acetyl-lysine sites in a given protein, often only one or two may be responsible for changes in protein function associated with this post-translational modification. An in depth analysis was conducted with the peptide data generated in these MS experiments for several validated SIRT3 targets to determine in our dataset the effect both SIRT3 deletion and SIRT3 overexpression had on published acetylation sites. Figure 5.4 shows the relative abundance of individual acetylated peptides detected for

141 CHAPTER 5: MS Analysis

5 known SIRT3 targets: 1) The TCA cycle/complex II enzyme succinate dehydrogenase a subunit (SDHa); 2) The urea cycle enzyme ornithine transcarbamylase (OTC); 3) The fatty acid oxidation enzyme long chain acyl-CoA dehydrogenase (LCAD); 4) the rate- limiting step in synthesis of the ketone β-hydroxybutyrate, 3-hydroxy-3-methylglutaryl CoA synthase 2 (HMGCS2); and 5) glutamate dehydrogenase (GDH) an important metabolic enzyme with roles in both nitrogen and glutamate metabolism. Figure 5.4A shows the relative abundance of all these peptides combined. Because this data is presented as a ratio, a value >1 denotes increased abundance of acetylated peptides as is the case with the SIRT3KO to WT comparison which shows that mean acetylation is increased with SIRT3 deletion. A ratio <1 which is seen in SIRT3OE compared to control liver mitochondria, denotes a decreased abundance of acetylated peptides with SIRT3 overexpression.

5.3.9.1 Acetylation at K179 of SDH is decreased with SIRT3 overexpression, increased with SIRT3 deletion Lysines (K) 179, 498 and 485 have been shown in the literature to be important sites for the enzyme activity of SDH (Cimen, Han et al. 2010, Finley, Haas et al. 2011, Hebert, Dittenhafer-Reed et al. 2013). These three sites were also detected by MS in our study. As seen in Figure 5.4, of these three residues, only K179 showed a large decrease in acetylation with SIRT3 overexpression, and a corresponding increase with SIRT3 deletion in liver mitochondria samples. The K179 site specifically showed a large increase in acetylation with CR in a previous study, that was also associated with changes in succinate levels in CR via a metabolite screen (Hebert, Dittenhafer-Reed et al. 2013). Acetylation at these lysines have been shown to be associated with reducing activity of SDH enzyme, via a proposed mechanism of blocking substrate entry to the active site (Cimen, Han et al. 2010, Xiong and Guan 2012).

142 CHAPTER 5: MS Analysis

A B MEAN SDHa

*K179 K335 SIRT3OE:CON * K485 * K498 KO:WT K547 K608 Lysine Residue Lysine

1 2 1 2 4 0.5 0.5 0.25 Ratio Ratio (Log Abundance Peptides) 2 (Log2 Abundance Peptides)

C OTC D LCAD

* K88 K81 K144 K221 K92 K231 K156 K243 Lysine Residue Lysine Lysine Residue Lysine

1 2 4 8 1 2 4 0.5 0.5 0.25 0.25 0.125 Ratio Ratio (Log2 Abundance Peptides) (Log2 Abundance Peptides)

HMGCS2 GDH E F K83 K84 K243 K90 * K310 K200 K342 K352 K415 K354 K457 K358 K480 K367 K503 K427 K527 Lysine Residue Lysine Lysine Residue Lysine * K447 K545 * K473 K548

1 2 4 8 1 2 4 8 0.5 0.5 0.25 0.25 0.125 0.125 Ratio Ratio (Log2 Abundance Peptides) (Log2 Abundance Peptides)

Figure 5.4 Acetylation of individual peptides in a selection of known SIRT3 targets from SIRT3KO and SIRT3OE liver mitochondria: A. Mean difference in abundance of acetylated peptides from all proteins presented in this figure presented as ratio of SIRT3KO/WT (black bars), and SIRT3 overexpression (SIRT3OE)/CON (grey bars). B-F. Relative abundance of individual peptides described by their acetylated lysine residue in (B) succinate dehydrogenase subunit (SDHa), (C) ornithine transcarbamylase, (D) long-chain acyl-CoA dehydrogenase, (E) 3-hydroxy-3-methylglutaryl-CoA synthase 2 (HMGCS2), and (F) glutamate dehydrogenase. *= Lysine residue known to confer change in activity when acetylated. Data are normalised abundance of individual peptides expressed as a ratio and log2 transformed. Ratio>1 denotes increased abundance of acetylated peptide, ratio<1 denotes decreased abundance of acetylated peptides.

143 CHAPTER 5: MS Analysis

5.3.9.2 Known SIRT3 targets OTC and LCAD show variable effects of SIRT3 overexpression and deletion All acetylated peptides detected in Figure 5.4C and D have been previously reported (Weinert, Schölz et al. 2013). The acetylation sites of OTC have been characterised in other studies and acetylation at K88 has been shown to be responsible for interference at the active site of the enzyme, reducing its activity (Yu, Lin et al. 2009, Hallows, Yu et al. 2011). Although on average the abundance of acetylated peptides was decreased with SIRT3 overexpression (extrapolated from Table 5.3), as can be seen in Figure 5.4C, the pattern of acetylation at K88 does not fit this theory. SIRT3 deletion appears to increase acetylation only at K221. SIRT3 overexpression shows decreased acetylation at three lysines, K88, K144 and K231.

The role of acetylation on the function of the fatty acid oxidation enzyme LCAD has been investigated (Hirschey, Shimazu et al. 2010, Bharathi, Zhang et al. 2013). Hirschey et al thoroughly investigated acetylation of LCAD, and concluded that K42 showed a 20-fold increase in acetylation in SIRT3KO liver (Hirschey, Shimazu et al. 2010). More recently, Bharathi et al performed mutation analyses on several residues of LCAD, where substitution of the lysine residue with an arginine mimics the effect of constituitive deacetylation, and showed the importance of K318 and K322 residues for LCAD conformation and activity (Bharathi, Zhang et al. 2013). However as can be seen in Figure 5.4D, only 3 acetylated peptides were detected in this analysis, all of which have been previously reported (Weinert, Schölz et al. 2013), although the crucial peptides containing K42 and K318/322 were not detected by MS in our dataset. K81 showed a modest increase with SIRT3 deletion and decrease with SIRT3 overexpression, however K92 showed a large decrease in abundance of acetylated peptide with SIRT3 deletion, illustrating that not all peptides match this general trend.

5.3.9.3 Known SIRT3 target HMGCS2 show variable effects of SIRT3 overexpression and deletion HMGCS2 is an important enzyme in ketone body synthesis, and its acetylation at K310, K447 and K473 in SIRT3KO mice has been associated with reduced activity via inducing conformational change of the enzyme, causing a reduction in levels of the ketone β-

144 CHAPTER 5: MS Analysis

hydroxybutyrate (Shimazu, Hirschey et al. 2010). We identified 10 acetyl sites in HMGCS2 that were differentially acetylated by SIRT3 (Figure 5.4E), including the three previously described. On average, there was increased acetylation with SIRT3KO compared to WT in liver mitochondria (Mean SIRT3KO:WT ratio of 2.6 ± 0.9), and decreased acetylation with SIRT3 overexpression, although the magnitude of change was much smaller (Mean SIRT3OE:CON ratio of 0.7 ± 0.1), however the pattern of acetylation in the two different models at the crucial sites described in the literature was variable.

5.3.9.4 Known SIRT3 target GDH shows decreased acetylation with SIRT3 overexpression As seen in Figure 5.4F, 11 acetylation sites were identified in the mitochondrial enzyme GDH. Of these, all had been described previously except for K457 (Weinert, Schölz et al. 2013). Acetylation sites at K90, K480 and K527 have also been identified in multiple previous studies as being affected by SIRT3 deletion or calorie restriction (Lombard, Alt et al. 2007, Schwer, Eckersdorff et al. 2009, Hebert, Dittenhafer-Reed et al. 2013). On average the abundance of acetylated peptides in GDH increased almost 2-fold with SIRT3 deletion (SIRT3KO:WT ratio of 1.7 ± 0.4) and decreased by ~40% with SIRT3 overexpression (Mean SIRT3OE:CON ratio of 0.6 ± 0.1). However, currently the precise mechanism by which acetylation affects GDH activity, and which residues are involved is unknown, hence further analysis of how the acetylation at these particular sites could effect function is needed to give these results further meaning.

5.3.10 MS analysis of muscle SIRT3 overexpression

Mass spectrometry analysis using similar methods was also performed on rat muscle overexpressing SIRT3 following intramuscular delivery of SIRT3-AAV [data not shown (see Chapter 3 for details of muscle overexpression model)]. Despite its importance as a metabolic tissue, skeletal muscle remains relatively unstudied in proteomics due to the difficulty of achieving thorough coverage of the muscle proteome (Burniston, Connolly et al. 2014). In the context of lysine acetylation, only two studies have

145 CHAPTER 5: MS Analysis

included skeletal muscle in their analysis of the acetylome (Yang, Vaitheesvaran et al. 2011, Lundby, Lage et al. 2012). Muscle proteomics are hindered by the fact that LC- MS techniques are biased towards detecting heavily abundant proteins in the sample unless technical modifications are made such as the use of inclusion lists during the MS/MS run. In skeletal muscle, approximately half of the total protein mass is occupied by only 10 highly abundant contractile proteins (Burniston, Connolly et al. 2014) including actins and myosins, which can obscure detection of less abundant proteins. In the MS analysis of SIRT3 overexpressing muscle, 7 out of the top 10 most abundant peptides were for contractile proteins, despite the fact that mitochondrial fractions were prepared and analysed. Due to the large mass of these contractile proteins, a large number of peptides were detected in the muscle overexpression analysis (1277 peptides), however they only represented 82 proteins, and only 76 of these peptides were acetylated. The only acetylated proteins of note detected in this analysis were from ATP synthase F(0) subunit and ATP synthase subunit d (data not shown). Further optimisation will need to be done to improve the method of subcellular fractionation prior to performing more muscle overexpression proteomics studies.

5.3.11 Limitations of non- MS techniques

Although SIRT3 deletion causes marked hyperacetylation that is discernible via western blot (Lombard, Alt et al. 2007), SIRT3 overexpression in liver did not produce changes in acetylation that were apparent via western blot with a pan-acetyl lysine antibody (data not shown). Even with the very high level of SIRT3 overexpression seen in skeletal muscle with SIRT3 AAV, there was only a subtle decrease in total acetylation as assessed by western blot techniques (see Chapter 3). Therefore, MS analysis is critical for determining the functional relevance of SIRT3 overexpression, due to the limitations of current biochemical methods such as western blot. It is important to note that some acetylated proteins have multiple lysine acetylation sites, and hence validation studies are necessary to delineate which of the modified lysines are

146 CHAPTER 5: MS Analysis

responsible for changes in activity of the target protein. In the case of LCAD, a SIRT3 target known to be involved in fatty acid oxidation, while there are more than 8 acetyl- lysine sites on the protein, only one has been shown to be associated with modified enzyme activity in the models studied thus far (Hirschey, Shimazu et al. 2010). The idea that there is a generally low level of acetylation at many lysine residues in a protein, but that only a select sub-set are important for function has been discussed in recent literature, and provides support for the hypothesis that acetylation is a non-enzymatic process (Choudhary, Weinert et al. 2014) (discussed in Chapter 1). Hence investigation of lysine acetylation using techniques such as western blotting or immunoprecipitation (IP) that cannot discriminate between different lysine sites is problematic. Until site- specific acetyl antibodies become commercially available for proteins of interest (such as is currently available for phosphorylation sites in insulin signalling pathways for example), then mass spectrometric methods remain an important method of targeting specific acetyl sites in relevant metabolic proteins.

5.3.12 Limitations of MS techniques (I): Functional annotation of acetylation sites is needed for more SIRT3 targets

Tables 5.2 and 5.3 summarised the top 50 most abundant peptides that were detected in each dataset. Although normalised abundances are provided, the quantification and maximum-fold change presented in these tables are for the sum of all peptides detected in each sample, including those that were un-acetylated. Recent evidence is emerging that acetylation may occur in a non-enzymatic manner (Wagner and Hirschey 2014), and that not all acetylation sites necessarily have a functional significance.

Hence a subset of published SIRT3 targets were manually investigated to see which precise peptides were acetylated, and whether these acetylated peptides showed changes in abundance that supported the hypothesis that SIRT3 overexpression decreased acetylation. Figure 5.4 shows 5 known targets of SIRT3 and the individual

147 CHAPTER 5: MS Analysis

acetylated peptides identified in the MS experiments. On average, the ratio of acetylated peptide abundance between SIRT3KO:WT was increased, while the ratio of SIRT3OE:CON was decreased, in line with predictions according to SIRT3’s role as a deacetylase. However, if the non-enzymatic theory of lysine acetylation is true, then it means that the combined acetylation on a particular protein is not necessarily going to have an effect on function, rather the precise site at which that acetylation is occurring will be the most informative. Figure 5.4 shows that in proteins where the site of acetylation that regulates function is known such as K179 of SDHa, then investigation of SIRT3 overexpression on function using mass spectrometry is relatively straightforward. In proteins such as HMGCS2 where the precise residues involved have not been delineated, then seeing an average decrease in acetylation such as we saw with SIRT3 overexpression in this chapter is encouraging, however it may not translate into a difference in enzyme function. In addition, there is significant overlap between PTMs on mitochondrial proteins (Weinert, Schölz et al. 2013), with many of the sites identified in this study also having the capacity to be succinylated, or modified by other lysine modifications. Further discovery in the field is needed to ascertain which acetylation sites are functionally relevant for each SIRT3 target in order to provide the full story of the role of acetylation in a particular pathway or disease model.

5.3.13 Limitations of MS techniques (II): The problem of stoichiometry of PTMs

One caveat of MS analysis using antibody enrichment strategies are that the levels of acetylation detected in this way can only ever be relative, as we can only compare the material that has been immnoprecipitated by the antibody, and get no information on the absolute amount of the PTM across the entire proteome. Attempts to measure the stoichiometry of acetyl modifications are now being made in lower organisms using new methodologies (Baeza, Dowell et al. 2014, Weinert, Lesmantavicius et al. 2014) and are revealing that lysine acetylation occurs at quite low stoichiometry at least in bacteria and yeast, but correlates well with acetyl-CoA levels in mitochondria. The significance of this is not entirely understood and does not preclude acetylation from

148 CHAPTER 5: MS Analysis

having a significant functional role in specific cellular compartments, however it is in contrast with other PTMs such as phosphorylation where high stoichiometry, site- specificity and tight kinase/phophatases regulation are important for the functional regulation by this PTM and may provide further support for low-level non-enzymatic acetylation (Choudhary, Weinert et al. 2014).

149 CHAPTER 5: MS Analysis

5.4 Summary and Conclusions

The aim of this chapter was to investigate the effect of 3 weeks of hepatic SIRT3 overexpression on lysine acetylation using mass spectrometry techniques. The well- studied SIRT3KO mouse was also included in this analysis both as a proof-of-principle for what is a new technique in our laboratory, and also to allow identification of those changes that are induced by both SIRT3 deletion and SIRT3 overexpression, potentially identifying proteins that are robustly regulated by SIRT3 levels in the liver. In Chapter 4, we showed that SIRT3 overexpression in liver had beneficial effects to decrease triglyceride accumulation and increase oxygen consumption in isolated primary hepatocyte cultures, but that these changes had little effect on in vivo parameters in HFD-fed mice including body weight gain, liver triglyceride accumulation or glucose tolerance. In this chapter we show that despite the lack of observed in vivo effects, SIRT3 overexpression does cause some changes in the acetylation of enzymes that contribute to diverse metabolic pathways such as the TCA cycle, the urea cycle and fatty acid oxidation among others. Although the role of SIRT3 in regulating acetylation in mouse liver has been investigated using MS techniques by many groups, this is the first SIRT3 overexpression model to look at acetylation using proteomics.

In conclusion, we have established a method for assessing changes in acetylation in the SIRT3 liver overexpression model using a label-free LC-MS/MS approach. The effect of SIRT3 deletion on acetylation was a generalised increase in acetylation in SIRT3KO compared to WT liver mitochondria, in line with published studies. Conversely, SIRT3 overexpression was seen to decrease protein acetylation in liver mitochondria, both in the abundance of total detected proteins containing acetyl modifications, the average acetylation for all acetylated peptides in a subset of known SIRT3 targets, and at the individual peptide level for known acetylation sites of functional importance. Hence, we have established that a lack of a downstream effect of SIRT3 overexpression on acetylation in liver does not appear to explain the lack of effect on metabolic parameters seen in vivo. Pathway analysis also revealed widespread enrichment of

150 CHAPTER 5: MS Analysis

acetylated proteins in metabolic pathways, particularly those associated with the electron transport chain, mitochondrial fatty acid oxidation, the TCA cycle, amino acid metabolism and ketone synthesis as has been previously reported to be modulated by SIRT3.

This trial analysis has shown that the MS technology is a promising avenue of future research to investigate the downstream functional effects of changing levels of SIRT3 in mouse liver. However, these results fail to fully account for the disparate effects seen in vivo in the mouse liver overexpression model, as quantifiable changes in acetylation do appear to be occurring in vivo in response to SIRT3 overexpression. The acetylation changes could explain the increased metabolism seen in isolated hepatocytes, but do not account for why there is no effect on metabolic parameters in the in vivo setting. However, due to the current limitations of what is known about specific SIRT3 targets, the true stoichiometry and occupancy of lysine PTMs, and the threshold of changes in acetylation that are required to produce functional change in vivo, more work is needed to unravel the complexity of acetylation and its role in liver mitochondrial metabolism.

151 CHAPTER 6: Discussion

CHAPTER 6 GENERAL DISCUSSION

The cost of treating obesity and its complications including type-2-diabetes and other metabolic diseases is a growing problem worldwide. Post-translational modification (PTMs) of proteins is a regulatory mechanism that is known to impact on many aspects of metabolism, with dysregulation in PTMs reported to be associated with metabolic disease (Norvell and McMahon 2010, Iyer, Fairlie et al. 2012, Peinado, Diaz-Ruiz et al. 2014). This thesis explored the role of the post-translational modification acetylation in liver and muscle metabolism using high-fat-fed animal models. Specifically, the role of SIRT3, a member of the sirtuin family of NAD+-dependent deacylases, was investigated for its role in regulating mitochondrial energy metabolism and insulin resistance in two important metabolic target tissues, skeletal muscle and liver. The goal of the present study was to determine whether overexpression of SIRT3 in vivo could prevent the metabolic defects induced by excess lipid supply.

Acute muscle-specific overexpression using a SIRT3 AAV to induce eleven-fold overexpression of SIRT3 was investigated as described in Chapter 3. SIRT3 overexpression of this magnitude had a slight but non-significant effect on global lysine acetylation as assessed by western blot. Acetylation of mitochondrial proteins is associated with a decrease in mitochondrial respiration, and it was postulated that decreasing acetylation (by overexpressing SIRT3) might increase mitochondrial respiration of fatty acids, reduce the accumulation of lipid in skeletal muscle, and improve insulin action in this tissue. While SIRT3 overexpression for 4 weeks was shown to increase mitochondrial respiration rates in isolated mitochondria, there was no effect over this timeframe to prevent the decrease in glucose uptake into muscle induced by a HFD, nor was there any effect of SIRT3 on muscle triglyceride and glycogen stores following the clamp. Hence, we showed that in our acute overexpression model that SIRT3-induced changes in mitochondrial metabolism as assessed in an ex vivo setting did not correlate with any measureable in vivo metabolic effects, at least in the context of skeletal muscle glucose metabolism during lipid- excess.

152 CHAPTER 6: Discussion

Chapters 4 and 5 dealt with the acute liver-specific overexpression of SIRT3 using the hydrodynamic tail vein injection (HTVI) method. Chapter 4 investigated the in vivo effect of increased protein levels of SIRT3 in the liver of mice fed a HFD over 3 and 13 weeks, and with long-term fasting. Using these models, it was found that in common with the muscle overexpression model already described, increasing SIRT3 levels by ~2-fold in liver had some effect on mitochondrial respiration rate in ex vivo conditions such as isolated mitochondria, liver homogenates and primary hepatocyte cultures, but no effect on in vivo parameters such as body weight, liver triglyceride accumulation or glucose tolerance. One physiological parameter that was increased with SIRT3 overexpression in liver, although only in the context of starvation, was levels of the ketone body β-hydroxybutyrate, which may point to greater SIRT3 regulation of pathways linked with calorie-restriction rather than calorie excess.

Chapter 5 used samples from the 3 week liver overexpression study to further investigate, using mass spectrometry techniques, the effect of increasing SIRT3 protein on mitochondrial acetylation. SIRT3KO liver tissue was used in this study to provide an important reference to the overexpression model. Decreases in acetylation of known SIRT3 targets within metabolic pathways were found to be associated with increased liver SIRT3 levels, while acetylation of metabolic target proteins was shown to be increased in SIRT3KO liver, in agreement with published studies (Lombard, Alt et al. 2007, Rardin, Newman et al. 2013). Pathway analysis confirmed the importance of acetylation in key metabolic pathways including an emphasis on fatty-acid oxidation and OXPHOS processes with SIRT3 overexpression in this model. The mass spectrometry study provided important evidence that the SIRT3 construct was able to induce downstream effects on acetylation in this model, however the functional relevance of these changes in acetylation both to specific SIRT3 targets and to mitochondrial metabolism more broadly, remains unresolved in the field as a whole.

In combination, Chapters 3, 4 and 5 of this thesis show that changes in acetylation do occur when SIRT3 is overexpressed specifically in liver and muscle, that the pathways that are affected by acetylation (at least in liver) do correspond with the ex vivo

153 CHAPTER 6: Discussion

increases in mitochondrial respiration, but that these changes have no effect on lipid- induced defects in the in vivo setting, at least in the timeframes that were studied in this thesis.

A number of different methods can be used to manipulate and investigate protein function in animal models, each with their own advantages and disadvantages. These include global, conditional and tissue-specific knockout and overexpression models, and acute manipulation models. This thesis employed acute, tissue-specific manipulation of SIRT3 to investigate its function in metabolism. Tissue specific overexpression studies have the advantage of allowing targeted investigation of the role of a protein in a developmentally mature tissue, without the confounding factors of central or developmental effects, and without the expense and time required to generate transgenic mouse lines. Another advantage of overexpression studies is that they may replicate physiological changes in protein levels more than deletion of a protein from early development, such as is the case with the SIRT3KO model. Such constitutive global knockout models that involve deletion of the gene from early development may represent a more extreme phenotype (Shen, Xiao et al. 2008, Savage 2009). For instance, calorie restriction (CR) has been shown to cause an increase in SIRT3 protein levels, while exercise can increase SIRT3 protein approximately 2-fold (Palacios, Carmona et al. 2009). Conversely, HFD-feeding for 13 weeks was reported to decrease SIRT3 levels by ~50% (Hirschey, Shimazu et al. 2011). These physiological fluctuations in SIRT3 level are similar to what is achieved in the acute SIRT3 liver overexpression model presented here, and is markedly different in a physiological sense to complete gene deletion.

Although studies in all chapters used tissue specific overexpression, there were significant differences in the model between each tissue. Liver overexpression was induced using the HTVI technique to induce overexpression from 1 -13 weeks. This technique relies on the unique properties of the high volume injection, and the capacity of the liver to absorb it, to allow naked DNA to become incorporated into hepatocytes (Budker, Subbotin et al. 2006). The level of overexpression on average produced with this technique ranged from 1.4 to 2.6-fold depending on the time point

154 CHAPTER 6: Discussion

studied. Conversely, the adeno-associated viral (AAV) technique to induce overexpression in skeletal muscle generated much higher average overexpression of up to 11-fold. In this thesis SIRT3 overexpression in muscle was only investigated in the acute time-frame of 4 weeks, however the AAV technique has the potential to allow longer-term overexpression studies to be carried out up to at least 23 weeks with the viral serotype used in this thesis (Wang, Louboutin et al. 2011). Even with quite dramatic differences in overexpression level between muscle and liver, remarkably similar results were obtained in the two tissues, with both showing a tendency for the overexpression to affect respiration in an isolated system but not glucose metabolism in the whole animal. Hence, although the liver overexpression model could be classed as a physiological increase in SIRT3 protein, and the muscle overexpression model as a supra-physiological model of SIRT3 overexpression, very similar results were achieved with each.

The fact that SIRT3 levels can be modulated so significantly in these studies and yet produce minimal effect in vivo highlights a potential caveat of these overexpression studies, that changing the total protein level is not necessarily sufficient to change the activity of the SIRT3 enzyme. Currently, the precise regulation of SIRT3 activity, independent of SIRT3 protein levels, is not well described. What is known is that SIRT3 protein can be transcriptionally regulated in brown adipose, muscle and liver cells by peroxisome proliferator-activated receptor gamma coactivator 1α (PGC-1α), which is itself activated through deacetylation by SIRT1 (Kong, Wang et al. 2010, Giralt, Hondares et al. 2011). Additionally, SIRT1 activity has been shown to be regulated post-translationally by phosphorylation, ubiquitination and SUMOylation (Yang, Fu et al. 2007, Nasrin, Kaushik et al. 2009, Lin, Yang et al. 2012), while SIRT6 activity is stimulated by physiological concentrations of free-fatty acids including myristic, oleic, and linoleic acids (Feldman, Baeza et al. 2013). Although these processes have not been formally identified with regard to SIRT3 it is possible that SIRT3 may also be regulated in a similar fashion (Feldman, Dittenhafer-Reed et al. 2012, Feldman, Baeza et al. 2013).

155 CHAPTER 6: Discussion

One regulatory process that is likely the most pertinent in the context of changing nutrient status and the regulation of SIRT3 is modulation by NAD+ levels. It has been known since the sirtuins were first identified that their activity is dependent on NAD+ (Onyango, Celic et al. 2002, Schwer, North et al. 2002). NAD+ binds to the NAD binding pocket of sirtuins, including SIRT3, and is used as a substrate in the deacylase reaction, which produces deacylated lysine, nicotinamide and 2’-O-acetyl-ADP-ribose (OAADPr) (Jackson and Denu 2002). Sirtuins can thus be regulated both by levels of the co- substrate NAD+, but also nicotinamide (NAM), which is an inhibitor of the deacetylase reaction (Feldman, Dittenhafer-Reed et al. 2012).

Research into the role of NAD+ metabolism in health has seen a renewed interest, partly due to the emergence of the NAD+ consuming enzymes including the sirtuins and poly (ADP-ribose) polymerases (PARPs) (Canto, Houtkooper et al. 2012, Houtkooper and Auwerx 2012). There is evidence that manipulating the NAD+ salvage pathway to increase levels of NAD+ may have important metabolic benefits, and increase the activity of the sirtuins (Houtkooper and Auwerx 2012). Specifically, the addition of salvage pathway enzymes such as nicotinamide phosphoribosyl-transferase (NAMPT), or proteins such as nicotinamide riboside (NR) or nicotinamide mononucleotide (NMN) have been shown to increase NAD+ levels and have beneficial metabolic effects (Yoshino, Mills et al. 2011, Gomes, Price et al. 2013, Cerutti, Pirinen et al. 2014, Harkcom, Ghosh et al. 2014), however muscle-specific NAMPT overexpression was recently shown to be ineffective in increasing mitochondrial oxidation or alleviating defects in HFD-fed mice (Frederick, Davis et al. 2015). Whatever the effects of NAD+ manipulation may be, it is likely that sirtuins may respond to or mediate these effects.

The results in this thesis suggest that simply increasing the level of SIRT3 protein by introducing an exogenous construct expressing SIRT3 does not affect metabolic phenotype in the whole animal. We can speculate based on our results that in the context of SIRT3 functional activity, it is not the absolute level of SIRT3 protein that is limiting the action of this enzyme, but possibly the highly regulated levels of co- substrate NAD+ or other regulating factors that are responsible for SIRT3’s effects. In

156 CHAPTER 6: Discussion

this regard NAD+ has also been shown to exist in discrete intracellular pools within the nucleus, the cytosol and the mitochondria, and because NAD+ cannot freely cross the mitochondrial membrane its levels in mitochondria appear to be highly regulated, although these processes have not been fully elucidated (Yang, Yang et al. 2007, White and Schenk 2012, Dolle, Rack et al. 2013). As more becomes known about how mitochondrial NAD+ levels are maintained and regulated, greater understanding will emerge on the role of SIRT3, SIRT4 and SIRT5 in the mitochondria under fluctuating nutritional conditions and NAD+ levels.

One recent theory that has been discussed in the literature is the idea that acetylation may occur by non-enzymatic processes in response to high levels of acetyl-CoA in the mitochondrial matrix (Wagner and Payne 2013, Wagner and Hirschey 2014, Weinert, Lesmantavicius et al. 2014). This model postulates that sirtuins may have evolved as a quality control mechanism to correct spontaneous acyl modifications that occur under conditions of metabolic stress, i.e. when levels of acyl-CoAs are elevated in the mitochondrial matrix following caloric restriction, fasting and caloric excess (Wagner and Hirschey 2014). Recent studies have also confirmed that in lower organisms at least, levels of acetylation correlate with the amount of acetyl-CoA that is present in the cell (Baeza, Dowell et al. 2014, Weinert, Lesmantavicius et al. 2014). Hence in addition to regulation of SIRT3 deacetylase by transcriptional, post-transcriptional and NAD+-dependency, acetylation may also be regulated by changes in the discrete mitochondrial and extra-mitochondrial pools of acetyl-CoA (Choudhary, Weinert et al. 2014).

Hand-in-hand with this theory of non-enzymatic acetylation, comes renewed interest in the true extent of acetylation and its functional significance. The stoichiometry of acetylation changes across the proteome is of particular importance in understanding the functional significance of acetylation. As illustrated in Figure 6.1, acetylation changes detected using MS immunoprecipitation enrichment methods such as we have described in Chapter 5 only show differences in relative acetylation between two disease states or genotypes. Thus, a difference in acetylation site occupancy from 1% to 5% on a given protein, and a change in acetylation occupancy from 10% to 50% will

157 CHAPTER 6: Discussion

be detected equally as a 5-fold increase in acetylation, despite potentially having very different effects on protein function. New methods have been recently developed that attempt to measure the true stoichiometry of acetyl modifications rather than its relative change (Baeza, Dowell et al. 2014, Weinert, Lesmantavicius et al. 2014). These new methodologies have revealed that the actual amount of acetylation modifications across the proteome is often quite low compared to other PTMs such as phosphorylation, although this was measured in lower organisms and has not been confirmed in mammals (Choudhary, Weinert et al. 2014, Weinert, Lesmantavicius et al. 2014).

These theories about the evolution and function of acetylation in the proteome cannot explain why the manipulation of SIRT3 at the protein level does not prompt more significant changes in mitochondrial metabolism readouts in animal models, but they do highlight that there are still many fundamental questions in the field that we are only just beginning to answer. It is possible that the changes in acetylation seen with SIRT3 overexpression, for example in fatty acid oxidation proteins, have not yet

Figure 6.1 Illustration of the importance of stoichiometry for functional effects: The circles represent the pool of available individual proteins that may be acetylated, blue circles do not have an acetyl modification (at a particular site of importance for instance), red circles depict acetyl-modified site or protein. Relative acetylation methods cannot differentiate between the stoichiometry differences of total acetyl occupancy between the top and bottom examples, with both examples showing a 5-fold increase despite potential large functional differences.

158 CHAPTER 6: Discussion

reached a critical threshold that can impact significantly on flux through β-oxidation and the actual amount of hepatic lipid that accumulates in vivo. Further studies that can correlate the precise amount of acetyl-coA in the mitochondria with acetylation change and then to a functional readout in terms of enzyme activities in metabolic pathways will be needed to address this question. At the very least, the existence of changes in acetylation levels in Chapter 5 rules out the possibility that the SIRT3 overexpression models employed in this thesis are producing a defective or non- functional deacetylase, but much more needs to be understood about the functional relevance of these acetylation changes, especially in the context of the low stochiometry of acetylation in general, so that we can explain the lack of phenotypic changes in both the liver and muscle overexpression models which still remain unresolved.

The timeline of how changes in acetylation are accompanied by the induction of mitochondrial metabolism defects and metabolic disease in response to lipid or calorie excess is also an important area about which little is currently known. The initial studies on the role of SIRT3 in liver found that it took some time for the defects observed in hepatic fat oxidation and whole body glucose tolerance to occur, even in the extreme case of constitutive global SIRT3 deletion (Hirschey, Shimazu et al. 2011). In that study, the reduction in SIRT3 protein levels and increase in relative acetylation following high fat feeding was not evident until 13 weeks after high fat feeding was initiated (Hirschey, Shimazu et al. 2011). Furthermore, the reduction in SIRT3 levels associated with high fat feeding continued to decrease between 4 months and 6 months of high fat feeding, long after the initial effects of the lipid excess had occurred (Hirschey, Shimazu et al. 2011). In the case of the SIRT3KO model, differences in body weight compared to the WT were not significantly different until animals were 33 weeks of age and had been on HFD since weaning (Hirschey, Shimazu et al. 2011). Significantly greater impairments in glucose tolerance in HFD-fed SIRT3KO mice were also not shown until 12 months of age (Hirschey, Shimazu et al. 2011). This is of interest because it has been reported that HFD-induced insulin resistance is apparent in liver tissue much earlier than the reported changes that occur in SIRT3 levels and acetylation, with defects in glucose tolerance and hepatic insulin resistance occuring in

159 CHAPTER 6: Discussion

mice as early as 1 week after HFD commences (Turner, Kowalski et al. 2013). Indeed in Chapter 4 we showed a clear defect in glucose tolerance was already apparent after 3 weeks of HFD, although at this timepoint SIRT3 overexpression had no effect on this measure. This temporal discrepancy between the induction of insulin resistance and acetylation/SIRT3 changes points to the fact that acetylation changes are unlikely to be a primary mechanism underlying the development of insulin resistance in the liver, but instead may contribute to chronic changes associated with obesity.

Overall, the studies in this thesis suggest that merely increasing protein levels of SIRT3 is not enough to cause widespread functional changes in rodents, and that the regulation of SIRT3 by acetyl-CoA and NAD+ levels and possibly other still to be discovered mechanisms may be important. Similar results have been seen in the related deacetylase SIRT1 where a muscle overexpression transgenic mouse model showed no effect of SIRT1 on body composition, energy expenditure or muscle insulin sensitivity (White, McCurdy et al. 2013), even though whole-body transgenic mice did show beneficial metabolic effects (Bordone, Cohen et al. 2007, Pfluger, Herranz et al. 2008).

This is not the first study to find little metabolic perturbation with SIRT3 manipulation in liver and muscle. One recent study that was published during the course of this thesis was the generation of tissue specific knockout models of SIRT3 in liver and skeletal muscle by Fernandez-Marcos et al (Fernandez-Marcos, Jeninga et al. 2012). Despite thoroughly investigating these muscle and liver SIRT3KOs for a metabolic phenotype in-line with what had previously been reported in the global SIRT3KO mouse, no evidence was found for any metabolic phenotype in the tissue-specific deletion models. Similarly to the liver studies in this thesis, the tissue-specific SIRT3KOs did show significant changes in acetylation, however this was not accompanied by changes in gene expression, body weight, energy expenditure, glucose metabolism, ketone metabolism or acylcarnitine profile among the other parameters shown to be effected by global SIRT3KO (Fernandez-Marcos, Jeninga et al. 2012). This raises the possibility that other tissues in the body other than liver or muscle may be important

160 CHAPTER 6: Discussion

sites of SIRT3-mediated regulation and may be driving the defects seen in the global SIRT3KO mice. Indeed, in the case of the well-studied SIRT1, it was the brain specific SIRT1 overexpressing transgenic model (BRASTO mouse) that was shown to recapitulate the beneficial effects of calorie restriction in a manner not seen in the whole-body SIRT1 overexpressing mice (Satoh, Brace et al. 2013). However in the case of SIRT3 deficiency there are still only a few models that have been generated and there are some differences between them. Two independent SIRT3KO mouse lines exist that were both derived from the 129/Sv genetic background (Lombard, Alt et al. 2007, Ahn, Kim et al. 2008). The mice used in the liver overexpression studies in this thesis are on a C57BL6 background, while mice generated for the liver and muscle specific SIRT3KO mice were backcrossed onto a C57BL6 background following generation (Fernandez-Marcos, Jeninga et al. 2012). Strain differences are known to underlie some major differences in metabolic measurements and may be a significant factor in this discrepancy between models (Montgomery, Hallahan et al. 2013).

Another detail to bear in mind when discussing SIRT3 animal models is that the initial SIRT3KO in vivo phenotypes were actually quite mild. Initial reports on one line of SIRT3KO mice reported no phenotype (Lombard, Alt et al. 2007). Both independent SIRT3KO models however showed significant reductions in ATP levels in tissues including heart, kidney, liver (Ahn, Kim et al. 2008) and fasted liver (Hirschey, Shimazu et al. 2010), which does not support the idea that another tissue may be responsible for the whole body defects seen in the KOs (Ahn, Kim et al. 2008, Hirschey, Shimazu et al. 2011). In addition, the reported in vivo phenotypes in the SIRT3KO mouse are often only manifested in response to an intervention of reasonable severity such as very long-term HFD, long-term fasting or cold-challenge (Hirschey, Shimazu et al. 2010, Qiu, Brown et al. 2010, Hallows, Yu et al. 2011). In our own colony of SIRT3KO mice derived from the same colony as many of these published studies, glucose and insulin tolerance tests have been performed that show no defect at all between chow-fed WT and SIRT3KO mice at various ages and under standard housing conditions as well as with the added condition of thermoneutral housing conditions (unpublished data not shown) that is required to uncover some intrinsic defects in metabolism in animal models (Lodhi and Semenkovich 2009, Cannon and Nedergaard 2011).

161 CHAPTER 6: Discussion

While it is still unclear what is the driving force of the discrepancy between different metabolic SIRT3 animal models, it is clear from this thesis that SIRT3 overexpression in liver and muscle in an acute time frame does not have a major effect on whole-body metabolic endpoints. Muscle-specific overexpression studies previously performed in our laboratory have shown that acute overexpression of genes such as PGC1β, manganese superoxide dismutase (MnSOD) and carnitine palmitoyltransferase 1 (CPT1) for short time frames of 1 – 3 weeks were able to ameliorate HFD-induced insulin resistance (Wright, Brandon et al. 2011, Boden, Brandon et al. 2012), and decrease intramuscular triglyceride stores (Bruce, Brolin et al. 2007). Conversely, changes in expression of genes expected to drive large changes in metabolism can often show little whole-body phenotype, for example the acetyl-CoA carboxylase 2 (ACC2) KO mice show changes in whole body fatty acid oxidation that do not affect adiposity or glucose tolerance (Hoehn, Turner et al. 2010), and the previously described example of the SIRT1 muscle overexpression model that had no significant phenotype (White, McCurdy et al. 2013). Thus in the context of metabolic homeostasis where metabolite and substrate levels are closely controlled, quite large changes in metabolic phenotype are probably required to drive whole-body effects.

In conclusion, the field of acetylation and its functional significance is only in its infancy compared to what has been learnt about other PTMs such as phosphorylation in the past decades. These are important questions to consider in the context of SIRT3 overexpression as investigated in this thesis. The fact that lysine acetylation is widespread, is evolutionarily conserved, and appears on multiple enzymes within discrete metabolic pathways as we saw in our pathway analysis in Chapter 5, points to acetylation playing an important role in mitochondrial biology. However, in a rapidly expanding new field such as this there are still many details to unravel, and it is clear from this thesis that merely increasing the levels of this important deacetylase, SIRT3, is not enough to shift the metabolic health of the entire organism. These findings have importance for those that look towards the sirtuins as potential therapeutics to treat diseases of ageing and metabolic disease, as modulating levels of SIRT3 will probably

162 CHAPTER 6: Discussion

not be of enormous benefit, and looking to activators (as is currently being done for SIRT1) or NAD+ manipulation as therapies may be of more benefit. As expanded on in a recent review, we still do not fully understand the true scope of lysine acetylation (Choudhary, Weinert et al. 2014), and indeed its regulation by SIRT3. This thesis reinforces the idea that metabolism is highly regulated by an orchestrated series of overlapping regulatory frameworks, and therefore the function of SIRT3 itself is probably regulated by several mechanisms, and more work needs to be done before we can reliably manipulate its activity in an in vivo setting.

163 REFERENCES

REFERENCES

Ahn, B. H., H. S. Kim, S. Song, I. H. Lee, J. Liu, A. Vassilopoulos, C. X. Deng and T. Finkel (2008). "A role for the mitochondrial deacetylase sirt3 in regulating energy homeostasis." Proc. Natl. Acad. Sci. U. S. A. 105(38): 14447-14452.

Ahuja, N., B. Schwer, S. Carobbio, D. Waltregny, B. J. North, V. Castronovo, P. Maechler and E. Verdin (2007). "Regulation of insulin secretion by sirt4, a mitochondrial adp-ribosyltransferase." J. Biol. Chem. 282(46): 33583-33592.

Alberti, K. G., P. Zimmet and J. Shaw (2005). "The metabolic syndrome--a new worldwide definition." Lancet 366(9491): 1059-1062.

Alberti, K. G. and P. Z. Zimmet (2014). "Diabetes: A look to the future." Lancet Diabetes Endocrinol 2(1): e1-2.

Alhazzazi, T. Y., P. Kamarajan, N. Joo, J. Y. Huang, E. Verdin, N. J. D'Silva and Y. L. Kapila (2011). "Sirtuin-3 (sirt3), a novel potential therapeutic target for oral cancer." Cancer 117(8): 1670-1678.

Allfrey, V. G., R. Faulkner and A. E. Mirsky (1964). "Acetylation and methylation of histones and their possible role in the regulation of rna synthesis." Proc. Natl. Acad. Sci. U. S. A. 51: 786-794.

Andersen, P. and B. Saltin (1985). "Maximal perfusion of skeletal muscle in man." The Journal of Physiology 366(1): 233-249.

Anderson, E. J., M. E. Lustig, K. E. Boyle, T. L. Woodlief, D. A. Kane, C. T. Lin, J. W. Price, 3rd, L. Kang, P. S. Rabinovitch, H. H. Szeto, J. A. Houmard, R. N. Cortright, D. H. Wasserman and P. D. Neufer (2009). "Mitochondrial h2o2 emission and cellular redox state link excess fat intake to insulin resistance in both rodents and humans." J. Clin. Invest. 119(3): 573-581.

Anderson, K. A. and M. D. Hirschey (2012). "Mitochondrial protein acetylation regulates metabolism." Essays Biochem. 52: 23-35.

Baeza, J., J. A. Dowell, M. J. Smallegan, J. Fan, D. Amador-Noguez, Z. Khan and J. M. Denu (2014). "Stoichiometry of site-specific lysine acetylation in an entire proteome." J. Biol. Chem. 289(31): 21326-21338.

164 REFERENCES

Banks, A. S., N. Kon, C. Knight, M. Matsumoto, R. Gutiérrez-Juárez, L. Rossetti, W. Gu and D. Accili (2008). "Sirt1 gain of function increases energy efficiency and prevents diabetes in mice." Cell Metabolism 8(4): 333-341.

Bao, J., I. Scott, Z. Lu, L. Pang, C. C. Dimond, D. Gius and M. N. Sack (2010). "Sirt3 is regulated by nutrient excess and modulates hepatic susceptibility to lipotoxicity." Free Radic. Biol. Med. 49(7): 1230-1237.

Barbi de Moura, M., R. Uppala, Y. Zhang, B. Van Houten and E. S. Goetzman (2014). "Overexpression of mitochondrial sirtuins alters glycolysis and mitochondrial function in hek293 cells." PLoS One 9(8): e106028.

Begriche, K., A. Igoudjil, D. Pessayre and B. Fromenty (2006). "Mitochondrial dysfunction in nash: Causes, consequences and possible means to prevent it." Mitochondrion 6(1): 1-28.

Begriche, K., J. Massart, M. A. Robin, F. Bonnet and B. Fromenty (2013). "Mitochondrial adaptations and dysfunctions in nonalcoholic fatty liver disease." Hepatology 58(4): 1497-1507.

Bell, E. L., B. M. Emerling, S. J. Ricoult and L. Guarente (2011). "Sirt3 suppresses hypoxia inducible factor 1alpha and tumor growth by inhibiting mitochondrial ros production." Oncogene 30(26): 2986-2996.

Benton, C. R., G. P. Holloway, X. X. Han, Y. Yoshida, L. A. Snook, J. Lally, J. F. Glatz, J. J. Luiken, A. Chabowski and A. Bonen (2010). "Increased levels of peroxisome proliferator-activated receptor gamma, coactivator 1 alpha (pgc-1alpha) improve lipid utilisation, insulin signalling and glucose transport in skeletal muscle of lean and insulin-resistant obese zucker rats." Diabetologia 53(9): 2008-2019.

Berg, J., J. Tymoczko and L. Stryer (2002). Biochemistry. New York, W H Freeman.

Berry, M. N. and D. S. Friend (1969). "High-yield preparation of isolated rat liver parenchymal cells: A biochemical and fine structural study." J. Cell Biol. 43(3): 506-520.

Bharathi, S. S., Y. Zhang, A. W. Mohsen, R. Uppala, M. Balasubramani, E. Schreiber, G. Uechi, M. E. Beck, M. J. Rardin, J. Vockley, E. Verdin, B. W. Gibson, M. D. Hirschey and E. S. Goetzman (2013). "Sirtuin 3 (sirt3) protein regulates long- chain acyl-coa dehydrogenase by deacetylating conserved lysines near the active site." J. Biol. Chem. 288(47): 33837-33847.

Biason-Lauber, A., M. Boni-Schnetzler, B. P. Hubbard, K. Bouzakri, A. Brunner, C. Cavelti-Weder, C. Keller, M. Meyer-Boni, D. T. Meier, C. Brorsson, K. Timper, G.

165 REFERENCES

Leibowitz, A. Patrignani, R. Bruggmann, G. Boily, H. Zulewski, A. Geier, J. M. Cermak, P. Elliott, J. L. Ellis, C. Westphal, U. Knobel, J. J. Eloranta, J. Kerr-Conte, F. Pattou, D. Konrad, C. M. Matter, A. Fontana, G. Rogler, R. Schlapbach, C. Regairaz, J. M. Carballido, B. Glaser, M. W. McBurney, F. Pociot, D. A. Sinclair and M. Y. Donath (2013). "Identification of a sirt1 mutation in a family with type 1 diabetes." Cell Metab 17(3): 448-455.

Biddinger, S. B., A. Hernandez-Ono, C. Rask-Madsen, J. T. Haas, J. O. Aleman, R. Suzuki, E. F. Scapa, C. Agarwal, M. C. Carey, G. Stephanopoulos, D. E. Cohen, G. L. King, H. N. Ginsberg and C. R. Kahn (2008). "Hepatic insulin resistance is sufficient to produce dyslipidemia and susceptibility to atherosclerosis." Cell Metab 7(2): 125-134.

Biddinger, S. B. and C. R. Kahn (2006). "From mice to men: Insights into the insulin resistance syndromes." Annu. Rev. Physiol. 68: 123-158.

Boden, G., X. Duan, C. Homko, E. J. Molina, W. Song, O. Perez, P. Cheung and S. Merali (2008). "Increase in endoplasmic reticulum stress-related proteins and genes in adipose tissue of obese, insulin-resistant individuals." Diabetes 57(9): 2438- 2444.

Boden, M. J., A. E. Brandon, J. D. Tid-Ang, E. Preston, D. Wilks, E. Stuart, M. E. Cleasby, N. Turner, G. J. Cooney and E. W. Kraegen (2012). "Overexpression of manganese superoxide dismutase ameliorates high-fat diet-induced insulin resistance in rat skeletal muscle." American journal of physiology. Endocrinology and metabolism 303(6): E798-805.

Bonnard, C., A. Durand, S. Peyrol, E. Chanseaume, M. A. Chauvin, B. Morio, H. Vidal and J. Rieusset (2008). "Mitochondrial dysfunction results from oxidative stress in the skeletal muscle of diet-induced insulin-resistant mice." J. Clin. Invest. 118(2): 789-800.

Bordone, L., D. Cohen, A. Robinson, M. C. Motta, E. van Veen, A. Czopik, A. D. Steele, H. Crowe, S. Marmor, J. Luo, W. Gu and L. Guarente (2007). "Sirt1 transgenic mice show phenotypes resembling calorie restriction." Aging cell 6(6): 759-767.

Boushel, R., E. Gnaiger, P. Schjerling, M. Skovbro, R. Kraunsoe and F. Dela (2007). "Patients with type 2 diabetes have normal mitochondrial function in skeletal muscle." Diabetologia 50(4): 790-796.

Boutant, M. and C. Canto (2014). "Sirt1 metabolic actions: Integrating recent advances from mouse models." Molecular metabolism 3(1): 5-18.

Boyle, K. E., J. P. Canham, L. A. Consitt, D. Zheng, T. R. Koves, T. P. Gavin, D. Holbert, P. D. Neufer, O. Ilkayeva, D. M. Muoio and J. A. Houmard (2011). "A high-fat diet

166 REFERENCES

elicits differential responses in genes coordinating oxidative metabolism in skeletal muscle of lean and obese individuals." J. Clin. Endocrinol. Metab. 96(3): 775-781.

Boyle, K. E., S. A. Newsom, R. C. Janssen, M. Lappas and J. E. Friedman (2013). "Skeletal muscle mnsod, mitochondrial complex ii, and sirt3 enzyme activities are decreased in maternal obesity during human pregnancy and gestational diabetes mellitus." J. Clin. Endocrinol. Metab. 98(10): E1601-1609.

Bradford, M. M. (1976). "A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding." Anal. Biochem. 72: 248-254.

Brand, Martin D. and David G. Nicholls (2011). "Assessing mitochondrial dysfunction in cells." Biochem. J. 435(Pt 2): 297-312.

Bruce, C. R., C. Brolin, N. Turner, M. E. Cleasby, F. R. van der Leij, G. J. Cooney and E. W. Kraegen (2007). "Overexpression of carnitine palmitoyltransferase i in skeletal muscle in vivo increases fatty acid oxidation and reduces triacylglycerol esterification." American journal of physiology. Endocrinology and metabolism 292(4): E1231-1237.

Bruce, C. R., A. J. Hoy, N. Turner, M. J. Watt, T. L. Allen, K. Carpenter, G. J. Cooney, M. A. Febbraio and E. W. Kraegen (2009). "Overexpression of carnitine palmitoyltransferase-1 in skeletal muscle is sufficient to enhance fatty acid oxidation and improve high-fat diet-induced insulin resistance." Diabetes 58(3): 550-558.

Budker, V. G., V. M. Subbotin, T. Budker, M. G. Sebestyen, G. Zhang and J. A. Wolff (2006). "Mechanism of plasmid delivery by hydrodynamic tail vein injection. Ii. Morphological studies." J. Gene Med. 8(7): 874-888.

Burnett, C., S. Valentini, F. Cabreiro, M. Goss, M. Somogyvari, M. D. Piper, M. Hoddinott, G. L. Sutphin, V. Leko, J. J. McElwee, R. P. Vazquez-Manrique, A. M. Orfila, D. Ackerman, C. Au, G. Vinti, M. Riesen, K. Howard, C. Neri, A. Bedalov, M. Kaeberlein, C. Soti, L. Partridge and D. Gems (2011). "Absence of effects of sir2 overexpression on lifespan in c. Elegans and drosophila." Nature 477(7365): 482-485.

Burniston, J. G., J. Connolly, H. Kainulainen, S. L. Britton and L. G. Koch (2014). "Label- free profiling of skeletal muscle using high-definition mass spectrometry." PROTEOMICS 14(20): 2339-2344.

Cannon, B. and J. Nedergaard (2011). "Nonshivering thermogenesis and its adequate measurement in metabolic studies." J. Exp. Biol. 214(Pt 2): 242-253.

167 REFERENCES

Canto, C., R. H. Houtkooper, E. Pirinen, D. Y. Youn, M. H. Oosterveer, Y. Cen, P. J. Fernandez-Marcos, H. Yamamoto, P. A. Andreux, P. Cettour-Rose, K. Gademann, C. Rinsch, K. Schoonjans, A. A. Sauve and J. Auwerx (2012). "The nad(+) precursor nicotinamide riboside enhances oxidative metabolism and protects against high-fat diet-induced obesity." Cell Metabolism 15(6): 838-847.

Caton, P. W., S. J. Richardson, J. Kieswich, M. Bugliani, M. L. Holland, P. Marchetti, N. G. Morgan, M. M. Yaqoob, M. J. Holness and M. C. Sugden (2013). "Sirtuin 3 regulates mouse pancreatic beta cell function and is suppressed in pancreatic islets isolated from human type 2 diabetic patients." Diabetologia 56(5): 1068- 1077.

Cerutti, R., E. Pirinen, C. Lamperti, S. Marchet, A. A. Sauve, W. Li, V. Leoni, E. A. Schon, F. ß. Dantzer, J. Auwerx, C. Viscomi and M. Zeviani (2014). "Nad+-dependent activation of sirt1 corrects the phenotype in a mouse model of mitochondrial disease." Cell Metabolism 19(6): 1042-1049.

Chang, H. C. and L. Guarente (2014). "Sirt1 and other sirtuins in metabolism." Trends Endocrinol. Metab. 25(3): 138-145.

Chappell, J. B. and S. V. Perry (1954). "Biochemical and osmotic properties of skeletal muscle mitochondria." Nature. 173(4414): 1094-1095.

Cheng, H.-L., R. Mostoslavsky, S. i. Saito, J. P. Manis, Y. Gu, P. Patel, R. Bronson, E. Appella, F. W. Alt and K. F. Chua (2003). "Developmental defects and p53 hyperacetylation in sir2 homolog (sirt1)-deficient mice." Proceedings of the National Academy of Sciences 100(19): 10794-10799.

Choudhary, C., C. Kumar, F. Gnad, M. L. Nielsen, M. Rehman, T. C. Walther, J. V. Olsen and M. Mann (2009). "Lysine acetylation targets protein complexes and co- regulates major cellular functions." Science 325(5942): 834-840.

Choudhary, C., B. T. Weinert, Y. Nishida, E. Verdin and M. Mann (2014). "The growing landscape of lysine acetylation links metabolism and cell signalling." Nat. Rev. Mol. Cell Biol. 15(8): 536-550.

Cimen, H., M. J. Han, Y. Yang, Q. Tong, H. Koc and E. C. Koc (2010). "Regulation of succinate dehydrogenase activity by sirt3 in mammalian mitochondria." Biochemistry (Mosc.) 49(2): 304-311.

Clark, L. C., Jr., R. Wolf, D. Granger and Z. Taylor (1953). "Continuous recording of blood oxygen tensions by polarography." J. Appl. Physiol. 6(3): 189-193.

Cleasby, M. E., J. R. Davey, T. A. Reinten, M. W. Graham, D. E. James, E. W. Kraegen and G. J. Cooney (2005). "Acute bidirectional manipulation of muscle glucose

168 REFERENCES

uptake by in vivo electrotransfer of constructs targeting glucose transporter genes." Diabetes 54(9): 2702-2711.

Cleasby, M. E., Q. Lau, E. Polkinghorne, S. A. Patel, S. J. Leslie, N. Turner, G. J. Cooney, A. Xu and E. W. Kraegen (2011). "The adaptor protein appl1 increases glycogen accumulation in rat skeletal muscle through activation of the pi3-kinase signalling pathway." The Journal of Endocrinology 210(1): 81-92.

Cooper, H. M., J. Y. Huang, E. Verdin and J. N. Spelbrink (2009). "A new splice variant of the mouse sirt3 gene encodes the mitochondrial precursor protein." PLoS One 4(3): e4986.

Cooper, H. M. and J. N. Spelbrink (2008). "The human sirt3 protein deacetylase is exclusively mitochondrial." Biochem. J. 411(2): 279-285.

Croft, D., A. F. Mundo, R. Haw, M. Milacic, J. Weiser, G. Wu, M. Caudy, P. Garapati, M. Gillespie, M. R. Kamdar, B. Jassal, S. Jupe, L. Matthews, B. May, S. Palatnik, K. Rothfels, V. Shamovsky, H. Song, M. Williams, E. Birney, H. Hermjakob, L. Stein and P. D'Eustachio (2014). "The reactome pathway knowledgebase." Nucleic Acids Res. 42(Database issue): D472-477.

Croft, D., G. O'Kelly, G. Wu, R. Haw, M. Gillespie, L. Matthews, M. Caudy, P. Garapati, G. Gopinath, B. Jassal, S. Jupe, I. Kalatskaya, S. Mahajan, B. May, N. Ndegwa, E. Schmidt, V. Shamovsky, C. Yung, E. Birney, H. Hermjakob, P. D'Eustachio and L. Stein (2011). "Reactome: A database of reactions, pathways and biological processes." Nucleic Acids Res. 39(Database issue): D691-697.

Dai, H., L. Kustigian, D. Carney, A. Case, T. Considine, B. P. Hubbard, R. B. Perni, T. V. Riera, B. Szczepankiewicz, G. P. Vlasuk and R. L. Stein (2010). "Sirt1 activation by small molecules: Kinetic and biophysical evidence for direct interaction of enzyme and activator." J. Biol. Chem. 285(43): 32695-32703.

Dai, S. H., T. Chen, Y. H. Wang, J. Zhu, P. Luo, W. Rao, Y. F. Yang, Z. Fei and X. F. Jiang (2014). "Sirt3 protects cortical neurons against oxidative stress via regulating mitochondrial ca2+ and mitochondrial biogenesis." International journal of molecular sciences 15(8): 14591-14609. de Moura, M. B. and B. Van Houten (2014). "Bioenergetic analysis of intact mammalian cells using the seahorse xf24 extracellular flux analyzer and a luciferase atp assay." Methods Mol. Biol. 1105: 589-602.

Dedkova, E. N. and L. A. Blatter (2014). "Role of beta-hydroxybutyrate, its polymer poly-beta-hydroxybutyrate and inorganic polyphosphate in mammalian health and disease." Front. Physiol. 5: 260.

169 REFERENCES

DeFronzo, R. A. and D. Tripathy (2009). "Skeletal muscle insulin resistance is the primary defect in type 2 diabetes." Diabetes Care 32 Suppl 2: S157-163.

Denu, J. M. and J. M. Gottesfeld (2012). "Minireview series on sirtuins: From biochemistry to health and disease." J. Biol. Chem. 287(51): 42417-42418.

Dole, V. P. (1956). "A relation between non-esterified fatty acids in plasma and the metabolism of glucose." J. Clin. Invest. 35(2): 150-154.

Dolle, C., J. G. Rack and M. Ziegler (2013). "Nad and adp-ribose metabolism in mitochondria." The FEBS journal 280(15): 3530-3541.

Dominy, J. E., Jr., Y. Lee, M. P. Jedrychowski, H. Chim, M. J. Jurczak, J. P. Camporez, H. B. Ruan, J. Feldman, K. Pierce, R. Mostoslavsky, J. M. Denu, C. B. Clish, X. Yang, G. I. Shulman, S. P. Gygi and P. Puigserver (2012). "The deacetylase sirt6 activates the acetyltransferase gcn5 and suppresses hepatic gluconeogenesis." Mol. Cell 48(6): 900-913.

Du, J., H. Jiang and H. Lin (2009). "Investigating the adp-ribosyltransferase activity of sirtuins with nad analogues and 32p-nad." Biochemistry 48(13): 2878-2890.

Du, J., Y. Zhou, X. Su, J. J. Yu, S. Khan, H. Jiang, J. Kim, J. Woo, J. H. Kim, B. H. Choi, B. He, W. Chen, S. Zhang, R. A. Cerione, J. Auwerx, Q. Hao and H. Lin (2011). "Sirt5 is a nad-dependent protein lysine demalonylase and desuccinylase." Science 334(6057): 806-809.

Dyck, D. J. and A. Bonen (1998). "Muscle contraction increases palmitate esterification and oxidation and triacylglycerol oxidation." Am. J. Physiol. 275(5 Pt 1): E888- 896.

Dzugaj, A. (2006). "Localization and regulation of muscle fructose-1,6-bisphosphatase, the key enzyme of glyconeogenesis." Adv. Enzyme Regul. 46: 51-71.

The economic cost of obesity (2006). A. Economics, Diabetes Australia.

Egnatchik, R. A., A. K. Leamy, Y. Noguchi, M. Shiota and J. D. Young (2014). "Palmitate- induced activation of mitochondrial metabolism promotes oxidative stress and apoptosis in h4iiec3 rat hepatocytes." Metabolism 63(2): 283-295.

Fabbrini, E., F. Magkos, B. S. Mohammed, T. Pietka, N. A. Abumrad, B. W. Patterson, A. Okunade and S. Klein (2009). "Intrahepatic fat, not visceral fat, is linked with metabolic complications of obesity." Proc. Natl. Acad. Sci. U. S. A. 106(36): 15430-15435.

170 REFERENCES

Favre, D., Y. Cherel, N. Provost, V. Blouin, N. Ferry, P. Moullier and A. Salvetti (2000). "Hyaluronidase enhances recombinant adeno-associated virus (raav)-mediated gene transfer in the rat skeletal muscle." Gene Ther. 7(16): 1417-1420.

Feldman, J. L., J. Baeza and J. M. Denu (2013). "Activation of the protein deacetylase sirt6 by long-chain fatty acids and widespread deacylation by mammalian sirtuins." J. Biol. Chem. 288(43): 31350-31356.

Feldman, J. L., K. E. Dittenhafer-Reed and J. M. Denu (2012). "Sirtuin catalysis and regulation." J. Biol. Chem. 287(51): 42419-42427.

Fernandez-Marcos, P. J., E. H. Jeninga, C. Canto, T. Harach, V. C. de Boer, P. Andreux, N. Moullan, E. Pirinen, H. Yamamoto, S. M. Houten, K. Schoonjans and J. Auwerx (2012). "Muscle or liver-specific sirt3 deficiency induces hyperacetylation of mitochondrial proteins without affecting global metabolic homeostasis." Sci. Rep. 2: 425. Epub 2012 May 2028.

Ferrannini, E. (2006). "Is insulin resistance the cause of the metabolic syndrome?" Ann. Med. 38(1): 42-51.

Ferrannini, E., S. M. Haffner, B. D. Mitchell and M. P. Stern (1991). "Hyperinsulinaemia: The key feature of a cardiovascular and metabolic syndrome." Diabetologia 34(6): 416-422.

Finley, L. W., A. Carracedo, J. Lee, A. Souza, A. Egia, J. Zhang, J. Teruya-Feldstein, P. I. Moreira, S. M. Cardoso, C. B. Clish, P. P. Pandolfi and M. C. Haigis (2011). "Sirt3 opposes reprogramming of cancer cell metabolism through hif1alpha destabilization." Cancer Cell 19(3): 416-428.

Finley, L. W., W. Haas, V. Desquiret-Dumas, D. C. Wallace, V. Procaccio, S. P. Gygi and M. C. Haigis (2011). "Succinate dehydrogenase is a direct target of sirtuin 3 deacetylase activity." PLoS One 6(8): e23295.

Finley, L. W. S., W. Haas, V. r. Desquiret-Dumas, D. C. Wallace, V. Procaccio, S. P. Gygi and M. C. Haigis (2011). "Succinate dehydrogenase is a direct target of sirtuin 3 deacetylase activity." PLoS One 6(8): e23295.

Finley, L. W. S. and M. C. Haigis (2012). "Metabolic regulation by sirt3: Implications for tumorigenesis." Trends Mol. Med. 18(9): 516-523.

Folch, J., M. Lees and G. H. Sloane Stanley (1957). "A simple method for the isolation and purification of total lipides from animal tissues." J. Biol. Chem. 226(1): 497- 509.

171 REFERENCES

Frangioudakis, G., J. M. Ye and G. J. Cooney (2005). "Both saturated and n-6 polyunsaturated fat diets reduce phosphorylation of insulin receptor substrate- 1 and protein kinase b in muscle during the initial stages of in vivo insulin stimulation." Endocrinology 146(12): 5596-5603.

Frederick, D. W., J. G. Davis, A. Davila, Jr., B. Agarwal, S. Michan, M. A. Puchowicz, E. Nakamaru-Ogiso and J. A. Baur (2015). "Increasing nad synthesis in muscle via nicotinamide phosphoribosyltransferase is not sufficient to promote oxidative metabolism." J. Biol. Chem. 290(3): 1546-1558.

Fritz, K. S., J. J. Galligan, M. D. Hirschey, E. Verdin and D. R. Petersen (2012). "Mitochondrial acetylome analysis in a mouse model of alcohol-induced liver injury utilizing sirt3 knockout mice." J. Proteome Res. 11(3): 1633-1643.

Fritz, K. S., M. F. Green, D. R. Petersen and M. D. Hirschey (2013). "Ethanol metabolism modifies hepatic protein acylation in mice." PLoS One 8(9): e75868.

Frye, R. A. (2000). "Phylogenetic classification of prokaryotic and eukaryotic sir2-like proteins." Biochem. Biophys. Res. Commun. 273(2): 793-798.

Giralt, A., E. Hondares, J. A. Villena, F. Ribas, J. Diaz-Delfin, M. Giralt, R. Iglesias and F. Villarroya (2011). "Peroxisome proliferator-activated receptor-gamma coactivator-1alpha controls transcription of the sirt3 gene, an essential component of the thermogenic brown adipocyte phenotype." J. Biol. Chem. 286(19): 16958-16966.

Glozak, M. A., N. Sengupta, X. Zhang and E. Seto (2005). "Acetylation and deacetylation of non-histone proteins." Gene 363: 15-23.

Gomes, A. P., N. L. Price, A. J. Ling, J. J. Moslehi, M. K. Montgomery, L. Rajman, J. P. White, J. S. Teodoro, C. D. Wrann, B. P. Hubbard, E. M. Mercken, C. M. Palmeira, R. de Cabo, A. P. Rolo, N. Turner, E. L. Bell and D. A. Sinclair (2013). "Declining nad(+) induces a pseudohypoxic state disrupting nuclear- mitochondrial communication during aging." Cell 155(7): 1624-1638.

Gonçalves, L. A., A. M. Vigário and C. Penha-Gonçalves (2007). "Improved isolation of murine hepatocytes for in vitro malaria liver stage studies." Malar. J. 6: 169- 169.

Goodpaster, B. H. (2013). "Mitochondrial deficiency is associated with insulin resistance." Diabetes 62(4): 1032-1035.

Gregor, M. F. and G. S. Hotamisligil (2011). "Inflammatory mechanisms in obesity." Annu. Rev. Immunol. 29: 415-445.

172 REFERENCES

Grunstein, M. (1997). "Histone acetylation in chromatin structure and transcription." Nature 389(6649): 349-352.

Guan, K.-L., W. Yu, Y. Lin, Y. Xiong and S. Zhao (2010). "Generation of acetyllysine antibodies and affinity enrichment of acetylated peptides." Nat. Protoc. 5(9): 1583-1595.

Gurd, B. J., G. P. Holloway, Y. Yoshida and A. Bonen (2012). "In mammalian muscle, sirt3 is present in mitochondria and not in the nucleus; and sirt3 is upregulated by chronic muscle contraction in an adenosine monophosphate-activated protein kinase-independent manner." Metabolism 61(5): 733-741.

Haas, J. T. and S. B. Biddinger (2009). "Dissecting the role of insulin resistance in the metabolic syndrome." Curr. Opin. Lipidol. 20(3): 206-210.

Hafner, A. V., J. Dai, A. P. Gomes, C. Y. Xiao, C. M. Palmeira, A. Rosenzweig and D. A. Sinclair (2010). "Regulation of the mptp by sirt3-mediated deacetylation of cypd at lysine 166 suppresses age-related cardiac hypertrophy." Aging (Albany NY) 2(12): 914-923.

Hagberg, C., A. Mehlem, A. Falkevall, L. Muhl and U. Eriksson (2013). "Endothelial fatty acid transport: Role of vascular endothelial growth factor b." Physiology (Bethesda) 28(2): 125-134.

Haigis, M. C., R. Mostoslavsky, K. M. Haigis, K. Fahie, D. C. Christodoulou, A. J. Murphy, D. M. Valenzuela, G. D. Yancopoulos, M. Karow, G. Blander, C. Wolberger, T. A. Prolla, R. Weindruch, F. W. Alt and L. Guarente (2006). "Sirt4 inhibits glutamate dehydrogenase and opposes the effects of calorie restriction in pancreatic beta cells." Cell 126(5): 941-954.

Hallows, W. C., B. N. Albaugh and J. M. Denu (2008). "Where in the cell is sirt3? - functional localization of an nad+-dependent protein deacetylase." Biochem. J. 411(2): e11-e13.

Hallows, W. C., S. Lee and J. M. Denu (2006). "Sirtuins deacetylate and activate mammalian acetyl-coa synthetases." Proc. Natl. Acad. Sci. U. S. A. 103(27): 10230-10235.

Hallows, W. C., W. Yu, B. C. Smith, M. K. Devries, J. J. Ellinger, S. Someya, M. R. Shortreed, T. Prolla, J. L. Markley, L. M. Smith, S. Zhao, K. L. Guan and J. M. Denu (2011). "Sirt3 promotes the urea cycle and fatty acid oxidation during dietary restriction." Mol. Cell 41(2): 139-149.

Hancock, C. R., D.-H. Han, M. Chen, S. Terada, T. Yasuda, D. C. Wright and J. O. Holloszy (2008). "High-fat diets cause insulin resistance despite an increase in muscle

173 REFERENCES

mitochondria." Proceedings of the National Academy of Sciences 105(22): 7815-7820.

Harkcom, W. T., A. K. Ghosh, M. S. Sung, A. Matov, K. D. Brown, P. Giannakakou and S. R. Jaffrey (2014). "Nad+ and sirt3 control microtubule dynamics and reduce susceptibility to antimicrotubule agents." Proc. Natl. Acad. Sci. U. S. A. 111(24): E2443-2452.

Hatano, E., C. A. Bradham, A. Stark, Y. Iimuro, J. J. Lemasters and D. A. Brenner (2000). "The mitochondrial permeability transition augments fas-induced apoptosis in mouse hepatocytes." J. Biol. Chem. 275(16): 11814-11823.

Haw, R., H. Hermjakob, P. D'Eustachio and L. Stein (2011). "Reactome pathway analysis to enrich biological discovery in proteomics data sets." Proteomics 11(18): 3598-3613.

He, W., J. C. Newman, M. Z. Wang, L. Ho and E. Verdin (2012). "Mitochondrial sirtuins: Regulators of protein acylation and metabolism." Trends Endocrinol. Metab. 23(9): 467-476.

Hebert, A. S., K. E. Dittenhafer-Reed, W. Yu, D. J. Bailey, E. S. Selen, M. D. Boersma, J. J. Carson, M. Tonelli, A. J. Balloon, A. J. Higbee, M. S. Westphall, D. J. Pagliarini, T. A. Prolla, F. Assadi-Porter, S. Roy, J. M. Denu and J. J. Coon (2013). "Calorie restriction and sirt3 trigger global reprogramming of the mitochondrial protein acetylome." Mol. Cell 49(1): 186-199.

Henriksen, P., S. A. Wagner, B. T. Weinert, S. Sharma, G. Bacinskaja, M. Rehman, A. H. Juffer, T. C. Walther, M. Lisby and C. Choudhary (2012). "Proteome-wide analysis of lysine acetylation suggests its broad regulatory scope in saccharomyces cerevisiae." Mol. Cell. Proteomics 11(11): 1510-1522.

Herranz, D. and M. Serrano (2010). "Sirt1: Recent lessons from mouse models." Nat. Rev. Cancer 10(12): 819-823.

Hewitt, N. J., M. J. Lechon, J. B. Houston, D. Hallifax, H. S. Brown, P. Maurel, J. G. Kenna, L. Gustavsson, C. Lohmann, C. Skonberg, A. Guillouzo, G. Tuschl, A. P. Li, E. LeCluyse, G. M. Groothuis and J. G. Hengstler (2007). "Primary hepatocytes: Current understanding of the regulation of metabolic enzymes and transporter proteins, and pharmaceutical practice for the use of hepatocytes in metabolism, enzyme induction, transporter, clearance, and hepatotoxicity studies." Drug Metab. Rev. 39(1): 159-234.

Hirschey, M. D., T. Shimazu, E. Goetzman, E. Jing, B. Schwer, D. B. Lombard, C. A. Grueter, C. Harris, S. Biddinger, O. R. Ilkayeva, R. D. Stevens, Y. Li, A. K. Saha, N. B. Ruderman, J. R. Bain, C. B. Newgard, R. V. Farese, Jr., F. W. Alt, C. R. Kahn and

174 REFERENCES

E. Verdin (2010). "Sirt3 regulates mitochondrial fatty-acid oxidation by reversible enzyme deacetylation." Nature 464(7285): 121-125.

Hirschey, M. D., T. Shimazu, E. Jing, C. A. Grueter, A. M. Collins, B. Aouizerat, A. Stancakova, E. Goetzman, M. M. Lam, B. Schwer, R. D. Stevens, M. J. Muehlbauer, S. Kakar, N. M. Bass, J. Kuusisto, M. Laakso, F. W. Alt, C. B. Newgard, R. V. Farese, Jr., C. R. Kahn and E. Verdin (2011). "Sirt3 deficiency and mitochondrial protein hyperacetylation accelerate the development of the metabolic syndrome." Mol. Cell 44(2): 177-190.

Hoehn, K. L., A. B. Salmon, C. Hohnen-Behrens, N. Turner, A. J. Hoy, G. J. Maghzal, R. Stocker, H. Van Remmen, E. W. Kraegen, G. J. Cooney, A. R. Richardson and D. E. James (2009). "Insulin resistance is a cellular antioxidant defense mechanism." Proc. Natl. Acad. Sci. U. S. A. 106(42): 17787-17792.

Hoehn, K. L., N. Turner, M. M. Swarbrick, D. Wilks, E. Preston, Y. Phua, H. Joshi, S. M. Furler, M. Larance, B. D. Hegarty, S. J. Leslie, R. Pickford, A. J. Hoy, E. W. Kraegen, D. E. James and G. J. Cooney (2010). "Acute or chronic upregulation of mitochondrial fatty acid oxidation has no net effect on whole-body energy expenditure or adiposity." Cell Metab 11(1): 70-76.

Hoeks, J., J. J. Briede, J. de Vogel, G. Schaart, M. Nabben, E. Moonen-Kornips, M. K. Hesselink and P. Schrauwen (2008). "Mitochondrial function, content and ros production in rat skeletal muscle: Effect of high-fat feeding." FEBS Lett. 582(4): 510-516.

Hokari, F., E. Kawasaki, A. Sakai, K. Koshinaka, K. Sakuma and K. Kawanaka (2010). "Muscle contractile activity regulates sirt3 protein expression in rat skeletal muscles." J Appl Physiol (1985) 109(2): 332-340.

Holloszy, J. O. (2013). ""Deficiency" of mitochondria in muscle does not cause insulin resistance." Diabetes 62(4): 1036-1040.

Hotamisligil, G. S., P. Arner, J. F. Caro, R. L. Atkinson and B. M. Spiegelman (1995). "Increased adipose tissue expression of tumor necrosis factor-alpha in human obesity and insulin resistance." J. Clin. Invest. 95(5): 2409-2415.

Houstis, N., E. D. Rosen and E. S. Lander (2006). "Reactive oxygen species have a causal role in multiple forms of insulin resistance." Nature 440(7086): 944-948.

Houtkooper, R. H. and J. Auwerx (2012). "Exploring the therapeutic space around nad+." J. Cell Biol. 199(2): 205-209.

Howitz, K. T., K. J. Bitterman, H. Y. Cohen, D. W. Lamming, S. Lavu, J. G. Wood, R. E. Zipkin, P. Chung, A. Kisielewski, L. L. Zhang, B. Scherer and D. A. Sinclair (2003).

175 REFERENCES

"Small molecule activators of sirtuins extend saccharomyces cerevisiae lifespan." Nature 425(6954): 191-196.

Hoy, A. J., C. R. Bruce, A. Cederberg, N. Turner, D. E. James, G. J. Cooney and E. W. Kraegen (2007). "Glucose infusion causes insulin resistance in skeletal muscle of rats without changes in akt and as160 phosphorylation." American journal of physiology. Endocrinology and metabolism 293(5): E1358-1364.

Huang da, W., B. T. Sherman and R. A. Lempicki (2009). "Bioinformatics enrichment tools: Paths toward the comprehensive functional analysis of large gene lists." Nucleic Acids Res. 37(1): 1-13.

Huang, D. W., B. T. Sherman and R. A. Lempicki (2008). "Systematic and integrative analysis of large gene lists using david bioinformatics resources." Nat. Protocols 4(1): 44-57.

Hubbard, B. P. and D. A. Sinclair (2014). "Small molecule sirt1 activators for the treatment of aging and age-related diseases." Trends Pharmacol. Sci. 35(3): 146-154.

Hulsen, T., J. de Vlieg and W. Alkema (2008). "Biovenn - a web application for the comparison and visualization of biological lists using area-proportional venn diagrams." BMC Genomics 9: 488.

Humphrey, S. J., G. Yang, P. Yang, D. J. Fazakerley, J. Stockli, J. Y. Yang and D. E. James (2013). "Dynamic adipocyte phosphoproteome reveals that akt directly regulates mtorc2." Cell Metab 17(6): 1009-1020.

IDF. (2014). "Idf diabetes atlas (updated)." 6th Edition. from http://www.idf.org/diabetesatlas.

Inoue, T., M. Hiratsuka, M. Osaki and M. Oshimura (2007). "The molecular biology of mammalian sirt proteins: Sirt2 in cell cycle regulation." Cell cycle (Georgetown, Tex.) 6(9): 1011-1018.

Iwahara, T., R. Bonasio, V. Narendra and D. Reinberg (2012). "Sirt3 functions in the nucleus in the control of stress-related gene expression." Mol. Cell. Biol. 32(24): 5022-5034.

Iyer, A., D. P. Fairlie and L. Brown (2012). "Lysine acetylation in obesity, diabetes and metabolic disease." Immunol. Cell Biol. 90(1): 39-46.

Iyer, V. V., H. Yang, M. G. Ierapetritou and C. M. Roth (2010). "Effects of glucose and insulin on hepg2-c3a cell metabolism." Biotechnol. Bioeng. 107(2): 347-356.

176 REFERENCES

Jackson, M. D. and J. M. Denu (2002). "Structural identification of 2'- and 3'-o-acetyl- adp-ribose as novel metabolites derived from the sir2 family of beta -nad+- dependent histone/protein deacetylases." J. Biol. Chem. 277(21): 18535-18544.

Jeong, S. M., C. Xiao, L. W. S. Finley, T. Lahusen, A. L. Souza, K. Pierce, Y.-H. Li, X. Wang, G. Laurent, N. J. German, X. Xu, C. Li, R.-H. Wang, J. Lee, A. Csibi, R. Cerione, J. Blenis, C. B. Clish, A. Kimmelman, C.-X. Deng and M. C. Haigis (2013). "Sirt4 has tumor-suppressive activity and regulates the cellular metabolic response to DNA damage by inhibiting mitochondrial glutamine metabolism." Cancer Cell 23(4): 450-463.

Jheng, H. F., P. J. Tsai, S. M. Guo, L. H. Kuo, C. S. Chang, I. J. Su, C. R. Chang and Y. S. Tsai (2012). "Mitochondrial fission contributes to mitochondrial dysfunction and insulin resistance in skeletal muscle." Mol. Cell. Biol. 32(2): 309-319.

Jiang, H., S. Khan, Y. Wang, G. Charron, B. He, C. Sebastian, J. Du, R. Kim, E. Ge, R. Mostoslavsky, H. C. Hang, Q. Hao and H. Lin (2013). "Sirt6 regulates tnf-[agr] secretion through hydrolysis of long-chain fatty acyl lysine." Nature 496(7443): 110-113.

Jing, E., B. Emanuelli, M. D. Hirschey, J. Boucher, K. Y. Lee, D. Lombard, E. M. Verdin and C. R. Kahn (2011). "Sirtuin-3 (sirt3) regulates skeletal muscle metabolism and insulin signaling via altered mitochondrial oxidation and reactive oxygen species production." Proc. Natl. Acad. Sci. U. S. A. 108(35): 14608-14613.

Jing, E., B. T. O'Neill, M. J. Rardin, A. Kleinridders, O. R. Ilkeyeva, S. Ussar, J. R. Bain, K. Y. Lee, E. M. Verdin, C. B. Newgard, B. W. Gibson and C. R. Kahn (2013). "Sirt3 regulates metabolic flexibility of skeletal muscle through reversible enzymatic deacetylation." Diabetes 62(10): 3404-3417.

Joseph, A. M., P. J. Adhihetty, T. W. Buford, S. E. Wohlgemuth, H. A. Lees, L. M. Nguyen, J. M. Aranda, B. D. Sandesara, M. Pahor, T. M. Manini, E. Marzetti and C. Leeuwenburgh (2012). "The impact of aging on mitochondrial function and biogenesis pathways in skeletal muscle of sedentary high- and low-functioning elderly individuals." Aging cell 11(5): 801-809.

Kamp, F. and J. A. Hamilton (2006). "How fatty acids of different chain length enter and leave cells by free diffusion." Prostaglandins Leukot. Essent. Fatty Acids 75(3): 149-159.

Kanfi, Y., S. Naiman, G. Amir, V. Peshti, G. Zinman, L. Nahum, Z. Bar-Joseph and H. Y. Cohen (2012). "The sirtuin sirt6 regulates lifespan in male mice." Nature 483(7388): 218-221.

177 REFERENCES

Kanzleiter, T., E. Preston, D. Wilks, B. Ho, A. Benrick, J. Reznick, L. K. Heilbronn, N. Turner and G. J. Cooney (2010). "Overexpression of the orphan receptor nur77 alters glucose metabolism in rat muscle cells and rat muscle in vivo." Diabetologia 53(6): 1174-1183.

Kelley, D. E., J. He, E. V. Menshikova and V. B. Ritov (2002). "Dysfunction of mitochondria in human skeletal muscle in type 2 diabetes." Diabetes 51(10): 2944-2950.

Kelley, D. E. and L. J. Mandarino (2000). "Fuel selection in human skeletal muscle in insulin resistance: A reexamination." Diabetes 49(5): 677-683.

Kendrick, A. A., M. Choudhury, S. M. Rahman, C. E. McCurdy, M. Friederich, J. L. Van Hove, P. A. Watson, N. Birdsey, J. Bao, D. Gius, M. N. Sack, E. Jing, C. R. Kahn, J. E. Friedman and K. R. Jonscher (2011). "Fatty liver is associated with reduced sirt3 activity and mitochondrial protein hyperacetylation." Biochem. J. 433(3): 505-514.

Kim, H. S., K. Patel, K. Muldoon-Jacobs, K. S. Bisht, N. Aykin-Burns, J. D. Pennington, R. van der Meer, P. Nguyen, J. Savage, K. M. Owens, A. Vassilopoulos, O. Ozden, S. H. Park, K. K. Singh, S. A. Abdulkadir, D. R. Spitz, C. X. Deng and D. Gius (2010). "Sirt3 is a mitochondria-localized tumor suppressor required for maintenance of mitochondrial integrity and metabolism during stress." Cancer Cell 17(1): 41- 52.

Kim, J. Y., R. C. Hickner, R. L. Cortright, G. L. Dohm and J. A. Houmard (2000). "Lipid oxidation is reduced in obese human skeletal muscle." American journal of physiology. Endocrinology and metabolism 279(5): E1039-1044.

Kim, S. C., R. Sprung, Y. Chen, Y. Xu, H. Ball, J. Pei, T. Cheng, Y. Kho, H. Xiao, L. Xiao, N. V. Grishin, M. White, X. J. Yang and Y. Zhao (2006). "Substrate and functional diversity of lysine acetylation revealed by a proteomics survey." Mol. Cell 23(4): 607-618.

Kim, Y.-B., S. E. Nikoulina, T. P. Ciaraldi, R. R. Henry and B. B. Kahn (1999). "Normal insulin-dependent activation of akt/protein kinase b, with diminished activation of phosphoinositide 3-kinase, in muscle in type 2 diabetes." The Journal of Clinical Investigation 104(6): 733-741.

Klaunig, J. E., P. J. Goldblatt, D. E. Hinton, M. M. Lipsky, J. Chacko and B. F. Trump (1981). "Mouse liver cell culture. I. Hepatocyte isolation." In Vitro 17(10): 913- 925.

178 REFERENCES

Kong, X., R. Wang, Y. Xue, X. Liu, H. Zhang, Y. Chen, F. Fang and Y. Chang (2010). "Sirtuin 3, a new target of pgc-1alpha, plays an important role in the suppression of ros and mitochondrial biogenesis." PLoS One 5(7): e11707.

Kopelman, P. G. (2000). "Obesity as a medical problem." Nature 404(6778): 635-643.

Kouzarides, T. (2000). "Acetylation: A regulatory modification to rival phosphorylation?" EMBO J. 19(6): 1176-1179.

Koves, T. R., J. R. Ussher, R. C. Noland, D. Slentz, M. Mosedale, O. Ilkayeva, J. Bain, R. Stevens, J. R. Dyck, C. B. Newgard, G. D. Lopaschuk and D. M. Muoio (2008). "Mitochondrial overload and incomplete fatty acid oxidation contribute to skeletal muscle insulin resistance." Cell Metab 7(1): 45-56.

Kowalski, G. M. and C. R. Bruce (2014). "The regulation of glucose metabolism: Implications and considerations for the assessment of glucose homeostasis in rodents." Am. J. Physiol. Endocrinol. Metab. 307(10): E859-871.

Koyama, T., S. Kume, D. Koya, S. Araki, K. Isshiki, M. Chin-Kanasaki, T. Sugimoto, M. Haneda, T. Sugaya, A. Kashiwagi, H. Maegawa and T. Uzu (2011). "Sirt3 attenuates palmitate-induced ros production and inflammation in proximal tubular cells." Free Radic. Biol. Med. 51(6): 1258-1267.

Kraegen, E. W., P. W. Clark, A. B. Jenkins, E. A. Daley, D. J. Chisholm and L. H. Storlien (1991). "Development of muscle insulin resistance after liver insulin resistance in high-fat-fed rats." Diabetes 40(11): 1397-1403.

Kurtz, D. M., P. Rinaldo, W. J. Rhead, L. Tian, D. S. Millington, J. Vockley, D. A. Hamm, A. E. Brix, J. R. Lindsey, C. A. Pinkert, W. E. O'Brien and P. A. Wood (1998). "Targeted disruption of mouse long-chain acyl-coa dehydrogenase gene reveals crucial roles for fatty acid oxidation." Proc. Natl. Acad. Sci. U. S. A. 95(26): 15592-15597.

Laemmli, U. K. (1970). "Cleavage of structural proteins during the assembly of the head of bacteriophage t4." Nature 227(5259): 680-685.

Lagouge, M., C. Argmann, Z. Gerhart-Hines, H. Meziane, C. Lerin, F. Daussin, N. Messadeq, J. Milne, P. Lambert, P. Elliott, B. Geny, M. Laakso, P. Puigserver and J. Auwerx (2006). "Resveratrol improves mitochondrial function and protects against metabolic disease by activating sirt1 and pgc-1alpha." Cell 127(6): 1109- 1122.

Lanza, I. R., D. K. Short, K. R. Short, S. Raghavakaimal, R. Basu, M. J. Joyner, J. P. McConnell and K. S. Nair (2008). "Endurance exercise as a countermeasure for aging." Diabetes 57(11): 2933-2942.

179 REFERENCES

Laplante, M. and D. M. Sabatini (2009). "Mtor signaling at a glance." J. Cell Sci. 122(20): 3589-3594.

Laurent, G., N. J. German, A. K. Saha, V. C. de Boer, M. Davies, T. R. Koves, N. Dephoure, F. Fischer, G. Boanca, B. Vaitheesvaran, S. B. Lovitch, A. H. Sharpe, I. J. Kurland, C. Steegborn, S. P. Gygi, D. M. Muoio, N. B. Ruderman and M. C. Haigis (2013). "Sirt4 coordinates the balance between lipid synthesis and catabolism by repressing malonyl coa decarboxylase." Mol. Cell 50(5): 686-698.

Lettner, A. and M. Roden (2008). "Ectopic fat and insulin resistance." Curr Diab Rep 8(3): 185-191.

Li, Y., J. C. Silva, M. E. Skinner and D. B. Lombard (2013). "Mass spectrometry-based detection of protein acetylation." Methods Mol. Biol. 1077: 81-104.

Lin, L., K. Chen, W. A. Khalek, J. L. Ward, 3rd, H. Yang, B. Chabi, C. Wrutniak-Cabello and Q. Tong (2014). "Regulation of skeletal muscle oxidative capacity and muscle mass by sirt3." PLoS One 9(1): e85636.

Lin, Z., H. Yang, Q. Kong, J. Li, S. M. Lee, B. Gao, H. Dong, J. Wei, J. Song, D. D. Zhang and D. Fang (2012). "Usp22 antagonizes p53 transcriptional activation by deubiquitinating sirt1 to suppress cell apoptosis and is required for mouse embryonic development." Mol. Cell 46(4): 484-494.

Listenberger, L. L., D. S. Ory and J. E. Schaffer (2001). "Palmitate-induced apoptosis can occur through a ceramide-independent pathway." J. Biol. Chem. 276(18): 14890-14895.

Liu, F., Y. Song and D. Liu (1999). "Hydrodynamics-based transfection in animals by systemic administration of plasmid DNA." Gene Ther. 6(7): 1258-1266.

Liu, M., M. K. Montgomery, C. E. Fiveash, B. Osborne, G. J. Cooney, K. Bell-Anderson and N. Turner (2014). "Pparalpha-independent actions of omega-3 pufas contribute to their beneficial effects on adiposity and glucose homeostasis." Sci. Rep. 4: 5538.

Lodhi, I. J. and C. F. Semenkovich (2009). "Why we should put clothes on mice." Cell Metabolism 9(2): 111-112.

Loh, K., H. Deng, A. Fukushima, X. Cai, B. Boivin, S. Galic, C. Bruce, B. J. Shields, B. Skiba, L. M. Ooms, N. Stepto, B. Wu, C. A. Mitchell, N. K. Tonks, M. J. Watt, M. A. Febbraio, P. J. Crack, S. Andrikopoulos and T. Tiganis (2009). "Reactive oxygen species enhance insulin sensitivity." Cell Metab 10(4): 260-272.

180 REFERENCES

Lombard, D. B., F. W. Alt, H. L. Cheng, J. Bunkenborg, R. S. Streeper, R. Mostoslavsky, J. Kim, G. Yancopoulos, D. Valenzuela, A. Murphy, Y. Yang, Y. Chen, M. D. Hirschey, R. T. Bronson, M. Haigis, L. P. Guarente, R. V. Farese, Jr., S. Weissman, E. Verdin and B. Schwer (2007). "Mammalian sir2 homolog sirt3 regulates global mitochondrial lysine acetylation." Mol. Cell. Biol. 27(24): 8807-8814.

Lowell, B. B. and G. I. Shulman (2005). "Mitochondrial dysfunction and type 2 diabetes." Science 307(5708): 384-387.

Lundby, A., K. Lage, B. T. Weinert, D. B. Bekker-Jensen, A. Secher, T. Skovgaard, C. D. Kelstrup, A. Dmytriyev, C. Choudhary, C. Lundby and J. V. Olsen (2012). "Proteomic analysis of lysine acetylation sites in rat tissues reveals organ specificity and subcellular patterns." Cell reports 2(2): 419-431.

Mammucari, C. and R. Rizzuto (2010). "Signaling pathways in mitochondrial dysfunction and aging." Mech. Ageing Dev. 131(7-8): 536-543.

Martin, B. C., J. H. Warram, A. S. Krolewski, J. S. Soeldner, C. R. Kahn, B. C. Martin and R. N. Bergman (1992). "Role of glucose and insulin resistance in development of type 2 diabetes mellitus: Results of a 25-year follow-up study." The Lancet Neurology 340(8825): 925 - 929.

Masri, S. and P. Sassone-Corsi (2014). "Sirtuins and the circadian clock: Bridging chromatin and metabolism." Science signaling 7(342): re6.

Mathias, R. A., T. M. Greco, A. Oberstein, H. G. Budayeva, R. Chakrabarti, E. A. Rowland, Y. Kang, T. Shenk and I. M. Cristea (2014). "Sirtuin 4 is a lipoamidase regulating pyruvate dehydrogenase complex activity." Cell 159(7): 1615-1625.

McBurney, M. W., X. Yang, K. Jardine, M. Hixon, K. Boekelheide, J. R. Webb, P. M. Lansdorp and M. Lemieux (2003). "The mammalian sir2alpha protein has a role in embryogenesis and gametogenesis." Mol. Cell. Biol. 23(1): 38-54.

McMahon, J. M., E. Signori, K. E. Wells, V. M. Fazio and D. J. Wells (2001). "Optimisation of electrotransfer of plasmid into skeletal muscle by pretreatment with hyaluronidase -- increased expression with reduced muscle damage." Gene Ther. 8(16): 1264-1270.

Michan, S. and D. Sinclair (2007). "Sirtuins in mammals: Insights into their biological function." Biochem. J. 404(1): 1-13.

Minor, R. K., J. A. Baur, A. P. Gomes, T. M. Ward, A. Csiszar, E. M. Mercken, K. Abdelmohsen, Y. K. Shin, C. Canto, M. Scheibye-Knudsen, M. Krawczyk, P. M. Irusta, A. Martin-Montalvo, B. P. Hubbard, Y. Zhang, E. Lehrmann, A. A. White, N. L. Price, W. R. Swindell, K. J. Pearson, K. G. Becker, V. A. Bohr, M. Gorospe, J.

181 REFERENCES

M. Egan, M. I. Talan, J. Auwerx, C. H. Westphal, J. L. Ellis, Z. Ungvari, G. P. Vlasuk, P. J. Elliott, D. A. Sinclair and R. de Cabo (2011). "Srt1720 improves survival and healthspan of obese mice." Sci. Rep. 1: 70.

MirusBio. "Plive® in vivo expression and reporter vectors." 2014, from https://www.mirusbio.com/products/in-vivo/plive-in-vivo-expression- and-reporter-vectors.

Mitchell, S. J., A. Martin-Montalvo, E. M. Mercken, H. H. Palacios, T. M. Ward, G. Abulwerdi, R. K. Minor, G. P. Vlasuk, J. L. Ellis, D. A. Sinclair, J. Dawson, D. B. Allison, Y. Zhang, K. G. Becker, M. Bernier and R. de Cabo (2014). "The sirt1 activator srt1720 extends lifespan and improves health of mice fed a standard diet." Cell reports 6(5): 836-843.

Montgomery, M. K., N. L. Hallahan, S. H. Brown, M. Liu, T. W. Mitchell, G. J. Cooney and N. Turner (2013). "Mouse strain-dependent variation in obesity and glucose homeostasis in response to high-fat feeding." Diabetologia 56(5): 1129- 1139.

Morris, B. J. (2013). "Seven sirtuins for seven deadly diseases of aging." Free Radic. Biol. Med. 56(0): 133-171.

Mostoslavsky, R., K. F. Chua, D. B. Lombard, W. W. Pang, M. R. Fischer, L. Gellon, P. Liu, G. Mostoslavsky, S. Franco, M. M. Murphy, K. D. Mills, P. Patel, J. T. Hsu, A. L. Hong, E. Ford, H. L. Cheng, C. Kennedy, N. Nunez, R. Bronson, D. Frendewey, W. Auerbach, D. Valenzuela, M. Karow, M. O. Hottiger, S. Hursting, J. C. Barrett, L. Guarente, R. Mulligan, B. Demple, G. D. Yancopoulos and F. W. Alt (2006). "Genomic instability and aging-like phenotype in the absence of mammalian sirt6." Cell 124(2): 315-329.

Muoio, D. M. and P. D. Neufer (2012). "Lipid-induced mitochondrial stress and insulin action in muscle." Cell Metabolism 15(5): 595-605.

Muoio, D. M., R. C. Noland, J. P. Kovalik, S. E. Seiler, M. N. Davies, K. L. DeBalsi, O. R. Ilkayeva, R. D. Stevens, I. Kheterpal, J. Zhang, J. D. Covington, S. Bajpeyi, E. Ravussin, W. Kraus, T. R. Koves and R. L. Mynatt (2012). "Muscle-specific deletion of carnitine acetyltransferase compromises glucose tolerance and metabolic flexibility." Cell Metabolism 15(5): 764-777.

Nakagawa, T., D. J. Lomb, M. C. Haigis and L. Guarente (2009). "Sirt5 deacetylates carbamoyl phosphate synthetase 1 and regulates the urea cycle." Cell 137(3): 560-570.

Nakatani, Y., H. Kaneto, D. Kawamori, K. Yoshiuchi, M. Hatazaki, T. A. Matsuoka, K. Ozawa, S. Ogawa, M. Hori, Y. Yamasaki and M. Matsuhisa (2005). "Involvement

182 REFERENCES

of endoplasmic reticulum stress in insulin resistance and diabetes." J. Biol. Chem. 280(1): 847-851.

Nasrin, N., V. K. Kaushik, E. Fortier, D. Wall, K. J. Pearson, R. de Cabo and L. Bordone (2009). "Jnk1 phosphorylates sirt1 and promotes its enzymatic activity." PLoS One 4(12): e8414.

Nasrin, N., X. Wu, E. Fortier, Y. Feng, O. C. Bare, S. Chen, X. Ren, Z. Wu, R. S. Streeper and L. Bordone (2010). "Sirt4 regulates fatty acid oxidation and mitochondrial gene expression in liver and muscle cells." J. Biol. Chem. 285(42): 31995-32002.

NHPA (2013). Healthy communities: Overweight and obesity rates across australia, 2011-12 (in focus). N. H. P. Authority.

Nielsen, J. N., W. Derave, S. Kristiansen, E. Ralston, T. Ploug and E. A. Richter (2001). "Glycogen synthase localization and activity in rat skeletal muscle is strongly dependent on glycogen content." The Journal of Physiology 531(3): 757-769.

Nobes, C. D., W. W. Hay, Jr. and M. D. Brand (1990). "The mechanism of stimulation of respiration by fatty acids in isolated hepatocytes." J. Biol. Chem. 265(22): 12910-12915.

Noguchi, Y., J. D. Young, J. O. Aleman, M. E. Hansen, J. K. Kelleher and G. Stephanopoulos (2009). "Effect of anaplerotic fluxes and amino acid availability on hepatic lipoapoptosis." J. Biol. Chem. 284(48): 33425-33436.

North, B. J., B. L. Marshall, M. T. Borra, J. M. Denu and E. Verdin (2003). "The human sir2 ortholog, sirt2, is an nad+-dependent tubulin deacetylase." Mol. Cell 11(2): 437-444.

Norvell, A. and S. B. McMahon (2010). "Cell biology. Rise of the rival." Science 327(5968): 964-965.

Oakes, N. D., G. J. Cooney, S. Camilleri, D. J. Chisholm and E. W. Kraegen (1997). "Mechanisms of liver and muscle insulin resistance induced by chronic high-fat feeding." Diabetes 46(11): 1768-1774.

Ogura, M., Y. Nakamura, D. Tanaka, X. Zhuang, Y. Fujita, A. Obara, A. Hamasaki, M. Hosokawa and N. Inagaki (2010). "Overexpression of sirt5 confirms its involvement in deacetylation and activation of carbamoyl phosphate synthetase 1." Biochem. Biophys. Res. Commun. 393(1): 73-78.

Onyango, P., I. Celic, J. M. McCaffery, J. D. Boeke and A. P. Feinberg (2002). "Sirt3, a human sir2 homologue, is an nad-dependent deacetylase localized to mitochondria." Proc. Natl. Acad. Sci. U. S. A. 99(21): 13653-13658.

183 REFERENCES

Osborne, B., G. J. Cooney and N. Turner (2014). "Are sirtuin deacylase enzymes important modulators of mitochondrial energy metabolism?" Biochim. Biophys. Acta 1840(4): 1295-1302.

Ozcan, U., E. Yilmaz, L. Ozcan, M. Furuhashi, E. Vaillancourt, R. O. Smith, C. Z. Gorgun and G. S. Hotamisligil (2006). "Chemical chaperones reduce er stress and restore glucose homeostasis in a mouse model of type 2 diabetes." Science 313(5790): 1137-1140.

Pacholec, M., J. E. Bleasdale, B. Chrunyk, D. Cunningham, D. Flynn, R. S. Garofalo, D. Griffith, M. Griffor, P. Loulakis, B. Pabst, X. Qiu, B. Stockman, V. Thanabal, A. Varghese, J. Ward, J. Withka and K. Ahn (2010). "Srt1720, srt2183, srt1460, and resveratrol are not direct activators of sirt1." J. Biol. Chem. 285(11): 8340-8351.

Paik, W. K., D. Pearson, H. W. Lee and S. Kim (1970). "Nonenzymatic acetylation of histones with acetyl-coa." Biochim. Biophys. Acta 213(2): 513-522.

Palacios, O. M., J. J. Carmona, S. Michan, K. Y. Chen, Y. Manabe, J. L. Ward, 3rd, L. J. Goodyear and Q. Tong (2009). "Diet and exercise signals regulate sirt3 and activate ampk and pgc-1alpha in skeletal muscle." Aging (Albany NY) 1(9): 771- 783.

Paulin, R., P. Dromparis, G. Sutendra, V. Gurtu, S. Zervopoulos, L. Bowers, A. Haromy, L. Webster, S. Provencher, S. Bonnet and E. D. Michelakis (2014). "Sirtuin 3 deficiency is associated with inhibited mitochondrial function and pulmonary arterial hypertension in rodents and humans." Cell Metab 20(5): 827-839.

Peinado, J. R., A. Diaz-Ruiz, G. Frühbeck and M. M. Malagon (2014). "Mitochondria in metabolic disease: Getting clues from proteomic studies." PROTEOMICS 14(4- 5): 452-466.

Peng, C., Z. Lu, Z. Xie, Z. Cheng, Y. Chen, M. Tan, H. Luo, Y. Zhang, W. He, K. Yang, B. M. Zwaans, D. Tishkoff, L. Ho, D. Lombard, T. C. He, J. Dai, E. Verdin, Y. Ye and Y. Zhao (2011). "The first identification of lysine malonylation substrates and its regulatory enzyme." Mol. Cell. Proteomics 10(12): M111 012658.

Perez-Carreras, M., P. Del Hoyo, M. A. Martin, J. C. Rubio, A. Martin, G. Castellano, F. Colina, J. Arenas and J. A. Solis-Herruzo (2003). "Defective hepatic mitochondrial respiratory chain in patients with nonalcoholic steatohepatitis." Hepatology 38(4): 999-1007.

Perseghin, G., P. Scifo, F. De Cobelli, E. Pagliato, A. Battezzati, C. Arcelloni, A. Vanzulli, G. Testolin, G. Pozza, A. Del Maschio and L. Luzi (1999). "Intramyocellular triglyceride content is a determinant of in vivo insulin resistance in humans: A

184 REFERENCES

1h-13c nuclear magnetic resonance spectroscopy assessment in offspring of type 2 diabetic parents." Diabetes 48(8): 1600-1606.

Petersen, K. F., D. Befroy, S. Dufour, J. Dziura, C. Ariyan, D. L. Rothman, L. DiPietro, G. W. Cline and G. I. Shulman (2003). "Mitochondrial dysfunction in the elderly: Possible role in insulin resistance." Science 300(5622): 1140-1142.

Pfluger, P. T., D. Herranz, S. Velasco-Miguel, M. Serrano and M. H. Tschop (2008). "Sirt1 protects against high-fat diet-induced metabolic damage." Proc. Natl. Acad. Sci. U. S. A. 105(28): 9793-9798.

Phielix, E., V. B. Schrauwen-Hinderling, M. Mensink, E. Lenaers, R. Meex, J. Hoeks, M. E. Kooi, E. Moonen-Kornips, J. P. Sels, M. K. Hesselink and P. Schrauwen (2008). "Lower intrinsic adp-stimulated mitochondrial respiration underlies in vivo mitochondrial dysfunction in muscle of male type 2 diabetic patients." Diabetes 57(11): 2943-2949.

Pillai, V. B., N. R. Sundaresan, G. Kim, M. Gupta, S. B. Rajamohan, J. B. Pillai, S. Samant, P. V. Ravindra, A. Isbatan and M. P. Gupta (2010). "Exogenous nad blocks cardiac hypertrophic response via activation of the sirt3-lkb1-amp-activated kinase pathway." J. Biol. Chem. 285(5): 3133-3144.

Price, N. L., A. P. Gomes, A. J. Ling, F. V. Duarte, A. Martin-Montalvo, B. J. North, B. Agarwal, L. Ye, G. Ramadori, J. S. Teodoro, B. P. Hubbard, A. T. Varela, J. G. Davis, B. Varamini, A. Hafner, R. Moaddel, A. P. Rolo, R. Coppari, C. M. Palmeira, R. de Cabo, J. A. Baur and D. A. Sinclair (2012). "Sirt1 is required for ampk activation and the beneficial effects of resveratrol on mitochondrial function." Cell Metab 15(5): 675-690.

Qiu, X., K. Brown, M. D. Hirschey, E. Verdin and D. Chen (2010). "Calorie restriction reduces oxidative stress by sirt3-mediated sod2 activation." Cell Metabolism 12(6): 662-667.

Rardin, M. J., W. He, Y. Nishida, J. C. Newman, C. Carrico, S. R. Danielson, A. Guo, P. Gut, A. K. Sahu, B. Li, R. Uppala, M. Fitch, T. Riiff, L. Zhu, J. Zhou, D. Mulhern, R. D. Stevens, O. R. Ilkayeva, C. B. Newgard, M. P. Jacobson, M. Hellerstein, E. S. Goetzman, B. W. Gibson and E. Verdin (2013). "Sirt5 regulates the mitochondrial lysine succinylome and metabolic networks." Cell Metabolism 18(6): 920-933.

Rardin, M. J., J. C. Newman, J. M. Held, M. P. Cusack, D. J. Sorensen, B. Li, B. Schilling, S. D. Mooney, C. R. Kahn, E. Verdin and B. W. Gibson (2013). "Label-free quantitative proteomics of the lysine acetylome in mitochondria identifies substrates of sirt3 in metabolic pathways." Proc. Natl. Acad. Sci. U. S. A. 110(16): 6601-6606.

185 REFERENCES

Ravussin, E. and S. R. Smith (2002). "Increased fat intake, impaired fat oxidation, and failure of fat cell proliferation result in ectopic fat storage, insulin resistance, and type 2 diabetes mellitus." Ann. N. Y. Acad. Sci. 967: 363-378.

Reaven, G. M. (1988). "Banting lecture 1988. Role of insulin resistance in human disease." Diabetes 37(12): 1595-1607.

Reaven, G. M. (2005). "The insulin resistance syndrome: Definition and dietary approaches to treatment." Annu. Rev. Nutr. 25: 391-406.

Ren, J., L. Pulakat, A. Whaley-Connell and J. R. Sowers (2010). "Mitochondrial biogenesis in the metabolic syndrome and cardiovascular disease." J. Mol. Med. (Berl.) 88(10): 993-1001.

Rodgers, J. T., C. Lerin, W. Haas, S. P. Gygi, B. M. Spiegelman and P. Puigserver (2005). "Nutrient control of glucose homeostasis through a complex of pgc-1alpha and sirt1." Nature 434(7029): 113-118.

Rodgers, J. T. and P. Puigserver (2007). "Fasting-dependent glucose and lipid metabolic response through hepatic sirtuin 1." Proceedings of the National Academy of Sciences 104(31): 12861-12866.

Rogers, G. W., M. D. Brand, S. Petrosyan, D. Ashok, A. A. Elorza, D. A. Ferrick and A. N. Murphy (2011). "High throughput microplate respiratory measurements using minimal quantities of isolated mitochondria." PLoS One 6(7): e21746.

Rogina, B. and S. L. Helfand (2004). "Sir2 mediates longevity in the fly through a pathway related to calorie restriction." Proc. Natl. Acad. Sci. U. S. A. 101(45): 15998-16003.

Rose, A. J. and E. A. Richter (2005). "Skeletal muscle glucose uptake during exercise: How is it regulated?" Physiology 20(4): 260-270.

Rothgiesser, K. M., S. Erener, S. Waibel, B. Luscher and M. O. Hottiger (2010). "Sirt2 regulates nf-kappab dependent gene expression through deacetylation of p65 lys310." J. Cell Sci. 123(Pt 24): 4251-4258.

Ryu, D., Y. S. Jo, G. Lo Sasso, S. Stein, H. Zhang, A. Perino, J. U. Lee, M. Zeviani, R. Romand, M. O. Hottiger, K. Schoonjans and J. Auwerx (2014). "A sirt7- dependent acetylation switch of gabpbeta1 controls mitochondrial function." Cell Metab 20(5): 856-869.

Saltiel, A. R. and C. R. Kahn (2001). "Insulin signalling and the regulation of glucose and lipid metabolism." Nature 414(6865): 799-806.

186 REFERENCES

Samant, S. A., H. J. Zhang, Z. Hong, V. B. Pillai, N. R. Sundaresan, D. Wolfgeher, S. L. Archer, D. C. Chan and M. P. Gupta (2014). "Sirt3 deacetylates and activates opa1 to regulate mitochondrial dynamics during stress." Mol. Cell. Biol. 34(5): 807-819.

Samocha-Bonet, D., L. V. Campbell, T. A. Mori, K. D. Croft, J. R. Greenfield, N. Turner and L. K. Heilbronn (2012). "Overfeeding reduces insulin sensitivity and increases oxidative stress, without altering markers of mitochondrial content and function in humans." PLoS One 7(5): e36320.

Samuel, V. T., K. F. Petersen and G. I. Shulman (2010). "Lipid-induced insulin resistance: Unravelling the mechanism." Lancet 375(9733): 2267-2277.

Samuel, V. T. and G. I. Shulman (2012). "Mechanisms for insulin resistance: Common threads and missing links." Cell 148(5): 852-871.

Satoh, A., C. S. Brace, N. Rensing, P. Cliften, D. F. Wozniak, E. D. Herzog, K. A. Yamada and S. Imai (2013). "Sirt1 extends life span and delays aging in mice through the regulation of nk2 homeobox 1 in the dmh and lh." Cell Metabolism 18(3): 416- 430.

Saunders, L. R. and E. Verdin (2007). "Sirtuins: Critical regulators at the crossroads between cancer and aging." Oncogene 26(37): 5489-5504.

Savage, D. B. (2009). "Mouse models of inherited lipodystrophy." Dis. Model. Mech. 2(11-12): 554-562.

Savage, D. B., K. F. Petersen and G. I. Shulman (2007). "Disordered lipid metabolism and the pathogenesis of insulin resistance." Physiol. Rev. 87(2): 507-520.

Schenk, S., C. E. McCurdy, A. Philp, M. Z. Chen, M. J. Holliday, G. K. Bandyopadhyay, O. Osborn, K. Baar and J. M. Olefsky (2011). "Sirt1 enhances skeletal muscle insulin sensitivity in mice during caloric restriction." J. Clin. Invest. 121(11): 4281-4288.

Scher, M. B., A. Vaquero and D. Reinberg (2007). "Sirt3 is a nuclear nad+-dependent histone deacetylase that translocates to the mitochondria upon cellular stress." Genes Dev. 21(8): 920-928.

Schilling, B., M. J. Rardin, B. X. MacLean, A. M. Zawadzka, B. E. Frewen, M. P. Cusack, D. J. Sorensen, M. S. Bereman, E. Jing, C. C. Wu, E. Verdin, C. R. Kahn, M. J. Maccoss and B. W. Gibson (2012). "Platform-independent and label-free quantitation of proteomic data using ms1 extracted ion chromatograms in skyline: Application to protein acetylation and phosphorylation." Mol. Cell. Proteomics 11(5): 202-214.

187 REFERENCES

Schlicker, C., M. Gertz, P. Papatheodorou, B. Kachholz, C. F. W. Becker and C. Steegborn (2008). "Substrates and regulation mechanisms for the human mitochondrial sirtuins sirt3 and sirt5." J. Mol. Biol. 382(3): 790-801.

Schwer, B., J. Bunkenborg, R. O. Verdin, J. S. Andersen and E. Verdin (2006). "Reversible lysine acetylation controls the activity of the mitochondrial enzyme acetyl-coa synthetase 2." Proc. Natl. Acad. Sci. U. S. A. 103(27): 10224-10229.

Schwer, B., M. Eckersdorff, Y. Li, J. C. Silva, D. Fermin, M. V. Kurtev, C. Giallourakis, M. J. Comb, F. W. Alt and D. B. Lombard (2009). "Calorie restriction alters mitochondrial protein acetylation." Aging cell 8(5): 604-606.

Schwer, B., B. J. North, R. A. Frye, M. Ott and E. Verdin (2002). "The human silent information regulator (sir)2 homologue hsirt3 is a mitochondrial nicotinamide adenine dinucleotide‚äìdependent deacetylase." J. Cell Biol. 158(4): 647-657.

Scott, I., B. R. Webster, J. H. Li and M. N. Sack (2012). "Identification of a molecular component of the mitochondrial acetyl transferase program; a novel role for gcn5l1." Biochem. J. 443(3): 655-661.

Sebastian, C., B. M. Zwaans, D. M. Silberman, M. Gymrek, A. Goren, L. Zhong, O. Ram, J. Truelove, A. R. Guimaraes, D. Toiber, C. Cosentino, J. K. Greenson, A. I. MacDonald, L. McGlynn, F. Maxwell, J. Edwards, S. Giacosa, E. Guccione, R. Weissleder, B. E. Bernstein, A. Regev, P. G. Shiels, D. B. Lombard and R. Mostoslavsky (2012). "The histone deacetylase sirt6 is a tumor suppressor that controls cancer metabolism." Cell 151(6): 1185-1199.

Sebastian, D., M. I. Hernandez-Alvarez, J. Segales, E. Sorianello, J. P. Munoz, D. Sala, A. Waget, M. Liesa, J. C. Paz, P. Gopalacharyulu, M. Oresic, S. Pich, R. Burcelin, M. Palacin and A. Zorzano (2012). "Mitofusin 2 (mfn2) links mitochondrial and endoplasmic reticulum function with insulin signaling and is essential for normal glucose homeostasis." Proc. Natl. Acad. Sci. U. S. A. 109(14): 5523-5528.

Shen, X. Z., H. D. Xiao, P. Li, S. Billet, C. X. Lin, S. Fuchs and K. E. Bernstein (2008). "Tissue specific expression of angiotensin converting enzyme: A new way to study an old friend." Int. Immunopharmacol. 8(2): 171-176.

Shi, T., G. Q. Fan and S. D. Xiao (2010). "Sirt3 reduces lipid accumulation via ampk activation in human hepatic cells." J. Dig. Dis. 11(1): 55-62.

Shi, T., F. Wang, E. Stieren and Q. Tong (2005). "Sirt3, a mitochondrial sirtuin deacetylase, regulates mitochondrial function and thermogenesis in brown adipocytes." J. Biol. Chem. 280(14): 13560-13567.

188 REFERENCES

Shimazu, T., M. D. Hirschey, L. Hua, K. E. Dittenhafer-Reed, B. Schwer, D. B. Lombard, Y. Li, J. Bunkenborg, F. W. Alt, J. M. Denu, M. P. Jacobson and E. Verdin (2010). "Sirt3 deacetylates mitochondrial 3-hydroxy-3-methylglutaryl coa synthase 2 and regulates ketone body production." Cell Metabolism 12(6): 654-661.

Shin, J., M. He, Y. Liu, S. Paredes, L. Villanova, K. Brown, X. Qiu, N. Nabavi, M. Mohrin, K. Wojnoonski, P. Li, H. L. Cheng, A. J. Murphy, D. M. Valenzuela, H. Luo, P. Kapahi, R. Krauss, R. Mostoslavsky, G. D. Yancopoulos, F. W. Alt, K. F. Chua and D. Chen (2013). "Sirt7 represses myc activity to suppress er stress and prevent fatty liver disease." Cell reports 5(3): 654-665.

Shulga, N., R. Wilson-Smith and J. G. Pastorino (2010). "Sirtuin-3 deacetylation of cyclophilin d induces dissociation of hexokinase ii from the mitochondria." J. Cell Sci. 123(Pt 6): 894-902.

Shulman, G. I. (2000). "Cellular mechanisms of insulin resistance." J. Clin. Invest. 106(2): 171-176.

Sinclair, D. A. and L. Guarente (1997). "Extrachromosomal rdna circles--a cause of aging in yeast." Cell 91(7): 1033-1042.

Smith, A. G. and G. E. Muscat (2005). "Skeletal muscle and nuclear hormone receptors: Implications for cardiovascular and metabolic disease." Int. J. Biochem. Cell Biol. 37(10): 2047-2063.

Sol, E. M., S. A. Wagner, B. T. Weinert, A. Kumar, H.-S. Kim, C.-X. Deng and C. Choudhary (2012). "Proteomic investigations of lysine acetylation identify diverse substrates of mitochondrial deacetylase sirt3." PLoS One 7(12): e50545.

Someya, S., W. Yu, W. C. Hallows, J. Xu, J. M. Vann, C. Leeuwenburgh, M. Tanokura, J. M. Denu and T. A. Prolla (2010). "Sirt3 mediates reduction of oxidative damage and prevention of age-related hearing loss under caloric restriction." Cell 143(5): 802-812.

Sookoian, S., M. S. Rosselli, C. Gemma, A. L. Burgueno, T. Fernandez Gianotti, G. O. Castano and C. J. Pirola (2010). "Epigenetic regulation of insulin resistance in nonalcoholic fatty liver disease: Impact of liver methylation of the peroxisome proliferator-activated receptor gamma coactivator 1alpha promoter." Hepatology 52(6): 1992-2000.

Still, A. J., B. J. Floyd, A. S. Hebert, C. A. Bingman, J. J. Carson, D. R. Gunderson, B. K. Dolan, P. A. Grimsrud, K. E. Dittenhafer-Reed, D. S. Stapleton, M. P. Keller, M. S. Westphall, J. M. Denu, A. D. Attie, J. J. Coon and D. J. Pagliarini (2013). "Quantification of mitochondrial acetylation dynamics highlights prominent sites of metabolic regulation." J. Biol. Chem. 288(36): 26209-26219.

189 REFERENCES

Storgaard, H., C. B. Jensen, M. Bjornholm, X. M. Song, S. Madsbad, J. R. Zierath and A. A. Vaag (2004). "Dissociation between fat-induced in vivo insulin resistance and proximal insulin signaling in skeletal muscle in men at risk for type 2 diabetes." J. Clin. Endocrinol. Metab. 89(3): 1301-1311.

Su, B., X. Wang, L. Zheng, G. Perry, M. A. Smith and X. Zhu (2010). "Abnormal mitochondrial dynamics and neurodegenerative diseases." Biochim. Biophys. Acta 1802(1): 135-142.

Sun, C., F. Zhang, X. Ge, T. Yan, X. Chen, X. Shi and Q. Zhai (2007). "Sirt1 improves insulin sensitivity under insulin-resistant conditions by repressing ptp1b." Cell Metab 6(4): 307-319.

Sundaresan, N. R., M. Gupta, G. Kim, S. B. Rajamohan, A. Isbatan and M. P. Gupta (2009). "Sirt3 blocks the cardiac hypertrophic response by augmenting foxo3a- dependent antioxidant defense mechanisms in mice." J. Clin. Invest. 119(9): 2758-2771.

Sundaresan, N. R., S. A. Samant, V. B. Pillai, S. B. Rajamohan and M. P. Gupta (2008). "Sirt3 is a stress-responsive deacetylase in cardiomyocytes that protects cells from stress-mediated cell death by deacetylation of ku70." Mol. Cell. Biol. 28(20): 6384-6401.

Szendroedi, J., M. Chmelik, A. I. Schmid, P. Nowotny, A. Brehm, M. Krssak, E. Moser and M. Roden (2009). "Abnormal hepatic energy homeostasis in type 2 diabetes." Hepatology 50(4): 1079-1086.

Tao, R., M. C. Coleman, J. D. Pennington, O. Ozden, S. H. Park, H. Jiang, H. S. Kim, C. R. Flynn, S. Hill, W. Hayes McDonald, A. K. Olivier, D. R. Spitz and D. Gius (2010). "Sirt3-mediated deacetylation of evolutionarily conserved lysine 122 regulates mnsod activity in response to stress." Mol. Cell 40(6): 893-904.

Thompson, A. L., M. Y. Lim-Fraser, E. W. Kraegen and G. J. Cooney (2000). "Effects of individual fatty acids on glucose uptake and glycogen synthesis in soleus muscle in vitro." Am. J. Physiol. Endocrinol. Metab. 279(3): E577-584.

Tissenbaum, H. A. and L. Guarente (2001). "Increased dosage of a sir-2 gene extends lifespan in caenorhabditis elegans." Nature 410(6825): 227-230.

Turner, N., C. R. Bruce, S. M. Beale, K. L. Hoehn, T. So, M. S. Rolph and G. J. Cooney (2007). "Excess lipid availability increases mitochondrial fatty acid oxidative capacity in muscle: Evidence against a role for reduced fatty acid oxidation in lipid-induced insulin resistance in rodents." Diabetes 56(8): 2085-2092.

190 REFERENCES

Turner, N., G. J. Cooney, E. W. Kraegen and C. R. Bruce (2014). "Fatty acid metabolism, energy expenditure and insulin resistance in muscle." J. Endocrinol. 220(2): T61-T79.

Turner, N. and L. K. Heilbronn (2008). "Is mitochondrial dysfunction a cause of insulin resistance?" Trends Endocrinol. Metab. 19(9): 324-330.

Turner, N., G. M. Kowalski, S. J. Leslie, S. Risis, C. Yang, R. S. Lee-Young, J. R. Babb, P. J. Meikle, G. I. Lancaster, D. C. Henstridge, P. J. White, E. W. Kraegen, A. Marette, G. J. Cooney, M. A. Febbraio and C. R. Bruce (2013). "Distinct patterns of tissue- specific lipid accumulation during the induction of insulin resistance in mice by high-fat feeding." Diabetologia 56(7): 1638-1648.

Turner, N., J. Y. Li, A. Gosby, S. W. To, Z. Cheng, H. Miyoshi, M. M. Taketo, G. J. Cooney, E. W. Kraegen, D. E. James, L. H. Hu, J. Li and J. M. Ye (2008). "Berberine and its more biologically available derivative, dihydroberberine, inhibit mitochondrial respiratory complex i: A mechanism for the action of berberine to activate amp-activated protein kinase and improve insulin action." Diabetes 57(5): 1414-1418.

Uysal, K. T., S. M. Wiesbrock, M. W. Marino and G. S. Hotamisligil (1997). "Protection from obesity-induced insulin resistance in mice lacking tnf-[alpha] function." Nature 389(6651): 610-614.

Vaag, A., J. E. Henriksen, S. Madsbad, N. Holm and H. Beck-Nielsen (1995). "Insulin secretion, insulin action, and hepatic glucose production in identical twins discordant for non-insulin-dependent diabetes mellitus." J. Clin. Invest. 95(2): 690-698.

Vempati, R. K., R. S. Jayani, D. Notani, A. Sengupta, S. Galande and D. Haldar (2010). "P300-mediated acetylation of histone h3 lysine 56 functions in DNA damage response in mammals." J. Biol. Chem. 285(37): 28553-28564.

Verdin, E., M. D. Hirschey, L. W. Finley and M. C. Haigis (2010). "Sirtuin regulation of mitochondria: Energy production, apoptosis, and signaling." Trends Biochem. Sci. 35(12): 669-675.

Verma, M., N. Shulga and J. G. Pastorino (2013). "Sirtuin-3 modulates bak- and bax- dependent apoptosis." J. Cell Sci. 126(Pt 1): 274-288.

Verma, M., N. Shulga and J. G. Pastorino (2013). "Sirtuin-4 modulates sensitivity to induction of the mitochondrial permeability transition pore." Biochim. Biophys. Acta 1827(1): 38-49.

191 REFERENCES

Wagner, G. R. and M. D. Hirschey (2014). "Nonenzymatic protein acylation as a carbon stress regulated by sirtuin deacylases." Mol. Cell 54(1): 5-16.

Wagner, G. R. and R. M. Payne (2013). "Widespread and enzyme-independent nepsilon-acetylation and nepsilon-succinylation of proteins in the chemical conditions of the mitochondrial matrix." J. Biol. Chem. 288(40): 29036-29045.

Wang, F., M. Nguyen, F. X. Qin and Q. Tong (2007). "Sirt2 deacetylates foxo3a in response to oxidative stress and caloric restriction." Aging cell 6(4): 505-514.

Wang, F. and Q. Tong (2009). "Sirt2 suppresses adipocyte differentiation by deacetylating foxo1 and enhancing foxo1's repressive interaction with ppargamma." Mol. Biol. Cell 20(3): 801-808.

Wang, L., J. P. Louboutin, P. Bell, J. A. Greig, Y. Li, D. Wu and J. M. Wilson (2011). "Muscle-directed gene therapy for hemophilia b with more efficient and less immunogenic aav vectors." J. Thromb. Haemost. 9(10): 2009-2019.

Wang, Q., Y. Zhang, C. Yang, H. Xiong, Y. Lin, J. Yao, H. Li, L. Xie, W. Zhao, Y. Yao, Z. B. Ning, R. Zeng, Y. Xiong, K. L. Guan, S. Zhao and G. P. Zhao (2010). "Acetylation of metabolic enzymes coordinates carbon source utilization and metabolic flux." Science 327(5968): 1004-1007.

Warren, B. E., P. H. Lou, E. Lucchinetti, L. Zhang, A. S. Clanachan, A. Affolter, M. Hersberger, M. Zaugg and H. Lemieux (2014). "Early mitochondrial dysfunction in glycolytic muscle, but not oxidative muscle, of the fructose-fed insulin- resistant rat." Am. J. Physiol. Endocrinol. Metab. 306(6): E658-667.

Wasserman, D. H. and J. E. Ayala (2005). "Interaction of physiological mechanisms in control of muscle glucose uptake." Clin. Exp. Pharmacol. Physiol. 32(4): 319- 323.

Watt, M. J., A. J. Hoy, D. M. Muoio and R. A. Coleman (2012). "Distinct roles of specific fatty acids in cellular processes: Implications for interpreting and reporting experiments." Am. J. Physiol. Endocrinol. Metab. 302(1): E1-3.

Weinert, B. T., V. Lesmantavicius, T. Moustafa, C. Scholz, S. A. Wagner, C. Magnes, R. Zechner and C. Choudhary (2014). "Acetylation dynamics and stoichiometry in saccharomyces cerevisiae." Mol. Syst. Biol. 10: 716.

Weinert, Brian T., C. Schölz, Sebastian A. Wagner, V. Iesmantavicius, D. Su, Jeremy A. Daniel and C. Choudhary (2013). "Lysine succinylation is a frequently occurring modification in prokaryotes and eukaryotes and extensively overlaps with acetylation." Cell reports 4(4): 842-851.

192 REFERENCES

White, A. T., C. E. McCurdy, A. Philp, D. L. Hamilton, C. D. Johnson and S. Schenk (2013). "Skeletal muscle-specific overexpression of sirt1 does not enhance whole-body energy expenditure or insulin sensitivity in young mice." Diabetologia 56(7): 1629-1637.

White, A. T. and S. Schenk (2012). "Nad(+)/nadh and skeletal muscle mitochondrial adaptations to exercise." Am. J. Physiol. Endocrinol. Metab. 303(3): E308-321.

WHO. (2011). "Mean body mass index- global health advisory." Global Health Advisory, 2014, from http://www.who.int/gho/ncd/risk_factors/bmi_text/en/.

Winnik, S., D. S. Gaul, F. Preitner, C. Lohmann, J. Weber, M. X. Miranda, Y. Liu, L. J. van Tits, J. M. Mateos, C. E. Brokopp, J. Auwerx, B. Thorens, T. F. Luscher and C. M. Matter (2014). "Deletion of sirt3 does not affect atherosclerosis but accelerates weight gain and impairs rapid metabolic adaptation in ldl receptor knockout mice: Implications for cardiovascular risk factor development." Basic Res. Cardiol. 109(1): 399.

Withrow, D. and D. A. Alter (2011). "The economic burden of obesity worldwide: A systematic review of the direct costs of obesity." Obes. Rev. 12(2): 131-141.

Wolff, J. A. and V. Budker (2005). "The mechanism of naked DNA uptake and expression." Adv. Genet. 54: 3-20.

Wright, L. E., A. E. Brandon, A. J. Hoy, G. B. Forsberg, C. J. Lelliott, J. Reznick, L. Lofgren, J. Oscarsson, M. Stromstedt, G. J. Cooney and N. Turner (2011). "Amelioration of lipid-induced insulin resistance in rat skeletal muscle by overexpression of pgc-1beta involves reductions in long-chain acyl-coa levels and oxidative stress." Diabetologia 54(6): 1417-1426.

Xiong, Y. and K. L. Guan (2012). "Mechanistic insights into the regulation of metabolic enzymes by acetylation." J. Cell Biol. 198(2): 155-164.

Xue, L., F. Xu, L. Meng, S. Wei, J. Wang, P. Hao, Y. Bian, Y. Zhang and Y. Chen (2012). "Acetylation-dependent regulation of mitochondrial aldh2 activation by sirt3 mediates acute ethanol-induced enos activation." FEBS Lett. 586(2): 137-142.

Yang, H., T. Yang, J. A. Baur, E. Perez, T. Matsui, J. J. Carmona, D. W. Lamming, N. C. Souza-Pinto, V. A. Bohr, A. Rosenzweig, R. de Cabo, A. A. Sauve and D. A. Sinclair (2007). "Nutrient-sensitive mitochondrial nad+ levels dictate cell survival." Cell 130(6): 1095-1107.

Yang, L., B. Vaitheesvaran, K. Hartil, A. J. Robinson, M. R. Hoopmann, J. K. Eng, I. J. Kurland and J. E. Bruce (2011). "The fasted/fed mouse metabolic acetylome:

193 REFERENCES

N6-acetylation differences suggest acetylation coordinates organ-specific fuel switching." J. Proteome Res. 10(9): 4134-4149.

Yang, Y., H. Cimen, M. J. Han, T. Shi, J. H. Deng, H. Koc, O. M. Palacios, L. Montier, Y. Bai, Q. Tong and E. C. Koc (2010). "Nad+-dependent deacetylase sirt3 regulates mitochondrial protein synthesis by deacetylation of the ribosomal protein mrpl10." J. Biol. Chem. 285(10): 7417-7429.

Yang, Y., W. Fu, J. Chen, N. Olashaw, X. Zhang, S. V. Nicosia, K. Bhalla and W. Bai (2007). "Sirt1 sumoylation regulates its deacetylase activity and cellular response to genotoxic stress." Nat. Cell Biol. 9(11): 1253-1262.

Ye, J. M., J. Tid-Ang, N. Turner, X. Y. Zeng, H. Y. Li, G. J. Cooney, E. M. Wulff, P. Sauerberg and E. W. Kraegen (2011). "Ppardelta agonists have opposing effects on insulin resistance in high fat-fed rats and mice due to different metabolic responses in muscle." Br. J. Pharmacol. 163(3): 556-566.

Yoshino, J., K. F. Mills, M. J. Yoon and S. Imai (2011). "Nicotinamide mononucleotide, a key nad(+) intermediate, treats the pathophysiology of diet- and age-induced diabetes in mice." Cell Metab 14(4): 528-536.

Yoshizawa, T., M. F. Karim, Y. Sato, T. Senokuchi, K. Miyata, T. Fukuda, C. Go, M. Tasaki, K. Uchimura, T. Kadomatsu, Z. Tian, C. Smolka, T. Sawa, M. Takeya, K. Tomizawa, Y. Ando, E. Araki, T. Akaike, T. Braun, Y. Oike, E. Bober and K. Yamagata (2014). "Sirt7 controls hepatic lipid metabolism by regulating the ubiquitin-proteasome pathway." Cell Metabolism 19(4): 712-721.

Yu, J. and J. Auwerx (2009). "The role of sirtuins in the control of metabolic homeostasis." Ann. N. Y. Acad. Sci. 1173 Suppl 1: E10-19.

Yu, W., Y. Lin, J. Yao, W. Huang, Q. Lei, Y. Xiong, S. Zhao and K. L. Guan (2009). "Lysine 88 acetylation negatively regulates ornithine carbamoyltransferase activity in response to nutrient signals." J. Biol. Chem. 284(20): 13669-13675.

Zhang, G., X. Gao, Y. K. Song, R. Vollmer, D. B. Stolz, J. Z. Gasiorowski, D. A. Dean and D. Liu (2004). "Hydroporation as the mechanism of hydrodynamic delivery." Gene Ther. 11(8): 675-682.

Zhang, Z., M. Tan, Z. Xie, L. Dai, Y. Chen and Y. Zhao (2011). "Identification of lysine succinylation as a new post-translational modification." Nat. Chem. Biol. 7(1): 58-63.

Zhao, S., W. Xu, W. Jiang, W. Yu, Y. Lin, T. Zhang, J. Yao, L. Zhou, Y. Zeng, H. Li, Y. Li, J. Shi, W. An, S. M. Hancock, F. He, L. Qin, J. Chin, P. Yang, X. Chen, Q. Lei, Y. Xiong

194 REFERENCES

and K. L. Guan (2010). "Regulation of cellular metabolism by protein lysine acetylation." Science 327(5968): 1000-1004.

Zimmet, P. Z., D. J. Magliano, W. H. Herman and J. E. Shaw (2014). "Diabetes: A 21st century challenge." Lancet Diabetes Endocrinol 2(1): 56-64.

195