xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001914233
007 cr mnu|||uuuuu
008 071019s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001739
Lagor, William Raymond.
Occupancy and function of the hepatic HMG-CoA reductase promoter
h [electronic resource] /
by William Raymond Lagor.
[Tampa, Fla] :
b University of South Florida,
ABSTRACT: HMG-CoA reductase (HMGR) catalyzes the rate controlling step in cholesterol production. This enzyme is highly expressed in the liver where it is subject to extensive hormonal and dietary regulation. This study was undertaken to examine the occupancy and function of the hepatic HMGR promoter in regards to insulin and sterol regulation. HMGR protein and mRNA are substantially decreased in diabetic animals and rapidly restored by administration of insulin. Nuclear run-on assays revealed that HMGR transcription was substantially reduced in the diabetic rats, and fully restored within two hours after insulin treatment. In vivo footprinting revealed several areas of protein binding as shown by dimethyl sulfate protection or enhancement. The CRE was heavily protected in all conditions including diabetes, cholesterol feeding, or statin treatment. Striking enhancements in footprints from diabetic animals were observed at -142 and at -161 (in the SRE). Protections at a newly ident ified NF-Y site at -70/-71 were seen in normal animals, and not in diabetics. This proximal NF-Y site was found to be required for efficient HMGR transcription. CREB-1 was able to bind the HMGR CRE in vitro, and to the promoter in vivo. The data supports an essential role for CREB in transcription of hepatic HMGR, and identifies at least two sites where in vivo occupancy is regulated by insulin. The technique of in vivo electroporation was utilized to perform the first functional analysis of the HMGR promoter in live animals. Analysis of a series of deletion constructs showed that deletion of the region containing the cyclic AMP response element (CRE) at -104 to -96 and the newly identified NF-Y site at -70 resulted in marked reduction of promoter activity. Possible sterol regulation of the promoter was investigated by raising tissue cholesterol levels by feeding cholesterol, or by inhibiting cholesterol synthesis with a statin (lovastatin). It was found that HMGR promoter constructs r esponded to lovastatin, in agreement with previous findings in cultured cells. This work sheds light on the regulation of the HMGR promoter in the liver, whose expression is a key determinant of serum cholesterol levels- a major risk factor for cardiovascular disease.
Dissertation (Ph.D.)--University of South Florida, 2006.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 137 pages.
Adviser: Gene C. Ness, Ph.D.
In vivo footprinting.
In vivo electroporation.
x Biochemistry and Molecular Biology
t USF Electronic Theses and Dissertations.
Occupancy and Function of the Hepatic HMG-CoA Reductase Promoter by William Raymond Lagor A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Biochemistry and Molecular Biology College of Medicine University of South Florida Major Professor: Gene C. Ness, Ph.D. George Blanck, Ph.D. Kenneth Wright, Ph.D. Ronald K. Keller, Ph.D., Larry P. Solomonson, Ph.D. Richard Hanson, Ph.D. Date of Approval: October 23, 2006 Keywords: In vivo footprinting, in vivo electroporation, insulin, cholesterol, lovastatin, rat, transcription, liver Copyright 2006, William Raymond Lagor
ACKNOWLEDGEMENTS I sincerely thank Dr. Gene C. Ness for his expert guidance, and financial support of the research contained in this dissertation. This work would not have been possible without his friendship, remarkable patience, and steadfast encouragement. Considerable credit should be given to Eric D. deGroh, who cloned the HMGR promoter-luciferase plas mids (-325, -228, -176, -123, and -58) used in this work. This greatly accele rated our progress, and I sincerely thank him for his effort. I am also grateful to Dr. Aar on Osborne for instruction in molecular biology, Dr. Dayami Lopez for help with nuclear run-on assays, Dr. Kenneth Wright, Jenny Wu and Sophie Bolick for advice on in vivo footprinting, Dr. Keller for constructive criticism and scientific discussion, and Dr. Heller for his expertise and technical assistance with the in vivo electroporation studies. Lindsey Jackson, Brittany Doupnik, Reed Holland, Mohammad Hassanyer, Nicolle Rodriguez, Jose Abisambra, and Dr. Veronica Pollock all contributed to this research as members of our laborator y team. I also thank my mother, father, brothers, sisters, and ex tended family for their unwavering support. I am particularly grateful to my beautiful wife Jamie, who is my true inspiration.
i TABLE OF CONTENTS LIST OF TABLES iv LIST OF FIGURES v LIST OFABBREVIATIONS vii ABSTRACT x INTRODUCTION 1 MATERIALS AND METHODS 21 Plasmids 21 Growth of Bacteria and Isolation of Plasmid DNA 24 Animals 25 Nuclei Isolation 25 Nuclear run-on assay 26 Serum Cholesterol and Glucose 27 Cell Culture 27 DMS treatment of H4IIE cells 29 DMS treatment of rat liver nuclei 29
ii DMS treatment of rat liver 30 DNA extraction 30 Preparation of control DNA and piperidine treatment 32 Ligation-Mediated PCR 33 Identification of protections and enhancements 36 Compilation of in vivo footprinting data 37 Nuclear extract 37 EMSA 38 Transient transfections 39 Chromatin preparation from rat liver 40 Chromatin Immunoprecipitation assays 41 Surgery 44 Electroporation 44 Luciferase Assays 45 Treatment of luciferase data 45 Isolation of Microsomes 47 Western Blotting 47 RNA Isolation and cDNA synthesis 48 Real-Time PCR 49 RESULTS 50 Effect of diabetes and insulin on HMGR transcription in the liver 51
iii Effects of insulin on HMGR mRNA levels, serum cholesterol, and serum glucose with time. 53 In vivo footprinting of the HMGR promoter in H4IIE cells 56 Footprinting of the HMGR promoter from rat liver nuclei 59 In vivo footprinting of the HMGR promoter in rat liver 61 In vitro identification of transcr iption factors binding to the footprinted regions of the HMGR promoter 69 Chromatin immunoprecipitation of CREB-1 bound to the HMGR promoter 75 Effect of the -70/-65 NF-Y site on transcription in H4IIE cells 77 HMGR promoter function in normal rats 80 Importance of the -70/-65 NF-Y site in driving transcription in the liver 83 Mapping of HMGR promoter sterol response by in vivo electroporation 86 The effect of lovastatin and cholesterol feeding on HMGR expression 87 DISCUSSSION 89 REFERENCES 103 APPENDIX 116 ABOUT THE AUTHOR End Page
iv LIST OF TABLES Table 1 Primer sequences used in experiments 23 Table 2 DNA ol igos sequences used as probes 40 in EMSA experiments
v LIST OF FIGURES Fig. 1 The SREBP pathway 10 Fig. 2 Sequence of the rat HMG-CoA reductase promoter 17 Fig. 3 Overview of in vivo genomic footprinting 29 Fig. 4 Overview of in vivo electroporation 44 Fig. 5. Measurement of HM G-CoA reductase transcription in normal and diabetic rats by nuclear run-on 52 Fig. 6 Time course of HMG-CoA reductase mRNA activation by insulin 54 Fig. 7 In vivo footprinting of the HMG-CoA reductase promoter in H4IIE cells 57 Fig. 8 Footprinting of the HMG-CoA reductase promoter from rat liver nuclei 60 Fig. 9 In vivo footprint of the hepatic HMGR promoter, bottom strand 62 Fig. 10 In vivo footprint of the hepatic HMGR promoter, top strand 64 Fig. 11 In vivo footprinting of the HMGR promoter in rats fed lovastatin or cholesterol 66 Fig. 12 Summary of in vivo DMS reactivity for the hepatic HMG-CoA reductase promoter 68 Fig. 13 EMSA analysis of f ootprinted regions of the HMG-CoA reductase promoter 70 Fig. 14 EMSA analysis of the promoter region from -152 to -119 71
vi Fig. 15 EMSA analysis of S p1-like factor binding to the -142/-119 region 73 Fig. 16 EMSA analysis of the CRE and the -70/-65 NF-Y site 74 Fig. 17 CREB is bound to the HMGR promoter in live animals 76 Fig. 18 Effect of mutating the proximal NF-Y site on transcription in H4IIE cells 78 Fig. 19 Schem atic of HMG-CoA reductase promoterluciferase constructs 79 Fig. 20 Functional mapping of the HMGR promoter in the livers of normal rats 81 Fig. 21 Effect of mutating the proximal NF-Y site on HMG-CoA reductase promoter activity in live animals 84 Fig. 22 Response of the hepatic HMG-CoA reductase promoter to lovastatin or cholesterol feeding 87 Fig. 23 Effect of lovastat in or cholesterol feeding on HMG-CoA reductase immunoreactive protein levels 88 Fig. 24 Effect of lovastatin or cholesterol feeding on HMGR mRNA levels 89 Fig. 25 Effect of lovastatin or cholesterol feeding on HMG-CoA synthase mRNA levels 90 Fig. 26. Model of the HMGR promoter in normal and diabetic rat liver 96
vii LIST OF ABBREVIATIONS ABCA1 ATP binding cassette transporter A 1 ABCG1 ATP binding cassette transporter G 1 ABCG5/8 ATP binding cassette transporters G 1 and G8 ALLN N -acetyl-leucyl-leucyl-norleucynal Apoapolipoprotein ATP adenosine triphosphate BCA bicinchoninic acid BLUE bluescript bp base pairs BSA bovine serum albumin CAT catalase CETP cholesteryl ester transfer protein ChIP chromatin immunoprecipitation CHO Chinese hamster ovary cells cm centimeter CRE cyclic AMP response element CREB-1 cyclic AMP response element binding protein CVD cardiovascular disease
viii DMS dimethyl sulfate dpm detections per minute DTT dithiolthreitol ECL enhanced chemiluminescence EDTA ethylenedia minetetraacetic acid EGTA ethylene glycol-bis [ -aminomethyl-etherN,N,N,Ntetraacetic acid] EMEM Eagles modification of essential media EMSA electrophoretic mobility shift assay ER endoplasmic reticulum HCS HMG-CoA Synthase HMG-CoA 3-hydroxy-3-methylglutaryl coenzyme A HMGR HMG-CoA reductase HDL high density lipoprotein HEPES N-[2-hydroxyethyl]pipe razine-N-[2ethanesulfonic acid] HPLC high performance liquid chromatography IDL intermediate density lipoprotein Insig insulin induced gene product LCAT lecithin: cholesterol acyl transferase LDL low density lipoprotein LDLR low density lipoprotein receptor LRP LDL receptor related protein LRH-1/FTF liver receptor homologue-1 / fetoprotein transcription factor
ix mRNA messenger ribonucleic acid PCR polymerase chain reaction PPAR-alpha peroxisome prolifer ator activator receptoralpha rATP ribosyl adenosine triphosphate SRE sterol response element NF-Y nuclear factorY NPCL1 Niemann Pick C like protein 1 Ox-LDL oxidized LDL PBS phosphate buffered saline PMSF phenylmethylsulfonyl fluoride RT-PCR reverse transcriptase polymerase chain reaction SCAP SREBP cleavage accessory protein SDS sodium dodecyl sulfate SRB-I scavenger receptor B one SRE sterol response element SREBP sterol response element binding protein TB1 thyroid receptor beta 1 TBE Tris borate EDTA TEMED N,N,N,N-Tetramethylethylenediamine Tris Tris [hydroxymethyl]aminomethanehydrochloride tRNA transferribonucleic acid VLDL very low density lipoprotein
x OCCUPANCY AND FUNCTIO N OF THE HEPATIC HMG-COA REDUCTASE PROMOTER William Raymond Lagor ABSTRACT HMG-CoA reductase (HMGR) catalyzes the rate controlling step in cholesterol production. This enzyme is highl y expressed in the liver where it is subject to extensive hormonal and dietary regulation. This study was undertaken to examine the occupancy and function of the hepatic HMGR promoter in regards to insulin and sterol regulation. HMGR protein and mRNA are substantially decreased in diabetic animals and rapidly restored by administration of insulin. Nuclear run-on assays revealed that HMGR transcription was substantially reduced in the diabetic rats, and fully re stored within two hours after insulin treatment. In vivo footprinting revealed several areas of protein binding as shown by dimethyl sulfate protection or e nhancement. The CRE was heavily protected in all conditions including diabetes, c holesterol feeding, or statin treatment. Striking enhancements in footprints from diabetic animals were observed at -142 and at -161 (in the SRE). Protections at a newly identified NF-Y site at -70/-71 were seen in normal animals, and not in diabetics. This proximal NF-Y site was
xi found to be required for efficient HMGR transcription. CREB-1 was able to bind the HMGR CRE in vitro and to the promoter in vivo The data supports an essential role for CREB in transcription of hepatic HMGR, and identifies at least two sites where in vivo occupancy is regulated by insulin. The technique of in vivo electroporation was utilized to perform the first functional analysis of the HMGR promoter in live animals. Analysis of a series of deletion constructs showed that deletion of the region c ontaining the cyclic AM P response element (CRE) at -104 to -96 and the newly identif ied NF-Y site at -70 resulted in marked reduction of promoter activity. Possible sterol regulation of the promoter was investigated by raising tissue cholesterol levels by feeding cholesterol, or by inhibiting cholesterol synthesis with a st atin (lovastatin). It was found that HMGR promoter constructs responded to lovastatin, in agreement with previous findings in cultured cells. This work sheds light on the regulation of the HMGR promoter in the liver, whose expression is a key deter minant of serum cholesterol levelsa major risk factor for card iovascular disease.
1 INTRODUCTION Cholesterol is an essential component of mammalian cells. In addition to its role in maintaining membrane structure, it is also required for the production of bile acids, steroid hormones, (1) and in at least one case, serves as a posttranslational modification (2). Newly synthesized cholesterol is essential for development of tissues as well as proper organ function. The cell expends a great deal of energy in the production of chol esterol. In contrast to fatty acids, proteins, and carbohydrates, there is no system for the regeneration of energy from the cholesterol molecu le, or breakdown of its ca rbon skeleton. Because of this, the body has evolved an intricate system to ensure that this valuable molecule is not lost. Cholesterol is pa ckaged into lipoprotein particles in the bloodstream and is then carried throughout the body. Maintaining an ample supply of cholesterol is crucial as even minor defects can have drastic consequences (3). The great irony is that the same lipoprotein particles which nourish the cells with cholesterol and other lipids, can also make a significant contribution to human disease. Elevated serum cholesterol is a ma jor risk factor for cardiovascular disease (CVD) (4), which continues to be the primary cause of death in the US
2 and most developed countries. While there is still much to learn about the role of cholesterol in the pathogenesis of CVD, a great deal is already known. There are two sources of cholesteroldietary cholesterol, and newly synthesized cholesterol. Dietary cholesterol enters the body through the intestine. This process is dependent on Niemann-Pick C1-lik e 1 protein (NPCL1) (5), a putative cholesterol transporter. Along with fa tty acids and triglycerides, the newly absorbed cholesterol is packaged into chylomicrons and enters the bloodstream. Chylomicrons are acted upon by lipoprotein lipase which cleaves triglycerides thereby liberating free fatty acids from the particle. The resulting chylomicron remnants are taken up by the LDL receptor related protein (LRP) in the liver (6). In this organ, dietary cholesterol t hen merges with the newly synthesized cholesterol pool. A portion of this choleste rol will be converted to bile acids and exit through the bile ducts. The bile acid s enter the intestine where they function as detergents to solubilize dietary fat. T he majority of the bile acid pool is recycled back to the liver, but a fraction of it is excreted (7). The ATP binding cassette G5 and G8 transporters (ABCG5/8) also function to remove part of the cholesterol pool from the liver into the bi le (8). In the intestine, ABCG5 and G8 are the primary proteins responsible for removal of plant st erols from the body, and are important players in the efflux of unm odified cholesterol. Cholesterol esters, cholesterol, and triglycerides are packaged into very low density lipoprotein particles (VLDL) in the liver and secreted into the bloodstream. The VLDL particle provides the peripheral tissues with a valuable
3 source of cholesterol, as well as energy in the form of triglycerides. The VLDL particle is large in size, and has apolipoproteins B-100, C-I, C-II, C-III, and E as its primary protein constit uents (9). These VLDL particl es carry triglycerides to the muscle and heart, which have a particular re ceptor that is specific for this lipoprotein. VLDL has an abundance of apolipoprotein C-II which prevents efficient interaction with the low density li poprotein receptor. Circulating VLDL is a target of lipoprotein lipase. This enzyme cleaves triglycerides into free fatty acids which then enter peripheral tissue, or bind to albumin to remain in the bloodstream. After interacting with lipoprotein lipase, the particle now contains a larger percentage of cholesterol, and is reclassified as an intermediate density lipoprotein (IDL) particle. IDL particles have lost some of their apolipoprotein C component, and are smaller in size than VLDL. These particles have apolipoprotein E (Apo E), the high affinity ligand for the LDL receptor. For this reason, IDL efficiently binds to the LD L receptor, and is taken into the liver. Without a doubt, this is the primary route of cholesterol clearance from the blood. Defects in either the LDL receptor or apoE genes, have devastating consequences for cholesterol balance. A defec t in LDLR is the basis for familial hypercholesterolemia type II, a disease in which subjects may have serum cholesterol levels in excess of 400 mg / dl, and often die fr om coronary artery disease in their twentie s to fifties (10). When the amount of total choleste rol in the blood exceeds the LDL receptors clearance capacity, low density lipoprotein particles (LDL) will
4 accumulate. LDL particles form as a result of the action of lipoprotein lipase on circulating IDL particles. Along with the loss of triglyceride mass, there is also a corresponding loss of Apo E. This leaves the LDL particle with only apolipoprotein B-100 (Apo B-100) as its primary protein ligand. The LDL particle is then less amenable to internalization by the LDL receptor. Circulating LDL not cleared by the LDLR are taken up by the scavenger receptor (SRB-I). These receptors are present in all cell types, including macrophages. LDL cholesterol tends to linger in the bloodstream longer than other lipoprotein particles. During this time, it is subject to oxidation by free radicals, as well as reactive oxygen and nitrogen species. The resulting oxidiz ed LDL (ox-LDL) are taken up by macrophages via the scavenger receptor. After ingesting more oxidized LDL, macrophages turn into foam cells that depos it themselves in the artery wall. The foam cells eventually die and lead to further inflammation exacerbating the problem (11). At this stage, large extr acellular deposits of cholesterol and lipid begin to stack up and form atherosclerotic plaques. Plaque formation is a progressive process that continues thr oughout life, as long as LDL cholesterol levels are sufficiently elevated. Over time the plaques can rupt ure, allowing clots to occlude blood vessels, leading to peri pheral artery disease, stroke, and heart failure. Reverse cholesterol transport is a key determinant in the rate of atherosclerotic plaque formation (12). Macrophages that have infiltrated the endothelium can also get rid of excess cholesterol. This is thought to happen
5 primarily through the actions of ATP bi nding cassette transporter A1 (ABCA1), and ATP binding cassette transporter G1 (ABCG1). These proteins can transport unesterified cholesterol from the macrophages to lipid-poor nascent HDL particles in the serum. The nascent HDL particles hav e apolipoprotein A-1 as their protein component. With the ac tion of Lecithin: Cholesterol Acyl Transferase (LCAT), the free cholesterol is then converted into cholesterol esters, and is sent to the core of the particle. This results in the formation of the high density lipoprotein particlesHDL 2 and HDL 3 which are rich in protein and phospholipids. These HDL particles can t hen be taken up by SR-BI in the liver and eliminated. HDL particles can also ex change lipids with circulating VLDL and IDL. Cholesterol Ester Transfer Protein (CET P) brings cholesterol esters from the HDL particle to the VLDL/IDL particle. This represents a significant action of the HDL cholesterol. VLDL particles will eventually become IDL, and direct their cholesterol load to the liver via the LDL receptor. Low HDL levels are an independent risk factor fo r CVD. For these reasons, high HDL levels are desirable, as they can favorably shift the equilibrium from cholesterol deposition to efflux. While cholesterol absorption is a majo r player in the development of CVD, the role of cholesterol synthesis is ev en more important. In human beings, sterol synthesis accounts for between 60 and 80% of the total cholesterol pool (1). Due to obvious technological and ethical barriers, it is still unclear what contribution each tissue makes to whole body choleste rol synthesis in humans (13). Early
6 measurements of 14 Cacetate incorporation show ed the liver and intestine were responsible for the bulk of cholestero l synthesis in rats (14), and squirrel monkeys (15). In these animals, only liver cholesterol synthesis was affected by fasting and refeeding, or feedback inhibited by dietary cholesterol. This contrasts with experiments with 114 Coctanoate labeling in r abbits where the liver made only a minor contribution to cholesterol synthesis (16). Rabbits are very susceptible to dietary cholesterol. Inte restingly, rabbits had feedback inhibition by dietary cholesterol in all the tissues examined. It was later discovered that labeling with 14 Cacetate or 114 Coctanoate may in fact misrepresent the true rates of sterol synthesis due to variations in the acetate pool (17). The development of a tritiated water-b ased assay greatly improved absolute measurements of sterol synthes is (18,19). Experiments with 3 HH 2 O showed that over 50% of cholesterol synthesis occurs in the livers of male Sprague Dawley rats (20). A more systematic study using rats, monkeys, hamsters, rabbits and guinea pigs, was also per formed using this method. It was discovered that in the rat, monkey and hamster, the liver contained more newly synthesized cholesterol than any other or gan (21), in contra st to rabbits and guinea pigs which had less total synthesis. More than any other organ, the liver pla ys a critical role in the control of cholesterol balance. Virtually all the chol esterol that enters or leaves the body passes through the liver. Dietary cholestero l is brought from the intestine to the liver where it mixes with newly synthesiz ed cholesterol. The cholesterol can then
7 be packaged into VLDL particles and secreted into the bloodstream. This particle provides a source of triglycerides and c holesterol for peripheral tissues and is the precursor of IDL and LDL. The liver guards against excess serum cholesterol through the action of the LDL receptor (LDLR) (22). The LDL receptor removes IDL, and to a lesser extent LDL particles from the bloodstream. On ce in the liver, the excess cholesterol can be converted to bile acids, esterified or directly effluxed as unmodified choleste rol. This clearance from the bile ducts into the gut is the primary route of cholesterol elimi nation from the body. The liver integrates all of these processes through a series of complex regulatory networks. At the center of this is 3-hydroxy-3-methylgl utaryl coenzyme A reductase (HMGR), the rate-controlling enzyme in cholesterol biosynthesis. HMGR is an enzyme that functions as a dimer in the endoplasmic reticulum. Each monomer consists of eight membrane spanning domains as well as a cytoplasmic catalytic domain (23). HMGR utilizes two reducing equivalents of NADPH to convert 3-hydroxy-3-methyl gutaryl coenzyme A to mevalonate. This is the rate-limiting reaction in the biosynt hesis of cholesterol and a critical control point for regulation. In addition to cholesterol, mevalonate is also the precursor of important non-sterol products including isopentenyl tRNA, heme A, ubiquinone, dolichol, farnesyl pyrophosphate, and ger anylgeranyl pyrophosphate. Since cells cannot survive without a source of me valonate or cholesterol, an elaborate regulatory network has evolved to ensure adequate rates of synthesis. Equally intricate mechanisms guard against over expression of HMGR, which would lead
8 to the toxic accumulation of excess cholesterol. These mechanisms act at transcriptional, as well as several posttr anscriptional levels to fine tune HMGR activity; this phenomenon has been term ed multivalent regulation (24). HMGR is highly expressed in the live r, the primary site of regulatable cholesterol synthesis. Hepatic HMGR activity is affected by cholesterol, insulin, thyroid hormone, bile acids, fasting and re feeding, and also varies diurnally (5). Like many other enzymes, HMGR is subj ect to phosphorylation (25,26). It has been shown that HMGR has less cata lytic activity when phosphorylated. However the phosphorylation status does not vary under most physiological conditions (27), and therefore does not pl ay a significant regulatory role. In contrast to many other en zymes, HMGR activity is c ontrolled almost exclusively by changes in expression. Cholesterol has, without a doubt, the most important influence on HMGR protein levels. Feedback inhibition of chol esterol biosynthesis was first observed by Schoenheimer and Breusch in 1933 when they noticed that mice produced more cholesterol on a low cholestero l diet. This accumulation stopped when placed on a cholesterol rich diet (28). Year s later, this observation was confirmed with 14 Cacetate labeling in dogs (29). Li scum and others discovered that rats expressed 33 times more HMGR when fed a diet containing colestipol (a bile acid binding resin) and mevinolin (an i nhibitor of HMGR) than those fed a chow diet. It was found that prot ein and mRNA levels were both significantly reduced
9 by cholesterol feeding (30) in these co lestipol and mevinolin fed animals. This was attributed to an inhibitory effect of cholesterol on HMGR transcription. Since that time, a great deal of work has been done in cultured cells to elucidate the exact mechanisms of this feedback inhibition. This ultimately led to the discovery of the SREBP pathway (31). In cultured cells, HMGR transcription is repressed by decreased binding of sterol regulatory element binding protei ns (SREBPs) to the promoter (32). SREBPs are transcription factors that are synthesized as membrane bound precursors in the endoplasmic reticulum (33). SREBPs normally associate with SREBP cleavage-activating protein (SCAP) which forms a tertiary complex with insulin induced gene product (Insig) protein in the presence of sterols (34). SCAP contains a YIYF motif that has been i dentified as a sterol sensing domain. Binding of sterols to this domain i nduces a conformational change in SCAP, causing it to associate tightly with Insig. Insig can be thought of as an ERretention protein. It serves as an anchor to keep the SREBP/SCAP complex pinned in the ER when sufficient sterols are present. When sterol levels fall, the SCAP/SREBP complex dissociates from Insig. SCAP then escorts SREBPs to the Golgi where they are cleaved by two pr oteases to liberate their mature, active N-terminal domains (35). These N-termi nal domains are basic helix-loop-helixleucine-zipper transcription factors which when released, migrate into the nucleus, dimerize, and bind to the SREs of target genes to activate transcription (Figure 1). There are three SREBP isof orms: 1a, 1c, and 2. SREBP-1a, and 1c
are splice variants that are involved in fatty acid metabolism (36) while SREBP-2 is specific for genes involved in cholesterol metabolism (37). ERCytoplasmGolgiSterols SCAPSCAP SREBP Insig Insig SREBP SREBP-NH2SRENucleus FIGURE 1. The SREBP pathway. SREBPs are first synthesized as membrane bound precursors in the ER. When sterol levels are sufficient, they form a tertiary complex with SCAP and Insig. When sterol levels fall sufficiently low, SCAP undergoes a conformational change and dissociates from Insig. The SCAP/SREBP complex is then escorted to the Golgi where SREBP is then acted on by two proteases (not shown). The SREBP NH 2 terminal domains dimerize and translocate to the nucleus. Once inside, they bind to the SREs of target genes to activate transcription. 10
11 In the liver, cholesterol decreases HMGR protein primarily by a posttranscriptional mechanism. Feeding rats a diet containing 2% cholesterol dramatically reduces HMGR protein levels, yet has little to no effect on mRNA levels (38). This is in contrast with prev ious findings in animals fed a basal diet of colestipol and mevinolin (30). It was found that under normal physiological circumstances, cholesterol does little to repress transcription in the liver. The transcriptional effects are typically obser ved only in the presence of cholesterol lowering agents (39). The exact mechanism for feedback inhibition by cholesterol in rat liver has not been determined. One factor appears to be slowing of translation of the HMGR message due to dietary cholesterol (27,40). RNA from cholesterol fed animals was found to be a ssociated almost exclusively with the inefficiently translating monosome fraction. Dietary cholesterol also inhibited incorporation of 35 S-methionine into HMGR in isolated liver slices. Consistent with this, there was no change in the half life of the HMGR protein when rats were treated with cycloheximi de to inhibit protein synthesis. At present it is unclear whether this effect on translation is a direct effect of dietary cholesterol or the result of the accumulation of an interm ediate. In addition, it is unclear whether the magnitude of this effect is suffici ent to account for the total reduction in HMGR protein. Other evidence suggests that choleste rol acts by accelerating degradation of HMGR (41-43). HMGR has a sterol s ensing domain (similar to that found in
12 SCAP) that allows it to interact with In sigs (44). Recently, it has been shown that lanosterol in particular promotes Insi g-dependent ubiquiti nation and degradation of HMGR in cultured cells (45). This occurs through recognition of the HMGR/Insig complex by gp78, a membranebound ubiquitin ligase. Gp78 in turn binds to VCP, an ATPase involved in re cognition and degradation of proteins in the ER. HMGR is then sent to the proteasome to be degraded. Currently, there is no direct evidence that this is the regul atory mechanism at work in the liver. However, the liver-specific Insig knockout mouse has now been generated (46). These animals have constitutive processing of SREBPs because the SCAP/SREBP complex cannot be retained in the ER. HMGR mRNA levels were elevated in these mice as expected, but protein levels were disproportionately higher. An overwhelming amount of HMGR protein was present, and this affect was attributed to the inability of Insig to turnover HMGR. Further experiments are needed to actually measure the rates of HMGR protein translation and degradation in these Insig deficient animals. Dietary intake of cholesterol can have a major impact on liver cholesterol levels. Conversely, there are several dr ugs currently in use that can markedly reduce liver and serum cholesterol levels. Ezetimibe, an inhibitor of cholesterol absorption in the intestine ( 47), is effective at lowering serum cholesterol levels in humans (48). Treatment with ezetimibe re sults in restoration of HMGR protein levels to normal in diabetic rats, ev en when fed a cholesterol-rich diet (49). Another class of drugs, the statins act by inhibiting HMG-CoA reductase activity,
13 thus slowing cholesterol synthesis in the liver. Millions of people worldwide are currently taking a statin. These drugs have been shown to reduce serum cholesterol levels (50), improve survival rates for those at risk for coronary events (51), and even reduce the size of atheroscl erotic lesions (52). In contrast to cholesterol feeding, statins cause a transi ent depletion of intrac ellular cholesterol levels. The effectiveness of these drugs is limited by the ability of HMGR to bounce back and compensate for the inhibition of cholesterol synthesis. When rats are fed a statin, HMGR protein levels may be increased by as much as 700fold (53). This compensation o ccurs at both transcriptional and posttranscriptional levels, likely by re versal of the mechanisms affected by cholesterol feeding. Research into t he effects of statins on hepatic HMGR expression is of critical importance to treatment of cardiovascular disease. Thyroid hormone levels have a major influence on HMGR expression. Hypophysectomized and thyroidectomized ra ts both have very low levels of HMGR protein. HMGR expr ession can be restored to normal by administering T 3 In contrast to cholesterol, this effect is entirely accounted for by the increase in HMGR mRNA levels (54). Thyroid hormone stimulates transcription of the HMGR gene in the liver about 5-fold based on nucl ear run-on assays (55). This effect on transcription takes about 96 hours to reach it maximum (56). The slow induction may be due to the synthesis of another thyroid hormone responsive transcription factor, such as SREBP-2 which is induced transcriptionally by Thyroid Receptor Beta 1 (TR1) (57). Thyroid hormone also st abilizes the HMGR message, by
14 four to six-fold (58,59) This c an be antagonized by the administration of synthetic glucocorticoids (44), and the ex act mechanism is not known. Both of these effects need further investigation. HMG-CoA reductase activity varies throughout the day in rodents (60) and humans (61). Although initiall y ascribed to alterations in the phosphorylation state of the enzyme (62), this was later f ound to be due to corresponding changes in HMGR mRNA and immunoreactive protein levels (63,64). Curiously, this diurnal rhythm is lost in PPAR-alpha deficient mi ce (65). It should be noted that HMGR expression is at its peak during t he first few hours of the dark cycle, corresponding with the primary feeding ti me for rats and mice Indeed, fasting and refeeding alone have a dramatic effect on HMGR expression that exceeds the magnitude of the diurnal rhythm. Fasted rats hav e considerably less HMGR activity and expression than fed rats (66,67). HMGR reductase levels return to normal shortly after refeeding It is therefore likely that changes in food intake explain the majority of the diurnal variation in HMGR. Bile acids are modified cholesterol deriv atives that emulsify cholesterol, lipids, and fat soluble vitamins in the in testine. These are secreted into the gut during feeding. Cholesterol is then needed to replenish the bile acid pool. This cholesterol is obtained by internalization via the LDL receptor, as well as new synthesis. This necessitates increased HMGR expression, and may have been a driving force in the regulation of this gene in mammals. Bile acid feeding has
15 been shown to exert a potent suppressive effect on HMGR (68). Conversely, removal of bile acids by cholestyrami ne, a bile acid sequesterant, greatly elevates HMGR protein levels (69). This effect may be due to lowering of liver cholesterol levels, or alternatively to a mo re direct effect of bile acids on HMGR expression (70). Insulin levels rise during feeding to promote the absorption of serum glucose, and also to stimulate glycogen and fatty acid synthesis. Conversely, during periods of fasting, glucagon levels rise to promote gluconeogenesis, fatty acid oxidation and glycogenolysis. The balance between insulin and glucagon signaling is a key regulator of many metabol ic processes (71). It is not surprising then that HMGR is also an insulin-regul ated gene (72). In this way insulin can activate HMGR during periods of feeding to provide new cholesterol for the liver as well as for the replenishment of t he bile acid pool. Type I diabetics have lower rates of cholesterol synthesis and increased absorption of dietary cholesterol (73). HMG-CoA reductase (HMGR) protein and mRNA levels are both decreased in streptozotocin-induced diabetic rats, and can be rapidly restored with insulin treatment (74) suggesting regulation at the transcriptional level. The effect of insulin on HMGR message levels can be observed even in the presence of cycloheximide (74). This suggests that the action of insulin is direct, and does not occur by increased synthesis of the responsible protein(s). Previous experiments in H4IIE cells (rat hepatoma), showed that the proximal reductase promoter could be activated by insulin, at levels great er than or equal to those seen in live
16 animals (75). Questions remain as to w hether the insulin activation observed in hepatoma cells mirrors the physiological regulation of the gene. The HMGR promoter has been well-characterized in cell culture. The promoter lacks a classical TATA box and CCAAT box found in most other eukaryotic genes. It is very GC-rich and exhibits 90% identity between the human, mouse, rat and hamster sequences. The major regulatory el ements are highly conserved between these species. Early work found that the hamster HMG-CoA reductase gene requires about 300 base pa irs of sequence upstream of the transcription start site for high level expr ession in HeLa cell extracts (76), and in mouse L cells (32). Further studies were successful in identifying the sterol responsive element in the HM GR promoter (77). The HM GR promoter contains a variant sterol response element (SRE) of the sequence GTG C GGTG (78) which is a 7/8 bp match to the consensus SR E found in the LDL receptor (79), and HMG-CoA synthase promoters (80), al beit on the opposite strand of the DNA. The HMGR promoter can be activat ed by SREBP-1 and SREBP-2 in cultured cells, and in transgenic mice overexpressi ng these proteins (11, 12, 13). The SREBP-2 knockout mouse is not viable, but the SREBP-1 knockout mice shows increased HMGR expression, owing to compensation by SREBP-2. Although SREBP-1c appears to be insulin-responsive at the mRNA level, recent evidence suggests this factor is more clos ely tied to lipogenesis than cholesterol biosynthesis (14-16). SRE BP-1a and 1c prefer an E-box binding motif which is not present in the HMGR promoter, although they have been shown to bind the
promoter in vitro (81). SREBP-2 binds to the classical SRE motif, and is a potent activator of the HMG-CoA reductase gene. The HMGR promoter also contains potential binding sites for NF-Y, Sp1, as well as a functional cyclic AMP response element (CRE) (82,83). Recently, an LRH-1/FTF site has been identified, that is proposed to mediate repression by bile acids on the human promoter (70) (Figure 2). -325CTGCAGGTCA AATTCTGAGT TCGGGGTACT CCACCCGCGG AATCCCCTGTLRH-1/FTF-275CCCCCGCGCG GGCGGCGTCC GGCAGGCGCC CCCACGGCTC GGGGACCAAT-225 For. primer reads CsAGGAAGGCCG CGATGCTGGG ACCCGACTAG CCATTGG TTG GCTCGGCCGTNF-Y-175GGTGAGAGAT GGTGCGGTGC C CGTTCTCCG CCCGGGTGCG AGCAGTGGGCSRE Sp1-like-125GGTTGTTAGG GCGACCGTTC GTGACGTAG G CCGTCAGGCT GAGCAGCCTCCRE-75CCGCCGATTGG CTAGGGGAT CGGACGATCC TTCCTTATTG GCGGCCGGTTNF-Y Rev. primer reads Gs-25GGCGGCCCGG AGCGTGCGTA AGCGCAGTTC CTTCCGCCCT GGTCTCCGTTTRANSCRIPTION START+26+70GGCTGGAGAC GGCGGCAGGG CCGGCTTGGT GGCCTCCATT GAGATRat HMG-CoA reductase promoter FIGURE 2 Sequence of the rat HMG-CoA reductase promoter. The rat HMG-CoA reductase promoter was cloned by PCR and sequenced. The major transcription factor binding sites are underlined. These include a recently identified LRH-1/FTF site at -321/-316, an NF-Y site at -204/-189, the sterol response element (SRE) at -164/+155, and the Cyclic AMP response element (CRE) at -105/-96. The Sp-1 like site at -147/-135 and the NF-Y site -70/-65 were first identified in this work. The boldface A denotes the transcription start site. Locations of the labeling primers for footprinting are indicated by the arrows. 17
18 In regards to insulin regulation of t he HMGR promoter, previous work from our laboratory was performed in H4IIE cells, a rat hepatoma cell line. The hamster HMG-CoA reductase promoter wa s evaluated for insulin-responsiveness by transient transfections of luciferase reporter genes. Addition of 0.1 M insulin to these cells resulted in a 10-fold activation of transcription with the full length promoter construct (-270 to +20). Muta tion of the SRE had no effect on insulin stimulation, while mutation of the CRE abolished the insulin effect. During subsequent work with the H4IIE cells, it was found that the magnitude and nature of the response was not very predictabl e. Passage time, and to a greater extent, cell density both seemed to affect the insuli n response. In many cases, treatment with insulin caused a noticeable change in the morphology of the cells. H4IIE cells would lose their squared-off appearance and develop into more of a spindle-shaped structure. It seems likely that in culture, insulin may cause these cells to differentiate and lose their hepatocyte character. Insulin is a hormone that is known to affect both cell growth and viability in culture (84,85). In addition, numerous tumor cell lines have lost the normal feedback regulation for HMGR (86,87). Aberrant regulation of HMGR in these cells may confer a selective advantage to them in terms of growth or viability (88) as cholesterol and other mevalonate-derived products are indispensa ble for growth. For these reasons, there is considerable doubt that the insulin effect observed in H4IIE cells is an accurate reflection of the in vivo situation in the liver.
19 Virtually everything known about the HMGR promoter has been gleaned from in vitro studies, or transfections in cultured cells. Previous in vitro footprinting studies of the HMGR prom oter provided a wealth of information about potential transcription factor binding sites (76), but left many unanswered questions about the in vivo occupancy of these footprinted areas. The only in vivo footprints of the HMGR promoter have been performed in Jurkat and HepG2 cells. Despite a 1.5 fold response to insulin, the HepG2 cells displayed no noticeable changes in prom oter occupancy (89). To date there has been no work on the regulation of the HMGR promoter within the physiological framework of the liver. To address these questions, we have chosen to examine the occupancy and function of the HMGR promoter in the liver. Performing these studies in live animals ensures that our results will reflect the physiol ogical regulation of HMGR, in the context of the many nutritional and hormonal stimuli that the liver receives. Toward this end, we have examined the re sponse of the HMGR promoter to two key stimuli: insulin and ster ols. In order to directly examine the occupancy of the HMGR promoter, we have adapted the in vivo genomic footprinting technique to isolated liver nuclei as well as fresh li ver. To evaluate the function of the footprinted regions, we have used in vivo electroporation (90,91) to deliver HMGR promoter-luciferase reporter genes to the livers of live animals. We have found that diabetes markedly alters the occupancy of the hepatic HMG-CoA reductase promoter. In addition, we have performed the first functional analysis
20 of the hepatic HMGR promot er and observed sterol responsiveness in live animals.
21 MATERIALS AND METHODS MaterialsMaterials used in this work were obtained from various suppliers, and are mentioned under the method secti on for each particular experiment. Radioisotopes were obtained from Perkin Elmer. Unless otherwise indicated, chemicals were from Sigma-Aldrich. PlasmidsThe HMG-CoA reductase promoter was obtained by PCR from rat genomic DNA. Primers were designed to am plify regions of t he HMGR promoter from -325, -228, -176, 123, and -58 on the 5 end to a common end extending to +70 on the 3 end (see acknow ledgements). An additional larger fragment of the promoter ranging from 770 to +441 was also cloned. Primers for PCR were designed to introduce an Mlu I site on the 5 end, and an Xho I site on the 3end. The PCR products were then digested and ligated into pGL3 basic. These plasmids encode the firefly luciferase gene driven by different fragments of the HMGR promoter. The resulting plasmids are named as follows: -770/+441, -325, -228, -176, -123, -58 and are referred to as A-F in the text respectively (Figure 19). HMGR promoter-luciferase plasmi ds based on the -325 plasmid, harboring point mutations in the -70/-65 NF-Y site, were generated using the Quik Change kit from Stratagene. Primers for cl oning were designed using the rat genome
22 reference sequence (Genbank accession no. NW_047617.1) and are listed in Table I. Clones were initially identif ied by restriction digestion, and then confirmed by DNA sequencing at either Retrogen (San Diego) or the Moffitt Molecular Biology Core Facility (University of South Florida, Tampa FL). Renilla luciferase plasmids phRL-TK (Thymidine Kinase promoter) and phRL-CMV (Cytomegalovirus promoter ) were from Promega.
Table I GTAACCCGTTGAACCCCATT18S RNAReverseReal time18S real revCCATCCAATCGGTAGTAGCG18S RNAForwardReal time18S real forAGGAGCATTTGGCCCAATTAGCAGH. SynthaseReverseReal timeHCS real revCTTGGGATGGACGATACGCTTTGGH. SynthaseForwardReal timeHCS real forCTTCAAATTTTGGGCACTCAExon 2ReverseReal timeHMGR real revTGTGGGAACGGTGACACTTAExon 2ForwardReal timeHMGR real forGCAGAGCCCACAAGATTCTTExon 12ReverseChIPChIPExon 12 revGGCGGTCAGTGGTAACTATTExon 12ForwardChIPChIPExon 12 forCGGAAGGAACTGCGCTTACGpromoterReverseChIPChIPHMGR revCAATAGGAAGGCCGCGATGCpromoterForwardChIPChIPHMGR forGAATTCAGATCsyn. linkerLinkerFootprinting Short liner primerGCGGTGACCCGGGAGATCTGAATTCsyn. linkerLinkerFootprinting Long Linker primerCGGCCGCCAATAAGGAAGGATCGTCCGATCpromoterLabelFootprinting Footprinting B-3AACCGGCCGCCAATAAGGAAGGATCpromoterPCRFootprinting Footprinting B-2CGGAAGGAACTGCGCTTACGpromoterExtensionFootprinting Footprinting B-1ATGCTGGGACCCGACTAGCCATTGGTTGpromoterLabelFootprinting Footprinting A-3ATGCTGGGACCCGACTAGCCATTGpromoterPCRFootprinting Footprinting A-2CAATAGGAAGGCCGCGATGCpromoterExtensionFootprinting Footprinting A-1CGATCCCCTAGCCCCTCGGCGGGAGGCTGCTCpromoterReverseMutagenesisNF-Y mutB revGAGCAGCCTCCCGCCGAGGGGCTAGGGGATCGpromoterForwardMutagenesisNF-Y mutB forCGATCCCCTAGCTTATCGGCGGGAGGCTGCTCpromoterReverseMutagenesisNF-Y mutA revGAGCAGCCTCCCGCCGATAAGCTAGGGGATpromoterForwardMutagenesisNF-Y mutA forGCGCCTCGAGGCCTGTATCTGGCTCTTCTCCATpromoterReverseCloning441GCGCCTCGAGATCTCAATGGAGGCCACCAAGCpromoterReverseCloning70GCGCACGCGTGATCGGACGATCCTTCCTTATTGpromoterForwardCloning-58GCGCACGCGTTTGTTAGGGCGACCGTTCGTGApromoterForwardCloning-123GCGCACGCGTTGGTGAGAGATGGTGCGGTpromoterForwardCloning-176GCGCACGCGTAATAGGAAGGCCGCGATGCTpromoterForwardCloning-228GCGCACGCGTCTGCAGGTCAAATTCTGAGTTCGpromoterForwardCloning-325GCCGACGCGTGCCAGAAGCAGAAGGTGTAAGCACpromoterForwardCloning-770SequenceTargetDirectionPurposePrimer Table 1. Primer sequences used in experiments. The name of each primer is included in column one for reference. Column two states the use of the primer. Column three shows the direction of the primer relative to the template being amplified. The target of the primers is listed in column four, and the sequence in column five. 23
24 Growth of Bacteria and Isolation of Plasmid DNAPlasmids were transformed into JM109 competent cells (prom ega) and spread onto LB-Agar plates containing 100 g / ml ampicillin. Single colonies were picked for propagation in larger liquid cultures in Luria-Bertani medium (10 g bactotryptone, 5 g yeast extract, and 10 g NaCl per liter, pH 7.5) wi th 100 g / ml ampicillin. Cultures were grown overnight with constant shaking. For cell culture work or spotting of membranes for nuclear run-ons, 500 ml cult ures were grown. This plasmid DNA was isolated by the alkaline lysis method and followed by chromatography using the QIAGEN Maxi-Prep kit. For in vivo electroporation, 2.5 L cultures were grown. Plasmid DNA for electroporation was isolated by the same method using the QIA-Filter GigaPrep kit (Q IAGEN). After precipitation, the DNA pellet was suspended in 400-1000 l of TE buffer pH 8.0 for the Maxi-Preps, and in an initial volume of 2 ml for the GigaPreps. The DNA concentrations were then quantified by measuring the absorbance of a 1:100 dilution of each sample (generally 5 l of DNA in 495 l of water), and multiplyi ng by a factor of 5. The DNA yield varied depending on the plasmid bei ng propagated. Typical yiel ds were around 0.5-1 mg for the MaxiPreps, and 3-9 mg for the GigaPreps. GigaPrep DNA was diluted in bacterostatic 0.9 % saline to a final concentration between 1.0 and 2.0 mg / ml, and passed through a syringe filter to remove undissolved particles. The DNA was then flash frozen in liquid nitrogen and stored at -20C until used. MaxiPrep DNA was stored at 4C. All plasmid preps were confirmed by restriction digestion and electrophoresis on 1% agarose gels.
25 AnimalsMale Sprague Dawley rats, 125150g (Harlan), were allowed free access to Harlan Teklad 22/5 rodent chow and water. Animals were kept in a reverse cycle lighting system and we re sacrificed at 9:00 to 10:00 am, corresponding to the third to fourth hour of the dark period when HMG-CoA reductase expression is at its diurnal high. Animals were rendered diabetic by a single subcutaneous injection of streptozot ocin (Sigma), 65 mg/kg. Diabetes was verified by the presence of urinary glucos e using Clinistix from Bayer. In cases where animals did not test positiv e for diabetes, another injection of streptozotocin was admini stered. Diabetic animals were typically used 3-5 days after injection. Where indicated, animals received a subcutaneous injection of 3.0 units/100 g of recombinant human insulin (Novolin 70/30, Novo Nordisk) two hours prior to sacrifice, unless a different time is indicated in the figure legend. Some animals were fed ground chow cont aining 1% cholesterol or 0.02-0.04% lovastatin for 3 days prior to sacrifice or electroporation. For the electroporation studies, animals weighed between 175 and 225 g. This size was found to be optimal for surgical exposure of the liver Following surgery, animals were kept on the indicated feeding regimen until time of death. In one experiment, WistarFurth rats were substituted for Sprague Dawley (Figure 21). All procedures were carried out according to the regulations and oversight of the USF Institutional Animal Care and Use Committee (IACUC), protocol 2317. Nuclei IsolationNuclei were prepared as previ ously described by centrifugation through dense sucrose (92). Briefly, between 2.0 and 2.5 g of rat liver was placed
26 on ice in a beaker containing Nuclei Isol ation Buffer15 mM Tris-HCl pH 7.5, 15 mM NaCl, 60 mM KCl, 2 mM EDTA, 0.5 mM EGTA, 0.15 mM spermine, 0.5 mM spermidine, 1.9 M sucrose, 15 mM -Mercaptoethanol, 0. 5 % Triton X-100 and 0.1 mM PMSF. The liver was then minced into small pieces with small scissors and homogenized in a glass homogenizing vessel with a serrated Teflon pestle using a drill press. The homogenate was di luted with an equal volume of Nuclei Isolation Buffer without Triton X-100 and centrifuged at 90,000 x g for 90 minutes at 4C in a Beckman ultracentrifuge with a 50.2Ti rotor. Follo wing centrifugation, the supernatant was decanted, and the nuclear pellet was resuspended in 1 ml of cold PBS with 3 mM MgCl 2 Nuclear run-on assayNuclear run-on assays were carried out essentially as previously described (39). After centrif ugation, nuclei from 2 g of liver were resuspended in 1 ml of PBS with 3 mM MgCl 2 Next, 100 l resuspended nuclei was mixed with 100 l of 2x run-on buffer (160 mM Tris pH 7.5, 20 mM MgCl 2 2 mg/ml heparin, 1% sarkosyl, 0.7 M ammoni um sulfate, 0.8 mM each of ATP, GTP, UTP) and 250 Ci of 32 P-CTP were added to each reaction. Samples were incubated at 37C to continue ext ension of RNA transcripts. Samples were then treated with 50 units DNase I, 40 g proteinase K, and lysed with the addition of 50 l of 10% SDS and 25 l of 0.2 M EDTA. After extraction and precipitation, equal counts of 32 P-labeled RNA were added to each membraneapproximately 5 x 10 6 dpm. These membranes were previously spotted with 5 g of cDNA encoding HMG-CoA reductase (pRed 227), catalase, or the bluescript
27 vector. cDNA probes were described prev iously (39). Hybridizations were performed overnight at 57C. The next day, membranes were washed with 2x SSC, 0.1% SDS at room temperature for 1 minute, then 0.2x SSC, 0.1% SDS at 60C for 30 minutes, followed by 2x SSC containing 250 g of RNAse A at 37C for 30 minutes. Membranes were given a quick final rinse in 2x SSC and dried, followed by exposure to autoradiography film with an enhancing screen at -70C for 1-5 days. Serum Cholesterol and Glucose Blood was obtained at t he time of sacrifice by collection in 1.5 ml microcentrifuge tubes. Approximately 1 ml of blood was centrifuged at >16,000 x g in a bench t op microcentrifuge for 2 minutes at room temperature. The resulting supernatant (serum) was removed and stored at -20 C until used. Serum cholesterol levels were determined using the cholesterol oxidase assay from Pointe Scientific, acco rding to the manufacturers instructions. Serum glucose levels were determined usin g the glucose oxidase assay kit from Sigma, according to the protocol provided. Cell CultureThe H4IIE cell line, was obtain ed from the Amer ican Type Tissue Collection (ATCC) and grown in Eagles modified minimal essential medium (EMEM) supplemented with 10% feta l bovine serum, 100 U penicillin streptomycin per milliliter, and 1 mM sodium pyruvate. Cells were maintained at 37 C and 5% CO 2 in a humidified tissue culture incubator.
+DMSPiperidineXXYYExtensionLigation-Mediated PCRLabel with 32P-primerSequencing gel 123 xyz FIGURE 3. Overview of in vivo genomic footprintingIn vivo genomic footprinting is a powerful technique that is used to examine promoter occupancy in live cells or tissue. This technique depends on the selective methylation of exposed guanine residues (and to a lesser extent adenine) in the genomic DNA. Live cells are treated with dimethyl sulfate (DMS) which readily passes through the plasma and nuclear membranes. Areas of DNA where protein is bound will be protected from methylation, while areas of accessible DNA will be methylated. This effectively takes a snapshot of where transcription factors are bound to promoters at a point in time. Following DMS treatment, high molecular weight DNA is extracted. Once all the proteins, RNA, and extraneous matter are removed, the DNA is treated in a reaction with piperidine. Piperidine generates single stranded nicks in the DNA at methylated guanine residues. This results in a continuum of different DNA fragments from a population of cells for a given promoter. In order to examine a particular region of DNA, a promoter-specific primer is hybridized to the DNA and filled in with a polymerase. The resulting double stranded fragments receive a piece of linker DNA to be used as a common end for PCR. These fragments are then amplified in a PCR reaction to magnify the signal for a particular promoter region. Next the DNA is labeled with another promoter-specific primer, and run on a sequencing gel. Each site of methylation and cleavage is represented by an individual band on the gel. DNA that is extracted first and then treated with DMS (naked DNA) is run as a control (lane 1). In vivo samples will show protectionsareas of lightened or missing bands, indicative of protein binding (lane 2: x, y). Alternatively, significantly darker bands in the in vivo samples are known as enhancements (lane 3: z). These areas of enhanced DMS reactivity usually indicate nearby protein binding, presumably due to bending of the DNA molecule, or possibly the local chromatin environment affecting the diffusion of the DMS. 28
29 DMS treatment of H4IIE cellsH4IIE cells were plated on 150 mm dishes the day before the experiment at a density of 3.5 x 10 6 cells per plate. Four plates were used per condition. The next day, cells were washed two times with PBS and switched to serum free media. One gro up of cells received 100 nM insulin in the media. Three hours late r, media was removed and the cells were incubated with serum free media contai ning dimethyl sulfate (10 l / ml) for 5 minutes. Four plates were set aside for preparation of naked DNA and did not receive DMS. Media was removed by aspiration and plates were washed 3 times with PBS. Each plate then received 1 ml of lysis buffer (100 mM EDTA, 60 mM Tris pH 7.5, 0.1% SDS and 500 g / ml proteinase K). Ce ll lysates were transferred to 50 ml conicals, received additional lysis buffer to a final volume of 10 ml, and were rocked gently at room temperature for 3 hours. DMS treatment of rat liver nucleiNuclei were prepared by centrifugation through dense sucrose as described previously from 2 to 2.5 g of rat liver (92). Following centrifugation, t he nuclear pellet was resuspended in 2 ml of Nuclei Wash Buffer (PBS with 3 mM MgCl 2 ) and transferred to 15 ml conicals. Nuclei suspensions were then treated with 5 l of DMS for exactly 30 seconds (Final concentration of 0.25%). A control samp le used for the preparation of naked DNA was also mock treated without DMS. To stop the methylation reaction, suspensions were rapidly diluted with 12 ml of cold Nuclei Wash Buffer Tubes were then centrifuged at 1, 000 g for 5 minutes at 4 C to pellet the nuclei. The supernatant was removed with an aspira tor, and the wash was repeated with 3
30 ml, and then with 1ml of Nuclei Wash Buffer. The pellets were then resuspended in 10 ml of a buffer cont aining 10 mM Tris-HCl pH 8.0, 140 mM NaCl, 10 mM EDTA, and then transferred to 50 ml c onicals. Ten percent SDS was added to give a final concentration of 0.5%. Prot einase K (100 l of a 10 mg / ml stock solution) was also added at this time Tubes were rocked gently at room temperature for 3 hours to completely lyse the nuclei. DMS treatment of rat liverRat liver (2.2g) was minced in 8 ml of ice cold PBS. Liver pieces were homogenized 4-5 times using a drill press with a serrated Teflon pestle in a glass homogenizing ve ssel. A 5 ml portion of filtered homogenate was placed in a 50 ml polypropylene centrifuge tube. Each sample was then treated with 5 l of dimethyl sulfate for 2 mi nutes at room temperature. The DMS reaction was slowed by rapid dilu tion with 40 ml of ice cold PBS. Tubes were centrifuged at 1,000 x g for 5 minutes at 4C. Pellets were resuspended in 20 ml of PBS and washed again. The pellet was then resuspended in 15 ml of lysis buffer (60 mM Tris pH 7. 5, 100 mM EDTA, 0.5% SDS and 100 g/ml proteinase K). Samples were rocked gently at room temperature for 3 hours to completely lyse the nuclei. DNA extraction Genomic DNA isolation and piperidine treatment were performed as described previously (93). After lysing nuclei, DNA was extracted by adding equal volumes of buffered pheno l pH 7.5 (Roche), and chloroform: isoamyl alcohol (24:1). Samples were rocked gently for 10 minutes, and
31 centrifuged at 1500 x g to separate the phases. If the aqueous phase was clear, the supernatant was carefully removed us ing a cut-off 1000 l pipetman. Care was taken to avoid the white interphase. In general, the aqueous phase was quite viscous due to the presence of very high molecular weight DNA. In situations where the DNA was cloudy, addi tional TE Buffer (10 mM Tris-HCl pH, 0.1 mM EDTA) pH 8.0 was added, and the extraction was attempted again. Next, the aqueous phase was re-extracted with 2 volumes of chloroform: isoamyl alcohol (24:1) as above. Following extraction, one-tenth volume of 3 M sodium acetate pH 7.0 was added to the sample and mixed until the sample was clear and blurry lines were no longer visible. Two and a half volumes of -20C 75% ethanol were gently layered on top of the sample. The DNA was then spooled using a hooked Pasteur pipet. (Alternatively, if the DNA had no viscosity during the extractions, it was precipitated with 2.5 volumes of 75% ethanol by incubating overnight at -20C.) DNA was spooled by dipping the curved head of the Pasteur pipet just below the ethanol layer, and swirling the solution in a circular motion. The high molecular weight DNA attac hed to the pipet and had the appearance of a white silken thread. If a large amount of DNA was recovered, then the DNA appeared clear and globular. Next, the DNA was washed off the Pasteur pipet into a 15 ml conical using 1-2 ml of TE buffer and a 1,000 l pipetman. The final volume was adjusted with TE Buffer to 3-5 ml if the DNA was spooled, and 1-2 ml if it was precipitated overnight. The next day, the DNA was digested with 100 units of Hind III, a non-cutte r for the HMGR pr omoter, in a minimal volume. The sample was incubated for 3-4 hours at 37C and then extracted as before. After
32 the DNA had been spooled a second time, it was gently scraped into a 1.5 ml microfuge tube using 200 l of TE buffer. If the DNA was not viscous at this stage, then it was precipitated overnight and resuspended in 200 l of TE buffer. Preparation of control DNA and piperidine treatmentDNA that was extracted as above, but not DMS treated was used as a control for the footprinting experiments. This DNA was treated in vitro with DMS as described previously (93). Briefly, 200 l of DNA in TE buffer was incubated for 20 seconds with 1 l of DMS. The methylation reaction was r apidly stopped by the addition of DMS stop buffer1.5 M sodium acetate pH 7.0, 1.0 M mercaptoethanol, and 100 g yeast tRNA. ( In vivo samples were mock treated with DMS stop buffer and processed alongside control samples from this step onward.) The sample was quickly mixed by inversion and immersed in a dry ice ethanol bath for 10 minutes. The samples were then centrifuged at 16,000 x g for 10 minutes at 4C in a bench top microcentrifuge. The supernatant was removed, and the pellet was washed with 500 l of cold 75% ethanol. The pellet was then dried in a speedvac for 2 minutes. Piperidine was added to the sample to cleave the DNA at methylated guanines. This was accomplished by diluting piperidine 1:10 in ice cold water and adding 200 l to each sample. Samples were then vortexed to dissolve the pellet and incubated for 30 minutes at 90C on a heat block (pellet often did not fully dissolve). The samples were then frozen in a dry ice / ethanol bath and lyophilized to dryness in the speedvac. To remove residual piperidine, the sample was resuspended in 100 l of distilled water and lyophilized again.
33 This was repeated once more with 50 l of water. After the pellet was dried, it was resuspended in 50 l of distilled wa ter. DNA was quantified by measuring the absorbance of a 1:100 dilution of each sa mple (generally 5 l of DNA in 495 l of water), and multiplying by a fact or of 5. Roughly 200-300 g of DNA was obtained per sample. DNA was stored at 4C until needed for footprinting. Ligation-Mediated PCRLigation-mediated PCR wa s performed according to the original method (93). Primers were ordered HPLC-purified from Integrated DNA Technologies (IDT) at a synthesis sca le of 250 nmole. Two sets of primers were used to examine the occupancy of th e HMGR promoter. T he first primer set faces towards the transcription start site and reads Cs beginning at -185. The second primer set faces away from the transcription start site and begins by reading Gs at -58. These primer sets are labeled Footprinting A-1, A-2, A-3 and Footprinting B-1, B-2, B3 respectively (Table 1). Six to nine micrograms of DNA containing single stranded breaks were made double-stranded by annealing a promoter-specific primer (A-1 or B1, Table 1) and extending with Sequenase 2.0 (USB Corp). To accomplish this, 0. 3 pmole of primer was annealed to the DNA in a 15 l reaction containing 200 mM Tris pH 7.7, and 250 mM NaCl. Reactions were incubated in a thermocycl er (DNA engine, MJ Research) at 97C for 4 minutes, 58C for 30 minutes, and t hen held at 4C. Next 7.5 l of Mg/DTT/dNTP solution (20 mM MgCl 2 20 mM DTT, 0.2 mM each dNTP) was added to the annealed DNA, along with 1.5 l of Sequenase 2.0 diluted 1:3 in ice cold TE. The tubes were incubated at 47C for 10 min., 60C for 5 min., and then
34 held at 22C. During the 22C hold, 6.0 l of 310 mM Tris-HCl pH 7.7 was added, and the reaction was stopped by incubation at 67C for 10 minutes. In order to provide a common end for amplification of the HMGR promoter fragments, a synthetic linker was ligated to the DNA. This linker consisted of two annealed complementary oligos with a single blunt end (Table1, long linker oligo, short linker oligo ). For the ligation reactions, 20 l of dilution solution (17.5 mM MgCl 2 42.3 mM DTT, 125 g / ml BSA) was added to each sample and mixed, followed by 25 l of ligation mix (10 mM MgCl 2 20 mM DTT, 50 g / ml BSA, 50 mM TrisHCl pH 7.7, 3 mM rATP, 100 pmoles of hy bridized linker per reaction, and 3 units of T4 DNA ligase (Promega) per reaction) Ligations were performed overnight at 16C in the thermocycler. The next day, ligation reactions were stopped with a 10 minute incubation at 70C. Samples then re ceived 8.4 l of 3 M sodium acetate pH 7.0, 1.0 l of yeast tRNA (10 mg / ml), and 220 l of 95% ethanol, and were precipitated by a 10 minute incubation at -70 C followed by 3 hours at -20 C. Samples were then centrifuged at 16,000 x g for 30 minutes at 4 C to pellet the DNA. The pellet was then washed with 200 l of 75% ethanol, centrifuged, and air dried for 10 to 15 minutes. Each sample was then resuspended in 70 l of water with occasional vo rtexing and kept on ice. A 5x Taq buffer was prepared containing 200 mM NaCl, 50 mM Tris pH 8.9, 25 mM MgCl 2 0.05% (w/v) gelatin. Next each sample received 29 l of PCR mix containing 20 l of 5x Taq buffer, 20 nmole dNTPs, 10 pmole primer 2 (2-A or 2-B, Table 1), 10 pmole long linker oligo (Table 1) and water to volume. Tubes were mixed by gentle flicking, and 1 l of AmpliTaq (Applied Biosystems) DNA polymerase was added to each. PCR
35 reactions were carried out in a thermocycl er at the following temperatures: 97 C 2 min.; then 16 cycles of 97 C for 1 min. 66 C for 2 min., 76 C for 3 min.; followed by 95C for 1 min.; 66 C for 2 mi n.; 76 C for 10 min. and a hold at 4 C. Prior to beginning the PCR, primer 3 ( 3-A or 3-B Table 1) was labeled with 32 P ATP in a master mix containing 20 pmole of primer (1 l of a 1: 5 dilution), 2 l of 10 x Kinase buffer (Promega), 1 l of water, 15 l 32 P ATP (6000 Ci / mM, Perkin Elmer), and 1 l of T 4 polynucleotide kinase (10 u / l, Promega) for a group of five samples using the same pr imer set. The reaction was incubated at 37 C for 45 minutes, followed by 68 C for 10 minutes. Unincorporated label was removed with the QIAGEN Nucleotide Removal Kit according to the manufacturers instructions. Labeled primer was eluted in 100 l of 3 mM Tris pH 8.9 and lyophilized to dryness in a speedvac. The labeled primer was then resuspended in water (2 l per reaction) and stored on ice. PCR reactions then received of 5 l of Labeling Mix (1 l of 5 x Taq Buffer, 1.5 l of 10 mM dNTP mix, 2.0 l of resuspended labeled primer 3, and 0.5 l of AmpliTaq) and were labeled by incubation at the following temperat ures: 97 C for 3 minutes, 68 C for 2 minutes, and 76 C for 10 minutes. This cycle was repeated two more times. Samples then received 296 l of Taq Stop Buffer (260 mM sodium acetate pH 7.0, 10 mM Tris pH 7.5, 4 mM EDTA, and 10 g of yeast tRNA per sample) and were mixed well. DNA was extracted with 800 l of phenol: chloroform: isoamyl alcohol pre-mixed at 25:24:1. The aqueous phase was transferred to a new tube and precipitated overnight at -20 C wi th 1.0 ml of 95 % ethanol. The next day, samples were centrifuged at 16,000 x g for 30 minutes to pellet the DNA. The
36 pellet was washed with 400 l of 75% ethanol and centrifuged again. After removing the wash, the pellet was dri ed in the speedvac (approximately 2-5 minutes) and resuspended in loading dye (80% formamide, 0.5 x TBE, 0.01% xylene cyanol, 0.01% bromophen ol blue) with occasional vortexing. DNA was heat denatured at 90 C for 5 minutes and ch illed in ice water. Samples were loaded onto a 6% denaturing polyacrylamide gel, set up in a 0.4 to 1.2 mm wedge sequencing apparatus (Gibco S2 sequencing apparatus ) and run until the xylene cyanol front was 4-5 cm from the bottom. The gel was composed of 6% polyacrylamide (19:1 acrylamide: bis-ac rylamide), 7.75 M urea, 1 x TBE) and was prepared in a 130 ml volume using 1 ml of 10% ammonium persulfate and 36 l of TEMED. After the run, the gel was transferred to whatman 3 mm filter paper and dried for 2 hours. The dri ed gels were allowed to expose autoradiography film (Kodak Biomax MS Fi lm 35 x 43 cm) with an intensifying screen (Kodak Biomax Transcreen LE) at -70 C for 1 to 7 days, or put on a Molecular Dynamics Storage Phosphor Sc reen and developed on a Molecular Dynamics Storm 860 Phosphor imager. In cases where overall signal intensity varied greatly between wells, another gel was run with loading adjusted accordingly. Identification of protections and enhancementsClear and legible in vivo footprints were carefully examined for the presence of protections or enhancements. The position of the mo st recognizable protections ( ) and enhancements ( ) are shown on the side of each figure. It should be noted that
37 any given footprint may contain unique pr otections or enhancements that were not consistently observed. These are mentioned in the text. Each figure contains a footprint that is representative for each condition (i.e. normal, diabetic, cholesterol or lovastatin) in regards to the major sites sh own in the insets. Compilation of in vivo footprinting dataAlthough several footprints were obtained for the bottom strand (as in Fi gure 9), there were not enough legible footprints to make a fair comparison between the normal and diabetic groups. The primer set used to reveal the top strand (as in Figure 10) gave the clearest and most interpretable footprints, with fi ve normal and five diabetic animals per group. Only bands that varied distinctively from the naked DNA were recorded as protections or enhancements. These positio ns were compiled into a spreadsheet using Microsoft Excel software. Protected areas that were observed in four or more animals in at least one of the gr oups were noted in Figure 12. The same criteria was used to identify enhancements. Therefore the bindi ng sites that are shown as occupied in figure 12 are all bas ed on footprints from the top strand. Nuclear extractNuclei isolated from 2 g rat live r were resuspended in 1 ml of PBS containing 3 mM MgCl 2 and centrifuged at 3,000 g for 5 minutes at 4C. Nuclear pellets were resuspended in 0.5 to 1.0 ml high salt buffer (420 mM NaCl, 20 mM HEPES pH 7.9, 1 mM EDTA, 1 mM EGTA, 20% glycerol, 20 mM NaF, 1 mM Na 3 VO 4 1 mM Na 4 P 2 O 7 1 mM DTT, 0.5 mM PMSF, 1x protease inhibitor cocktail [Sigma]). Nuclei were lysed by ro tating slowly at 4C for 30 minutes. The
38 lysates were then centrifuged at 16,000 g for 15 minutes to pellet nuclear debris. The supernatant (nuclear extract) was collected and stored at -70C until needed. Protein concentrations were determined using the BCA assay (Pierce). EMSAElectrophoretic mobility shift a ssays were performed as previously described (94). Briefly, probes corr esponding to the HMG-CoA reductase promoter footprinted r egions were generated by anneal ing two complementary oligonucleotides (IDT). The oligo sequences are given in Table 2. One pmol of probe was labeled by the Klenow fill in reaction using 20 Ci of 32 P-dCTP, along with cold 0.125 mM dATP, dGTP, and dTTP. Each probe (25 fmol) was incubated with 10 g of rat liver nucl ear extract in a binding buffer (10 mM HEPES pH 7.9, 25 mM KCl, 0.5 mM EDTA, 50 g/ml poly dI:dC, 5% glycerol, 0.5 mM DTT, 125 g/ml BSA) for 20 minutes at room temperature. One to two g of the following antisera were added to t he binding reactions: Sp1 (sc-59x), Sp1 (sc-420x), Egr (sc-110x), CREB-1 (sc-187x), NF-Y (sc-7711x), NF-Y (sc-13045x), NF-1 (sc-870), from Santa Cruz Biotec hnology. Binding reactions were run on a 6% polyacrylamide gel in 0.25x TBE. Gels were dried on 3 mm Whatman paper and allowed to expose film overnight at -70 C.
Table 2 CTG ACC CTA GCC AAT CGG CGGGAG GCT GCAG CCT CCC GCC GAT TGG CTA GGG-59-82ICTG ACT GAC GGC CTA CGT CAC GAA CGG TACC GTT CGT GAC GTA GGC CGT CAG-88-112HCTG AAC GTC ACG AAC GGT CGC CCT AAC ATGT TAG GGC GAC CGT TCG TGA CGT-99-122GCTG ACG GTC GCC CTA ACA ACC GCC CAC TAGT GGG CGG TTG TTA GGG CGA CCG-109-132FCTG AAA CAA CCG CCC ACT GCT CGC ACC CGGG TGC GAG CAG TGG GCG GTT GTT-119-142ECTG ACA CTG CTC GCA CCC GGG CGG AGA ATTC TCC GCC CGG GTG CGA GCA GTG-129-152DCTG ACA CCC GGG CGG AGA ACG GGC ACC GCGG TGC CCG TTC TCC GCC CGG GTG-139-161CCTG AGG AGA ACG GGC ACC GCA CCA TCT CGAG ATG GTG CGG TGC CCG TTC TCC-147-170BCTG AAC GGC CGA GCC AAC CAA TGG CTA GCTA GCC ATT GGT TGG CTC GGC CGT-176-199ABottomTopEndStartProbe Table 2. DNA oligo sequences used as probes in EMSA experiments. The probe name is given in the first column. The second and third columns show which nucleotides in the HMGR promoter that are covered by the probe. Columns three and four show the nucleotide sequences of the oligos used to generate the probes. Transient transfectionsH4IIE cells were plated to an initial density of 100,000 cells per well in 24 well plates the day before the experiment. The following day, the media was removed and the cells were washed one time with PBS. Cells were transfected with 1 g of DNA/ well using Promegas Transfast reagent in the recommended 2:1 ratio. Cells were co-transfected with reporter construct and pRL-TK in a 4:1 ratio. One hour after transfection, the 200 l of transfection mix in each well was diluted with 800 l of growth media. 12-16 hours later, cells were harvested in 100 l of passive lysis buffer and assayed for luciferase activity using the Dual Luciferase Assay Kit (Promega). Data are shown as the average ratio of firefly to Renilla luciferase counts +/standard deviation of the 39
40 mean. At least six independent yet identical transfections were performed per condition. All plasmid concentrations were checked by A 260 prior to transfection. Chromatin preparation from rat liverRat liver (2.2 g) was placed in a beaker containing 10 ml of ice-cold PBS. The liver was minced into small pieces and diluted with an equal volume of 2% fo rmaldehyde in PBS, followed by a tenminute incubation at room temperatur e. Formaldehyde cross-linking was stopped by the addition of 2 ml of 1.25 M glycine. Liver pieces were washed 3 times with 10 ml of ice-cold PBS. Samples were then homogenized in 12 ml of nuclei isolation buffer + Triton X-100 using a serrated Teflon-glass homogenizer with a drill press. Nuclei were isolated by centrifugation through dense sucrose according to protocol. Nuclei were re suspended in 1-2 ml of PBS containing 3 mM MgCl 2 and centrifuged at 3,000 g, for 5 minutes in 1.5 ml tubes. The nuclear pellet occupied about 50 l volume and was white in color. Nuclei were then resuspended in 1 ml of nuclei lysis buffer (50 mM Tris pH 8.1, 10 mM EDTA, 1% SDS, 1 g/ml leupeptin, 1 g/ml aprotinin, and 0.1 M PMSF). Sonication was accomplished with a Heat Systems-Ultr asonics W375 sonicater with a microtip. Ten 10-second bursts at a pow er setting of 4, 40% duty cycle, were sufficient to shear chromatin to an average size of 200-600 bp. Shearing of chromatin was checked by electrophoresis on a 3% agarose gel. Chromatin concentrations were equalized to 0.3 mg/ml by A 260 and stored at 70C until needed.
41 Chromatin Immunoprecipitation assaysChromatin immunoprecipitations were performed essentially as previously described (95). One hundred microliters of chromatin suspension was diluted with 900 l of IP dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris-HCl pH 8.1, and 167 mM NaCl) in a 1.5 ml microcentrifuge tube. Twenty microliters of 10 mg / ml salmon sperm DNA and 20 l of 10 mg / ml yeast tRNA were also added to each tube. Samples were pre-cleared with 80 l (160 l of a 50% slurry) of ssDNA/Protein A agarose beads (Upstate), by rocking for 30 minut es at 4 C and then centrifuging for 16,000 g for 5 minutes at 4 C. One hal f of the pre-cleared supernatant was transferred to a new tube for immunoprec ipitation. Each immunoprecipitation reaction (containing 15 g of chromatin) received 5 g of the appropriate antibody: USF-2 (sc-862), CREB-1 (sc-187x), phospho-CREB (sc7978r). All antibodies were rabbit polyclonal IgG from Santa Cruz. Tubes were rocked overnight at 4 C to bind antibody to the chromatin. The next day, immunocomplexes were collected by adding 30 l (60 l of slurry) of ssDNA/ Protein A agarose for one hour at 4 C. Tubes were then centrifuged at 1,000 g for 1 minute. The supernatant was removed and saved as the Input. Immunoprecipitations were then washed twic e with 1 ml of Wa sh Buffer A (0.1% SDS, 1 % Triton X-100, 2 mM EDTA, 20 mM Tris-HCl pH 8.1, 150 mM NaCl), Wash Buffer B (0.2 % SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl pH 8.1, 500 mM NaCl), Wash Buffer C (0 .25 M LiCl, 1% NP-40, 1% sodium deoxycholate, 1 mM EDTA, 10 mM Tris-HCl pH 8.1) and then TE Buffer pH 8.0. Immunocomplexes were eluted by adding IP elution buffer (1% SDS, 0.1 M
42 NaHCO 3 ) and shaking on a vortex platfo rm for 15 minutes. Tubes were centrifuged at 16,000 g for 5 minutes to pellet the beads. The supernatant was transferred to a new tube. This elut ion was repeated, and the eluates were pooled. To reverse the crosslinks, each reaction received 20 l of 5 M NaCl and was incubated overnight at 65C. The next day, eluates were treated with 10 l of 0.5 M EDTA, 20 l of 1 M Tris-HCl pH 6. 5, and 2 l of proteinase K (10 mg / ml) and incubated at 50C for one hour. DNA was extracted with 2 volumes of phenol: chloroform. The aqueous phase was kept (300 l) and precipitated with 3 volumes of 95% ethanol for 3 hours to overnight at -20C. Tubes were centrifuged at 16,000 g for 10 minutes to pellet the DNA. The supernatant was removed and the pellet was washed with 600 l of 70% ethanol. The pellet was then dried and resuspended in 30 l of st erile water. PCR was performed using primers for the HMG-CoA reductase promoter (82) ( ChIP HMGR for, rev ) or Exon 12 ( ChIP Exon 12 for, rev ) of the HMGR gene (Table 1) Fifty microliter PCR reactions were set up containing 3 l of immunoprecipitat ed DNA, 10 pmol of each primer, 10 pmol each dNTP, 10 l of 10 x PCR buffer with MgCl 2 (Fisher), and 1.25 units of Taq DNA polymerase (Fisher). Input DNA was diluted 1:100 and used as a positive control in a separ ate reaction. PCR was carried out in a thermocycler under the following conditions: 95 C for 5 min., 40 cycles of 95 C, 58 C, 72 C, each for 45 seconds, followed by 95 C for 1 min., 58 C for 2 min., 72 C for 10 min., and held at 4 C. Ten microliters of each PCR reaction was run on a 2.2% agarose gel with ethidium bromide, next to a 100 bp DNA ladder (Promega).
x Harvest liverAssay luciferaseactivity In VivoElectroporation FIGURE 4. Overview of in vivo electroporation. Rats are anesthetized and kept under general anesthesia using isofluorane. The liver is then surgically exposed by making a transverse incision starting from the mid-sagittal position, approximately 1 cm caudal to the xiphoid process, extending 3 to 4 cm toward the dorsal surface of the rat. The left, right and median lobes of the liver are gently pushed out of the incision over a piece of sterile gauze. Plasmid is then injected beneath the capsular surface of the liver. A circular six node electrode (0.75 cm diameter) is placed on the liver such that the needle points line the perimeter of the injection site. A BTX T830 square wave electroporator is used to administer 6 x 150 ms pulses at 100 V/cm, with a 150 ms rest between pulses. This electric pulsing temporarily permeabilizes the hepatocytes and also electrophoretically drives the DNA through. Minor scarring occurs where the electrode needles are sunk into the liver. This is useful when locating the exact area of the liver that has been treated. Each animal receives several different promoter constructs, all at different locations in the liver. Following surgery, animals have their wounds closed, receive analgesic, and are returned to standard housing. One day later, the livers are harvested and assayed for luciferase activity. 43
44 SurgeryRats were anesthetized with 5% is ofluorane in oxygen. The rats were then fitted with a standard rodent mask and kept under general anesthesia using 3% isofluorane. Next, the surgical site was trimmed with electric clippers and scrubbed with 70% isopropyl alcohol followed by betadine. The liver was surgically exposed by making a transverse incision starting from the mid-sagittal position, approximately 1 cm caudal to the xiphoid process, extending 3 to 4 cm toward the dorsal surface of the rat. T he left, right and median lobes of the liver were exposed by gently pushing them out of the incision over a piece of sterile gauze. ElectroporationElectroporation was performed essentially as previously described (96). A subcapsular injection of approximately 15 g of plasmid in 40 l of sterile saline was performed using a 26 gauge, 3/8 inch length needle. After injection the plasmid was visible below the capsular surface as a blanched out area. A circular six node electrode (0 .75 cm diameter) was placed on the liver such that the needle points line the perimeter of the area where the plasmid was visible. Electrodes were sunk to a controlled depth of 2 mm using a rubber spacer. After placement of the electrodes, a BTX T830 square wave electroporator was used to administer 6 x 150 ms pulses at 100 V/cm, with a 150 ms rest between pulses. These setti ngs were found to be optimal for gene delivery to the liver with minimal ti ssue damage. Several HMG-CoA reductase promoter constructs linked to the lucifera se reporter gene were tested at different
45 locations in the liver. To control for transfection efficiency, a plasmid encoding renilla luciferase driven by either the th ymidine kinase promoter (phRL-TK) or the CMV promoter (phRL-CMV) wa s co-administered. Relative to reporter construct, phRL-TK was mixed in a 1:4 ratio, and phRL-CMV was mixed in a 1:1000 ratio. These ratios were found to be optimal for de tection of firefly and renilla luciferase at the same luminometer sensitivity. Fo llowing electroporation, the liver was placed back in the abdomen, and the wound was closed with surgical staples. At the time of surgery, animals were gi ven a single subcutaneous injection of ketoprofen (5 mg / kg) for analgesia. Luciferase AssaysTwenty four hours after electr oporation, the livers were harvested and tested for expression of t he HMG-CoA reductase constructs by measuring luciferase activity. Livers were removed from the animal, and the electroporated region was extracted using a si ze 3 cork borer (0.6 cm diameter). Approximately 0.1 to 0.15 g of liver was placed in 600 l of passive lysis buffer (Promega) and homogenized us ing a polytron tissue disruptor. The lysate was then centrifuged at 16,000 x g for 5 minut es and the supernatant was assayed for luciferase activity using the dual luciferase assay kit from Promega, and a Turner Designs 20/20 luminometer. Treatment of Luciferase DataFor each animal, an untreated area of liver was assayed for both firefly and renilla luci ferase activity. The average firefly and renilla background counts for all the anima ls on a given day were subtracted from
46 each sample. Any sample with firefly or renilla luciferase counts below background was excluded. Due to the inherent variability in DNA injection and electroporation efficiency, all luciferase values were normalized to renilla luciferase. Renilla luciferase was co-administered in the plasmid mixes using either phRL-TK (thymidine kinase promoter) or phRL-CMV (CMV promoter). Similar results were obtained with both r enilla luciferase vectors, and expression of these plasmids did not vary with inje ction site or feedi ng regimen. For the mapping of HMGR promoter activity in normal rats, all animals received phRL-TK (Figure 20). In the experiment examini ng the role of the -70/-65 NF-Y site, animals received phRL-CMV (Figure 21). For the mapping of t he sterol response, three separate experiments were perform ed. Two experiments were performed with phRL-TK in which animals were fed 0.02% lovastatin or 1% cholesterol. These experiments used 3 and 2 animals respectively for each feeding regimen. Another experiment was performed using phRL-CMV. These animals were fed either 0.04% lovastatin or 1% cholesterol in the chow for 3 days. Data from these three experiments were pooled by taki ng the average ratio obtained for the shortest promoter construct (-58/+70) of both feeding conditions, and setting this value to one. Remaining numbers in each experiment were multiplied by a constant so that the fold response to cholesterol loading or depletion could be compared for all the different promoter constructs. The resulting pattern of promoter activity was the same as each of the three indivi dual experiments, and data from both feeding regimens were tr eated identically. Calculations were performed using Microsoft Excel, and Sigm a Plot 8.0 software. Pooled data was
47 graphed as the mean ratio of firefly to reni lla luciferase +/standard error of the mean (Figure 22). P-values were ca lculated using the students t-test. Isolation of MicrosomesLivers from electroporated animals were used to prepare microsomes as previously descr ibed (27). Approxim ately one gram of liver was minced in 10 volumes of ic e cold 0.25 M sucrose solution and homogenized using a drill press with a serrated Teflon pestle in a glass homogenizing vessel. The homogenate was then centrifuged at 15,000 x g for 15 minutes at 4 C. The resulting supernat ant was centrifuged at 90,000 x g for one hour at 4 C. After the second centri fugation, the microsomal pellet was resuspended in 0.25 M sucrose at a concent ration of 5-20 mg / ml and stored at 70 C. Protein concentrations were determined using the BCA protein assay (Pierce) according to the manufacturers instructions. Western BlottingThirty micrograms of microsom al protein was combined with at least 3 volumes of Western Sample Bu ffer (2% SDS, 60 mM Tris pH 6.8, 8 M Urea, 0.1 M sucrose, 5 % -Mercaptoethanol, 0.005% bromophenol blue) and heated to 95-100 C for 5 minutes. Samples were cooled on ice for two minutes and then subjected to SDS-PAGE on 4-15% gradient gels (BioRad). These gels do not contain SDS. It was found that pre-running these gels for 30 minutes prior to loading of the sample, resulted in sharper bands and avoided solubility problems for HMGR. Following electrophores is, proteins were transferred to a PVDF membrane, and blocked in 5 % nonfat dry milk with 0.1 % Tween 20 for
48 one hour. The membrane was then probed with an antibody to HMG-CoA reductase A9 mouse monoclonal from ATCC (30), or -actin as a loading control (Sigma catalog A5441) for 2 hours to overnight. Anti-mouse horseradish peroxidase-conjugated secondary antibody from Amersham was used in a 1:5,000 dilution. Membranes were was hed 3 x 10 minutes after each antibody incubation with wash buffer (PBS wi th 0.1% Tween 20), and developed using SuperSignal West Pico ECL reagent (Pie rce) according to the manufacturers instructions. Each lane represents a samp le of liver microsomes from a separate animal. RNA Isolation and cDNA synthesisTotal RNA was isolat ed from 0.4 g of rat liver using Tri-Reagent (MRC) according to the manufacturers instructions. Ethanol pellets were stored at -70 C until the day of the experiment. RNA concentrations were then determined by measuring the absorbance of a 1:100 diluted sample at 260 nm. Prior to synthes is of cDNA for use in Real-Time PCR, the isolated RNA was DNase treated us ing TURBO DNA-Free kit (Ambion). Some modifications of the published pr otocol were made (49). Briefly, 23 g of RNA was treated using 2 l of Turbo DNase enzyme in a reaction volume of 30 l. Next, 5 g of the isolated and DNase-treat ed total RNA was used to generate cDNA for use in Real-Time PCR. The reverse transcriptase reaction was carried out with the Superscript II First Strand Synthesis system for RT-PCR (Invitrogen) according to the manufacturers instructions.
49 Real-Time PCRReal-Time PCR was conducted acco rding to the protocol from iQ SYBR green supermix using an iCycler (B io-Rad) with minor modifications. The total reaction volume was adjusted to 25 l down from 50 l. Two microliters of cDNA was used as template for r eactions with the HMG-CoA reductase primers ( HMGR real for, rev, Table 1), which corresponded to regions of exon 2 (Genbank accession number NM_013134). R eactions to detect HMG-CoA synthase used two microliters of cDNA with primers that have been described previously ( HCS real for, rev, Table I) (97). Two microliters of a 1:40 dilution of each cDNA was used for reactions with primers ( 18S real for, rev, Table 1) to the rat 18S ribosomal RNA sequences (G enbank accession number X01117). All primers were used at a final c oncentration of 100 nM. The annealing temperature was 61 C, and 40 amplific ation cycles were performed. Melting curves were done after each run and a single distinct peak was obtained for each primer set. The data were processed by iCycler IQ optical system software 3.0 (Bio-Rad) and analyzed by the CT method, using Microsoft Excel statistical programs and SigmaP lot (version 8.0).
50 RESULTS Nuclear run-on assays were performed to determine if insulin acts to increase transcription of the HMG-CoA reductase gene in the liver. Male Sprague-Dawley rats were injected with streptozotocin (65 mg/kg) to induce diabetes. Animals were sacrificed during the third hour of the dark cycle, at the diurnal high for hepatic HMG-CoA reductase expression. Nuclei were isolated from the livers of these animals, and nuclear run-on assays were performed as described in the methods section. Probes for HMGR, Catalase (CATa housekeeping gene) or pBluescript (BLU Ebackground hybridization) were spotted onto nylon membranes. Radiolabeled RNA from the nuclear run-on reactions was then hybridized to these membranes. The signal intensity of each band is indicative of that genes relative ra te of transcription. As seen in Figure 5, panel A, HMGR transcription was greatly diminished in the diabetic animals. It was also found that administration of insulin to diabetic animals restored HMGCoA reductase transcription to normal in just two hours (Fig. 5, Panel B).
D2D1N2N1HMGRCATBlueD1D2D3D4I1I2I3I4HMGRCATBlue A.B. FIGURE 5. Insulin activates HMG-CoA reductase transcription in diabetic rats. Nuclei were isolated from normal (N), diabetic (D), or insulin-injected diabetic rats (I). Extension of RNA transcripts was carried out with 32 Plabeled CTP. Equal dpm of RNA were hybridized to membranes containing 5 g of cDNA for HMG-CoA reductase (HMGR), catalase (CAT), or the Bluescript vector (Blue) A, Nuclear run-ons from two normal (N 1 ,N 2 ), and two diabetic rats (D 1, D 2 ). B, Nuclear run-ons from four diabetic rats (D 1-4 ), and four insulin-replenished diabetic rats (I 1-4 ). 51
52 To determine if the effect of insulin on transcription is related to changes in serum cholesterol or serum glucose levels, a time course was performed. Diabetic rats were injected with insulin (3 units / 100 g) and sacrificed 15, 30, 45, 60, or 75 minutes later. Relative HMGR mRNA levels were measured by real time RT-PCR using 18S ribosomal RNA as an internal control. It was found that insulin did increase HMGR message levels in a very rapid fashion (Figure 6, Panel A). It is likely that animal to ani mal variations in basal HMGR expression contributed to the scatter in the data. Nonetheless there was a clear correlation between time after insulin injection and HMGR expression. The effect appeared to be maximal in as little as 45 minutes. Serum cholesterol levels did not show any particular trend with time (Figure 6, panel B). Serum glucose however, showed a time-dependent decrease with insulin treatment, approaching normal levels only at 75 minutes (Figure 6, panel C) From this experiment it is clear that insulin rapidly acts to increase HMGR message levels. This effect is not dependent on changes in serum cholestero l, and precedes the normalization of serum glucose.
010020030040050060070080001020304050607080 02040608010012014001020304050607080 0.002.004.006.008.0010.0012.0001020304050607080 Serum Cholesterol mg / dlSerum Glucose mg / dlTimeA.B.C.Relative mRNA FIGURE 6. Time course of HMG-CoA reductase mRNA activation by insulin. Six diabetic rats were injected with insulin and sacrificed at 0, 15, 30, 45, 60 or 75 minutes later. A. HMGR mRNA levels were determined relative to 18S ribosomal RNA by real time RT-PCR. B. Final serum cholesterol levels were determined using the cholesterol oxidase assay. C. Final serum glucose was measured using the glucose oxidase assay. 53
54 These experiments demonstrate that in sulin exerts a rapid effect on HMGR transcription. We hypothesized that this is accomplished by the binding of an insulin-responsive transcription factor to the promoter to activate transcription. In order to directly examine the occupan cy of the HMGR promoter, we decided to employ the technique of in vivo genomic footprinting. This method has been used successfully to map where transcription factors are bound to DNA in vivo (93,98) (Figure 3). In vivo genomic footprinting allows for a complete and unbiased survey of the HMGR promoter. Performing this technique in animals ensures that the footprint reflects physiological regulat ion of the gene, in the context of the many nutritional and hormonal st imuli that the liver receiv es. In order to do this, we designed a primer set to reveal the bottom str and of the HMGR promoter which reads cytosines beginn ing at -185. Another primer set was designed to reveal the guanines on the top strand beginning at -58. It should be noted that these primers were designed against t he rat HMGR promoter sequence, which varies slightly (about 10 bp) from the hamster (83). The primer design was based on PCR and sequencing of the HMGR promoter from the Sprague-Dawley rats used in our experiments. Minor differ ences from the published rat sequence included an extra G at -15 and a reversal of the CG at -3,4. Numbering is therefore -1 bp relative to the prev iously published sequence, based on the transcription start site (Fig. 2). Previous work on from our laborator y was performed in H4IIE cells, an insulin-responsive rat hepatoma line (75). Questions have been raised as to
55 whether these cells are an accurate model of HMGR promoter regulation in the liver. For this reason, we chose to examine the occupancy of the HMGR promoter first in H4IIE cells, and then comp are this to the physi ological situation in the liver. To accomplish this, H4IIE cells were grown in media containing 10% fetal calf serum and switched to serum fr ee media on the day of the experiment. One group of cells also received 100 nM insulin. Three hours later, the cells were treated with DMS and the DNA was prepared as described in the methods. This DNA was used for ligation-mediated PCR and the resulting labeled products were run on a sequencing gel (Figure 7) Lane 1 contains naked DNA that was first extracted and then treated with DMS in vitro This lane represents all possible cleavage sites, and the sequence of the promoter can be followed based upon it. With this pr imer set, the Cs in the sequence can be read as a result of the reactive guanines on the bottom strand. Lane 2, is an in vivo sample (-) from the cells treated wi thout insulin, while the sample in lane 3 received insulin (+). Bands that are absent or reduced in intensity in the in vivo samples, represent protections where protein binding shields the DNA from dimethyl sulfate attack. These protections are noted with a filled triangle ( ) in the figures. Exact nucleotide positions are given at the right.
-161-154-148-137-126-114-106-101-95-87-71CRE SRE -+Naked FIGURE 7. In vivo footprinting of the HMG-CoA reductase promoter in H4IIE cells. H4IIE cells were switched to serum free media at time zero, and treated with (+) or without (-) 100 nM insulin for four hours. Cells were then treated with dimethyl sulfate and subjected to in vivo genomic footprinting. The first lane contains Naked DNA, or DNA that was first extracted and then treated with dimethyl sulfate. This represents all possible cleavage sites for this strand of the promoter. The position of the CRE and the SRE are underlined. The major protection at -101 of the CRE is shown with a solid triangle (). 56
57 There is a strong protection visible at -101 of the CRE, and less obvious protections at -114 and -71. Despite num erous attempts, the H4IIE cells could only be footprinted in one direction. It is likely that the other primer set did not work due to differences from the rat genome sequence. It is interesting that occupancy of the promoter in H4IIE cells di d not vary with insulin treatment. This agrees with previous observations in t he literature. It was found that insulin treatment did not affect HMGR promoter occupancy in HepG2 cells, despite a 1.5fold increase in mRNA (89) (HepG2 are an immortalized human hepatoma cell line). Questions remain about the relev ance of insulin regulation of HMGR in cultured tumor cells, to the physiological situation in the liver. Until now, this question has not been addressed. In order to examine the occupancy of the HMGR pr omoter in rat liver, nuclei were first isolated from norma l or diabetic rats. The nuclei were resuspended and then treated with DMS. Following this, the DNA was extracted and subjected to ligation-m ediated PCR as before. The footprints obtained from these samples showed some distinct areas of protection and enhancement (Figure 8). Protections are noted with a filled triangle ( ). Bands that are significantly darker in the in nuclei samples, or new bands that appear in these lanes, are known as enhancements. Enhanced DMS reactivity is indicative of protein binding in the nearby area, alt hough not necessarily on that particular residue. Enhancements have been marked with an open triangle ( ). On the bottom strand (Figure 8, Panel A), there is a very strong protec tion at -101 of the
58 CRE in both normal and diabetic footprints. This was always present and did not vary with treatment. On the top strand as well, the CRE element is protected in all cases at -109, -105, -103, and -100 (F igure 8, Panel B). There is also a particularly striking enhancem ent at -142 in the diabet ic nuclei. This was a consistent observation for all the di abetic samples and was never observed in footprints from normal nucle i. It should be noted that the overall signal in the diabetic samples fades towards the top of the gels (Figure 8, Panel B, lanes 2 and 3). This is indicative of over-treat ment with DMS relative to the normal and control samples. Excessive methylation results in more frequent cleavage of the DNA, and consequently selects for shorter PCR products. This tends to overrepresent the smaller fragm ents, and likely explains the presence of lesser enhancements flanking the CRE seen only in the diabetic samples. Nevertheless, it is clear that the CRE is occupied in nuclei from both normal and diabetic rats. In addition there is a very strong enhancement at -142 only in the diabetic footprints, suggesting that an insulin-responsive factor may act in this area.
-161-154-148-137-126-114-106-101-95-87-71ND1D3nakedCRE SRE -55 -142-80-100-105-70-109-124nakedND2D1-153SRE CRE -165-59A.B. FIGURE 8. Footprinting of the HMG-CoA reductase promoter from rat liver nuclei. Nuclei isolated from normal (N) or diabetic (D) rats were treated with dimethyl sulfate to methylate exposed guanine residues. DNA was treated with piperidine to cleave at these positions, and footprinted by ligation-mediated PCR. Naked refers to DNA that was first purified and then treated with DMS. This lane represents all reactive guanine residues. Protections are shown with a solid triangle () and enhancements with an open triangle (). The position of the reactive guanines is shown relative to the transcription start site at the right side of each gel. The sterol response element (SRE), and cyclic AMP response element (CRE) are highlighted. A. Footprint for the bottom strand B. Footprint for top (coding) strand. 59
60 Isolation of nuclei from rat liver ta kes approximately two hours for a single animal. It is actually somewhat surp rising that protections and enhancements would still be visible given the long time gap before DMS treatment. Although there was a striking difference at -142 in t he footprints from these nuclei, it was possible that other important contacts we re being missed. In order to get a truer in vivo look at promoter occupancy, it was necessary to have a more rapid treatment with DMS. In order to accompli sh this, livers were removed from the rats, quickly dounced to disaggregate the hepatocytes, and then immediately treated with DMS. Overall, the treatment of fresh liver by this method gave more consistent data, and resulted in the disco very of several important protections and enhancements that were miss ed when working with nuclei. In vivo footprinting was performed on live rs from normal or diabetic rats. On the bottom strand, several areas of DMS protection or enhancement were detected (Fig. 9). Once more, protections are indicated by a filled triangle ( ), and enhancements by an open triangle ( ). In both the normal and diabetic footprints, the CRE was completely protected at -101, and showed enhanced DMS reactivity at -104, -99 and -95/-94 (F ig. 9, inset). This pattern was seen in all animals regardless of treatment. A si gnificant protection seen only in normal animals occurred at -71, as shown in the inset. Other protections, such as those seen at -137 and -147 were not consistently observed.
NakedND 161-154-137-126-114-106-101-87-71-55-46-3410410199NakedND 71SRE CRE 95/94 FIGURE 9. In vivo footprint of the hepatic HMGR promoter, bottom strand. Livers from normal (N), or diabetic (D) rats were treated with dimethyl sulfate to methylate exposed guanine residues. DNA was footprinted by ligation-mediated PCR. The naked lane shows all reactive guanine residues. Protected bases are indicated by a solid triangle (), while areas of enhanced DMS reactivity are denoted with an open triangle (). Due to weak overall signal, the naked lane was given a longer exposure to allow for comparison to the in vivo samples. The -71 region and CRE have been enlarged in the insets. 61
62 On the top strand, the CRE is heavily protected at -100, -105, -103 and 109 (Fig. 10, bottom inset) in both normal and diabetic animals. The A at -102 showed up as an enhancement in both ca ses. As with the nuclei, a key difference between the in vivo footprints is a very obvious enhancement at -142 seen only in the diabetic samples (Fig. 10, middle inset). This particularly dark band, indicating enhanced DMS reactivity, wa s seen in 4/5 diabetic footprints and 0/5 normal footprints. Conversely, the nearby enhancement at -138 of the normal lane was not seen in diabetic footprints suggesting possible competition for a binding site in this region. Another obvious difference is an enhancement at -161 in the diabetic lane (Fig. 10, top inset). Th is residue is in the middle of the sterol response element (SRE) located between -164 and -155. The SRE appears unoccupied under normal conditions, and enhanced in diabetic samples (4/5 animals). Protections at -189/-190 were seen in all groups, while those at -70 were not observed in the diabetics. Both of these areas contain potential NF-Y binding sites, with the sequence ATTGG. A partial protection was always visible at -85 regardless of treatm ent. In addition, protections at less reactive guanines such as -121 and -93 were observed whenever the band was detectable in the naked lane. These also did not vary with treatment.
ND142-85-100-105-73-109-129-153-165-NakedSRECRE175-142161102109100ND 105CRESRENaked 138103 FIGURE 10. In vivo footprint of the hepatic HMGR promoter, top strand. Livers from normal (N), or diabetic (D) rats were treated with dimethyl sulfate to methylate exposed guanine residues. DNA was footprinted by ligation-mediated PCR. Thenaked lane shows all reactive guanine residues. Protected bases are indicated by a solid triangle (), while areas of enhanced DMS reactivity are denoted with an open triangle (). The SRE, -142 region, and the CRE have been enlarged in the insets. 63 Since there was an enhancement at -161 of the SRE in four of the five diabetic animals, we wondered if insulin activation could be a result of sterol regulation through the SRE. It is conceivable that insulin treatment indirectly activates HMGR transcription by altering liver cholesterol levels. A rapid
64 decrease in liver cholesterol, due to c onversion to bile acids during feeding, might result in increased SREBP cleavage, and binding of these factors to the SRE. To investigate this possibility, we ex amined livers of rats fed lovastatin or cholesterol to alter liver cholesterol leve ls. Animals were fed 0.02% lovastatin or 1% cholesterol for five days (Figure 11). Previous research in our lab has shown that a similar dose of lovastatin elevates HMGR transcription 4-6 fold (39); while dietary cholesterol reduces HMGR protein levels to about 1% of control. It should be noted that dietary cholesterol has only a minor repressive effect on the rate of HMGR transcription in these animals (99) We predicted that the lovastatin-fed animals would show strong protections at the SRE, because of elevated cleavage of SREBPs, induced by sterol deprivation. Curiously, no definitive protections or enhancements at the SRE were visible when animals were fed either lovastatin or cholesterol (Fig. 11, inset). The footprints were the same for the two animals in each group. The CRE was also heavily protected in these animals, but unchanged by either treatm ent. The enhancement at -138 seen in normal animals was also seen in both lo vastatin and cholesterol-fed rats. The enhancement at -142 seen in diabetic ra ts was noticeably absent from these animals. In addition, the NF-Y sites at -189/-190 and -70 are readily visible in this footprint, and protected in both cases. Invariant protections are also visible at -125, -121 and -85 in this footprint.
LCh142-85-100-105-73-109-129-153-165-175-NakedSRECRE 190LChNaked SRE FIGURE 11. In vivo footprinting of the HMGR promoter in rats fed lovastatin or cholesterol. Rats were fed either 0.02% Lovastatin (L) or 1% cholesterol (Ch) chow for 5 days. Livers were treated with dimethyl sulfate and footprinted by ligation-mediated PCR. Protected bases are indicated by a solid triangle (), while areas of enhanced DMS reactivity are denoted with an open triangle (). The SRE (-164/-155) has been enlarged in the inset. 65
66 In vivo footprinting was repeated on the livers of several normal and diabetic animals. The primer set used to reveal the top strand (as in Figure 10) gave the clearest and most interpretable footprints, with five normal and five diabetic animals per group (Figure 12). T he data from these footprints was compared to establish the major areas of difference between normal and diabetic rats. Protections or enhancements that were observed in four or more animals in at least one of the groups were noted in Fi gure 12. In all cases, protections were observed around -85, as well as -109, -105 and -100 of the CRE. Enhancements consistently appeared at -102 of the CRE. These were always seen and did not vary with treatment. There are several noteworthy areas of difference between the normal and diabetic animals Four out of five normal in vivo footprints showed an obvious protection at -70. This prot ection was never observed in the diabetic samples (0/5). As with the nuclear footprinting, the enhancement at -142 was seen only in the diabetic footprints (4/5 ), and never in the normal (0/5). An interesting protection at -161 in the SR E coincided with the enhancement at -142, suggestive of a factor binding between these positions. The enhancement at the SRE was seen in four of the five diabetic footprints, and never in the normal.
Position ND-1610/54/5-1420/54/5FactorSREBP??Sp1?-704/50/5NF-YSequenceGGTGCCGGTGC C CCGCCCGGGTGCG CCGCCGGATTGG CT-105-102-109-100NDFactorSequence5/5CGTTCGGTGGAACGCGTAG GCREB-855/55/5CAGGCTGGAGCAGC?Common ElementsMajor Differences PositionCGGTTCGGTGACGGACGTAG GCGGTTCGTGACGGACGTAG GCGTTCGGTGACGACGTAG G -190/-1894/52/5NF-YCCATTGGGG TTG A.B. 5/5 FIGURE 12. Summary of in vivo DMS reactivity for the hepatic HMG-CoA reductase promoter. Many areas of DMS protection and enhancement were observed by in vivo footprinting. The most common and reproducible sites are noted in this table. Protected bases are indicated by a solid triangle (), while areas of enhanced DMS reactivity are denoted with an open triangle (). The nucleotide position is given relative to the transcription start site. In the next column, the sequence is given, showing the reactive residue that was detected. Five normal (N) and five diabetic (D) animals were footprinted. The number of times each site was affected is expressed as a fraction of the total animals in the group. Possible factors that may bind are listed in the last column. In order to identify the major factors bound to footprinted regions, we performed electrophoretic mobility shift assays (EMSAs). The EMSA is a straightforward technique used to evaluate in vitro binding of transcription factors to DNA sequences. A radiolabeled DNA probe is incubated in a binding reaction 67
68 with nuclear proteins and run on a polyacrylamide gel. The free probe will run the fastest and gives a dark signal at the bo ttom of the gel. Prot ein DNA complexes migrate considerably slower, and are visibl e as bands closer to the top of the gel. Addition of antibody can further slow the migration of these complexes, resulting in a supershift up the gel. Short oligon ucleotide probes were designed to cover the major footprinted elements (Table 2, methods section). These probes were radiolabeled and incubated with nuclear ex tracts from the livers of normal or diabetic rats. An invariant band was observ ed in all lanes about a third of the way down the gel (Figure 13, marked with a star [*]). This band appeared regardless of the probe sequence used, and is likely a DNA binding protein that is not sequence-specific. Aside from this, most of the probes to the footprinted regions showed distinct binding patterns. This was particularly true for the regions from 152 to -109 (D, E, F), the CRE probe fr om -112 to -88 (H), and again for the region from to -59 (I). These regions were selected for characterization of the factors binding to them (Although visible in this particular gel, probe F did not consistently show strong binding and was not pursued further).
A -199/-176B -170/-147C -161/-134ABCDEFGHID -152/-129E -142/-119F -132/-109G -122/-99H -112/-88I -82/-59 Nuclear Extracts:Lane 1-normalLane 2-diabeticProbes:Fig. 13* FIGURE 13. EMSA analysis of footprinted regions of the HMG-CoA reductase promoter. Short DNA probes corresponding to the major footprinted areas were used to detect in vitro protein binding in normal and diabetic rat liver nuclear extracts. Ten micrograms of nuclear extract from Normal (first lane), or Diabetic (second lane) rats was incubated with 25 fmol of a 32 P-labeled probe and electrophoresed on 6% polyacrylamide gels. Probes tested were A -199/-176, B -170/-147, C -161/-134, D -152/-129, E -142/-119, F -132/-109, G -122/-99, H -112/-88, and I -82/-59. The nonspecific band is marked by a star (*). In an effort to identify the factor binding to the /-119 region, we performed additional EMSAs. Figure 14, Panel A shows an experiment in which probes to the /-129 and -142/-119 regions were incubated in a binding reaction with normal (N) or diabetic (D) rat liver nuclear extract. This gel confirms that the same band is present with both probes, albeit with stronger binding to the -142/-119 region. In an effort to identify this factor, we performed another 69
EMSA using nuclear extract from a normal rat (Figure 15, Panel B). Antisera to either Sp1, Sp3 or Egr-1 were added to the binding reactions. All three of these proteins are good candidates based on the sequence of the probe. However, none of these antibodies resulted in a supershift. Once again, the lower band is nonspecific as seen in figure 13. -142/-119 -152/-129 -ND-NDNuclear Extract:Probe:? -142/-119 NNNN-Nuclear Extract:Probe:Antibody:Sp1Sp3Egr1A.B.** FIGURE 14. EMSA analysis of the region from -152 to -119. Short DNA probes corresponding to the regions from -152 to -129 and from -142 to -119 were used to detect in vitro binding in rat liver nuclear extracts. Ten micrograms of nuclear extract from Normal (N), or Diabetic (D) rats was incubated with 25 fmol of a 32 P-labeled probe and electrophoresed on 6% polyacrylamide gels. The first lane in each gel (-) is a binding reaction without nuclear extract (probe only). A. Protein binding to the -152/-129 and the -142/-119 probes in normal or diabetic rat liver nuclear extracts. B. Protein binding in normal rat liver nuclear extracts is unaffected by the addition of Sp1, Sp3 or Egr1 antisera. The nonspecific band is marked by a star (*). 70
71 An experiment was then designed to characterize the sequence specificity of this unknown protein. This was done by performing an additional EMSA with cold DNA competitor probes (Figure 15). The first lane contains a binding reaction with probe only, while the second received ten micrograms of rat liver nuclear extract from a norma l rat. In the third lane, the same unlabeled -142/-119 probe was used in a 10-fold molar excess as a cold competitor. In the remaining lanes, mutant cold competitor probes were added to the binding reactions. Each of these competitors harbor s a single base pair mutation at a different place in the sequence. These were added in 10-fold molar excess to the hot probe. The exact base change substitution is shown above each lane, with the wild type sequence on top and the mutant sequence on the bottom. The upper band can be efficiently competed away with wild type competitor (lane 3), and most of the mutant oligos. Noteworthy exceptions are C _AG from to and the entire sequence GGGCGGTT between and The presence of two binding sites explains why this band is s een as well in the overlapping /-129 probe, albeit with weaker signal. T he sequence GGGCGGTT seems like a good match for Sp1, whose consensus sequenc e is GGGGCGGGGC with a strong requirement for the core GGGCGG(100). However the inability of Sp1 antisera to shift this band, suggests that it is ac tually a different factor with a similar sequence specificity. In vitro binding ability of this fa ctor did not vary with diabetes. The nonspecific band seen in other fi gures (13-16) is visible only in the lane without competitor (Figure 15, none *). This band is a non-sequence specific DNA binding protein, as it was effect ively competed away in all other lanes.
gtcagggcagtgggcggttgtgactttactgtttattggtw.t.-none10x comp.-140-121-142/-119 FIGURE 15. EMSA analysis of Sp1-like factor binding to the -142/-119 region. A short DNA probe corresponding to the footprinted region between -142 and -119 was used to detect in vitro binding in rat liver nuclear extracts. Competitor oligos were used in a 10-fold molar excess to the probe where indicated. The first lane (-) is a binding reaction without nuclear extract. The second lane (none) is a binding reaction with rat liver nuclear extract but no competitor. The third lane (w.t.) is a binding reaction with the wild type sequence used as a competitor. The remaining lanes are labeled with the wild type promoter sequence on top, and an arrow pointing to the point mutation in the cold competitor oligo. The nonspecific band is marked by a star (*). Two areas consistently bound in vivo were the CRE, and the region surrounding a putative NF-Y site from -70 to -65. In order to identify the factors responsible for the in vivo footprints in these areas, additional EMSAs were performed (Figure 16). The probe from -115 to -85, which covers the entire CRE, showed strong binding with nuclear extracts from both normal and diabetic rats 72
(Figure 16). Virtually all of this band could be shifted further up the gel with the addition of CREB-1 antisera (Figure 16, Panel A), confirming previous observations (83). Strong binding had also been observed with the probe from -82 to -59. These bands could be shifted with two different antibodies to NF-Y, but not by an irrelevant antibody (NF-1) (Figure 16, Panel B). This experiment therefore identifies the footprinted region from -82 to -59, containing the sequence GATTGG as an NF-Y binding site, As with the CRE and Sp1-like sites, in vitro binding to this region did not vary with insulin. N-NNN-NDNDD-82/-59 -115/-85 CRENF-Y #1NF-Y #2NF-1CREB-1 Antibody:CREB NF-Y Nuclear Extract:Probe:Fig. 16** FIGURE 16. EMSA analysis of the CRE and the -70/-65 NF-Y site. Short DNA probes corresponding to the major footprinted areas were used to detect in vitro protein binding in normal (N) and diabetic (D) rat liver nuclear extracts. A. Probe for the CRE (-85/-115) with CREB-1 antisera added where indicated. B. Probe to the -59/-82 region, with two different NF-Y antisera or NF-1 antisera added as indicated. The nonspecific band is marked by a star (*). 73
74 Given the strong protection of the CRE seen in all the in vivo footprints, and the ability of CREB to bind in vitro we wanted to find out if CREB was in fact bound in vivo to the hepatic HMGR promoter. To accomplish this, chromatin immunoprecipitation (ChIP) assays were performed. This method involves chemically cross-linking proteins to DNA to assay in vivo binding. Once the chromatin has been cross-linked, it is sheared into small pieces (typically less than 500 bp in size), and immunoprecipitated with the antibody of interest. Following immunoprecipitation and extensive washing, the DNA-protein-antibody complexes are eluted. The DNA is extrac ted, and PCR is performed to detect the promoter in question. To see if CR EB is bound to the HMGR promoter in vivo liver pieces from normal and diabetic rats were cross-linked with formaldehyde. ChIP was performed and the HMGR pr omoter was detected by PCR. CREB-1 antibody was able to pull down the HMGR promoter from both normal and diabetic chromatin (Figure 17, Panel A) while an isotype-matched antibody to an irrelevant nuclear protein was not (Figur e 17, Panel B). This immunoprecipitation was not able to pull down Exon 12 of the HMGR gene, confirming that DNA was sheared to an appropriate size. The differ ence in intensity of the band for CREB between normal and diabetic samples was not reproducible, although the promoter was clearly pulle d down in both cases. In other experiments the CREB immunoprecipitation gave a band of equal intensity for both normal and diabetic chromatin. It should be noted that ChIP analysis by end point PCR is not quantitative, but merely provides qualit ative information about the presence or
absence of factors on a given promoter. The inability of p-CREB to pull down the HMGR promoter could be a result of poor antibody-antigen interaction. We would expect that at least some of the endogenous CREB is phosphorylated. Regardless, this data confirms reports that CREB is bound to the HMGR promoter in vivo (101), and validates this observation in the context of the live animal. beadsbeadsCREB-1Irr. Abp-CREBCREB-1p-CREBinput1:100input 1:100beadsCREB-1p-CREBbeadsCREB-1p-CREBNormalDiabeticNormalDiabeticIP:HMGR promoterHMGRExon 12230 bp-Irr. AbIrr. AbIrr. Ab650 bpN input1:100D input 1:100 IP:A.B. FIGURE 17. CREB is bound to the HMGR promoter in live animals. Liver pieces from normal (N) and diabetic (D) rats were cross-linked in 1% formaldehyde. Nuclei were isolated and lysed. Chromatin was sonicated to an average size of 200-600 bp. Chromatin 15 g was immunoprecipitated with no antibody (beads), an irrelevant antibody (USF-2), CREB-1 antibody, or phospho-CREB antibody. Forty cycles of PCR were performed, and products were run on a 2.2% agarose gel stained with ethidium bromide. A. PCR to detect the HMG-CoA reductase promoter. B. PCR to detect HMG-CoA reductase Exon 12. 75
76 Another point of interest in the in vivo footprints was the consistent protection of the NF-Y site identified at /-65. Since this site was strongly protected in 4/5 of t he normal footprints, and none of the diabetics, we hypothesized that this element might play a critical role in activation of transcription. In order to investigate a po ssible functional role for this site, we constructed luciferase reporter plasmi ds containing the full length HMG-CoA reductase promoter starting at and ending at +70 of the 5 untranslated region. Two identical plasmids harboring mutations in the NF-Y site were also made. These plasmids were transfected into H4IIE cells. The next day, the cells were harvested and assayed for luciferase activity (renilla luciferase was cotransfected for normalization purposes). As seen in Figure 18, the wild type promoter shows a high level of activity relative to the vector backbone. Both of the mutants significantly inhi bited luciferase production, indicating that this NF-Y site is required for effici ent transcription of HMGR.
Firefly: RenillaR.L.U.pGL3 basicW.T.MutAMutBWild type:Mutant A:Mutant B:ATTGGATAAGAGGGG-325+70-70 Luc. HMGR 432150 A.B.C. FIGURE 18. Effect of mutating the proximal NF-Y site on transcription in H4IIE cells. A, H4IIE cells were transfected with pGL3 basic, the wild type HMG-CoA reductase promoter-luciferase construct (W.T.), or the same plasmid harboring mutations in the NF-Y site (Mut A, Mut B). For normalization purposes, each construct was co-transfected with pRL-TK, a plasmid containing the renilla luciferase gene driven by the thymidine kinase promoter. Relative luciferase units (R.L.U.) are expressed as the ratio of firefly to renilla luciferase +/standard deviation of the mean. B, Schematic of the HMG-CoA reductase promoter-luciferase construct in pGL3. C, Sequence of the NF-Y site in the mutant plasmids from -69 to -65. These experiments provided invaluable information about the occupancy of the HMGR promoter, and identified some of the factors that bind to it. In order to assess the functional relevance of these sites, reporter gene assays were performed. The rat HMG-CoA reductase promoter was obtained by PCR of rat genomic DNA. A series of nested deletions of this sequence was cloned in front 77
of firefly luciferase. These promoter constructs include a large piece extending from -770 to +441 relative to the transcription start site, as well as five smaller pieces from -325, -228, -176, -123, or -58 on the 5 end to +70 on the 3 end (Figure 19). These promoter fragments contain all relevant transcription factor binding sites previously identified by studies of this promoter in cultured cells. The major transcription factor binding sites are displayed, as well as their locations in the reporter constructs (Figure 19, lower panel). -58-123-176-228-325-770A.B.C.D.E.F. +70+441 ++ CRESRENF-Y NF-YSp1-likeLRH-1/FTF 78 FIGURE 19. Schematic of HMG-CoA reductase promoterluciferase constructs. Top panel: Different fragments of the rat HMG-CoA reductase promoter were obtained by PCR and cloned into the pGL3 vector. The resulting plasmids encode firefly luciferase driven by the various fragments of the HMGR promoter: A -770 to + 441, B -325 to +70, C -228 to +70, D -176 to +70, E -123 to +70 and F -58 to +70. All numbers are given relative to the transcription start site. Bottom Panel: The location of the major previously identified transcription factor binding sites: LRH-1/FTF, NF-Y, SRE, Sp1-like, and the CRE.
79 Next, we sought to determine the in vivo relevance of these promoter regions in driving HMGR transcription in the liver. To accomplish this, we used in vivo electroporation to deliver reporter constr ucts to the livers of live animals. This is an established technique for the delivery of plasmid DNA to the liver (96). This approach has advantages over the other existing methods for introducing reporter genes into animals. Transgenic mi ce are expensive and time consuming to produce, and only one or two reporter genes can be tested at a time, limiting the effectiveness of this technique (95) Other methods for the introduction of DNA such as adenoviral infection (102), or hydrodynamic in jection primarily target the liver (103). While these are very effective methods of gene delivery, they generally require a large number of animals because each animal can receive only one promoter construct. Using in vivo electroporation, the expression of reporter genes can be limit ed to a particular anatomical area of liver where the plasmid is injected. This allows the testing of multiple promoter constructs in a single animal. This not only reduces the number of animals needed for a promoter study, but also helps to control for changes in transgene expression due to animal to animal variability. Animals were anesthetized and surgica lly opened to expose the liver. Plasmids were then delivered to the liver by subcapsular injection. A six needle electrode array was used to administer elec tric pulses and drive the DNA into the hepatocytes (104,105). Each animal received al l six of the promoter constructs in separate areas of the liver. A plasmi d encoding renilla luciferase was co-
administered for normalization of transfection efficiency. Following surgery, the animals were allowed to eat and recover overnight. The next day, livers were harvested and assayed for luciferase activity (Figure 20). For each animal, an area of liver that was not electroporated or injected with DNA was taken as a control for luciferase background. This is represented by the minus sign in figure 20. The pattern of promoter activity looked essentially the same in an individual rat (Figure 20, Panel A), as it did for a group of normal rats (Figure 20, Panel B). ABCDEF0510152025 ABCDEF 0510152025 30 ** *R.L.U. ratioR.L.U. ratioSingle animalGroupA.B. FIGURE 20. Functional mapping of the HMGR promoter in the livers of normal rats. Normal Sprague-Dawley (175-200 g) rats were electroporated with the HMGR-luciferase promoter constructs listed in figure 19, along with Renilla luciferase for normalization. Each animal received all six promoter-luciferase constructs A-F. A. The pattern of HMGR promoter activity for a single animal. The minus sign, - indicates an area of liver that was not electroporated but was assayed for background luciferase activity. B. The pattern of HMGR promoter activity for a group of six animals. Data is reported as the mean +/standard error of the mean for each promoter construct. Statistically significant differences in promoter activity are noted as follows: p< 0.05 *, p< 0.01 **. 80
81 Although construct A (-770/+441) contains the largest fragment of genomic sequence, it did not result in mo re activity than construct B (-325/+70). This is consistent with previous findings in the literature showing that 277 bp of the hamster promoter is requi red for high level expression in vitro (32). When the promoter construct was shortened from -325 to -228 (compari ng B to C), there was a significant fall-off in activity. This region contains a recently identified LRH1/FTF site, proposed to mediate a repressive effect of bile acids on HMGR promoter activity (70). Mo ving from -228 to -176 (Constr uct C to D) excludes a previously identified NF-Y site that was constitutively occupied in the footprints. Deletion of this site had little to no effect on promoter activity in normal rats. The region from -176 to -123 wh ich contains the SRE and Sp1-like sites, appeared to have a repressive effect on promoter acti vity. Construct E, which lacks this region gave consistently more activity, although th is difference did not reach statistical significance. This is in agreement with in vivo footprinting data which showed that this region was preferentially occupied in the diabetic state, where transcription is dramatically reduced (106), suggesting the presence of a transcriptional repressor. There is a major reduction in promoter activity when comparing construct E to Construct F. The deleted region contains the CRE (-104 to -96), and a newly identified NF-Y site (-70 to -65) This fall-off in promoter activity was highly significant with a p-value of less than 0.01. The most dramatic difference in pr omoter activity was observed between constructs E and F. This region contai ns both the CRE and the newly identified
82 NF-Y site. The CRE was occupied in all the footprints regardless of treatment, and CRE was the major factor bound in vitro and in vivo. Interestingly, the NF-Y site at -70 to -65 was bound only in the presence of insulin. This site was found to be required for efficient transcription in H4IIE cells, as two different mutations inhibited luciferase production. In order to determine the role of the NF-Y site in the liver, a similar experiment was performed by in vivo electroporation. WistarFurth rats were electroporated with the wild type construct B (-325 to +70), or two constructs harboring double point mutations in the ATTGG core of this element. As shown in figure 21, both mutants re sulted in substantially less promoter activity (mutant B, p < 0. 05). This generally agrees with observations in H4IIE cells (Figure 18) although the relative reduc tion in activity was not as drastic in the liver.
0.00.20.40.60.81.01.2 MutAMutBRLU ratioWild type:Mutant A:Mutant B:ATTGGATAAGAGGGGWT FIGURE 21. Effect of mutating the proximal NF-Y site on HMG-CoA reductase promoter activity in live animals. Normal Wistar-Furth rats were electroporated with plasmids containing the HMG-CoA reductase promoter from -325 to +70 along with Renilla luciferase for normalization. Each animal received an injection of the wild type construct (WT), and an injection of each construct harboring point mutations in the NF-Y site (A), (B). Data is reported as the mean +/standard error of the mean for each promoter construct. There were four animals in the group. The activity for construct B was significantly lower than the wild type (p < 0.05). Considerable effort was made to identify which regions of the promoter are insulin-responsive in the liver. Experiments were performed with groups of normal or diabetic rats, as well as diabetic rats replenished with insulin. Despite numerous attempts, under an array of different experimental conditions, an insulin response was not observed. In some cases, insulin actually appeared to repress HMGR promoter activity (data not shown). This may have been due to glucagon exerting a dominant effect on the relatively short promoter constructs 83
84 through the CRE. It is therefore possible that the true insulin-responsive element(s) may lie outside of the cloned pr omoter fragments. Efforts to clone larger pieces of the promoter extending out to -5 kb were also unsuccessful. PCR amplicons of genomic DNA were obtained and sequenced, but ligation into the reporter vector proved impossible. The ex act reason for this is unclear, but may be due to the presence of highly repetitive upstream sequences preventing selection of positive clones. The sterol response of the HMGR pr omoter has been well characterized in cell culture. It was expected that in the live animal, sterol regulation would also be mediated through the SRE, although this issue has not been directly addressed. To accomplish this, male Sprague Dawley rats were fed diets containing either cholesterol or lovast atin for three days. These two feeding regimens were intended to cause the greatest possible differences in HMGR transcription. Cholesterol feeding has been shown to drastically decrease HMGR protein levels, while having little or no effect on mRNA levels in rats (99). Conversely, lovastatin feeding is expected to transiently deplete the cell of cholesterol and mevalonate-derived metabol ites, resulting in a compensatory increase in HMGR expression. This upr egulation happens transcriptionally (about a 4 fold effect) (39), and to a far greater extent, post-transcriptionally (53). The promoter activity in the cholesterol-fed ani mals followed a patte rn similar to that observed for normal rats, while the ani mals fed lovastatin had considerably greater activity in constructs B, C, and D. Three separate experiments were
85 performed, two with 0.02 % lovastatin and 1% cholesterol, and another with 0.04% lovastatin and 1% cholesterol. All three experiments gave the same pattern for promoter activity and sterol response, although none achieved significance on their own due to the low number of animals. When the data sets are combined (Figure 22), a clear pict ure of sterol r egulation emerges. Constructs B, C, and D, all contain t he SRE and would be expected to respond to lovastatin feeding. These constructs responded 2.8, 2.3, and 2.2-fold to lovastatin treatment, with p-values of 0.02, 0.04 and 0.04 respectively. Construct E, which has had the SRE and Sp1-like site deleted, did not respond to lovastatin (Ratio = 1.3, p=0.44). Unex pectedly, the shortest prom oter construct (F) which extends from -58 to +70 had a significant sterol response (2.6 fold, p= 0.003), despite low overall activity. This region does not contain any pr eviously identified response elements.
R.L.U. ratioBCDEF 024681012 FIGURE 22. Response of the HMG-CoA reductase promoter to lovastatin of cholesterol feeding. Rats were fed either lovastatin (gray bars) or cholesterol (black bars) for three days. Animals were then electroporated with constructs B-F, along with Renilla luciferase. Animals were maintained on their respective diets overnight, and livers were harvested the next day for luciferase measurements. The data is represented as the mean ratio of firefly to Renilla luciferase +/standard error of the mean. There were a total of 10 animals in each group. In order to verify that lovastatin and cholesterol actually had an impact on the expression of endogenous HMGR, we isolated microsomes from the livers of electroporated animals fed a normal diet, 1% cholesterol, or 0.02 % lovastatin. These samples were subjected to Western Blotting and probed with antibodies to HMGR and -actin. As seen in figure 23, cholesterol drastically reduced HMGR protein levels, while lovastatin resulted in a remarkable induction of HMGR protein (Figure 23). This is consistent with previous observations in rats, and 86
shows that surgical manipulation of the animal does not adversely affect liver-wide expression or regulation of HMGR. HMGR-ActinNormalCHL FIGURE 23. Effect of lovastatin or cholesterol feeding on HMG-CoA reductase immunoreactive protein and mRNA levels. Western blot of HMG-CoA reductase in microsomes from rats fed either normal chow (N), 1% cholesterol (CH), or 0.02% lovastatin (L) for three days. Each lane represents a separate animal. -actin was blotted as a loading control. Since lovastatin and cholesterol can have effects on translation of the HMGR message, or stability of the protein, measurements of mRNA levels were also made. Real time RT-PCR was performed on RNA isolated from livers of electroporated animals. Animals were fed either 1% cholesterol (CH) or 0.04% lovastatin (L) for 3 days. In agreement with previous nuclear run-on experiments (39), lovastatin feeding significantly up regulated HMGR mRNA levels greater than 6-fold (Figure 24, p=6.0 x 10 -6 ). 87
0510152025 LCHrelative mRNAp = 0.000006HMGR FIGURE 24. Effect of lovastatin or cholesterol feeding on HMGR mRNA levels. Total RNA was isolated from the livers of electroporated animals fed diets containing 0.04% lovastatin or 1% cholesterol. Relative mRNA levels were determined by Real time RT-PCR using 18S ribosomal RNA for normalization. Data is expressed as the mean +/standard error of the mean for 5 animals per group. In the same experiment, real time RT-PCR was performed using primers to detect the HMG-CoA synthase (HCS) message. HCS is a classic sterol-regulated gene with a consensus SRE in its promoter(107). This promoter has been shown to respond to fluvastatin in CHO cells, and the effect has been attributed to increased SREBP processing (108). The relative difference in HCS between the two groups was 6.3 fold, with a p-value of 0.0025 (Figure 25). Interestingly, HMGR and HCS both showed a nearly identical fold-induction by lovastatin, suggesting a common mechanism may be responsible for both. 88
0510152025 relative mRNALCHp= 0.0025HCS FIGURE 25. Effect of lovastatin or cholesterol feeding on HMG-CoA synthase mRNA levels. Total RNA was isolated from the livers of electroporated animals fed diets containing 0.04% lovastatin or 1% cholesterol. Relative mRNA levels were determined by Real time RT-PCR using 18S ribosomal RNA for normalization. Data is expressed as the mean +/standard error of the mean for 5 animals per group. 89
90 DISCUSSION These studies address the regulat ion of the HMG-CoA reductase promoter by insulin and sterols in live animals. Here we present the first evidence that insulin acts to directly increas e transcription of the HMG-CoA reductase gene in rat liver. Diabetic rats have lower rates of HMG-CoA reductase transcription than normal rats. With only two hours of insulin treatment, transcription was restored to normal. Previo us work from our lab showed that the corresponding increase in mRNA could be accomplished even in the presence of cycloheximide (74). Taken together, these results suggest that insulin acts directly to stimulate the HMG-CoA reductase promoter, and does not require new protein synthesis. In vivo footprinting revealed numerous protections and enhancements throughout the HMGR promot er. The most pronounced of these was at the CRE, which was occupied under all conditions tested. EMSA analysis confirmed that CREB-1 present in nuclear extracts from normal or diabetic rat livers could bind to this element in vitro, in agreement with observations in FRTL-5 cells (83) ChIP analysis of rat liver confirms the prev ious finding that CREB is bound to the HMGR promoter in vivo (82). Given the overwhelming and invariant occupancy of the CRE in vivo, it seems unlikely that CREB binding is the regulated event in
91 insulin activation. This may differ from st erol regulation in CHO cells, in which SREBP binding has been shown to select ively recruit CREB to the promoter (101). Although we previously showed t hat the CRE was required for insulin activation of this promoter in H4IIE cells (75), in vivo occupancy of this site did not vary in rat liver. In addition, footprints from H4IIE cells lacked important protections and enhancements seen in thos e obtained from liver. Based on this information, it is clear t hat insulin regulation in cultured rat hepatoma cells differs from the physiological regulati on of this gene seen in animals. The enhancement at -138 in normal foot prints was not seen in diabetic samples. Given the GC-rich content of the nearby sequence, it is likely that this may be due to binding of an Sp1-like factor. In fact, strong in vitro binding activity was observed with the probe from to The sequence GGGCGGCTT is a close match to the consensus binding sequence for Sp1. Though generally regarded as a more basal transcription factor, Sp1 has been invoked in the insulin regulation of several genes, including SREBP-1a (101). Four of the five diabetic animals showed a particularly striking enhancement at -142. This enhancement was never seen under the other conditions examined, including cholesterol and lovastat in treatment. In addition, only diabetic samples showed a change in th e DMS reactivity of the SRE. This enhancement at -161 of the SRE coinci ded with the enhancement at -142, and may be a result of binding of a repressive factor in the -161/ -142 region. Binding
92 of a factor in this region could distort the DNA in such a way that both -142 and 161 are more susceptible to dimethyl sulfate attack. In addition, this factor may preclude binding to the Sp1 site downstr eam of This competition would explain why enhancements at were only seen in the normal animals (due to binding at Sp1 sites), and the enhancem ent at only in the diabetics. When animals were fed either lovastatin or cholesterol, the footprint looked essentially identical to that seen in normal animals. This is peculiar because a similar dose of lovastatin was shown to elevate HMGR transcription 46 fold (39). Three possible reasons for t he lack of occupancy at the SRE come to mind: 1) SREBP binding has been proposed to be a rapid and transient event. It is known that SREBPs are by themselv es weak binders of DNA, suggesting that nearby factors are needed to stabilize t hem (82). Therefore SREBPs may not remain bound to the promoter long enough to show up in the footprints. 2) Perhaps not all of the cells in the li ver respond to lovastatin. It has been previously shown by immunostaining of livers from rats fed me vinolin (lovastatin), that detectable HMGR expression is clustered around the blood supply (109). Uneven distribution of the drug could m ean that some cells show a drastic upregulation of HMGR message, while a larger percentage remain unaffected. Since in vivo footprinting examines the net promoter occupancy of a population of all liver cells, the effect could be muted even though mRNA is markedly higher. 3) Lovastatin activation of this gene could occur through another site elsewhere in the promoter. HMGR is known to have a variant SRE, with lower affinity for
93 SREBPs than those found in the LDL re ceptor and HMG-CoA synthase genes. It has also been demonstrated t hat SREBP-1 can bind to the HMGR promoter at sites other than the SRE (81) In any case, the data str ongly suggests that insulin activation of this promoter occurs by a me chanism distinct from sterol regulation. The protections at -71 / -70 were not observed in footprints from diabetic animals. It was found that NF-Y from nuclear extracts could bind to this element. Curiously, in vitro binding of NF-Y from nuclear extracts did not vary with diabetes suggesting that other factor s may be necessary to stabilize its interaction with the promoter in vivo This is the first identification of this proximal NF-Y site in the HMGR promoter. It is al so clear that this site is of particular functional importance. Mutations to this site substantially decreased overall transcription in H4IIE cells, as well as in vivo in the liver. The area around -70 is the classical position for a CCAAT box (110). In this case, the ATTGG sequence recognized by NF-Y is actually an invert ed CAATT box. Given its proximity to the transcription start site, recruitment of NF-Y is probably a key event in insulin activation of HMGR. The areas of protection or enhancement identified in this study generally correspond with the large protected regions seen previously in DNase I footprinting studies of the hamster promoter (76). Previous in vivo footprints of the human promoter in HepG2 cells did not find any differences with insulin treatment, despite a 1.5 fold increas e in mRNA (89). These previous in vivo
94 studies also identified the SRE as a pr otected region, something that was not observed in rat liver. T hese reports were useful in both helping us design the experiments, and in allowing us to com pare results from cultured tumor cells to rat liver. Using the information obtai ned from these expe riments, we can construct a basic model of promoter o ccupancy in normal and diabetic rats (Fig. 26). CREB is bound to the promoter at t he CRE in both normal and diabetic rats. The previously described NF-Y site around -189/-190 is occupied in both situations. Sp1 or a related factor binds in the /-119 region, possibly accounting for the enhancement at -138 in normal rats. This binding is likely prevented by the presence of a repressi ve factor that occupies the region between -142 and -161 in the diabetic anima ls. The SRE did not have noticeable protections. The diagram also shows a newly identified NF-Y site at -70/-65 that is preferentially occupied in the normal animals.
SRE CRE -200-180-160-140-120-100-80 TATT CREB SRE NF-YCRE ++TATT CREB Sp1-like? SREBP Repressor ?NormalDiabetic-138-142 NF-Y NF-Y FIGURE 26. Model of the HMGR promoter in normal and diabetic rat liver. Schematic of the HMGR promoter showing the major elements that are occupied in vivo in normal or diabetic rats, along with putative factors bound to those elements. The nucleotide positions from -200 to -60 are drawn to scale relative to the transcription start site, and are shown on the ruler at the bottom of the figure. Through the use of in vivo electroporation, we were able to examine the function of the footprinted regions in driving HMGR transcription in the liver. In agreement with previous studies, about 300 bp of upstream sequence was needed for high level expression. There was no benefit from additional flanking sequence extending back to -770 or to +441 past the transcription start site. This is consistent with previous studies in cultured cells (32). In agreement with in vivo footprinting studies, inclusion of the region from -176 to -123 (contains SRE and Sp1-like sites) appeared to have a minor repressive effect. This area was previously found to have protein binding only in situations where the gene was 95
96 not transcriptionally active (106). T he CRE and NF-Y sites around -100 and -70 respectively, likely explain the increase in activity seen in construct E. Deletion of the region from -123 to -58 dr astically reduced promoter activity, consistent with a critical role for CREB and NF-Y in activation of the HMGR gene (82,83,106). In an effort to map the insulin re sponsive regions of the HMGR promoter in the liver, we performed numerous el ectroporation experiments. Normal and diabetic rats were electroporated with the HMGR promoter luciferase constructs shown in Figure 18. In most cases there was no difference in promoter activity between these groups (data not shown). This was found to be true comparing diabetic animals to those replenished with insulin. On occasion, insulin actually appeared to repress promoter activity in the reporter constructs. This does not agree with HMGR protein or mRNA measurements, Furthermore, nuclear run-on assays, a direct measure of rates of transcription, showed a clear effect of insulin (Figure 4). There are at least three like ly explanations for this discrepancy: 1) In vivo electroporation disturbs the hepatocytes in such a way that insulin signal transduction is negatively affected. Th is seems reasonable as the insulin receptor functions as a dimer in the plas ma membrane. This would contrast with sterol regulation that is not dependent on an extracellular signaling cascade. 2) The HMGR insulin response may require features of the native chromatin environment which cannot be replicated by a transiently transfected reporter plasmid. This is a difficult theory to test but has proven to be the case for other genes (111). Due to the negat ive nature of such data, such a problem is probably
97 under-reported in the literature. 3) The in sulin responsive element(s) lies outside of the cloned promoter constructs. This seems to be the simplest and most likely explanation. Changes even at distant bindi ng sites can have a substantial impact on transcription. This may occur thr ough changes in chromatin structure or through recruitment of other factors to the promoter. Pe rhaps recruitment of NF-Y in particular is an important event. Another intriguing observation is the loss of HMGR diurnal variation in PPAR-alpha knockout mice. The diurnal variation correlates well with the major and minor feeding times for t he animals. Most, if not all, of this affect may be explained by the accompanying changes in insulin levels. Therefore it is reasonable to hypothesize that PPAR-alpha is in fact required for insulin regulation of the gene. This is inte resting because PPA R-alpha is a ligandactivated nuclear receptor. It is possible that a fatty acid or other metabolite is released when insulin levels rise. This could then bind to PPAR-alpha and activate the gene. A similar mechanism has been proposed for insulin regulation of the SREBP-1 promoter by LXR (112). Indeed, several fatty acids have already been identified as ligands for PPAR-alpha (113-115). We have searched as far as 5 kb upstream of the HMGR gene for consensus PPAR-alpha sites, and there do not appear to be any. It is possibl e that PPAR-alpha functions through a variant site, but without more upstream promoter sequence cloned, this cannot be tested. It is also possible that PPARalpha knockout mice lose the diurnal
98 variation simply because it is requir ed for the expression of another insulinresponsive factor. It has also been suggested that bile ac ids play a role in transcriptional regulation of HMGR. Administration of an FXR agonist, GW4064 to mice potently suppresses HMGR mRNA levels in the live r (70). Bile acids bind to the farnesoid X receptor (FXR). This is a ligand-acti vated nuclear receptor that affects a number of important genes in bile ac id metabolism. FXR activates small heterodimeric partner (SHP), by increasing transcription of this gene. SHP is a non-ligand binding nuclear re ceptor partner that can pair with other nuclear receptors to repress transcription, such as the liver receptor homologue-1 / fetoprotein transcription factor (LRH1/FTF). It has been proposed that under conditions of bile acid accumulation, SH P will bind to LRH-1/FTF and prevent its activation of the HMGR promoter (70) Deletion of the region containing the alleged LRH-1/FTF site significantly reduced HMGR promoter activity (Figure 20). However, we saw no change in promoter activity when electroporated animals were fed either 1% cholic acid (bile acid ) or 2% colestipol (bile acid sequesterant) for three days (data not shown). The sterol response element is one of the key determinants of HMGR promoter activity in cultured cells (77) There has been a considerable amount of work in transgenic mouse models that point to its importance in vivo For example, mice overexpressing the matu re form of SREBP-2 have drastically elevated HMGR levels, as well as increased cholesterol synthesis (37). These
99 animals rapidly develop fatty livers, show ing that when over expressed, SREBP2 can activate the hepatic HMGR promoter. Recent evidence from the liver specific SCAP knockout mouse points to the importance of SREBPs in maintaining appropriate levels of HMGR expression (116). These animals lack SCAP, the accessory protein required for processing of SREBPs to their mature forms. Thus, SREBPs can not be transporte d to the Golgi for processing, and consequently they do not activate tar get genes. These animals exhibit an 80% reduction in HMGR expression. In addi tion, both SREBP-2 and HMGR knockout mice are embryonic lethal (117,118) s uggesting that the lethal effect of SREBP2 deletion might be ascribed to its inabi lity to activate the HMGR gene during development. However, until now, the sterol-responsiveness of the HMGR promoter has not been examined in the context of the liver. In order to examine sterol regulation of the promot er in the liver, we fed rats diets containing cholesterol or lo vastatin. These diets were intended to maximize the possible sterol dependent di fferences in HMGR transcription. It should be noted that this dose of cholesterol results in little or no change of HMGR mRNA levels compared to normal chow despite larger differences in total liver cholesterol (99). In contrast, lovast atin significantly upregulates transcription of the HMGR gene (39), in response to t he temporary depleti on of intracellular cholesterol (Due to the ensuing compens ation by HMGR and LDLR, total liver cholesterol levels are seldom changed by statin treatment). It is therefore likely that the difference in promoter activity between the cholesterol and lovastatin-fed
100 animals is largely due to activation by lovastatin, rather than repression by dietary cholesterol. With this in mind, in vivo electroporation showed that all three constructs containing the HMGR SRE re sponded to lovastatin. The promoter construct from -123 to +70 which la cks the SRE did not have a significant response to lovastatin. This finding is consistent with the predicted 2-3 fold response to sterol depletion seen for the HM GR promoter in cultured cells (78). It is unclear why the shortest promoter construct from -58 to +70 responded to lovastatin, as this promoter construc t lacks any of the pr eviously identified enhancer or sterol response elements. It is worth noting that this construct extends out to +70, past t he transcription start site. Th is represents the majority of the 100 bp of 5 untranslated regi on present in rat liver HMGR mRNA (unpublished observations). This sequence contains several predicted hairpin loops (119) and has been suggested to play a role in translational control of the gene either by a mevalonate-derived nons terol product (120), or by dietary cholesterol (27). The effect of lovastat in or cholesterol feeding on HMGR protein levels was greater than the observed increas e in mRNA levels. This is consistent with the considerable amount of posttranscriptional regulat ion of HMGR by dietary cholesterol. The observed effect of sterol depletion on promoter activity was less than the 6-fold effect on mRNA seen by real time RT-PCR. There may be other regulatory elements that help enhance the promot ers response to sterol depletion that are not incl uded in our promoter constructs. One such candidate SRE exists at -820 to -813 with t he consensus sequence GTGGGGTG. This differs from the SRE in the proximal promoter with a single mismatch-
101 GTGC GGTG. The location of this putativ e distal SRE calls for further examination and functional analysis. Alter natively, it is also possible that lovastatin treatment may have an effect on the stability of the HMGR mRNA. We demonstrate that the H MGR promoter responds to statin treatment in the livers of live animals. The exact mechani sm responsible for this effect is not known. Based on other work in the literatu re, there is at least one likely scenario that would explain the lovastatin-mediated increase in HMGR transcription. Lovastatin treatment slows de novo cholesterol synthesis by inhibiting HMGR. The decrease in cholesterol synthesis transiently reduces intracellular cholesterol levels. The hepatocytes must then respond by obtaining more cholesterol. This compensation can occur through increased synthesis, and/or by receptormediated endocytosis (10). To accomplish this, low sterol levels cause SCAP to dissociate from Insig. The SCAP/SREBP co mplex is then carried from the ER to the Golgi for processing. After proteolytic cleavage, the mature SREBPs migrate to the nucleus to activate target g enes (31). HMGR and LDLR are two well established SREBP target genes (121,122) The nuclear SREBPs then bind to the HMGR SRE and activate tran scription. This is consistent with the fact that the three largest promoter constructs, all containing the SRE, responded positively to lovastatin (Figure 22). While largely based on cell culture studies, there is some supporting in vivo evidence for this model.
102 Previous studies in our lab did not detect changes in nuclear SREBP levels in the livers of rats fed cholesterol or lovastatin (123). However, several other studies have addressed sterol-dependent processing of SREBPs in the liver. Hamsters fed a basal diet of 4% co lestipol show greater levels of nuclear SREBP-2 when fed increasi ng amounts of mevinolin (lovastatin) (124). In addition, hamsters fed a chow diet with as little as 0.01% cholesterol had decreased SREBP cleavage and nuclear ab undance (125). Mice are also known to exhibit a statin effect on nuclear SREB P-2 levels in the liver. There was about a two-fold increase in mice fed 0.2% lo vastatin (126). Likewise, fluvastatin has been shown to increase the mRNA levels of both HMG-CoA reductase and HMG-CoA synthase in rat liver (108). The effect of fluvastatin on the HMG-CoA synthase promoter was dependent on the presence of intact SREs and NF-Y boxes. N -acetyl-leucyl-leucyl-norleucyl (ALLN) a protease inhibitor that prevents degradation of SREBPs, could also mimic the effect of fluvastatin on the HMGCoA synthase promoter. The authors attribut ed the fluvastatin e ffect to increased SREBP cleavage caused by cholesterol depletion. Since HMGR and HCS mRNA were elevated to the same degree with statin treatment (Figure 25), it is likely that the same factor is responsible. It should be noted that while total liver cholesterol levels do not vary greatly with statin treatment (127), ultimately it is the intracellular cholesterol in the regul atory pool that makes the difference in controlling transcription.
103 In summary, we have examined the occ upancy of the HMGR promoter in live animals. Diabetes dramatically alte rs the pattern of transcription factor binding to the promoter. Both CREB and NF-Y play important roles in the regulation of the gene. A newly identified NF-Y site at -70/-65, is required for efficient transcription, and recruitment of th is factor to the pr oximal promoter is likely a key event in insulin activation. Data suggests that expression of HMGR plays an important role in determining ones susceptibility to dietary cholesterol (128). HMGR expression is severely reduced in diabetic rats, and these animals have a corresponding loss of cholesterol buffering capacity (129). Likewise, HMGR may be insulin-regulated in humans based on serum and urinary mevalonate measurements (73). If insulin -regulation of the gene is compromised, cholesterol balance could be adversely affe cted in these individuals. This is of great relevance to the millions of people who suffer from diabetes today. This research represents a step forwar d towards understanding the complex relationships between diabetes and cardiovascular disease. Promoter regions responsible for dr iving HMGR expression in the liver have been evaluated. We have found that the HMGR promoter can be sterolresponsive in the liver, as it is in cult ured cells. The ability to upregulate HMGR in response to statins may be a factor in determining their cholesterol-lowering capacity. A recent study has identifi ed polymorphisms in the HMGR gene that are associated with decreased response to a statin (130). These polymorphisms were found in intronic regions of the gene, and the mechanism of these effects is
104 not yet known. It will be interesti ng to see if polymor phisms in the HMGR promoter, or other genes in the SREBP pathway, will be good predictors of a persons response to cholesterol-lowering drugs. We hope that the basic research presented here will help guide fu ture investigators in discovering more effective treatments for hypercholestero lemia and cardiovascular disease.
105 REFERENCES 1. Turley, S. D., and Di etschy, J.M. (1988) in Biology and Pathobiology Raven Press Ltd., New York 2. Goetz, J. A., Suber, L. M., Zeng, X., and Robbins, D. J. (2002) Bioessays 24, 157-165 3. Porter, F. D. (2003) Curr Opin Pediatr 15, 607-613 4. Kannel, W. B., Dawber, T. R., Friedman, G. D., Glennon, W. E., and McNamara, P. M. (1964) Ann Intern Med 61, 888-899 5. Altmann, S. W., Davis, H. R., Jr., Zhu, L. J., Yao, X., H oos, L. M., Tetzloff, G., Iyer, S. P., Maguire, M., Golovko, A., Zeng, M., Wang, L., Murgolo, N., and Graziano, M. P. (2004) Science 303, 1201-1204 6. Cooper, A. D. (1997) J Lipid Res 38, 2173-2192 7. Galatola, G., Jazrawi, R. P., Bridges C., Joseph, A. E., and Northfield, T. C. (1991) Gastroenterology 100, 1100-1105 8. Sudhop, T., Sahin, Y., Lindenthal, B., Hahn, C., Luer s, C., Berthold, H. K., and von Bergmann, K. (2002) Gut 51, 860-863 9. Brewer, H. B., Jr. (1981) Klin Wochenschr 59 1023-1035
106 10. Goldstein, J. L., Hobbs, H.H., Brown, M.S. (1995) in The Metabolic and Molecular Basis of Inherited Disease (Scriver, C. R., Beaudet, A.L., Sly, W.S., Valle, D., ed) Vol. 2, pp. 1981-2030, 3 vols., McGraw-Hill, Inc. 11. Llorente, V., and Badimon, L. (1998) Rev Esp Cardiol 51 633-641 12. Breslow, J. L. (1995) in The Metabolic and Molecu lar Basis of Inherited Disease (Scriver, C. R., Beaudet, A.L., Sly, W.S., Valle, D., ed) Vol. 2, pp. 2031-2052, 3 vols., McGraw-Hill, Inc., New York 13. Dietschy, J. M., Turley, S. D., and Spady, D. K. (1993) J Lipid Res 34, 1637-1659 14. Dietschy, J. M., and Si perstein, M. D. (1967) J Lipid Res 8, 97-104 15. Dietschy, J. M., and Wilson, J. D. (1968) J Clin Invest 47 166-174 16. Andersen, J. M., Turley, S. D., and Dietschy, J. M. (1982) Biochim Biophys Acta 711, 421-430 17. Dietschy, J. M., and McGarry, J. D. (1974) J Biol Chem 249, 52-58 18. Lakshmanan, M. R., and Veech, R. L. (1977) J Biol Chem 252, 4667-4673 19. Dietschy, J. M., and Spady, D. K. (1984) J Lipid Res 25, 1469-1476 20. Turley, S. D., Andersen, J. M., and Dietschy, J. M. (1981) J Lipid Res 22, 551-569 21. Spady, D. K., and Dietschy, J. M. (1983) J Lipid Res 24, 303-315 22. Dietschy, J. M. (1984) Klin Wochenschr 62, 338-345 23. Roitelman, J., Olender, E. H., Bar-Nun, S., Dunn, W. A., Jr., and Simoni, R. D. (1992) J Cell Biol 117 959-973 24. Brown, M. S., and Go ldstein, J. L. (1980) J Lipid Res 21, 505-517
107 25. Sato, R., Goldstein, J. L., and Brown, M. S. (1993) Proc Natl Acad Sci U S A 90, 9261-9265 26. Beg, Z. H., Stonik, J. A ., and Brewer, H. B., Jr. (1987) Metabolism 36, 900-917 27. Chambers, C. M., and Ness, G. C. (1998) Arch Biochem Biophys 354, 317-322 28. Schoenheimer R, B. F. (1933) J. Biol. Chem. 103, 439-448 29. Gould, R. G., Taylor, C. B., Hagerman, J. S., Wa rner, I., and Campbell, D. J. (1953) J Biol Chem 201, 519-528 30. Liscum, L., Luskey, K. L., Chin, D. J., Ho, Y. K., Goldstein, J. L., and Brown, M. S. (1983) J Biol Chem 258, 8450-8455 31. Brown, M. S., and Go ldstein, J. L. (1999) Proc Natl Acad Sci U S A 96, 11041-11048 32. Osborne, T. F., Goldstein, J. L., and Brown, M. S. (1985) Cell 42, 203-212 33. Hua, X., Sakai, J., Brown, M. S., and Goldstein, J. L. (1996) J Biol Chem 271, 10379-10384 34. Yang, T., Espenshade, P. J., Wright, M. E., Yabe, D., Gong, Y., Aebersold, R., Goldstein, J. L., an d Brown, M. S. (2002) Cell 110, 489-500 35. Dobrosotskaya, I. Y., Goldstein, J. L., Brown, M. S., and Rawson, R. B. (2003) J Biol Chem 278 35837-35843 36. Shimano, H. (2001) Prog Lipid Res 40, 439-452 37. Horton, J. D., Shimomur a, I., Brown, M. S., Hamme r, R. E., Goldstein, J. L., and Shimano, H. (1998) J Clin Invest 101 2331-2339
108 38. Ness, G. C., Eales, S., Lopez, D., and Zhao, Z. (1994) Arch Biochem Biophys 308 420-425 39. Lopez, D., Chambers, C. M., and Ness, G. C. (1997) Arch Biochem Biophys 343 118-122 40. Chambers, C. M., and Ness, G. C. (1997) Biochem Biophys Res Commun 232, 278-281 41. Gil, G., Faust, J. R., Chin, D. J., Go ldstein, J. L., and Brown, M. S. (1985) Cell 41, 249-258 42. Jingami, H., Brown, M. S., Goldstei n, J. L., Anderson, R. G., and Luskey, K. L. (1987) J Cell Biol 104, 1693-1704 43. Edwards, P. A., Lan, S. F., Tanaka, R. D., and Fogelma n, A. M. (1983) J Biol Chem 258, 7272-7275 44. Sever, N., Yang, T., Br own, M. S., Goldstein, J. L., and DeBose-Boyd, R. A. (2003) Mol Cell 11 25-33 45. Song, B. L., Javitt, N. B., and DeBose-Boyd, R. A. (2005) Cell Metab 1, 179-189 46. Engelking, L. J., Liang, G., Hammer R. E., Takaishi, K., Kuriyama, H., Evers, B. M., Li, W. P., Horton, J. D., Goldstein, J. L., and Brown, M. S. (2005) J Clin Invest 115 2489-2498 47. Van Heek, M., France, C. F., Compt on, D. S., McLeod, R. L., Yumibe, N. P., Alton, K. B., Sybertz, E. J. and Davis, H. R., Jr. (1997) J Pharmacol Exp Ther 283, 157-163
109 48. Gagne, C., Bays, H. E., Weiss, S. R., Mata, P., Quinto K., Melino, M., Cho, M., Musliner, T. A ., and Gumbiner, B. (2002) Am J Cardiol 90 10841091 49. Ness, G. C., Holland, R. C., and Lopez, D. (2006) Exp Biol Med (Maywood) 231, 559-565 50. Tobert, J. A., Hitzenberger, G., Kukove tz, W. R., Holmes, I. B., and Jones, K. H. (1982) Atherosclerosis 41, 61-65 51. (1994) Lancet 344, 1383-1389 52. Nissen, S. E., Tuzcu, E. M., Sch oenhagen, P., Brown, B. G., Ganz, P., Vogel, R. A., Crowe, T., Howard, G., Cooper, C. J., Brodie, B., Grines, C. L., and DeMaria, A. N. (2004) Jama 291, 1071-1080 53. Ness, G. C., Chambers, C. M., and Lopez, D. (1998) J Lipid Res 39, 75-84 54. Sample, C. E., Pendleton, L. C., and Ness, G. C. (1987) Biochemistry 26, 727-731 55. Simonet, W. S., and Ness, G. C. (1988) J Biol Chem 263, 12448-12453 56. Ness, G. C., Pendleton, L. C., Li, Y. C., and Chia ng, J. Y. (1990) Biochem Biophys Res Commun 172, 1150-1156 57. Shin, D. J., and Osborne, T. F. (2003) J Biol Chem 278 34114-34118 58. Simonet, W. S., and Ness, G. C. (1987) Biochem Biophys Res Commun 146, 1033-1039 59. Simonet, W. S., and Ness, G. C. (1989) J Biol Chem 264, 569-573 60. Shapiro, D. J., and Rodwell, V. W. (1969) Biochem Biophys Res Commun 37, 867-872
110 61. Parker, T. S., McNamara, D. J., Br own, C., Garrigan, O., Kolb, R., Batwin, H., and Ahrens, E. H., Jr. (1982) Proc Natl Acad Sci U S A 79, 3037-3041 62. Easom, R. A., and Zammit, V. A. (1987) Biochem J 241 183-188 63. Ness, G. C., and C hambers, C. M. (1996) Arch Biochem Biophys 327, 4144 64. Clarke, C. F., Fogelman, A. M., and Edwards, P. A. (1984) J Biol Chem 259, 10439-10447 65. Patel, D. D., Knight, B. L., Wiggi ns, D., Humphreys, S. M., and Gibbons, G. F. (2001) J Lipid Res 42 328-337 66. Ness, G. C., Eales, S. J., Pendl eton, L. C., and Smith, M. (1985) J Biol Chem 260 12391-12393 67. Ness, G. C., Zhao, Z., and Wiggins, L. (1994) J Biol Chem 269 2916829172 68. Duckworth, P. F., Vlahcevic, Z. R., St uder, E. J., Gurley, E. C., Heuman, D. M., Beg, Z. H., and Hy lemon, P. B. (1991) J Biol Chem 266, 9413-9418 69. Tanaka, R. D., Edwards, P. A., L an, S. F., Knoppel, E. M., and Fogelman, A. M. (1982) J Lipid Res 23, 1026-1031 70. Datta, S., Wang, L. Moore, D. D., and Os borne, T. F. (2006) J Biol Chem 281, 807-812 71. Nepokroeff, C. M., Lakshmanan, M. R., Ness, G. C., Muesing, R. A., Kleinsek, D. A., and Porter, J. W. (1974) Arch Biochem Biophys 162 340344
111 72. Lakshmanan, M. R., Dugan, R. E ., Nepokroeff, C. M., Ness, G. C., and Porter, J. W. (1975) Arch Biochem Biophys 168, 89-95 73. Miettinen, T. A., Gylling, H., Tuomi nen, J. A., Simonen, P., and Koivisto, V. (2004) Duodecim 120 721-723 74. Ness, G. C., Wiggins L., and Zhao, Z. (1994) Arch Biochem Biophys 309, 193-194 75. Osborne, A. R., Pollock, V. V., Lagor, W. R., and Ness, G. C. (2004) Biochem Biophys Res Commun 318 814-818 76. Osborne, T. F., Gil, G ., Brown, M. S., Kowal, R. C., and Goldstein, J. L. (1987) Proc Natl Acad Sci U S A 84, 3614-3618 77. Osborne, T. F., Gil, G., Goldstei n, J. L., and Brown, M. S. (1988) J Biol Chem 263 3380-3387 78. Osborne, T. F. (1991) J Biol Chem 266 13947-13951 79. Sudhof, T. C., Russell, D. W., Brown, M. S., and Goldstein, J. L. (1987) Cell 48, 1061-1069 80. Metherall, J. E., Goldstein, J. L ., Luskey, K. L., and Brown, M. S. (1989) J Biol Chem 264, 15634-15641 81. Vallett, S. M., Sanchez, H. B., Rosenfeld, J. M., and Osborne, T. F. (1996) J Biol Chem 271 12247-12253 82. Bennett, M. K., and Osborne, T. F. (2000) Proc Natl Acad Sci U S A 97, 6340-6344 83. Bifulco, M., Perillo, B., Saji, M., Laezza, C., Tedesco, I., Kohn, L. D., and Aloj, S. M. (1995) J Biol Chem 270 15231-15236
112 84. Fernyhough, P., Willars, G. B., Linds ay, R. M., and Tomlin son, D. R. (1993) Brain Res 607, 117-124 85. Finlay, D., Healy, V., Furlong, F., O'Connell, F. C., Keon, N. K., and Martin, F. (2000) Cell Death Differ 7, 302-313 86. Beirne, O. R., and Watson, J. A. (1976) Proc Natl Acad Sci U S A 73 2735-2739 87. Hentosh, P., Yuh, S. H., Els on, C. E., and Peffley, D. M. (2001) Mol Carcinog 32 154-166 88. Elson, C. E., Peffley, D. M. Hentosh, P., an d Mo, H. (1999) Proc Soc Exp Biol Med 221 294-311 89. Lloyd, D. B., and T hompson, J. F. (1995) J Biol Chem 270, 25812-25818 90. Heller, L. C., Ugen, K., and Heller, R. (2005) Expert Opin Drug Deliv 2, 255-268 91. Heller, R. (2003) Preclinica 1, 198-208 92. Mano, H., Mori, R., Ozawa, T., Take yama, K., Yoshizawa, Y., Kojima, R., Arao, Y., Masushige, S., and Kato, S. (1994) J Biol Chem 269, 1591-1594 93. Mueller, P. R., and Wold, B. (1989) Science 246, 780-786 94. Yu, C. L., Meyer, D. J., Campbell, G. S., Larner, A. C., Carter-Su, C., Schwartz, J., and Jove, R. (1995) Science 269, 81-83 95. Latasa, M. J., Griffin, M. J., Moon, Y. S., Kang, C., and Sul, H. S. (2003) Mol Cell Biol 23, 5896-5907 96. Heller, R., Jaroszeski, M., Atkin, A., Moradpour, D., Gilbert, R., Wands, J., and Nicolau, C. (1996) FEBS Lett 389, 225-228
113 97. Wang, L., and Menon, K. M. (2005) Endocrinology 146, 423-431 98. Bossard, P., McPherson, C. E., and Zaret, K. S. (1997) Methods 11, 180188 99. Ness, G. C., Keller, R. K., and Pendleton, L. C. (1991) J Biol Chem 266, 14854-14857 100. Kriwacki, R. W., Schultz, S. C., St eitz, T. A., and Cara donna, J. P. (1992) Proc Natl Acad Sci U S A 89, 9759-9763 101. Ngo, T. T., Bennett, M. K., Bourgeois, A. L., Toth, J. I., and Osborne, T. F. (2002) J Biol Chem 277 33901-33905 102. Haviv, Y. S., van Houdt, W. J., Lu, B., Curiel, D. T., and Zhu, Z. B. (2004) Mol Cancer Ther 3, 687-691 103. Al-Dosari, M. S., Knapp, J. E., and Liu, D. (2006) Mol Pharm 3, 322-328 104. Gilbert, R. A., Jaroszeski, M. J., a nd Heller, R. (1997) Biochim Biophys Acta 1334 9-14 105. Jaroszeski, M. J., Gilbert, R. A., and Heller, R. (1997) Biochim Biophys Acta 1334 15-18 106. Lagor, W. R., de Groh, E. D., and Ness, G. C. (2005) J Biol Chem 280, 36601-36608 107. Inoue, J., Sato, R. and Maeda, M. (1998) J Biochem (Tokyo) 123, 11911198 108. Mascaro, C., Ortiz, J. A., Ramos, M. M., Haro D., and Hegardt, F. G. (2000) Arch Biochem Biophys 374 286-292
114 109. Singer, II, Kawka, D. W., Kazazis, D. M., Alberts, A. W., Chen, J. S., Huff, J. W., and Ness, G. C. (1984) Proc Natl Acad Sci U S A 81, 5556-5560 110. Grosveld, G. C., Shewmaker, C. K ., Jat, P., and Flavell, R. A. (1981) Cell 25, 215-226 111. Cook, J. L., Zhang, Z., Al am, J., and Re, R. N. (1999) Oncogene 18, 23732379 112. Chen, G., Liang, G., Ou, J., Goldst ein, J. L., and Brown, M. S. (2004) Proc Natl Acad Sci U S A 101 11245-11250 113. Hostetler, H. A., Kier, A. B., and Schroeder, F. (2006) Biochemistry 45, 7669-7681 114. Hostetler, H. A., Petrescu, A. D., Kier, A. B., and Schroeder, F. (2005) J Biol Chem 280, 18667-18682 115. Schroeder, F., Huang, H., Hostetler, H. A., Petrescu, A. D., Hertz, R., BarTana, J., and Kier, A. B. (2005) Lipids 40, 559-568 116. Kuriyama, H., Liang, G., Engelking, L. J., Horton, J. D., Goldstein, J. L., and Brown, M. S. (2005) Cell Metab 1, 41-51 117. Shimano, H., Shimomura, I., Hammer, R. E., Herz, J., Goldstein, J. L., Brown, M. S., and Ho rton, J. D. (1997) J Clin Invest 100 2115-2124 118. Ohashi, K., Osuga, J., Tozawa, R., Kitamine, T., Yagyu, H., Sekiya, M., Tomita, S., Okazaki, H ., Tamura, Y., Y ahagi, N., Iizuka, Y., Harada, K., Gotoda, T., Shimano, H., Yamada, N., and Ishibashi, S. (2003) J Biol Chem 278 42936-42941
115 119. Reynolds, G. A., Basu, S. K., Osborne, T. F., Chin, D. J., Gil, G., Brown, M. S., Goldstein, J. L., an d Luskey, K. L. (1984) Cell 38, 275-285 120. Nakanishi, M., Goldstein, J. L., and Brown, M. S. (1988) J Biol Chem 263 8929-8937 121. Hua, X., Yokoyama, C., Wu, J., Briggs M. R., Brown, M. S., Goldstein, J. L., and Wang, X. (1993) Proc Natl Acad Sci U S A 90, 11603-11607 122. Yokoyama, C., Wang, X., Briggs, M. R., Admon, A., Wu, J., Hua, X., Goldstein, J. L., and Brown, M. S. (1993) Cell 75, 187-197 123. Lopez, D., and Ness, G. C. (1997) Arch Biochem Biophys 344, 215-219 124. Sheng, Z., Otani, H., Brown, M. S., and Goldstein, J. L. (1995) Proc Natl Acad Sci U S A 92, 935-938 125. Shimomura, I., Bashmako v, Y., Shimano, H., Horton, J. D., Goldstein, J. L., and Brown, M. S. (1997) Proc Natl Acad Sci U S A 94, 12354-12359 126. Rashid, S., Curtis, D. E., Garuti, R., Anderson, N. N., Bashmakov, Y., Ho, Y. K., Hammer, R. E., Moon, Y. A., and Horton, J. D. (2005) Proc Natl Acad Sci U S A 102, 5374-5379 127. Kamisako, T., and Ogawa, H. (2004) J Gastroenterol Hepatol 19, 879-883 128. Ness, G. C., and Ge rtz, K. R. (2004) Exp Biol Med (Maywood) 229 412416 129. Ness, G. C., and Ge rtz, K. R. (2004) Exp Biol Med (Maywood) 229 407411 130. Chasman, D. I., Posada, D., Subrahm anyan, L., Cook, N. R., Stanton, V. P., Jr., and Ridker, P. M. (2004) JAMA 291, 2821-2827
116 APPENDIX Manipulation of Hepatic HMGR Expression by Hydrodynamic Tail Vein Injection
117 Appendix: Manipulation of Hepatic HMGR Expression by Hydrodynamic Tail Vein Injection INTRODUCTION HMG-CoA reductase catalyzes the rate limiting step in cholesterol production. As such it is a critical cont rol point for metabolic regulation. HMGR expression varies greatly with respect to various dietary and hormonal states. Dietary cholesterol has a particularly potent ability to reduce HMGR activity and protein levels. Other factors such as diabetes, thyroid hormone status, age, and genetic variation all have an influenc e on the basal expression of HMGR. Interestingly, hepatic HMGR expressi on inversely correlates with serum cholesterol levels in different strains and species of rodents (1). Animals that express high levels of HMGR are resistant to the serum cholesterol raising action of dietary cholesterol. Accordingly, anima ls that express low levels of HMGR, are particularly susceptible to a dietary c holesterol insult. This holds true for situations where basal HMGR expre ssion is compromised due to hormonal deficiencies. Diabetic and hypothyroid rats have markedly diminished expression of HMGR, and increased sensitivity to diet ary cholesterol (2). For example, male Spargue Dawley diabetic rats may have a basal cholesterol level of 130 mg/dl, which can be rapidly elevated to 400 mg / dl or more within just a few days of cholesterol feeding. Normal animals on the same diet will exhibit virtually no change in serum cholesterol levels. In humans, Type I diabetics have lower rates
118 Appendix: (Continued) of cholesterol synthesis, and increased cholesterol absorption (3,4). This underscores the importance of HMGR in cardiovascular disease. It seems plausible that HMGR expression could play a protective role in guarding against dietary cholesterol. It has been proposed that HMGR can function as a cholesterol buffer to prevent undesirable incr eases in serum cholesterol (5). It is suspected that animals expressing high le vels of HMGR can more effectively downregulate protein levels. This would provide a mechanism for fine tuning cholesterol synthesis in response to di etary changes. Animals with highest basal HMGR expression, would have the greatest abi lity to adjust thei r synthetic rates. With the ability to drastically downregulate synthesis, these animals could better manage the influx of cholesterol from the diet. It is not known whether HMGR expressi on is a cause or effect of this resistance to cholesterol insult. The purpose of this study is to evaluate the role of HMGR in the maintenance of desirable serum cholesterol levels. To do this, we have use hydrodynamic tail vein in jection to introduce DNA plasmids and short interfering RNAs to the livers of live mice and rats. The objective was to artificially manipulate HMGR levels without the constrai nts of the normal regulatory mechanisms present in the endogenous gene. If HMGR is a cause of resistance to dietary cholesterol, t hen overexpression of HMGR should lend animals resistant. Conversely, knockdown of HMGR in the liver should lend animals more susceptible to dietary cholesterol insult.
119 Appendix: (Continued) METHODS AnimalsMale Balb/cJ mice were obtained from Jackson (figures 1,3). Male Sprague Dawley rats 50-75 g were ordered from Harlan (Figure 2). Male FVB (JAX control no. 001800, Figure 4,5,6) and male C57BL/6 mice (JAX control no. 000664, Figure 7,8) were obtained from Jackson labs. The Balb/c AnHSD mice used in Figures 9 and 10 were from Harlan. All animals except those in figures 1 and 3 were kept on reverse cycle lighting, and sacrificed at 9:00-10:00 am. This corresponds to the third to fourth hours of the dark cycle, when HMGR expression is at its diurnal high. All ex periments were carried out in accordance with the regulations of the USF Institutional Animal Care and Use Committee, protocol numbers 2440, 2976 and 2317. PlasmidspRed227 was obtained from ATCC. p5Luc3, a plasmid encoding firefly luciferase flanked by the rat HM GR 5 and 3 untranslated regions was generated using standard molecular biology techniques. Briefly, rat liver mRNA was reverse transcribed to make cDNA. Th is cDNA was used as a template for PCR to generate the 3 and 5 ends of the HMGR message based on the rat genome reference sequence. These were cloned into pGL3 control (contains an SV40 promoter) on either side of the luci ferase gene. On the 5 end, Hind III and
120 Appendix: (Continued) Nco I were used to ligate in the 5 end. Once this clone was obtained (p5UTR), it was digested with Xba I on the 3 end of the luciferase gene. The HMGR 3UTR was then ligated into this site using co mplementary Nhe I ends. Lindsey Jackson, cloned the 3 UTR of HMGR (from +2765 to +4202) to generate the final plasmid. phRL-CMV was obtained from Promega. siRNAsiRNA to the firefly luciferase gene has been described previously (6). siRNA to the rat HMGR 5 end of the message was designed using the Whitehead Institute of Bi omedical Research at MIT design tool at http://jura.wi.mit.edu/bioc/siRNAext/ home/php. This siRNA was generated by annealing two complementary oligonuc leotides of the sequence 5AAggacuguguagcuacaaugTT-3, and 5-cau uguagcuacacaguccTT-3 where the uppercase letters are deoxyribonucleot ides, and the lowercase letters are ribonucleotides. SiRNAs were ordered fr om IDT as annealed du plexes and later diluted in sterile saline. Hydrodynamic tail vein injectionPlasmids were diluted in 0.9% sterile saline to a volume corresponding to 9% of the body weight of the animal. For example, a 22 g mouse received plasmid DNA diluted in 1.8 ml of sterile saline. Animals were restrained using a standard rodent rest rainer. To visualize the tail veins, animals were placed under infrared light fo r about a minute. The tails were then
121 Appendix: (Continued) rinsed with 70% alchohol. For mice, the pl asmid solution was delivered using a 3 ml syringe equipped with a 27 G needle. For rats, the plasmid solution was delivered using a 10 ml syringe with a 23 G needle. In every case the plasmid solution was rapidly injected in 5-10 seconds. Animals that received less than 80% of the injection volume were excluded from the study. Injections for the animals in figures 1-3 were performed by Dr. Dexi Liu, an expert in the technique who graciously assisted and trained laboratory personnel. Other MethodsMicrosome preparation, western blotting, and serum cholesterol determinations were performed as described in the main methods section of the dissertation. The only exception is Figure 6. For this Western Blot, a large 7.5% gel was poured using the BioRad Prot ean II xi gel appa ratus. One hundred micrograms of microsomal protein was loaded in each lane. Following electrophoresis and transfer, the memb rane was cut above 100 KDa, and blotted for LDL Receptor and HMGR. The LDLR antibody has been described previously (7), and was used in a 1:5000 dilution.
122 Appendix: (Continued) RESULTS This study was undertaken to determine if HMGR expression is a cause or consequence of resistance to dietary cholesterol. An expression vector for hamster HMGR, pRed 227 was used in this study. This plasmid encodes a cDNA for HMG-CoA reductase of about 4.5 kb in size. The vector contains a constitutively active SV40 promoter that is not subject to r egulation by sterols. The mRNA generated includes a 163 bp 5 un translated region, the 2.7 kb coding sequence, and 1650 bp of the 3 untranslate d region (8). The first task was to determine if delivery of pRed227 to the li ver could result in noticeable HMGR overexpression. To accomplish this, BALB/c mice were given a hydrodynamic injection of 40 g of a control plasmid mi x (p5Luc3, phRL-CMV, siRNA to firefly luciferase) or one to overexpress HM GR (pRed 227, phRL-CMV). Eighteen hours later, the livers were harvested and western blotting was performed to measure HMGR protein levels (Figure 1). HMGR was substantially overexpressed in these two animals (lanes 5 and 6), compared to the control group (lanes 1-4). It should be noted that these animals were sacrificed near the diurnal low for HMGR activity, in the first few hours of the light cycle. These mice also received firefly luciferase (p5Luc3) and renilla lucife rase for normalization (phRL-CMV). The first four animals received an siRNA to firefly luciferase and the last two did not.
Appendix: (Continued) As seen in Figure 2, the mice had considerably lower luciferase expression when the siRNA was co-administered. This shows that hydrodynamic tail vein injection of siRNAs can effectively silence target genes in the transfected cells. FIGURE 1. Overexpression of HMGR in Balb/cJ mice by hydrodynamic tail vein injection. Balb/cJ mice were injected with 40 g of a plasmid mix containing either p5Luc3, phRL-CMV (1:500 dilution), and 2 nmol of siRNA against luciferase (control, lanes 1-4) or pRed227 and phRL-CMV (1:500 dilution) (pRed 227, lanes 5 and 6). Microsomes were prepared 18 hours after injection (during the third to fourth hour of the light cycle). Fifty micrograms of microsomal protein was subjected to SDS PAGE and Western blotting. Blots were probed with an antibody to HMGR, and then to -Actin as a loading control. 123
Appendix: (Continued) FIGURE 2. Knockdown of luciferase by hydrodynamic tail vein injection. Livers were harvested from the Balb/cJ mice in figure 1, and assayed for luciferase activity. The first four animals received an siRNA to firefly luciferase (), while the last two did not. Relative luciferase units represents the ratio of firefly to renilla luciferase for each animal. In the next experiment, small (50-75 g) male Srague Dawley rats were treated with an siRNA to HMGR by hydrodynamic injection. Eighteen hours later, the livers were harvested, and microsomes were prepared. HMGR expression was detected by western blotting. As seen in Figure 3, the two animals receiving the siRNA had noticeably higher HMGR expression. This result was unexpected, and may be due to a short term inhibition of HMGR expression causing a compensatory increase in protein levels. 124
Appendix: (Continued) FIGURE 3. Compensatory increase in HMGR protein following siRNA injection. Rats were injected with 40 ug of a plasmid mix containing p5Luc3, phRL-CMV (1:500) (lanes 1 and 2) or these plasmids with 2 nmol of an siRNA targeting the 5 end of the HMGR message (lanes 3 and 4). Microsomes were prepared 18 hours after injection (during the third to fourth hour of the dark cycle). Fifty micrograms of microsomal protein was subjected to SDS PAGE and Western blotting. Blots were probed with an antibody to HMGR, and then to -Actin as a loading control. Since HMGR was overexpressed after one day in the Balb/c mice (Figure 1), we decided to see what influence this would have on serum cholesterol levels. We also wanted to determine what effect overexpression would have on mice fed a cholesterol rich diet. In previous experiments, feeding FVB and C57BL/6 mice a diet containing 1% cholesterol for three days did not significantly change serum cholesterol levels. For this reason, male FVB mice were fed 2% cholesterol and 0.5% cholic acid for three days following the hydrodynamic injection. The cholic acid is a bile acid intended to increase the rate of cholesterol absorption in the intestine, as well as decrease clearance of excess cholesterol via conversion to bile acids. This diet caused a noticeable yellowing of the liver. As seen in Figure 4, mice injected with pRed227 and fed a chow diet (lanes 4-6) overexpressed 125
Appendix: (Continued) HMGR relative to the uninjected controls (lanes 10 and 11). Interestingly, even when overexpressed, dietary cholesterol was still able to decrease HMGR protein levels (lanes 1-3 vs. 7-9). Since pRed227 does not contain a sterol-regulated promoter, this must be due to posttranscriptional effects. Serum cholesterol levels were increased in both the overexpressed and control animals when fed the cholesterol rich diet (Figure 5). This difference reached statistical significance with p<0.05 only in the overexpressed mice. Curiously, the animals receiving pRed227 appeared to have slightly lower serum cholesterol levels than controls when on a chow diet. This difference did not reach significance, p > 0.05. It was also found that LDL receptor protein levels were unaffected by the HMGR overexpression or feeding of dietary cholesterol (Figure 6). This agrees with previous findings in rats. 126 FIGURE 4. Overexpression of HMGR in FVB mice by hydrodynamic tail vein injection after three days. FVB mice were injected with 30 ug of plasmid mix containing either pRed227 and phRL-CMV (1:500 dilution) (lanes 1-6), or 30 ug of p5Luc3, and phRL-CMV (1:500 dilution) (lane 7). The remaining control mice were not injected (lanes 8-11). Mice were then fed 2% cholesterol and 0.5% cholic acid (lanes 1-3, 7-9) or chow diet (lanes 4-6, 10,11) for 3 days. The livers were harvested at the third to fourth hour of the dark cycle, and microsomes were prepared. Sixty micrograms of microsomal protein was subjected to SDS PAGE and Western blotting using an antibody to HMGR.
Appendix: (Continued) 0 100200300400500 pRed-227controlSerumCholesterolmg / dl* 2% CH0.5% C.A.:++-FIGURE 5. The effects of HMGR overexpression on serum cholesterol levels in FVB mice after three days. Serum was obtained from the mice in the experiment shown in Figure 4. Serum cholesterol levels were determined using the cholesterol oxidase assay, by comparison to a standard curve. The gray bars represent groups of animals fed 2% cholesterol and 0.5% cholic acid. The black bars represent the animals fed a chow diet. Serum cholesterol is shown as the average +/standard error of the mean for 3 animals in each group. The * indicates a statistically significant difference with p < 0.05. 127
Appendix: (Continued) FIGURE 6. LDL receptor protein levels are unaffected by HMGR overexpression, or cholesterol and cholic acid feeding in FVB mice. Microsomes (100 ug of protein) from the FVB mice in the previous experiment were subjected to Western Blotting using antibodies to LDL Receptor (LDLR), or HMG-CoA reductase (HMGR). Mice were injected with a plasmid mix containing 30 ug of pRed227 and phRL-CMV (1:500 dilution) (lanes 1-6), or 30 ug of p5Luc3, and phRL-CMV (1:500 dilution) (lane 7). The remaining control mice were not injected (lanes 8-12). Mice were then fed 2% cholesterol and 0.5% cholic acid (lanes 1-3, 7-9) or chow diet (lanes 4-6, 10-12) for 3 days. Since bile acid flux may be important for the buffering capacity of HMGR, we decided to remove cholic acid from the diet. C57BL/6 mice were fed diets containing 2% cholesterol for 5 days following hydrodynamic tail vein injection. At this point, the livers had a pale yellow color due to the accumulation of cholesterol in the tissue. Microsomes were prepared, and western blotting was performed to detect HMGR protein levels (Figure 7). After 5 days, it appears that expression of pRed227 wanes considerably (lanes 4-6 vs. 10, 11). As seen 128
Appendix: (Continued) previously, cholesterol feeding was able to completely suppress HMGR protein levels in the animals receiving pRed227 (lanes 1-3 vs. 7-9). The animals injected with pRed227 had a significant increase in serum cholesterol levels when fed the 2% cholesterol chow (Figure 8). This was not seen in the control animals. For unknown reasons, the control animals had higher serum cholesterol levels on a chow diet than the overexpressed mice. FIGURE 7. Overexpression of HMGR in c57BL/6 mice five days after injection. C57BL/6 mice were injected with a plasmid mix containing 20 ug of pRed227 and phRL-CMV (1:100 dilution) (lanes 1-6), or not injected and left as controls (lanes 7-11). Animals were fed 2% cholesterol (lanes 1-3, 7-9) or a chow diet (lanes 4-6, 10 and 11) for five days. The livers were harvested at between the third and fourth hour of the dark cycle, and microsomes were prepared. Sixty micrograms microsomal protein was subjected to SDS PAGE and Western blotting using an antibody to HMGR. 129
Appendix: (Continued) 050100150200250300 pRed227controlSerumCholesterolmg / dl*** 2% CH:++-FIGURE 8. The effects of HMGR overexpression on serum cholesterol levels in C57BL/6 mice after five days. Serum was obtained from the mice used in the experiment shown in Figure 7. Serum cholesterol levels were determined using the cholesterol oxidase assay by comparison to a standard curve. The gray bars represent groups of animals fed 2% cholesterol. The black bars represent the animals fed a chow diet. Serum cholesterol is shown as the average +/standard error of the mean for 3 animals in each group. The * indicates a statistically significant difference with p < 0.05, ** indicates p< 0.01. A similar experiment was also performed using Balb/c mice, since these gave excellent overexpression after just one day. These Balb/c mice were fed 2 % cholesterol and 0.5% cholic acid for three days following hydrodynamic injection. As expected there was a considerable yellowing of the livers in these animals. Western blotting was performed to detect HMGR protein levels (Figure 9). Only one of the mice still had detectable overexpression of HMGR after 3 days (lane 6). As seen before, the cholesterol rich diet was able to completely 130
Appendix: (Continued) suppress HMGR protein levels in the overexpressed animals, to levels equal to or lower than the controls (lanes 1-3 vs. 7-9). Both groups of animals had a modest increase in serum cholesterol levels when fed the cholesterol rich diet (Figure 10) although these differences did not reach statistical significance. The animal represented in lane 6 of Figure 9, had final serum cholesterol levels approximately the same as those in lanes 4 and 5. 131 FIGURE 9. Overexpression of HMGR in Balb/c mice three days after injection. Balb/c mice were injected with a plasmid mix containing 50 g of pRed227 (lanes 1-6), 50 g of p5Luc3 (lane 10), or not injected and left as controls (lanes 7-9,11 and12). Animals were fed 2% cholesterol and 0.5% cholic acid (lanes 1-3, 7-9) or a chow diet (lanes 4-6, 10 and 11) for five days. The livers were harvested at between the third and fourth hour of the dark cycle, and microsomes were prepared. Forty micrograms of microsomal protein was subjected to SDS PAGE and Western blotting using an antibody to HMGR. Dark and light exposures are shown.
Appendix: (Continued) FIGURE 10. The effects of HMGR overexpression on serum cholesterol levels in Balb/c mice after three days. Serum was obtained from the Balb/c mice used in the experiment in figure 9. Serum cholesterol levels were determined using the cholesterol oxidase assay by comparison to a standard curve. The gray bars represent groups of animals fed 2% cholesterol and 0.5% cholic acid. The black bars represent the animals fed a chow diet. Serum cholesterol is shown as the average +/standard error of the mean for 3 animals in each group. 132
133 Appendix: (Continued) DISCUSSION This study sought to determine whether HMGR expression is a cause or consequence of resistance to dietary c holesterol. We hypothesized that high HMGR expression lends resistance to dietary cholesterol. This could happen through a number of possible mechanisms. Three likely scenarios come to mind: 1) High basal HMGR expression results in increased bile acid production, in turn resulting in a net increase in cholesterol excretion as bile acids. 2) High basal HMGR expression may affect ABCG5/8 expression. ABCG5 and ABCG8 are cholesterol transporters in the liver and intestine. Higher expression of these proteins would help clear free cholesterol from the body. 3) High HMGR expression could lend a cholesterol bufferi ng capacity to the liver. Animals with high levels of HMGR would have higher levels of cholesterol synthesis and depend less on dietary cholesterol. When c hallenged with a cholesterol-rich diet, these animals can more effectively downregulate HMGR expression and better manage the incoming cholesterol. It is perhaps equally likely that high HMGR expression and the corresponding resistance to dietary choles terol is a result of the activity or expression of some other fact or. Animals with low rates of cholesterol absorption
134 Appendix: (Continued) may inherently express more HMGR than animals with greater rates of absorption. Likewise, animals with greater rates of cholesterol efflux may also have higher basal HMGR expression. In both of these situations, a clear correlation would still exist between hi gh HMGR expression and resistance to dietary cholesterol insult. In this work we were able to overexpress HMGR in the liver by hydrodynamic tail vein injection. The expression was optimal at shorter time periods (18 hrs), and fell considerably by three days. Five days post injection, HMGR levels generally returned to normal. Unfortunately, it was very difficult to realize a significant increase in serum chol esterol levels with short term feeding. The greatest increase was achieved when FVB mice were fed a diet containing 2% cholesterol and 0.5% cholic acid for 3 days (Figure 5). In this experiment HMGR protein levels were still overex pressed about 3-fold at time of death (Figure 4). In terms of serum cholesterol levels, there was no benefit to overexpressing HMGR in the FVB mice f ed a cholesterol rich diet. All other things being equal, this suggests that high HMGR expression per se is not sufficient to lend resistance to dietary c holesterol insult. A similar result was achieved with the Balb/c mice (Figures 9 and 10) although the final difference in HMGR expression was not as great. This suggests that high HMGR expression is not causative, but rather is dependent on some other factor which lends resistance to dietary cholesterol.
135 Appendix: (Continued) These factors may be ABCG5 or ABCG8, which have been identified as a cause of strain specific differences in response to dietary cholesterol (9). There may be cross talk between the cholestero l absorption pathway and cholesterol synthesis in the liver. This might occur th rough cholesterol itself or a metabolite in the cholesterol biosynthetic pathway that can traverse the bloodstream. Recently FGF15 has been identified as enterohepatic signal that connects bile acid absorption in the intestine to bile acid production in the liver (10). This pathway, or one like it, could modulate HMGR expression in response to changes in intestinal cholesterol absorpt ion or bile acid efflux. It is also possible that our system is not ideally suited for testing the role of HMGR in the maintenance of serum choles terol levels. A system that would allow more prolonged or tunable HMGR expression such as adenoviral infection, may provide more definitive answers to thes e questions. Likewise, if possible, liver specific knockdown of HMGR by RNA interference could yield a wealth of information about the functional relevanc e of HMGR to whole body cholesterol balance.
136 Appendix: (Continued) REFERENCES 1. Ness, G. C., and Gertz, K. R. (2 004) Exp Biol Med (Maywood) 229, 412416 2. Ness, G. C., and Gertz, K. R. (2 004) Exp Biol Med (Maywood) 229, 407411 3. Gylling, H., Tuominen, J. A., Koivis to, V. A., and Mietti nen, T. A. (2004) Diabetes 53, 2217-2222 4. Miettinen, T. A., Gylling, H., Tuom inen, J., Simonen, P., and Koivisto, V. (2004) Diabetes Care 27, 53-58 5. Ness, G. C., and Chambers, C. M. (2000) Proc Soc Exp Biol Med 224, 819 6. Elbashir, S. M., Harborth, J., Lendeckel, W., Yalcin, A., Weber, K., and Tuschl, T. (2001) Nature 411, 494-498 7. Ness, G. C., and Zhao, Z. ( 1994) Arch Biochem Biophys 315, 199-202 8. Chin, D. J., Gil, G., Russell, D. W., Liscum, L., Luskey, K. L., Basu, S. K., Okayama, H., Berg, P., Goldstein, J. L., and Brown, M. S. (1984) Nature 308, 613-617
137 Appendix: (Continued) 9. Wittenburg, H., Lyons, M. A., Li, R ., Kurtz, U., Wang, X., Mossner, J., Churchill, G. A., Carey, M. C., an d Paigen, B. (2006) J Lipid Res 47, 17801790 10. Inagaki, T., Choi, M., Moschetta, A ., Peng, L., Cummins, C. L., McDonald, J. G., Luo, G., Jones, S. A., Goodwin, B., Richardson, J. A., Gerard, R. D., Repa, J. J., Mangelsdorf, D. J., and Kl iewer, S. A. ( 2005) Cell Metab 2, 217-225
ABOUT THE AUTHOR William (Bill) Raymond Lagor was born in Sanford, Florida on May 25, 1978. He grew up in Clearwate r and Tampa Florida. He attended Jesuit High School in Tampa, and graduated magna cum laude in May of 1996. He then moved to Texas and studied at the University of Dallas. During this time he spent a semester in Rome studying history, philosophy, art and literature. He graduated cum laude with a Bachelor of Science degree in Biochemistry in May of 2000. In 2002 he entered the graduate program in Medical Scienc es at the University of South Florida College of Medicine. Bill earned his doctoral degree in October of 2006. He is married to Jamie Goh, a Chinese, Ma laysian citizen whom he met in 2000. Bill plans to pursue a career in re search as an independ ent investigator.