Cholesterol-Dependent Degradation and Unsaturated Fatty Acid-Dependent Stabilisation of Squalene Monooxygenase in the Control of Cholesterol Synthesis

Cholesterol-dependent degradation and unsaturated fatty acid-dependent stabilisation of squalene monooxygenase in the control of cholesterol synthesis

Julian Stevenson

A thesis in fulfilment of the requirements for the degree of

Doctor of Philosophy

School of Biotechnology and Biomolecular Sciences Faculty of Science

Supervisor: Professor Andrew Brown

March 2014

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

Surname or Family name: Stevenson

First name: Julian Other name/s:

Abbreviation for degree as given in the University calendar: PhD

School: Biotechnology and Biomolecular Sciences Faculty: Science

Title: Cholesterol-dependent degradation and unsaturated fatty acid-dependent stabilisation of squalene monooxygenase in the control of cholesterol synthesis

Abstract 350 words maximum: (PLEASE TYPE) Exquisite control of cholesterol synthesis is crucial for maintaining homeostasis of this vital lipid. Squalene monooxygenase (SM) catalyses the first oxygenation step in cholesterol synthesis, acting on squalene before cyclisation into the basic steroid structure. Using model cell systems, we found that cholesterol caused the accumulation of the substrate squalene, suggesting that SM may serve as a flux-controlling enzyme beyond 3- hydroxy-3-methylglutaryl-coenzyme A reductase (HMGCR, considered as rate-limiting). Cholesterol accelerated the proteasomal degradation of SM, which required the N-terminal domain (N100), partially conserved in vertebrates, but not lower organisms. Unlike HMGR, SM degradation is not mediated by Insig, 24,25-dihydrolanosterol or side-chain oxysterols, but rather by cholesterol itself. Importantly, SM’s N-terminal domain conferred cholesterol-regulated turnover on heterologous fusion proteins. Furthermore, proteasomal inhibition almost totally eliminated squalene accumulation, highlighting the importance of this degradation mechanism for the control of SM and cholesterol synthesis after mevalonate production. On the contrary, treatment with unsaturated fatty acids such as oleate, but not saturated fatty acids, increased protein levels of SM or N100, as well as reversing cholesterol-dependent squalene accumulation. Notably, the stabilisation occurred through reduced ubiquitination by the E3 ubiquitin ligase, MARCH6. Maximum stabilisation required activation of fatty acids, but not triglyceride or phosphatidylcholine synthesis. Stabilisation of a cholesterol biosynthetic enzyme by unsaturated fatty acids may help maintain a constant cholesterol/phospholipid ratio. In addition, we optimised and compared highly efficient ligation-independent cloning techniques, identifying important parametres for project design using either polymerase incomplete primer extension (PIPE) cloning, sequence and ligation-independent cloning (SLIC), or overlap extension cloning (OEC), including the need to avoid PCR artefacts such as primer-dimers and vector plasmid background. Experiments made use of a common reporter vector and a set of modular primers to clone DNA fragments of increasing size. Overall, PIPE achieved cloning efficiencies of ~95% with few manipulations, whereas SLIC provided a much higher number of transformants, but required additional steps. Our data suggest that for small inserts (<1.5 kb), OEC is a good option, requiring only two new primers, but performs poorly for larger inserts. We believe that these ligation-independent cloning approaches constitute an essential part of the researcher's molecular-tool kit

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 OR BE THE FIRST PAGE INSERTED III

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 ......

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 ……………………………………………......

VII

VIII

Abstract

Exquisite control of cholesterol synthesis is crucial for maintaining homeostasis of this vital lipid. Squalene monooxygenase (SM) catalyses the first oxygenation step in cholesterol synthesis, acting on squalene before cyclisation into the basic steroid structure. Using model cell systems, we found that cholesterol caused the accumulation of the substrate squalene, suggesting that SM may serve as a flux-controlling enzyme beyond 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGCR, considered as rate-limiting). Cholesterol accelerated the proteasomal degradation of SM, which required the N-terminal domain (N100), partially conserved in vertebrates, but not lower organisms. Unlike HMGCR, SM degradation is not mediated by Insig, 24,25-dihydrolanosterol or side-chain oxysterols, but rather by cholesterol itself. Importantly, SM’s N-terminal domain conferred cholesterol-regulated turnover on heterologous fusion proteins. Furthermore, proteasomal inhibition almost totally eliminated squalene accumulation, highlighting the importance of this degradation mechanism for the control of SM and cholesterol synthesis after mevalonate production. On the contrary, treatment with unsaturated fatty acids such as oleate, but not saturated fatty acids, increased protein levels of SM or N100- GFP, as well as reversing cholesterol-dependent squalene accumulation. Notably, the stabilisation occurred through reduced ubiquitination by the E3 ubiquitin ligase, MARCH6. Maximum stabilisation required activation of fatty acids, but not triglyceride or phosphatidylcholine synthesis. Stabilisation of a cholesterol biosynthetic enzyme by unsaturated fatty acids may help maintain a constant cholesterol/phospholipid ratio. In addition, we optimised and compared highly efficient ligation-independent cloning techniques, identifying important parametres for project design using either polymerase incomplete primer extension (PIPE) cloning, sequence and ligation-independent cloning (SLIC), or overlap extension cloning (OEC), including the need to avoid PCR artefacts such as primer-dimers and vector plasmid background. Experiments made use of a common reporter vector and a set of modular primers to clone DNA fragments of increasing size. Overall, PIPE achieved cloning efficiencies of ~95% with few manipulations, whereas SLIC provided a much higher number of transformants, but required additional steps. Our data suggest that for small inserts (<1.5 kb), OEC is a good option, requiring only two new primers, but performs poorly for larger inserts. We believe that these ligation- independent cloning approaches constitute an essential part of the researcher's molecular-tool kit.

List of publications

Stevenson, J., Kristiana, I., Luu, W., and Brown, A. J. (2014) Squalene monooxygenase, a key enzyme in cholesterol synthesis, is stabilised by unsaturated fatty acids. The Biochemical Journal 461, 435-442

Zelcer, N., Sharpe, L. J., Loregger, A., Kristiana, I., Cook, E. C., Phan, L., Stevenson, J., and Brown, A. J. (2014) The E3 ubiquitin ligase MARCH6 degrades squalene monooxygenase and affects 3-hydroxy-3-methyl-glutaryl coenzyme A reductase and the cholesterol synthesis pathway. Molecular and Cellular Biology 34, 1262-1270

Stevenson, J.*, Krycer*, J. R., Phan, L., and Brown, A. J. (2013) A practical comparison of ligation-independent cloning techniques. PLOS ONE 8, e8388 *Equal first author

Kristiana, I., Luu, W., Stevenson, J., Cartland, S., Jessup, W., Belani, J. D., Rychnovsky, S. D., and Brown, A. J. (2012) Cholesterol through the looking glass: ability of its enantiomer also to elicit homeostatic responses. The Journal of Biological Chemistry 287, 33897-33904

Luu, W., Sharpe, L. J., Stevenson, J., and Brown, A. J. (2012) Akt acutely activates the cholesterogenic transcription factor SREBP-2. Biochimica et Biophysica Acta 1823, 458-464

Gill, S.*, Stevenson, J.*, Kristiana, I., and Brown, A. J. (2011) Cholesterol- dependent degradation of squalene monooxygenase, a control point in cholesterol synthesis beyond HMG-CoA reductase. Cell Metabolism 13, 260-273 *Equal first author

Stevenson, J., and Brown, A. J. (2009) How essential is cholesterol? The Biochemical Journal 420, e1-4

XII

List of conference poster presentations

Stevenson, J., Krycer, J. R., Phan, L., and Brown, A. J. A comparison of ligation- independent cloning techniques ASMR NSW Scientific Meeting 2013 (Australian Society for Medical Research), Sydney, Australia, June 2013

Stevenson, J., Kristiana, I., and Brown, A. J. Squalene monooxygenase, a key enzyme in cholesterol synthesis, is stabilised by unsaturated fatty acids Membrane Transporters and Their Role in Human Disease (Australian Society for Biochemistry and Molecular Biology), Sydney, Australia, December 2012

Stevenson, J., Kristiana, I., and Brown, A. J. Squalene monooxygenase, a key enzyme in cholesterol synthesis, is stabilized by unsaturated fatty acids Frontiers in Lipid Biology (American Society for Biochemistry and Molecular Biology, International Conference on the Bioscience of Lipids, Canadian Lipoprotein Conference), Banff, Canada, September 2012

Stevenson, J., Gill, S., Kristiana, I., and Brown, A. J. Cholesterol-dependent proteasomal degradation of squalene monooxygenase, a second control point in cholesterol synthesis Experimental Biology 2011 (American Society for Biochemistry and Molecular Biology), Washington D. C., U. S. A., April 2011

Stevenson, J., Gill, S., Kristiana, I., and Brown, A. J. A novel control point in cholesterol synthesis OzBio2010 (Australian Society for Biochemistry and Molecular Biology), Melbourne, Australia, September 2010

XIII

XIV

Acknowledgements

Firstly, I’d like to thank my supervisor, Professor Andrew Brown. You are an outstanding mentor and extremely kind man. I have learnt so much about the craft of academic research and benefitted greatly from your advice and encouragement.

I thank Saloni Gill for collaborating with me on the SM project, especially for processing my radioactive samples. Together with Ika, we made a formidable lab team.

I was extremely fortunate to have access to the excellent technical assistance of the lab’s RA, Ika Kristiana. I hope you can fully engage your passion for the arts and hands-on craftsmanship beyond science, maybe with some cholesterol- themed artwork?

I thank Dr Jenny Wong for doing the groundwork that sowed the seed for my project.

I’d like to thank James Krycer for being an inspiring colleague and supportive friend. Discussing science at a very high level and sharing my ideas with you is one of my favourite things to do, so I hope we can continue this for many years to come.

I thank the other existing and previous members of the Brown lab for all their help and friendship, including Lisa Phan, Laura Sharpe, Winnie Luu, Cyril Burr, Eser Zerenturk, Vicky Burns, and Anika Prabhu. It was great to be in a lab where I got on very well with everyone. You were always willing to assist or listen to my advice and ideas, even if we disagreed.

I would also like to thank Dr Maaike Kockx for her assistance with the 35S pulse-chase experiments.

XV I thank Dr Gabriel Perrone, Professor Paul Curmi, Dr Louise Lutze-Mann and Associate Professor Noel Whitaker for taking an interest in my research, and for their helpful advice.

Thanks also to the department administration staff for being so helpful and friendly.

I would like to thank you, the reader, for taking an interest in this work. This thesis is largely adapted from the manuscripts I drafted, where I made a central intellectual contribution. I had hoped to submit my thesis as a series of publications, but the faculty now requires four accepted first author publications to allow this. For the rare new student reading this thesis, I plan to publish all of the raw data and abridged commentary in academic journals, but I believe you will find my additional analysis here useful.

I thank Endemol Southern Star Pty Limited and the Seven Network for providing a great experience and broadcasting the cover for my Cell Metabolism paper (beautifully illustrated by Dr Anne Galea) to 1.5 million viewers.

I’d like to thank Tracey and all the other community carers and respite nurses for their invaluable assistance.

I thank my family for their unfailing support and encouragement, providing truly life-saving help.

I dedicate this thesis to the memory of my mother, Anna Stevenson (1951-2012), a saintly parent and brilliant clinician, educator and entertainer.

XVI

Table of contents

Thesis/Dissertation Sheet ...... III Originality statement ...... V Copyright statement ...... VII Authenticity statement ...... VII Abstract ...... IX List of publications ...... XI List of conference poster presentations ...... XIII Acknowledgements ...... XV List of figures ...... XXI List of tables ...... XXIII Abbreviations ...... XXV

1 General introduction ...... 3 1.1 Cholesterol ...... 3 1.1.1 Background ...... 3 1.1.2 Functions of cholesterol ...... 3 1.1.3 Cholesterol excess is toxic ...... 4 1.2 Cellular cholesterol levels are controlled through regulation of cholesterol uptake, synthesis and efflux ...... 5 1.2.1 The Scap/SREBP-2 pathway ...... 6 1.2.2 SREBP-2 processing may be increased by the important signalling kinase, Akt…… ...... 8 1.2.3 The liver X receptor ...... 8 1.2.4 Cholesterol is synthesised by the mevalonate pathway ...... 9 1.2.5 Regulation of HMGCR ...... 10 1.3 Endoplasmic reticulum-associated degradation (ERAD) ...... 11 1.4 Squalene monooxygenase ...... 12 1.4.1 Background ...... 12 1.4.2 Structure of SM ...... 13 1.4.3 SM as a drug target ...... 14 1.4.4 SM may catalyse a rate-limiting step in cholesterol synthesis ...... 14 1.4.5 Regulation of SM ...... 15 1.5 Ligation-independent cloning ...... 15 1.6 Aims and hypotheses ...... 16

2 General materials and methods ...... 19 2.1 General materials ...... 19 XVII 2.1.1 Media recipes ...... 21 2.2 Cell culture ...... 21 2.3 Metabolic labelling of squalene and cholesterol ...... 22 2.4 Quantitative real-time PCR ...... 22 2.5 Immunoblot analysis ...... 23 2.6 Transfection ...... 24 2.7 Cholesterol mass determination ...... 24 2.8 Data presentation ...... 24

3 A practical comparison of ligation-independent cloning techniques ...... 27 3.1 Introduction ...... 27 3.2 Materials and methods ...... 30 3.2.1 Primer combinations ...... 30 3.2.2 PCR ...... 32 3.2.3 PIPE cloning ...... 32 3.2.4 SLIC ...... 32 3.2.5 OEC ...... 33 3.2.6 Colony counting and screening ...... 33 3.2.7 Assembly of the pBI-CMV-FRB-Akt-Myc-2xFKBP-HA-FRT inducible Akt construct and creation of a stable Flp-In cell-line ...... 34 3.3 Results ...... 36 3.3.1 Design of the reporter system ...... 36 3.3.2 Generation of nicked vector plasmid reduces PIPE cloning efficiency...... 37 3.3.3 Effect of cycle number and insert:vector ratio on PIPE cloning efficiency .. 39 3.3.4 Optimisation of SLIC ...... 43 3.3.5 Effect of megaprimer concentration, PCR template and cycle number on OEC……...... 44 3.3.6 OEC does not tolerate the presence of primer-dimers ...... 47 3.3.7 Direct comparison ...... 49 3.3.8 LIC can be performed without product quantification or purification ...... 52 3.3.9 An example of using LIC to solve a biological problem: Akt/Protein Kinase B acutely activates SREBP-2 ...... 53 3.4 Discussion ...... 56

4 Cholesterol-dependent degradation of squalene monooxygenase, a novel control point in cholesterol synthesis beyond HMG-CoA reductase ...... 63 4.1 Introduction ...... 63 4.2 Materials and methods ...... 64 4.2.1 Construction of expression plasmids ...... 64 4.2.2 Preparation of lysine mutants by multiple SDM ...... 67

XVIII

4.2.3 Ubiquitination of human SM ...... 68 4.2.4 Metabolic labelling of N100-GST with [35S]-methionine/cysteine ...... 68 4.2.5 Cell fractionation ...... 69 4.3 Results ...... 70 4.3.1 Cholesterol treatment causes squalene to accumulate, suggesting rate- limiting activity of SM ...... 70 4.3.2 Cholesterol-dependent squalene accumulation occurs in a variety of cell types…...... 72 4.3.3 Analysis of squalene accumulation and the cholesterol-dependent reduction in message levels suggests post-transcriptional regulation of SM...... 72 4.3.4 Squalene accumulates in SRD-1 cells upon cholesterol treatment, consistent with post-transcriptional regulation of SM ...... 74 4.3.5 Cholesterol-dependent degradation of SM ...... 76 4.3.6 Accelerated degradation occurs when cholesterol exceeds basal levels ..... 78 4.3.7 Cholesterol-dependent degradation of SM is mediated by the ubiquitin- proteasome system ...... 80 4.3.8 Proteasomal degradation affects flux through the mevalonate pathway at SM……...... 84 4.3.9 Insig and Scap are not required for the cholesterol-dependent degradation of SM...... 86 4.3.10 Cholesterol-dependent proteasomal degradation of SM requires its N-terminal domain ...... 88 4.3.11 N100 is membrane associated ...... 97 4.3.12 Regulated degradation can be mediated by multiple ubiquitination sites 98 4.3.13 A putative CRAC motif in SM and other motifs similar to those found in Scap and HMGCR are not functionally significant ...... 99 4.4 Discussion ...... 101

5 SM is degraded in response to cholesterol itself, and stabilised by unsaturated fatty acids ...... 105 5.1 Introduction ...... 105 5.2 Materials and methods ...... 106 5.2.1 Materials ...... 106 5.2.2 Preparation of the N100-GFP stable cell-line ...... 106 5.2.3 siRNA knockdown ...... 106 5.2.4 Ubiquitination of N100-GFP ...... 107 5.2.5 Statistical analysis ...... 107 5.3 Results ...... 108 5.3.1 Endogenous sterols can cause degradation of SM ...... 108 XIX 5.3.2 Cholesterol is the principal degradative signal ...... 110 5.3.3 SM is not degraded in response to side-chain oxysterols ...... 112 5.3.4 SM turnover is increased by the enantiomer of cholesterol ...... 114 5.3.5 Oleate treatment stabilises SM, which may increase flux through the cholesterol synthesis pathway ...... 115 5.3.6 Fatty acid specificity - unsaturated fatty acids stabilise N100-GFP ...... 116 5.3.7 Stabilisation is not mediated through reduction of free-cholesterol ...... 118 5.3.8 Maximum protection requires activation of fatty acids, but not triglyceride synthesis ...... 120 5.3.9 Stabilisation does not require protein kinase C activity ...... 123 5.3.10 Oleate stabilisation of SM can occur independently of UBXD8 ...... 124 5.3.11 Oleate blunts polyubiquitination by MARCH6 ...... 126 5.4 Discussion ...... 128

6 General discussion ...... 133 6.1 Summary of findings ...... 133 6.2 Ligation-independent cloning and beyond ...... 134 6.3 Regulated degradation of SM is distinct from that of HMGCR and does not appear to require lipid droplets ...... 136 6.4 N100 is sufficient for degradation ...... 138 6.5 SM tertiary structure ...... 140 6.6 SM is ubiquitinated by the E3 ubiquitin ligase, MARCH6 ...... 140 6.7 SM dislocation ...... 142 6.8 Advantages of additional control points in cholesterol synthesis ...... 143 6.9 Conclusion ...... 146

7 References ...... 147

List of Figures

Figure 1.1. Structure of cholesterol...... 3 Figure 1.2. Overview of transcriptional control in cellular cholesterol homeostasis. .. 6 Figure 1.3. The Scap/SREBP pathway...... 7 Figure 1.4. Schematic of cholesterol synthesis by the mevalonate pathway...... 10 Figure 1.5. Schematic of the reaction catalysed by SM...... 13 Figure 2.1. Media formulations...... 21 Figure 3.1. Principles of polymerase incomplete primer extension (PIPE) cloning, sequence and ligation-independent cloning (SLIC) and overlap extension cloning (OEC)...... 28 Figure 3.2. pUC18/Kan reporter plasmid...... 36 Figure 3.3. Generation of nicked vector plasmid can reduce PIPE cloning efficiency...... 38 Figure 3.4. OEC does not tolerate primer-dimers...... 48 Figure 3.5. Primer-dimer can be removed with T4 DNA polymerase exonuclease treatment...... 48 Figure 3.6. Rapalog-activated Akt increases SREBP-2 activation...... 55 Figure 3.7. Technique selection flowchart for a new cloning project...... 57 Figure 4.1. Cholesterol treatment causes squalene to accumulate...... 71 Figure 4.2. Squalene accumulates in a variety of human cell types...... 72 Figure 4.3. Squalene accumulates whilst message levels of SM and HMGCR remain high...... 73 Figure 4.4. Cholesterol treatment causes squalene to accumulate in SRD-1 cells, independent of transcriptional regulation...... 75 Figure 4.5. SM is regulated at the post-translational level by cholesterol-dependent degradation...... 77 Figure 4.6. Cholesterol-dependent degradation of ectopic SM is blunted with higher over-expression...... 78 Figure 4.7. Degradation is accelerated as cholesterol levels rise above a basal threshold...... 79 Figure 4.8. Cholesterol-Dependent degradation is reversed by proteasomal but not lysosomal inhibition...... 81 Figure 4.9. Addition of Proteasome-inhibitory GA-repeats blunt degradation of SM...... 82 Figure 4.10. Cholesterol treatment increases polyubiquitination of SM...... 83

XXI Figure 4.11. Proteasomal degradation affects flux through the mevalonate pathway at SM...... 85 Figure 4.12. Insig and Scap are not required for accelerated degradation of SM...... 87 Figure 4.13. Multiple sequence alignment of SM protein sequences reveals the presence of a conserved N-terminal region only found in vertebrates...... 89 Figure 4.14. The first 100 amino acids of SM are required to mediate cholesterol- dependent degradation...... 90 Figure 4.15. Overexpression of SM lacking the first hundred amino acids increases conversion of squalene to 2,3-monoxidosqualene...... 91 Figure 4.16. The N-terminal domain (N100) confers cholesterol-dependent degradation upon fusion proteins...... 93 Figure 4.17. Pulse-chase analysis reveals rapid degradation of N100-GST...... 94 Figure 4.18. N100-GST is targeted by the ubiquitin-proteasome system similarly to full-length SM and as early as 1 hr...... 97 Figure 4.19. N100 is sufficient, but not necessary, for SM membrane localisation. .... 97 Figure 4.20. Lysine substitution mutants retain cholesterol-dependent degradation...... 99 Figure 4.21. Mutation of motifs similar to Insig and Scap in N100 do not affect SM regulation...... 100 Figure 5.1. Degradation occurs in response to sterols, but not earlier intermediates...... 109 Figure 5.2. SM degradation specificity for sterol intermediates...... 111 Figure 5.3. SM degradation specificity for oxysterols...... 113 Figure 5.4. SM is degraded in response to ent-cholesterol...... 114 Figure 5.5. Oleate treatment reverses squalene accumulation and cholesterol- dependent degradation of SM...... 116 Figure 5.6. Fatty acid specificity of N100-GFP stabilisation...... 118 Figure 5.7. Oleate-dependent stabilisation is not mediated through reduction of free cholesterol levels...... 119 Figure 5.8. Stabilisation requires activation of fatty acids, but not triglyceride synthesis...... 122 Figure 5.9. Protein kinase C inhibition has no effect on oleate-dependent stabilisation of SM...... 123 Figure 5.10. Protection can occur independently of UBXD8...... 125 Figure 5.11.Oleate blunts polyubiquitination by MARCH6...... 127 Figure 6.1. Segmental control of the mevalonate pathway...... 144

XXII

List of tables

Table 2.1. Primer sequences for quantitative real-time PCR analysis ...... 23 Table 3.1. Primer sequences for LIC comparisons...... 31 Table 3.2. Plasmid names and gene accession numbers...... 32 Table 3.3. Primer sequences for assembly of a Flp-In inducible Akt construct...... 35 Table 3.4. Effect of template concentration on PIPE cloning efficiency...... 39 Table 3.5. Cycle number for insert product does not affect PIPE cloning efficiency. 40 Table 3.6. Cycle number for vector product does not affect PIPE cloning efficiency. 41 Table 3.7. Effect of insert:vector ratio on PIPE cloning efficiency...... 42 Table 3.8. Effect of T4 exonuclease treatment time on colony number...... 43 Table 3.9. Effect of megaprimer concentration on overlap extension cloning efficiency...... 44 Table 3.10. Increasing PCR template can generate insert template background in OEC for high megaprimer concentrations...... 46 Table 3.11. Colony number increases with additional cycles of overlap extension. .. 47 Table 3.12. Direct comparison of PIPE, SLIC and OEC for various insert sizes...... 50 Table 3.13. Complete screening for the comparison of PIPE, SLIC and OEC for increasing insert sizes...... 51 Table 3.14. Cloning of a 350 bp fragment without quantification or purification...... 52 Table 3.15. Summary of effectiveness and resource use...... 57 Table 4.1. Plasmids prepared in this study...... 64 Table 4.2. Primer sequences for cloning and SDM...... 65 Table 6.1. Sterol specificity for selected proteins in cholesterol homeostasis...... 137

XXIII

XXIV

Abbreviations

19HC 19-hydroxycholesterol 24,25DHL 24,25-dihydrolanosterol 24,25EC 24(S),25-epoxycholesterol 25HC 25-hydroxycholesterol 27HC 27-hydroxycholesterol 7AHC 7A-hydroxycholesterol 7BHC 7B-hydroxycholesterol 7DHC 7-dehydrocholesterol ABCA1 ATP-binding cassette transporter A1 ABCG1 ATP-binding cassette transporter G1 ATP adenosine triphosphate BSA bovine serum albumin CD methyl-β-cyclodextrin CHO Chinese Hamster Ovary Chol cholesterol CMV cytomegalovirus COPII coatomer protein complex-II CT CTP:phosphocholine cytidylyltransferase DGAT diglyceride acyltransferase DMEM/F12 Dulbecco’s Modified Eagle’s Medium/Ham’s Nutrient Mixture F-12 DOS 2,3(S):22(S),23-dioxidosqualene ER endoplasmic reticulum ERAD ER-associated protein degradation FAD flavin adenine dinucleotide FCS fetal calf serum GC-MS gas chromatography tandem mass spectrometry GFP green flourescent protein GST gluthione S transferase HEK Human Embryonic Kidney HMG-CoA 3-hydroxy-3-methylglutaryl coenzyme A HMGCR 3-hydroxy-3-methylglutaryl-coenzyme A reductase

XXV Idol inducible degrader of the LDL receptor IGF-1 insulin-like growth factor-1 Insig insulin-induced gene IPTG Isopropyl β-D-1-thiogalactopyranoside LDL low-density lipoprotein LXR liver X receptor MOS 2,3(S)-monooxidosqualene N100 first 100 amino acids of SM NADPH nicotinamide adenine dinucleotide phosphate NCS newborn calf serum OEC overlap extension cloning PBS phosphate buffered saline PDK phosphoinositide-dependent kinase PI3K phosphatidylinositol 3′-kinase PIP3 phosphatidylinositol (3,4,5)-trisphosphate PIPE polymerase primer incomplete primer extension S1P site 1 protease S2P site 2 protease SDM site-directed mutagenesis SDS sodium dodecyl sulphate SLIC sequence and ligation-independent cloning SM squalene monooxygenase SPF supernatant protein factor SRE sterol-response element SREBP sterol regulatory element-binding protein TK thymidine kinase

XXVI

Chapter 1

General introduction

1 General introduction

1.1 Cholesterol

1.1.1 Background Cholesterol was first isolated from bile and gallstones by Poulletier de la Salle in 1784, then later rediscovered by Chevreul in the early 19th Century, who named it ‘cholesterine’ (1). The complex structure of this neutral lipid - four fused rings, aliphatic side-chain and polar hydroxyl group (Figure 1.1) - was elucidated by Wieland and Windaus, for which they were awarded the Nobel Prize in 1927-8 (2). Cholesterol has proved so important and interesting that it has played a large role in the research of thirteen Nobel Prize winners, and was hence described as the ‘most decorated small molecule in biology’ by the 1985 Prize winners, Michael Brown and Joseph Goldstein (3). Cholesterol is also seen as a two-faced molecule: something thought of as ‘bad’ for health in the popular imagination, but which is nevertheless essential.

Figure 1.1. Structure of cholesterol.

1.1.2 Functions of cholesterol Cholesterol is an essential lipid constituent of higher eukaryotic cells (4), playing a vital barrier role by maintaining membrane fluidity and integrity (5) and assisting membrane curvature, fusion and budding (6). It lies in the membrane parallel to fatty acyl-chains, such that the polar head-group points

3 out, while the steroid nucleus interacts with the hydrocarbons of neighbouring lipids, with its side-chain extending toward the boundary of the two membrane leaflets. Cholesterol also stabilises the structure of lipid rafts, cholesterol and sphingolipid enriched microdomains that act as a platform for the association of proteins and lipid to regulate signalling and membrane trafficking (7-9). It also plays a part in signalling due to its covalent attachment to the protein hedgehog (10), required for embryonic development and involved in cellular differentiation (11,12). Cholesterol also serves as a precursor for incorporation into bile acids, vitamin D, oxysterols, and steroid hormones (13-15). Bile acids can work as amphipathic salts - a ‘natural detergent’ - to solubilise lipids and fat-soluble vitamins in the small intestine to facilitate their adsorption, as well as assist in excretion of excess free-cholesterol from the body. Vitamin D is primarily involved in adsorption of dietary calcium and phosphate. The cholecalciferol form is synthesised in the skin from 7-dehydrocholesterol in the presence of light, hence it is the ‘sunshine vitamin’. Oxysterols are oxygenated forms of cholesterol found at very low concentrations, in the order of a hundred to a million fold lower than cholesterol, with enhanced hydrophilicity, and in many cases, a significantly more potent regulatory effect compared to cholesterol, acting through specific protein binding and potentially by modulating membrane structure (15,16). Steroid hormones are another important class of signalling molecules derived from cholesterol, involved in processes such as inflammation and development of sexual characteristics (17,18).

1.1.3 Cholesterol excess is toxic Although cholesterol is required for numerous processes, excess levels are toxic, disrupting normal cell function, and can lead to disease, for example by promoting apoptosis (19) or preventing lateral movement of membrane proteins (20). Excess serum cholesterol levels - in the form of low-density lipoprotein (LDL) - contribute to atherosclerosis in humans (21), which underlies most cardiovascular disease and can lead to complications such as stroke and myocardial infarction (22-24). Cardiovascular disease is the primary cause of death in the Western world, so improvement in our understanding of the regulation of cholesterol levels is important to provide new therapeutic strategies to target this risk factor. Cardiovascular disease is a major economic

burden, with health care and pharmaceuticals costing the general public and governments billions of dollars each year. At the molecular level, synthesis of cholesterol is also very energetically expensive, with a single cholesterol molecule requiring the use of, or forgoing production of, more than one hundred ATP molecules that cannot be recouped (25).

1.2 Cellular cholesterol levels are controlled through regulation

of cholesterol uptake, synthesis and efflux

As cholesterol is indispensable, but toxic in excess, cholesterol levels are tightly controlled by regulation of uptake, efflux and synthesis, a state of homeostasis. Uptake is mediated by the LDL receptor, which enables acceptance of cholesterol located in lipoproteins circulating in the bloodstream (26). Cholesterol is synthesised via the mevalonate pathway, which is the focus of this thesis. Cholesterol can also be exported from the cell, such as through efflux via ATP-binding cassette transporter A1 (ABCA1), or converted to bile acid or oxysterol for excretion (27,28). These three processes are regulated primarily by two transcription factors, the sterol regulatory element-binding proteins (SREBPs) and liver X receptor (LXR) (27,29,30) (Figure 1.2).

Figure 1.2. Overview of transcriptional control in cellular cholesterol homeostasis. Cholesterol levels are regulated by control of cholesterol uptake, synthesis and efflux. These processes are regulated at the transcriptional level by two master transcription factors, the sterol regulatory element-binding protein 2 (SREBP-2) and liver X receptor (LXR). See text for details. Modified from (31).

1.2.1 The Scap/SREBP-2 pathway SREBPs are master regulators of lipid homeostasis, such as by controlling the expression of the LDL receptor and a whole suite of genes required for the synthesis of cholesterol, fatty acids, triglyceride and phospholipids (32). There are three isoforms of SREBP in humans: SREBP-2, SREBP-1c and SREBP-1a. SREBP-2 primarily increases the expression of cholesterogenic genes, SREBP-1c primarily regulates fatty acid metabolism, and SREBP-1a targets both. SREBP-2 also activates expression of PCSK9 (Proprotein convertase subtilisin/kexin type 9), an enzyme that induces degradation of the LDL receptor (33). The elegant mechanism whereby sterols feed back on their own synthesis has been elucidated in the laboratory of Brown and Goldstein (30) (Figure 1.3). SREBPs are synthesised as inactive precursors in the endoplasmic reticulum (ER) membrane, where they bind to SREBP cleavage-activating protein (Scap). Scap function requires its sterol-sensing domain, which consists of five transmembrane α-helices, with cholesterol binding to a luminal loop (34,35).

When cholesterol levels are low, Scap escorts SREBP from the ER to the Golgi apparatus via coatomer protein complex-II (COPII) vesicles, where SREBP is proteolytically processed to liberate the N-terminal domain, the mature transcription factor, which can then enter the nucleus to activate lipogenic genes, recognised by short sterol-response element DNA sequences (30,36). When cholesterol levels in the ER are high, cholesterol binds to Scap, inducing a conformational change, causing it to bind to Insig via Scap’s sterol- sensing domain (34,35,37). The changed conformation of Scap hides a sorting signal required for transport vesicle assembly (38). In this way, Insig behaves as a retention protein. Alternatively, oxysterols such as 25-hydroxycholesterol can stimulate complex formation by binding to Insig itself, inducing binding to Scap and causing a similar conformational change (39).

Figure 1.3. The Scap/SREBP pathway. Sterol regulatory element-binding protein (SREBP) exists as an inactive precursor in the endoplasmic reticulum (ER). When cholesterol levels are low, SREBP cleavage-activating protein (Scap) escorts SREBP to the Golgi, where it is sequentially cleaved by site 1 (S1P) and site 2 protease (S2P). The active N- terminal transcription factor can then enter the nucleus, bind to sterol-response elements (SRE) and drive expression of lipogenic target genes. When cholesterol levels are sufficient, Scap binds to Insig, retaining the Scap/SREBP complex in the ER and preventing gene expression. Modified from (10). 7

1.2.2 SREBP-2 processing may be increased by the important signalling kinase, Akt. SREBP-2 activation is inhibited by sterols, but evidence from our laboratory and others suggests that SREBP-2 processing can be increased by signalling pathways, namely the phosphatidylinositol 3′-kinase (PI3K)/Akt pathway (40). The PI3K/Akt pathway plays a central role in regulating growth and proliferation in response to external signals, such as insulin and insulin-like growth factor-1 (IGF-1) (41,42). Hence, its involvement in cholesterol synthesis might be expected, as cholesterol is required for building new membranes (40). PI3K is first activated by growth factor receptors, causing it to convert the plasma membrane lipid phosphatidylinositol (4,5)-bisphosphate to phosphatidylinositol (3,4,5)-trisphosphate (PIP3). The important downstream kinase Akt is then recruited to the plasma membrane through binding to PIP3, where it is in turn activated by phosphoinositide-dependent kinase (PDK)-1 and PDK2. Pharmacological inhibition of PI3K or over-expression of a dominant negative form of Akt inhibited the increase in SREBP-2 processing by IGF-1, as well as ER to Golgi trafficking of GFP-tagged Scap (40). However, PI3K has other important targets other than Akt. Similarly, dominant negative Akt inhibits endogenous Akt by competing for upstream kinases that activate it, but this could prevent activation of other kinases (43). Hence, the direct involvement of Akt activity in SREPB-2 processing remains to confirmed.

1.2.3 The liver X receptor The LXR transcription factors modulate cholesterol homeostasis in the opposite direction to SREBP-2, promoting cholesterol reduction by increasing the expression of efflux genes such as ABCA1 and ABCG1 (ATP-binding cassette transporter G1), as well as MYLIP that encodes the inducible degrader of the LDL receptor (Idol), reducing cholesterol uptake (44). LXRs are activated by oxysterols and desmosterol rather than directly by cholesterol itself (45).

1.2.4 Cholesterol is synthesised by the mevalonate pathway Cholesterol is synthesised from acetyl-CoA via the mevalonate pathway, which involves more than thirty enzymes (Figure 1.4). The pathway also synthesises important isoprenoids through one branch off point, as well as a unique oxysterol via another shunt. The first rate-limiting step in the pathway is the synthesis of mevalonate, catalysed by 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGCR) (46). HMGCR is the target of the blockbuster statin class of drugs, which inhibit cholesterol synthesis in the liver. This consequently leads to the compensatory up-regulation of the LDL receptor by SREBP-2, reducing serum cholesterol levels by increasing uptake whilst de novo synthesis is blocked. A potential problem with the statins is that the synthesis of important non-sterol products that branch off after mevalonate is also inhibited (Figure 1.4). These include isoprenoids used for post-translational modification of signal transduction proteins such as Rho (10), and ubiquinone (coenzyme Q10), potentially interfering with the electron transport chain (47). This may explain some of the adverse side-effects of statin treatment, which can include muscle and liver damage (48). Hence, greater understanding of the regulation and properties of later enzymes in the pathway could reveal novel therapeutic targets for effectively lowering cholesterol levels that do not present these problems.

Figure 1.4. Schematic of cholesterol synthesis by the mevalonate pathway. Abbreviations: 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGCR), 2,3(S)-monooxidosqualene (MOS), 2,3(S):22(S),23-dioxidosqualene (DOS). Squalene monooxygenase also catalyses the conversion of MOS into 2,3(S):22(S),23-dioxidosqualene, the precursor for the potent oxysterol 24(S),25- epoxycholesterol. Lanosterol synthase yields the first sterol in the pathway. See text for details. Modified from (15).

1.2.5 Regulation of HMGCR Study of the regulation of cholesterol synthesis has focussed on HMGCR, as the first rate-limiting step and important drug target (49). Like the majority of the enzymes in the pathway, it is associated with the ER membrane. HMGCR is regulated at numerous levels, including transcriptionally by SREBP-2, through inhibition of translation, and by phosphorylation (50,51). At the post- translational level, HMGCR is also rapidly degraded by the ubiquitin-

proteasome system, an example of endoplasmic reticulum associated degradation (ERAD), stimulated by side-chain oxysterols, particularly 27- hydroxycholesterol (52), 24,25-dihydrolanosterol and non-sterol derivatives of mevalonate (53,54). This is facilitated by Insig, playing an alternative role to retention of the Scap/SREBP complex, where Insig-1 binds to HMGCR via reductase’s own sterol-sensing domain. In the absence of Scap, Insig-1 can form a complex with a ubiquitin ligase such as Gp78 and associated machinery, mediating the polyubiquitination of HMGCR and its targeting to the proteasome for degradation (55). Membrane extraction involves VCP/p97 and dislocation via a lipid droplet-like compartment (56,57). Since Scap and HMGCR compete for Insig binding, and similar amino acid substitution mutations disrupt this, it is likely that they bind to the same site (58). In comparison to HMGCR, very little is known about the post-transcriptional regulation of the other cholesterol biosynthetic enzymes.

1.3 Endoplasmic reticulum-associated degradation (ERAD)

Just as cholesterol levels need to be kept within a narrow range for optimal function, the cell also requires mechanisms to control the level of different proteins and thereby maintain proteostasis. One way to avoid the toxic buildup of excess or misfolded proteins is to specifically degrade them. The major pathway by which this occurs is the ubiquitin-proteasome system. For proteins within the ER lumen or buried in the ER membrane, this process is referred to as ER-associated degradation (ERAD). A well studied example of ERAD is the regulated of degradation of HMGCR, discussed above. A more general example of ER protein quality control can involve binding of molecular chaperones to a misfolded protein and attempts to refold it. If the correct conformation is not achieved, the protein can be transported from the ER into the cytosol, polyubiquitinated and delivered to the proteasome - a large multi protein complex with proteolytic activity that degrades the target into peptides (59-61). Ubiquitination requires activation of the small and highly conserved ubiquitin protein by an E1 ubiquitin-acitivating enzyme, then action of an E2 ubiquitin- conjugating enzyme and an E3 ubiquitin ligase to join ubiquitin to the target protein (62). Substrate recognition for ubiquitination appears to be mediated primarily by the E3 ubiquitin ligase. Further ubiquitins are then added to the

11 original to form the polyubiquitin chain, which can be recognised by the proteasome, particularly for chains of four or more joined by K48 linkages of the first ubiquitin to the N-terminus of the next. The ubiquitins are then released prior to degradation of the substrate protein (61).

1.4 Squalene monooxygenase

1.4.1 Background Squalene monooxygenase (SM), formerly known as squalene epoxidase, performs the first oxygenation step in cholesterol synthesis, converting squalene to 2,3(S)-monooxidosqualene (MOS) (Figure 1.5), which is committed solely to sterol synthesis. Squalene monooxygenase also catalyses the conversion of MOS into 2,3(S):22(S),23-dioxidosqualene (Figure 1.4), the precursor for the potent oxysterol 24(S),25-epoxycholesterol, which fine tunes acute cholesterol synthesis (63). This oxysterol can enhance degradation of HMGCR, block SREBP processing, is more easily trafficked than cholesterol, and can act as a very potent ligand for the LXR nuclear receptor, stimulating cholesterol efflux (15,63,64). SM is a mixed-function (with oxidation of squalene and FAD/NADPH), flavin monooxygenase bound to the ER membrane, requiring oxygen, FAD, NADPH, and an electron transfer partner, such as cytochrome P450 reductase for activity (65-69). The 64 kDa mammalian forms derived from rat and human (70) also require anionic phospholipids and a supernatant protein factor (SPF) for activity using in vitro assays. SPF was first isolated from the cytosolic fraction of rat liver, and is thought to be involved in lipid transport, with strong binding to the phospholipid phosphatidylinositol, but poor binding to squalene itself (71). These components can be replaced with the detergent Triton X-100 for in vitro assays (72).

Figure 1.5. Schematic of the reaction catalysed by SM. Diagram of the reaction catalysed by SM, including necessary cofactors. The oxygen in the epoxide group will eventually become the 3β-hydroxyl group of cholesterol. SM can also act a second time, on MOS, to synthesise 2,3(S):22(S),23-dioxidosqualene (DOS).

1.4.2 Structure of SM The tertiary structure of SM is currently unknown. At the level of primary structure, the Saccharomyces cerevisiae (S. cerevisiae) enzyme shares ~30% amino- acid identity with the vertebrate, but lacks the first 77 N-terminal residues, thought to contain one or more transmembrane domains (70). Alternative membrane topology models with multiple transmembrane domains have been suggested for S. cerevisiae (73). A truncated form of the recombinant rat enzyme missing the first 99 residues retains activity (74). This suggests that the vertebrate N-terminal region is a structurally and functionally distinct domain that may play a role in regulation. In contrast, both the vertebrate and S. cerevisiae enzyme contain numerous highly conserved aromatic residues required for normal activity (75). Photoaffinity labelling and site-directed mutagenesis experiments for the rat enzyme have identified putative active site residues towards the C-terminal half of the enzyme (76,77). All homologues

13 contain highly conserved FAD binding domains; the dinucleotide binding GxGxxG (β1-sheet-α-helix-β2-sheet) motif, found in distantly related hydroxylases, DG and GD motifs, categorising SM as a flavoprotein.

1.4.3 SM as a drug target Unlike HMGCR, SM has not been used clinically as a drug target to treat high serum cholesterol. However, squalene is a stable and non-toxic intermediate in the sterol-committed branch of the mevalonate pathway, making it attractive for drug design. Preclinical trails of mammalian inhibitors were promising, but use in dogs was associated with skin toxicity, possibly due to non-specific effects of the drug, NB-598 (78-80). In contrast, SM is an important target for anti-fungals such as terbinafine, which is ineffective against the mammalian enzyme (81).

1.4.4 SM may catalyse a rate-limiting step in cholesterol synthesis SM has been proposed to be a second rate-limiting enzyme in the mevalonate pathway, but this is not widely appreciated. It had an extremely low specific activity compared to HMGCR in Hep G2 (liver) cells, and levels of protein and mRNA were also low (65,82). Incubation of Chinese hamster ovary (CHO) cells with labelled mevalonate resulted in the accumulation of squalene, but the effect was decreased in the presence of lipid-depleted media (83). Accumulation of squalene was also observed in renal carcinoma cells cultured in the presence of cholesterol (55, 56). This is unusual in light of the rate- limiting properties of HMGCR, since squalene is downstream of mevalonate (Figure 1.4). One would expect that the flux through the pathway would be lowered at the heavily regulated step catalysed by HMGCR, preventing the production of sufficient precursors required for conversion to accumulating squalene. Thus, this rate-limiting behaviour suggests the presence of additional levels of regulation of SM in addition to transcriptional control by SREBP-2, the only mechanism previously documented in the literature (84). In the yeast S. cerevisiae, the major product of the mevalonate pathway is ergosterol instead of cholesterol. Overexpression of truncated HMGCR lacking a sterol-sensing domain in this organism significantly enhanced accumulation of squalene, but with little or no increase in the level of ergosterol product (85), again highlighting the possible rate-limiting nature of SM.

1.4.5 Regulation of SM As an SREBP-2 target gene, SM is regulated at the transcriptional level in response to changes in cholesterol levels, with a potential sterol response element close to (~250 bp) the transcriptional start site (82,84). Regulation might also involve stimulation by supernatant protein factor (SPF), which can be activated by phosphorylation in response to changes in cAMP levels, but apparently not on the basis of cholesterol status (86,87). SPF increased the activity of HMGCR and SM in a rat cell-free system, but the hamster SPF only had a low stimulatory effect on activity (87-89). Furthermore, it was found that substitution of the cytosolic fraction isolated from either rat liver or CHO cells grown in lipid-depleted media did not increase activity compared to the equivalent fraction from rich media, but rather the activity was solely dependent on the amount of microsomal SM, in turn determined by cholesterol status (88). These data suggest that SPF does not represent a sterol-dependent regulator of SM activity. In contrast, one group found that a prolonged high fat diet was associated with a halving of the stimulatory effect of SPF in rat (90), but did not demonstrate the short-term changes in activity that would be required to explain rapid squalene accumulation upon sterol addition. Reduction of activity might also be achieved by degradation of the enzyme, which would have the regulatory advantage of rapidity, as seen for the proteasomal degradation of HMGCR. Studies of knockout mice where the first known electron partner for SM, cytochrome P450 reductase, was absent, showed that whilst SM and HMGCR mRNA levels were higher than normal, protein levels were decreased for both enzymes, again suggesting that SM is post-transcriptionally regulated (69).

1.5 Ligation-independent cloning

The cloning and manipulation of DNA in plasmid vectors are essential steps in the projects of most molecular biology laboratories. Traditional cloning involves amplification of a DNA fragment of interest using PCR, followed by restriction digestion of insert and vector, and ligation in vitro. This can be slow and inefficient, so ligation-independent techniques have since been developed. These include polymerase incomplete primer extension (PIPE) cloning (91), sequence and ligation-independent cloning (SLIC) (92), and overlap extension 15 cloning (OEC) (93,94). Although they are simpler and highly efficient, they have not all been fully optimised and their relative effectiveness remains unclear. Hence, gaining and sharing such knowledge should accelerate the research for this thesis, as well as assist many others working on unrelated topics.

1.6 Aims and hypotheses

The primary aim of this work is to characterise the regulation of squalene monooxygenase. This may be facilitated by optimisation, analysis and the practical application of ligation-independent cloning techniques. We hypothesise that in addition to transcriptional regulation via SREBP-2, sterols regulate SM at the post-transcriptional level. Consequently, the aims of this thesis are:

• Aim 1. To optimise and compare the simplest and most effective ligation-independent cloning techniques, and use them to help answer important questions in cholesterol homeostasis (Chapter 3 and throughout).

• Aim 2. To investigate the post-transcriptional regulation of SM by cholesterol, including work to uncover the molecular mechanism (Chapters 4 and 5).

• Aim 3. To study post-transcriptional regulation of SM by other lipids (Chapter 5).

Chapter 2

General materials and methods

2 General materials and methods

2.1 General materials

Chemicals and reagents used are listed below with the supplier. Dulbecco’s Modified Eagle’s Medium/Ham’s Nutrient Mixture F-12 (DMEM/F12), Dulbecco’s Modified Eagle’s Medium (DMEM), DMEM low glucose, DMEM high glucose (L-methionine, and L-cysteine free), Fetal Bovine Serum (FBS), hygromycin B, Newborn Calf Serum (NCS), penicillin- streptomycin, TRIzol Reagent, SuperScript III First Strand cDNA Synthesis Kit, Lipofectamine LTX, Opti-MEM I medium, Dynabeads, Amplex Red Cholesterol Assay kit, and anti-V5 antibody were purchased from Invitrogen (Carlsbad, CA). [1-14C]-acetic acid sodium salt (specific radioactivity: 56 mCi/mmol) and Glutathione Sepharose 4B beads were purchased from GE Healthcare (Chalfont St. Giles, UK). [2-14C]-mevalonolactone (mevalonate) (specific radioactivity: 40- 60 mCi/mmol) and [35S]-Protein Labeling Mix (EXPRE35S35S Protein Labeling Mix, specific radioactivity: >1000 Ci/mmol) were purchased from Perkin Elmer (Waltham, MA). Anti-SQLE (SM) antibody was purchased from Protein Tech Group (Chicago, IL). HA.11 monoclonal antibody was purchased from Covance (Princeton, NJ). Peroxidase-conjugated AffiniPure Donkey Anti-Mouse IgG and Peroxidase-conjugated AffiniPure Donkey Anti-Rabbit IgG were obtained from Jackson ImmunoResearch Laboratories (West Grove, PA). Lipoprotein-Deficient Serum (LPDS) was prepared from NCS as described previously (95). LDL (d=1.019-1.063 g/ml) was isolated by standard ultracentrifugation techniques from the plasma of healthy male volunteers (96). N-Acetyl-Leu-Leu-Norleu-al (ALLN), anti-α-tubulin antibody, butylated hydroxytoluene, chloroquine, compactin (also called mevastatin), cycloheximide, desmosterol, Dulbecco’s phosphate buffered saline (PBS), isopropyl β-D-1-thiogalactopyranoside (IPTG), lactacystin, methyl-β-cyclodextrin, mevalonate, Z-Leu-Leu-Leu-al (MG132), primers, protease inhibitor cocktail, N-ethylmaleimide, sodium oleate, sodium dodecyl sulphate (SDS), IGEPAL CA-630, sodium deoxycholate, Zaragozic acid A trisodium salt (squalene synthase inhibitor, SSi), methionine and cysteine, and X-gal were obtained from Sigma-Aldrich (St. Louis, MO). The squalene epoxidase inhibitor, GR144000X (squalene monooxygenase inhibitor, SMi) was kindly donated by Glaxo-Smith Kline (Middlesex, UK). iProof High- 19 Fidelity DNA Polymerase and Precision Plus Kaleidoscope protein marker were from Bio-Rad Laboratories (Hercules, CA). Phusion High-fidelty DNA Polymerases and restriction enzymes were from New England Biolabs (Ipswich, MA). SYBR Green SensiMix dT was from Quantace (Norwood, MA). 24(S),24-Epoxycholesterol (24,25EC) was obtained from Enzo Life Sciences (Farmingdale, NY). Cholesterol, lanosterol, lathosterol, 7-dehydrocholesterol (7DHC), 24,25-dihydrolanosterol (24,25DHL), 7α-hydroxycholesterol (7αHC), 7β-hydroxycholesterol (7βHC), 7-ketocholesterol (7KC), 19-hydroxycholesterol (19HC), 25-hydroxycholesterol (25HC), and 27-hydroxycholesterol (27HC) were obtained from Steraloids (Newport, RI). If not otherwise mentioned, oxysterols were delivered in ethanol. Sterols and oxysterols complexed with methyl-β- cyclodextrin (CD) were prepared as described (97). Sterol/CD complexes were diluted without addition of further cyclodextrin, so a constant molar ratio of ~0.1 sterol to methyl-β-cyclodextrin was used. All solvents used for thin layer chromatography (TLC) were analytical reagent grade from Ajax Finechem (Taren Point, NSW, Australia). Chinese hamster ovary-7 (CHO-7), SRD-1, SRD- 13A, and HEK293 cells were generous gifts of Drs. Michael S. Brown and Joseph L. Goldstein (UT Southwestern Medical Center, Dallas, TX). SRD-15 cells were generously donated by Dr Russell DeBose-Boyd (UT Southwestern Medical Center, Dallas, TX). HepG2 cells and primary human fibroblasts were kind gifts from the Centre for Vascular Research (UNSW, Sydney, NSW, Australia). BE(2)C cells were generously donated by Dr Louise Lutze-Mann (UNSW, Sydney, NSW, Australia). The HA-tagged ubiquitin plasmid, pMT123, encoding 8 tandem HA-ubiquitins (98), was a gift from Dr Dirk Bohmann (University of Rochester Medical Center, Rochester, NY).

2.1.1 Media recipes

Figure 2.1. Media formulations. Medium*

A DMEM/F12 supplemented with 5% LPDS B Medium A containing 5 µM compactin, and 50 M mevalonate C DMEM/F12 supplemented with 5% NCS D DMEM low glucose supplemented with 10% FBS E DMEM low glucose supplemented with 5% LPDS F Medium E containing 5 M compactin, and 50 M mevalonate G DMEM high glucose supplemented with 10% FBS H DMEM high glucose supplemented with 5% LPDS I Medium H containing 5 M compactin, and 50 M mevalonate

*All media containing penicillin (100 U/ml), streptomycin (100 g/mL), and L-glutamine (2 mM)

2.2 Cell culture

In general, various Chinese hamster ovary (CHO) cell-lines were employed (kind gifts of Drs. Goldstein, Brown, and DeBose-Boyd, UT Southwestern, Dallas). Unless otherwise stated, cells were statin pretreated overnight (16 hr) to deplete sterols through incubation in medium containing LPDS, the HMGCR inhibitor compactin (5 M), and a low level of mevalonate (50 M) that allows synthesis of essential non-sterol isoprenoids but not of cholesterol (57). After statin pretreatment, cells were washed once with PBS, which was sufficient to remove any residual compactin. For treatment, the media was refreshed to include test agents as described in the figure legends for the times indicated, followed by cell harvesting for the assays described below. Flp-In stable cell lines were maintained in 200 g/mL hygromycin B.

21 2.3 Metabolic labelling of squalene and cholesterol

Accumulation of [14C]-squalene and [14C]-cholesterol were determined by radio-TLC (thin layer chromatography) as described (63) with minor modifications. Following statin pretreatment, cells were metabolically labelled with 1 Ci/well [14C]-acetate or [14C]-mevalonate added to the existing media for the last 2 or 4 hr of treatment, as indicated. Cells were washed once with

PBS, lysed in 500 L 0.1 M NaOH, and rinsed with 1.25 mL H2O. Protein concentrations were measured by the Bicinchoninic Acid method (Pierce, Rockford, IL). Lysates were saponified with 500 L 20% KOH (w/v) in methanol, butylated hydroxytoluene (1 L, 20 mM), and EDTA (20 L, 20 mM) at 70 ºC for 1 hr. After cooling, the lipids were extracted with 2 mL hexane and evaporated to dryness. Extracts were re-dissolved in 60 L hexane and aliquots corresponding to equivalent amounts of protein separated on Silica Gel 60 F254 plates (Merck, Whitehouse Station, NJ) with a mobile phase of hexane: diethyl ether: glacial acetic acid (60:40:1, v/v/v). Bands corresponding to cholesterol and squalene (with relative Rf values of ~0.4 and ~0.9, respectively) were visualised using the FLA-5100 phosphorimager (Fujifilm, Tokyo, Japan). The relative intensities of bands were quantified using Sciencelab ImageGauge 4.0 Software (Fujifilm).

2.4 Quantitative real-time PCR

As previously described (63), RNA was harvested in triplicate using TRIzol reagent, reverse transcribed to yield cDNA using the SuperScript III First Strand cDNA Synthesis kit (Invitrogen), and mRNA levels determined relative to the housekeeping gene by quantitative real-time PCR using SYBR Green and a Corbett Rotorgene 3000. Primers (Table 2.1) were directed against SM (SQLE), UBXD8 (FAF2), MARCH6 (MARCH6) and HMGCR (HMGCR), with porphobilinogen deaminase (Pbgd) as the housekeeping gene.

Table 2.1. Primer sequences for quantitative real-time PCR analysis Gene Direction Primer Sequence (5’-3’) Reference

SQLE Forward TCTGATACACGGCTACATAG Present study

Reverse ACTTGCCATGGTGGAAAGCAAC

HMGCR Forward CTGGTGATGGGAGCTTGCTGTG (40)

Reverse AATCACAAGCACGAGGAAGACG

Pbgd Forward AGATTCTTGATACTGCACTC (99)

Reverse TGAAAGACAACAGCATCACA

FAF2 Forward GCTGCTTGGATGGGGTTATTAC Present study

Reverse GGCCGTATAAAACGAAGAGCAAA

MARCH6 Forward GAGGTACTTCGACCTGGTGTC (100) Reverse CACTGTAGAGCATGACATTG

2.5 Immunoblot analysis

Immunoblot analysis was performed as described (101) with minor modifications. Cells were lysed in 100 L 10% (w/v) SDS with 5 L protease inhibitor cocktail. Samples (usually 40 g of protein) were analysed by 10% SDS-PAGE and immunoblotted with the following antibodies: anti-V5 (1:10,000), anti-SM (1:5000), anti-HA (1:10,000), and anti-α-tubulin (1:200,000). The observed endogenous protein bands migrated according to their calculated molecular weight: 64 kDa for SM, 42 kDa for SM N100-GFP/GST and 50 kDa for α-tubulin. The relative intensities of bands were quantified by densitometry using ImageJ Software (1.36b).

23 2.6 Transfection

CHO cells in 6-well plates (unless otherwise stated) were grown in medium A (but without antibiotics) and transfected the following day using Lipofectamine LTX reagent (Invitrogen) according to the manufacturer’s instructions, using a ratio of 1 g of DNA: 4 L of transfection reagent. DNA was equalised with empty vector between different conditions. Following 24 hr transfection, cells were statin pretreated overnight, and then treated as indicated.

2.7 Cholesterol mass determination

Cells were washed, harvested in modified RIPA buffer (50 mM NaCl, 1.0% IGEPAL CA-630, 0.5% sodium deoxycholate, 0.1% SDS, 200 mM Tris, pH 8.0) and passed through an 18G needle 20 times. Total cellular cholesterol content was determined using the Amplex Red Cholesterol assay kit (Invitrogen), according to the manufacturer's instructions (with an Fmax microplate spectrofluorometer (Molecular Devices, CA) or POLARstar Omega (BMG Labtech, Ortenberg, Germany), excitation λ = 544 nm, emission λ = 590 nm), and expressed relative to protein (measured by the Bicinchoninic Acid assay, Pierce).

2.8 Data presentation

Unless otherwise indicated, values are normalised to the vehicle-treated control condition. Quantitative data are presented as averages and all error bars are SEM. Other data are comprised or representative of at least n separate experiments, as noted in the figure legends.

Chapter 3

A practical comparison of ligation- independent cloning techniques

Julian Stevenson, James Krycer and Lisa Phan carried out the DNA cloning work. Julian Stevenson prepared the Akt stable cell-line. Winnie Luu performed the mammalian experiment and immunoblotting.

Data in Figure 3.6 has been presented previously in the PhD thesis of co-author Winnie Luu (102).

This work has been published in:

Stevenson J.,* Krycer J. R.*, Phan L., and Brown A. J. (2013) A practical comparison of ligation-independent cloning techniques. PLOS ONE 8(12): e83888. doi: 10.1371/journal.pone.0083888 *Equal first author

Luu, W., Sharpe, L. J., Stevenson, J., and Brown, A. J. (2012) Akt acutely activates the cholesterogenic transcription factor SREBP-2. Biochimica et Biophysica Acta 1823, 458-464

3 A practical comparison of ligation-independent cloning techniques

3.1 Introduction

The precise assembly of specific DNA sequences is a critical technique in molecular biology. Traditional cloning makes use of restriction enzymes and ligation of DNA in vitro. Restriction endonuclease digestion and ligation increase the complexity of cloning projects, for example by requiring selection of appropriate restriction-sites and inefficient ligation steps. Consequently, several ligation-independent cloning (LIC) methods have since been developed that are simpler, faster, and highly efficient. These strategies rely on the generation of DNA fragments with single-stranded complementary ends to allow directional cloning of any insert, independent of restriction enzymes and in vitro ligation. The most effective and convenient methods include polymerase incomplete primer extension (PIPE) cloning (91), sequence and ligation- independent cloning (SLIC) (92), and overlap extension cloning (OEC) (93,94) (Figure 3.1). The relative effectiveness and suitability of these techniques for different types of cloning project are currently unclear. In this chapter, we will empirically compare these cloning strategies.

Figure 3.1. Principles of polymerase incomplete primer extension (PIPE) cloning, sequence and ligation-independent cloning (SLIC) and overlap extension cloning (OEC). In PIPE, incomplete extension during PCR generates 3'-recessed ends. In SLIC, purified PCR products are treated with T4 DNA polymerase (DNAP) so that the exonuclease activity will increase the proportion of recessed ends. In both these techniques, by amplifying vector and insert with primers containing complementary 5'-tails and mixing the products, the overhangs can anneal and are joined in vivo after transformation into Escherichia coli. In OEC, the insert PCR product acts a megaprimer to generate a nicked plasmid by overlap extension in vitro in a second round of amplification. The nicks are also repaired in vivo. For all techniques, a DpnI digestion step is included to remove plasmid template (not shown).

All three techniques amplify the gene of interest by polymerase chain reaction (PCR). The 3’ ends of the PCR primers are template-specific, whilst the 5’ ends incorporate tails specific for the cloning junction. PIPE relies on the observation that a significant portion of PCR products are incomplete, having 3’-recessed ends (91), particularly in the absence of a final extension step. SLIC uses a brief treatment of purified PCR products with the 3’ → 5’ exonuclease activity of T4 DNA polymerase to generate a higher proportion of recessed ends (92,103). Cloning is possible due to the complementary 5’ ends of the insert and vector fragments. These are analogous to the much shorter ‘sticky ends’ generated by restriction enzymes. After mixing, these single-stranded overhangs anneal and can be efficiently ligated

in vivo after transformation by the bacterial DNA recombination and repair machinery. In contrast, OEC has two rounds of amplification. Firstly, the insert is PCR-amplified using primers with 5’ ends complementary to the target site in a circular destination vector. The 3’ ends of this product (‘megaprimer’) can consequently anneal and amplify the destination vector by overlap extension. This yields a nicked plasmid that is repaired after transformation. A variant of the OEC strategy is utilised by the commonly used QuikChange site-directed mutagenesis kit (Agilent Technologies). In an age of high-throughput molecular biology, there is a need to move away from traditional cloning methods. We believe that LIC methods offer a more robust, efficient means for DNA cloning. Furthermore, variations between strategies suggest that certain techniques may be better suited for particular situations. Consequently, here we characterise and compare the efficiency, convenience, and utility of three major LIC techniques. This knowledge will allow the easiest preparation of various molecular tools necessary to answer specific biological questions throughout this thesis, such as the example of a Flp-In inducible Akt construct in this chapter.

29 3.2 Materials and methods

3.2.1 Primer combinations

Primer sequences are listed in Table 3.1. Gene accession numbers and plasmid vector backbones are listed in Table 3.2. The pUC18/Kan reporter vector was prepared using PIPE cloning. The vector backbone of pUC18 was obtained by PCR with primers pUC18–F and pUC18–R, and the kanamycin resistance cassette from pEGFP-C1 (Clontech) with primers KanR-in-pUC-F and KanR-in- pUC-R. A 24 bp FLAG epitope insertion (84 bp fragment) was achieved through PIPE or SLIC with primers FLAG-pUC18-F and FLAG-pUC18-R to amplify pUC18/Kan. The 85 bp FLAG megaprimer for OEC was prepared through 40 L thermal cycling of the primers alone, with 4 M FLAG-pUC18-overlap-F and FLAG-pUC18-overlap-R with 1x HF Buffer, 0.8 U of Phusion Hot Start II High-Fidelity DNA Polymerase (New England Biolabs) and cycling conditions: (98 °C 30 s, 63 °C 30 s, 72 °C 1 min) x 5. The pUC18/Kan vector backbone was amplified with primers pUC18- L30-F and pUC18-L30-R, which are the reverse complement of the common 5’-tails of the insert primers. A 350 bp fragment (Gluc) was amplified from FLAG-hLXRβ-hGluc(1) (31) with primers Gluc-pUC18-F and BGH-pUC18-R. LXRβ (1.4 kb) was similar amplified with primers LXRb-pUC18-F and LXRb- pUC18-R. A 2.2 kb sequence (AR) from pTK-AR-V5 (31) was obtained with primers AR-pUC18-F and AR-pUC18-R. SRC-1 (4.3 kb) was amplified with primers SRC1-pUC18-F and SRC1-pUC18-R from pCR3.1-SRC1 (104). 1.4 kb Insig-1 and 4.3 kb SCAP fragments were obtained with primers T7-pUC18-F and BGH-pUC18-R from pCMV-Insig-1-Myc or pCMV-SCAP respectively (105).

Table 3.1. Primer sequences for LIC comparisons. Vector tail sequences are indicated by boldface. Overlapping reverse- complementary sequences are underlined. Primer Name Primer Sequence (5’ → 3’) KanR-in-pUC-F CGTGCCAGCTGCATTAATGAATCGGCCAACGCGGAACCCCTATTT GTTTA KanR-in-pUC-R CGAGCGCAGCGAGTCAGTGAGCGAGGAAGCTCATTTCGAACCCC AGAGTC pUC18–F GCTTCCTCGCTCACTGACTC pUC18–R GTTGGCCGATTCATTAATGC pUC18-L30-F CGAGCTCGAATTCGTAATCATGGTCATAGC pUC18-L30-R CGCATATGGTGCACTCTCAGTACAATCTGC FLAG-pUC18-F GACTACAAGGACGACGACGACAAACGAGCTCGAATTCGTAATCAT GGTCATAGC FLAG-pUC18- TTTGTCGTCGTCGTCCTTGTAGTCCGCATATGGTGCACTCTCAGTA R CAATCTGC FLAG-pUC18- GCAGATTGTACTGAGAGTGCACCATATGCGGACTACAAGGACGAC overlap-F GACGACAAA FLAG-pUC18- GCTATGACCATGATTACGAATTCGAGCTCGTTTGTCGTCGTCGTC overlap-R CTTGTAGTC LXRb-pUC18-F GCAGATTGTACTGAGAGTGCACCATATGCGTCCTCTCCTACCACG AGTTC LXRb-pUC18-R GCTATGACCATGATTACGAATTCGAGCTCGCTCGTGGACGTCCCA GATCT Gluc-pUC18-F GCAGATTGTACTGAGAGTGCACCATATGCGCAACGAAGACTTCAA CATCGTG BGH-pUC18-R GCTATGACCATGATTACGAATTCGAGCTCGTAGAAGGCACAGTCG AGG T7-pUC18-F GCAGATTGTACTGAGAGTGCACCATATGCGCGCAAATGGGCGGT AGGCGTG SRC1-pUC18- GCAGATTGTACTGAGAGTGCACCATATGCGAGTGGCCTCGGGGA F CAGTT SRC1-pUC18- GCTATGACCATGATTACGAATTCGAGCTCGTTCAGTCAGTAGCTG R CTGAAGGA AR-pUC18-F GCAGATTGTACTGAGAGTGCACCATATGCGATGGAAGTGCAGTTA GGGCTGGGAA AR-pUC18-R GCTATGACCATGATTACGAATTCGAGCTCGCTGGGTGTGGAAATA GATGG

31 Table 3.2. Plasmid names and gene accession numbers. Gene sequences can be found on the NCBI GenBank or RefSeq databases. Vector backbone information is available on the Addgene Vector Database (http://www.addgene.org/vector-database/). Plasmid Gene NCBI Gene Vector Reference Accession Number Backbone pTK-AR-V5 AR NM_000044.3 pcDNA3.1D/V5- (31) His-TOPO FLAG-hLXRβ- hGluc(1) HQ388295.1 pCMX (31) hGluc(1) hLXRβ NM_007121.5 pCMV-Insig-1- Insig-1 NM_005542 pcDNA3 (105) Myc pCMV-SCAP Scap NM_012235.2 pcDNA3 (105) pCR3.1-SRC-1 SRC-1 NM_003743.4 pCR3.1 (104) pUC18/Kan See text for plasmid map. Current work

3.2.2 PCR PCR was performed in 50 L reactions using 0.5 ng of template, 0.5 M forward and reverse primers, 5% DMSO, HF Buffer and 1 U of the non-strand displacing enzyme Phusion Hot Start II High Fidelity DNA polymerase, and cycling conditions: 98 °C 3 min, (98 °C 30 s, 63 °C 30 s, 72 °C 3 min for products >1.5 kb or 1 min for <1.5 kb) x 30. Reaction products were column purified using a QIAquick PCR purification or gel extraction kit (Qiagen) and quantified by spectroscopy.

3.2.3 PIPE cloning Vector (0.025 pmol) and insert (0.0625 pmol) purified products (2.5:1 insert:vector ratio) were digested for 3 hr at 37 °C with 10 U of DpnI in 10 L CutSmart Buffer (NEBuffer 4.1/0.1 mg/mL BSA), diluted 1:1 with 1x HF buffer (to control for buffer composition between purified and unpurified products, as 1x HF buffer halved transformation efficiency), and 2 L used to transform 18 L of XL10 Gold ultracompetent cells (Agilent Technologies) according to manufacturer’s instructions.

3.2.4 SLIC DpnI digested purified PIPE products were incubated at 25 °C for 5 min with 0.75 U of T4 DNA polymerase, immediately placed on ice for 10 min, diluted 1:1

with ice-cold 1x HF Buffer, and 2 L used for transformation as described above.

3.2.5 OEC OEC reactions were performed using 10-250 fmol of insert product - 84 bp: 250 fmol (12.5 nM); 350 bp: 100 fmol (5 nM); 1.4 kb: 25 fmol (1.25 nM); 4.3 kb: 10 fmol (0.5 nM) - as megaprimer and 25 ng of pUC18/Kan vector template with 5% DMSO, 0.4 U of Phusion Hot Start II High Fidelity DNA polymerase in a volume of 20 L, and cycling conditions: 72 °C 5 min (to blunt the megaprimer), 98 °C 3 min, (98 °C 30 s 63°C 30 s 72 °C 3 min) x 30. Fractions (5 L) were diluted 1:1 with CutSmart Buffer (to control for buffer and assist DpnI activity), digested for 3 hr at 37 °C with 20 U of DpnI, and 2 L used for transformation as described above.

3.2.6 Colony counting and screening Serial dilutions of transformation mixture were spread onto ampicillin selective plates to allow counting of the number of colony forming units (CFU), adjusted to account for total volume, and rounded to 3 significant figures. 40 colonies were picked and streaked onto kanamycin selective plates in the presence of X-gal/IPTG to identify recombinants. For one replicate experiment of each set, colonies were also tested by colony PCR across the cloning junctions or sequenced to validate the blue/white screening.

3.2.7 Assembly of the pBI-CMV-FRB-Akt-Myc-2xFKBP-HA-FRT inducible Akt construct and creation of a stable Flp-In cell-line A plasmid containing a FRT recombination site and encoding FRB-Akt-Myc and myristoylated (Myr)-2xFKBP-HA driven by a bidirectional CMV promoter was created using PIPE with the primers listed in Table 3.3. An intermediate destination plasmid was first prepared by replacing the CMV promoter/enhancer to polyadenylation region of pcDNA5/FRT/TO (Invitrogen, Carlsbad, CA), amplified with vector primers pc5 F and pc5 R, with the bidirectional CMV promoter and multiple-cloning sites of pBI-CMV1 (Clontech Laboratories, Inc, Mountain View, CA), using insert primers BI CMV MCS F and BI CMV MCS R. Bovine AKT1 with a C-terminal Myc tag was amplified using Akt F and Akt Myc R, and subcloned into into the pC4-RHE plasmid encoding the FRB domain (ARIAD Pharmaceuticals, Cambridge, MA) using FRB Akt F and FRB Akt R. The splicing site, FRB Akt Myc and terminator was then inserted into the destination plasmid using insert primers FRB F and FRB R, with vector primers MCS1 F and MCS1 R. Myr-2xFKBP-HA from pC4M-F2E (ARIAD Pharmaceuticals, Cambridge, MA) was similarly introduced using insert primers FKBP F and FKBP R and vector primers MCS2 F and MCS2 R in a second cloning step, yielding the complete expression vector. The resulting pBI-CMV-FRB-Akt-Myc-Myr-2xFKBP-HA-FRT construct was verified by sequencing and used to prepare CHO-7 stable cells with the Flp-In system according to manufacturer’s instructions (Invitrogen, Carlsbad, CA), selecting for single colonies with 200 g/mL hygromycin B. Empty vector cells were prepared using a pcDNA5/FRT empty expression plasmid.

Table 3.3. Primer sequences for assembly of a Flp-In inducible Akt construct. Primer Name Primer Sequence (5’ → 3’) pc5 F AAAACTTGATTAGGGTGATGGTTC pc5 R AGGGAGCAGATACTGGCTTAACTA BI CMV MCS F AAGCCAGTATCTGCTCCCTCTAGACTGCAGCCTCAGGAGAT BI CMV MCS R TCACCCTAATCAAGTTTTCTCTGGAGATATCGTCGACAAGC Akt F ACGAATCTCAAAGACTAGTATGAACGACGTGGCCATC Akt Myc R CTGGTACGTCGTACGGATACAGATCCTCTTCTGAGATGAGT FRB Akt F TATCCGTACGACGTACCAGACTACGCA FRB Akt R ACTAGTCTTTGAGATTCGTCGGAAC FRB F TCCAGAGAAAACTTGATTAGTCAGATCGCCTGGAGACG FRB R AGGTACGTGAACCATCACCGAGGGATCTTCATAAGAGAAGAGG MCS1 F GGTGATGGTTCACGTACCTAG MCS1 R CTAATCAAGTTTTCTCTGGAGATATC FKBP F AATTCACCGGTCATATGCCTCAGATCGCCTGGAGACG FKBP R TGCCGCATAGTTAAGCCAGAGGGATCTTCATAAGAGAAGAGG MCS2 F ATCTGAGGCATATGACCGGTGAATTCTC MCS2 R CCTCTGGCTTAACTATGCGGCATCAG

35 3.3 Results

3.3.1 Design of the reporter system To efficiently compare the different cloning techniques, we prepared a reporter vector plasmid, pUC18/Kan (Figure 3.2A). It contains resistance cassettes for ampicillin and kanamycin, as well as a polylinker encoding the lacZα fragment. Each pair of insert-specific primers had the same 30 bp 5’-tail – this modular design allowed replacement of the multiple-cloning site of pUC18/Kan with a collection of inserts. This allows simple identification of the type of plasmid in each colony – insert plasmid, vector plasmid or desired recombinant plasmid – as the insert plasmids lack kanamycin resistance, and the vector plasmid template contains undisrupted lacZα. Hence, bacteria were first plated onto ampicillin plates and then patched onto kanamycin selective plates in the presence of IPTG for blue/white colony screening. Thus, positive colonies are white and resistant to both ampicillin and kanamycin (Figure 3.2B), whilst blue colonies possess pUC18/Kan and Kan-sensitive colonies possess insert- containing plasmids.

Figure 3.2. pUC18/Kan reporter plasmid. (A) Design of the reporter vector, encoding resistance for ampicillin and kanamycin, and the alpha-fragment of beta-galactosidase. (B) Colonies from ampicillin plates were patched onto kanamycin/X-gal plates to distinguish recombinants from unwanted insert vector or empty pUC18/Kan background colonies.

3.3.2 Generation of nicked vector plasmid reduces PIPE cloning efficiency. We began our optimisation with PIPE. After first transforming the DpnI- digested vector PCR product alone, we observed a significant number of blue (empty pUC18/Kan) colonies, but no colonies without the PCR (data not shown). Given that DpnI only digests methylated DNA, this indicates that vector plasmid is being generated during the PCR. This can occur because extension of the vector primers all the way around the plasmid template can regenerate the empty vector plasmid (Figure 3.3A). Note that this is linear amplification, generating less product in contrast to the desired PCR product between the primers, which is amplified exponentially by PCR. However, the vector PCR products can also re-anneal to the vector and extend analogously to the megaprimers in OEC, although this appeared to be less important, as addition of three 5’ non-template thymidines to the vector primers did not affect cloning efficiency (data not shown). Nevertheless, these situations can generate nicked copies of the vector in vitro that will escape DpnI digestion (Figure 3.3A). This leads to unwanted vector background colonies. To confirm this mechanism, we hypothesised that nicked vector background should be reduced by 1) increasing the gap between the 5’ ends of the vector primers, or 2) cutting the template in between the vector primers (Figure 3.3A). To test this, we replaced 350 bp, 1.4 kb, or 4.3 kb inserts in pUC18/Kan (i.e. different gap sizes) with the same 2.2 kb insert. The insert primer 5’ vector overhangs were complementary to the vector backbone, independent of the pre-existing insert sequences. This allows us to determine PIPE cloning efficiency – the proportion of successful recombinants - based on insert size differences using colony PCR. This was performed with or without linearising the vector using PstI. We found that both increasing the gap and linearisation reduced background (Figure 3.3C). However, because these strategies restrict the utility of PIPE-cloning, we sought a more flexible solution. We reasoned that reducing the template should dilute out the background – since the complete vector is linearly amplified, it is more dependent on template concentration and thus a lower template concentration would favour PCR amplification of the desired product. Accordingly, PIPE cloning efficiency increased from approximately 40% to greater than 95% when using ten-fold less template (Figure 2B), with similar results for insertion of the 350 bp, 1.4 kb and 37 4.3 kb inserts into pUC18/Kan (Table 3.4). We adopted this latter strategy in subsequent experiments.

Figure 3.3. Generation of nicked vector plasmid can reduce PIPE cloning efficiency. Unwanted copies of the original vector plasmid template can be generated by overlap extension in the process of obtaining vector-backbone PCR product (A). This can be prevented by reducing the template concentration (B), cutting the plasmid template first or cloning into a vector with an existing large insert (C). See main text for details.

Table 3.4. Effect of template concentration on PIPE cloning efficiency. Vector products amplified from 5 ng or 0.5 ng of template per 50 L reaction were combined with purified insert at a 5:1 I:V ratio. 40 colonies were patched onto kanamycin/X-gal/IPTG plates to screen for potential recombinants (white colonies), rather than empty vector (blue colonies) or insert vector (kanamycin sensitive, Kan (-), colonies). Cloning efficiency is the percentage of successful recombinants. Template Insert Colonies Cloning White Blue Kan (-) Efficiency 350 bp 2420 28% 11 29 0 5 ng 1.4 kb 1620 30% 12 28 0 4.3 kb 1070 8% 3 37 0 350 bp 967 88% 35 3 2 0.5 ng 1.4 kb 814 93% 37 2 1 4.3 kb 140 88% 35 4 1

3.3.3 Effect of cycle number and insert:vector ratio on PIPE cloning efficiency PIPE relies on incomplete extension in PCR. Although the proportion of incomplete products is sufficient at 25 cycles (91), we reasoned that additional cycles might deplete PCR components or generate DNA products that inhibit extension, leading to a greater proportion of recessed ends and thus more clones. However, PIPE cloning efficiency did not increase using equal amounts of DNA product taken from 25 to 40 PCR cycles (Table 3.5 & Table 3.6). We maintained subsequent reactions at 35 cycles to ensure that good product yields are achieved, particularly for difficult templates or inefficient primers.

39 Table 3.5. Cycle number for insert product does not affect PIPE cloning efficiency. Insert products purified from 25-40 cycle reactions were combined at a 5:1 I:V ratio with 40 cycle vector product. Insert Exp. # Cycles Colonies Cloning White Blue Kan (-) Efficiency 1 25 1080 63% 25 14 1 350 bp 30 830 78% 31 9 0 35 935 65% 26 14 0

40 880 68% 27 13 0

2 25 2220 83% 33 5 2 30 2610 73% 29 10 1 35 1860 60% 24 13 2

40 2600 78% 31 8 1 3 25 1550 50% 20 24 0 30 1450 58% 23 19 0

35 2430 73% 29 11 0

40 2200 63% 25 15 0 1 25 1200 80% 32 8 0 1.4 kb 30 1510 93% 37 3 0 35 1510 78% 31 9 0

40 1230 88% 35 5 0

2 25 685 48% 19 21 0 30 830 48% 19 20 1 35 1130 68% 27 12 1

40 1520 78% 31 9 0 3 25 646 78% 31 6 3 30 1400 75% 30 9 1

35 1140 85% 34 4 2

40 1020 80% 32 7 1 1 25 715 85% 34 6 0 4.3 kb 30 600 80% 32 8 0

35 605 68% 27 13 0

40 600 75% 30 10 0 2 25 330 40% 16 24 0 30 357 45% 18 22 0 35 961 75% 30 10 0

40 1960 95% 38 2 0 3 25 655 85% 34 2 4 30 740 73% 29 10 1

35 860 78% 31 8 1

40 940 80% 32 7 1

Table 3.6. Cycle number for vector product does not affect PIPE cloning efficiency. Vector products purified from 25-40 cycle reactions were combined at a 5:1 I:V ratio with 40 cycle insert products. Insert Cycles Colonies Cloning White Blue Kan (-) Efficiency 25 2450 93% 37 2 1 1.4 kb 30 2560 88% 35 5 0 35 3560 88% 35 5 0 40 3360 88% 35 5 0 25 638 95% 38 2 0 4.3 kb 30 660 88% 35 5 0 35 605 93% 37 3 0 40 412 75% 30 10 0

41 PIPE cloning tolerated a wide range of insert:vector molar ratios for a range of insert sizes (Table 3.7), with an optimum of 2.5:1, similar to previously reported values for SLIC (92,103).

Table 3.7. Effect of insert:vector ratio on PIPE cloning efficiency. Purified products were combined at the indicated molar ratios. Insert Exp. # I:V Colonies Cloning White Blue Kan (-) Efficiency 1 1:1 585 40% 16 24 0 350 bp 2.5:1 1180 75% 30 9 1 5:1 1250 80% 32 8 0 10:1 2240 90% 36 4 0 1 1:1 2700 68% 27 13 0 1.4 kb 2.5:1 2490 90% 36 4 0

5:1 2420 90% 36 4 0 7.5:1 3380 90% 36 4 0 10:1 3290 93% 37 3 0 2 1:1 1600 90% 36 4 0 2.5:1 1980 93% 37 3 0 3:1 3300 83% 33 3 0 5:1 3170 88% 35 5 0 10:1 2400 85% 34 6 0 3 1:1 775 90% 36 4 0 2.5:1 1190 95% 38 2 0 5:1 785 100% 40 0 0 7.5:1 975 88% 35 4 1 10:1 720 80% 32 8 0 1 1:1 516 78% 31 9 0 4.3 kb 2.5:1 715 90% 36 4 0 5:1 620 75% 30 10 0 10:1 500 85% 34 6 0

3.3.4 Optimisation of SLIC We made use of a one-tube version of SLIC (103,106). For a convenient volume of T4 polymerase (0.75 U/0.25 L), the highest efficiency was observed after 5-10 min treatment at 25 °C, although 5 min was most robust (Table 3.8), followed by immediate incubation on ice to halt the reaction.

Table 3.8. Effect of T4 exonuclease treatment time on colony number. Vector (0.025 pmol) and insert (0.0625 pmol) purified products were digested with DpnI then treated for 5 or 10 min with 0.75 U of T4 DNA polymerase, as described in the Materials and Methods. Untreated (0 min) conditions are identical to PIPE cloning. Insert Exp. # T4 Colonies Cloning White Blue Kan (-) treatment Efficiency (min) 1 0 2650 98% 39 0 1 85 bp 5 77700 100% 40 0 0 10 42000 100% 40 0 0 1 0 176 98% 39 1 0 350 bp 5 18500 98% 39 0 1 10 5440 100% 40 0 0 2 0 1040 91% 19 1 1 5 17000 100% 40 0 0 10 6300 100% 40 0 0 1 0 2420 100% 40 0 0 1.4 kb 5 20000 98% 39 0 1 10 20000 100% 40 0 0 2 0 136 100% 40 0 0 5 8360 100% 40 0 0 10 8340 100% 40 0 0 3 0 483 100% 40 0 0 5 10200 100% 40 0 0 10 11400 100% 40 0 0 1 0 714 95% 38 1 1 4.3 kb 5 13000 100% 40 0 0 10 23300 100% 40 0 0 2 0 474 100% 40 0 0 5 4150 98% 39 1 0 10 10700 95% 38 2 0 3 0 309 100% 40 0 0 5 6420 100% 40 0 0 10 2880 98% 39 1 0 43 3.3.5 Effect of megaprimer concentration, PCR template and cycle number on OEC OEC uses linear amplification, generating less product than the exponential amplification in PCR. We consequently used 25 ng of template for OEC, 50-fold higher than for the PCR-based PIPE and SLIC. This required doubling the concentration of DpnI to account for the increased template. The effectiveness of OEC was dependent upon megaprimer concentration, with the optimum being inversely proportional to megaprimer size: high concentrations of small insert and vice versa (Table 3.9). Thus concentrations of ~25-50 fmol (1.25 – 2.5 nM) were optimal for products >1.5 kb, ~50-100 fmol (2.5 – 5 nM) <1.5 kb, and ~100-300 fmol (5 – 15 nM) <350 bp. This is likely due to the increasing propensity of larger products to anneal to themselves rather than to the plasmid template.

Table 3.9. Effect of megaprimer concentration on overlap extension cloning efficiency. OEC reactions (20 L) were performed with 25 ng template and 30a or 18b thermal cycles using the indicated amount of purified megaprimer. cAnomalous result. Megaprimer Exp. # fmol Colonies Cloning White Blue Kan (-) Efficiency 1a 100 2480 85% 34 5 1 84 bp 250 15100 95% 38 1 1 500 11900 93% 37 3 0 750 1600 88% 35 4 1 1000 9300 88% 35 5 0 2a 100 5690 70% 28 12 0 250 6380 93% 37 3 0 500 4710 88% 35 5 0 750 3470 80% 32 8 0 1000 1730 80% 32 7 1 1b 10 299 23% 9 31 0 350 bp 33 2420 45% 18 22 0

100 3330 53% 21 18 1 330 7640 70% 28 12 0 Table 3.9 continues on the opposite page.

Megaprimer Exp. # fmol Colonies Cloning White Blue Kan (-) Efficiency 2b 10 525 15% 6 33 1 350 bp 33 2400 38% 15 24 1 (continued) 100 7720 45% 18 19 3 330 4410 43% 17 13 10 1000 3870 10% 4 2 34 3330 62 5% 2 15 23 3b 10 89 30% 12 27 1 33 473 58% 23 17 0 100c 42 10% 4 36 0 330 9840 63% 25 14 1 1000 4440 20% 8 15 17 3330 51 10% 4 7 29 4b 10 331 20% 8 32 0 33 3360 40% 16 23 1 100 10900 65% 26 12 2 330 11600 70% 28 12 0 1000 1240 38% 15 21 4 3330 10 10% 1 9 0 1a 10 306 3% 1 37 2 1.4 kb 25 408 18% 7 33 0 50 861 65% 26 14 0 100 1450 73% 29 11 0 2a 10 145 20% 8 32 0 25 19 24% 4 13 0 50 13 15% 2 11 0 100 0 n/a n/a n/a n/a 1a 10 251 45% 18 22 0 4.3 kb 25 135 35% 14 26 0 50 20 21% 5 18 1 100 0 n/a n/a n/a n/a 2a 10 211 13% 5 37 1 25 151 5% 2 38 0 50 57 0 0 39 1 100 3 0 0 3 0

45 We observed more kanamycin-sensitive colonies when the megaprimer was originally prepared from insert PCRs with higher template (5 ng). This is likely to be insert-vector background from the insert PCR, and was removed by reducing the template amount to 0.5 ng (Table 3.10).

Table 3.10. Increasing PCR template can generate insert template background in OEC for high megaprimer concentrations. OEC reactions (20 L) were performed with 25 ng template and 18 thermal cycles using the indicated amount of purified 350 bp Gluc megaprimer obtained from a PCR using 5 or 0.5 ng of template. PCR fmol Colonies Cloning White Blue Kan (-) template Efficiency 10 532 40% 16 24 0 5 ng 33 1920 38% 15 25 0 100 3800 58% 23 12 5 330 2940 50% 20 12 8 1000 2670 15% 6 6 28 3330 19 0% 0 6 13 10 557 18% 7 33 0 0.5 ng 33 2600 43% 17 23 0 100 5010 60% 24 15 1 330 2830 48% 19 20 1 1000 408 40% 16 16 8 3330 0 n/a n/a n/a n/a

Increasing the number of cycles of overlap extension gave progressively more colonies, without clearly increasing cloning efficiency (Table 3.11). Hence, 30 thermal cycles were used for comparisons to the other techniques.

Table 3.11. Colony number increases with additional cycles of overlap extension. OEC reactions (20 L) were performed with 25 ng template, 100 fmol of purified 350 bp Gluc megaprimer and the indicated number of thermal cycles. Exp. # Cycles Colonies Cloning White Blue Kan (-) Efficiency 1 14 3380 83% 33 7 0 16 5170 58% 23 17 0 18 9550 73% 29 11 0 20 13100 73% 29 11 0 22 15100 80% 32 6 2 2 14 4740 40% 16 24 0 18 8320 53% 21 18 1 22 14400 60% 24 14 2 26 28700 65% 26 12 2 30 33200 78% 31 6 3 3 14 2880 38% 15 23 2 18 6700 53% 21 18 1 22 12800 58% 23 16 1 26 15800 65% 26 14 0 30 21500 78% 31 9 0

3.3.6 OEC does not tolerate the presence of primer-dimers To test the ability of the techniques to tolerate the presence of primer-dimer, we attempted to clone a 1.4 kb gene product where the PCR product also included a marked amount of ~150 bp primer-dimer or mispriming product. Since these products contain 5’ ends complementary to the vector, they may be incorporated into recombinant clones, also yielding white colonies in our screening assay. Primer-dimer had little effect on PIPE or SLIC, as screened white colonies contained the correct insert, confirmed by colony PCR and sequencing (data not shown). In contrast, for OEC, most white colonies contained unwanted primer-dimer instead (Figure 3.4). Prior gel extraction of

47 the insert megaprimer and using this in the overlap extension step ensured that nearly all white colonies contained the desired insert instead.

Figure 3.4. OEC does not tolerate primer-dimers. Megaprimer and primer-dimer contaminant or 1.4 kb LXR megaprimer alone were gel purified and used for OEC. Ten white colonies for each were screened by PCR across the cloning junctions.

An alternative method to remove primer-dimer before OEC was to pretreat the purified insert mixture with T4 DNA polymerase in the absence of dNTPs (Figure 3.5). This allowed digestion of the primer-dimer, leaving the larger desired product to be end-filled - during the 5 min, 72 °C step at the start of the overlap extension reaction - and cloned.

Figure 3.5. Primer-dimer can be removed with T4 DNA polymerase exonuclease treatment. 1.4 kb LXR purified PCR product (0.25 pmol) was treated for 30 min at 25 °C with 3 U of T4 DNA polymerase.

3.3.7 Direct comparison To directly compare the effectiveness of the different cloning strategies, the reporter vector and different inserts were first amplified and purified, then fed into PIPE, SLIC and OEC with equivalent amounts of DNA using the optimised protocols (as described in the Materials and Methods). Reaction products were diluted with either HF buffer or NEBuffer to control for transformation buffer composition, as the Phusion 1x HF PCR buffer halved transformation efficiency (data not shown). The amount of DNA (25 fmol vector and 62.5 fmol insert) used for transformation was equivalent to that of unpurified PIPE cloning from a single pair of PCRs. 24 bp, 350 bp, 1.4 kb and 4.3 kb fragments were chosen to gauge the effect of increasing size on cloning efficiency for each strategy. The 24 bp insertion is a special case due to its small size, in that PIPE and SLIC used a single primer pair of reverse design for PCR – with 3’ ends complementary to the vector and 5’-tails containing the entire insert – thus requiring only one PCR, such that the product anneals to itself to form a circular product. For very small inserts, using two primer sets to amplify vector and insert would be wasteful and very inefficient for PIPE and SLIC. The standard insert primer design was used for OEC, with annealing and 3’ end- filling of partially overlapping primers alone to first prepare the insert megaprimer. PIPE consistently performed well, yielding hundreds of colonies with close to 100% cloning efficiency (table, Table 3.7, Table 3.12, Table 3.13). Addition of T4 DNA polymerase exonuclease treatment for SLIC increased the number of transformants by ~4-100 fold for all inserts, retaining the high efficiency (Table 3.12). OEC yielded very high numbers of transformants for the 24 bp (84 bp megaprimer) insertion, with close to 100% efficiency. Moderate to high efficiencies were also observed for the 350 bp insertion. Colony number fell proportionally as insert size increased, which was associated with a corresponding decline in cloning efficiency (Table 3.12). Cloning efficiencies and colony numbers were far more variable for larger fragments, reflecting lower robustness of OEC compared to the other LIC techniques (Table 3.9, Table 3.12, Table 3.13).

Table 3.12. Direct comparison of PIPE, SLIC and OEC for various insert sizes. See main text for details. Techniques were performed using the optimised conditions as described in the Materials and methods. aSingle representative experiments are shown for FLAG (85 bp), Gluc (350 bp), Insig-1 (1.4 kb), and Scap (4.3 kb). bThe increase in colonies relative to PIPE is shown.

Insert Cloning Fold Technique Colonies Sizea Efficiency Increaseb

85 bp PIPE 100% 1560 1

SLIC 100% 15800 10 OEC 95% 15100 10 350 bp PIPE 98% 176 1 SLIC 98% 18500 105 OEC 90% 9200 52 1.4 kb PIPE 95% 705 1 SLIC 100% 2760 4 OEC 73% 1450 2 4.3 kb PIPE 100% 309 1 SLIC 100% 6420 21 OEC 45% 251 1

Table 3.13. Complete screening for the comparison of PIPE, SLIC and OEC for increasing insert sizes. Techniques were performed as described in Table 3.12. Insert Exp. # Technique Colonies Cloning White Blue Dead Efficiency 1 PIPE 1680 100% 40 0 0 85 bp SLIC 15100 100% 40 0 0 OEC 15100 95% 38 1 1 2 PIPE 806 100% 40 0 0 SLIC 8740 100% 40 0 0 OEC 6700 100% 40 0 0 3 PIPE 2650 98% 39 0 1 SLIC 77700 100% 40 0 0 OEC 10600 93% 37 3 0 1 PIPE 284 93% 37 2 1 350 bp SLIC 1440 98% 39 1 0 OEC 15800 85% 34 6 0 2 PIPE 262 95% 38 1 1 SLIC 1870 98% 39 0 1 OEC 11400 78% 31 9 0 3 PIPE 1030 91% 19 1 1 SLIC 17000 100% 40 0 0 OEC 6970 90% 36 0 4 4 PIPE 176 98% 39 1 0 SLIC 18500 98% 39 0 1 OEC 9200 90% 36 1 3 1 PIPE 705 95% 38 0 2 1.4 kb SLIC 2760 100% 40 0 0 OEC 1450 73% 29 11 0 2 PIPE 183 100% 40 0 0 SLIC 1580 100% 40 0 0 OEC 145 20% 8 32 0 1 PIPE 309 100% 40 0 0 4.3 kb SLIC 6420 100% 40 0 0 OEC 251 45% 18 22 0 2 PIPE 110 95% 38 2 0 SLIC 1640 100% 40 0 0 OEC 211 13% 5 37 1 51

3.3.8 LIC can be performed without product quantification or purification For short fragments, a small fraction of PCR product from the average reaction will easily fall into the optimal range for PIPE, SLIC or OEC. For larger products (>1.5 kb), PIPE and SLIC will also perform well without quantification. Thus, investigators can skip quantification or purification for routine cloning. This can be important when higher throughput is required. PCR buffers can be inhibitory to transformation, as observed for the Phusion HF buffer (Table 3.14 and data not shown), but purification without quantification and short incubations with T4 polymerase are likely to provide a large number of colonies.

Table 3.14. Cloning of a 350 bp fragment without quantification or purification. 5 ng of pUC18/Kan pre-cleaved with PstI was amplified by PCR. a5 L of 350 bp Gluc insert and vector PCR products were combined and digested with DpnI without dilution or purification. b30 L of insert and vector PCR product were combined, purified and equal volumes treated with T4 for SLIC or left untreated for PIPE. c1 L of unpurified insert PCR product was used as megaprimer for OEC. d100 fmol of purified megaprimer was used for OEC. eThe number of colonies relative to unpurified, unconcentrated PIPE. Cloning Technique Cloning Colonies Folde Efficiency Increase PIPEa (Unpurified) 95% 226 1 PIPEb (Purified) 95% 631 3 SLICb (5 min) 100% 1920 8 SLICb (10 min) 100% 2220 10 OECc (Unpurified) 80% 44 0.2 OECd (Purified) 76% 206 1

3.3.9 An example of using LIC to solve a biological problem: Akt/Protein Kinase B acutely activates SREBP-2 Because LIC techniques are not constrained by the availability of restriction sites, nearly any DNA construct can easily be prepared. In this case, we wanted to directly confirm that activity of the kinase Akt (also known as Protein Kinase B) increases processing of SREBP-2, without confounding by upstream kinases. To achieve this, we prepared a specific rapid inducible heterodimerisation system (83), by co-expressing FRB-Akt-Myc and myristoylated (Myr)-2xFKBP- HA in cells. Our laboratory does not have facilities for retroviral transfection, so PIPE was used to incorporate a bidirectional CMV promoter, FRB-AKT fusion and FKBP expression cassettes into a vector containing a FRT site. This allows site-specific recombination by Flp recombinase of a single copy of the bicistronic vector into a previously prepared hamster cell-line containing a characterised FRT site. This one construct (Figure 3.6A) contains elements from five different plasmids, which would have been extremely difficult to assemble using restriction digestion and ligation, but was successfully prepared using successive rounds of PIPE cloning. Stable cells were rapidly generated using transient coexpression of Flp-recombinase (data not shown). The cells were then treated with rapalog (a non-immunosuppressive derivative of rapamycin) to induce heterodimerisation of the stably expressed genes, recruiting the FRB- Akt-Myc to the membrane-bound myristolated FKBP fragment. This brings the kinase into close proximity with activating proteins (PDK1 and mTOR complex 2), which phosphorylate it at the plasma membrane, activating Akt (Figure 3.6B). Indeed, rapalog activated FRB-Akt-Myc, as shown by increased phosphorylation (Figure 3.6C). Importantly, SREBP-2 was also activated upon rapalog addition in the FRB-Akt-Myc stable, but not in the control cell-line. This confirms that Akt activates SREBP-2 acutely, by increasing SREBP-2 processing, consistent with less specific lines of evidence using kinase inhibitors or dominant-negative Akt (40,107).

Figure 3.6. Rapalog-activated Akt increases SREBP-2 activation. (A) A schematic of the vector encoding FRB-Akt-Myc and Myr-2xFKBP-HA driven by a bidirectional CMV promoter with a pcDNA5/FRT backbone. (B) Under basal conditions, FKBP is expressed and anchored to the membrane via the myristoylation (Myr) anchor, leaving Akt inactive in the cytosol. When added, rapalog rapidly induces heterodimerisation of the FKBP and FRB fragments, recruiting Akt to the plasma membrane to be activated by PDK1 and mTOR complex 2 (mTORC2). (C) Stable cell-lines (expressing inducible Akt or empty vector; EV) were serum-starved overnight in 0.1% BSA (essentially fatty acid free) in DMEM (low glucose), and then treated with or without rapalog (0.5 M) for 1 hr. Whole cell lysates were subjected to SDS-PAGE and then transferred to nitrocellulose membranes. Membranes were probed for SREBP-2, pAkt, Akt, and α-tubulin. The indicated bands corresponding to the 56 kDa endogenous pAkt/Akt and the ~ 75 kDa FRB-Akt. The immunoblot is representative of at least two separate experiments, and the relative intensity (M) value represents the densitometric quantification of the mature form of SREBP-2, where the maximal condition (lane 2) has been set to 1.

55 3.4 Discussion

Ligation-independent cloning techniques can be used to introduce DNA fragments into cloning vector plasmids quickly, cheaply, and with high efficiency. Anything that can be amplified by PCR can be introduced into any position of any vector of choice in a single cloning step without unwanted additional nucleotides, so called ‘scarless cloning’. We observed that PIPE worked very well with limited manipulations, as long as the template concentration was kept to a minimum to avoid overlap extension vector background. SLIC consistently achieved the highest efficiencies and number of transformants, but required additional resources. OEC worked well for smaller fragments, but was less effective for larger fragments. It was also very vulnerable to the presence of primer-dimers. The methods compared here work either through the generation of complementary single-stranded overhangs for in vivo homologous recombination (PIPE, SLIC), or by generating a nicked plasmid in vitro by overlap extension (OEC). This can require the design of new primers with longer complementary tail sequences than used for restriction cloning, but the time savings and increased robustness outweigh these nominal costs, even for routine applications. Modular primers can also be designed to allow parallel cloning of a gene into different vectors with identical linkers (used often throughout this thesis), as well as inserting different genes with the same 5’-tails into the same vector, as in this chapter. The identical insert primer design also allowed the use of the same PCR product with the three different techniques. This can enable rescue of a failed cloning attempt using a different strategy with a higher efficiency but requiring additional steps and resources (Figure 3.7 & Table 3.15).

Figure 3.7. Technique selection flowchart for a new cloning project. The number of fragments, their size, primer availability and the presence of primer-dimers will determine the optimal cloning strategy. See main text for details.

Table 3.15. Summary of effectiveness and resource use. See main text for details. aOnly 1 primer pair is required for SLIC if cleaved plasmid vector is used. bOptional for small megaprimers (~500 bp).

LIC Techniques PIPE SLIC OEC Cloning Efficiency <1.5 kb Very High Very High High Colony Number <1.5 kb High Very High High Cloning Efficiency >1.5 kb Very High Very High Low Colony Number > 1.5 kb High Very High Low Primer Pairs 2 2a 1 Purification Optional Yes Yesb Thermal Cycling Rounds 1 1 2 T4 Treatment Step No Yes No

The first method of choice is dependent upon the nature of the insert, the availability of existing primers and the level of efficiency required. For genes of up to 1.5 kb, OEC is a good choice, as useful efficiencies (50%+) can be achieved with one primer pair. This will often be suitable for insertion of affinity tags or

57 fluorescent proteins, such as the green fluorescent protein. However, cloning efficiency and robustness for OEC decrease with increasing insert size, introducing the risk of failure for larger genes. The mechanism to explain this is currently unclear, but may involve the increased distance that the megaprimer ends can move during the annealing and extension steps, reducing the efficiency of correct binding. Polymerase-dependent unwanted background colonies might also result from uncharacterised structural changes to the template plasmid, such as alterations to supercoiling or permanent denaturation in vitro that provide resistance to DpnI digestion, or extension from a nicked end. Overlap extension also requires relatively long (~30 bp) target-vector-tail sequences in the primers to allow stable annealing, whereas PIPE and SLIC can work well with overhangs of ~15 bp, allowing synthesis of shorter primers (91,92,103). Purification of the insert PCR is recommended for OEC, but for smaller inserts, a wide range of megaprimer concentrations are tolerated, safely allowing use of a small fraction of unpurified PCR product as megaprimer (Table 3.14 and data not shown). Heterostagger or mixed PCR cloning only requires primers and polymerase to precisely engineer recessed ends similar to PIPE and SLIC (92,108). Although very effective, it is also costly as it requires double the number of PCRs and primers used in SLIC, and long denaturing and annealing steps, without performing better (92). PIPE cloning is generally simpler and robustly performs well, at the cost of requiring a second primer pair over OEC. For cloning projects where vector primers are already available or higher efficiencies are desired to clone inserts greater than 1.5 kb, PIPE is usually the best choice, achieving high efficiencies for a wide range of gene sizes (91) (Table 3.12 & Table 3.13). Unpurified PIPE PCR products can be used for transformation directly after DpnI treatment, making it extremely fast, especially useful for high-throughput cloning projects (91). This simple approach is the mainstay LIC technique in our laboratory and our default choice for any fragment size. However, purification to remove inhibitory buffer components and concentrate the DNA provides better results (Table 3.14 and data not shown). If PIPE cloning fails or a greater number of transformants are required, such as for library construction, then SLIC is superior, typically increasing the number of transformants by more than 5-fold in one step after purification, DpnI and T4 exonuclease treatment. SLIC is also effective after

combining purified fragments without quantification (Table 3.14). SLIC can also be performed using restriction enzyme-cleaved plasmid, minimising primer design at the cost of additional manipulations over PIPE (92,103,106). SLIC has the advantage that it can be used to assemble multiple insert fragments in a single cloning step, particularly with addition of recombinant RecA to enhance annealing in vitro, albeit with reduced efficiency compared to single-fragment cloning (92). This can also be achieved with high efficiency using the Gibson method (109), which uses similar primer design to PIPE and SLIC, but involves use of a 5’ exonuclease, DNA ligase and DNA polymerase in vitro to join previously generated PCR fragments. This has proved to be particularly useful for synthetic biology projects requiring assembly of very large DNA fragments. It would have made it possible to assemble the Flp-In inducible Akt construct weeks sooner. However, for single fragment insertion, we do not believe that the increased resources required are justified when PIPE and SLIC robustly achieve high efficiencies. However, Gibson assembly (109) can prove a highly effective alternative when even SLIC fails (Figure 3.7). The LIC techniques tested in this study can be highly efficient and technically simple, but can be compromised by problems such as nicked vector plasmid background, cloning of primer-dimers and amplification of non- specific PCR products. However, these can be overcome through careful design, checking PCR products, and taking additional measures accordingly. Primer- dimers and non-specific products are only likely to be a problem if visible on the agarose gel or with a Bioanalyzer. However, nicked copies of the vector are often below the limit of detection, yet can still generate significant unwanted background (data not shown). Vector background colonies are generated when the vector plasmid templates are amplified rather than just the insert, or desired vector-backbone product (Figure 3.3). This potential problem is most relevant to PIPE, due to a relatively low number of transformants. The simplest effective method to reduce this background for PIPE is to use less template. If that fails or if greater robustness is desired, a restriction enzyme can be used to first cut the template outside of the primer binding sites. Increasing the gap between the primers was also beneficial for our pUC18/Kan reporter, likely due to the difficulty of amplifying larger products in their entirety. We have previously observed higher efficiencies when replacing genes rather than PIPE cloning into an empty vector (31). Using a vector containing an existing insert of 59 distinguishable size can be particularly beneficial if it includes the lethal ccdB gene to select against vector background (91). Non-specific PCR products and primer-dimers can be a significant problem. Primer-dimers had little effect on PIPE or SLIC, likely due to their small size, such that they would lack the 5’-tails required for single-strand annealing and recombination. However, even low concentrations of primer- dimer are of concern for OEC, since cloning is extremely efficient for very small inserts, which clone preferentially. Primer-dimers could be removed by gel extraction or through pre-incubation with T4 polymerase to digest primer- dimer. Larger non-specific products will also compete with the desired insert in PIPE and SLIC, reducing cloning efficiency, which may require gel extraction. In our experience, these non-specific products can also be reduced through design of longer 3’ ends, increasing annealing temperatures, and reducing primer concentration. To avoid the requirement for gel extraction or PCR optimisation, alternative primer design can be used for Quick and Clean Cloning (106), a more specific variant of SLIC. This involves designing longer primers for one of the cloning junctions where the vector overhang is complementary to the region within the target insert PCR product, rather than insert primer. The resulting fragments can still be joined because short (~20 bp) ends of non-complementary sequence adjacent to the desired sequence are removed and successful recombination can occur in vivo, as long as the internal complementary sequence is single-stranded (92,106). Ligation-independent cloning approaches constitute an essential part of the biomedical researcher's molecular-tool kit. With the extremely high fidelity of modern polymerases and availability of modular vector-specific primers, we find that ligation-independent methods are preferable even for simple subcloning projects. Due to their robustness, speed and low cost, they may largely supplant restriction enzyme and ligation-dependent cloning in many laboratories.

Chapter 4

Cholesterol-dependent degradation of squalene monooxygenase, a novel control point in cholesterol synthesis beyond HMG-CoA reductase

Julian Stevenson performed the DNA cloning, site-directed mutagenesis, membrane topology experiments and cholesterol assays. Julian Stevenson, Saloni Gill and Ika Kristiana conducted the pulldown/immunoprecipitation and immunoblotting experiments. Saloni Gill and Ika Kristiana performed the thin layer chromatography and pulse-chase labelling experiments. Saloni Gill performed quantitative real-time PCR.

The data in this chapter have previously been presented or referred to in the PhD thesis of co-author Saloni Gill (110).

This work has been published in:

4 Cholesterol-dependent degradation of squalene monooxygenase, a novel control point in cholesterol synthesis beyond HMG-CoA reductase

4.1 Introduction

Cholesterol is a vital lipid in animals, but also can be toxic in excess. Consequently, elaborate homeostatic mechanisms have evolved, with exquisite control of cholesterol levels occurring at multiple levels within the cell (30). Research into the regulation of cholesterol synthesis has centered on 3-hydroxy- 3-methylglutaryl-coenzyme A reductase (HMGCR). In contrast, relatively little is known about the regulatory role and control of other enzymes in the pathway, although most are known SREBP-2 target genes (111). One such enzyme is squalene monooxygenase (SM). As an SREBP-2 target, SM expression is modulated by sterols, increasing under lipid-depleted conditions (82,84). Although not widely appreciated, SM has been proposed to be a second rate-limiting enzyme in cholesterol synthesis (82,112). The precursor squalene accumulates when Chinese hamster ovary (CHO) cells (83), human fibroblasts (113), rat hepatoma cells and renal carcinoma cells (112) are incubated with radiolabelled mevalonate and exogenous sterols. Similar accumulation has also been observed using rat and dog kidney slices (114). Furthermore, the concept that SM may be a largely overlooked control point in cholesterol synthesis is suggested by its much lower specific activity in liver cells compared to that of HMGCR (82). These data suggest that SM is under post-transcriptional control, in addition to transcriptional regulation by SREBP-2. In this chapter, we aim to investigate if SM is regulated by cholesterol at the post-transcriptional level, including testing the potential control mechanism, and attempt to identify the relevant parts of the protein involved in this regulation.

63 4.2 Materials and methods

4.2.1 Construction of expression plasmids PCRs were performed using Phusion polymerase, with verification by sequencing. Plasmids constructed in this study are listed in Table 4.1. Primer sequences for PIPE/OEC cloning and site-directed mutagenesis (SDM) are included in Table 4.2.

Table 4.1. Plasmids prepared in this study. Plasmid Name Description pTK-SM-V5 Moderate expression of SM pTK-EV Empty moderate expression vector pCMV-EV Empty high expression vector pTK-SMΔ(W -K )-V5 2 100 SM lacking the first 100 amino acids pCMV-SMΔ(W2-K100)-V5 pTK-SM-V5-GAr SM with proteasome inhibitory GA pTK-GAr-SM-V5-GAr repeats pTK-SM-N100-GFP-V5 First 100 amino acids of SM fused to pTK-SM-N100-GST-V5 GFP or GST pTK-Ub-WT-SM-N100-GFP-V5 N100-GFP with an automatically pTK- Ub-K48R-SM-N100-GFP-V5 cleaved ubiquitin protein pTK-SM-K15,16,157,268,293R-V5 pTK-SM-K318,399,400,429R-V5 Full-length lysine mutants pTK-SM-K436,496,536,570R-V5 pTK-SM-K15,16,82,90,100R-V5 pTK-SM-N100-K15,16,82,90,100R- N100-GST lysine mutant GST-V5 pTK-SM-N100-Y12,14S-GFP-V5 N100-GFP motif point mutants pTK-SM-N100-Y44S-GFP-V5

Table 4.2. Primer sequences for cloning and SDM. Name Primer Sequence (5' → 3') topoSQEF ATGTGGACTTTTCTGGGCATTGC topoSQER ATGAACCATATACTTCATTTCTGAG pcTOPO TK F GCTTAGGGTTAGGCGTTTTGCGCTGCTTCGGCAGCTGCTT CATCCCCGTGAC pcTOPO TK R GCTTGGGTCTCCCTATAGTGAGTCGTATTATACAATTCCG CAGCTTTTAGAGC SM Ntrm KO F CCAGTGTGGTGGAATTGCCCTTATGGGAACCAATATTTCA GAAACAAGC pcV155 R TCTTCATGCAATTGTCGGTC pcDNA3 MCS C F AAGGGCAATTCTGCAGATATCCAGCACAGTGG pcTOPO EV R CTGCAGAATTGCCCTTAAGGGCAATTCCACCACACTGGAC GAr oligo F GCTGGAGCAGGCGGTGGAGCAGGTGCTGGAGGTGCAGG TGGAGCAGGCGGTGCAGGAGCA GAr oligo R ACCTGCTCCACCTCCAGCACCTGCACCACCTGCTCCTGCA CCGCCTGCTCCACCTGCACC SM N GAr F CCAGTGTGGTGGAATTGCCCTTATGGCTGGAGCAGGCGG TGGAGC SM N GAr R GAAAGTGGCAATGCCCAGAAAAGTCCAACCTGCTCCACCT CCAGCAC pcC GAr F CGTACCGGTCATCATCACCATCACCATGCTGGAGCAGGC GGTGGAGC pcC GAr R GAGGCTGATCAGCGGGTTTAAACTCAACCTGCTCCACCTC CAGCAC pcGFP F GCCGCTCGAGTCTAGAGGGCCCGCGGTTCGAAATGGTGA GCAAGGGCGAGGAG pcGFP R CGAGACCGAGGAGAGGGTTAGGGATAGGCTTACCCTTGT ACAGCTCGTCCATGCC Cd GFP F GCCGGCAGCGGCGCCGTGAGCAAGGGCGAGGAGC SM d476 GFP R GGCGCCGCTGCCGGCTTTTCTGCGCCTCCTGGCCTC C GST F GCCGGCAGCGGCGCCTCCCCTATACTAGGTTATTGGAAA ATT C GST R GTTAGGGATAGGCTTACCGTCACGATGCGGCCGCTCG C V5 F GGTAAGCCTATCCCTAACCCTCTCCTCGGTCTCG SM N Ub F GTGGTGGAATTGCCCTTATGGGCTACCCCTATGATGTG

SM N Ub R CCCAGAAAAGTCCACATACCACCTCTGAGACGGAGGACC SM M1 F ATGTGGACTTTTCTGGGCATTGC pcDNA3 MCS n R AAGGGCAATTCCACCACACTGG SM 2K15R F CTATTTTTATAGGAGGTTCGGGGACTTC SM K82R R GGGGATCTGGCCCAGAAGAAG SM K90R F CAGAAAATAGGGAGCAGCTC SM K100R R TATTGGTTCCTCTTCTGCGCC SM d476 K0 GFP R GGCGCCGCTGCCGGCCCGTCTGCGCCTCCTGGCCTC Table 4.2 continues on the next page.

65 Table 4.2 continued. Name Primer Sequence (5' → 3') SM K157R R CTGTCAGGCTCCCTTAAGTCTCTC SM K268R F GGGAGTTCAGTACAAGGATAGGGAGACTGGAGATATCAAGG SM K293R R GGAGACCAGGCTCCTCCTGAACTTGG SM K318R F CTTTCTTATGAAGAATGCACCACAGTTTAGGGCAAATCATGC TGAAC SM 2K400R R GAAGAACACCTCGCCTCCTCACTGATGAAGG SM K429R R CAGTTTTCTCCATAGCCTTATATCTTTAAAAGC SM K436R F GAAAACTGCTAAGGGGTATCCCTGACC SM K496R R CATTCGCCACCAAGCCTGAAATAAAGAAAAC SM K536R F GTATTTTTGCTTTAGGTCAGAACCTTGG SM K570R R GAACCATATACCTCATTTCTGAG SM L5A F CCTTATGTGGACTTTTGCCGGCATTGCC SM SFS F GCATTGCCACTTTCACCAGCTTTAGCAAGAAGTTCGGGGAC TTCAT SM Y44S F CTCGCTGGGCCTGGTGCTCTCCAGCCGCTGTCGCCACC pCMV-SM-V5 contains the protein coding sequence of human squalene monooxygenase (identical to NM_003129.3 gi.62865634, 927-2651, NP_003120) with a C-terminal V5 epitope and His tag. It was previously prepared by Dr Jenny Wong by TA cloning the human gene into pcDNA3.1-V5-His TOPO vector (Invitrogen) using the primers topoSQEF and topoSQER for PCR. pTK- SM-V5 is identical to pCMV-SM-V5, but with expression driven by the thymidine kinase promoter. It was prepared through amplification of the promoter and splicing region from pTK-beta (Invitrogen) with primers pcTOPO TK F, pcTOPO TK R, using the resulting megaprimer for OEC SDM of pCMV- SM-V5 to replace the cytomegalovirus enhancer/promoter. Empty vectors were generated using PIPE with pcDNA3 MCS C F and pcTOPO EV R.

SM∆(W2-K100)-V5 consists of an initiating methionine followed by amino acids 101-574 of human SM, a multiple cloning site and V5-His tag.

Corresponding deletions, pTK-SM∆(W2-K100)-V5 and pCMV-SM∆(W2-K100)-V5 were prepared from the respective vectors above with PCR and SDM (115) with SM Ntrm KO F and pcV155 R. Substitution point mutations were similarly prepared with pcV155 R, SM L5A F, SM SFS F (Y12,14S) or SM Y44S F. pTK-SM-V5-GAr contains a 30 amino-acid repeat from Epstein-Barr virus nuclear antigen-1 after the V5-His tag. pTK-GAr-SM-V5-GAr contains an additional copy of the repeat after the initiating methionine. These were constructed using PCR of the repeat and OEC SDM of pTK-SM-V5. The oligos

GAr oligo F and GAr oligo R were annealed and extended, then amplified to generate mutagenic megaprimers targeting the N- or C-termini using primers SM N GAr F and SM N GAr R or pcC GAr F and pcC GAr R, respectively. pTK-SM-N100-GFP-V5 encodes the first 100 amino acids of human followed by a linker ‘AGSGA’, the enhanced green fluorescent protein and the V5-His tag. It was prepared through OEC SDM of pTK-SM-V5 with a GFP megaprimer derived from pEGFP-N1 (Clontech) with the primers pcGFP F and pcGFP R, followed by a deletion using PIPE SDM with the primers Cd GFP F and SM d476 GFP R. pTK-SM-N100-GST-V5 instead includes a glutathione S-transferase (GST) fusion, amplified from pGEX-4T-1 (GE Healthcare) with PIPE insert primers C GST F and C GST R, combined with the TK vector product from primers C V5 F and SM d476 GFP R. Wild-type or mutant ubiquitin was fused to the N-terminus of SM-N100- GFP-V5 with amplification of the insert with SM N Ub F and SM N Ub R from pRK5-HA-Ubiquitin-WT or -K48R (116) (Ted Dawson, Johns Hopkins University School of Medicine, Addgene Plasmids 17608 and 17604 respectively), and the vector with SM M1 F and pcDNA3 MCS n R, with PIPE cloning yielding pTK-Ub-WT-SM-N100-GFP-V5 or pTK-Ub-K48R-SM-N100- GFP-V5, respectively.

4.2.2 Preparation of lysine mutants by multiple SDM Multiple SDM was achieved using a novel, unpublished method. It involves preparing multiple phosphorylated PCR products containing mutations at the ends. These are used as megaprimers to amplify the target plasmid with a non- strand displacing polymerase, with repair of the nicks by a thermostable DNA ligase. Pairs of primers (SM 2K15R F, SM K82R R, K90R F, SM K100R R, SM d476 K0 GFP R, SM K157R R, SM K268R F, SM K293R R, SM K318R F, SM 2K400R R, SM K429R R, SM K436R F, SM K496R R, SM K536R F, SM K570R R, and flanking vector primers if necessary, Table 4.2) containing mutations in the centre were phosphorylated in a 10 L reaction with 1X Fermentas T4 DNA ligase buffer, 10 U of NEB polynucleotide kinase (PNK) and 3 M of each primer, incubated at 37 °C for 30 min. PCR was performed in 25 L reactions using 0.5 ng of template, 0.5 M forward and reverse phosphorylated primers, 5% DMSO, HF Buffer and 0.5 U

67 of Phusion Hot Start II High Fidelity DNA polymerase, and cycling conditions: 98 °C 3’ (98 °C 30'' 53 °C 1' 72 °C x') x 35, 72 °C 5'. Pooled phosphorylated PCR products were purified using the Qiagen PCR purification kit according to manufacturer’s instructions, with an elution volume of 30 L. SDM was performed using 0.5 L of megaprimer products, 25 ng of template, 5% DMSO, 0.9 x HF buffer, 0.1 x NEB Taq ligase buffer, 20 U of Taq ligase and 0.5 U Phusion Hot Start II in a 25 uL reaction with cycling conditions: 95 °C 2' (95 °C 1' 53 °C 1' 65 °C 4') x 35, with the products used similarly to OEC.

4.2.3 Ubiquitination of human SM Following transfection and treatment (as indicated in the figure legend), CHO-7 cells were lysed in modified RIPA buffer (50 mM NaCl, 1.0% IGEPAL CA-630, 0.5% sodium deoxycholate, 0.1% SDS, 200 mM Tris, pH 8.0) supplemented with N-ethylmaleimide (10 mM), protease inhibitor cocktail and ALLN (25 g/ml). Lysates with equal cell protein were immunoprecipitated with monoclonal anti- V5-conjugated Dynabeads (Invitrogen), according to the manufacturer’s instructions. Pull-down of the N100-GST fusion protein from lysates with equal cell protein was achieved using glutathione sepharose beads. Following 4 washes with modified RIPA buffer, beads were resuspended in 50 L of ‘loading buffer’ (2 vol RIPA: 2 vol 10% SDS: 1 vol 5x Laemmli buffer). Pellets were subjected to 7.5% SDS-PAGE, followed by immunoblot analysis with anti- V5 (for SM) and anti-HA (for ubiquitin) antibodies.

4.2.4 Metabolic labelling of N100-GST with [35S]-methionine/cysteine CHO-7 cells were transiently transfected as indicated in the figure legend. After statin pretreatment, cells were labelled in methionine-free medium (Invitrogen) supplemented with 5 mM compactin and 50 mM mevalonate containing 250 mCi/ml [35S]-Protein Labeling Mix (Perkin Elmer) for 3 hr, then washed and chased in medium B containing 2 mM methionine and cysteine with or without Chol/CD, for 0-8 hr. [35S]-labelled N100-GST was pulled down from lysates with equal cell protein using glutathione sepharose beads, and pellets were subjected to 4-20% or 10% SDS-PAGE. Bands were visualised by phosphorimaging, and their relative intensities were quantified using

Sciencelab ImageGauge 4.0 Software (Fujifilm). We would like to thank Dr Maaike Kockx for providing access to and assistance with [35S] facilities.

4.2.5 Cell fractionation Fractions were prepared according to (117), with minor modifications. CHO-7 cells in 10 cm dishes were grown in medium A (without antibiotic) and transfected with 5 µg of DNA. Cells were harvested after 24 hr by scraping into ice-cold PBS, washed, resuspended in Buffer A (10 mM HEPES-KOH pH 7.4, 10 mM KCl, 1.5 mM MgCl2, 100 mM NaCl, 5 mM Na EDTA, 5 mM Na EGTA, and 250 mM sucrose), and passed through an 18 G needle 50 times. The lysate of equalized protein content was centrifuged at 1,000 x g for 5 min, 4°C, and the post-nuclear supernatant centrifuged at 100,000 x g for 30 min, 4°C, with resuspension of the resulting membrane pellet in an equal volume of the same buffer. The 1,000 x g pellet was resuspended in Buffer B (20 mM Hepes-KOH

(pH 7.6), 25% (v/v) glycerol, 0.42 M NaCl, 1.5 mM MgCl2, 5 mM sodium EDTA, and 5 mM sodium EGTA), rotated at 4°C for 1 hr, and similarly centrifuged at 100,000 x g for 30 min at 4°C, with the supernatant yielding the nuclear fraction. Fractions of equal volume were analysed with SDS-PAGE and immunoblotting.

69 4.3 Results

4.3.1 Cholesterol treatment causes squalene to accumulate, suggesting rate-limiting activity of SM A key observation that highlights the rate-limiting activity of SM is squalene accumulation when cholesterol levels are high (83,112,113). These experiments were conducted using radiolabelled mevalonate, consequently bypassing HMGCR. We labelled CHO-7 cells for 4 hr with [14C]-acetate which feeds into the beginning of the cholesterol biosynthetic pathway. CHO-7 cells were chosen because they can be maintained in lipoprotein-deficient serum (LPDS) (118) which offers considerable flexibility in manipulating cholesterol levels. Importantly, we found that treatment of CHO-7 cells with sterols and [14C]-acetate also led to the accumulation of [14C]-squalene (Figure 4.1). This band was absent upon inhibition of squalene synthase (lane 1), and accumulated when SM was inhibited (lane 2). The most striking squalene accumulation resulted from addition of cholesterol complexed with methyl-β- cyclodextrin (Chol/CD, lane 5), followed by low-density lipoprotein (LDL) (lane 4), and then the oxysterol 25-hydroxycholesterol (25HC) (lane 6). Thus in this system, cholesterol treatment induces squalene accumulation, raising the possibility of a rate-limiting step after HMGCR involving SM.

Figure 4.1. Cholesterol treatment causes squalene to accumulate. CHO-7 cells were statin pretreated overnight, then treated and labelled with [14C]-acetate in medium A with the following test agents for 4 hr: squalene synthase inhibitor (SSi, 150 M), SM inhibitor (SMi, 10 M), LDL (50 g/ml), cholesterol complexed with methyl-β-cyclodextrin (Chol/CD, 20 g cholesterol/ml), or 25-hydroxycholesterol (25HC, 1 g/ml), and assayed for [14C]-cholesterol and [14C]-squalene accumulation. [14C]-Squalene accumulation was expressed relative to the maximal condition (Chol/CD), which was set to 1 (n=5).

71 4.3.2 Cholesterol-dependent squalene accumulation occurs in a variety of cell types The effect was not restricted to CHO-7 hamster cells (Figure 4.2): a cholesterol- dependent increase in squalene accumulation was also seen in human cell-lines of hepatic (HepG2), neuronal (BE(2)C), and renal origin (HEK293), as well as in primary human fibroblasts (Fb). Hence, it is likely to occur in all cell types.

Figure 4.2. Squalene accumulates in a variety of human cell types. Cells in medium E (HepG2) or medium H (BE(2)C, HEK293, and fibroblasts (Fb)) were statin pretreated overnight, labelled with [14C]-acetate, treated with or without Chol/CD (20 g/ml) for 4 hr, and assayed for [14C]-cholesterol and [14C]-squalene accumulation (n=2).

4.3.3 Analysis of squalene accumulation and the cholesterol- dependent reduction in message levels suggests post- transcriptional regulation of SM. When examined over time, cholesterol treatment led to the progressive accumulation of squalene and decreased de novo cholesterol synthesis (Figure 4.3A), most notably after 4 hr (lane 5), but also as early as 2 hr (lane 3). In the absence of exogenous cholesterol, the levels of newly synthesized cholesterol were relatively constant, with no observed squalene. These results suggest that SM becomes rate-limiting due to an acute cholesterol-dependent regulatory mechanism. To uncover this mechanism, we first examined transcriptional regulation using quantitative (real-time) PCR, and compared the mRNA levels of SQLE (SM) to that of HMGCR (HMGCR). In the absence of added cholesterol, the mRNA levels of both genes were constant and remained unaffected over 16 hr

(Figure 4.3B). The addition of cholesterol caused a decrease in expression levels of both genes with similar kinetics and magnitude. Squalene rapidly accumulated with message levels remaining high and in spite of HMGCR being a post-translational regulated rate-limiting enzyme upstream of SM. The dissociation of mRNA levels from flux through the pathway similarly suggests post-transcriptional regulation of SM.

Figure 4.3. Squalene accumulates whilst message levels of SM and HMGCR remain high. (A) Overnight statin pretreated CHO-7 cells were treated in medium B with or without Chol/CD (20 g/ml) as indicated and labelled with [14C]-acetate in medium A during treatment 2 hr prior to harvesting. Relative accumulation was calculated so that [14C]-squalene + [14C]-cholesterol = 1 at each time point. Error bars (±SEM) are contained within the symbols (n=6). (B) CHO-7 cells were treated as in (A) without radiolabelling and mRNA levels for hamster SM and HMGCR determined by quantitative real-time PCR (n=3, each performed in triplicate).

4.3.4 Squalene accumulates in SRD-1 cells upon cholesterol treatment, consistent with post-transcriptional regulation of SM To exclude transcriptional regulation, we utilised SRD-1 cells, a mutant line of CHO cells which exhibit sterol-independent expression of SREBP-2 target genes (119). These cells overexpress the N-terminal transcription factor fragment of SREBP-2, bypassing the sterol-regulated proteolytic step, leading to constant transcriptional activation regardless of sterol levels. Thus, mRNA levels of HMGCR and SQLE were similarly unaffected by cholesterol addition (Figure 4.4A). Nevertheless, cholesterol-dependent squalene accumulation was still evident in SRD-1 cells (Figure 4.4B). This is consistent with SM being post- transcriptionally regulated by cholesterol, which may impact on the control of cholesterol synthesis.

Figure 4.4. Cholesterol treatment causes squalene to accumulate in SRD-1 cells, independent of transcriptional regulation. (A) Overnight statin pretreated SRD-1 cells were treated in medium B with or without Chol/CD (20 g/ml) as indicated and mRNA levels for hamster SM and HMGCR determined by quantitative real-time PCR (n=4, each performed in triplicate). (B) SRD-1 cells were treated as in (A) and labelled with [14C]- acetate in medium A during treatment 2 hr prior to harvesting. Relative accumulation was calculated so that [14C]-squalene + [14C]-cholesterol = 1 at each time point (n=3, each performed in triplicate).

75 4.3.5 Cholesterol-dependent degradation of SM One explanation for the accumulation of squalene is that cholesterol also acts post-translationally, accelerating the degradation of SM. We investigated this possibility in SRD-1 cells (Figure 4.5). Chol/CD caused a reduction in endogenous SM protein from 4 hr, down to negligible levels by 8 hr (lane 7). This effect was not observed when cells were treated with methyl-β-cyclodextrin (CD) without cholesterol (data not shown). Protein synthesis was inhibited with cycloheximide (CHX) during cholesterol treatment, showing that the regulation was post-translational, through degradation of SM, and did not require nascent gene expression. This cholesterol-mediated degradation of SM was on a comparably acute timescale to squalene accumulation (cf. Figure 2B). Overall, addition of cholesterol increased SM turnover several-fold, with the estimated half-life decreasing from ~14 hr to ~4 hr.

Figure 4.5. SM is regulated at the post-translational level by cholesterol- dependent degradation. SRD-1 cells were statin pretreated overnight and treated as indicated in medium A containing cycloheximide (10 g/ml) with or without Chol/CD (20 g/ml) for up to 16 hr. Cell lysates were assayed for endogenous SM by immunoblotting (n=3). A second band (~50 kDa) sometimes evident in the immunoblots of the endogenous protein, likely reflects partial degradation (66), but not an intermediate in sterol regulated degradation, since it is unaffected by cholesterol treatment.

Consistent with our results for endogenous hamster protein, there was a cholesterol-dependent reduction in human SM (with expression in CHO-7 cells driven by a thymidine kinase (TK) promoter) (Figure 4.6, lanes 3-4). In contrast, a construct with a higher expression cytomegalovirus (CMV) promoter did not show clear sterol regulation (lanes 7-8).

Figure 4.6. Cholesterol-dependent degradation of ectopic SM is blunted with higher over-expression. CHO-7 cells were transfected with 1 g of empty vector (EV) and plasmids encoding human SM with expression driven by the thymidine kinase (TK) or stronger cytomegalovirus (CMV) promoter, as indicated. Following statin pretreatment overnight, cells were treated for 8 hr in medium B containing cycloheximide (10 g/ml) with or without Chol/CD (20 g/ml). SM-V5 protein was analysed by immunoblotting (n=3).

4.3.6 Accelerated degradation occurs when cholesterol exceeds basal levels To investigate the level of cholesterol required to mediate degradation, we manipulated cellular cholesterol by adding defined amounts in CD complexes and examined protein levels. The inverse curvilinear relationship between cholesterol and SM protein levels suggests the presence of a threshold required to trigger rapid degradation, approximately 15-20 g total cholesterol/mg total protein in CHO-7 cells (Figure 4.7A). This threshold is close to the basal cholesterol value of ~15 g/mg protein observed in full serum without statin pretreatment (Figure 4.7B), consistent with a physiological feedback role for SM degradation in control of cholesterol synthesis.

Figure 4.7. Degradation is accelerated as cholesterol levels rise above a basal threshold. (A) CHO-7 cells were transfected with 1 g of pTK-SM-V5, statin pretreated overnight and treated with varying concentrations of Chol/CD (0-20 g/ml) for 8 hr. SM-V5 protein was analysed by immunoblotting (n=8). Total cellular cholesterol levels were measured in parallel experiments (n=3). The error bars (±SEM) presented are sometimes contained within the symbols. (B) Basal cholesterol levels for CHO-7 cells were measured with or without overnight statin pretreatment and 8 hr treatment in DF12 media supplemented with lipoprotein deficient serum (LPDS) or newborn calf serum (NCS) (n=3). 79 4.3.7 Cholesterol-dependent degradation of SM is mediated by the ubiquitin-proteasome system Since SM is associated with the endoplasmic reticulum (ER), and the fate of many ER-bound proteins is destruction by the proteasome (120), it is likely that the cholesterol-dependent degradation of SM is proteasomal. In support of this, a range of proteasomal inhibitors (ALLN, MG132, and lactacystin) preserved endogenous SM protein levels (Figure 4.8A & data not shown). On the contrary, a lysosomal inhibitor had no effect (lanes 6 vs 9). As observed with endogenous hamster SM, the cholesterol-dependent reduction in ectopic human enzyme was rescued by the addition of MG132 (Figure 4.8B, lanes 2 vs 4).

Figure 4.8. Cholesterol-Dependent degradation is reversed by proteasomal but not lysosomal inhibition. (A) CHO-7 cells were statin pretreated overnight and treated in medium B containing cycloheximide (10 g/ml), Chol/CD (20 g/ml) and MG132 (10 M), ALLN (20 g/ml) and chloroquine (200 M), as indicated for 4 hr (n=2). Lysates were assayed for endogenous SM by immunoblotting (n=3). (B) CHO-7 cells were transfected with 1 g of plasmid as indicated, statin pretreated overnight and treated in medium B containing cycloheximide (10 g/ml), Chol/CD (20 g/ml) and MG132 (10 M), as indicated for 8 hr (n=4 each). Cell lysates were assayed for SM-V5 by immunoblotting (n=2).

Alternatively, glycine-alanine repeats (GAr) from Epstein-Barr virus nuclear antigen-1 were added to the ends of the human protein to provide resistance to proteasomal degradation (121). This allows specific targeting of SM, avoiding indirect effects of systemic proteasome inhibition. A single 30 amino-acid repeat at the C-terminus slightly increased protein levels (Figure 4.9, lanes 3 vs 1), whereas repeats on both the N- and C-termini blunted cholesterol-dependent

81 degradation (lanes 5-6). This is consistent with the hypothesis that SM undergoes proteasomal degradation.

Figure 4.9. Addition of Proteasome-inhibitory GA-repeats blunt degradation of SM. CHO-7 cells were transfected with 1 g of plasmid as indicated, statin pretreated overnight and treated for 8 hr in medium B containing cycloheximide (10 g/ml), Chol/CD (20 g/ml) and MG132 (10 M), as indicated. The SM constructs contain glycine-alanine repeats at the C- and/or N-termini of wild-type (WT) SM-V5 as indicated. Cell lysates were assayed for SM-V5 by immunoblotting (n=2).

Furthermore, in co-expression experiments with HA-tagged ubiquitin, human SM was polyubiquitinated in the presence of MG132, which was greater in the presence of cholesterol (Figure 4.10, lanes 4 vs 3). Thus, SM is degraded by the ubiquitin-proteasome system, with cholesterol stimulating ubiquitination of the protein, leading to its rapid degradation.

Figure 4.10. Cholesterol treatment increases polyubiquitination of SM. CHO-7 cells (in 6 cm dishes) were transfected with 1.5 g of pTK-SM-V5 and 0.5 g of pMT123 (pUb-HA, HA-tagged ubiquitin), statin pretreated, treated in medium B with or without Chol/CD (20 g/ml) and MG132 (10 M) as indicated for 4 hr. The SM-V5 was immunoprecipitated from cell lysates. Immunoprecipitated pellets were assayed for SM-V5 or HA-ubiquitin by immunoblotting (n=2).

83 4.3.8 Proteasomal degradation affects flux through the mevalonate pathway at SM To establish if proteasomal degradation is the mechanism that explains cholesterol-dependent squalene accumulation, we observed flux through the pathway at SM in the presence of cholesterol and MG132. Importantly, in SRD-1 cells, proteasomal inhibition dramatically reduced the accumulation of squalene after 4 hr of cholesterol treatment and increased cholesterol synthesis from acetate (Figure 4.11A, lane 4 vs 3). The same effect was also observed using radiolabelled mevalonate in both SRD-1 and CHO-7 cells (Figure 4.11B & C, respectively), thus bypassing HMGCR. This is consistent with the cholesterol- dependent proteasomal degradation of SM having functional consequences, by contributing to the observed squalene accumulation.

Figure 4.11. Proteasomal degradation affects flux through the mevalonate pathway at SM. Statin pretreated SRD-1 (A, C) or CHO-7 (B) cells were treated in medium A with or without Chol/CD (20 g/ml) and/or MG132 (10 M), and labelled with [14C]-acetate (A) or [14C]-mevalonate (B, C) for 4 hr. [14C]-Squalene accumulation was expressed relative to the maximal condition (Chol/CD), which was set to 1 (n=4 (A); 5 (B); or 3 (C)).

85 4.3.9 Insig and Scap are not required for the cholesterol-dependent degradation of SM Sterol-dependent degradation of HMGCR requires its sterol-sensing domain, which binds to the Insig retention protein, which in turn brings a ubiquitin ligase into contact with HMGCR (51). To determine whether Insig is required for degradation of SM, we observed lipid synthesis and SM turnover in SRD-15 cells. These mutant CHO cells are deficient in Insig, with no functional Insig-1 isoform, and an extremely low level of Insig-2 (122). When SRD-15 cells were treated with cholesterol, squalene still accumulated (Figure 4.12A) and protein degradation was still observed for both endogenous SM (Figure 4.12B, lane 2) and ectopic human SM (Figure 4.12C, left panel, lane 4). Cholesterol-dependent degradation of SM also occurred in SRD-13A cells (123), mutant CHO cells which lack the cholesterol-sensing protein Scap and consequently do not have a functional SREBP pathway (Figure 4.12C, right panel, lane 2). Thus, Insig and Scap are not required for degradation of SM. Furthermore, if additional regulatory machinery is necessary, then the genes do not appear to be strict SREBP-2 targets.

Figure 4.12. Insig and Scap are not required for accelerated degradation of SM. (A) Statin pretreated SRD-15 cell, deficient in Insig, were treated in medium B with or without Chol/CD (20 g/ml) as indicated and labelled with [14C]-acetate in medium A during treatment 2 hr prior to harvesting. Relative accumulation was calculated so that [14C]-squalene + [14C]-cholesterol = 1 at each time point. Error bars (±SEM) are contained within the symbols (n=5). (B) SRD-15 cells were statin pretreated overnight and treated in medium B containing cycloheximide (10 g/ml), Chol/CD (20 g/ml) and MG132 (10 M), as indicated. Cell lysates were assayed for endogenous SM by immunoblotting (n=4). (C) SRD-13A cells are deficient in Scap. The indicated cell-lines were transfected with 1 g of pTK-SM-V5, statin pretreated overnight and treated in medium B containing cycloheximide (10 g/ml) with or without Chol/CD (20 g/ml) for 8 hr. SM-V5 protein was analysed by immunoblotting (each n=2).

87 4.3.10 Cholesterol-dependent proteasomal degradation of SM requires its N-terminal domain The N-terminus of SM, encoded by the first exon, is partially conserved in vertebrates but lacking in lower organisms (Figure 4.13). In addition, a recombinant truncated rat enzyme missing the first 99 residues retains full activity (66). This led us to suspect that the vertebrate N-terminal region is a structurally and functionally distinct domain that may play a role in post- transcriptional regulation.

Figure 4.13. Multiple sequence alignment of SM protein sequences reveals the presence of a conserved N-terminal region only found in vertebrates. Multiple sequence alignment of SM protein for selected species, constructed using ClustalW2. Alignment of human (Homo sapiens, NP_003120), rat (Rattus norvegicus, P52020), finch (Taeniopygia guttata, XP_002187271), zebra fish (Danio rerio, NP_001103509), lancelet (Branchiostoma floridae, XP_002594656), sea urchin (Strongylocentrotus purpuratus, XP_001199544), slime mould (Dictyostelium discoideum, XP_629022) and yeast (Saccharomyces cerevisiae, P32476).

We constructed a corresponding human version that is missing the N-terminal region (Δ(W2-K100); Figure 4.14A), and examined its response to cholesterol treatment (Figure 4.14B). After cell fractionation, the truncated enzyme remains associated with the membrane fraction (Figure 4.19, lane 5), consistent with wild-type localization. However, unlike full-length SM (WT, Figure 4.14B, lane 2 vs 1), turnover of this deletion construct was unaffected by cholesterol (lane 5 vs 4), and levels were further increased by addition of MG132 (lane 6 vs 4).

Figure 4.14. The first 100 amino acids of SM are required to mediate cholesterol-dependent degradation. (A) Schematic of wild-type full-length SM and deletion constructs. (B) CHO-7 cells were transfected with 1 g pTK-SM-V5 (WT) or pTK-SM-∆(W2-K100)-V5 [∆(W2-K100)] as indicated (n=2). Following overnight statin pretreatment, cells were treated in medium B containing cycloheximide (10 g/ml), with or without Chol/CD (20 g/ml), and/or MG132 (10 M) for 8 hr. Cell lysates were assayed for V5-tagged constructs by immunoblotting.

We then determined if the lack of cholesterol-regulation of the deletion construct has functional consequences, by acutely labelling transfected cells with [14C]-acetate. Importantly, we observed that MOS, the product of SM, accumulates preferentially in cholesterol-treated cells transfected with the N-terminal deletion construct compared to full-length SM (Figure 4.15, lanes 4 vs 2). Accordingly, densitometric analysis showed that the squalene precursor to MOS product ratio was 30-40% lower in the cholesterol- treated cells transfected with the truncated construct, in keeping with the failure of cholesterol to accelerate degradation of this variant.

Figure 4.15. Overexpression of SM lacking the first hundred amino acids increases conversion of squalene to 2,3-monoxidosqualene. Statin pretreated CHO-7 cells transfected the previous day with 1 g pTK-SM- 14 V5 (WT) or pTK-SM-∆(W2-K100)-V5 [∆(W2-K100)] were labelled with [ C]- acetate in medium A, and treated with or without Chol/CD (20 g/ml) for 4 hr. Cells were assayed for accumulation of [14C]-cholesterol, [14C]-squalene and [14C]-2,3-monooxidosqualene (MOS). The [14C]-MOS band is indicated with an arrow. The [14C]-squalene to [14C]-MOS ratio was expressed relative to the maximal condition (cells transfected with pTK-SM-V5 and treated with Chol/CD), which was set to 1. For the Chol/CD conditions, the ∆(W2-K100) construct produced a lower [14C]-squalene to [14C]-MOS ratio than the wild- type (WT) construct (p<0.05 by t-test; n=5).

We next prepared a complementary construct of the first 100 amino acids of epitope-tagged human SM, but it failed to express (data not shown). This may be due to the extremely hydrophobic character of this region, previously proposed to contain transmembrane domains (66,73). However, protease protection experiments conducted in our laboratory suggested that it does not

91 contain membrane spanning α-helices. For example, an N-terminal Myc or internal FLAG epitope tag within this sequence were degraded when protease- impermeable membrane vesicles were treated with trypsin (data not shown). To assist folding or solubilization and enable expression, we fused the first 100 amino acids (N100) to green fluorescent protein (N100-GFP) or glutathione S-transferase (N100-GST) under the control of the TK promoter (Figure 4.16A). Importantly, turnover of both heterologous fusion constructs was robustly regulated by cholesterol (Figure 4.16B, lane 4 vs 3; Figure 4.18A, lane 2 vs 1), whereas expression of either GFP or GST alone was unaffected by cholesterol treatment (data not shown). When the full-length protein and truncated chimera were co-transfected, cholesterol-dependent turnover of each was also unaffected (Figure 4.16B, lane 6 vs 5). This was surprising in light of the marked blunting of regulation seen for the CMV system earlier, because expression of the TK-driven truncated fusion protein (N100-GFP) was much higher than the full-length non-fusion protein (data not shown).

Figure 4.16. The N-terminal domain (N100) confers cholesterol-dependent degradation upon fusion proteins. (A) Schematic of fusion constructs, where the first 100 amino acids of SM (N100) is fused to GFP or GST. (B) CHO-7 cells were transfected with 0.5 g pTK-SM-V5 (WT) and/or 0.5 g pTK-SM-N100-GFP-V5 (N100-GFP) as indicated. Following overnight statin pretreatment, cells were treated in medium B containing cycloheximide (10 g/ml), with or without Chol/CD (20 g/ml), and/or MG132 (10 M) for 8 hr. Cell lysates were assayed for V5-tagged constructs by immunoblotting (n=3).

A pulse-chase approach permits the study of protein degradation without the use of cycloheximide. This confirmed that Chol/CD treatment accelerates degradation of the N100-GST (Figure 4.17), and suggests a comparable half-life (~3 hr) to that seen in the cycloheximide-containing studies on full-length endogenous SM (~4 hr in Figure 4.5). Without added cholesterol, the N100-GST appeared remarkably stable over the 8 hr. Cholesterol began to degrade this

93 truncated version of SM by 2 hr (Figure 4.17), in line with the timing observed for squalene accumulation (Figure 4.3 & Figure 4.4).

Figure 4.17. Pulse-chase analysis reveals rapid degradation of N100-GST. CHO-7 cells were transfected with 0.25 g of pTK-SM-N100-GST-V5 (N100- GST). Following statin pretreatment, cells were pulsed for 3 hr with [35S]-methionine/cysteine and then chased in medium B with or without Chol/CD (20 g/ml) for 0-8 hr. [35S]-N100-GST protein was pulled down with glutathionine sepharose, analysed by SDS-PAGE and the band visualised by phosphorimaging (n=2). For the image shown, densitometric values were plotted for the +Chol/CD conditions. The –Chol/CD values, which did not change over the 8 hr, were each set to 1. Right hand panel: Cholesterol significantly degraded N100-GST at 2 hr (p<0.05 by t-test; n=4).

Polyubiquitin chains recognised by the proteasome are made up of G76-K48 inter-ubiquitin linkages, so the K48R mutant causes premature chain termination, inhibiting degradation (124). Expression of TK-driven N100-GFP was abrogated when co-transfected with CMV-driven mutant ubiquitin (data not shown), possibly due to transcriptional squelching by the CMV promoter. Hence to avoid this, ubiquitin was delivered under the control of the TK promoter by fusing it to the N-terminus of SM. This approach is possible because ubiquitin fusions are efficiently processed to liberate free ubiquitin by deubiquitinating enzymes (98). Thus, the size of the immunoblotted N100-GFP did not shift and there was no new second band. Co-expression of a chain-

terminating K48R ubiquitin mutant through fusion to N100-GFP blunted cholesterol-dependent degradation and increased SM protein levels compared to wild-type ubiquitin (Figure 4.18A), again implicating the ubiquitin- proteasome system. Moreover, GST pull-down of the N100-GST chimera co-expressed with HA-tagged ubiquitin revealed clear cholesterol-dependent polyubiquitination (Figure 4.18B, lane 1 vs 2), which was observed as early as one hour after cholesterol treatment (Figure 4.18C, lane 1 vs 2).

Figure 4.18. N100-GST is targeted by the ubiquitin-proteasome system similarly to full-length SM and as early as 1 hr. (A) CHO-7 cells were transfected with 0.5 g pTK-Ub-WT-SM-N100-GFP-V5 (UbWT) or pTK-Ub-K48R-SM-N100-GFP-V5 (UbK48R), statin pretreated and treated in medium B containing cycloheximide (10 g/ml) with or without Chol/CD (20 g/ml) for 8 hr. V5-tagged construct protein was analysed by immunoblotting (n=2). The N-terminal ubiquitin is efficiently cleaved from the N100-GFP fusion upon expression. (B, C) CHO-7 cells (in 10 cm dishes) were transfected with 1.5 g of pTK-SM-N100-GST-V5 (N100-GST) and 0.5 g of pMT123 (pUb-HA, HA-tagged ubiquitin). Following statin pretreatment, cells were treated in medium B with or without Chol/CD (20 g/ml) and MG132 (10 M) for 4 hr (B) or 1 hr (C). N100-GST protein was pulled down with glutathionine sepharose and immunoblotted for V5 (N100-GST) and HA- ubiquitin (n=2).

Together, these data suggest that the N-terminal region forms a regulatory domain, which is necessary and sufficient for post-translational regulation of SM by the proteasome, and which in turn may help to regulate flux through the cholesterol biosynthetic pathway.

4.3.11 N100 is membrane associated Despite the apparent absence of membrane spanning α-helices, the chimeric N100 fusion protein was still membrane-associated (Figure 4.19, right panel, lane 2), similarly to the rest of the protein (left panel, lanes 2 & 5).

Figure 4.19. N100 is sufficient, but not necessary, for SM membrane localisation. CHO-7 cells were transfected as indicated and harvested for cell fractionation after 24 hr as described in the Materials and Methods. Cell lysates were assayed for V5-tagged constructs by immunoblotting (n=3). N, nuclear; M, membrane (100,000 x g pellet); C, cytosol (100,000 x g supernatant).

97 4.3.12 Regulated degradation can be mediated by multiple ubiquitination sites Our preliminary experiments using protease protection are consistent with the N-terminal domain being accessible to the cytoplasmic ubiquitin-proteasome system, albeit membrane bound. To identify the ubiquitin site or sites, we mutated groups of lysines to arginine (Figure 4.20) using a novel multiple SDM method, which performed well, but was not directly compared to commercial kits or previously published methods. We mutated selected conserved lysines, or all from the first 100 amino acids of the full-length protein (data not shown) and N100 region of the GST fusion protein (Figure 4.20), which all retained cholesterol-regulated turnover. These data suggest that a specific regulated ubiquitination site may be unnecessary, and/or that ubiquitination may also occur on the N-terminal amino group or a lysine on the GST or V5 tag.

Figure 4.20. Lysine substitution mutants retain cholesterol-dependent degradation. CHO-7 cells were transfected with (A) 1 g of pTK-SM-V5, pTK-SM- K15,16,157,268,293R-V5 (NK→R), pTK-SM-K318,399,400,429R-V5 (MK→R), pTK- SM-K436,496,536,570R-V5 (CK→R), or (B) 0.5 g of pTK-SM-N100-GST-V5 or pTK-SM-N100-K15,16,82,90,100R-GST-V5 (SM-N100-KO-GST-V5), as indicated. Following statin pretreatment, cells were treated in medium B containing cycloheximide (10 g/ml) with or without Chol/CD (20 g/ml) for 8 hr. SM- V5 protein was analysed by immunoblotting (n=2). SM-N100-KO-GST-V5 (KO) contains no lysines in the first 100 hundred amino acids, but others remain within GST and the V5 tag.

4.3.13 A putative CRAC motif in SM and other motifs similar to those found in Scap and HMGCR are not functionally significant The N100 region contains the short sequences 'YFY' and 'LGIA' in close proximity, near-identical to motifs found in transmembrane helices of the sterol-sensing domain of HMGCR and Scap, shown to be functionally important for Insig binding (105,125,126) The YIYF peptide has also been

99 suggested to recruit cholesterol on the basis of phase transition experiments (127). However, mutation of the tyrosines to serine or the leucine to alanine did not abolish regulation, although they appeared to destabilize the protein (Figure 4.21, lanes 3-4 of each panel). It is not known if regulated degradation requires direct binding of cholesterol to the N-terminus. Mutation of a single tyrosine from a further motif, fulfilling the cholesterol recognition/interaction amino acid consensus (CRAC) sequence (127), associated with cholesterol binding, also had no effect on regulation (Figure 4.21, left panel, lanes 5-6).

Figure 4.21. Mutation of motifs similar to Insig and Scap in N100 do not affect SM regulation. CHO-7 cells were transfected with 0.5 g of plasmid as indicated. Following overnight statin pretreatment, cells were treated in medium B containing cycloheximide (10 g/ml) with or without Chol/CD (20 g/ml) for 8 hr. Cell lysates were assayed for V5-tagged constructs by immunoblotting (n=2).

100

4.4 Discussion

Using model cell systems, we found that cholesterol caused the accumulation of the substrate squalene, suggesting that SM may serve as a flux-controlling enzyme beyond HMGCR (Figure 4.1 & Figure 4.3). Out data confirm that SM is a rate-limiting enzyme, as suggested by previous studies, particularly when cholesterol levels are high. Sterol-dependent squalene accumulation has been reported before (83,112,113), but not investigated or tested as a specific regulatory point. We demonstrated that control occurs through a post- transcriptional mechanism using a mutant cell-line that lacks sterol-regulated transcription by SREBP-2, as squalene still accumulated in a cholesterol- dependent manner (Figure 4.4). Cholesterol accelerated the degradation of the protein, as shown by increased turnover in the presence of cycloheximide (Figure 4.5). Protein levels could be rescued through incubation with proteasome inhibitors or addition of viral GA-repeats to the protein that provide resistance to proteasomal degradation (Figure 4.8 & Figure 4.9). Cholesterol treatment also increased ubiquitination of SM (Figure 4.10). Thus, SM is subjected to ER associated degradation (ERAD) through the ubiquitin-proteasome system. Our studies suggest that accelerated SM degradation in response to cholesterol has functional consequences, by contributing to flux control through the cholesterol synthesis pathway. Importantly, inhibition of the proteasome reduced the cholesterol-dependent accumulation of squalene (Figure 4.11), consistent with the argument that degradation of the protein reduces flux through the pathway at SM. Notably, squalene accumulation occurred with acetate labelling, which feeds into the beginning of the cholesterol synthesis pathway before HMGCR, highlighting the importance of the control point at SM. Squalene accumulation also occurred with mevalonate labelling, the product of HMGCR, identifying SM as a flux controlling enzyme beyond HMGCR. We often used a sterol-depleting pretreatment step to help dissect the effect and better control cholesterol levels in experiments, but cholesterol- dependent squalene accumulation also occurred similarly under more physiological conditions, without statin. Rapid degradation was triggered when

101 cholesterol levels rose above normal basal levels (Figure 4.7), reflecting a physiological feedback mechanism. On the basis of high sequence conservation in higher organisms, its absence in yeast and lack of catalytic function, we hypothesised that the first 100 amino acids of SM constitute a regulatory domain, which we denote as N100 (Figure 4.13). Indeed, deleting this region abolished cholesterol-regulation (Figure 4.14), increasing flux through the pathway at SM (Figure 4.15). Accordingly, SM’s N-terminal domain (N100) conferred cholesterol-regulated turnover on heterologous fusion proteins, with similar kinetics to that of full- length SM (Figure 4.16 & Figure 4.17). Degradation was acute, occurring within 1-2 hr, as shown by increased ubiquitination and pulse-chase analysis (Figure 4.18 & Figure 4.17), timing consistent with the rapid accumulation of squalene. Whilst generally comparable, the time-course of squalene accumulation did not exactly mirror falling SM protein levels for the full-length. However, the more rapid timing observed for N100 is compatible with that observed for squalene accumulation (obtained without cycloheximide). It is possible that cycloheximide may alter the kinetics of SM degradation, as has been noted previously for HMGCR (128). Nevertheless, there is a chance that other post- transcriptional mechanisms may be involved in the control of mammalian SM activity. SM degradation was not mediated by Insig (Figure 4.12), which is required for degradation of HMGCR in response to side-chain oxysterols, suggesting that ubiquitination of SM occurs via a different mechanism. This is consistent with the lack of a sterol-sensing domain and classic alpha-helical transmembrane domains within the membrane bound N100. Thus, cholesterol-dependent degradation constitutes a novel control point in the cholesterol synthesis pathway. This would allow more rapid shutdown than though inhibition of transcription by SREBP-2. The presence of this mechanism downstream of HMGCR might also enable differential control of sterol and isoprenoid synthesis. The exact sterol specificity of SM degradation and relevant E3 ligase remains to be determined, although potential regulatory proteins do not appear to be SREBP-2 target genes, as regulated turnover of SM was unaffected in Scap deficient mutant cells (Figure 4.12). These questions will be explored in Chapter 5.

102

Chapter 5

SM is degraded in response to cholesterol itself, and stabilised by unsaturated fatty acids

Julian Stevenson performed the DNA cloning, immunoprecipitation/pulldown and cholesterol assays. Julian Stevenson, Ika Kristiana and Winnie Luu conducted the siRNA and immunoblotting experiments. Julian Stevenson and Lisa Phan prepared the N100-GFP stable cell-line. Julian Stevenson and Laura Sharpe performed quantitative real-time PCR. Ika Kristiana performed the thin layer chromatography.

Data in figures 5.1-5.3 have previously been presented in the PhD thesis of co- author Saloni Gill (110).

This work has been submitted for publication in:

Stevenson, J., Kristiana, I., Luu, W., and Brown, A. J. (2014) Squalene monooxygenase, a key enzyme in cholesterol synthesis, is stabilised by unsaturated fatty acids. The Biochemical Journal [Under revision]

103

104

5 SM is degraded in response to cholesterol itself, and stabilised by unsaturated fatty acids

5.1 Introduction

Cellular cholesterol synthesis is carefully controlled, such as through transcription by SREBP-2 (30) and post-transcriptionally at HMGCR (51) and SM (Chapter 4). However, additional layers of complexity continue to be uncovered that affect the synthesis of this essential lipid. In the previous chapter, we described the regulated ER-associated protein degradation (ERAD) of SM in response to cholesterol accumulation, which appears to be distinct from the well characterised sterol-regulated ubiquitination and proteasomal degradation of HMGCR, as it does not require Insig. HMGCR is degraded in response to dihydrolanosterol and side-chain oxysterols (52) (64), which can involve activity of the E3 ubiquitin ligases Gp78 and Trc8 (129). However, the precise sterols and E3 that mediate degradation of SM are currently unknown. The turnover of other proteins involved in lipid metabolism is also affected by the level of fatty acids. Levels of the triglyceride lipase PNPLA3 are increased post-translationally by addition of saturated or monounsaturated free fatty acids (130). The lipid droplet protein Fsp27 is protected in response to addition of unsaturated fatty acid (oleate), triglyceride synthesis and lipid droplet formation, likely through translocation of the protein to lipid droplets, preventing ERAD (131). Oleate also stabilised the droplet proteins ADRP (132) and perilipin (133,134). Unsaturated fatty acids stabilise the essential sterol regulator, Insig-1, through inhibition of binding to UBXD8 (135), part of the protein degradation machinery (136). UBXD8 also acts as sensor of long-chain unsaturated fatty acids, diverting lipid synthesis towards phospholipid production, in addition to effects on transcription through Insig/SREBP (137). In contrast, the cholesterol export protein ABCA1 undergoes accelerated degradation in response to oleate (138). In this chapter, we aim to investigate the lipid-specificity of SM turnover, in response to sterols and fatty acids, as well as uncover further details of the molecular mechanism that leads to accelerated degradation.

105 5.2 Materials and methods

5.2.1 Materials Fatty acids were complexed with bovine serum albumin (BSA), as described (139). Fatty acids and siRNA were from Sigma-Aldrich. A 922500 (DGAT-1 inhibitor) was from Tocris Biosciences (Bristol, UK).

5.2.2 Preparation of the N100-GFP stable cell-line HEK N100-GFP cells stably express the first 100 amino acids of SM fused to the enhanced green fluorescent protein (GFP) with a V5 epitope tag. The expression vector was prepared by PIPE subcloning the region between the T7 and BGH primer binding sites from pTK-SM-N100-GFP (Chapter 4) to replace the equivalent region of pcDNA5/FRT (Life Technologies). This was then used to create a stable cell-line using the Flp-In system with the Human Embryonic Kidney (HEK) 293 T-REx cell line, according to manufacturer’s instructions (Life Technologies), where selection made use of 400 g/ml hygromycin B. Due to silencing of transgene expression over time (data not shown), the cell population was subjected to fluorescence activated cell sorting for GFP-positive cells with the assistance of the flow cytometry scientist Dr Christopher Brownlee (Biological Resources Imaging Laboratory, UNSW).

5.2.3 siRNA knockdown Cells were transfected for 24 hr according to the manufacturer’s instructions (Life Technologies), using a ratio of 37.5 nmol siRNA: 5 L of Lipofectamine RNAiMAX reagent. After transfection, cells were statin pretreated overnight, and then treated as indicated in the same media. MISSION siRNA (sense strand provided only) directed against: FAF2 (UBXD8), 5′-CGGUUUACCUAUUACACGATT-3′; DGAT-2, 5′-CCAUCUACUCCUUUGGAGATT-3′ and 5′-GACUAUUUGCUUUCAAAGATT-3′; PCYT1 (CTP:phosphocholine cytidylyltransferase alpha), 5′-CCGAAUUGUGCGGGAUUAUTT-3′ and 5′-CCUAUGUCAGGGUAACUAUTT-3′; MARCH6, 5′-GCATCTACAAGTGCTTGTT-3′; and universal negative control 1, were designed by Sigma-Aldrich.

106

5.2.4 Ubiquitination of N100-GFP Following treatment, cells were lysed in 1% digitonin in PBS, supplemented with N-ethylmaleimide (10 mM), protease inhibitor cocktail, and MG132 (10 M). Lysates were immunoprecipitated with GFP-trap beads (Chromotek), according to the manufacturer’s instructions. Samples were subjected to 10% SDS-PAGE followed by immunoblot analysis with anti-V5 (for SM) and anti- ubiquitin antibodies.

5.2.5 Statistical analysis Statistical differences for Figure 5.4, Figure 5.8A and Figure 5.11 were determined by the Student’s paired t-test (two-tailed), where p values of < 0.05 were considered statistically significant. For Figure 5.10, a two-way ANOVA with repeated measures was performed for control vs UBXD8 knockdown.

107 5.3 Results

5.3.1 Endogenous sterols can cause degradation of SM Cholesterol accumulation leads to the accelerated degradation SM (Chapter 4), but it is unclear if this is due to cholesterol itself or a derivative, as well as if other sterols and intermediates might cause degradation of SM. To address this, we first examined whether endogenously synthesised sterols, oxysterols or pathway intermediates could regulate SM. We pretreated cells for 16 hr with a statin to inhibit sterol synthesis, lowering cholesterol levels, and stabilising SREBP-2 target genes. The removal of the statin from the media then allows increased sterol synthesis due to the higher level of biosynthetic enzymes (63), and conversion of accumulated HMG-CoA and other non-sterol intermediates. Total cellular cholesterol at 8 hr increased by ~20% when statin was withdrawn from SRD-1 cells (an increase of 7 g/mg from 34 g/mg cell protein). This newly synthesized sterol was sufficient to accelerate the degradation of SM (Figure 5.1A). Inhibiting later steps in the mevalonate pathway, at SM and lanosterol synthase (the first sterol generating enzyme), preserved SM protein levels to the same extent as statin (Figure 5.1B, lanes 1, 3 and 5), indicating that the degradative signal is a sterol.

108

Figure 5.1. Degradation occurs in response to sterols, but not earlier intermediates. (A and B) SRD-1 cells were statin pretreated overnight and treated as indicated in medium A containing cycloheximide (10 g/ml) and either statin (compactin, 5 M, thus identical to medium B), SM inhibitor (SMi, 10 M), lanosterol synthase inhibitor (LSi, 10 M), and/or Chol/CD (20 g/ml) for up to 16 hr (A), or 8 hr (B). Cell lysates were assayed for SM by immunoblotting (n=4).

109 5.3.2 Cholesterol is the principal degradative signal To examine the sterol specificity of this regulatory mechanism, we tested the ability of a selection of sterol pathway intermediates and oxysterols to cause degradation of ectopic, human SM. A range of sterols and oxysterols were delivered complexed in methyl-β-cyclodextrin at the same high concentration as cholesterol (20 g/ml; Figure 5.2 & Figure 5.3). For the intermediate pathway sterols, the major methyl-sterol intermediates lanosterol and 24,25-dihydrolanosterol (24,25DHL) – the latter of which also accelerates degradation of HMGCR (52) – had no effect on SM (Figure 5.2, lanes 7-8). 7-Dehydrocholesterol (7DHC) and lathosterol with double bonds on the steroid ring also did not reduce SM protein levels (lane 5-6). Desmosterol, which differs only by a double bond on the side chain, was as effective as cholesterol at reducing SM protein (lane 4 vs 3).

110

Figure 5.2. SM degradation specificity for sterol intermediates. CHO-7 cells were transfected with 1 g of pTK-SM-V5. Following overnight statin pretreatment, cells were treated for 8 hr in medium B containing cycloheximide (10 g/ml) with or without: an equivalent concentration of CD alone or CD complexed with the following sterols (20 g/ml): cholesterol, desmosterol, 7-dehydrocholesterol (7DHC), lathosterol, 24,25-dihydrolanosterol (24,25DHL), or lanosterol. SM-V5 protein was analysed by immunoblotting (each n=4). Structure (140) names for sterols that accelerate degradation of SM are indicated in boldface.

111 5.3.3 SM is not degraded in response to side-chain oxysterols The TK-SM-V5 construct was next used to test oxysterols, which have one or more additional oxygen-containing groups on the carbon backbone of cholesterol or related sterols (Figure 5.3). Treatment with the side-chain oxysterols 24,25EC, 25HC and 27-hydroxycholesterol (27HC) again had no effect on SM protein levels (lanes 6-8), in contrast to what has been observed for HMGCR (64). However, the steroid-ring oxysterols 7α-hydroxycholesterol (7αHC), 7β-hydroxcholesterol (7βHC), 7-ketocholesterol (7KC) and synthetic 19-hydroxycholesterol (19HC) induced degradation (lanes 2-5). Significantly, the likely cellular concentration of these oxysterols would be dramatically supraphysiological and the most potent oxysterol tested, 19HC, is not found in nature (141,142). Thus, due to its much greater relative abundance compared to 7-oxygenated sterols (141) or desmosterol (45), cholesterol itself appears to be the primary signal that mediates accelerated turnover of SM.

112

Figure 5.3. SM degradation specificity for oxysterols. CHO-7 cells were transfected with 1 g pTK-SM-V5. Following overnight statin pretreatment, cells were treated for 8 hr in medium B containing cycloheximide (10 g/ml) with or without the following oxysterols (20 g/ml): 7α- hydroxycholesterol (7αHC), 7β-hydroxycholesterol (7βHC), 7-ketocholesterol (7KC), 19-hydroxycholesterol (19HC), 24,25EC, 25HC, 27-hydroxycholesterol (27HC). SM-V5 protein was analysed by immunoblotting (n=4). Structure (140) names for sterols that accelerate degradation of SM are indicated in boldface.

113 5.3.4 SM turnover is increased by the enantiomer of cholesterol Lipids can affect cholesterol homeostasis by directly binding to regulatory proteins, such as to a specific binding pocket. However, bulk membrane effects can also be important, such as changes in membrane fluidity. To test if SM turnover is influenced by cholesterol’s modulation of the local lipid environment, we used the enantiomer of cholesterol (ent-cholesterol). The enantiomer is the mirror image of cholesterol and thus has the same physicochemical properties as cholesterol and same effect on membrane properties. However, it should not directly bind to proteins, which are chiral and thus enantiospecific. Both endogenous SM in SRD-1 cells and full-length ectopic protein were degraded upon treatment with ent-cholesterol, albeit less potently than cholesterol (Figure 5.4). This is evidence that cholesterol binds directly to SM, but also that membrane affects can play a role, suggesting that non-sterol lipids may affect SM protein levels.

Figure 5.4. SM is degraded in response to ent-cholesterol. (A) SRD-1 cells were statin pretreated overnight and treated in medium B containing cycloheximide (10 g/ml) and/or cyclodextrin alone, Chol/CD (nat-C, 20 g/ml) or ent-cholesterol (enantiomer of cholesterol in cyclodextrin complex (20 g/ml)). Cell lysates were assayed for SM or SM-V5 by immunoblotting. (B) CHO-7 cells were transfected with 1 g of pTK-SM-V5 for 24 hr and treated as in (A). Densitometric values of SM are expressed relative to the non-treated condition, which was set to 1 (n ≥ 4, ±S.E. *, p ≤ 0.05 for ent- cholesterol vs the non-treated condition; **, p ≤ 0.01 for nat-cholesterol vs the non-treated condition.

114

5.3.5 Oleate treatment stabilises SM, which may increase flux through the cholesterol synthesis pathway To investigate the effect of lipids other than cholesterol on post-transcriptional control of SM, cells in lipoprotein-deficient media were incubated with the unsaturated fatty acid oleate complexed with BSA, and SM protein levels analysed using immunoblotting. For CHO-7 cells transfected with full-length SM, 100 M oleate markedly increased SM and partially reversed cholesterol- dependent degradation, albeit not to the extent of proteasomal inhibition using MG132 (Figure 5.5A). As described in Chapter 4, protection against degradation using MG132 reduced cholesterol-dependent squalene accumulation (Figure 5.5B & C), suggesting increased SM activity. Similarly, oleate treatment decreased cholesterol-dependent squalene accumulation in CHO-7 cells and SRD-1 cells. Thus, unsaturated fatty acids may prevent the post-transcriptional reduction in SM activity, maintaining flux through the pathway.

115

Figure 5.5. Oleate treatment reverses squalene accumulation and cholesterol- dependent degradation of SM. (A) CHO-7 cells were transfected with 1 g of pTK SM V5 encoding full-length human SM for 24 hr, statin pretreated overnight, and then treated in medium A with or without Chol/CD (20 g/ml), MG132 (10 M), oleate (100 M fatty acid delivered complexed with bovine serum albumin (BSA)), and cycloheximide (10 g/ml) for 8 hr. Cell lysates were assayed for SM-V5 by immunoblotting (n=4). (B, C) CHO-7 (B) or SRD-1 (C) cells were statin pretreated overnight and treated in medium A with or without Chol/CD (20 g/ml), oleate (100 M), MG132 (10 M), and cycloheximide (10 g/ml) for 6 hr, and then radiolabelled with [14C]-acetate in the absence of cycloheximide and statin for 2 hr. Lipid extracts were assayed for squalene accumulation by thin layer chromatography (n=2, mean + half-range).

5.3.6 Fatty acid specificity - unsaturated fatty acids stabilise N100-GFP To help refine the effect of oleate on SM protein levels, we employed an HEK293 cell line stably expressing the first 100 amino acids of human SM fused to an EGFP V5 epitope tag (N100-GFP, Figure 5.6A). This comprises the regulatory domain of SM that mediates cholesterol-dependent turnover (Chapter 4), with similar kinetics to the full-length protein. Treatment with a range of oleate concentrations confirmed that marked stabilisation was observed at 100 M, although there was a hint of increased protein levels at

116

lower concentrations (data not shown). Oleate stabilisation was clearest when comparing cholesterol-accelerated degradation, although this effect was also observed without cholesterol treatment. Treatment with higher concentrations of oleate (up to 400 M) did not further increase SM protein (Figure 5.6B). To determine if stabilisation was confined to oleate and other unsaturated fatty acids, the effect of a panel of fatty acids (Figure 5.6C) were examined. The saturated fatty acids, palmitate (16:0) and stearate (18:0), did not increase N100-GFP protein levels, in contrast to the monounsaturated fatty acids, palmitoleate (16:1) and oleate (18:1) (Figure 5.6D). Polyunsaturated fatty acids also protected N100-GFP against cholesterol-dependent degradation to various extents; linoleate (18:2 n-6) and α‑linolenate (18:3 n-3) appeared to provide the greatest protection, although this was still observed with omega-6 arachidonate (20:4 n-6), omega‑3 eicosapentaenoate (20:5 n-3), and docosahexaenoate (22:6 n-3) (Figure 5.6E). Thus, stabilisation of SM by fatty acids appears to be mediated solely by unsaturated fatty acids, including polyunsaturated fatty acids.

117

Figure 5.6. Fatty acid specificity of N100-GFP stabilisation. (A) Schematic of full length wild-type (WT) SM, and N100-GFP fusion expression constructs. (C) Lipid numbers (number of carbons:number of double bonds, position of first double bond) for the panel of fatty acids tested and whether they stabilise SM, indicated in boldface. (B, D, E) Overnight statin pretreated HEK293 cells stably overexpressing N100-GFP were treated in medium I with or without Chol/CD (20 g/ml) and fatty acid (salt/BSA complex for all fatty acids) for 8 hr, with: (B), 0–400 M oleate; (D) palmitate (16:0), stearate (18:0), palmitoleate (16:1), oleate (18:1); (E) oleate (18:1), linoleate (18:2), α-linolenate (18:3), arachidonate (20:4), eicosapentaenoate (20:5), and docosahexaenoate (22:6). Cell lysates were assayed for V5-tagged constructs by immunoblotting (n=4). The higher fatty acid concentration (400 M) was used in (D,E) to increase robustness and consistency between preparations.

5.3.7 Stabilisation is not mediated through reduction of free- cholesterol Accelerated degradation of SM requires cholesterol. Oleate can act as a substrate for cholesterol esterification, which could cause nascent free- cholesterol to be sequestered into a regulatory-inactive pool (143,144). To address this possibility, total free and esterified cholesterol were measured using an enzyme assay. Without cholesterol, negligible cholesteryl ester was detected in this system, whereas treatment with cholesterol or oleate caused cholesteryl ester to appear (Figure 5.7). With cholesterol treatment, oleate reduced the level of free cholesterol, although this effect was only modest, and levels still exceeded the threshold that accelerates degradation of SM.

118

Importantly, free-cholesterol levels were unaffected by oleate addition, whereas SM protein levels still increased under these conditions. It has been reported that ER cholesterol levels were similarly unaffected by saturated fatty acid loading, although oleate increased phospholipid synthesis (16). Therefore, loss of ER cholesterol due to cholesterol esterification is unlikely to play a part in oleate stabilisation of SM. SM turnover appears to be affected by both membrane effects and direct binding of cholesterol, as the enantiomer of cholesterol increases turnover of the protein, albeit less effectively than cholesterol itself (Figure 5.4). SM has been shown to also bind cholesterol specifically, identified first via a proteomic screen (145). Hence, unsaturated fatty acids may increase SM protein by modulating membrane properties.

Figure 5.7. Oleate-dependent stabilisation is not mediated through reduction of free cholesterol levels. Overnight statin pretreated CHO-7 cells were treated in medium A with or without Chol/CD (20 g/ml) and oleate (100 M) for 8 hr, as indicated. Free and esterified cholesterol levels were determined using the Amplex Red cholesterol assay kit (n=2, mean + half-range).

119 5.3.8 Maximum protection requires activation of fatty acids, but not triglyceride synthesis Next, we determined whether incorporation of oleate into other lipids is required for SM stabilisation. We employed Triacsin C to inhibit acyl-CoA synthetases, which disrupts multiple points in lipid metabolism, including cholesterol esterification and triglyceride synthesis. Triacsin C appeared to counteract stabilisation by oleate upon cholesterol treatment (Figure 5.8A), with a ~28% Triacsin C-dependent reduction in N100-GFP protein levels (lane 4 vs 8, p < 0.01 by paired t-test). This suggests that stabilisation involves incorporation of free unsaturated fatty acids into a lipid derivative. The major metabolic fate for exogenous oleate is incorporation into triglyceride (146). Triglyceride exists within lipid droplets in the cell, and treatment with oleate is commonly used to induce lipid droplet formation (139). The other major core constituent of lipid droplets is esterified cholesterol. Triacsin C treatment inhibits lipid droplet formation, chiefly by potently preventing triglyceride synthesis (147). SM has previously been found to associate with lipid droplets, including in mammalian cells (148,149). These findings raised the possibility that oleate was directing SM to lipid droplets, such that the different lipid droplet membrane environment or protein complement might halt ubiquitination. On the contrary, stabilisation was still observed with Triacsin C treatment in the absence of exogenous cholesterol (Figure 5.8A), whereas triglyceride synthesis is virtually eliminated by this drug (147). Triacsin C has been reported to be far less effective at inhibiting phospholipid synthesis, in part due to alternative remodelling pathways (147). Triglyceride synthesis was more specifically targeted through inhibition of diglyceride acyltransferase (DGAT) activity, the committed and final step in the pathway. Pharmacological inhibition and genetic silencing of the two DGAT isoforms further increased N100-GFP protein levels, rather than reversing oleate-dependent stabilisation (Figure 5.8B). Taken together, these data suggest that triglyceride synthesis and lipid droplet formation are not required for stabilisation of SM by unsaturated fatty acids. Inhibition of triglyceride synthesis causes acute diversion of diglyceride into phospholipids, chiefly phosphatidylcholine, as seen after oleate loading of adipocytes derived from DGAT knockout mice (150). Even without DGAT

120

inhibition, most oleate is acutely incorporated into phosphatidylcholine (151). This could lead to significant remodelling of ER lipids, altering membrane structure so as to reduce turnover of SM, such as by altering the activity of cholesterol or bulk membrane effects that also modulate SM degradation. Palmitate loading of hamster cells was reported to increase the saturation in the acyl chains of ER phospholipids and attenuate membrane-oxysterol cholesterol homeostatic responses, without changing ER cholesterol levels (16). Although ER lipid side chains are enriched for unsaturated fatty acids compared to the plasma membrane, oleate loading might be expected to analogously increase the proportion of unsaturated phospholipids. Phosphatidylcholine is also the major lipid species in the ER (~50%) (152,153), so we first inhibited phosphatidylcholine synthesis via the Kennedy pathway through siRNA knockdown of CTP:phosphocholine cytidylyltransferase (CT) alpha. Knockdown of CT did not reduce oleate-dependent stabilisation, and appeared to slightly increase N100-GFP protein levels (Figure 5.8C), albeit the increase in the presence of cholesterol was far less striking in the other versions of this experiment (data not shown).

121

Figure 5.8. Stabilisation requires activation of fatty acids, but not triglyceride synthesis. (A) Overnight statin pretreated HEK293 N100-GFP cells were treated in medium I with or without Chol/CD (20 g/ml), Triacsin C (5 M) and oleate (400 M) for 8 hr, as indicated. (B) HEK293 N100-GFP stable cells were transfected with siRNA (scrambled control or DGAT-2 for DGAT inhibition), statin pretreated overnight, then treated in medium I with or without Chol/CD (20 g/ml), oleate (400 M) and DGAT-1 inhibitor (1 M A 922500) for 8 hr, as indicated. (C) HEK293 N100-GFP stable cells were transfected with siRNA directed against CTP:phosphocholine cytidylyltransferase (CT) alpha, statin pretreated overnight, then treated in medium I with or without Chol/CD (20 g/ml) and oleate (400 M) for 8 hr, as indicated. Cell lysates were assayed for V5-tagged constructs by immunoblotting (n=4).

122

5.3.9 Stabilisation does not require protein kinase C activity We next explored the possible involvement of regulation by signalling through protein kinase C. Diglyceride and unsaturated fatty acids potently activate protein kinase C in vitro (154). Unsaturated fatty acids negatively regulate the cholesterol export protein, ABCA1, through activation of protein kinase C delta, leading to its phosphorylation and accelerated degradation (138), increasing cholesterol status. Conversely, phosphorylation of SM or another regulatory protein after activation of phospholipases or stimulation of diglyceride synthesis by oleate might protect SM, also maintaining higher cell cholesterol. However, protein kinase C inhibition with Bisindolylmaleimide I or Calphostin C, and activation with the diglyceride analogue, phorbol 12- myristate 13-acetate, had no effect on SM protein levels (Figure 5.9 & data not shown). This suggests that protein kinase C activity is not required for oleate stabilisation of SM.

Figure 5.9. Protein kinase C inhibition has no effect on oleate-dependent stabilisation of SM. Overnight statin pretreated HEK293 N100-GFP cells were treated in medium I with or without Chol/CD (20 g/ml), Bisindolylmaleimide I (5 M) and oleate (400 M) for 8 hr, as indicated. Cell lysates were assayed for V5-tagged constructs by immunoblotting (n=4).

123 5.3.10 Oleate stabilisation of SM can occur independently of UBXD8 Another important protein in cholesterol homeostasis is Insig-1, involved in HMGCR degradation (51) and ER retention of precursor SREBPs when cellular sterol status is high (30). In contrast to SM, Insig-1 is stabilised when cholesterol levels are high (155). By an independent mechanism, unsaturated free fatty acids protect Insig-1 against proteasomal degradation. This was reported to occur through inhibition of UBXD8 binding to Insig-1, preventing the membrane extraction of ubiquitinated Insig-1 to cytosolic proteasomes by VCP/p97, as UBXD8 is required for recruitment of VCP/p97 (135). UBXD8 has been observed to form part of an ERAD complex including the ubiquitin E3 ligase Gp78 (136). Hence, an analogous situation may exist for SM that leads to the stabilisation in response to free fatty acids. Thus, the effect of oleate was tested in combination with siRNA knockdown of UBXD8 in the N100-GFP stable cell line. Knockdown of UBXD8 increased overall N100-GFP protein levels (Figure 5.10, p=0.0002 by two-way ANOVA), where knockdown alone produced a similar effect to oleate treatment (lanes 3 vs 5, p=0.96 by paired t-test). In the presence of cholesterol, treatment with oleate in combination with UBXD8 knockdown further increased SM protein by 58% (lanes 4 vs 8, p=0.02 by paired t-test). Thus, UBXD8 can play a minor role in the turnover of SM, either directly or indirectly. This may occur by preventing UBXD8 from mediating extraction of ubiquitinated SM, similarly to the situation observed for Insig-1. However, the fact that the oleate effect and UBXD8 knockdown was additive (Figure 5.10) suggests that oleate-induced stabilisation operates by an additional, independent mechanism. Thus, UBXD8 is unlikely to play an essential role in oleate-mediated SM protection, as marked degradation still occurs in the presence of UBXD8 knockdown.

124

Figure 5.10. Protection can occur independently of UBXD8. HEK293 N100-GFP stable cells were transfected with siRNA to knock down (~90% message) UBXD8. Overnight statin pretreated cells were treated in medium I with or without Chol/CD (20 g/ml) and oleate (100 M) for 8 hr. A representative immunoblot is shown with densitometry for additional experiments. Relative N100-GFP protein was quantified by densitometry, and normalised to the non-treated control siRNA condition, which has been set to 1 (n=3, mean + S.E.M.).

125 5.3.11 Oleate blunts polyubiquitination by MARCH6 An alternative mechanism to explain SM stabilisation would be that the unsaturated fatty acids prevent its polyubiquitination, particularly in rapid response to cholesterol accumulation. In support of this, oleate reduced polyubiquitination of N100-GFP and reversed cholesterol-dependent ubiquitination (Figure 5.11A). Thus, unsaturated fatty acids can protect SM at the ubiquitination step. This is the simplest and likely most important explanation. Recently, cholesterol-dependent degradation of SM has been shown to be mediated by the E3 ubiquitin ligase, MARCH6, by our lab and others (100,156). To determine if MARCH6 is involved in oleate stabilisation of SM, we knocked down expression of this E3 ligase. MARCH6 knockdown greatly stabilised SM N100 (Figure 5.11B), as shown previously. However, this was not further enhanced by oleate treatment (lanes 5 vs 7, p=0.18 by paired t-test), in contrast to the stabilisation by oleate observed without knockdown (lanes 1 vs 3, p=0.005 by paired t-test). This is consistent with oleate-dependent stabilisation of SM occurring primarily through inhibition of ubiquitination by MARCH6.

126

Figure 5.11.Oleate blunts polyubiquitination by MARCH6. (A) Overnight statin pretreated HEK293 N100-GFP stable or untransfected cells were treated in medium I with or without Chol/CD (20 g/ml), MG132 (10 M) and oleate (400 M) for 2 hr. N100-GFP was pulled down with GFP- trap magnetic beads and SM and Ubiquitin analysed by immunoblotting (n=3). (B) HEK293 N100-GFP stable cells were transfected with siRNA to knock down (~60% message) MARCH6. Overnight statin pretreated cells in medium I were treated with or without Chol/CD (20 g/ml) and oleate (100 M) for 8 hr. A representative immunoblot is shown with densitometry for additional experiments. Relative N100-GFP protein was quantified by densitometry, and normalised to the non-treated MARCH6 siRNA condition, which has been set to 1 (n=4, mean + S.E.M.) 127 5.4 Discussion

In this chapter, we confirmed that cholesterol itself causes the accumulation of the squalene, and degradation of SM (Figure 5.1). Unlike HMGCR, other mevalonate pathway intermediates and side-chain oxysterols had no effect on SM turnover (Figure 5.2 & Figure 5.3). The ability of desmosterol to cause SM degradation reflects its close similarity to cholesterol, such that it can replace cholesterol completely in maintaining cell viability and modulate Scap normally ((97,157)). The potent effect of the steroid-ring oxysterols may reflect their increased solubility (15), whilst remaining sufficiently similar to cholesterol in terms of protein binding contacts or biophysical properties to enable accelerated degradation of SM. Direct binding of cholesterol to SM (145) appears to be more important than membrane effects, as the enantiomer of cholesterol was markedly less effective at causing SM degradation (Figure 5.4). In contrast to accelerated degradation by cholesterol, we observed stabilisation of SM by a wide range of unsaturated fatty acids (Figure 5.5 & Figure 5.6), independently of changes to free cholesterol levels (Figure 5.7) or protein kinase C activity (Figure 5.9). SM resides in the ER, but it is unclear if it directly binds unsaturated fatty acids that confer stabilisation. Our experience with the enantiomer of cholesterol suggest that both effects will be important, although this is difficult to demonstrate experimentally. Direct interaction of unsaturated fatty acids with UBXD8 may play a part in SM stabilisation (Figure 5.10), but if it does, it appears to be a minor one. The major mechanism appears to be attenuation of polyubiquitination by MARCH6 (Figure 5.11). It is plausible that unsaturated fatty acids might prevent a conformational change in SM that leads to the observed polyubiquitination by MARCH6 and proteasomal degradation. However, it is unlikely that they change the accessibility of MARCH6 by altering its conformation or activity, as MARCH6 targets multiple substrates, although this remains to be confirmed. Testing SM instead for conformational changes upon cholesterol and unsaturated fatty acid treatment would likely be more fruitful. A technically simpler experiment to strengthen the link between oleate-dependent stabilisation of SM and MARCH6 would be to use co-immunoprecipitation to confirm that unsaturated fatty acids reduce the interaction between the two proteins.

128

Maximum stabilisation of SM by unsaturated fatty acids required activation of free fatty acids, but not triglyceride synthesis (Figure 5.8). Stabilisation did not appear to require phosphotidylcholine synthesis either, as knockdown of one of the major enzymes that produces it had no effect on oleate-stabilisation (Figure 5.8). Confirmation and further investigation with strain 58 CHO temperature sensitive CTP:phosphocholine cytidylyltransferase activity mutant cells (158) could be useful, but the hint of increased protein levels (Figure 5.8) strongly suggests that increased phosphatidylcholine does not mediate SM stabilisation. The next major ER lipid is phosphatidylethanolamine, which albeit less than half as abundant compared to phosphatidylcholine (152,153), appears to play an important part in modulating lipid homeostasis in flies, which lack cholesterol (159). In Drosophila, phosphatidylethanolamine is thought to fulfill the role of cholesterol in mammalian cells, causing a change in ER membrane properties that is sensed by Scap, preventing lipid synthesis. However, this effect was most potent when the phosphatidylethanolamine was derived from saturated palmitate, in contrast to palmitoleic acid. Nevertheless, the degree of unsaturation in ER lipid tails is high in mammalian systems (16), which suggests that stabilisation is not mediated by increased phosphatidylethanolamine through reducing a cholesterol-like membrane structure. It is also unlikely to be diglyceride, as it is very rapidly turned over, such as into phospholipid and triglyceride (151,160). Thus, the identity of the precise lipid/s that confers stabilisation of SM independently of free fatty acids remains to be defined, although the presence of unsaturation is critical. In this chapter, we have demonstrated that SM is degraded in response to cholesterol itself via the activity of the E3 ligase MARCH6, which can be reversed by unsaturated fatty acids. However, further work is required to precisely define the molecular mechanisms of cholesterol-dependent degradation and unsaturated fatty acid-dependent stabilisation. Even so, these observations represent another link between cholesterol synthesis and fatty acid metabolism. The potential for a regulatory role of end product inhibition of cholesterol synthesis at SM appears more obvious, but the need for coordination with levels of other lipids can also be imagined. On the one hand, unsaturated fatty acids can increase triglyceride synthesis through inhibition of UBXD8 (137), as well as eventually decreasing fatty acid synthesis at the 129 transcriptional level through inhibition of the SREBPs. When the levels of unsaturated fatty acids are still high, stabilisation of SM protein may help maintain cholesterol synthesis and an appropriate cholesterol/phospholipid ratio. This could prevent disruption of membrane and protein structure and function when the cell is stressed by sudden changes to its lipid profile. When the unsaturated fatty acid is turned over or sequestered into triglyceride, excess SM could then be rapidly degraded once again.

130

Chapter 6

General discussion

131

132

6 General discussion

In this chapter, we reiterate our major findings and discuss their implications, with recommendations for future work to help resolve the many questions they raise.

6.1 Summary of findings

Cholesterol synthesis is controlled at multiple levels, including transcriptionally by the SREBPs (51), and post-transcriptionally, notably at HMGCR (51). In this thesis, we investigated the regulation of squalene monooxygenase (SM), a neglected cholesterogenic enzyme beyond HMGCR. We presented evidence that cholesterol can stimulate with considerable specificity the proteasomal degradation of SM. Consistent with this, cholesterol treatment caused the non-toxic isoprenoid squalene to accumulate in a variety of cell-lines (Chapter 4), preceding transcriptional downregulation of SM. Indeed, cholesterol-dependent squalene accumulation was also observed in mutant cells lacking transcriptional regulation by SREBP-2. At the post-translational level, this was accompanied by a reduction in endogenous SM protein levels. Cholesterol- dependent degradation of SM also required the ubiquitin-proteasome system: protein levels were rescued through proteasomal inhibition with MG132, and polyubiquitination increased on cholesterol treatment. This mechanism did not require Insig or Scap, but was mediated by the N-terminal region of human SM, which also conferred sterol-regulated turnover on heterologous fusion proteins (Chapter 4). Importantly, MG132 treatment reversed the accumulation of squalene, suggesting that accelerated proteasomal degradation of SM may help to acutely control flux through the cholesterol biosynthetic pathway. In an example of end-product inhibition, turnover was accelerated by cholesterol itself, rather than methylated sterols or side-chain oxysterols (Chapter 5). Both direct binding and bulk membrane effects were implicated in turnover of SM, as the enantiomer of cholesterol still caused degradation, albeit to a lesser extent. On the contrary, SM protein levels were increased when cells were treated with unsaturated, but not saturated fatty acids. This appeared to

133 be mediated primarily through inhibition of polyubiquitination of SM by the E3 ubiquitin ligase, MARCH6. This thesis also optimised and examined the utility of a range of ligation- independent cloning techniques, identifying important practical considerations for cloning project design (Chapter 3). These include observing that overlap extension cloning (OEC) is good for short products, but that PIPE cloning performs best for most applications. OEC and PIPE cloning efficiencies could suffer due to the presence of primer-dimers or generation of nicked copies of the vector template, respectively, but these problems can easily be overcome, such as through purification and reduction of the template concentration. These molecular tools were extremely useful for answering important biological questions throughout this work, such as for preparing an inducible dimerisation system, which was used to confirm the stimulatory effect of the important kinase Akt on SREBP-2 processing (Chapter 3).

6.2 Ligation-independent cloning and beyond

Ligation-independent cloning methods proved invaluable in conducting the work of this thesis. This included making modifications such as adding small to large fusion-protein domains, such as GFP, GA repeats or ubiquitin, or introducing sequence elements to change the behaviour of existing vectors, such as the incorporation of a FRT recombination site or weaker thymidine kinase promoter. Whilst most expression vectors could have been obtained using traditional restriction digestion and ligase dependent methods, using LIC instead, and PIPE in particular, ensured that preparation was simple, highly efficient and robust. We found that PIPE, SLIC and OEC were also very efficient for site- directed mutagenesis. PIPE or SLIC are optimal for large deletions (91), and extremely effective for substitutions or small insertions like epitope tags (Table 3.12 and data not shown). However, to create point mutations we routinely used an alternative two-stage cycling strategy similar to OEC, adapted from QuikChange (115,161). It uses one mutagenic primer and a shorter pre-existing one, often a sequencing primer, to create a megaprimer by PCR, which then generates the mutant plasmid by overlap extension. This normally provides adequate efficiencies at minimal expense and can be superior to PIPE for difficult templates, such as highly GC rich regions ((115) and data not shown).

134

There is still a place for in vitro ligation when performing simple subcloning steps, which minimises cost, but also as required for some restriction enzyme-independent cloning methods. The much higher cloning efficiencies observed with Gibson assembly ((109) and data not shown) are likely due to the creation of closed phosphodiester bonds. This precludes the need for the separate DNA ends to meet and be joined in vivo, which may be limited by the availability of DNA recombination and repair machinery. It may also reflect the increased stability of a closed circular plasmid in contrast to linear DNA. This suggests that the addition of ligases to OEC may significantly increase cloning efficiencies, such as by adding a short incubation with T5 exonuclease and Taq ligase after thermal cycling. Alternatively, free phosphates could be included in the products through prior phosphorylation of the primers, followed by addition of Taq ligase, which might rescue cloning of large inserts with a single pair of primers. Further optimisation of the multiple SDM method introduced in this thesis may involve similar modifications. Another streamlined cloning method which uses in vitro ligation is TA cloning (162). PCR products are tailed with a single overhanging 3’ A by Taq polymerase, which can base-pair with overhanging Ts of a linear vector and be covalently joined by T4 ligase. An improved version of this could be directional GC cloning. GC base-pairs have three rather than two hydrogen bonds, which should increase the cloning efficiency through improved binding. Inserts are normally randomly incorporated in both orientations, but directional cloning could be achieved by dephosphorylating a linear vector, cutting it a second time to reveal a free phosphate, blunting and C-tailing the backbone, then mixing with a G-tailed PCR amplicon containing a single phosphorylated primer. This has the advantage over LIC strategies of only requiring short primers. However, the numerous steps to prepare the vector and insert are time consuming and we have yet to test this method. A number of commercial cloning kits make use of site-specific recombination, requiring special vectors and costly proprietary enzymes, such as the Gateway® system (Life Technologies). Libraries of open reading frames are available for such systems or using ready-made vectors, albeit limited in design. As the cost of oligos and artificial DNA synthesis falls, LIC and Gibson cloning should become increasingly widespread, up to the point where it might be easiest to purchase complete custom vectors, with almost any DNA 135 sequence. However, to minimise cost or clone new sequences from biological samples, LIC techniques will likely remain important for many years to come.

6.3 Regulated degradation of SM is distinct from that of HMGCR

and does not appear to require lipid droplets

Clearly, while this mechanism is a second example of regulated degradation for the control of cholesterol synthesis, it is distinct from the well-characterised ER- associated degradation of HMGCR: SM lacks the five transmembrane sterol- sensing domain and its degradation is not mediated by Insig, 24,25-dihydrolanosterol or side-chain oxysterols such as 27HC. Indeed, the sterol specificity of SM degradation is also strikingly different to what has previously been shown to inhibit SREBP processing, and/or bind to Scap or Insig (summarised in Table 6.1). This may be explained by the observation that SM turnover appears to be influenced both by direct sterol-binding (145) and bulk membrane effects (Chapter 5). This sensing of the overall membrane lipid environment may also partly explain the stabilisation effect of unsaturated fatty acids.

136

Table 6.1. Sterol specificity for selected proteins in cholesterol homeostasis. Comparison of the ability (high, + ; moderate, +/- ; negligible, - ; not determined, nd ) of the listed sterols to inhibit SREBP processing, bind to Scap or Insig, and stimulate ubiquitination (or degradation) of HMGCR or SM. Modified from (39), and partly derived from data from (37,54,64,163). Sterols are grouped as either ring oxysterols (including the man-made 19-hydroxysterol), side-chain oxysterols, cholesterol and ring double-bond variant sterols, and methylated sterols. The last column summarises the results from Chapter 5. SREBP In vitro Binding Ub/Deg. Sterol Inhibition Scap Insig HMGR SM 7α-Hydroxycholesterol +/- - +/- - + 7β-Hydroxycholesterol +/- - +/- - + 7-Ketocholesterol +/- - +/- - + 19-Hydroxycholesterol - - - + + 24(S),25-Epoxycholesterol + - + + - 25-Hydroxycholesterol + - + + - 27-Hydroxycholesterol + - + + - Cholesterol + + - - + Desmosterol + + - - + 7-Dehydrocholesterol - nd nd +/- - Lathosterol - nd nd nd - 24,25-Dihydrolanosterol - nd nd + - Lanosterol - - - - -

The post-ubiquitination pathway for SM also appears to be different to that of HMGCR. HMGCR requires temporary association with a lipid droplet-like compartment for dislocation (57) and eventual degradation by the proteasome, whereas SM did not appear to, as inhibition of droplet formation did not rescue SM protein levels (Chapter 5). It has been proposed that lipid droplets might help move proteins from the ER membrane to the hydrophilic environment of the proteasome (164), but this is not universal (165), as observed here for SM. Nevertheless, SM has been shown to associate with lipid droplets (148,149). This may be artefactual, reflecting the close association of the ER with and around droplets. True association with lipid droplets as opposed to contaminating ER might be explained by bulk sequestration of proteins from nearby ER on the basis of common structural features. The lack of transmembrane domains in SM could be well suited to the monolayer surface of a droplet. From our own preliminary lipid droplet isolation experiments, negligible N100-GFP protein was found in the droplet fraction unless cells were overloaded with lipid overnight, with nearly all remaining in the ER membrane

137 fraction (data not shown). Thus, translocation to lipid droplets does not appear to play a role in regulation of SM.

6.4 N100 is sufficient for degradation

To begin delineating the mechanism of accelerated degradation of SM, we hypothesised that the N-terminus was involved, based on its lack of requirement for catalytic activity (66) and on its evolutionary conservation. This region is particularly well conserved between birds and mammals, suggesting that this mechanism may have evolved to assist cholesterol regulation in higher animals. Indeed, we showed that the first 100 amino acids (N100, comprising 17% of the protein) was sufficient to mediate cholesterol- regulated turnover of GST or GFP fusion proteins (Chapter 4), when GFP alone for example is otherwise remarkably stable (166). Furthermore, it is not known if regulated degradation requires a conformational change mediated by altered membrane structure or requires a second sensing protein. A cholesterol-dependent conformational change might be detected by ER microsomes with or without cholesterol and a labelling agent or protease, as performed previously for Scap (38,97). Alternatively, a fluorescence sensor system could be used, such as by adding probes and fluorescent tags to check for a cholesterol-dependent fluorescence resonance energy transfer signal in vivo. Similar strategies could be attempted to test the effect of unsaturated fatty acids on a conformational change, including use of artificial lipid micelles in vitro. The blunted regulation seen for the CMV-driven full-length expression vector (Chapter 4) hints at the presence of additional regulatory proteins. By contrast, competition for a limiting factor(s) did not appear to occur when we co-transfected full-length SM and the massively over-expressed N100-GFP fusion protein. N100-GFP under the control of the TK promoter is robustly regulated, but appears to be expressed at a similar protein level to that of the full-length SM driven by the CMV promoter. One possibility is that the smaller size and greater solubility of N100-GFP makes it easier to process by the ERAD machinery. Similarly, localisation to a subdomain of the ER of the larger and more hydrophobic full-length protein might create a bottleneck in ERAD of SM. Alternatively, the solubilising properties of the GFP tag might also make dislocation easier. The presence of a bottleneck appears unlikely though, as

138

turnover of the N100-GFP and full-length driven by the moderate expression TK promoter was unaffected by their coexpression (Chapter 4). However, a better test of this would be to co-express the original CMV-driven full-length and TK N100-GFP. SM is unlikely to require a second cholesterol-sensing protein for regulated degradation, since it has been shown to bind cholesterol (145). It is plausible to expect that cholesterol will bind to N100 alone, rather than the catalytic domain, since N100 fusion proteins recapitulate the post-translational regulation of wild-type SM. Stabilisation by unsaturated fatty acids may also be indirect however, including via UBXD8 or other proteins, although changing the conformation of SM itself appears plausible in light of the reduced polyubiquitination observed upon oleate treatment. Indeed, a preliminary experiment suggested differential labelling of a cysteine with polyethylene glycol maleimide in N100 (data not shown). The precise amino acids within N100 that sense cholesterol are currently unknown. We investigated this by testing a series of 12-amino-acid truncations or internal deletions for expression and regulation. This strategy failed to clearly identify the region of interest, as it did not reveal a mutant of similar stability and expression that lacked cholesterol-dependent regulation (data not shown). Such large deletions are likely to significantly change tertiary structure, leading to proteasomal degradation or folding into quite different non- functional but nevertheless stable structures. Smaller truncations selected on the basis of homology or systematic alanine-scanning mutagenesis to highlight critical residues would be less likely to disrupt the overall fold of this putative degron domain. The effect of such mutations could be tested on the interaction with other regulatory proteins, such as through co-immunoprecipitation experiments for over-expressed candidate proteins, or using SILAC (stable isotope labelling with amino acids in cell culture) for quantitative proteomics, comparing a GFP transfected line with an N100-GFP over-expresser with or without cholesterol for a more sensitive, blind screen. To determine if the residues play a role in direct binding of cholesterol, cells could be treated with a photoactivatable derivative of cholesterol (145), SM isolated and analysed for binding of the probe. Delivery of such a molecule is difficult, due to its low concentration and poor solubility. Thus, the efficiency of delivery could be improved using a cyclodextrin complex containing lanosterol, which should not 139 bind to SM, but spiked with the small amount of photo-cholesterol probe. The probe can be visualised using radiolabelled sterol with a scintillant or alternatively by using a photoactivatable sterol which can be specifically labelled with a fluorophore using copper-catalysed azide-alkyne cycloaddition (click chemistry) (145), which would be preferable due to reduced cost and improved safety.

6.5 SM tertiary structure

Relatively little is known about SM structure, particularly the first 100 amino acids. Although the membrane topology of SM is currently unknown, our preliminary investigations have determined that the hydrophobic N-terminal domain is sufficient but not necessary for membrane association (Chapter 3). Further structural information could be obtained using cysteine-derivitisation and labelling to map membrane topology. For example, by using polyethylene glycol bound to a maleimide active group, which will only label exposed cysteine residues in the cytosol, which appears to be working well in preliminary experiments (data not shown). The strong association with the membrane suggests that traditional structural determination techniques such as X-ray crystallography would prove unsuccessful. This is because recombinant expression of membrane proteins in sufficient quantities, their purification, solubilisation and refolding with the right detergent is very challenging (167). Structural information could also be obtained using site-directed spin-label electron paramagnetic resonance or nuclear magnetic resonance spectroscopy, although these techniques are also technically difficult or suffer from low resolution (167). On the contrary, the relatively small size of this degron domain suggests that even limited distance constraints obtained by fluorescence based techniques in vivo could prove very fruitful in combination with computational prediction.

6.6 SM is ubiquitinated by the E3 ubiquitin ligase, MARCH6

Work in our laboratory ((100) and Chapter 5) has recently established that SM is targeted for ubiquitination by the E3 ubiquitin ligase MARCH6/TEB4, the homologue of the important yeast ERAD protein, Doa10, the major E3 ligase in yeast together with Hrd-1. The yeast enzyme lacks N100 and degradation is thought to occur in response to lanosterol (156), so would likely occur via a

140

different mechanism, beyond using similar core components. Hrd1 and Gp78 do not appear to play an important role in cholesterol-dependent degradation of human SM (100). In mammalian cells, MARCH6 has only been reported to target Type 2 Iodothyronine Deiodinase and itself for ubiquitination (168,169). It is likely that the UBC7 homologue Ube2g2 is again the cognate E2 to MARCH6 for ubiquitination of SM. SM has been shown to have multiple ubiquitination sites scattered across the protein (170-172), including one within the N100 region. We observed that turnover of N100-GFP was unaffected when we substituted the lysines for arginine in the first hundred amino acids. This may suggest that ubiquitination occurs when MARCH6 binds to SM with one recognition domain, whereas the catalytic RING domain of the E3 will ‘search’ for the closest lysine to target, or possibly an N-terminal amino acid. Similar results have been reported for HMGCR, although it has strongly preferred sites (55,125), which may suggest that this is a common feature of E3s involved in ERAD. Sterol accelerated degradation of HMGCR has been reported to occur primarily through the action of Gp78 (55), although this is uncertain, with degradation occurring normally in a gp78−/− s genetic background, and as knockdown using lower concentrations of siRNA than the original studies failed to prevent degradation although protein levels of Gp78 were successfully depleted (173). In yeast, certain isoprenoids can cause a conformational change in HMGCR, which is thought to be recognised as less folded by quality control machinery (174), leading to its degradation. Whether MARCH6 binds to a specific motif on SM or senses more general properties, such as hydrophobic or disordered patches in the cytosol, is currently unknown. If partial unfolding occurs in response to cholesterol, then this might be reversed by treating cells with ‘chemical chaperones’ such as glycerol or trimethylamine N-oxide (175). It is also possible that a chaperone-like or other adapter protein binds to a cholesterol-dependent conformation in SM, acting as a stable scaffold to recruit MARCH6. This could be investigated by searching for interacting partners using proteomic techniques and/or knocking down candidate proteins using siRNA as mentioned above. MARCH6 has been shown to interact with SM using immunoprecipitation and immunoblotting (100), but that does not preclude the presence of additional core proteins to make up a ternary complex. Use of in vitro binding assays to demonstrate specific interaction between 141 recombinant MARCH6 and SM mutants alone in model membranes or in detergent solution could also be employed. However, evidence of interaction in such an unphysiological system should be treated with caution, even with good controls. Proteomic approaches using mass spectrometry and knockdown of candidate interactors would be more physiologically relevant. This would likely be unsuccessful using MARCH6 as a bait protein, as seen previously (136), as it targets itself for destruction and thus is extremely unstable (168). A better strategy would be to use the catalytically inactive MARCH6 mutant instead, which is very stable, but still binds SM (100).

6.7 SM dislocation

Relatively little is known about the mechanism by which SM is translocated to the proteasome, although it does not appear to involve lipid droplets nor strictly require UBXD8 (Chapter 4). It is possible that UBXD8 may interact with SM as part of the ERAD machinery, but that this occurs very transiently and at a low frequency. In a series of proteomic screens (136), SM interacted with the ERAD-associated proteins, Hrd1 and Derlin-1 – both of which interacted with UBXD8 and VCP/p97 – as well as FAM8A1, but not with UBXD8 or VCP/p97 directly. However, knockdown of the important E3 ligases, Hrd1 and Gp78 had no effect on SM degradation, in contrast to MARCH6 (100). This suggests that VCP/p97 and UBXD8 may be involved in basal SM turnover as opposed to cholesterol-dependent degradation, although we cannot rule out a redundant role for these ERAD proteins in regulated degradation of SM. Preliminary experiments for siRNA knockdown of VCP/p97 unexpectedly suggested no effect on cholesterol-dependent degradation of SM (data not shown). VCP/p97 usually plays an essential role in multiple processes, such as retrotranslocation of proteins for ERAD. The apparent lack of this requirement warrants further investigation, but it may be that dislocation of SM is mediated by the AAA ATPases of the proteasome itself. This has been previously suggested for a number of substrates (176,177). This might be tested using siRNA or isolating microsomes and incubating them with defined cytosolic components such as the regulatory unit of the proteasome and assaying for translocation (177). Hrd1 and Doa10 have been proposed to act as a channel for retrotranslocation, as an alternative to Sec61p, but the most important putative interaction partner for SM dislocation is likely Derlin-1, which plays an essential role in this for other

142

proteins (177-179). This could be confirmed in the in vitro assay using an inhibitory Derlin-1 antibody or through silencing in live cells using siRNA, although this approach would be vulnerable to indirect effects. Thus, dislocation of SM may in part be facilitated by Derlin-1, acting to apply membrane tension or recruit necessary machinery, possibly with assistance of MARCH6 or Hrd1 to help provide a docking environment on the surface of the ER. Such a mechanism may again be determined by a lack of classical transmembrane domains.

6.8 Advantages of additional control points in cholesterol

synthesis

What might be the evolutionary advantage of having an additional post- translational control point for regulating cholesterol synthesis beyond HMGCR? Transcriptional downregulation of the SREBP pathway is relatively slow, with mRNA levels of target genes only appreciably decreasing hours after treatment (Chapter 4), potentially leaving active enzyme with a relatively long half-life. More rapid shutdown of cholesterol synthesis requires post- transcriptional control, such as the well-documented proteasome-mediated degradation of HMGCR (51). However, HMGCR activity can vary widely (51), such that sterol synthesis could be difficult to dampen in a timely manner. Moreover, some HMGCR activity is needed for isoprenoid production (113). In contrast, SM appears to have much lower activity than HMGCR (82) and is committed to sterol production, making it suitable for rapid modulation of the cholesterol biosynthetic pathway independently from isoprenoid synthesis. We propose a model whereby the mevalonate pathway is controlled rapidly and segmentally at HMGCR and SM (Figure 6.1). One situation where this could be useful would be to respond to a sudden influx of exogenous cholesterol while the demand for isoprenoids such as farnesol or geranylgeraniol remains high. This argument is consistent with the observation that the physiologically relevant cholesterol molecule itself appears to be the major signal for SM degradation (Chapter 5). Our findings may explain the often observed positive association between serum levels of squalene and cholesterol in humans, under a variety of pathophysiological and pharmacological conditions [e.g. (180)]. It is likely that unsaturated fatty acids will also act differentially on HMGCR, such

143 as by increasing degradation through increased levels of Insig-1, but this remains to be tested.

Figure 6.1. Segmental control of the mevalonate pathway. Sterol-dependent, post-translational control of the mevalonate pathway is mediated largely by the ubiquitin-proteasome system. HMGCR is degraded in response to 24,25-dihydrolanosterol and side-chain oxysterols, which will consequently also reduce flux into the non-sterol branch of the pathway. In contrast, turnover of SM is accelerated by the end-product, cholesterol itself, specifically inhibiting sterol production. This cholesterol-dependent degradation of SM requires its N-terminal 100 amino acids, here designated as the regulatory domain (Reg.), which is separate from the catalytic portion of the enzyme, and also activity of the E3 ubiquitin ligase MARCH6. On the contrary, SM is stabilised by unsaturated fatty acids (UFAs), which appears to occur through inhibition of ubiquitination by MARCH6.

The location of the control point at SM in the pathway has intriguing consequences for other regulatory molecules beyond cholesterol and its oxysterol derivatives, although the significance of this is unknown. The reduction of SM activity will lower the production of the endogenous regulator and potent LXR agonist 24(S),25-epoxycholesterol, which is sensitive to the ratio of SM to the subsequent more active enzyme, lanosterol synthase (63,181). 24(S),25-epoxycholesterol helps to smooth cholesterol homeostatic responses,

144

activating LXR, accelerating degradation of HMGCR and inhibiting SREBP-2 processing, but SM degradation may prevent overcompensation, fulfilling an analogous moderating role. Work in our laboratory revealed that the final step in the Bloch pathway of cholesterol synthesis catalysed by 3β-hydroxysterol ∆24-reductase (DHCR24) is directly inhibited by 24(S),25-epoxycholesterol (182), which can lead to the accumulation of desmosterol, thought to be the most physiologically relevant ligand for LXR in some contexts (183). Thus, small changes in SM activity may have important effects on LXR signalling. These possibilities might be investigated through lentiviral over expression of a cholesterol-resistant mutant, such as the catalytic domain of SM, and measuring transcription of relevant targets. Isotopic- or radio-labelling could also be used to measure acute changes in flux and the size of metabolite pools over time under different SM regulatory conditions, achieved most efficiently using gas chromatography tandem mass spectrometry (GC-MS) and deuterated labelling. For the greatest physiological relevance, studies should be repeated in vivo, such as using animal models, although this would increase the scale of experiments. To better understand such complex regulatory networks, it may also be beneficial to mathematically model the pathway using increasing knowledge of regulatory steps and flux changes, which could also generate new hypotheses. It is likely that other post-transcriptional regulatory mechanisms remain to be uncovered in the control of cholesterol synthesis, as we have reported in this thesis and elsewhere (182,184). Other candidate enzymes could be found by investigating acute sterol-dependent accumulation of cholesterol synthesis intermediates using GC-MS as above, in combination with quantitive proteomics. SRD-1 cells may simplify and increase the sensitivity of the metabolomic approach by removing transcriptional regulation by SREBP-2 and increasing the levels of each pool. Together with squalene, lanosterol is a major intermediate, but is unlikely to be as important in acute control of cholesterol synthesis, as accumulation did not increase upon sterol treatment. Another intermediate of particular interest would be 7-dehydrocholesterol, due to its position in the pathway as a possible penultimate sterol in the mevalonate pathway and precursor for vitamin D.

145 6.9 Conclusion

Our results reveal yet another layer of complexity in the control of cholesterol synthesis. In an example of end-product inhibition, cholesterol feeds back to stimulate the proteasomal degradation of SM in a mechanism that is distinct from that previously established for HMGCR. This indicates that SM constitutes a second important control point in cholesterol synthesis, serving as a reminder that flux control of metabolic pathways tends to be shared by multiple enzymes (185), which could facilitate flexible control of different intermediates. However, to what extent degradation of SM normally regulates cholesterol synthesis in vivo remains to be determined. The observed stabilisation of SM by unsaturated fatty acids further highlights the complexity of the regulation of metabolic pathways, also suggesting functionally significant cross-talk between sterol and fatty acid metabolism, although the precise nature of this interaction remains to be defined. This thesis also demonstrated the advantages of optimising and utilising improved DNA cloning technologies to increase the pace of research. Furthermore, our work may have important implications for human health and disease, notably in relation to cardiovascular disease, but also other conditions (e.g., neurodegenerative diseases and certain cancers) in which cholesterol has also been implicated. Moreover, it may stimulate interest in SM as a potential therapeutic target for cholesterol-related diseases (78), possibly by stimulating the degradation mechanism itself with a small- molecule drug.

146

7 References

1. Chevreul, M. (1816) Examen des graisses d’homme, de mouton, de boeuf, de jaguar et d’oie. Ann. Chim. Phys 2, 339-372 2. Lunde, G. (1930) The 1927 and 1928 Nobel chemistry prize winners, Wieland and Windaus. Journal of Chemical Education 7, 1763 3. Fisher, E. A. (2011) Introduction for National Cholesterol Month. Arteriosclerosis, thrombosis, and vascular biology 31, 1939-1940 4. Simons, K., and Ikonen, E. (2000) How cells handle cholesterol. Science 290, 1721-1726 5. Simons, K., and Vaz, W. L. (2004) Model systems, lipid rafts, and cell membranes. Annual review of biophysics and biomolecular structure 33, 269- 295 6. Churchward, M. A., Rogasevskaia, T., Hofgen, J., Bau, J., and Coorssen, J. R. (2005) Cholesterol facilitates the native mechanism of Ca2+-triggered membrane fusion. Journal of cell science 118, 4833-4848 7. Dreja, K., Voldstedlund, M., Vinten, J., Tranum-Jensen, J., Hellstrand, P., and Sward, K. (2002) Cholesterol depletion disrupts caveolae and differentially impairs agonist-induced arterial contraction. Arteriosclerosis, thrombosis, and vascular biology 22, 1267-1272 8. Echarri, A., Muriel, O., and Del Pozo, M. A. (2007) Intracellular trafficking of raft/caveolae domains: insights from integrin signaling. Seminars in cell & developmental biology 18, 627-637 9. Lu, X., Kambe, F., Cao, X., Yoshida, T., Ohmori, S., Murakami, K., Kaji, T., Ishii, T., Zadworny, D., and Seo, H. (2006) DHCR24-knockout embryonic fibroblasts are susceptible to serum withdrawal-induced apoptosis because of dysfunction of caveolae and insulin-Akt-Bad signaling. Endocrinology 147, 3123-3132 10. Brown, A. J. (2007) Cholesterol, statins and cancer. Clinical and experimental pharmacology & physiology 34, 135-141 11. Lewis, P. M., Dunn, M. P., McMahon, J. A., Logan, M., Martin, J. F., St- Jacques, B., and McMahon, A. P. (2001) Cholesterol modification of sonic hedgehog is required for long-range signaling activity and effective modulation of signaling by Ptc1. Cell 105, 599-612 12. Chen, H. W. (1984) Role of cholesterol metabolism in cell growth. Federation proceedings 43, 126-130 13. Hsu, H. J., Hsu, N. C., Hu, M. C., and Chung, B. C. (2006) Steroidogenesis in zebrafish and mouse models. Molecular and cellular endocrinology 248, 160-163 14. Norlin, M., and Wikvall, K. (2007) Enzymes in the conversion of cholesterol into bile acids. Current molecular medicine 7, 199-218 15. Gill, S., Chow, R., and Brown, A. J. (2008) Sterol regulators of cholesterol homeostasis and beyond: the oxysterol hypothesis revisited and revised. Progress in lipid research 47, 391-404 16. Gale, S. E., Westover, E. J., Dudley, N., Krishnan, K., Merlin, S., Scherrer, D. E., Han, X., Zhai, X., Brockman, H. L., Brown, R. E., Covey, D. F., Schaffer, J. E., Schlesinger, P., and Ory, D. S. (2009) Side chain oxygenated cholesterol regulates cellular cholesterol homeostasis through direct sterol-membrane interactions. The Journal of biological chemistry 284, 1755-1764 147 17. Falkenstein, E., Tillmann, H. C., Christ, M., Feuring, M., and Wehling, M. (2000) Multiple actions of steroid hormones--a focus on rapid, nongenomic effects. Pharmacological reviews 52, 513-556 18. Wilson, C. A., and Davies, D. C. (2007) The control of sexual differentiation of the reproductive system and brain. Reproduction 133, 331-359 19. Feng, B., Yao, P. M., Li, Y., Devlin, C. M., Zhang, D., Harding, H. P., Sweeney, M., Rong, J. X., Kuriakose, G., Fisher, E. A., Marks, A. R., Ron, D., and Tabas, I. (2003) The endoplasmic reticulum is the site of cholesterol-induced cytotoxicity in macrophages. Nature cell biology 5, 781-792 20. Rawicz, W., Olbrich, K. C., McIntosh, T., Needham, D., and Evans, E. (2000) Effect of chain length and unsaturation on elasticity of lipid bilayers. Biophysical journal 79, 328-339 21. Guyton, J. R., and Klemp, K. F. (1996) Development of the lipid-rich core in human atherosclerosis. Arteriosclerosis, thrombosis, and vascular biology 16, 4-11 22. Pekkanen, J., Linn, S., Heiss, G., Suchindran, C. M., Leon, A., Rifkind, B. M., and Tyroler, H. A. (1990) Ten-year mortality from cardiovascular disease in relation to cholesterol level among men with and without preexisting cardiovascular disease. The New England journal of medicine 322, 1700-1707 23. Maxfield, F. R., and Tabas, I. (2005) Role of cholesterol and lipid organization in disease. Nature 438, 612-621 24. Cullen, P., Rauterberg, J., and Lorkowski, S. (2005) The pathogenesis of atherosclerosis. Handbook of experimental pharmacology, 3-70 25. Harris, J. J., and Attwell, D. (2012) The energetics of CNS white matter. The Journal of neuroscience : the official journal of the Society for Neuroscience 32, 356-371 26. Brown, M. S., and Goldstein, J. L. (1986) A receptor-mediated pathway for cholesterol homeostasis. Science 232, 34-47 27. Repa, J. J., and Mangelsdorf, D. J. (2002) The liver X receptor gene team: potential new players in atherosclerosis. Nature medicine 8, 1243-1248 28. Oram, J. F., and Vaughan, A. M. (2006) ATP-Binding cassette cholesterol transporters and cardiovascular disease. Circulation research 99, 1031-1043 29. Tontonoz, P., and Mangelsdorf, D. J. (2003) Liver X receptor signaling pathways in cardiovascular disease. Molecular endocrinology 17, 985-993 30. Goldstein, J. L., DeBose-Boyd, R. A., and Brown, M. S. (2006) Protein sensors for membrane sterols. Cell 124, 35-46 31. Krycer, J. R., and Brown, A. J. (2011) Cross-talk between the androgen receptor and the liver X receptor: implications for cholesterol homeostasis. The Journal of biological chemistry 286, 20637-20647 32. Eberle, D., Hegarty, B., Bossard, P., Ferre, P., and Foufelle, F. (2004) SREBP transcription factors: master regulators of lipid homeostasis. Biochimie 86, 839-848 33. Horton, J. D., Cohen, J. C., and Hobbs, H. H. (2007) Molecular biology of PCSK9: its role in LDL metabolism. Trends in biochemical sciences 32, 71-77 34. Radhakrishnan, A., Sun, L. P., Kwon, H. J., Brown, M. S., and Goldstein, J. L. (2004) Direct binding of cholesterol to the purified membrane region of SCAP: mechanism for a sterol-sensing domain. Molecular cell 15, 259- 268 35. Motamed, M., Zhang, Y., Wang, M. L., Seemann, J., Kwon, H. J., Goldstein, J. L., and Brown, M. S. (2011) Identification of luminal Loop 1

148

of Scap protein as the sterol sensor that maintains cholesterol homeostasis. The Journal of biological chemistry 286, 18002-18012 36. Brown, M. S., and Goldstein, J. L. (1999) A proteolytic pathway that controls the cholesterol content of membranes, cells, and blood. Proceedings of the National Academy of Sciences of the United States of America 96, 11041-11048 37. Adams, C. M., Reitz, J., De Brabander, J. K., Feramisco, J. D., Li, L., Brown, M. S., and Goldstein, J. L. (2004) Cholesterol and 25- hydroxycholesterol inhibit activation of SREBPs by different mechanisms, both involving SCAP and Insigs. The Journal of biological chemistry 279, 52772-52780 38. Sun, L. P., Seemann, J., Goldstein, J. L., and Brown, M. S. (2007) Sterol- regulated transport of SREBPs from endoplasmic reticulum to Golgi: Insig renders sorting signal in Scap inaccessible to COPII proteins. Proceedings of the National Academy of Sciences of the United States of America 104, 6519-6526 39. Radhakrishnan, A., Ikeda, Y., Kwon, H. J., Brown, M. S., and Goldstein, J. L. (2007) Sterol-regulated transport of SREBPs from endoplasmic reticulum to Golgi: oxysterols block transport by binding to Insig. Proceedings of the National Academy of Sciences of the United States of America 104, 6511-6518 40. Du, X., Kristiana, I., Wong, J., and Brown, A. J. (2006) Involvement of Akt in ER-to-Golgi transport of SCAP/SREBP: a link between a key cell proliferative pathway and membrane synthesis. Molecular biology of the cell 17, 2735-2745 41. Engelman, J. A., Luo, J., and Cantley, L. C. (2006) The evolution of phosphatidylinositol 3-kinases as regulators of growth and metabolism. Nature reviews. Genetics 7, 606-619 42. Krycer, J. R., Sharpe, L. J., Luu, W., and Brown, A. J. (2010) The Akt- SREBP nexus: cell signaling meets lipid metabolism. Trends in endocrinology and metabolism: TEM 21, 268-276 43. Vanhaesebroeck, B., and Alessi, D. R. (2000) The PI3K-PDK1 connection: more than just a road to PKB. The Biochemical journal 346 Pt 3, 561-576 44. Zelcer, N., Hong, C., Boyadjian, R., and Tontonoz, P. (2009) LXR regulates cholesterol uptake through Idol-dependent ubiquitination of the LDL receptor. Science 325, 100-104 45. Yang, C., McDonald, J. G., Patel, A., Zhang, Y., Umetani, M., Xu, F., Westover, E. J., Covey, D. F., Mangelsdorf, D. J., Cohen, J. C., and Hobbs, H. H. (2006) Sterol intermediates from cholesterol biosynthetic pathway as liver X receptor ligands. The Journal of biological chemistry 281, 27816- 27826 46. Rodwell, V. W., Nordstrom, J. L., and Mitschelen, J. J. (1976) Regulation of HMG-CoA reductase. Advances in lipid research 14, 1-74 47. Langsjoen, P. H., and Langsjoen, A. M. (2003) The clinical use of HMG CoA-reductase inhibitors and the associated depletion of coenzyme Q10. A review of animal and human publications. BioFactors 18, 101-111 48. Robinson, J. G. (2008) LDL reduction: how low should we go and is it safe? Current cardiology reports 10, 481-487 49. Goldstein, J. L., and Brown, M. S. (1990) Regulation of the mevalonate pathway. Nature 343, 425-430

149 50. Jo, Y., and Debose-Boyd, R. A. (2010) Control of cholesterol synthesis through regulated ER-associated degradation of HMG CoA reductase. Critical reviews in biochemistry and molecular biology 45, 185-198 51. DeBose-Boyd, R. A. (2008) Feedback regulation of cholesterol synthesis: sterol-accelerated ubiquitination and degradation of HMG CoA reductase. Cell research 18, 609-621 52. Lange, Y., Ory, D. S., Ye, J., Lanier, M. H., Hsu, F. F., and Steck, T. L. (2008) Effectors of rapid homeostatic responses of endoplasmic reticulum cholesterol and 3-hydroxy-3-methylglutaryl-CoA reductase. The Journal of biological chemistry 283, 1445-1455 53. Ravid, T., Doolman, R., Avner, R., Harats, D., and Roitelman, J. (2000) The ubiquitin-proteasome pathway mediates the regulated degradation of mammalian 3-hydroxy-3-methylglutaryl-coenzyme A reductase. The Journal of biological chemistry 275, 35840-35847 54. Song, B. L., Javitt, N. B., and DeBose-Boyd, R. A. (2005) Insig-mediated degradation of HMG CoA reductase stimulated by lanosterol, an intermediate in the synthesis of cholesterol. Cell metabolism 1, 179-189 55. Song, B. L., Sever, N., and DeBose-Boyd, R. A. (2005) Gp78, a membrane- anchored ubiquitin ligase, associates with Insig-1 and couples sterol- regulated ubiquitination to degradation of HMG CoA reductase. Molecular cell 19, 829-840 56. Leichner, G. S., Avner, R., Harats, D., and Roitelman, J. (2009) Dislocation of HMG-CoA reductase and Insig-1, two polytopic endoplasmic reticulum proteins, en route to proteasomal degradation. Molecular biology of the cell 20, 3330-3341 57. Hartman, I. Z., Liu, P., Zehmer, J. K., Luby-Phelps, K., Jo, Y., Anderson, R. G., and DeBose-Boyd, R. A. (2010) Sterol-induced dislocation of 3- hydroxy-3-methylglutaryl coenzyme A reductase from endoplasmic reticulum membranes into the cytosol through a subcellular compartment resembling lipid droplets. The Journal of biological chemistry 285, 19288-19298 58. Sever, N., Yang, T., Brown, M. S., Goldstein, J. L., and DeBose-Boyd, R. A. (2003) Accelerated degradation of HMG CoA reductase mediated by binding of insig-1 to its sterol-sensing domain. Molecular cell 11, 25-33 59. Meusser, B., Hirsch, C., Jarosch, E., and Sommer, T. (2005) ERAD: the long road to destruction. Nature cell biology 7, 766-772 60. Olzmann, J. A., Kopito, R. R., and Christianson, J. C. (2013) The mammalian endoplasmic reticulum-associated degradation system. Cold Spring Harbor perspectives in biology 5 61. Finley, D. (2009) Recognition and processing of ubiquitin-protein conjugates by the proteasome. Annual review of biochemistry 78, 477-513 62. Kerscher, O., Felberbaum, R., and Hochstrasser, M. (2006) Modification of proteins by ubiquitin and ubiquitin-like proteins. Annual review of cell and developmental biology 22, 159-180 63. Wong, J., Quinn, C. M., Gelissen, I. C., and Brown, A. J. (2008) Endogenous 24(S),25-epoxycholesterol fine-tunes acute control of cellular cholesterol homeostasis. The Journal of biological chemistry 283, 700-707 64. Song, B. L., and DeBose-Boyd, R. A. (2004) Ubiquitination of 3-hydroxy- 3-methylglutaryl-CoA reductase in permeabilized cells mediated by cytosolic E1 and a putative membrane-bound ubiquitin ligase. The Journal of biological chemistry 279, 28798-28806

150

65. Ono, T., and Bloch, K. (1975) Solubilization and partial characterization of rat liver squalene epoxidase. The Journal of biological chemistry 250, 1571-1579 66. Sakakibara, J., Watanabe, R., Kanai, Y., and Ono, T. (1995) Molecular cloning and expression of rat squalene epoxidase. The Journal of biological chemistry 270, 17-20 67. Yamamoto, S., and Bloch, K. (1970) Studies on squalene epoxidase of rat liver. The Journal of biological chemistry 245, 1670-1674 68. Tai, H. H., and Bloch, K. (1972) Squalene epoxidase of rat liver. The Journal of biological chemistry 247, 3767-3773 69. Li, L., and Porter, T. D. (2007) Hepatic cytochrome P450 reductase-null mice reveal a second microsomal reductase for squalene monooxygenase. Archives of biochemistry and biophysics 461, 76-84 70. Laden, B. P., Tang, Y., and Porter, T. D. (2000) Cloning, heterologous expression, and enzymological characterization of human squalene monooxygenase. Archives of biochemistry and biophysics 374, 381-388 71. Panagabko, C., Morley, S., Hernandez, M., Cassolato, P., Gordon, H., Parsons, R., Manor, D., and Atkinson, J. (2003) Ligand specificity in the CRAL-TRIO protein family. Biochemistry 42, 6467-6474 72. Shibata, N., Arita, M., Misaki, Y., Dohmae, N., Takio, K., Ono, T., Inoue, K., and Arai, H. (2001) Supernatant protein factor, which stimulates the conversion of squalene to lanosterol, is a cytosolic squalene transfer protein and enhances cholesterol biosynthesis. Proceedings of the National Academy of Sciences of the United States of America 98, 2244-2249 73. Ono, T. (2002) The first step of oxygenation in cholesterol biosynthesis. Biochemical and biophysical research communications 292, 1283-1288 74. Nagumo, A., Kamei, T., Sakakibara, J., and Ono, T. (1995) Purification and characterization of recombinant squalene epoxidase. Journal of lipid research 36, 1489-1497 75. Abe, I., Abe, T., Lou, W., Masuoka, T., and Noguchi, H. (2007) Site- directed mutagenesis of conserved aromatic residues in rat squalene epoxidase. Biochemical and biophysical research communications 352, 259-263 76. Lee, H. K., Denner-Ancona, P., Sakakibara, J., Ono, T., and Prestwich, G. D. (2000) Photoaffinity labeling and site-directed mutagenesis of rat squalene epoxidase. Archives of biochemistry and biophysics 381, 43-52 77. Lee, H. K., Zheng, Y. F., Xiao, X. Y., Bai, M., Sakakibara, J., Ono, T., and Prestwich, G. D. (2004) Photoaffinity labeling identifies the substrate- binding site of mammalian squalene epoxidase. Biochemical and biophysical research communications 315, 1-9 78. Chugh, A., Ray, A., and Gupta, J. B. (2003) Squalene epoxidase as hypocholesterolemic drug target revisited. Progress in lipid research 42, 37- 50 79. Horie, M., Sawasaki, Y., Fukuzumi, H., Watanabe, K., Iizuka, Y., Tsuchiya, Y., and Kamei, T. (1991) Hypolipidemic effects of NB-598 in dogs. Atherosclerosis 88, 183-192 80. Horie, M., Tsuchiya, Y., Hayashi, M., Iida, Y., Iwasawa, Y., Nagata, Y., Sawasaki, Y., Fukuzumi, H., Kitani, K., and Kamei, T. (1990) NB-598: a potent competitive inhibitor of squalene epoxidase. The Journal of biological chemistry 265, 18075-18078 81. Ryder, N. S. (1992) Terbinafine: mode of action and properties of the squalene epoxidase inhibition. The British journal of dermatology 126 Suppl 39, 2-7 151 82. Hidaka, Y., Satoh, T., and Kamei, T. (1990) Regulation of squalene epoxidase in HepG2 cells. Journal of lipid research 31, 2087-2094 83. Eilenberg, H., and Shechter, I. (1984) A possible regulatory role of squalene epoxidase in Chinese hamster ovary cells. Lipids 19, 539-543 84. Nagai, M., Sakakibara, J., Nakamura, Y., Gejyo, F., and Ono, T. (2002) SREBP-2 and NF-Y are involved in the transcriptional regulation of squalene epoxidase. Biochemical and biophysical research communications 295, 74-80 85. Donald, K. A., Hampton, R. Y., and Fritz, I. B. (1997) Effects of overproduction of the catalytic domain of 3-hydroxy-3-methylglutaryl coenzyme A reductase on squalene synthesis in Saccharomyces cerevisiae. Applied and environmental microbiology 63, 3341-3344 86. Mokashi, V., and Porter, T. D. (2005) Supernatant protein factor requires phosphorylation and interaction with Golgi to stimulate cholesterol synthesis in hepatoma cells. Archives of biochemistry and biophysics 435, 175-181 87. Mokashi, V., Singh, D. K., and Porter, T. D. (2004) Rat supernatant protein factor-like protein stimulates squalene monooxygenase and is activated by protein kinase A. Biochemical and biophysical research communications 316, 688-692 88. Eilenberg, H., and Shechter, I. (1987) Regulation of squalene epoxidase activity and comparison of catalytic properties of rat liver and Chinese hamster ovary cell-derived enzymes. Journal of lipid research 28, 1398-1404 89. Mokashi, V., Singh, D. K., and Porter, T. D. (2005) Supernatant protein factor stimulates HMG-CoA reductase in cell culture and in vitro. Archives of biochemistry and biophysics 433, 474-480 90. Singh, D. K., Mokashi, V., Elmore, C. L., and Porter, T. D. (2003) Phosphorylation of supernatant protein factor enhances its ability to stimulate microsomal squalene monooxygenase. The Journal of biological chemistry 278, 5646-5651 91. Klock, H. E., Koesema, E. J., Knuth, M. W., and Lesley, S. A. (2008) Combining the polymerase incomplete primer extension method for cloning and mutagenesis with microscreening to accelerate structural genomics efforts. Proteins 71, 982-994 92. Li, M. Z., and Elledge, S. J. (2007) Harnessing homologous recombination in vitro to generate recombinant DNA via SLIC. Nature methods 4, 251- 256 93. Bryksin, A. V., and Matsumura, I. (2010) Overlap extension PCR cloning: a simple and reliable way to create recombinant plasmids. BioTechniques 48, 463-465 94. Unger, T., Jacobovitch, Y., Dantes, A., Bernheim, R., and Peleg, Y. (2010) Applications of the Restriction Free (RF) cloning procedure for molecular manipulations and protein expression. Journal of structural biology 172, 34- 44 95. Goldstein, J. L., Basu, S. K., and Brown, M. S. (1983) Receptor-mediated endocytosis of low-density lipoprotein in cultured cells. Methods in enzymology 98, 241-260 96. Brown, A. J., Dean, R. T., and Jessup, W. (1996) Free and esterified oxysterol: formation during copper-oxidation of low density lipoprotein and uptake by macrophages. Journal of lipid research 37, 320-335 97. Brown, A. J., Sun, L., Feramisco, J. D., Brown, M. S., and Goldstein, J. L. (2002) Cholesterol addition to ER membranes alters conformation of

152

SCAP, the SREBP escort protein that regulates cholesterol metabolism. Molecular cell 10, 237-245 98. Treier, M., Staszewski, L. M., and Bohmann, D. (1994) Ubiquitin- dependent c-Jun degradation in vivo is mediated by the delta domain. Cell 78, 787-798 99. Wong, J., Quinn, C. M., and Brown, A. J. (2006) SREBP-2 positively regulates transcription of the cholesterol efflux gene, ABCA1, by generating oxysterol ligands for LXR. The Biochemical journal 400, 485-491 100. Zelcer, N., Sharpe, L. J., Loregger, A., Kristiana, I., Cook, E. C., Phan, L., Stevenson, J., and Brown, A. J. (2014) The E3 ubiquitin ligase MARCH6 degrades squalene monooxygenase (SM), and affects HMG-CoA reductase (HMGCR) and the cholesterol synthesis pathway. Molecular and cellular biology 101. Kristiana, I., Sharpe, L. J., Catts, V. S., Lutze-Mann, L. H., and Brown, A. J. (2010) Antipsychotic drugs upregulate lipogenic gene expression by disrupting intracellular trafficking of lipoprotein-derived cholesterol. The pharmacogenomics journal 10, 396-407 102. Luu, W. (2013) The role of signalling and sensing in cellular cholesterol homeostasis PhD thesis, UNSW 103. Jeong, J. Y., Yim, H. S., Ryu, J. Y., Lee, H. S., Lee, J. H., Seen, D. S., and Kang, S. G. (2012) One-step sequence- and ligation-independent cloning as a rapid and versatile cloning method for functional genomics studies. Applied and environmental microbiology 78, 5440-5443 104. Jenster, G., Spencer, T. E., Burcin, M. M., Tsai, S. Y., Tsai, M. J., and O'Malley, B. W. (1997) Steroid receptor induction of gene transcription: a two-step model. Proceedings of the National Academy of Sciences of the United States of America 94, 7879-7884 105. Yabe, D., Brown, M. S., and Goldstein, J. L. (2002) Insig-2, a second endoplasmic reticulum protein that binds SCAP and blocks export of sterol regulatory element-binding proteins. Proceedings of the National Academy of Sciences of the United States of America 99, 12753-12758 106. Thieme, F., Engler, C., Kandzia, R., and Marillonnet, S. (2011) Quick and clean cloning: a ligation-independent cloning strategy for selective cloning of specific PCR products from non-specific mixes. PloS one 6, e20556 107. Luu, W., Sharpe, L. J., Stevenson, J., and Brown, A. J. (2012) Akt acutely activates the cholesterogenic transcription factor SREBP-2. Biochimica et biophysica acta 1823, 458-464 108. Tillett, D., and Neilan, B. A. (1999) Enzyme-free cloning: a rapid method to clone PCR products independent of vector restriction enzyme sites. Nucleic acids research 27, e26 109. Gibson, D. G., Young, L., Chuang, R. Y., Venter, J. C., Hutchison, C. A., 3rd, and Smith, H. O. (2009) Enzymatic assembly of DNA molecules up to several hundred kilobases. Nature methods 6, 343-345 110. Gill, S. (2011) Squalene monooxygenase: A novel control point in cholesterol synthesis UNSW 111. Horton, J. D., Shah, N. A., Warrington, J. A., Anderson, N. N., Park, S. W., Brown, M. S., and Goldstein, J. L. (2003) Combined analysis of oligonucleotide microarray data from transgenic and knockout mice identifies direct SREBP target genes. Proceedings of the National Academy of Sciences of the United States of America 100, 12027-12032

153 112. Gonzalez, R., Carlson, J. P., and Dempsey, M. E. (1979) Two major regulatory steps in cholesterol synthesis by human renal cancer cells. Archives of biochemistry and biophysics 196, 574-580 113. Brown, M. S., and Goldstein, J. L. (1980) Multivalent feedback regulation of HMG CoA reductase, a control mechanism coordinating isoprenoid synthesis and cell growth. Journal of lipid research 21, 505-517 114. Raskin, P., and Siperstein, M. D. (1974) Mevalonate metabolism by renal tissue in vitro. Journal of lipid research 15, 20-25 115. Sanchis, J., Fernandez, L., Carballeira, J. D., Drone, J., Gumulya, Y., Hobenreich, H., Kahakeaw, D., Kille, S., Lohmer, R., Peyralans, J. J., Podtetenieff, J., Prasad, S., Soni, P., Taglieber, A., Wu, S., Zilly, F. E., and Reetz, M. T. (2008) Improved PCR method for the creation of saturation mutagenesis libraries in directed evolution: application to difficult-to- amplify templates. Applied microbiology and biotechnology 81, 387-397 116. Lim, K. L., Chew, K. C., Tan, J. M., Wang, C., Chung, K. K., Zhang, Y., Tanaka, Y., Smith, W., Engelender, S., Ross, C. A., Dawson, V. L., and Dawson, T. M. (2005) Parkin mediates nonclassical, proteasomal- independent ubiquitination of synphilin-1: implications for Lewy body formation. The Journal of neuroscience : the official journal of the Society for Neuroscience 25, 2002-2009 117. Feramisco, J. D., Goldstein, J. L., and Brown, M. S. (2004) Membrane topology of human insig-1, a protein regulator of lipid synthesis. The Journal of biological chemistry 279, 8487-8496 118. Metherall, J. E., Goldstein, J. L., Luskey, K. L., and Brown, M. S. (1989) Loss of transcriptional repression of three sterol-regulated genes in mutant hamster cells. The Journal of biological chemistry 264, 15634-15641 119. Yang, J., Sato, R., Goldstein, J. L., and Brown, M. S. (1994) Sterol-resistant transcription in CHO cells caused by gene rearrangement that truncates SREBP-2. Genes & development 8, 1910-1919 120. Hampton, R. Y. (2002) ER-associated degradation in protein quality control and cellular regulation. Current opinion in cell biology 14, 476-482 121. Sharipo, A., Imreh, M., Leonchiks, A., Imreh, S., and Masucci, M. G. (1998) A minimal glycine-alanine repeat prevents the interaction of ubiquitinated I kappaB alpha with the proteasome: a new mechanism for selective inhibition of proteolysis. Nature medicine 4, 939-944 122. Lee, P. C., Sever, N., and Debose-Boyd, R. A. (2005) Isolation of sterol- resistant Chinese hamster ovary cells with genetic deficiencies in both Insig-1 and Insig-2. The Journal of biological chemistry 280, 25242-25249 123. Rawson, R. B., DeBose-Boyd, R., Goldstein, J. L., and Brown, M. S. (1999) Failure to cleave sterol regulatory element-binding proteins (SREBPs) causes cholesterol auxotrophy in Chinese hamster ovary cells with genetic absence of SREBP cleavage-activating protein. The Journal of biological chemistry 274, 28549-28556 124. Ward, C. L., Omura, S., and Kopito, R. R. (1995) Degradation of CFTR by the ubiquitin-proteasome pathway. Cell 83, 121-127 125. Sever, N., Song, B. L., Yabe, D., Goldstein, J. L., Brown, M. S., and DeBose-Boyd, R. A. (2003) Insig-dependent ubiquitination and degradation of mammalian 3-hydroxy-3-methylglutaryl-CoA reductase stimulated by sterols and geranylgeraniol. The Journal of biological chemistry 278, 52479-52490 126. Yang, T., Espenshade, P. J., Wright, M. E., Yabe, D., Gong, Y., Aebersold, R., Goldstein, J. L., and Brown, M. S. (2002) Crucial step in cholesterol

154

homeostasis: sterols promote binding of SCAP to INSIG-1, a membrane protein that facilitates retention of SREBPs in ER. Cell 110, 489-500 127. Epand, R. M. (2006) Cholesterol and the interaction of proteins with membrane domains. Progress in lipid research 45, 279-294 128. Chun, K. T., Bar-Nun, S., and Simoni, R. D. (1990) The regulated degradation of 3-hydroxy-3-methylglutaryl-CoA reductase requires a short-lived protein and occurs in the endoplasmic reticulum. The Journal of biological chemistry 265, 22004-22010 129. Jo, Y., Lee, P. C., Sguigna, P. V., and DeBose-Boyd, R. A. (2011) Sterol- induced degradation of HMG CoA reductase depends on interplay of two Insigs and two ubiquitin ligases, gp78 and Trc8. Proceedings of the National Academy of Sciences of the United States of America 108, 20503- 20508 130. Huang, Y., He, S., Li, J. Z., Seo, Y. K., Osborne, T. F., Cohen, J. C., and Hobbs, H. H. (2010) A feed-forward loop amplifies nutritional regulation of PNPLA3. Proceedings of the National Academy of Sciences of the United States of America 107, 7892-7897 131. Nian, Z., Sun, Z., Yu, L., Toh, S. Y., Sang, J., and Li, P. (2010) Fat-specific protein 27 undergoes ubiquitin-dependent degradation regulated by triacylglycerol synthesis and lipid droplet formation. The Journal of biological chemistry 285, 9604-9615 132. Xu, G., Sztalryd, C., Lu, X., Tansey, J. T., Gan, J., Dorward, H., Kimmel, A. R., and Londos, C. (2005) Post-translational regulation of adipose differentiation-related protein by the ubiquitin/proteasome pathway. The Journal of biological chemistry 280, 42841-42847 133. Brasaemle, D. L., Barber, T., Kimmel, A. R., and Londos, C. (1997) Post- translational regulation of perilipin expression. Stabilization by stored intracellular neutral lipids. The Journal of biological chemistry 272, 9378- 9387 134. Xu, G., Sztalryd, C., and Londos, C. (2006) Degradation of perilipin is mediated through ubiquitination-proteasome pathway. Biochimica et biophysica acta 1761, 83-90 135. Lee, J. N., Zhang, X., Feramisco, J. D., Gong, Y., and Ye, J. (2008) Unsaturated fatty acids inhibit proteasomal degradation of Insig-1 at a postubiquitination step. The Journal of biological chemistry 283, 33772-33783 136. Christianson, J. C., Olzmann, J. A., Shaler, T. A., Sowa, M. E., Bennett, E. J., Richter, C. M., Tyler, R. E., Greenblatt, E. J., Harper, J. W., and Kopito, R. R. (2012) Defining human ERAD networks through an integrative mapping strategy. Nature cell biology 14, 93-105 137. Lee, J. N., Kim, H., Yao, H., Chen, Y., Weng, K., and Ye, J. (2010) Identification of Ubxd8 protein as a sensor for unsaturated fatty acids and regulator of triglyceride synthesis. Proceedings of the National Academy of Sciences of the United States of America 107, 21424-21429 138. Wang, Y., and Oram, J. F. (2007) Unsaturated fatty acids phosphorylate and destabilize ABCA1 through a protein kinase C delta pathway. Journal of lipid research 48, 1062-1068 139. Listenberger, L. L., and Brown, D. A. (2007) Fluorescent detection of lipid droplets and associated proteins. Current protocols in cell biology / editorial board, Juan S. Bonifacino ... [et al.] Chapter 24, Unit 24 22 140. Sud, M., Fahy, E., Cotter, D., Brown, A., Dennis, E. A., Glass, C. K., Merrill, A. H., Jr., Murphy, R. C., Raetz, C. R., Russell, D. W., and

155 Subramaniam, S. (2007) LMSD: LIPID MAPS structure database. Nucleic acids research 35, D527-532 141. Brown, A. J., and Jessup, W. (1999) Oxysterols and atherosclerosis. Atherosclerosis 142, 1-28 142. Brown, A. J., Leong, S. L., Dean, R. T., and Jessup, W. (1997) 7- Hydroperoxycholesterol and its products in oxidized low density lipoprotein and human atherosclerotic plaque. Journal of lipid research 38, 1730-1745 143. Du, X., Pham, Y. H., and Brown, A. J. (2004) Effects of 25- hydroxycholesterol on cholesterol esterification and sterol regulatory element-binding protein processing are dissociable: implications for cholesterol movement to the regulatory pool in the endoplasmic reticulum. The Journal of biological chemistry 279, 47010-47016 144. Chang, T. Y., Chang, C. C., Ohgami, N., and Yamauchi, Y. (2006) Cholesterol sensing, trafficking, and esterification. Annual review of cell and developmental biology 22, 129-157 145. Hulce, J. J., Cognetta, A. B., Niphakis, M. J., Tully, S. E., and Cravatt, B. F. (2013) Proteome-wide mapping of cholesterol-interacting proteins in mammalian cells. Nature methods 10, 259-264 146. Tansey, J. T., Huml, A. M., Vogt, R., Davis, K. E., Jones, J. M., Fraser, K. A., Brasaemle, D. L., Kimmel, A. R., and Londos, C. (2003) Functional studies on native and mutated forms of perilipins. A role in protein kinase A-mediated lipolysis of triacylglycerols. The Journal of biological chemistry 278, 8401-8406 147. Igal, R. A., Wang, P., and Coleman, R. A. (1997) Triacsin C blocks de novo synthesis of glycerolipids and cholesterol esters but not recycling of fatty acid into phospholipid: evidence for functionally separate pools of acyl-CoA. The Biochemical journal 324 ( Pt 2), 529-534 148. Liu, P., Ying, Y., Zhao, Y., Mundy, D. I., Zhu, M., and Anderson, R. G. (2004) Chinese hamster ovary K2 cell lipid droplets appear to be metabolic organelles involved in membrane traffic. The Journal of biological chemistry 279, 3787-3792 149. Umlauf, E., Csaszar, E., Moertelmaier, M., Schuetz, G. J., Parton, R. G., and Prohaska, R. (2004) Association of stomatin with lipid bodies. The Journal of biological chemistry 279, 23699-23709 150. Harris, C. A., Haas, J. T., Streeper, R. S., Stone, S. J., Kumari, M., Yang, K., Han, X., Brownell, N., Gross, R. W., Zechner, R., and Farese, R. V., Jr. (2011) DGAT enzymes are required for triacylglycerol synthesis and lipid droplets in adipocytes. Journal of lipid research 52, 657-667 151. Florin-Christensen, J., Florin-Christensen, M., Delfino, J. M., Stegmann, T., and Rasmussen, H. (1992) Metabolic fate of plasma membrane diacylglycerols in NIH 3T3 fibroblasts. The Journal of biological chemistry 267, 14783-14789 152. Canuto, R. A., Biocca, M. E., Muzio, G., and Dianzani, M. U. (1989) Fatty acid composition of phospholipids in mitochondria and microsomes during diethylnitrosamine carcinogenesis in rat liver. Cell biochemistry and function 7, 11-19 153. Zambrano, F., Fleischer, S., and Fleischer, B. (1975) Lipid composition of the Golgi apparatus of rat kidney and liver in comparison with other subcellular organelles. Biochimica et biophysica acta 380, 357-369 154. Shinomura, T., Asaoka, Y., Oka, M., Yoshida, K., and Nishizuka, Y. (1991) Synergistic action of diacylglycerol and unsaturated fatty acid for

156

protein kinase C activation: its possible implications. Proceedings of the National Academy of Sciences of the United States of America 88, 5149-5153 155. Gong, Y., Lee, J. N., Lee, P. C., Goldstein, J. L., Brown, M. S., and Ye, J. (2006) Sterol-regulated ubiquitination and degradation of Insig-1 creates a convergent mechanism for feedback control of cholesterol synthesis and uptake. Cell metabolism 3, 15-24 156. Foresti, O., Ruggiano, A., Hannibal-Bach, H. K., Ejsing, C. S., and Carvalho, P. (2013) Sterol homeostasis requires regulated degradation of squalene monooxygenase by the ubiquitin ligase Doa10/Teb4. eLife 2, e00953 157. Rodriguez-Acebes, S., de la Cueva, P., Fernandez-Hernando, C., Ferruelo, A. J., Lasuncion, M. A., Rawson, R. B., Martinez-Botas, J., and Gomez-Coronado, D. (2009) Desmosterol can replace cholesterol in sustaining cell proliferation and regulating the SREBP pathway in a sterol-Delta24-reductase-deficient cell line. The Biochemical journal 420, 305-315 158. Esko, J. D., Wermuth, M. M., and Raetz, C. R. (1981) Thermolabile CDP- choline synthetase in an animal cell mutant defective in lecithin formation. The Journal of biological chemistry 256, 7388-7393 159. Dobrosotskaya, I. Y., Seegmiller, A. C., Brown, M. S., Goldstein, J. L., and Rawson, R. B. (2002) Regulation of SREBP processing and membrane lipid production by phospholipids in Drosophila. Science 296, 879-883 160. Igal, R. A., Caviglia, J. M., de Gomez Dumm, I. N., and Coleman, R. A. (2001) Diacylglycerol generated in CHO cell plasma membrane by phospholipase C is used for triacylglycerol synthesis. Journal of lipid research 42, 88-95 161. Tseng, W. C., Lin, J. W., Wei, T. Y., and Fang, T. Y. (2008) A novel megaprimed and ligase-free, PCR-based, site-directed mutagenesis method. Analytical biochemistry 375, 376-378 162. Mead, D. A., Pey, N. K., Herrnstadt, C., Marcil, R. A., and Smith, L. M. (1991) A universal method for the direct cloning of PCR amplified nucleic acid. Bio/technology 9, 657-663 163. Fitzky, B. U., Moebius, F. F., Asaoka, H., Waage-Baudet, H., Xu, L., Xu, G., Maeda, N., Kluckman, K., Hiller, S., Yu, H., Batta, A. K., Shefer, S., Chen, T., Salen, G., Sulik, K., Simoni, R. D., Ness, G. C., Glossmann, H., Patel, S. B., and Tint, G. S. (2001) 7-Dehydrocholesterol-dependent proteolysis of HMG-CoA reductase suppresses sterol biosynthesis in a mouse model of Smith-Lemli-Opitz/RSH syndrome. The Journal of clinical investigation 108, 905-915 164. Ploegh, H. L. (2007) A lipid-based model for the creation of an escape hatch from the endoplasmic reticulum. Nature 448, 435-438 165. Olzmann, J. A., and Kopito, R. R. (2011) Lipid droplet formation is dispensable for endoplasmic reticulum-associated degradation. The Journal of biological chemistry 286, 27872-27874 166. Corish, P., and Tyler-Smith, C. (1999) Attenuation of green fluorescent protein half-life in mammalian cells. Protein engineering 12, 1035-1040 167. Arora, A., and Tamm, L. K. (2001) Biophysical approaches to membrane protein structure determination. Current opinion in structural biology 11, 540-547 168. Hassink, G., Kikkert, M., van Voorden, S., Lee, S. J., Spaapen, R., van Laar, T., Coleman, C. S., Bartee, E., Fruh, K., Chau, V., and Wiertz, E.

157 (2005) TEB4 is a C4HC3 RING finger-containing ubiquitin ligase of the endoplasmic reticulum. The Biochemical journal 388, 647-655 169. Zavacki, A. M., Arrojo, E. D. R., Freitas, B. C., Chung, M., Harney, J. W., Egri, P., Wittmann, G., Fekete, C., Gereben, B., and Bianco, A. C. (2009) The E3 ubiquitin ligase TEB4 mediates degradation of type 2 iodothyronine deiodinase. Molecular and cellular biology 29, 5339-5347 170. Wagner, S. A., Beli, P., Weinert, B. T., Nielsen, M. L., Cox, J., Mann, M., and Choudhary, C. (2011) A proteome-wide, quantitative survey of in vivo ubiquitylation sites reveals widespread regulatory roles. Molecular & cellular proteomics : MCP 10, M111 013284 171. Wagner, S. A., Beli, P., Weinert, B. T., Scholz, C., Kelstrup, C. D., Young, C., Nielsen, M. L., Olsen, J. V., Brakebusch, C., and Choudhary, C. (2012) Proteomic analyses reveal divergent ubiquitylation site patterns in murine tissues. Molecular & cellular proteomics : MCP 11, 1578-1585 172. Kim, W., Bennett, E. J., Huttlin, E. L., Guo, A., Li, J., Possemato, A., Sowa, M. E., Rad, R., Rush, J., Comb, M. J., Harper, J. W., and Gygi, S. P. (2011) Systematic and quantitative assessment of the ubiquitin-modified proteome. Molecular cell 44, 325-340 173. Tsai, Y. C., Leichner, G. S., Pearce, M. M., Wilson, G. L., Wojcikiewicz, R. J., Roitelman, J., and Weissman, A. M. (2012) Differential regulation of HMG-CoA reductase and Insig-1 by enzymes of the ubiquitin- proteasome system. Molecular biology of the cell 23, 4484-4494 174. Shearer, A. G., and Hampton, R. Y. (2005) Lipid-mediated, reversible misfolding of a sterol-sensing domain protein. The EMBO journal 24, 149- 159 175. Brown, C. R., Hong-Brown, L. Q., Biwersi, J., Verkman, A. S., and Welch, W. J. (1996) Chemical chaperones correct the mutant phenotype of the delta F508 cystic fibrosis transmembrane conductance regulator protein. Cell stress & chaperones 1, 117-125 176. Lipson, C., Alalouf, G., Bajorek, M., Rabinovich, E., Atir-Lande, A., Glickman, M., and Bar-Nun, S. (2008) A proteasomal ATPase contributes to dislocation of endoplasmic reticulum-associated degradation (ERAD) substrates. The Journal of biological chemistry 283, 7166-7175 177. Wahlman, J., DeMartino, G. N., Skach, W. R., Bulleid, N. J., Brodsky, J. L., and Johnson, A. E. (2007) Real-time fluorescence detection of ERAD substrate retrotranslocation in a mammalian in vitro system. Cell 129, 943-955 178. Lilley, B. N., and Ploegh, H. L. (2004) A membrane protein required for dislocation of misfolded proteins from the ER. Nature 429, 834-840 179. Greenblatt, E. J., Olzmann, J. A., and Kopito, R. R. (2011) Derlin-1 is a rhomboid pseudoprotease required for the dislocation of mutant alpha-1 antitrypsin from the endoplasmic reticulum. Nature structural & molecular biology 18, 1147-1152 180. Rajaratnam, R. A., Gylling, H., and Miettinen, T. A. (1999) Serum squalene in postmenopausal women without and with coronary artery disease. Atherosclerosis 146, 61-64 181. Wong, J., Quinn, C. M., and Brown, A. J. (2007) Synthesis of the oxysterol, 24(S), 25-epoxycholesterol, parallels cholesterol production and may protect against cellular accumulation of newly-synthesized cholesterol. Lipids in health and disease 6, 10 182. Zerenturk, E. J., Kristiana, I., Gill, S., and Brown, A. J. (2012) The endogenous regulator 24(S),25-epoxycholesterol inhibits cholesterol

158

synthesis at DHCR24 (Seladin-1). Biochimica et biophysica acta 1821, 1269- 1277 183. Spann, N. J., Garmire, L. X., McDonald, J. G., Myers, D. S., Milne, S. B., Shibata, N., Reichart, D., Fox, J. N., Shaked, I., Heudobler, D., Raetz, C. R., Wang, E. W., Kelly, S. L., Sullards, M. C., Murphy, R. C., Merrill, A. H., Jr., Brown, H. A., Dennis, E. A., Li, A. C., Ley, K., Tsimikas, S., Fahy, E., Subramaniam, S., Quehenberger, O., Russell, D. W., and Glass, C. K. (2012) Regulated accumulation of desmosterol integrates macrophage lipid metabolism and inflammatory responses. Cell 151, 138-152 184. Gill, S., Stevenson, J., Kristiana, I., and Brown, A. J. (2011) Cholesterol- dependent degradation of squalene monooxygenase, a control point in cholesterol synthesis beyond HMG-CoA reductase. Cell metabolism 13, 260-273 185. Thomas, S., and Fell, D. A. (1998) The role of multiple enzyme activation in metabolic flux control. Advances in enzyme regulation 38, 65-85

159