THE REGULATION OF TXNIP GENE EXPRESSION
FAXING YU
(B.Sc. (Hons.), NUS)
A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF BIOCHEMISTRY NATIONAL UNIVERSITY OF SINGAPORE 2009 Acknowledgments
I first thank my supervisor, Dr. Yan Luo, for his valuable comments on how to design and perform experiments, for his generous giving of freedom to explore new fields, for his enthusiastic encouragement when I encountered difficulties, and for his critical reading and suggestions on my manuscripts and thesis. It is impossible to have this thesis presented here without Dr. Luo’s patient guidance throughout my graduate study.
I wish to thank Drs. Edwin Cheung and Xinmin Cao, my thesis advisory committee members, for sharing their knowledge and wisdom. I also wish to thank
Drs. Thilo Hagen and Ruping Dai for their constructive comments and friendly discussions on my projects.
I would also like to thank past and present members of Luo lab; all of us have constructed a harmonious and productive working environment. Special thanks to
Shuangru Goh, she has experimentally contributed to a portion of results described in this thesis, and Drs. Hongpeng He, Liling Zheng and Mingji Jin, who have given me wonderful suggestions on thesis writing.
Finally, I thank my loving wife, my parents, my siblings, my grandpa, and other family members for their understanding, supports and sacrifices in these years.
I Table of Contents
page
Acknowledgments I
Table of Contents II
Summary IX
List of Tables XI
List of Figures XII
Abbreviations XVII
Chapter 1 Introduction 1
1.1 A General Overview of Eukaryotic Gene Transcription 1
1.1.1 RNA Polymerases 1
1.1.2 Core Promoters and General Transcription Factors (GTF) 2
1.1.3 Sequence‐specific Transcription Factors 3
1.1.4 Cofactors 4
1.1.5 Other Regulatory Mechanisms 6
1.2 Extracellular Signals Regulate Gene Transcription 7
1.2.1 Internal Sensors Regulate Transcription Factors 7
1.2.2 Cell Permeable Ligands Regulate Transcription Factors 9
1.2.3 Plasma Membrane Receptors Regulate Nuclear Transcription 9 Factors
1.2.4 Plasma Membrane Receptors Regulate Cytoplasmic Transcription 10 Factors
II 1.3 Thioredoxin Interacting Protein (Txnip) 10
1.4 Txnip Functions 14
1.4.1 Txnip and Redox State 14
1.4.2 Txnip and Cell Proliferation and Cell Death 15
1.4.3 Txnip and Cell Differentiation 18
1.4.4 Txnip and Cellular Metabolism 18
1.5 Txnip Expression in Response to Different Signals 20
1.6 Objectives 23
Chapter 2 Materials and Methods 25
2.1 Chemicals and Buffers 25
2.2 Plasmid Constructs 25
2.3 Purification of Bacterially Expressed Recombinant Proteins 26
2.4 Mammalian Cell Culture 27
2.5 Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis 27 (SDS‐PAGE)
2.6 Immuno‐Blot 27
2.7 RNA Extraction, RT (Reverse Transcription)‐PCR and Real‐Time PCR 28
2.8 Genomic DNA Extraction 29
2.9 Small Interfering RNAs (siRNAs) and RNA Interference Assays 29
2.10 Transfection 30
2.11 Promoter Activity (Reporter) Assays 30
2.12 Immunocytometry 30
2.13 Fluorescence‐activated Cell Sorting (FACS) 31
III
2.14 Electrophoresis Mobility Shift Assays (EMSA) 31
2.15 Chromatin Immunoprecipitation (ChIP) Assays 34
2.16 Thioredoxin Activity Assays 35
2.17 Glucose Transport Assays 35
2.18 Statistical Analyses 36
Chapter 3 Identification of Molecules Modulating Txnip Expression 37
3.1 Preface 37
3.2 Results 38
3.2.1 Adenosine‐containing Molecules Induce Txnip Expression 38
3.2.1.1 NAD(H) and ATP Stimulate Txnip Expression 38
3.2.1.2 Molecules Containing Adenosine Group Induce Txnip 40 Expression
3.2.1.3 Adenosine is Necessary and Sufficient for Inducing Txnip 40 Expression
3.2.1.4 Adenosine‐containing Molecules Induce Txnip Expression 43 in a Dose‐dependent Manner
3.2.1.5 Long Term Effect of Adenosine‐containing Molecules on 44 Txnip Expression
3.2.1.6 Adenosine‐containing Molecules Induce Txnip Expression 45 at the Transcriptional Level
3.2.1.7 Adenosine‐containing Molecules Induce Txnip Expression 47 Is Mediated by an Earlier Defined ChoRE
3.2.1.8 The MLX/MondoA Complex Mediates the Induction of 49 Txnip Expression by Adenosine‐containing Molecules
3.2.1.9 Adenosine‐containing Molecules Facilitate MondoA 52 Nuclear Translocation
IV
3.2.1.10 Glucose Is Required for the Induction of Txnip Expression 54 by Adenosine‐containing Molecules
3.2.1.11 Glucose Induced Txnip Expression Is Amplified by 56 Adenosine‐containing Molecules
3.2.1.12 Potential Plasma Membrane Target(s) of Adenosine‐ 57 containing Molecules
3.2.1.12.1 Purinergic Receptors Are Not Required for the 59 Induction of Txnip Expression
3.2.1.12.2 Adenosine‐containing Molecules May Target 59 Adenosine Transporters
3.2.1.13 Signaling Pathway(s) Evoked by Adenosine‐containing 62 Molecules for Regulating Txnip Expression
3.2.1.13.1 The Induction of Txnip Expression Requires 62 Intracellular Ca2+
3.2.1.13.2 The Induction of Txnip Expression Does Not 64 Require cAMP
3.2.1.13.3 The Involvement of MAPK in the Induction of 64 Txnip Expression
3.2.1.13.4 Non‐involvement of AMPK in the Induction of 66 Txnip Expression
3.2.1.14 Adenosine‐containing Molecules Repress Thioredoxin 67 Activity and Glucose Transport
3.2.1.15 Adenosine‐containing Molecules Affect Cell Cycle 69 Progression
3.2.2 Effects of Glucose Analogs on Txnip Expression 70
3.2.2.1 Effects of Selected Monosaccharides and Disaccharides on 70 Txnip Expression
3.2.2.2 Effect of PD169316 on Glucose, 2DG or Maltose/NAD+ 75 Induced Txnip Expression
V
3.2.2.3 A Ca2+ Chelator Abolishes the Stimulatory Effect of Glucose 75 on Txnip Expression
3.2.3 Inhibitors of Oxidation Phosphorylation Repressed Txnip 76 Expression
3.2.3.1 Nitric Oxide (NO) and Sodium Azide (NaN3) Repress Txnip 76 Expression
3.2.3.2 Inhibition of Oxidative Phosphorylation Represses Txnip 80 Expression
3.2.3.3 A Ca2+ Chelator Rescues Txnip Expression in the Presence of 82 Oxidative Phosphorylation Inhibitors
3.3 Discussion 83
3.3.1 Adenosine‐containing Molecules May Remain Extracellular to 83 Induce Txnip Expression
3.3.2 Potential Membrane Targets for Adenosine‐containing Molecules 84
3.3.3 Signaling Pathway(s) Involved in the Induction of Txnip 85 Expression by Adenosine‐containing Molecules
3.3.4 The MondoA/MLX Complex Mediates Txnip Expression 87
3.3.5 Physiological Significance 89
3.3.6 Two Signaling Pathways Evoked by Glucose for Inducing Txnip 91 Expression
3.3.7 Txnip Expression Is a Sensor of Oxidative Phosphorylation Status 93
3.4 Conclusion and Perspectives 94
Chapter 4 Regulatory Mechanisms Underlying the Induction of Txnip 96 Expression by Glucose or Adenosine‐containing Molecules
4.1 Preface 96
4.2 Results 98
VI 4.2.1 Regulatory Mechanisms at the Promoter Level Governing Glucose or Adenosine‐containing Molecules Induced Txnip 98 Expression
4.2.1.1 Txnip Promoter Regions Critical for Expression Induction 98 by NAD+ or Glucose
4.2.1.2 Tandem ChoREs on Txnip Promoters 102
4.2.1.3 The MondoA/MLX Complex Binds to Both ChoREs in vitro 105
4.2.1.4 Both ChoREs are Required for Optimal Txnip Promoter 108 Activity
4.2.1.5 ChoREs Are Not Sufficient for the Induction of Txnip 111 Expression
4.2.1.6 Tandem NF‐Y Binding Sites Are Required for the Induction 111 of Txnip Expression
4.2.1.7 NF‐Y Mediated Induction of Txnip Expression by SAHA 114 Requires MondoA/MLX
4.2.1.8 Txnip Promoter Recruits MondoA/MLX Complex in an 116 NF‐Y Dependent Manner
4.3 The Role of USFs in Txnip Expression 121
4.3.1 Expression and Purification of His‐tagged USF1 121
4.3.2 USF1 Interacts with ChoRE Sites in vitro 122
4.3.3 Down‐regulation of USFs Reduces Txnip Expression 125
4.3.4 Over‐expression of USFs Induces Txnip Promoter Activity 127
4.3.5 USF Is Not Involved in the Txnip Induction by 129 Adenosine‐containing Molecules
4.3.6 USF Is Not Involved in the Txnip Induction by Glucose 132
4.4 Discussion 133
4.4.1 Tandem ChoREs on the Txnip Gene Promoter 133
VII
4.4.2 Full Induction of Txnip Expression Requires Both ChoREs and 134 CCAAT Boxes
4.4.3 NF‐Y and MondoA/MLX Cooperate to Stimulate Txnip 136 Expression
4.4.4 The Role of USFs on Txnip Expression 138
4.5 Conclusion and Perspectives 140
References 143
Appendices 162
Appendix I, Buffers/Gels Used in This Study 162
Appendix II, RNA Integrity 165
Appendix III, Primer Specificity 166
Appendix IV, Primers Used in This Study 167
Appendix V, Paper 1 (Abstract) 172
Appendix VI, Paper 2 (Abstract) 173
VIII Summary
Thioredoxin interacting protein (Txnip) is a multifunctional protein involved in regulation of cell cycle events and cellular metabolism. The expression of Txnip is induced by glucose, and this induction is mediated by a carbohydrate response element (ChoRE) on Txnip promoter and its associated transcription factors,
MondoA and Max‐like protein X (MLX). In this study, I have discovered that the transcription of the Txnip gene is induced by an array of adenosine‐containing
molecules, of which an intact adenosine moiety is necessary and sufficient. The
induction of Txnip expression by adenosine‐containing molecules is glucose
dependent, and MondoA and MLX have been shown to convey stimulatory signals from extracellular molecules to the Txnip promoter. Therefore, the regulatory role of adenosine‐containing molecules is exerted via amplifying glucose signaling, and
this suggests that these molecules may modulate the kinetics of glucose homeostasis.
I have also studied the underlying regulatory mechanisms of glucose and
adenosine‐containing molecules on Txnip expression. An additional ChoRE on the promoter of Txnip gene has been identified, and this ChoRE is able to recruit
MondoA and MLX in a similar fashion as the previously identified ChoRE in vitro and in vivo. Both ChoREs function cooperatively to mediate optimal Txnip expression under glucose or adenosine‐containing molecules treatment. However, these two ChoREs are not sufficient to mediate the induction of Txnip expression by glucose or adenosine‐containing molecules, and two CCAAT boxes, both can recruit
IX nuclear factor Y (NF‐Y) to the Txnip promoter, are also required for the induction. I also found that the function of ChoREs and associated factors is contingent on tandem CCAAT boxes, in that the occupancy of the Txnip promoter by the CCAAT box‐associated NF‐Y is a prerequisite for efficacious recruitment of MondoA/MLX to
ChoREs under glucose stimulation. Such a strategy suggests a synergy between
NF‐Y and MondoA/MLX in enhancing Txnip expression presumably through inducing dynamic chromatin changes in response to diverse physiological inducers.
X List of Tables
Page
Table 1. Oligonucleotides used for EMSA. 32
Table 2. Primers used in ChIP assay. 33
Table 3. The effects of different molecules on Txnip mRNA levels. 41
XI List of Figures
Page
Figure 1. A simplified version of eukaryotic transcription machinery and 5 transcription regulatory mechanisms.
Figure 2. Possible mechanisms for the regulation of gene transcription by 8 extracellular molecules and internal environment changes.
Figure 3. A phylogenetic tree of Txnip orthologs from different organisms. 11
Figure 4. Gene structures of Txnip, arrestin β2 and ARRDC genes. 13
Figure 5. Protein primary structure of Txnip, arrestin β2 and ARRDCs. 14
Figure 6. Signaling pathways regulating Txnip expression and their 21 targeting cis‐regulatory elements on Txnip promoter.
Figure 7. NAD(H) or ATP induced Txnip mRNA level. 38
Figure 8. NAD+ induced Txnip protein level. 39
Figure 9. NAD(H) or ATP induced Txnip expression in diverse cell lines. 40
Figure 10. NAD+ and ATP share a common adenosine group. 41
Figure 11. Structure of ATP, dATP, Bz‐ATP or Tnp‐ATP. 42
Figure 12. Titrations of adenosine‐containing molecules on Txnip expression. 43
Figure 13. Time course of adenosine‐containing molecules on Txnip 44 expression.
Figure 14. Adenosine‐containing molecules induce Txnip promoter activity. 45
Figure 15. Adenosine‐containing molecules did not induce Txnip expression 46 in the presence of actinomycin D.
Figure 16. NAD+ or ATP did not induce activity of ChoRE‐mutated Txnip 48 promoters.
XII Figure 17. NAD+ or ATP could not induce Txnip promoter activity in the 48 presence of dominant negative MLX.
Figure 18. Effects of ectopic MLX and/or Mondo on Txnip promoter activity. 50
Figure 19. Effects of siRNAs against MLX or MondoA on Txnip expression. 51
Figure 20. Localization of HA‐MondoAin L6 cells under control or 2DG 53 treatment.
Figure 21. MondoA nuclear translocation was facilitated by adenosine‐ 53 containing molecules.
Figure 22. Glucose‐dependent induction of Txnip expression by NAD+ or ATP. 55
Figure 23. NAD+ induced Txnip expression in a wide titration of glucose. 56
Figure 24. The induction of Txnip expression was not abolished by ARL 57 67156.
Figure 25. Inhibitors for purinergic receptors did not inhibit the induction of 59 Txnip expression by NAD+ or ATP.
Figure 26. Structures of Adenosine, NBTI, Dipyridamole or Dilazep. 60
Figure 27. The effect of NAD+ or ATP on Txnip expression was blocked by 61 inhibitors of adenosine transporters.
Figure 28. Effect of Ca2+ chelators on Txnip expression. 62
Figure 29. cAMP signaling pathway did not mediate the induction of Txnip 63 expression by NAD+.
Figure 30. The effect of MAPK inhibitors on Txnip expression and its response 65 to NAD+ or ATP in U2OS cells.
Figure 31. The effect of MAPK inhibitors on Txnip expression and its response 66 to NAD+ or ATP in HeLa cells.
Figure 32. AMPK was not involved in the induction of Txnip expression by 67 adenosine‐containing molecules.
Figure 33. Adenosine‐containing Molecules Repressed Thioredoxin Activity. 68
XIII Figure 34. Adenosine‐containing Molecules Inhibited Glucose Transport. 68
Figure 35. NAD(H) treatment repressed cell cycle progression. 69
Figure 36. NAD(H) treatments elevated p21cip1 expression level. 70
Figure 37. Structures of glucose analogs and their fate in cellular metabolism. 71
Figure 38. The effect of glucose and glucose homologs on Txnip expression in 72 the absence or presence of NAD+.
Figure 39. The effect of PD169316 on Txnip induction by different 74 carbohydrates and/or NAD+.
Figure 40. Glucose induced Txnip expression was repressed by BAPTA‐AM. 76
Figure 41. Effects of NO donors and GC inhibitor on Txnip expression. 77
Figure 42. The effect of NO on Txnip expression was not mediated by GC. 78
Figure 43. The effect of NaN3 on Txnip mRNA Levels. 79
Figure 44. A simplified representation of oxidative phosphorylation. 80
Figure 45. Effects of oxidative phosphorylation inhibitors on Txnip or H2B 81 expression.
Figure 46. Oxidative phosphorylation inhibitors did not repress Txnip 82 expression in the presence BAPTA‐AM.
Figure 47. A negative feed‐back loop for glucose uptake. 90
Figure 48. A schematic representation of truncated Txnip promoters. 99
Figure 49. Responses of Txnip promoters to NAD+. 99
Figure 50. Responses of Txnip promoters to NAD+ or Glucose. 100
Figure 51. The minimal Txnip promoter sequence required for mediating the 101 stimulatory effect of NAD+.
Figure 52. cis‐regulatory elements on Txnip promoter. 102
Figure 53. Alignment of Txnip promoters from different species. 103
XIV
Figure 54. Phylogenetic tree of Txnip promoters. 104
Figure 55. Sequence alignment of fish and frog Txnip promoters with the 105 human Txnip promoter.
Figure 56. EMSAs using two Txnip promoter segments containing ChoRE‐a 106 or ChoRE‐b.
Figure 57. Responses of wild‐type (WT) or ChoRE mutant Txnip promoters to 109 NAD+ or glucose.
Figure 58. Effect of siRNAs against MLX or MondoA on Txnip promoter 110 activities.
Figure 59. The response of hybrid Txnip promoters to NAD+ or glucose. 112
Figure 60. Responses of wild‐type (WT) or CCAAT box mutant Txnip 113 promoters to NAD+ or glucose.
Figure 61. The effect of SAHA. 115
Figure 62. Stable cell lines for ChIP assays. 117
Figure 63. Recruitment of MLX and NF‐Y to respective targets (ChoREs and 118 CCAAT boxes) assessed by ChIP assays.
Figure 64. MLX is not recruited onto CCAAT boxes‐mutated Txnip promoter. 119
Figure 65. The interaction between MLX and ChoRE is glucose dependent. 121
Figure 66. Bacterial expression of USF1. 121
Figure 67. His‐USF1 interacts with ChoRE‐a or ChoRE‐b. 122
Figure 68. ChoRE‐containing probes interact with a protein(s) in HeLa nuclear 124 extract (NE).
Figure 69. Bands shift by USF antibodies. 125
Figure 70. The effects of USF1‐ or USF2‐specific siRNAs on Txnip expression. 126
Figure 71. The induction of Txnip promoter activity by over‐expression of 127 USFs.
XV
Figure 72. The response of Txnip promoter to NAD+ or ATP under USF 128 over‐expression.
Figure 73. Effects of USF over‐expression on the activities of truncated or 130 mutant Txnip promoters.
Figure 74. A‐USF did not repress the induction of Txnip promoter activity by 131 NAD+ or ATP.
Figure 75. A‐USF did not repress the induction of Txnip promoter activity by 132 glucose.
Figure 76. A model for the transcriptional regulation of the Txnip gene 137 promoter by NF‐Y, MondoA/MLX and other (co)factors.
XVI Abbreviations
2DG 2‐deoxy‐glucose
3OMG 3‐O‐Methylglucose
5‐aza‐CdR 5‐AZA‐2ʹ‐deoxycytidine
AC Adenylyl cyclase
AICAR 5‐aminoimidazole‐4‐carboxamide ribonucleoside
AP‐1 Activator protein 1
ARRDC Arrestin domain containing
ASK‐1 Apoptosis signal‐regulating kinase 1
BrdU Bromodeoxyuridine cAMP Cyclic adenosine monophosphate
CDK4 Cyclin‐dependent kinase 4 cDNA Complementary DNA
ChIP Chromatin immunoprecipitation
ChoRE Carbohydrate response element
ERK Extracellular signal‐regulated kinases
EMSA Electrophoresis mobility shift assay
FACS Fluorescence‐activated cell sorting
FCHL Familial combined hyperlipidemia
FOXO1 Forkhead box O1
G6P Glucose‐6‐phosphate
XVII GC Guanylate cyclase
Glut Glucose transporter
GPCR G‐protein coupled receptors
GR Glucocorticoid receptor
GRE Glucocorticoid response element
GTF General transcription factors
HA Hemagglutinin
HAT Histone acetyltransferases
HDAC Histone deacetylases
HDMT Histone demethylases
HIF‐1 Hypoxia‐inducing factor 1
HMT Histone methyltransferases
HRE HIF‐responsive element
HSE Heat shock response elements
HSF‐1 Heat shock factor 1
HSP Heat shock proteins
HXT Hexose transporter
IP Immunoprecipitation
IPTG Isopropyl‐beta‐D‐thiogalactopyranoside
Jab‐1 Jun activation domain‐binding protein 1
JAK Janus kinase
JNK c‐Jun N‐terminal kinases
XVIII MAPK Mitogen‐activated protein kinase
MLX Max‐like protein X mRNA Messenger RNA
mTOR Mammalian target of rapamycin
NE Nuclear extract
NF‐κB Kappa‐light‐chain‐enhancer of activated B cells
NF‐Y Nuclear factor Y
NO Nitric oxide
Oct‐1 Octamer‐1
PIC Pre‐initiation complex
PKA Protein kinase A
PKB/AKT Protein kinase B
PKG Protein kinase G
Pol II RNA polymerase II
PP2A Protein phosphatase 2A
PPAR Peroxisome proliferator‐activated receptor
PPRE Peroxisome proliferator hormone response element
PTEN Phosphatase and tensin homolog
ROS Reactive oxygen species
RTK Receptor tyrosine kinases
SAHA Suberoylanilide hydroxamic acid
SDS‐PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis
XIX SGLT Na+/glucose symporter
siRNA Small interference RNA
STAT Signal transducers and activator of transcription
TBP‐2 Thioredoxin‐binding protein 2
TGF‐β Transforming growth factor beta
Txnip Thioredoxin Interacting Protein
USF Upstream stimulatory factor
VDRE Vitamin D response element
VDUP1 Vitamin D3 up‐regulated protein 1
XX
Chapter 1
Introduction Chapter 1
1.1 A General Overview of Eukaryotic Gene Transcription
Genetic information is stored in the form of DNA. To exert their functions, genes are required to be expressed to RNA or proteins. Transcription, or RNA synthesis, is the process of transcribing sequence information from DNA to RNA in a complementary manner. RNA in turn can be used as template for protein synthesis, this process is called translation.
1.1.1 RNA Polymerases
Transcription is a biochemical reaction catalyzed by RNA polymerases. In eukaryotes, three types of RNA polymerases have been identified and respectively dubbed as RNA polymerase I, II and III (Roeder and Rutter, 1969). Among these polymerases, RNA polymerase I is responsible for the synthesis of ribosomal RNAs
(rRNA); RNA polymerase II (Pol II, or RNAP II) mediates the synthesis of messenger
RNAs (mRNA), some small nuclear RNAs (snRNA), small nucleolar RNAs (snoRNA) and microRNAs; and RNA polymerase III is responsible for the synthesis of transfer
RNAs (tRNA), 5S rRNA and some snRNAs (Roeder and Rutter, 1970; Lee et al., 2004).
Novel RNA polymerases have also been discovered recently that are involved in synthesis of some specific RNAs (e.g., synthesis of certain small interfering RNAs
[siRNA] in plants [Herr et al., 2005]).
Pol II is a protein complex containing 12 subunits encoded by different genes; among these subunits, some are Pol II‐specific while others are shared with other
RNA polymerases (Sklar et al., 1975). The synthesis of protein‐coding mRNAs by Pol
II is a complicated process facilitated by many other factors, and is controlled by multiple regulatory pathways to meet the spatial and temporal requirement for the expression of specific genes. The last four decades have witnessed extensive studies
1 Chapter 1 on the structure, function and regulation of Pol II using biochemical, structural and
genetic approaches (reviewed in Roeder, 2003; Levine and Tjian, 2003; and Kornberg,
2007).
1.1.2 Core Promoters and General Transcription Factors (GTF)
For a given mRNA encoding gene, a core promoter is defined as the minimal
DNA sequence required for accurate transcription initiation. Core promoters are
typically positioned around the transcription start site (~35 nucleotides upstream or downstream) and contain signature elements including, to name a few, TATA box,
initiator (Inr) and downstream core promoter element (DPE) (Butler and Kadonaga,
2002). There appears to be no universal core promoter elements; for instance, few
house keeping genes, such as the one encoding histone 1, contain a TATA box in the
promoters (Cooper et al., 2006; Isogai et al., 2007). I am working on a TATA box‐
containing gene, therefore below I will focus on the introduction of TATA‐regulated
promoters/genes.
Although RNA polymerases are enzymes responsible for RNA synthesis,
they alone are unable to initiate accurate transcription (Weil et al., 1979). Six ancillary protein or protein complexes, namely TFIIA, TFIIB, TFIID, TFIIE, TFIIF and TFIIH
(respectively, transcription factors A, B, D, E, F and H for Pol II), have been isolated and characterized from whole cell or nuclear extracts; they (with a total of 32 distinct polypeptides) are collectively called general transcription (initiation) factors (GTFs,
Figure 1; Matsui et al., 1980; Samuels et al., 1982; Reinberg and Roeder, 1987a;
Reinberg and Roeder, 1987b; Flores et al., 1990; Flores et al., 1992). Pol II and GTFs can form a pre‐initiation complex (PIC) important for accurate transcription initiation from a core promoter at a basal level (Van Dyke et al., 1988; Buratowski et al., 1989).
2 Chapter 1
GTFs are widely expressed and required for transcription of majority, if not all, mRNA genes; however, a subset of cellular genes may require a subset of the
GTFs that along with certain specialized (co)factors brings about tissue‐ and/or gene‐ specific transcriptional activation (reviewed in Green, 2000; Berk, 2000).
1.1.3 Sequence‐specific Transcription Factors
Many evolutionary conserved short DNA sequences have been identified on promoter region of most genes, and these short DNA elements (cis‐regulatory elements) are able to mediate transcription regulation by recruiting sequence‐specific transcription factors. The proximal sequences upstream (~250) of the transcription start site usually contain primary regulatory elements, and these regions are called proximal promoters. Less frequently, some elements are located proximally or distally downstream, or at distal regions upstream, of the transcription start site.
These elements are known as distal promoter or enhancer elements. Among these
cis‐regulatory elements, some can recruit positive regulatory factors (transcription
activators); on the other hand, some can recruit negative regulatory factors
(transcription repressors). For convenience, these two types of cis‐regulatory
elements are dubbed, respectively, enhancer and repressor (Figure 1).
Sequence‐specific transcription factors occupy enhancers or repressors via
multiple interactions, such as hydrogen bonds, ionic bonds and hydrophobic
interactions. These transcription factors usually contain structures (with multiple α
helices, loops or β sheets) which can bind to the major groove of DNA. Such
structures include Helix‐Turn‐Helix motifs, Zinc Fingers, Leucine Zippers, Helix‐
Loop‐Helix motifs and possibly others (Alberts et al., 2002). For instances, the
homeodomain of octamer‐1 (Oct‐1) is a Helix‐Turn‐Helix motif, and this motif is
3 Chapter 1 important for the binding of Oct‐1 to an octamer sequence (ATGCAAAT, Klemm et al., 1994). Activator protein 1 (AP‐1) is a dimer formed by c‐jun and c‐fos, and it contains a leucine zipper motif which is critical for the recruitment of AP‐1 to its binding site (TGAGTCA; Glover and Harrison, 1995); Upstream stimulatory factor
(USF; Gregor et al., 1990) and c‐Myc (Blackwood and Eisenman, 1991) contain Helix‐
Loop‐Helix motifs which guide these transcription factors to bind with E‐box sequence (CANNTG). Although less common, the Minor groove of DNA may also
participate in the interaction between transcription factors and their cognate DNA
binding sites (Panne et al., 2007).
1.1.4 Cofactors
Many sequence‐specific transcription factors, while bound with cis‐regulatory elements, have been documented to contact components of the PIC thus, according to the conventional wisdom, able to regulate gene transcription by direct modulation of GTF function (Roy et al., 1991); however, reconstituted transcription reactions that contained gene‐specific transcription factors, GTFs and Pol II failed to recapitulate gene activation at physiological level (reviewed in Luo and Roeder, 1999), prompting a search for cofactors that facilitate functional interplay between transcription factors and GTFs and/or Pol II (Ge et al., 2002; represented by “X” in Figure 1).
Recently, a large repertoire of proteins or protein complexes has been identified to regulate gene transcription together with sequence‐specific transcription factors and GTFs/Pol II. A commonly accepted nomenclature of these proteins is not available currently; for convenience, we call them cofactors here. There are at least 5 classes of cofactors, which include TAFs (TATA‐binding‐protein Associated Factors) in TFIID, some cell/gene specific cofactors (e.g., OCA‐B), USA (Upstream Stimulatory
4 Chapter 1
Activity)‐derived cofactors that contain the Mediator complex, chromatin remodelers, and histone modifiers (reviewed in Luo and Roeder, 1999). The former 3 classes of
cofactors were largely defined by activities on naked DNA templates, and the later 2
are required on in vitro assembled chromatin‐‐thus more natural‐‐DNA templates.
For a prevailing relevance to my study, the Mediator complex, chromatin remodelers and histone modifiers are briefly described below.
Figure 1. A simplified version of eukaryotic transcription machinery and transcription regulatory mechanisms. X: cofactors.
Native eukaryotic DNA is organized into chromatin that contains histones and other nucleoproteins assembled in a higher order structure called nucleosome, which imposes an intrinsic barrier for gene transcription. Prior to PIC assembly and transcription initiation, this barrier must be overcome. A class of ATP‐dependent
chromatin remodelers (e.g., SWI/SNF) can destabilize and restructure nucleosomes
(reviewed in Clapier and Cairns, 2009). Histone proteins can undergo modifications
5 Chapter 1 such as acetylation/deacetylation (by histone acetyltransferases [HAT] and histone
deacetylases [HDAC]), methylation/demethylation (by histone methyltransferases
[HMT] and histone demethylases [HDMT]), phosphorylation and ubiquitination.
These epigenetic marks in turn trigger configuration changes in chromatin templates,
which recruit stimulatory or repressive regulatory factors to modulate transcription
initiation and elongation (reviewed in Kouzarides, 2007; Li et al., 2007a).
Another important cofactor is the Mediator (reviewed in Malik and Roeder,
2005; Kornberg, 2005; Roeder, 2005). Mediator is a protein complex comprising of
~25 subunits; more than half of these subunits are transcription activators or
repressors as previously shown using genetic approaches. These subunits can be
divided into 4 sub‐complexes: a head complex and a middle complex can interact
directly with Pol II, a tail complex can interact with sequence‐specific transcription factors or some adaptor proteins, and a fourth complex called cyclin‐dependent CTD kinase activity (CDK). The Mediator can thus receive signals from sequence‐specific transcription factors and probably other signaling molecules, and transmits these
signals to Pol II or GTFs to regulate transcription, either positively or negatively
(Taatjes et al., 2004).
1.1.5 Other Regulatory Mechanisms
Other than mechanisms mentioned above, the transcription can be regulated at many other levels: by genes being located at a specific sub‐nuclear domain (Matera,
1999), at the transcription elongation level (Conaway, 2000) and at the transcription termination level (Buratowski, 2005). The mRNA expression levels can also be
regulated at the post‐transcriptional (Sharp, 2005; Gu and Lima, 2005; Mangus et al.,
2003) and RNA transport (Martin and Ephrussi, 2009) levels. These mechanisms are
6 Chapter 1 not reviewed in detail.
1.2 Extracellular Signals Regulate Gene Transcription
From unicellular yeast to multi‐cellular human beings, cells can respond to environmental factors, such as nutritional conditions, light, temperature and oxygen levels to adjust gene activities. In addition, cells in multi‐cellular organisms are required to respond to signals from neighboring cells, extracellular cellular matrix, endocrine systems, neurons and infections. In responding to these signals, a cell can modulate its gene expression network to regulate physiological processes such as cell proliferation, cell differentiation, apoptosis and metabolic switches.
In order to regulate gene expression, extracellular signals can activate a number of signaling pathways to modulate the activity and location of sequence‐ specific transcription factors or cofactors by diverse modifications (phosphorylation, sumoylation [Hay, 2005; Gill, 2005], ubiquitination [Conaway et al., 2002], etc.), and the interplay of signal transduction and transcription machineries provides a complex gene regulatory network that precisely dictates gene transcription under different environmental or physiological conditions. Usually, sequence‐specific transcription factors or cofactors can be regulated by specific signaling pathways to
modulate gene transcription (Brivanlou and Darnell, 2002).
1.2.1 Internal Sensors Regulate Transcription Factors
Eukaryotic cells have some internal apparatus, which can sense the change of environments, such as temperature and oxygen levels. In the heat shock response
(Wu, 1995; Prahlad and Morimoto, 2009), high temperature and some other stresses can cause mis‐folding of certain proteins; mis‐folded proteins are recognized by preexisting heat shock proteins (HSP) that normally sequester the heat shock factor 1
7 Chapter 1
(HSF‐1) in the cytoplasm; and the released HSF‐1 in turn forms a trimer, translocates to the nucleus, and binds to the heat shock response elements (HSE) in the heat shock gene promoters to activate their transcription (Figure 2, A). Another example is a heterodimeric protein complex, hypoxia‐inducing factor 1 (HIF‐1, Semenza, 2004;
Bruegge et al., 2007), in which the alpha subunit (HIF‐1α) is normally unstable but is dramatically stabilized when the environmental oxygen level is low, thus promoting the accumulation of dimeric HIF‐1. This in turn induces transcription of several HIF‐ responsive genes, mediated by the HIF‐responsive element (HRE) on their promoters; the induced expression of these genes can promote survival of host cells under hypoxia conditions.
Figure 2. Possible mechanisms for the regulation of gene transcription by extracellular molecules and internal environment changes.
8 Chapter 1
1.2.2 Cell Permeable Ligands Regulate Transcription Factors
Some small molecules (e.g., retinoids and vitamin D) or sex hormones (e.g., androgen and estrogen) can enter cells as they are permeable to plasma membrane; once inside cells, these molecules can bind to nuclear receptors to trigger gene transcription (Perissi and Rosenfeld, 2005). Nuclear receptors are a large family of proteins containing ligand‐binding domain and DNA‐binding domain, many of them can regulate gene transcription in a ligand‐dependent manner. In the absence of ligands, most nuclear receptors reside in nucleus and some (e.g., glucocorticoid receptor, GR) reside in cytosol. When bound with ligands, monomer, homodimer or heterodimer of nuclear receptors can be recruited onto promoters of target genes via different cis‐elements and regulate gene transcription. For GR, when bound with glucocorticoids it can translocate into nucleus and regulate gene transcription (Zhou and Cidlowski, 2005; see Figure 2, B).
1.2.3 Plasma Membrane Receptors Regulate Nuclear Transcription Factors
Many transcription factors (e.g., AP‐1 and cAMP response element‐binding protein [CREB]) are latent even when residing in nucleus; these transcription factors can respond to certain extracellular signals transmitted into nucleus via signaling pathways initiated by ligands‐receptor interaction at the plasma membrane. Once occupied by ligands, cell surface receptors like G‐protein coupled receptors (GPCR) and receptor tyrosine kinases (RTK) can increase cellular concentration of second messengers or initiate a cascade of kinase reactions. Protein kinase A (PKA) can be activated by cyclic adenosine monophosphate (cAMP, a second messenger) and translocate into nucleus to phosphorylate and activate CREB (Mayr and Montminy,
2001). One subunit of AP‐1, c‐jun, can be phosphorylated by c‐Jun N‐terminal
9 Chapter 1 kinases (JNKs), a downstream kinase in the mitogen‐activated protein kinase (MAPK) cascade which can enter nucleus when phosphorylated; the phosphorylation of c‐jun will enhance the transactivation potential of AP‐1(Karin et al., 1997; see Figure 2, C).
1.2.4 Plasma Membrane Receptors Regulate Cytoplasmic Transcription Factors
In contrast to latent nuclear transcription factors, some other transcription factors (e.g., Signal Transducers and Activator of Transcription [STAT] and nuclear
factor kappa‐light‐chain‐enhancer of activated B cells [NF‐κB; Baeuerle and
Baltimore, 1996; Perkins, 2000]) are normally sequestrated in cytoplasm; in the
presence of extracellular stimuli, cell surface receptors initiate signaling pathways to modify these transcription factors, which then enter nucleus to activate transcription of target genes. For example, extracellular cytokines bind to their receptors and
activate intracellular Janus kinase (JAK), JAK in turn phosphorylates a STAT tyrosine
residue, which then promotes the STAT dimerization; the phophorylated dimer is
transported into nucleus and actively induce cytokine response genes (Darnell, 1997;
Stark et al., 1998; see Figure 2, D).
The integration of signaling pathway and gene transcription is briefly
classified into four models (Figure 2); however, other modes of gene regulation do
exist, such as the retrograde signaling (Butow and Avadhani, 2004), which are not
discussed here.
1.3 Thioredoxin Interacting Protein (Txnip)
Txnip was initially identified as Vitamin D3 up‐regulated protein 1 (VDUP1), for its mRNA expression level was induced by 1,25‐dihydroxyvitamin D3
(1,25(OH)2D3) in HL60 cells (Chen and DeLuca, 1994). Later, the expression of Txnip
10 Chapter 1 has been shown to be modulated by many extracellular signals, and the product of
this gene exhibited diverse physiological and pathological functions.
Figure 3. A phylogenetic tree of Txnip orthologs from different organisms. The phylogenetic tree was generated using software in UCSC genome browser.
11 Chapter 1
The Txnip gene can be found in all vertebrates (from fish to human) with available genome sequence information; a phylogenetic tree has been constructed based on the similarity of different Txnip orthologs, and clearly, this tree fit well with the evolutionary tree (fish Æ amphibian Æ lower placental mammals Æ primates,
Figure 3). Currently, human and mice Txnip genes have been cloned (Chen and
DeLuca, 1994; Ludwig et al., 2001), these two genes are identical in both gene
structure and nucleotide sequences of promoter and coding regions.
The human Txnip gene is located on chromosome one (locus: 1q21.1), the gene contains 4167 base pair (bp) of nucleotides in which eight exons are identified
(Figure 4). In the human genome, five genes show homologies to the Txnip gene;
they are arrestin domain containing 1‐5 (ARRDC1‐5, of which ARRDC3 is also
known as TLIMP [Oka et al., 2006a]; and ARRDC4 is also known as DRH1
[Yamamoto et al., 2001]). Among ARRDC genes, four (ARRDC1‐4) contain eight exons, and three (ARRDC2‐4) have shown high homologies to Txnip when compared exon by exon (Figure 4); this suggests that Txnip and (at least) ARRDC2‐4 are generated by duplications of a common ancestor gene (Oka et al., 2006a). Txnip orthologs exist in diverse vertebrates (Figure 3), and its homologues have not been identified in lower organisms. When the human Txnip sequence is used to perform
BLAST against Drosophila melanogaster and Caenorhabditis elegans genes, some arrestin domain containing genes are identified but these genes have exhibited higher level of homologies to the ARRDC genes. A candidate Drosophila Txnip gene identified
(Mandalaywala et al., 2008) is more homologous to ARRDC genes, and hence should be considered as an ARRDC ortholog. Therefore, ARRDC genes might be more ancient than Txnip in evolution.
12 Chapter 1
Figure 4. Gene structures of Txnip, arrestin β2 and ARRDC genes. Drawings on the left indicate the size (in scale) and distribution of exons for each gene. Table on the right contains name, ID and chromosomal positions for each gene. †: ARRDC2, isoform 1 and 2; ‡: arrestin β2, isoform 1. All genes shown are human genes.
The expression products of Txnip and ARRDC genes also shared high level of similarity (with highest similarity to ARRDC2‐4, Figure 5). All these proteins contain one arrestin N‐terminal domain and one arrestin C‐terminal domain, and the overall primary structure of these proteins are similar to arrestin β2, although the gene structure of arrestin β2 is very different (Figures 4 and 5). Currently, studies on
ARRDC1‐5 are very limited, the functions of arrestin domains in these proteins are not clear. Unlike Txnip, ARRDC3 is largely localized at inner layer of plasma membrane, endosomes and lysosomes; it does not interact with thioredoxin; and its expression is induced by phorbol 12‐myristate 13‐acetate (PMA) (Oka et al., 2006a).
These findings suggest that arrestin domain containing proteins (Txnip, ARRDC and arrestin) may have remarkably different functions.
13 Chapter 1
Figure 5. Protein primary structure of Txnip, arrestin β2 and ARRDCs . Drawings on the left are representations of structures of different proteins, boxes are predicted domains identified, and numbers near tails are the length of each protein. Table on the right indicate the homologies of Txnip to each protein (all human proteins). *: Patwari et al., 2006; † and ‡: isoform 1.
1.4 Txnip Functions
Txnip is a multifunctional protein involved in many physiological and pathological processes (reviewed in Kim et al., 2007; Kaimul et al., 2007; Yoshida et al.,
2006; Chung et al., 2006; Yoshida et al., 2005; Nishiyama et al., 2001). The functions of the human and mouse Txnip have been extensively studied in the last decade, and it appears that Txnip is critical in regulating the cellular redox state and cell proliferation, differentiation and death. Moreover, Txnip is important in modulating glucose and lipid metabolism, and is probably associated with some metabolic diseases, such as diabetes mellitus (Price et al., 2006; van Greevenbroek et al., 2007;
Parikh et al., 2007) and familial combined hyperlipidemia (FCHL, Bodnar et al., 2002).
1.4.1 Txnip and Redox State
Thioredoxin is an important cellular reducing agent, and it reduces target proteins via cysteine thiol‐disulfide exchange (Powis and Montfort, 2001). Txnip
14 Chapter 1
(also known as thioredoxin‐binding protein 2 [TBP‐2]), as the name suggests, can
interact with thioredoxin, leading to an inhibition of the thioredoxin activity; therefore, Txnip is considered to be a negative regulator of thioredoxin (Nishiyama et al., 1999; Yamanaka et al., 2000; Junn et al., 2000). The interaction of Txnip and thioredoxin is probably mediated by forming disulfide between Txnip cysteine 247 and thioredoxin cysteine 32 (Patwari et al., 2006).
Thioredoxin has also been shown to interact with diverse transcription factors and cofactors, such as NF‐κB (Qin et al., 1995), Oct‐4 (Guo et al., 2004), Jun activation domain‐binding protein 1 (Jab‐1, Hwang et al., 2004), redox factor 1 (Ref‐1, Qin et al.,
1996) and GR (Makino et al., 1999), suggesting a link between cellular redox state
(thioredoxin activity) and gene expression. Due to an inhibitory role on the thioredoxin activity, Txnip may also play a role regulating redox‐sensitive gene expression‐‐a notion that is supported by observations of nuclear localization of
Txnip in certain cell types (Nishinaka et al., 2004a).
1.4.2 Txnip and Cell Proliferation and Cell Death
Ectopic expression of Txnip in HeLa cells, aortic smooth muscle cells or
HTLV‐I‐positive T cells suppressed cell growth and significantly reduced S‐phase cells (Joguchi et al., 2002; Schulze et al., 2002; Nishinaka et al., 2004b). On the other hand, Txnip deficient lung fibroblasts developed from Txnip knock‐out mice grow faster than wild type cells, and the percentage of S‐phase cells in Txnip‐/‐ cells is increased about 2‐fold (Jeon et al., 2005). The cell growth and cell cycle progression are repressed in some tumor cells treated with transforming growth factor beta (TGF‐
β) and 1,25(OH)2D3; and these agents are known to be able to up‐regulate Txnip expression (Han et al., 2003). Hence, Txnip is considered to be an anti‐proliferative
15 Chapter 1 protein.
The anti‐proliferative function of Txnip might be mediated by several cell cycle mediators. Txnip can inhibit the expression of cyclin A by recruiting repressor complex (e.g., HDAC1) to the promoter of cyclin A gene, and down‐regulation of
cyclin A leads to arrest of cells at G0/G1 phases (Han et al., 2003). Over‐expression of
Txnip is able to induce the expression of p16, which inhibits cyclin‐dependent kinase
4 (CDK4) and subsequently retinoblastoma protein (Rb); this affects transition of cells from G1 phase to S phase (Nishinaka et al., 2004b). Txnip can also interact with Jab‐1 and prevent p27kip1 from the jab‐1 mediated degradation, thus low expression of
Txnip in tumors leads to increased cell proliferation by reducing the stability of p27kip1 (Tomoda et al., 1999; Jeon et al., 2005).
The role of Txnip in apoptosis appears controversial given that its effect is cell type dependent. Over‐expression of Txnip in primary cardiomyocyte or WEHI7.2 T cell lymphoma promotes apoptosis of these cells (Wang et al., 2002; Wang et al., 2006).
However, in 293, MCF‐7, NIH3T3 and SNU gastric tumor cells, over‐expression of
Txnip is not sufficient to induce apoptosis (Junn et al., 2000; Han et al., 2003;
Nishinaka et al., 2004b). At this stage, it is at least safe to claim that Txnip possesses a pro‐apoptotic activity, as the up‐regulated expression of Txnip is coupled with induction of apoptosis by diverse physiological or pharmacological agents such as glucocorticoids (e.g., dexamethasone), peroxisome proliferator‐activated receptor
(PPAR) agonists, oxidative stress, suberoylanilide hydroxamic acid (SAHA) and high concentration of glucose (see below, section 1.5); in some cases, it has been shown that apoptosis is mediated by the up‐regulated expression of Txnip (Minn et al.,
2005b; Wang et al., 2006; Chen et al., 2008b).
16 Chapter 1
The mechanism for the pro‐apoptotic function of Txnip is not fully
understood. Recently, a role of apoptosis signal‐regulating kinase 1 (ASK‐1) in
mediating Txnip induced apoptosis has been proposed. ASK‐1 is a MAPK kinase
kinase that induces apoptosis by activating JNK and p38 MAPK (Ichijo et al., 1997).
Thioredoxin can bind with ASK‐1, which inhibits the ASK‐1 kinase activity and
promotes ubiquitination and degradation of ASK‐1 (Saitoh et al., 1998; Liu and Min,
2002). In turn, the up‐regulated Txnip expression can quench the cellular thioredoxin activity and induce the ASK‐1‐MAPK signaling pathway that can initiate apoptosis (Junn et al., 2000; Xiang et al., 2005; Yamawaki et al., 2005; Chen et al., 2008c).
Over expression of Txnip may increase cellular oxidative stress (e.g., reactive oxygen species, ROS) which will trigger apoptosis in MAPK‐dependent or ‐independent pathways (Schulze et al., 2004; Li et al., 2009).
The anti‐proliferative and pro‐apoptotic roles suggest that Txnip might play a role in cancer development. The expression of Txnip is dramatically reduced in breast, lung, gastric, and colon tumors (Butler et al., 2002; Ikarashi et al., 2002; Jeon et al., 2005) and in many cancer cell lines (Song et al., 2003; Nishinaka et al., 2004b; Dutta et al., 2005). The CpG island on Txnip promoter is highly methylated in renal cell carcinoma and adult T‐cell Leukemia, which might be a potential mechanism underlying reduced Txnip expression in tumors (Dutta et al., 2005; Ahsan et al., 2006).
Suggesting a role in metastasis, Txnip has been shown to inhibit cell migration (Ng et al., 2007), up‐regulate KiSS‐1 metastasis‐suppressor (Goldberg et al., 2003), and its expression is repressed in malignant tumor (Ohta et al., 2005). Moreover, Txnip‐ deficiency can lead to the predisposition and onset of hepatocellular carcinoma
(Sheth et al., 2006). Taken together, Txnip may function as a tumor‐suppressor.
17 Chapter 1
Indeed, the locus of human Txnip gene, 1q21, is a genomic region frequently mutated or lost in tumors (Keung et al., 1998; Ludwig et al., 2001).
1.4.3 Txnip and Cell Differentiation
Txnip is highly expressed in thymus and spleen, indicating a role of Txnip in immune system (Junn et al., 2000). In Txnip‐/‐ mice, there is a severe reduction of
natural killer (NK) cell population and cytotoxicity (Lee et al., 2005). Dendritic cells
derived from Txnip‐/‐ mice are defective in inducing T cell responses under
stimulation of lipopolysaccharide (Son et al., 2008). Other than immune systems,
Txnip has been shown to be an important regulator of human osteoclast
differentiation (Aitken et al., 2004); in another study, Txnip was shown to mediate the
osteogenic differentiation of mesenchymal stem cells (Li et al., 2007b). Moreover,
Txnip may also play a role in the differentiation process of keratinocytes
(Champliaud et al., 2003) and alveolar epithelial cells (Filby et al., 2006). The above lines of evidence suggest that Txnip can play a role in cell differentiation, although at this stage the underlying molecular mechanisms remain to be investigated.
1.4.4 Txnip and Cellular Metabolism
A wealth of accumulated information suggests an intimate involvement of
Txnip in regulating glucose metabolism. The expression of Txnip is induced by glucose (Hirota et al., 2002; Shalev et al., 2002; Kobayashi et al., 2003; Schulze et al.,
2004; Turturro et al., 2007), and this induction is mediated by a carbohydrate response element (ChoRE, defined by a tandem E‐box separated 5 nucleotides) and its associated transcription factor complex comprising MondoA and Max‐like protein
X (MLX) (Minn et al., 2005b; Stoltzman et al., 2008). On the other hand, Txnip can function as a repressor for glucose transport (Parikh et al., 2007; Yoshioka et al., 2007).
18 Chapter 1
The Txnip expression is elevated in patients with diabetes and prediabetes (Schulze
et al., 2004; Minn et al., 2005a; Parikh et al., 2007), and this could account for reduced glucose uptake and utilization in peripheral tissues (e.g., skeletal muscle and adipose tissue; Parikh et al., 2007). In Txnip‐/‐ mice, the hepatic glucose production is reduced
(Chutkow et al., 2008), and the myocardial glucose uptake is robustly increased
(Yoshioka et al., 2007). These observations indicate a negative feed‐back regulatory loop between Txnip expression and glucose transport at the cellular level. In the pancreas, Txnip induces the apoptosis of β‐cells (Minn et al., 2005b; Chen et al., 2008b), which are master regulators for glucose homeostasis via their production of insulin.
Txnip may also have an inhibitory role on insulin secretion (Hui et al., 2004; Minn et al., 2005a; Masson et al., 2009; Oka et al., 2009); however, Txnip expression is repressed by insulin (Parikh et al., 2007; Shaked et al., 2009), thus the expression of
Txnip and secretion of insulin might be two processes reciprocally restricted. Taken together, Txnip is an important protein in the network of glucose metabolism, and probably a link for glucose metabolism and cellular redox state (Muoio, 2007).
Txnip is also important in lipid metabolism. A mutant mouse strain, HcB‐19, exhibits characteristics of FCHL, such as hypertriglyceridemia and hypercholesterolemia; a nonsense mutation in Txnip gene has been identified in
HcB‐19 mice, this mutation produces a truncated Txnip which may account for the
FCHL‐like phenotype of HcB‐19 mice (Bodnar et al., 2002). A more recent study shows that Txnip is associated with hypertriglyceridemia and blood pressure in diabetes mellitus (van Greevenbroek et al., 2007). However, in some other gene association studies, USF1 has been identified as the causative gene for the FCHL‐like phenotype of HcB‐19 mice (Pajukanta et al., 2004; Coon et al., 2004; van der Vleuten et
19 Chapter 1 al., 2004). A clearer role of Txnip in lipid metabolism has been revealed in a study
using Txnip‐/‐ mice (Oka et al., 2006b); in this model, mice exhibited hyperlipidemia, hypoglycemia and high levels of plasma ketone bodies, lactate and pyruvate, indicating that Krebs cycle‐mediated fatty acid utilization in Txnip‐/‐ mice is severely compromised. In another study using Txnip knock‐out mice, the Txnip deficiency enhances lipogenesis, which contributes to a fatty liver phenotype (Donnelly et al.,
2004).
Other than regulating glucose and lipid metabolism, Txnip may also be involved in some other metabolic related processes. For instance, the fasting‐feeding metabolic transition is disrupted in Txnip deficient mice (Sheth et al., 2005). Recently, it has been shown that the signaling pathway of phosphatase and tensin homolog
(PTEN) and Protein Kinase B (PKB, or AKT) is regulated by Txnip, suggesting that
Txnip may have a broader role in metabolic control given widely recognized roles of
PTEN and PKB in cell growth and metabolism (Chen et al., 2008a; Hui et al., 2008).
1.5 Txnip Expression in Response to Different Signals
The expression of Txnip gene is sensitive to a diverse environmental factors and physiological cues. The modulation of Txnip expression by these signals is important for mediating their physiological functions.
Txnip is considered to be an early response gene to some stress conditions.
The expression level of Txnip is induced by heat shock, high cell density, serum deprivation (Kim et al., 2004), hypoxia (Xiang et al., 2005; Le Jan et al., 2006; Karar et al., 2007; Baker et al., 2008), ultraviolet (UV) radiation (Cheng et al., 2004) and cell senescence (Joguchi et al., 2002). On the other hand, Txnip expression is repressed by
20 Chapter 1
Ca2+ influx, H2O2 treatment (Saitoh et al., 2001; Wang et al., 2002), nitric oxide
treatment (Schulze et al., 2006) and fluid shear stress (Yamawaki et al., 2005). Among
these modulators, the effect of heat shock is mediated by an HSE on Txnip promoter
and its associated HSF (Figure 6; Kim et al., 2004).
Some drugs against DNA and histone modification enzymes can also regulate
Txnip expression. For instance, Txnip is induced by 5‐AZA‐2ʹ‐deoxycytidine (5‐aza‐
CdR, a DNA demethylating agent; Ahsan et al., 2006) and SAHA (a HDAC inhibitor;
Huang and Pardee, 2000; Butler et al., 2002; Ahsan et al., 2006; Xu et al., 2006). The
stimulatory effect of SAHA on Txnip expression is mediated by an inverted CCAAT box on Txnip promoter (Figure 6; Butler et al., 2002).
Figure 6. Signaling pathways regulating Txnip expression and their targeting cis‐ regulatory elements on Txnip promoter. TSS: transcription start site; iCAT: inverted CCAAT box; b.s.: binding site. Numbers indicate the distance to TSS.
Txnip expression is sensitive to extracellular glucose levels (Hirota et al., 2002;
Shalev et al., 2002; Kobayash et al., 2003; Shulze et al., 2004; Minn et al., 2005b; Cheng et al., 2006; Turturro et al., 2007; Stoltzman et al., 2008). The effect of glucose on Txnip expression is mediated by the ChoRE on Txnip promoter and its associated
21 Chapter 1 transcription factors, MondoA and MLX; MondoA and MLX form a protein complex
(di‐dimer) which may be sensitive to intracellular glucose‐6‐phophate (G6P) levels and shuttle between cytoplasm (probably mitochondria) and nucleus (Figure 6; Minn et al., 2005b; Stoltzman et al., 2008). The Txnip expression can be induced by D‐allose
(Yamaguchi et al., 2008), glucosamine (an amino sugar; Cheng et al., 2006) and 3‐O‐
Methylglucose (3OMG; Minn et al., 2006) as well. 3OMG is a glucose homolog that is resistant to hexokinase, hence is not to be phosphorylated and will accumulate in cells. This suggests that the glucose‐induced Txnip expression may not be solely dependent on cellular G6P levels.
Several steroid hormones have been shown to modulate Txnip expression.
Txnip gene was initially identified as a vitamin D up‐regulated gene (Chen and
Deluca, 1994); glucocorticoid was later found to induce Txnip expression (Kolbus et al., 2003; Wang et al., 2006; Tissing et al., 2007). On the other hand, estradial (E2, a major estrogen in human) can repress the Txnip expression (Deroo et al., 2004;
Simmons and Kennedy, 2004). On Txnip promoters, candidate glucocorticoid response element (GRE) and vitamin D response element (VDRE) have been identified, which should be able to recruit cognate factors (nuclear receptors) to stimulate Txnip expression in glucocorticoid‐ or vitamin D‐treated cells (Figure 6;
Wang et al., 2006; Butler et al., 2002). Other than these hormones, Txnip expression is also up‐regulated by PPARα or PPARγ agonists (Oka et al., 2006a; Rajhshandehroo et al., 2007; Billiet et al., 2008a; Billiet et al., 2008b). PPAR agonists may exert their function by activating PPAR bound to a peroxisome proliferator hormone response element (PPRE) on the Txnip promoter, activating Txnip transcription (Figure 6;
Rajhshandehroo et al., 2007; Billiet et al., 2008b). PPAR is able to sense intracellular
22 Chapter 1 free fatty acid levels (Berger and Moller, 2002), again suggesting that Txnip is linked to fatty acid metabolism.
Txnip expression is also affected by some well‐known extracellular signaling molecules. The expression level of Txnip gene is repressed by insulin (Parikh et al.,
2007), IGF‐1 (Saitoh et al., 2001) and NMDA receptor activation (Papadia et al., 2008) and induced by TGF‐β (Shalev et al., 2002; Han et al., 2003). Insulin may activate
PI3K and PKB that in turn can modulate the activity of Forkhead box O1 (FOXO1), for which the Txnip promoter has a binding site (Figure 6; Schulze et al., 2004; Parikh et al., 2007; de Candia et al., 2008).
Importantly, some drugs in clinical trails also impact Txnip expression. For instance, exenatide, an anti‐diabetic drug, can reduce Txnip expression (Chen et al.,
2006); and 5‐fluorouracil, an anti‐cancer drug can increase Txnip expression
(Takahashi et al., 2002). These observations suggest that Txnip might be a good pharmacological target for some metabolic or cell cycle related diseases.
1.6 Objectives
The expression of Txnip is very sensitive to environmental and physiological conditions, and has critical roles in multiple physiological processes. One of the most attractive phenomena to me is that the Txnip expression is induced by glucose and that the induction is tightly correlated with extracellular glucose levels.
Therefore, the Txnip expression is considered to be an intracellular sensor for extracellular glucose levels. In the human genome, expression of many glucose‐ responsive genes required both glucose and insulin signals (Dentin et al., 2004); however, glucose (positive) and insulin (negative) impose opposing effects on, and
23 Chapter 1 glucose signaling alone is sufficient to induce, Txnip expression. Hence, I reason that the Txnip expression provides a unique model to study the glucose responsiveness in mammalian systems.
While the major focus of the research community has been on the physiology of Txnip at the organism level, the study on how the gene encoding Txnip is expressed and induced by glucose has been rudimentary. In this study, I wish to contribute towards understanding the underlying mechanisms by which the Txnip expression is regulated, under conditions with or without of glucose. The outcome of this study may help us to understand how the gene regulatory network in mammalian cells responds to the most basic nutritional signal, glucose.
Given that Txnip expression is related to a diverse number of physiological functions and pathology of some diseases, the identification of molecules that can modulate Txnip expression has important clinical relevance. Therefore, in this study,
I also aim to identify new molecules that can impact Txnip expression, and study the underlying mechanisms of the newly identified molecules on transcriptional switches; in this project, attention will be given to metabolic‐related molecules. I hope that identification of such molecules may not only provide a tool for regulating
Txnip expression, probably in conjunction with glucose, but also set a stage for selection and development of drugs for treatment of metabolic disorders and cancer.
24
Chapter 2
Materials and Methods Chapter 2
2.1 Chemicals and Buffers
All chemicals were purchased from Sigma except for the following: SAHA from Cayman Chemical, myristoylated PKI (14‐22) from Biomol International, and 9‐
β‐D arabinofuranoside (araA) from Calbiochem. Compositions of all buffers used in
this study are shown in Appendix I.
2.2 Plasmid Constructs
High fidelity DNA polymerase, Phusion (Finnzymes), was used to amplify
DNA fragments to construct all plasmids. From a long Txnip promoter fragment
obtained by PCR of genomic DNA, various promoter fragments were generated by
subsequent PCR and inserted into the pGL3 vector using the Xho1 and Nhe1 sites.
Mutations at ChoREs, CCAAT box, inverted CCAAT box or FOXO binding site were
introduced into the Txnip promoter (‐269) by a PCR‐based method. Shorter Txnip
promoter DNA, or Shuffle, dChoRE‐a, dChoRE‐b, ChoRE‐ab, ChoRE‐a and ChoRE‐b
(double‐stranded DNAs annealed from single‐stranded oligonucleotides) fragments,
were inserted into a fire fly luciferase reporter driven by a minimum (65 base pair
nucleotides containing a TATA box) core promoter, using the Mlu1 and Nhe1 sites.
Three partial cDNAs of MondoA were amplified from total cDNA prepared
from U2OS cells; joined in order, the three fragments form the full‐length MondoA
cDNA that was inserted into the expression vector pdHA, pdMyc or pdFLAG (pCI‐
neo with tandem HA‐, Myc‐ or FLAG‐tags). ChREBP was similarly cloned. MLXβ,
USF1, USF2 and NF‐YA cDNAs were directly amplified using RT‐PCR and inserted
into the pdHA, pdMyc or pdFLAG vectors. Dominant negative MLX was created as
described (Ma et al., 2005), and the RNAi‐resistant MondoA (MondoAr) was made by
25 Chapter 2 a PCR‐based approach. A dominant negative USF (A‐USF) construct was provided by Dr. C. Vinson (NCI). His‐tagged USF1 cDNA sequence was inserted in the pET vector. Partial coding sequences of MondoA or MLX were inserted into the pGEX‐
4T1 vector to generate constructs for producing GST fusion proteins as antigens.
Sequences of all plasmids were confirmed by DNA sequencing using BigDye
(Applied Biosystems). Primers used for cloning, mutagenesis and sequencing were synthesized by Proligo; the sequence information for these primers is shown in
Appendix IV.
Qiagen kits were used in this study to purify plasmids (QIAprepSpin
Miniprep Kit or QIAFilter Plasmid Midi Kit) or DNA fragments (QIAquick PCR
Purification Kit or QIAquick Gel Extraction Kit).
2.3 Purification of Bacterially Expressed Recombinant Proteins
A picked single colony of transformed bacteria (E. coli, BL‐21) was amplified in LB medium (with appropriate antibiotics) overnight at 37oC. Then the culture was diluted 10‐fold with fresh LB medium (with antibiotics) and incubated at 37oC until the optical density (OD; at 600 nm) reading reached 0.4. Protein expression was then induced by 0.5 mM (final concentration) of isopropyl‐beta‐D‐thiogalactopyranoside
(IPTG) with further incubation at room temperature for 6 hrs.
Bacteria were collected by centrifugation (4,000 rpm, 20 min), and pellets re‐ suspended in bacteria lysis buffer followed by sonication. Insoluble materials were then spun down (15,000 rpm, 20 min), and the supernatant containing recombinant proteins collected. To purify GST‐tagged proteins, Glutathione Sepharose beads (GE
Healthcare, 50% slurry in lysis buffer) were added into the lysate and incubated for 1
26 Chapter 2 hr at 4oC. The beads were washed by lysis buffer (without lysozyme) 4 times and
PBS 2 times; GST‐tagged proteins were eluted with PBS containing 10 mM reduced
glutathione. His‐tagged recombinant proteins were purified on Ni Sepharose beads
(GE Healthcare), and proteins eluted with PBS containing 400 mM imidazole. The
concentrations of purified proteins were estimated using BSA standards.
2.4 Mammalian Cell Culture
Cells were maintained at 37oC with 5% CO2. Jurkat T and Namalwa B cells
were grown in RPMI‐1640 (Sigma), and all other cells in DMEM (Sigma, with 1 g/L
glucose); medium were supplemented with antibiotics (Invitrogen), L‐glutamine
(Invitrogen) and 10% fetal bovine serum (HyClone). The glucose‐free DMEM
medium (Sigma) was supplemented with (additional) 2 mM sodium pyruvate.
2.5 Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (SDS‐PAGE)
Polyacrylamide gels used for reducing SDS‐PAGE were prepared according to the recipe shown in Appendix I. Protein samples were made soluble in the SDS‐
PAGE sample buffer, denatured by boiling (10 min), and loaded onto prepared gels.
Electrophoresis was carried out in SDS‐PAGE running buffer at 100 V for 20 min, and the voltage was increased to 180 V until the dye front reached the bottom. The resolved proteins were detected by Coomassie Blue staining or immuno‐blot.
2.6 Immuno‐Blot
Following SDS‐PAGE, resolved proteins were transferred onto nitrocellulose membrane via electro‐blotting. Electro‐blotting was carried out at 100 V for 1 hr in
27 Chapter 2 transfer buffer. The membrane was blocked with blocking buffer (5% skim milk in
TBST) for 1 hr and then incubated with primary antibodies for 1 hr at room temperature. After 3 washes with TBST, the membrane was incubated with secondary antibodies (1:5,000 dilution; GE Healthcare) for 1 hr. Finally, after 3 washes with TBST, enhanced chemiluminescence substrate (GE healthcare) was applied evenly onto the membrane, and chemiluminescence was detected using Fuji
Medical X‐ray film (Fuji) and Kodak developer.
The expression of HA (hemagglutinin)‐tagged proteins was detected with mouse monoclonal anti‐HA antibodies (1:1,000; Millipore), and the control GAPDH level detected by polyclonal anti‐GAPDH antibodies (1:5,000; raised with in‐house rabbits). Txnip antibodies (JY1 and JY2) were purchased from MBL, and p21cip1 antibodies were purchased from BD biosciences. Anti‐mouse IgG κ chain secondary
antibody was from Invitrogen.
2.7 RNA Extraction, RT (Reverse Transcription)‐PCR and Real‐Time PCR
RNA was extracted using RNeasy Mini Kit (Qiagen), and the quality of RNA prepared was checked by agarose gel electrophoresis (Appendix II). Complementary
DNA (cDNA) was synthesized from RNA samples using reverse transcription with the SuperScript III reverse transcriptase and random hexamers, in the presence of
RNaseOUT to maximize RNA stability (all reagents were from Invitrogen). PCR was carried out using Taq DNA polymerase (New England Biolabs) with gene‐specific primers that were synthesized by Proligo (Sigma). Real‐Time PCR employed the
Sybr Green Core Reagents and the 7300 Real‐Time PCR system (Applied Biosystems) using the same primers as RT‐PCR, the sequences of which are shown in Appendix
28 Chapter 2
IV. The PCR product exhibited a unique peak in dissociation curve (Appendix III),
suggesting that the primers are specific.
2.8 Genomic DNA Extraction
Around half million cells were harvested and washed with PBS. Cell pellet was then re‐suspended in 40 μl of Genomic DNA extraction buffer with 0.5 μl of proteinase K (20 mg/ml stock) and incubated at 50oC. After 4 hrs, 150 μl of distilled water was added and the mixture was boiled for 10 min. The mixture was then centrifuged at 6,000 rpm for 10 min, and 1 μl of supernatant was used for PCR.
2.9 Small Interfering RNAs (siRNAs) and RNA Interference Assays
Duplexes of siRNA were synthesized by Proligo (Sigma). Oligofectamine
(Invitrogen) was used to delivered siRNAs into cells. Typically, 5 μl of stock siRNAs
(20 μM) and 5 μl of Oligofactamine was mixed with 100 μl of OPTI‐MEM
(Invitrogen) at room temperature for 10 min, and then added drop‐wise into the well with cells (30% confluency) incubated in 0.9 ml OPTI‐MEM. At 4 hrs, 0.5 ml of
complete DMEM containing 30% serum was added into each well. At 48 hrs, cells
were harvested for analysis of RNA or protein levels. Sequences of siRNAs (sense
strand) are:
MLX GAGUAUGCCUACAGCGACAdTdT,
MondoA GAACAACUGCUCAGGGAAAdTdT
USF1 GACCCAACCAGUGUGGCUAdTdT
USF2 CCUCCACUUGGAAACGGUAdTdT
Random UCAGUGUCAUACGUACGACdTdT
29 Chapter 2
2.10 Transfection
Lipofactamine 2000 (Invitrogen) was used in all transfection experiments.
Cells were cultured in 6‐well plates to about 80% confluency with antibiotics‐free medium. During transfection, 4 μl of lipofactamine 2000, 1 μg of Txnip promoter plasmid and 10 ng of simian virus 40‐renilla luciferase plasmid, with or without 0.5
μg of expression plasmids, were mixed with 100 μl of OPTI‐MEM (Invitrogen) for 20 minutes, and the mixture was then added drop‐wise into the well with cells incubated in 0.9 ml OPTI‐MEM. At 5 hrs, the medium was changed to complete medium. For plasmids and siRNA co‐transfection experiment, 5 μl of lipofactamine
2000, 1 μg of Txnip promoter plasmid, 100 pM siRNA and 10 ng of renilla luciferase plasmid, with or without 50 ng of expression plasmid, were used.
2.11 Promoter Activity (Reporter) Assays
The Promega Dual‐Luciferase Reporter (DLR) Assay System was used to determine the activities of Txnip promoters. Unless otherwise indicated, cells were treated with 0.2 mM of NAD(H) or ATP for 16 hrs after transfection, and cell lysates were prepared to measure the firefly or renilla luciferase activities.
2.12 Immunocytometry
HA‐MondoA and Myc‐MLX were co‐transfected into L6 cells grown on glass slips, and the cells cultured in glucose‐free DMEM with 2% FBS. Following different treatments, cells were fixed with 4% paraformaldehyde and subjected to indirect immunostaining. Anti‐HA antibody (1:500; Millipore) and Alexa Fluor 488 anti‐
30 Chapter 2 mouse IgG (1:500; Invitrogen) were used to label HA‐tagged MondoA; DAPI,
included in mounting media (Vector Shield), was used to locate the nuclei. The HA‐
MondoA subcellular localizations (cytoplasmic [cyt], nuclear [nuc] or heterogeneous
[cyt/nuc]) in 100 randomly selected cells were determined using confocal microscopy
(Olympus). Results from three experiments were used for the data analyses.
2.13 Fluorescence‐activated Cell Sorting (FACS)
Cells grown in 6‐well plates were treated with bromodeoxyuridine (BrdU, 10
μM) for 45 min before harvesting. Cells were then fixed with 70% cold ethanol and treated with 3N HCl to denature DNA. HCl was removed by 3 washes using PBST.
To stain incorporated BrdU, cells were incubated with the mouse anti‐BrdU monoclonal antibody (1:50; BD) for 30 min. After washing, cells were incubated with the Alexa Fluor 488 anti‐mouse IgG (1:60; Invitrogen) for 30 min. Finally, cells were treated with propidium iodide (50 μg/ml) and RNase A (20 μg/ml) for 30 min and subjected to analyses using FACS Scan (BD Biosciences), 10,000 cells were counted for each sample.
2.14 Electrophoresis Mobility Shift Assays (EMSA)
Oligonucleotides used for EMSA was synthesized by Proligo (Sigma). Sense and anti‐sense oligonucleotides were annealed to double stranded DNAs (dsDNA), which were then labeled with 33P by Klenow reaction using exo‐ Klenow enzyme
(New England Biolabs) and 33P dATP (GE Healthcare). Labeled dsDNAs were
purified using nick columns equilibrated in 2× GS buffer.
31 Chapter 2
Table 1. Oligonucleotides used for EMSA Probes Sequence GACCGGGCAGCCAATGGGAGGGATGTGCACGAGGGCAGC ACGAGCCTCCGGGCCAGC WT‐a TGCTGGCCCGGAGGCTCGTGCTGCCCTCGTGCACATCCCTC CCATTGGCTGCCCGGTC GACCGGGCAGCCAATGGGAGGGATGTGTATGAGGGCAGT ATGAGCCTCCGGGCCAGC MutCho‐a TGCTGGCCCGGAGGCTCATACTGCCCTCATACACATCCCT CCCATTGGCTGCCCGGTC GACCGGGCAGCTACTGGGAGGGATGTGCACGAGGGCAGC ACGAGCCTCCGGGCCAGC MutNFY‐a TGCTGGCCCGGAGGCTCGTGCTGCCCTCGTGCACATCCCTC CCAGTAGCTGCCCGGTC CAGCCAGGAGCACACCGTGTCCACGCGCCACAGCGATCT CACTGATTGGTCGGGCTC WT‐b TGAGCCCGACCAATCAGTGAGATCGCTGTGGCGCGTGGAC ACGGTGTGCTCCTGGCTG CAGCCAGGAGTATACCGTGTCTATGCGCCACAGCGATCTC ACTGATTGGTCGGGCTC MutCho‐b TGAGCCCGACCAATCAGTGAGATCGCTGTGGCGCATAGAC ACGGTATACTCCTGGCTG CAGCCAGGAGCACACCGTGTCCACGCGCCACAGCGATCT CACTGAGTAGTCGGGCTC MutNFY‐b TGAGCCCGACTACTCAGTGAGATCGCTGTGGCGCGTGGAC ACGGTGTGCTCCTGGCTG sCho‐a GGGAGGGATGTGCACGAGGGCAGCACGAGCCTCCGGGCC (EM‐a) TGGCCCGGAGGCTCGTGCTGCCCTCGTGCACATCCCTCCC MsCho‐a GGGAGGGATGTGTATGAGGGCAGTATGAGCCTCCGGGCC (mEM‐a) TGGCCCGGAGGCTCATACTGCCCTCATACACATCCCTCCC sCho‐b CCCAGCCAGGAGCACACCGTGTCCACGCGCCACAGCGAT (EM‐b) TATCGCTGTGGCGCGTGGACACGGTGTGCTCCTGGCTGGG MsCho‐b CCCAGCCAGGAGTATACCGTGTCTATGCGCCACAGCGAT (mEM‐b) TATCGCTGTGGCGCATAGACACGGTATACTCCTGGCTGGG GATCTCACTGATTGGTCGGGCTC sNFY GAGCCCGACCAATCAGTGAGATC GGGAGGGATGTGCACGTGGGCAGCACGAGCCTCCGGGCC E‐box TGGCCCGGAGGCTCGTGCTGCCCACGTGCACATCCCTCCC
32 Chapter 2
A typical EMSA reaction (12.5 μl) was set up as shown below:
Labeled probe ~10 fmol, top up with 2× GS buffer to 2.5 μl H2O 2.5 μl 1× GS buffer 5 μl Proteins recombinant protein or 5 μg of whole cell extract, top up with BC100/BSA to 2.5 μl
The EMSA reaction was carried out at 30oC for 30 min. In competition or
super‐shift assays, the 2.5 μl of H2O was replaced by cold probes (100× in excess) or antibodies. Free probes and protein‐DNA complexes were separated on native 4% polyacrylamide gels (Acrylamide/Bisacrylamide, 59:1 linkage, Bio‐Rad; see recipe in
Appendix I) in TGEMN running buffer. Gels were then dried and exposed to X‐ray film (Kodak) overnight with intensifier screen at ‐80°C.
Table 2. Primers used in ChIP assay. Primers Remarks Primers a and b target at left and right sides of ChIP‐a the multiple cloning site (MCS) of pEGFP‐1 GTTCTTTCCTGCGTTATCCC respectively. PCR Amplifies promoters in ChIP‐b pEGFP‐1 (~100 bp product for empty vector, GACCAGGATGGGCACCAC and ~700 bp product for plasmids with insertion of Txnip promoter. ChIP‐a Primer a targets left site of MCS of pEGFP‐1, GTTCTTTCCTGCGTTATCCC and primer d targets Txnip promoter. PCR ChIP‐d amplifies the ectopic Txnip promoters. CTGCCCGGTCCTTGTTTAC Primers c and d target Txnip promoter ChIP‐c sequences, but primer c does not target AGGTTTTAGGGTCAGTGGGAT sequences of Txnip promoters in the pEGFP‐1 ChIP‐d vector. PCR amplifies endogenous Txnip CTGCCCGGTCCTTGTTTAC promoter. Neg‐f ATGGTTGCCACTGGGGATCT Negative control primers. Target chromosome Neg‐r 12, NT_009759.15 TGCCAAAGCCTAGGGGAAGA
33 Chapter 2
2.15 Chromatin Immunoprecipitation (ChIP) Assays
Wild‐type or mutant Txnip promoters (same as promoters used in luciferase reporter assays) were inserted into the pEGFP‐1 vector, which along with the empty control vector were transfected into HeLa cells to establish G418 resistant stable cell lines. Chromosome integration of the constructs and stable EGFP expression were confirmed by PCR and/or fluorescence microscopy. These stable cell lines, following different treatments, were used in ChIP assays.
Cells were firstly cross‐linked with 1% formaldehyde (room temperature, 10 min), and formaldehyde quenched by glycine. Cells were then harvested, washed with PBS twice and incubated in cold MC buffer (Aparicio et al., 2005) for 30 min.
Afterwards, cell pellets were sonicated prior to immuno‐precipitation with the ChIP
Assay Kit (Upstate). Briefly, after sonication, lysates were centrifuged at 15,000 rpm for 15 min at 4oC, the supernatants collected, diluted (10×) with ChIP dilution buffer and pre‐cleared using 100 μl (for 3 ml of diluted lysate) of agarose beads (coated with protein A and G in 1:1 ratio in 50% slurry buffered in salmon sperm DNA) for 2 hrs at 4oC. Antibodies (4 μg) were then added into the pre‐cleared lysate and rotated
overnight at 4oC; 2% of the lysate was used as input control. Next morning, 50 μl of
protein A/G agarose beads were added, and rotated for 1 hr at 4oC. The beads were washed, respectively, twice with low salt buffer, high salt buffer, LiCl buffer and TE at 4oC. Protein‐DNA complexes were eluted with 200 μl of elution buffer (1% SDS in
100 μM of NaHCO3) at room temperature. To reverse the cross‐linking, 200 mM
(final concentration) of NaCl was added and incubated at 65oC overnight. To remove
RNA and proteins, RNAse and Proteinase K were added and incubate at 45oC for 2 hrs. Recovered DNA fragments were then purified using PCR clean up kit (Qiagen).
34 Chapter 2
Finally, the abundances of DNA fragments of interest were analyzed by PCR using
Taq DNA polymerase.
Antibodies used in ChIP assays are: anti‐human MLX (goat, R&D Systems), anti‐NF‐YA (Santa Cruz), anti‐NF‐YB (Diagenode), Rabbit and Goat IgGs (Santa
Cruz). Information of primers used for PCR is shown in Table 2.
2.16 Thioredoxin Activity Assays
Following NAD+ or ATP treatment, cellular thioredoxin activity was measured according to a published method (Arner and Holmgre, 2000). Briefly, a thioredoxin assay master mix containing insulin and NADPH was prepared. To
measure cellular thioredoxin activity, 5 μg of cell lysate (top up to 7 μl using 200 mM
of HEPES buffer, pH 7.6) was mixed with 33 μl of master mix and 10 μl of
thioredoxin reductase (250 nM). At the same time, a control reaction in which
thioredoxin reductase was replace by HEPES buffer was set up. All reactions were
carried out at 37oC for 20 min and stopped by adding 500 μl of DTNB/guanidine solution (1 mM of DTNB [5,5′‐dithio‐bis(2‐nitrobenzoic acid] and 5.4 M guanidine hydrochloride, pH 8.0). Absorbance (412 nm) was measured in a spectrophotometer, and the thioredoxin activity of a sample was calculated by comparing the net reading
(after deduction from a control reading) with a corresponding reading in a standard curve constructed using the commercial thioredoxin.
2.17 Glucose Transport Assays
U2OS cells (in 24‐well plate) were treated with 0.1 mM of NAD+ for 0, 2, 4, 8
or 16 hr(s), and the cellular uptake of H3‐2DG was measured. Briefly, cells were
35 Chapter 2 washed twice with KRH/0.1% BSA and incubated in 450 μl of KRH/0.1% BSA, and then 50 μl of 10× START was added for 10 min. After two washes with cold PBS, cells
were lysed by 0.5 ml of 0.05% SDS. To determine the uptake of H3‐2DG, 0.4 ml of
lysate was counted in scintillation fluid, 50 μl of 10× START was counted for specific activity calculations. One group of samples were treated with cytochalasin B (10 μM,
final concentration), and the reading was taken as non‐specific uptake.
2.18 Statistical Analyses
Data were analyzed by unpaired homoscedastic t‐test; and two tailed P value less than 0.05 was considered statistically significant. Single asterisk (*), diamond (♦) or dot (●) correspond to P values less than 0.05; double asterisk, diamond or dot indicate P values less than 0.005. Data without these indicators, or labeled with NS
(not significant), suggest P values at ≥0.05. Unless otherwise indicated, duplicated data were used in statistical analysis.
36
Chapter 3
Identification of Molecules Modulating Txnip Expression Chapter 3
3.1 Preface
In order to cope with nutritional or metabolic changes, eukaryotic cells
developed molecular sensors, either on plasma membrane or inside cells, to sense extracellular nutrients or intracellular metabolic status; these sensors can activate some signaling pathways to modulate gene expression. For instance, in yeast, Ssy1 can sense levels of multiple external amino acids and regulate expression of AGP1, encoding an amino acid permease (Iraqui et al., 1999); Snf3 and Rgt2 can sense extracellular glucose concentration and regulate expression of hexose transporter
(HXT) genes (Ozcan et al., 1998). In mammals, VDR can be activated by vitamin D to regulate transcription of TRPV6, a membrane calcium channel (Taparia et al., 2006);
AMPK can sense intracellular AMP/ATP ratio (Scott et al., 2004) and mammalian target of rapamycin (mTOR) is a ATP sensor (Dennis et al., 2001), and both can
regulate many genes involved in metabolism. Deregulation of gene expression in response to nutritional or metabolic status has been linked to metabolic diseases.
Txnip is a multifunctional protein that is involved in many cell cycle events and metabolic processes (reviewed in Kim et al., 2007). The expression level of Txnip is linked to several diseases, such as cancer (Butler et al., 2002; Ikarashi et al., 2002;
Jeon et al., 2005), diabetes mellitus (Price et al., 2006; van Greevenbroek et al., 2007;
Parikh et al., 2007), hyperlipidemia (Bodnar et al., 2002; Oka et al., 2006b) and atherosclerosis (Harrison, 2005). Therefore, identification of small molecules that can modulate the Txnip expression has important clinical relevance; this is one primary objective of my study.
The expression of Txnip is downstream of glucose signaling, and Txnip plays important role in cellular metabolism, suggesting a critical role of Txnip in nutrients
37 Chapter 3 sensing and cellular metabolism. Thus, in searching for molecules modulating Txnip expression, I have paid more attention to small molecules available in natural dietary sources, and some molecules related to cell metabolism. The identification of such
Txnip expression modulators can also help us to better understand role of Txnip in cellular metabolism, especially its response to metabolic signals other than glucose.
3.2 Results
3.2.1 Adenosine‐containing Molecules Induce Txnip Expression
3.2.1.1 NAD(H) and ATP Stimulate Txnip Expression
In a screen for small molecules that impact Txnip expression, I identified
NAD(H) and ATP as compounds that could induce the Txnip mRNA expression levels. HeLa cells were treated with NAD+, NADH or ATP for 4 hrs, and the Txnip mRNA levels were measured by RT‐PCR. As shown in Figure 7, the mRNA levels in cells treated with NAD(H) or ATP were dramatically increased; on the other hand, the mRNA level of a control gene, β‐actin, was largely constant.
Figure 7. NAD(H) or ATP induced Txnip mRNA expression. HeLa cells were treated with 0.1 mM of NAD(H) or ATP for 4 hrs or left untreated, total RNA was extracted and reverse transcribed, and Txnip cDNA levels measured by PCR using human Txnip specific primers; β‐actin gene was used as an endogenous control.
38 Chapter 3
Two commercial antibodies against Txnip, JY1 or JY2, did not give rise to strong signals in immuno‐blots, presumably because the Txnip protein level in HeLa cell lysates was low. To circumvent this, I first carried out immuno‐precipitation (IP) using JY1 or control IgGs; the immuno‐precipitates were analyzed by immuno‐blot with JY2. While the input protein (GAPDH control) levels in immunoprecipitation were similar, the levels of Txnip protein were clearly induced in cells treated with
NAD+ for 4 or 8 hrs (Figure 8). Therefore, both mRNA and protein levels of Txnip
can be up‐regulated by NAD+.
Figure 8. NAD+ induced Txnip protein level. HeLa cells were treated with 0.1 mM NAD+ for 4 or 8 hrs, or left untreated. The amount of Txnip protein in these samples was measured by immunoprecipitations (IP) using mouse IgGs or anti‐Txnip (JY1) antibodies. Immuno‐precipitates were detected by immuno‐blot using anti‐Txnip (JY‐2) antibodies. Total input protein levels were similar as shown by immuno‐blot (5% of input lysates) using anti‐GAPDH antibodies.
The mRNA levels were also quantified by Real‐Time PCR. HeLa cells treated with 0.1 mM of NAD(H) or ATP for 4 hrs exhibited >4‐fold increase in Txnip mRNA
levels (Figure 9). The response of Txnip expression to NAD(H) or ATP were also tested in several other mammalian cell lines such as U2OS, WI‐38, HepG2, C2C12,
293, Namalwa B and Jurkat T cells. In all tested cells lines, of different tissue origins, the Txnip mRNA expression level was induced by NAD(H) or ATP, suggesting that the induction of Txnip expression by NAD(H) or ATP is a ubiquitous phenomenon.
39 Chapter 3
12 * * - 10 * NAD+ 8 NADH ATP * 6 * * * *** 4 * * * * 2 ** Txnip mRNA Level Txnip mRNA 0 3 12 B eLa 2OS 29 H WI-38 2C U HepG2 C Jurkat T Namalwa Figure 9. NAD(H) or ATP induced Txnip expression in diverse cell lines. The mRNA levels in samples treated by adenosine‐containing molecules (0.1 mM; 4 hrs) or untreated control cells were measured by Real‐Time PCR. Asterisk: comparison with untreated sample.
3.2.1.2 Molecules Containing Adenosine Group Induce Txnip Expression
As shown in Figure 10, NAD(H) and ATP share a common adenosine group.
This prompted me to speculate that the adenosine group in these molecules is the functional group responsible for the induction of Txnip expression, and evaluate the effects of other adenosine‐containing molecules on Txnip expression levels. Thus,
HeLa cells were incubated with NADP(H), ADP, ADP‐ribose, FAD, AMP and adenosine (0.1 mM, 4 hrs), and the Txnip mRNA levels were similarly induced as observed in the NAD(H)‐ or ATP‐treated cells (Table 3). Therefore, a class of small molecules containing adenosine group have been identified as inducers for Txnip expression.
3.2.1.3 Adenosine is Necessary and Sufficient for Inducing Txnip Expression
The above results (Figure 9 and Table 3) suggest that an adenosine moiety is sufficient for the induction of Txnip expression, but it is not clear whether the entire adenosine moiety is necessary for this induction.
40 Chapter 3
Figure 10. NAD+ and ATP share a common adenosine group. The adenosine group in NAD(H) or ATP was highlighted using circle with green, dotted lines. Chemical structures are from ChemSpider.
Table 3. The effects of different molecules on Txnip mRNA levels. Stimulatory Non‐Stimulatory Molecule Fold ± SD Molecule Fold ± SD NAD+ 4.81±0.27 ** Ethano‐NAD 1.05±0.07 NADH 4.85±0.21 ** NGD 1.03±0.18 NADP+ 4.70±0.28 ** FMN 0.95±0.21 NADPH 4.65±0.35 ** NMN 1.14±0.12 α‐NAD 4.32±0.40 * CTP 0.90±0.14 FAD 4.38±0.32 ** TTP 0.70±0.28 ATP 4.55±0.64 * UTP 0.70±0.42 ATPγS 4.20±0.42 * Bz‐ATP 1.35±0.35 ADP 5.00±0.71 * Tnp‐ATP 1.77±0.37 ADP‐ribose 4.53±0.67 * dATP 1.85±0.49 AMP 4.75±0.36 ** dCTP 1.00±0.14 3’AMP 5.05±0.49 * dTTP 0.65±0.21 2’AMP 4.77±0.66 * dGTP 0.76±0.22 cAMP 4.70±0.99 * dUTP 0.73±0.25 AMPS 4.65±0.21** IDP 0.90±0.14 Adenosine 4.38±0.45 * Adenine 1.84±0.34 HeLa cells were treated with 0.1 mM of different molecules for 4 hrs, and Txnip mRNA level was measured using Real‐Time PCR. Txnip expression was not affected by H2O, DMSO or HEPES buffer (20 mM; pH5.4 or pH8.4) used as solvents. Asterisk: comparison with untreated sample.
41 Chapter 3
When HeLa cells were treated with 0.1 mM of Ethano‐NAD, NGD, FMN,
CTP, UTP, TTP, GTP or IDP for 4 hrs, Txnip mRNA levels were not changed significantly (Table 3). These molecules share structural similarities with NAD(H),
FAD, ATP or ADP; however, their base groups are distinct from adenine. Moreover,
adenine itself was not effective in inducing Txnip expression (Table 3). These results
indicate that a base component of the adenosine group is essential but not sufficient,
which led us to examine the ribose component by using AMP and ATP analogs that differ from AMP or ATP at the ribose moiety.
Figure 11. Structure of ATP, dATP, Bz‐ATP or Tnp‐ATP. Chemical structures are from ChemSpider.
Cyclic‐AMP (cAMP), 3’AMP, and 2’AMP stimulated Txnip expression to a similar degree as 5’AMP; however, ATP analogs with the ribose modified by bulky
groups such as benzoylbenzoyl ATP (Bz‐ATP) and trinitrophenyl ATP (Tnp‐ATP), or with a deoxidized ribose, dATP, showed marginal effects (Table 3, see Figure 11 for
42 Chapter 3 molecular structures). Thus, an adenosine moiety (adenine and ribose) is required
and sufficient for inducing Txnip expression, which tolerates phosphate groups but
not bulky modifications or deoxidization at the ribose (–OH positions).
A 10 B 10 NADP+ Adenosine AMP 8 NADPH 8 NAD+ ADP NADH ATP 6 6
4 4 Txnip mRNA Level Txnip mRNA Level 2 2
0 0 0 1 5 5 5 5 5 0 1 5 5 .5 .25 2. 0 05 2 2 2 1 .0 .0 .1 .6 1.25 0.00 0.00 0.02 0.12 0.62 0 0 0.0250 0 mM mM
Figure 12. Titrations of adenosine‐containing molecules on Txnip expression. HeLa cells were treated with different molecules (with different doses) for 4 hrs, and Txnip mRNA levels were measured using Real‐Time PCR.
3.2.1.4 Adenosine‐containing Molecules Induce Txnip Expression in a Dose‐ dependent Manner
Results in Figure 9 and Table 3 represent observations in cells treated with 0.1 mM of various compounds; I also examined effective dosage ranges of a selective set of adenosine‐containing molecules in inducing Txnip expression.
As shown in Figure 12A, 0.025 mM of NAD(H) was sufficient for maximum induction, and up to 2.5 mM of NAD(H) remained stimulatory; on the other hand,
NADP(H) stimulated Txnip expression at low concentrations, and its effect diminished at 0.625 mM or above. Adenosine, AMP, ADP and ATP exerted their optimal effects on the Txnip expression at 0.125 mM; AMP, but not adenosine, ADP
43 Chapter 3 or ATP, sustained the induction at higher dosages (Figure 12B). Most of these molecules contain phosphate group(s) carrying negative charge(s); for instance, at maximum, ATP and NADP(H) carry four, ADP three, and AMP and NAD(H) two, negative charges. These results suggest that the negative charges around the ribose moiety can affect the efficacies of adenosine‐containing molecules on inducing Txnip expression.
25 - * * + 20 NAD * NADH 15 ATP ADP * AMP 10 ** Adenosine * ** * *** 5 ** * Txnip mRNA Level mRNA Txnip
0 0 1 2 4 8 16 Hr(s)
Figure 13. Time course of adenosine‐containing molecules on Txnip expression. HeLa cells were treated with different molecules (0.1 mM) for various time, Txnip mRNA levels were measured using Real‐Time PCR. Asterisk: comparison with untreated sample.
3.2.1.5 Long Term Effect of Adenosine‐containing Molecules on Txnip Expression
When examining the time course of the induction of Txnip expression, all
tested adenosine‐containing molecules exhibited ~4‐fold stimulatory effects at 4 hrs, but their long term effects varied greatly (Figure 13). At 8 hrs, while NAD+, NADH,
ATP, ADP and AMP exhibited an approximately 8‐fold induction, the effect of adenosine became insignificant. At 16 hrs, NAD+, NADH and ATP dramatically
induced Txnip expression (~20‐fold); however, the effect of ADP was moderate and
44 Chapter 3 those of AMP and adenosine were marginal. This indicates that, although all
stimulatory, the above adenosine‐containing compounds are able to exhibit different
kinetics on Txnip expression.
5 A ** B ** * ** ** 4 ** ** **** 3
2
Promoter Activity 1
0 + + + - P P P e - P AD ine in NGD F FMNATPTT CT GT UTP n NADNADH-NAD os NAD α NADNADPH n Ade denosine Ade A ethano-NAD
Figure 14. Adenosine‐containing molecules induce Txnip promoter activity. U2OS cells transfected with Txnip promoter‐luciferase reporters were treated with 0.2 mM of different molecules for 16 hrs, or left untreated (‐). The promoter activities were measured using luciferase assays. (B) Medium was replaced every 4‐hrs with fresh medium containing adenosine or NAD+ (n=3). Asterisk: comparison with untreated sample.
3.2.1.6 Adenosine‐containing Molecules Induce Txnip Expression at the
Transcriptional Level
To test whether Txnip up‐regulation was at the transcription level, the effect of adenosine‐containing molecules on the expression of an ectopic Txnip promoter
was examined. Txnip promoter (269 base pairs upstream of the transcriptional start site) was obtained by PCR from genomic DNA and fused to a luciferase reporter, and the response of this promoter to different adenosine‐containing molecules was tested.
As shown in Figure 14A, with the exception of adenosine, all adenosine‐containing
45 Chapter 3 molecules induced the reporter activity. In this reporter assay, 16 hrs were used to manifest the luciferase activities. Thus, the marginal effect of adenosine on the Txnip promoter activity may be explained by the uptake and metabolism of extracellular adenosine by cells over 16 hrs. Indeed, cells transfected with the reporter and treated
with fresh adenosine‐containing medium at 4 hr intervals exhibited a significant induction of the promoter activity (Figure 14B). Other molecules lacking an intact adenosine group, e.g., NGD, FMN, TTP and adenine, failed to stimulate the transcription from the Txnip promoter (Figure 14A). This observation is in agreement with the earlier specified requirement for an intact adenosine moiety (Table 3).
12 - 10 NAD+ 8 ATP Adenosine 6 4 2 Txnip mRNA Level 0 - Actinomycin D
Figure 15. Adenosine‐containing molecules did not induce Txnip expression in the presence of actinomycin D. U2OS cells were treated with 0.1 mM of NAD+, ATP or adenosine for 4 hrs or left untreated. One group of cells was pre‐treated with 2 μg/ml of actinomycin D for 10 min. Txnip mRNA levels were measured by Real‐ Time PCR.
To further confirm that Txnip expression induction by adenosine‐containing molecules is at the transcription level but not other levels such as mRNA stability, the impact of actinomycin D, which inhibits transcription of all cellular genes, on the induction of Txnip expression by NAD+, ATP or adenosine was examined. Txnip expression in U2OS cells was no longer responsive to stimulation by NAD+, ATP or
46 Chapter 3 adenosine in the presence of actinomycin D (Figure 15). In control cells, actinomycin
D down‐regulated (~2‐fold) the basal Txnip mRNA expression (defined as expression level in the absence of adenosine‐containing molecules; Figure 15). Therefore, both the basal and the adenosine‐containing molecule‐stimulated Txnip expression are regulated at the transcriptional level.
3.2.1.7 Adenosine‐containing Molecules Induce Txnip Expression Is Mediated by an Earlier Defined ChoRE
The transcription of Txnip is induced by glucose, and this induction is
mediated by an earlier defined ChoRE (see Figure 6) and its associated transcription factors MondoA and MLX (Minn et al., 2005b; Stoltzman et al., 2008). To test if the target for adenosine‐containing molecule‐evoked signaling pathway is this ChoRE, I examined the effects of selective molecules on wild type or ChoRE mutant Txnip promoters. The wild‐type Txnip promoter, and a Txnip promoter with mutations at the FOXO‐binding site, was activated by both NAD+ and ATP, whereas the ChoRE
mutant Txnip promoter was not sensitive to either NAD+ or ATP (Figure 16). Thus,
the earlier‐defined ChoRE on Txnip promoter is also important for mediating the induction of Txnip expression by adenosine‐containing molecules. The ChoRE, as the name suggests, is a critical regulatory element that mediates the stimulatory effects of glucose on several genes involved in metabolic control (Ishii et al., 2004;
Minn et al., 2005b); I hence reasoned that glucose may play a role in the Txnip expression induction by adenosine‐containing molecules, which may be mediated by the ChoRE and associated transcription factors.
47 Chapter 3
5 ** - 4 ** + * NAD 3 ** ATP 2
1 Promoter Activity Promoter 0
WT RE OXO mF mCho
Figure 16. NAD+ or ATP did not induce activity of ChoRE‐mutated Txnip promoters. U2OS cells transfected with wild‐type (WT), FOXO binding site mutant (mFOXO) or ChoRE mutant (mChoRE) Txnip promoters were treated with 0.2 mM of NAD+ or ATP, or left untreated. At 16 hrs, the promoter activities were measured. Asterisk: comparison with untreated sample.
Figure 17. NAD+ or ATP could not induce Txnip promoter activity in the presence of dominant negative MLX. HA‐tagged versions of wild type MLX (MLX) and dominant negative MLX (DN) were ectopically expressed (see immuno‐blots) in a co‐ transfection assay with the wild‐type (WT) or ChoRE mutant (ChoREmut) Txnip promoter‐luciferase construct. Transfected cells were either untreated or treated with 0.2 mM of NAD+ or ATP; the effects of ectopic MLX on the Txnip promoter‐ driven reporters were analyzed at 16 hrs. Asterisk: comparison with untreated sample; diamond: comparison with empty vector transfected samples.
48 Chapter 3
3.2.1.8 The MLX/MondoA Complex Mediates the Induction of Txnip Expression by Adenosine‐containing Molecules
During glucose‐induced gene transcription, ChoRE elements can recruit transcription factor complexes, e.g., a tetramer of two molecules each of ChoRE binding protein (ChREBP) and MLX (Yamashita et al., 2001; Ishii et al., 2004;
Stoeckman et al., 2004; Ma et al., 2005; Billin and Ayer, 2006; Ma et al., 2006). MLX is also able to form a complex with MondoA, a ChREBP homolog (Stoltzman et al., 2008;
Sans et al., 2006). Thus, MondoA or ChREBP in association with MLX might be critical in mediating the induction of Txnip expression by adenosine‐containing molecules.
To test this out, I first used a dominant negative MLX mutant (DN), which can block glucose‐induced Txnip expression (Ma et al., 2005). When ectopically expressed in U2OS cells, this mutant diminished the effects of NAD+ or ATP on
Txnip promoter activation (Figure 17). Therefore, the induction of Txnip expression by adenosine‐containing molecules requires MLX.
When MLX, ChREBP or MondoA were ectopically‐expressed, basal Txnip promoter activity was not significantly induced by MLX or ChREBP, but MondoA induced a >10‐fold increase (Figure 18). When co‐expressed with MLX, MondoA and
ChREBP increased Txnip promoter activity by >30‐fold and ~4‐fold, respectively; the promoter responded to NAD+ in all cases (Figure 18). When co‐expressed with the
dominant negative MLX, neither MondoA nor ChREBP supported the basal or
NAD+‐induced activities of the wild‐type Txnip promoter; moreover, the activity of
the ChoRE‐mutant Txnip promoter (ChoREmut) was not affected by the over‐
expression of MondoA or co‐expression of MLX with MondoA or ChREBP (Figure
49 Chapter 3
18).
I next employed an RNAi approach to study the roles of MLX, MondoA and
ChREBP. MLX and MondoA mRNA levels were dramatically repressed in U2OS cells treated with the MLX‐ or MondoA‐specific siRNA, respectively. In cells with silenced expression of MLX or MondoA, a strong reduction in the basal and NAD+‐ induced activities of the wild‐type Txnip promoter was observed; however, silencing the MLX or MondoA expression did not impede the ChoRE mutant Txnip promoter activity (Figure 19A).
Figure 18. Effects of ectopic MLX and/or Mondo on Txnip promoter activity. HA‐ tagged wild type MLX (MLX), dominant negative MLX (DN), MondoA or ChREBP were transfected into cells together with wild‐type or ChoRE mutant Txnip promoter‐luciferase construct. Following transfection, cells were treated with NAD+ or left untreated; at 16 hrs, the promoter activities were measured. The insets are 5‐ fold augmentation of the images below. The expression levels of ectopic proteins were detected by HA‐antibody, and GAPDH level was used as internal control. Asterisk: comparison between NAD+ treated and untreated samples; diamond: comparison between basal activities (without NAD+).
A B
50 Chapter 3
Figure 19. Effects of siRNAs against MLX or MondoA on Txnip expression. (A) MLX or MondoA deficiencies (as shown by RT‐PCR) down‐regulated the wild‐type but not the ChoRE mutant Txnip promoter. NAD+ was unable to induce Txnip promoter activity when siRNAs against MLX or MondoA were present. R, random (control) siRNA. (B) Ectopic MondoAr, but not ChREBP, rescued the Txnip promoter activity in MondoA‐deficient cells. HA‐tagged MondoA (siRNA‐resistant copy, MondoAr) or ChREBP was co‐transfected with wild type Txnip promoter‐luciferase construct and siRNAs against MondoA. Ectopic MondoAr, but not ChREBP, restored the basal Txnip promoter activity, and the promoter activity was further boosted by NAD+ treatment. Ran, random siRNA. The expression levels of ectopic proteins were detected by HA‐antibody, and GAPDH level was used as internal control. Asterisk: comparison with untreated sample. Diamond: comparison with cells treated with random siRNA (A) or between samples connected by lines (B); n=4 for (A).
Given that ChREBP expression in U2OS cells is extremely low (not shown), and that ectopic ChREBP exhibited a much lower efficacy than did MondoA in a functional assay (Figure 18), it is impractical to employ RNAi to study the effects of further reduced ChREBP expression. I generated an RNAi‐resistant MondoA
(MondoAr) with nucleotide changes at the siRNA target site while retaining a normal peptide sequence, and analyzed the functions of ectopic MondoAr and ChREBP in
U2OS cells in which the expression of endogenous MondoA was silenced by RNAi.
As seen in Figure 19B, the ectopic MondoAr rescued, boosted, and supported NAD+
51 Chapter 3 in stimulating Txnip promoter activation. In contrast, even with a higher ectopic expression than MondoAr, ChREBP failed to rescue the Txnip promoter activity
(Figure 19B). The above results suggest that the MondoA/MLX complex plays a
major role for the ChoRE‐dependent induction of Txnip expression by adenosine‐
containing molecules.
3.2.1.9 Adenosine‐containing Molecules Facilitate MondoA Nuclear Translocation
MondoA and MLX are largely cytoplasmic and likely associated with the
outer mitochondrial membrane (Sans et al., 2006; Stoltzman et al., 2008); however, to
stimulate Txnip expression under high glucose concentration, they need to enter the
nucleus and bind to the ChoRE on Txnip promoter to exert stimulatory effects. As
ChoRE and MondoA/MLX were shown to be essential for the induction of Txnip
expression by adenosine‐containing molecules (see above), it is possible that these
molecules can modulate the cellular localization of MondoA/MLX.
HA‐tagged MondoA was co‐expressed with Myc‐tagged MLX in L6 cells, and
the localization of HA‐MondoA was determined by immuno‐staining (anti‐HA) and confocal microscopy. Confirming a previous study (Stoltzman et al., 2008), MondoA was localized in cytosol when cells were cultured in glucose‐free medium; however when cells were treated with 2‐deoxyglucose (2DG, which can sustain a robust glucose signaling [see section 3.2.2]), majority cells exhibited nuclear localization of
MondoA (Figure 20 and 21).
52 Chapter 3
Figure 20. Localization of HA‐MondoAin L6 cells under control or 2DG treatment. HA‐MondoA and Myc‐MLX were co‐transfected in L6 cells; following transfection, cells were incubated in glucose‐free DMEM for 24 hrs. Control cells (untreated) and treated cells (10 mM 2DG for 4 hrs) were fixed and the localization of HA‐MondoA was determined by immunocytometry using anti‐HA antibodies (green). Nuclear regions were visualized using DAPI (blue).
cyt nuc/cyt nuc 100 ** 80 * * 60
40 % of% cells 20
0 - + + D A Glc 2DG N NAD lc G Figure 21. MondoA nuclear translocation was facilitated by adenosine‐containing molecules. Localization (cytoplasmic [cyt], nuclear [nuc] or heterogeneous [cyt/nuc]) of HA‐MondoA in control cells or cells treated with glucose (Glc), 2DG or NAD+ for 4 hrs was determined by immunocytometry. The concentrations of the compounds when used were: glucose and 2DG, 10 mM; NAD+, 0.1 mM. Asterisk: compare the ratio of nuclear staining positive cells between cells under different treatments (as indicated by dotted lines); n=3.
53 Chapter 3
The localization of HA‐MondoA in L6 cells treated with different chemicals
were then analyzed; for each treatment, 100 HA‐MondoA expressing cells were randomly picked, and these cells were classified into three groups according to the distribution of HA‐MondoA: cytosolic (cyt), nuclear (nuc) or even distribution
(cyt/nuc). As shown in Figure 21, NAD+ or glucose alone exerted a marginal effect on the nuclear redistribution of HA‐MondoA; however, in the presence of both glucose and NAD+, significantly more cells exhibited nuclear staining of HA‐
MondoA. This suggests that adenosine‐containing molecules can facilitate MondoA nuclear Translocation in a glucose‐dependent manner.
3.2.1.10 Glucose Is Required for the Induction of Txnip Expression by Adenosine‐ containing Molecules
ChoRE and associated MondoA/MLX are involved in the induction of Txnip expression by adenosine‐containing molecules, suggesting that this induction might be glucose‐dependent. Indeed, when HeLa cells were cultured in glucose‐free medium supplemented with pyruvate as a carbon source, Txnip expression was not stimulated by NAD+ or ATP; however, if cells were freshly‐fed with glucose, Txnip mRNA levels were dramatically elevated and further induced by NAD+ or ATP
(Figure 22A). A similar induction pattern was also observed when glucose was present throughout (Figure 22A). Similarly, Txnip promoter activity was not stimulated by NAD+ or ATP when cells were cultured in glucose free medium, and the promoter activity was induced by NAD+ or ATP when glucose was freshly added; the ChoRE mutant Txnip promoter was not sensitive to NAD+ or ATP (Figure 22B).
54 Chapter 3
A 6 - 5 evel NAD+ 4 ATP 3 0.1 - 2
1 L Txnip mRNA 0 w/o glucose glucose glucose fresh throughout B 6 - 5 + NAD 4 ATP Glc 3 Glc+NAD+ 2 Glc+ATP 1 Promoter Activity 0 WT ChoREmut C 7 6 - NAD+ 5 ATP 4 3 0.15- 2 1 Txnip Level mRNA 0 - in in toB dz y CytoB C hloret hlori P in/ P ret o hl P
Figure 22. Glucose‐dependent induction of Txnip expression by NAD+ or ATP. (A) Txnip expression in HeLa cells without or with glucose (5 mM), NAD+ or ATP (0.1 mM) for 4 hrs. (B) Induction of Txnip promoter activity by NAD+ or ATP requires glucose; HeLa cells transfected with Txnip promoter (WT or ChoRE mutated) were incubated in glucose free medium, and the effect of glucose (5 mM) and/or NAD+/ATP (0.2 mM) was tested (16 hrs). (C) Txnip mRNA expression was not induced by NAD+ or ATP in the presence of glucose transporter inhibitors; HeLa cells were pretreated with 100 μM of phloretin or phloridzin or 10 μM of cytochalasin B (half dosages for two drugs used in combination) for 10 min, and then 0.1 mM of NAD+ or ATP was added for 4 hrs. The insets in (A) and (C) are 20‐ and 15‐fold enlargements, respectively, of the images below.
55 Chapter 3
The involvement of the cellular glucose uptake system in Txnip expression
was also tested. Indeed, separately or in combination, phloretin and cytochalasin B, inhibitors of the facilitated glucose transporters (Glut), dramatically repressed the basal and NAD+‐ or ATP‐stimulated Txnip expression; however, phloridzin, an
inhibitor of the tissue‐restricted Na+/glucose symporters (SGLT), did not exert
corresponding effects (Figure 22C). The presence of glucose and a functional glucose
transport system therefore appears to be essential and a prerequisite for the Txnip
expression that is inducible by adenosine‐containing molecules.
100 * - 80 * * NAD+ 60
40 y * y y 20 Txnip mRNA Level mRNA Txnip 0 0 0.2 1 5 10 25 Glc (mM)
Figure 23. NAD+ induced Txnip expression in a wide titration of glucose. Txnip mRNA levels in HeLa cells treated with a glucose (Glc) titration with or without 0.1 mM of NAD+. Asterisk: comparisons between NAD+ treated and untreated samples; Dot: comparisons between glucose treated and control (without glucose) samples.
3.2.1.11 Glucose Induced Txnip Expression Is Amplified by Adenosine‐containing
Molecules
The requirement of glucose (Figure 22) and mobilization of glucose‐sensitive transcription factors (Figure 21) suggests that the adenosine‐containing molecules enhance Txnip expression by amplifying the glucose signaling pathway. Indeed, over a wide glucose titration, Txnip mRNA level was further stimulated by treating cells with NAD+ (Figure 23). Therefore, enhanced nuclear accumulation of HA‐
56 Chapter 3
MondoA by NAD+ (Figure 21) might explain the general functionality of adenosine‐
containing molecules in inducing the Txnip expression in a glucose‐dependent
manner.
3.2.1.12 Potential Plasma Membrane Target(s) of Adenosine‐containing Molecules
Cells can respond to many extracellular/environmental factors, and these
factors include many adenosine‐containing molecules. Some adenosine‐containing
molecules can convert to others; hence it is hard to dissect the exact functions and
regulatory mechanisms of these molecules. For example, ATP can be converted to
ADP, AMP or adenosine by extracellular nucleotidase (Zimmermann, 2001). ATP,
ADP and AMP are charged and usually impermeable to plasma membrane; however,
these molecules can function as extracellular signaling molecules to activate some
intracellular signaling pathways. On the other hand, adenosine can enter cells and
modulate some cellular activities in a more direct fashion.
- ARL 67156 5
4
3
2
1 Txnip mRNA Level Txnip mRNA 0 + - P P ATP M NAD AD A
Adenosine
Figure 24. The induction of Txnip expression was not abolished by ARL 67156. HeLa cells were pre‐treated with 0.1 mM of ARL 67156 for 10 min, and then adenosine‐containing molecules (0.1 mM) were added for another 4 hrs. Txnip mRNA levels were measured using Real‐Time PCR.
57 Chapter 3
The induction of Txnip expression by adenosine‐containing molecules,
especially in the long‐term, requires these molecules to remain extracellular, and my
data in Figures 13 and 14 are in strong support of this view. Hence, adenosine transport/internalization may not be critical in inducing the Txnip expression.
Adenosine can be taken‐up by cells via adenosine transporters; however, majority adenosine‐containing molecules are charged (e.g., NAD[H], ATP) and impermeable to plasma membrane. To rule out the possibility that charged adenosine‐containing molecules act on Txnip expression via degradation to adenosine that then enter cells to exert stimulatory function, I tested the effects of a number of ATPase inhibitors on
Txnip expression in the presence of different adenosine‐containing molecules, and
Figure 24 shows an example: an ecto‐ATPase inhibitor, ARL 67156, affected neither the basal Txnip expression, nor the induction of Txnip expression by NAD+, ATP,
ADP, AMP and adenosine. Thus the conversion of adenosine‐containing molecules
to adenosine is not necessary for the induction of Txnip expression. This is in line
with the fact that non‐hydrolysable ATP‐γ‐S and AMP‐α‐S were as potent as other
adenosine containing molecules in stimulating Txnip expression (Table 3).
It is possible that the engagement of adenosine‐containing molecules with
their plasma membrane targets can trigger certain signaling pathways that ultimately lead to redistribution of MondoA/MLX and enhanced Txnip expression. Known
membrane targets that have potential affinity for adenosine‐containing molecules
include the purinergic receptors (P1, P2X or P2Y) (Burnstock, 2008) and adenosine transporters (Thorn and Jarvis, 1996). Thus I tested the involvement of these plasma
membrane proteins in inducting Txnip expression.
58 Chapter 3
3.2.1.12.1 Purinergic Receptors Are Not Required for the Induction of Txnip
Expression
Purinergic receptors are a family of plasma membrane molecules belonging
to GPCR (P1 or P2Y) or ligand‐gated ion channels (P2X) ; these receptors have important functions in immune or neuronal systems, in essence acting as receptors for purines, or certain nucleosides/nucleotides (Burnstock, 2008).
Most of purinergic receptors do not discriminate against the non‐adenine nucleotides such as UTP, CTP, GTP and IDP; however, these molecules did not stimulate the Txnip expression (Table 3). Moreover, an inhibitor of the P1 receptor
(caffeine) and inhibitors of the P2 receptor (RB‐2 or Suramin) failed to impede the induction of Txnip expression by NAD+ or ATP (Figure 25). Thus, purinergic receptors could not be the plasma membrane target of adenosine‐containing molecules in inducing Txnip expression.
7 - + ATP 6 NAD 5 * ** * ** * ** * 4 3 2 1 Txnip mRNA Level mRNA Txnip 0 - Caffeine Rb-2 Suramin
Figure 25. Inhibitors for purinergic receptors did not inhibit the induction of Txnip expression by NAD+ or ATP. HeLa cells were pre‐treated with 100 μM of suramin, Rb‐2 or caffeine for 10 min, then NAD+ or ATP (0.1 mM) were added for 4 hrs. Txnip mRNA levels were measured using Real‐Time PCR. Asterisk: compare with untreated sample.
3.2.1.12.2 Adenosine‐containing Molecules May Target Adenosine Transporters
Interestingly, an adenosine transporter inhibitor, nitrobenzylthioinosine
59 Chapter 3
(NBTI), repressed basal Txnip mRNA expression and abolished the stimulatory effect of adenosine‐containing molecules in a dose‐dependent manner (Figure 27A). This
inhibitor is an adenosine analog (Figure 26) and might be able to, with high potency,
compete with adenosine‐containing molecules for adenosine transporters, or other yet unidentified plasma membrane targets, hence acting as an antagonist.
Dipyridamole and Dilazep are two adenosine transporter inhibitors structurally unrelated to adenosine (Figure 26). Dipyridamole did not affect basal
Txnip mRNA levels but impeded the induction by NAD+ (Figure 27B). Dilazep also
impeded the induction of Txnip expression by NAD+, although the basal level of
Txnip expression was also repressed (Figure 27C). These results suggest that the adenosine transporter could be the primary target of the adenosine‐containing molecules in inducing the Txnip expression.
Figure 26. Structures of Adenosine, NBTI, Dipyridamole or Dilazep. Chemical structure are from ChemSpider.
60 Chapter 3
A 5 - *** 4 NAD+
Le ATP 3 2 * 1 y y y Txnip mRNATxnip vel 0 0 5 10 20 40 80 NBTI (μM) B 6 * vel 5 - NAD+ Le 4
3 2 * 1 Txnip mRNA Txnip 0 0 5 08 16 2.5 10 0. 0. 0.31 0.63 1.25 Dipyridamole (μM) C 6 * 5 - +
Leve NAD 4 3 * * 2 1 Txnip mRNA mRNA Txnip l 0 0 0.6 2.5 10 40 160 Dilazep (μM)
Figure 27. The effect of NAD+ or ATP on Txnip expression was blocked by inhibitors of adenosine transporters. HeLa cells were pre‐treated with different inhibitors of adenosine transporters (with different concentrations) for 10 min, and then NAD+ or ATP (0.1 mM) was added for 4 hrs. The Txnip mRNA expression levels were measured using Real‐Time PCR. Asterisk: comparison with untreated (NAD+ or ATP) sample; Diamond: comparison between NBTI (A), dipyridamole (B) or Dilazep (C) treated samples with control samples in the presence of NAD+ or ATP.
61 Chapter 3
3.2.1.13 Signaling Pathway(s) Evoked by Adenosine‐containing Molecules for
Regulating Txnip Expression
3.2.1.13.1 The Induction of Txnip Expression Requires Intracellular Ca2+
Extracellular nucleotides and nucleosides are known to affect intracellular
Ca2+ levels (Bruzzone et al., 2006; Hanley et al., 2004), prompting me to explore whether cellular Ca2+ homeostasis is critical for the induction of Txnip expression by adenosine‐containing molecules. To this end, I found that in the presence of BAPTA‐
AM, a cell‐permeable calcium chelator, the induction of Txnip expression by NAD(H) or ATP was abolished; however, an extracellular calcium chelator, EGTA, did not show any effect (Figure 28). Therefore, the intracellular Ca2+ signaling/homeostasis
might be required for the induction of Txnip expression by adenosine‐containing
molecules.
6 - 5 * * * ** ATP ** ** + 4 NAD NADH
3 2 1 Txnip mRNA Level Txnip mRNA 0 - EGTA BAPTA-AM
Figure 28. Effect of Ca2+ chelators on Txnip expression. HeLa cells were pre‐treated with BAPTA‐AM (10 μM) or EGTA (0.5 mM) for 10 min, and then with 0.1 mM of NAD(H) or ATP for 4 hrs. Txnip mRNA levels were measured with Real‐Time PCR. Asterisk: comparison with untreated sample; Diamond: comparison between BAMPTA‐AM treated samples with control samples in the presence of NAD(H) or ATP.
62 Chapter 3
A 5 4
3
2
1 Txnip mRNA Level 0 - + + + + + - + + + + + NAD+ 0 0 5 0 0 10 20 40 25 50 100 12.5 μM SQ22536 μM 9-CP-AdeM
B 5 4
3
2
1 Txnip mRNA Level 0 - + - + - + - + - + NAD+
- H-89 rp PKI KT
C 2.5 2.0
1.5
1.0
0.5 Txnip mRNA Level Txnip mRNA 0.0 0 1 2 4 0 1 2 4 Hrs 8-Br-cAMP Forskorlin
Figure 29. cAMP signaling pathway did not mediate the induction of Txnip expression by NAD+. (A and B) HeLa cells were pre‐treated with different inhibitors for 10 min (H‐89: 10 μM; rp: rp‐8‐Br‐cAMP, 100 μM; my‐PKI: myristoylated PKI [14‐ 22] amide, 100 μM; KT: KT5720, 5 μM; 9‐CP‐AdeM: 9‐Cyclopentyladenine mesylate), and with 0.1 mM NAD+ for 4 hrs. (C), HeLa cells were treated with 100 μM of 8‐Br‐ cAMP or forskorlin for 0‐4 hr(s). Txnip mRNA levels were measured using Real‐ Time PCR.
63 Chapter 3
3.2.1.13.2 The Induction of Txnip Expression Does Not Require cAMP
It was reported that adenosine, ATP, ADP, AMP and ATPγS, but not dATP,
UTP and GTP could up‐regulate the cellular cAMP levels (Matsuoka et al., 1995). The effects of these molecules on cellular cAMP levels correlated with their effects on
Txnip expression (Table 3). Thus, the cellular cAMP levels and related downstream pathways might be involved in the induction of Txnip expression. However, cells in which the activity of adenylyl cyclase (AC, for production of cAMP) was inhibited by
SQ22536 or 9‐Cyclopentyladenine mesylate exhibited a normal Txnip expression pattern (Figure 29A). When Protein Kinase A (cAMP activated kinase) was inhibited by H‐89, rp‐8‐Br‐cAMP or myristoylated PKI (14‐22) amide, the induction of Txnip expression by NAD+ was not affected (Figure 29B). Another PKA inhibitor, KT5720, induced Txnip basal expression slightly and abolished the effect of NAD+ (Figure
29B); this could be due to a non‐specific effect of this compound. Moreover, as
shown in Figure 29C, a cell‐permeable cAMP analog, 8‐Br‐cAMP, failed to induce
Txnip expression significantly; and Forskolin, which can induce intracellular cAMP
production, dramatically reduced Txnip expression. The latter might be due to the
strong inhibitory role of Forskolin on glucose transport (Sergeant and Kim, 1985).
Taken together, these results exclude an involvement of intracellular cAMP levels
and related downstream pathways in the induction of Txnip expression.
3.2.1.13.3 The Involvement of MAPK in the Induction of Txnip Expression
It has been shown that the induction of Txnip expression by glucose is mediated by p38/MAPK (Schulze et al., 2004), and glucose can induce ERK activity
(Bandyopadhyay et al., 2000). In this study, I have also tested different MAPK inhibitors on the stimulatory effect of adenosine‐containing molecules on Txnip
64 Chapter 3 expression. Although some controversies exist, it is known that PD169316 and
SB202190 are p38 MAPK inhibitors, PD98059 and U0126 are ERK1/2 inhibitors, and
SP600125 is a JNK inhibitor. As shown in Figure 30, in U2OS cells, both PD169316
and PD98059 had abolished the stimulatory effect of NAD+ and ATP on Txnip expression; U0126 had repressed Txnip expression and retarded the Txnip induction by NAD+ and ATP; SB202190 and SP600125 had stimulated basal Txnip expression
dramatically. Based on these results, a solid conclusion on the effect of MAPK on
Txnip expression is unable to be made; activation of p38/MAPK and/or ERK1/2 could
probably mediate the effects of adenosine‐containing molecules.
In HeLa cells (Figure 31), PD169316 fully abolished the effect of NAD+ or ATP on Txnip expression; however, PD98059 and U0126 revealed marginal effects, this was different from results obtained in U2OS cells (Figure 30). Thus, the involvement of MAP kinases in Txnip expression needs further investigations, and their effects might be cell type dependent.
18 16 14 - NAD+ ATP 12 10 8 6 4 Txnip mRNA level Txnip 2 0 Control PD169316 PD98059 U0126 SB202190 SP600125
Figure 30. The effect of MAPK inhibitors on Txnip expression and its response to NAD+ or ATP in U2OS cells. Cells were pre‐treated with 10 μM of MAPK inhibitors for 10 min, and with 0.1 mM NAD+ or ATP for another 4 hrs. Txnip mRNA levels were measured by Real‐Time PCR.
65 Chapter 3
8 7 - NAD+ ATP 6 5 4 3 2 Txnip mRNA Txnip level 1 0 Control PD169316 PD98059 U0126 SB202190 SP600125
Figure 31. The effect of MAPK inhibitors on Txnip expression and its response to NAD+ or ATP in HeLa cells. Same procedure was used as shown in Figure 30.
3.2.1.13.4 Non‐involvement of AMPK in the Induction of Txnip Expression
The 5’ AMP‐activated protein kinase, AMPK, is a master metabolic switch regulating glucose uptake, fatty acid utilization and biogenesis of mitochondria, and its activity is regulated by intracellular AMP/ATP ratios (McGee and Hargreaves,
2008). Recently, it has been shown that uptake of adenosine can activate AMPK (da
Silva et al., 2006; Aymerich et al., 2006). In this study, I investigated a potential role of
AMPK in the induction of Txnip expression by adenosine‐containing molecules. An
AMPK activator, 5‐aminoimidazole‐4‐carboxamide ribonucleoside (AICAR), was able to stimulate Txnip expression (Figure 32B). AICAR is slightly different from adenosine (Figure 32A); the dose response to AICAR (Figure 32B) is very similar to
that of adenosine (Figure 12B), suggesting that either AMPK is in the pathway
evoked by adenosine‐containing molecules, or AICAR, as a structural analog, can be
engaged to membrane targets (see above) as are adenosine‐containing molecules to ultimately stimulate Txnip expression. The fact that 9‐β‐D arabinofuranoside (AraA), an AMPK inhibitor, was unable to impede the induction of Txnip expression by
NAD+, supports the latter scenario and suggests that AMPK is unlikely involved in
66 Chapter 3 the pathway initiated by adenosine‐containing molecules to induce Txnip expression.
A B 5 C 6 l l - 4 5 NAD+ 4 3 3 2 2 1 1 Txnip mRNA Leve mRNA Txnip Txnip mRNA Leve mRNA Txnip 0 0 - 0.5 1 0 5 .5 1 2 0 625 0. 0 0.125 AraA (mM) 0. mM
Figure 32. AMPK was not involved in the induction of Txnip expression by adenosine‐containing molecules. (A) Structures of adenosine or AICAR, chemical structures are from ChemSpider. (B) The effect of AICAR on Txnip expression. (C) The effect of AraA on the induction of Txnip expression by NAD+.
3.2.1.14 Adenosine‐containing Molecules Repress Thioredoxin Activity and
Glucose Transport
Txnip can inhibit thioredoxin activity (Nishiyama et al., 1999; Yamanaka et al.,
2000; Junn et al., 2000). In this study, I had tested whether cellular thioredoxin activity is affected in cells treated with adenosine‐containing molecules. As shown in
Figure 33, NAD+ can effectively inhibit the activity of thioredoxin after 8 hrs of
treatment; short term treatment (4 hrs) was not effective.
Txnip has a critical role in regulating energy homeostasis and can inhibit the glucose uptake system (Parikh et al., 2007; Yoshioka et al., 2007). I therefore examined the effect of adenosine‐containing molecules on the glucose uptake. As seen in Figure 34, 8 or 16 hr treatments with NAD+ significantly repressed glucose
67 Chapter 3 uptake. Shorter treatments failed to inhibit glucose uptake (Figure 34; similar to thioredoxin activity [Figure 33]). Thus, in order to be efficacious, Txnip protein levels may need to reach a certain threshold.
1 A B 120 0.8 100 NS 0.6 80 * 0.4 60
OD (412 nm) 0.2 40 20 0
0 10203040 Thioredoxin Activity (%) 0 0 4 8 thioredoxin (ng) Hour(s)
Figure 33. Adenosine‐containing Molecules Repressed Thioredoxin Activity. (A) Standard curve (using commercial thioredoxin) for thioredoxin activity measurement. (B) HeLa cells were treated with 0.1 mM NAD+ for 0, 4 or 8 hr(s), the thioredoxin activities were measured; asterisk: compared with untreated sample; n=3. NS stands for “not significant”.
120 100 * 80 ** 60 40 20 Glucose Uptake (%) Uptake Glucose 0 0 2 4 8 16 Hour(s)
Figure 34. Adenosine‐containing Molecules Inhibited Glucose Transport. HeLa cells were treated with 0.1 mM of NAD+ for 0‐16 hr(s), and the glucose (3H‐2DG) transport were measured. Asterisk: compared with untreated sample; n=4.
The role of NAD+ or ATP on thioredoxin activity or glucose uptake is most likely mediated by up‐regulated Txnip protein level as a result of Txnip mRNA
68 Chapter 3 expression that is induced by adenosine‐containing molecules, and suggests that the induction of Txnip expression by these molecules has important physiological functions in regulating cellular redox state and metabolism.
3.2.1.15 Adenosine‐containing Molecules Affect Cell Cycle Progression
Txnip plays an important role in cell proliferation, apoptosis and the development of cancer (reviewed in Kim et al., 2007). In this study, the role of adenosine‐containing molecules in cell cycle events was also investigated. As shown in Figure 35, HeLa cells treated with NAD(H) for a short term (4 hrs) maintained a normal ratios of G1‐, S‐, or G2‐phase cells; however, when cells were exposed to
NAD(H) for 24 hrs, S‐phase cells were dramatically increased. It is possible that
some cell cycle checkpoints are activated by NAD(H) treatment, and cells, although are able to enter, are unable to progress through, S‐phase. This can be characterized as an intra‐S‐phase arrest.
Figure 35. NAD(H) treatment repressed cell cycle progression. HeLa cells were treated with 0.5 mM of NAD(H) for 4 or 24 hrs and subjected to BrdU‐FCAS analyses. Top panel, a dot‐plot of BrdU signals against propidium iodide (PI) signals; numbers on top indicate the ratio of S‐phase cells (cells within boxes). Bottom panel, a histogram presentation of PI staining.
69 Chapter 3
Several cell cycle regulatory proteins (e.g., p27kip1 and p16) can be activated by
Txnip (see section 1.4.2). I also checked the protein levels of several cell cycle
regulators in cells treated with NAD(H) for various time, as shown in Figure 36, the protein level of p21cip1 was robustly up‐regulated when cells were treated for >16 hrs; however, the protein levels of p27kip1 and p16 was not significantly modulated by
NAD(H) treatment (not shown).
Figure 36. NAD(H) treatments elevated p21cip1 expression level. HeLa cells were treated with 0.5 mM of NAD(H) for 0‐24 hrs. GAPDH and p21cip1 protein levels were measured using immuno‐blots.
Txnip has been shown to induce cell arrest at G0/G1 or G1/S transition (Han et al., 2003; Nishinaka et al., 2004b), this is very different from NAD(H) induced intra‐S‐
phase arrest. Thus, whether Txnip is the mediator in the process of NAD(H) induced
cell cycle arrest requires further resolutions; testing the response of Txnip‐deficient
cells to adenosine‐containing molecules might be helpful to clarify this issue.
3.2.2 Effects of Glucose Analogs on Txnip Expression
3.2.2.1 Effects of Selected Monosaccharides and Disaccharides on Txnip
Expression
Txnip expression is induced by glucose, and this induction is mediated by
MondoA and MLX that can sense intracellular glucose‐6‐phosphate (G6P) levels
(Stoltzman et al., 2008); in this model, glucose is required to be taken up by cells and phosphorylated to G6P by hexokinases. A glucose analog, 2‐deoxy‐glucose (2DG), is
70 Chapter 3 able to enter the cells and be phosphorylated to 2DG6P by hexokinases; however, the
2DG6P molecule cannot further participate in glycolysis thus can accumulate in cells
(Figure 37). This can dramatically induce the nuclear translocation of MondoA hence
Txnip expression (Stoltzman et al., 2008; also see Figures 20, 21 and 38); however,
Txnip expression can also be induced by a glucose analog 3OMG (Minn et al., 2006), which can enter cells but cannot be used as a substrate for hexokinases (Figure 37).
Thus, an alternative (G6P independent) pathway may also exist that additionally accounts for the induction of Txnip expression by glucose.
Figure 37. Structures of glucose analogs and their fate in cellular metabolism. Glcuose can be taken up by cells, phosphorylated by Hexokinase to G6P, G6P in turn can participate in downstream glycolysis. 2DG can be taken up by cells and phosphorylated, but the phosphorylated 2DG cannot be consumed, thus accumulates in cells. 3OMG can enter the cells, but cannot be phosphylated, thus accumulates in cells. Disaccharides cannot be taken up by cells. Structures of glucose, two analogs and a representative disaccharide (maltose) are also shown.
In this study, I have tested the effects of glucose, 2DG and 3OMG on Txnip expression, and the effects of adenosine‐containing molecules on Txnip expression in the presence of these glucose analogs in conjunction with other monosaccharides (e.g., fructose, galactose and mannose) and some disaccharides (e.g., maltose, lactose and sucrose; which are not permeable to mammalian cells [Figure 37]).
71 Chapter 3
A 100 B 100 - - vel 80 80 NAD+ NAD+ Le 60 60 RNA 40 40 p m 20 20 Txni mRNA Level Txnip 0 0 00.2151025 0 0.2 1 5 10 25 Glucose (mM) 3OMG (mM)
C 300 D 100 - - 250 80 NAD+ NAD+ 200 60 150 10 40 100 5 20
50 0 mRNATxnip Level 0 Txnip mRNA Level Txnip 00.04 0 e e e s s 00.21 51025 nno ructos F alacto Glucose 2DG (mM) G Ma
E 3 F 3.5 - 2.5 3 2.5 NAD+ 2 2 1.5 1.5 1 - 1 0.5 NAD+ 0.5 Txnip mRNA Level Txnip Txnip mRNA Level Txnip 0 0 0 5 10 25 0 5 10 25 Sucrose (mM) Lactose (mM) G H 25 70 - 60 20 - NAD+ 50 15 NAD+ 40 10 30 20 5
Txnip mRNA Level Txnip 10 Txnip mRNA Level Txnip 0 0 0 0.2 1 5 10 25 0 0.2 1 5 10 25 Maltose (mM) G6P
Figure 38. The effect of glucose and glucose homologs on Txnip expression in the absence or presence of NAD+. HeLa cells were incubated in glucose‐free medium overnight. Glucose (Glc) or other glucose analogs (with various concentrations) were added for 4 hrs; in one selected group, cells were also co‐treated with 0.1 mM NAD+. Txnip mRNA levels were measured using Real‐Time PCR. In (C), the inset is a 15‐ fold enlargement, which emphasizes the induction effect by NAD+ in the presence of 0.04 mM of 2DG.
72 Chapter 3
Glucose induced Txnip expression in a dose‐dependent manner, and its effect
on Txnip expression was further boosted by NAD+ treatment (Figure 38A). 3OMG
induced the Txnip expression at higher concentrations (Figure 38B). This might be explained by a low affinity of 3OMG with glucose transporters; nevertheless, the effect of 3OMG on Txnip expression was also amplified by NAD+ (Figure 38B). 2DG dramatically induced the Txnip expression even at low concentrations, and this induction was marginally affected by NAD+; only when very low concentration (e.g.,
0.04 mM) of 2DG was used could NAD+ significantly boost the effect of 2DG on
Txnip expression (Figure 38C).
When other monosaccharides were tested, fructose and galactose exhibited marginal effect on Txnip expression, and NAD+ was not effective in the presence of these two monosaccharides; on the other hand, mannose can induce Txnip
expression, which is amplified by NAD+; the effects of mannose are similar to glucose
(Figure 38D). When disaccharides were tested, sucrose and lactose did not show any effects on Txnip expression, and in the presence of these two disaccharides, NAD+ did not show significant effect on Txnip expression (Figure 38E and 38F). Maltose alone only marginally induced Txnip expression; however, it synergized with NAD+ to dramatically induce Txnip expression (Figure 38G). The fact that NAD+ induces
Txnip expression in the presence of maltose (di‐glucose, cell impermeable; see Figure
37 for the structure) suggests that the induction of Txnip expression by adenosine‐
containing molecules is dependent on glucose, but not necessarily glucose transport.
Interestingly, I also found that G6P (which is charged and cell impermeable) also
induced Txnip expression at high concentration (25 mM), and this induction was
73 Chapter 3 slightly induced by NAD+ (Figure 38H). Again, this suggests that glucose transport is not a prerequisite for the induction of Txnip expression by adenosine‐containing molecules, although the membrane glucose transporters are clearly involved (Figure
22C).
A 140 120 PD NAD+ 100 NAD+ 80 PD 60 - 40 20 Txnip mRNA Level 0 0 0.2 1 5 10 25 Glc (mM)
B 250 PD NAD+ 200 evel NAD+ 150 PD 100 -
50
L mRNA Txnip 0 0 0.2 1 5 10 25
2DG (mM)
C 12 10 PD NAD+ NAD+ 8 PD 6 - 4 2 Txnip mRNA Level 0 00.215 1025 Maltose (mM)
Figure 39. The effect of PD169316 on Txnip induction by different carbohydrates and/or NAD+. HeLa cells were incubated in glucose‐free medium overnight. PD (PD169316, 10 μM) were added in selected groups for 10 min, and then NAD+ and/or sugars were added for 4 hrs. Txnip mRNA levels were measured using Real‐Time PCR.
74 Chapter 3
3.2.2.2 Effect of PD169316 on Glucose, 2DG or Maltose/NAD+ Induced Txnip
Expression
PD169316 can repress the induction of Txnip expression by glucose (Schulze et al., 2004) as well as impede the effect of adenosine‐containing molecules (section
3.2.1.13.3), although it is not clear if the effect of PD169316 is purely mediated by p38
MAPK. Here, I tested the effect of PD169316 on the induction of Txnip expression by glucose, 2DG or maltose plus NAD+. As shown in Figure 39, in the presence of
PD169316, ~80% of Txnip expression induced by glucose was repressed (A); ~50% of
Txnip expression induced by 2DG was repressed (B); and ~100% of maltose/NAD+ induced Txnip expression was repressed (C); in all cases, PD169316 abolished the
effect of NAD+ on Txnip expression (A‐C).
Based on these results, it is suggested that glucose may evoke more than one signaling pathways to induce Txnip expression. One pathway is glucose metabolism
(or G6P) dependent (Schulze et al., 2004; Stoltzman et al., 2008), which is not repressed by PD169316; and the other pathway is metabolism independent (e.g., maltose/NAD+ induced), and this pathway is efficiently repressed by PD169316. The
amplified metabolism independent pathway by adenosine‐containing molecules may be abolished by PD169316 to reduce the overall Txnip expression output.
3.2.2.3 Ca2+ Chelator Abolishes the Stimulatory Effect of Glucose on Txnip
Expression
It has been shown that Ca2+ played an important role in adenosine‐containing
molecules induced Txnip expression (Figure 28). Here, the effect of Ca2+ chelator on the induction of Txnip expression by glucose and its analogs has also been tested. As
shown in Figure 40, in the presence of BAPTA‐AM, the mRNA level of Txnip was
75 Chapter 3 not effectively induced by glucose, 3OMG or 2DG. This suggests that the
intracellular Ca2+ homeostasis is also critical for inducing Txnip expression by
glucose.
75 - Glc 50 3OMG 2DG 25 Txnip mRNA Level mRNA Txnip 0 - BAPTA-AM
Figure 40. Glucose induced Txnip expression was repressed by BAPTA‐AM. HeLa cells were incubated overnight in glucose (Glc) ‐free medium. Selected groups were pre‐treated with 10 μM of BAPTA‐AM for 10 min, and 10 mM of Glc, 2DG, or 25 mM of 3OMG, was added for another 2 hrs.
3.2.3 Inhibitors of Oxidation Phosphorylation Repressed Txnip Expression
3.2.3.1 Nitric Oxide (NO) and Sodium Azide (NaN3) Repress Txnip Expression
It has been earlier shown that the expression of Txnip is repressed by NO
(Schulze et al., 2006); however, the underlying mechanism for this repression is not clear. In this study, I also investigated the effect of NO on Txnip expression, with an
aim to understand the underlying mechanism. As shown in Figure 41A, when cells
were incubated with an NO donor DETA/NO, the Txnip mRNA expression level was
down‐regulated in a dose‐dependent manner. In a time course, the Txnip expression was repressed by DETA/NO after 0.5 hr incubation; the repression peaked at 4 hrs but was slightly recovered at 8 hrs (Figure 41B). Another NO donor, S‐nitroso‐N‐
acetylpenicillamine (SNAP), was able to impose a similar effect on Txnip expression
(Figure 41C).
76 Chapter 3
A 1.2 B 1.2 DETA/NO DETA/NO 1.0 1.0 evel 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 Txnip mRNA L Txnip mRNA Level Txnip mRNA 0.0 0.0 0 0.250.5 1 2 0 0.5 1 2 4 8 mM Hr(s)
C 1.2 D 1.2 DETA/NO 1.0 1.0 evel 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 Txnip mRNAL Txnip mRNA Level Txnip mRNA 0.0 0.0 0 1 2 - 0 5 10 20 40 SNAP (mM) μM LY83583
Figure 41. Effects of NO donors and GC inhibitor on Txnip expression. Titration (2 hrs, A) and time course (1 mM, B) of DETA/NO on Txnip mRNA levels. SNAP (C) repressed Txnip mRNA levels (2 hrs). DETA/NO‐repressed Txnip expression was rescued by LY83583 (D); cells were pre‐treated with LY83583 for 10 min before adding DETA/NO (2 hrs). HeLa cells were used in all experiments, and Txnip mRNA levels were measured using Real‐Time PCR.
NO is known to be able to activate soluble guanylate cyclase (GC) that in turn produces cyclinc guanosine monophosphate (cGMP) and activates protein kinase G
(PKG) (reviewed in Villalobo, 2006). I found that the repression of Txnip expression by DETA/NO was relieved by a guanylate cyclase inhibitor, LY83583 (Figure 41D), implying a mediating role of GC in the Txnip expression repression by NO.
77 Chapter 3
1.4 1.2 1.0 0.8 0.6 0.4 0.2 TxnIP mRNA Level TxnIP mRNA 0.0 - O Q /N LY DQ O OD O+ + /N DETA LY83583 ATA DETA/NO+rp ETA/NO8-pCPT-cGMP DE rp-pCPT-cGMPD
Figure 42. The effect of NO on Txnip expression was not mediated by GC. HeLa cells were pre‐treated for 10 min with 10 μM of LY83583 (LY), rp‐pCPT‐cGMP (rp) or ODQ, and then 1 mM DETA/NO was added for another 2 hrs. For 8‐pCPT‐cGMP treatment, cells were treated with 50 μM of the chemical for 2 hrs.
However, given the observations described below, the classical signaling pathway evoked by NO (GC, cGMP and PKG) might not be involved in the Txnip
expression repression, and the effect of LY83583 (Figure 41D) might be due to certain
non‐specific effects. Another GC inhibitor, 1H [1,2,4] oxadiazolo[4,3,‐a]quinoxalin‐1‐
one (ODQ), could not rescue the Txnip expression in the presence of DETA/NO, and
ODQ alone slightly repress Txnip expression (Figure 42). A PKG inhibitor, rp‐pCPT‐
cGMP, also failed to rescue the Txnip expression repressed by DETA/NO (Figure 42).
Moreover, a PKG activator, 8‐pCPT‐cGMP, did not shown significant effect on Txnip
expression (Figure 42). These observations are not in agreement with the possibility
that the effect of NO on Txnip expression is mediated by the GC‐PKG pathway.
78 Chapter 3
A B NaN NaN 1.00 3 1.00 3
0.75 0.75
0.50 0.50
0.25 0.25
Level Txnip mRNA Level Txnip mRNA 0.00 0.00 0 0.250.5 1 2 0 0.5 1 2 4 8 mM Hr(s)
C 120 - DETA/NO NaN3 100 80 60 40 % of % of Control 20 0
La t T 2OS I-38 a He U W rk Ju Namalwa B
Figure 43. The effect of NaN3 on Txnip mRNA Levels. (A) Titration, with 2 hrs treatment. (B) Time course, with 1 mM of NaN3. In both experiments HeLa cells were used. (C) DETA/NO or NaN3 repressed Txnip expression in diverse cell lines. Different cells were treated with 1 mM of DETA/NO or NaN3 for 2 hrs.
NaN3 and NO share certain similarities in structure and functions, including inhibiting the mitochondria function (Figure 44). In this study, the effect of NaN3 on
Txnip expression was tested. As shown in Figure 43, NaN3 down‐regulated Txnip mRNA levels in a similar fashion as DETA/NO did, in both dose‐response and time course (Figure 43A and 43B, respectively). In addition to HeLa cells, DETA/NO and
NaN3 repressed Txnip expression in a variety of mammalian cell lines of diverse
tissue origins (Figure 43C). This suggests that the repression of Txnip expression by
DETA/NO or NaN3 is a ubiquitous phenomenon.
79 Chapter 3
Figure 44. A simplified representation of oxidative phosphorylation. The protein complexes in the electron transport chain, the ATP synthesizer, and their inhibitor(s) are indicated. CCCP is able to disrupt the H+ gradient.
3.2.3.2 Inhibition of Oxidative Phosphorylation Represses Txnip Expression
Both NO and NaN3 can inhibit the activity of complex IV of the electron
transport chain in mitochondria (Figure 44), thus repression of Txnip expression could be due to disrupted oxidative phosphorylation. To test this hypothesis, cells were treated with other types of inhibitors (Figure 44) of oxidative phosphorylation to examine their effects on Txnip expression. Among the tested inhibitors, rotenone
(complex I inhibitor), antimycin A and myxothiazol (complex III inhibitors) could effectively block Txnip expression; the time course of these inhibitors were almost identical to that of DETA/NO or NaN3 (Figure 45). In contrast, a complex II inhibitor,
2‐thenoyltrifluoroacetone (TTFA), was unable to repress Txnip expression (Figure 45);
this is line with a non‐essential role of complex II for electron transfer (Figure 44).
I have also tested oligomycin A, an inhibitor against complex V (FoF1‐ATP synthase). Similar to most tested inhibitors, oligomycin A swiftly repressed Txnip
expression (Figure 45). In addition, carbonyl cyanide 3‐chlorophenylhydrazone
(CCCP), an un‐coupler that disrupts the H+ gradient cross the inner mitochondrial membrane, also repressed Txnip expression (Figure 45).
80 Chapter 3
Rotennone TTFA Antimycin A Myxothiazol 1.2 1.2 1.2 1.2 1.0 1.0 1.0 1.0
0.8 Txnip 0.8 0.8 0.8 0.6 H2B 0.6 Txnip 0.6 Txnip 0.6 Txnip H2B H2B H2B 0.4 0.4 0.4 0.4 0.2 0.2 0.2 0.2 0.0 0.0 0.0 0.0 0 1 2 4 0 1 2 4 0 1 2 4 0 1 2 4 Hr(s) Hr(s) Hr(s) Hr(s)
NaN3 DETA/NO Oligomycin A CCCP 1.2 1.2 1.2 1.2 1.0 1.0 1.0 1.0 Txnip Txnip 0.8 H2B 0.8 H2B 0.8 0.8 0.6 0.6 0.6 Txnip 0.6 Txnip H2B H2B 0.4 0.4 0.4 0.4 0.2 0.2 0.2 0.2 0.0 0.0 0.0 0.0 0 1 2 4 0 1 2 4 0 1 2 4 0 1 2 4 Hr(s) Hr(s) Hr(s) Hr(s)
Figure 45. Effects of oxidative phosphorylation inhibitors on Txnip or H2B expression. HeLa cells were treated with different inhibitors for 0‐4 hrs (1 μM of rotenone, myxothiazol or CCCP; 2 μM of TTFA; 1 μg/ml of antimycin A or oligomycin A; 1 mM of DETA/NO or NaN3). While Txnip expression inhibition by DETA/NO or NaN3 is due to repressed mitochondrial function as shown here, the repressed H2B transcription by DETA/NO or NaN3 is likely attributed to a unique mechanism related to perturbed cellular redox (Dai et al., 2008).
It seems that all chemicals which can inhibit oxidation phosphorylation can
effectively repress Txnip expression. Thus the expression of Txnip is positively
correlated to the status of cellular oxidation phosphorylation. When oxidation
phosphorylation is disrupted, mitochondria will unable to synthesis ATP, it is
possible that the shortage of ATP can down‐regulate Txnip expression.
81 Chapter 3
- BAPTA-AM 1.4 1.2 1.0 0.8 0.6 0.4
TxnIP mRNA Level TxnIP mRNA 0.2 0.0
- 3 P A zol A C in NaN cin CC myc othia my Rotenone DETA/NO nti yx A M Oligo
Figure 46. Oxidative phosphorylation inhibitors did not repress Txnip expression in the presence BAPTA‐AM. HeLa cells were pre‐treated with 10 μM of BAPTA‐ AM for 10 min, and then with different oxidative phosphorylation inhibitors (at the same concentrations used as in Figure 45) for another 2 hrs.
3.2.3.3 A Ca2+ Chelator Rescues Txnip Expression in the Presence of Oxidative
Phosphorylation Inhibitors
Ca2+ chelator, BAPTA‐AM, abolished the stimulatory effect of adenosine‐ containing molecules or glucose on Txnip expression (Figures 29 and 40). The role of
BAPTA‐AM in the repression of Txnip expression by oxidative phosphorylation inhibitors was also studied. As shown in Figure 46, when cells were pre‐incubated with BAPTA‐AM, oxidative phosphorylation inhibitors were unable to repress Txnip expression. This suggests that Ca2+ signaling/homeostasis is also important for the repression of Txnip expression by oxidation phosphorylation inhibitors, and for the
induction of Txnip expression by glucose and adenosine‐containing molecules.
Therefore, Ca2+ plays a master role in regulating Txnip expression.
82 Chapter 3
3.3 Discussion
3.3.1 Adenosine‐containing Molecules May Remain Extracellular to Induce Txnip
Expression
The majority of tested adenosine‐containing molecules are charged and
generally not cell‐permeable. On the other hand, adenosine is permeable (via
adenosine transporters) and, once inside the cells, metabolized. For molecules that
remain extracellular, e.g., ATP or ADP, stimulatory effects were sustainable; however,
the effect of adenosine was not (Figures 13 and 14). Extracellular ATPase or
nucleotidase (Haag et al., 2007) might gradually remove phosphate groups from
adenine nucleotides, forming adenosine that can then be taken up by cells. This
would deplete extracellular adenosine or adenine nucleotides with fewer phosphate
groups (e.g., AMP), thus the diminishing efficacies of these molecules in long‐term
assays (Figure 13; >8 hrs). NAD(H) can be degraded, by extracellular NADase (Haag et al., 2007), to ADP‐ribose, which contains an intact adenosine moiety and induces
Txnip expression (Table 3). This explains the long‐term sustainable effects of
NAD(H) (Figures 13 and 14).
Adenosine‐containing molecules might not need to be converted to adenosine
to induce the Txnip expression because ATPγS, which resists ATPase (Beukers et al.,
1995), is also stimulatory (Table 3). Supporting this notion, adenosine and adenine
nucleotides show similar kinetics in a time course (Figure 13; short‐term 0‐4 hrs). I also found that the effect of ATP or ADP on Txnip expression was not affected by an
ATPase inhibitor, ARL 67156 (Figure 24). The fact that α‐NAD, a β‐NAD analogue that cannot function as a co‐enzyme, and non‐hydrolysable ATPγS, induce the Txnip expression (Table 3), suggests that cellular redox state or metabolic pathways
83 Chapter 3 utilizing adenine nucleotides‐ or NAD(H)‐consuming enzymes are also unlikely directly involved in mediating the induction of Txnip expression by adenosine‐ containing molecules. Finally, AMPS (AMPαS), an AMP analog that is resistant to
degradation at the alpha position, remains competent in inducing Txnip expression
(Table 3). Taken together, internalization of adenosine‐containing molecules, and processing of adenosine phosphate(s) into adenosine, might not be required for
inducing Txnip expression.
3.3.2 Potential Membrane Targets for Adenosine‐containing Molecules
Known membrane targets that have potential affinity for adenosine‐ containing molecules include the purinergic receptors (P1, P2X or P2Y) (Burnstock,
2008) and adenosine transporters (Thorn and Jarvis, 1996), of which purinergic receptors are unlikely involved in the induction of Txnip expression (see section
3.2.1.12.1 ).
On the other hand, adenosine transporter inhibitors abolished stimulatory effect of NAD+ on Txnip expression (Figure 28), suggesting that adenosine transporters or other membrane targets, which share functional domains that are
similarly inhibited by NBTI, Dipyridamole or Dilazep, play a critical role in the induction of Txnip expression by adenosine‐containing molecules. Adenosine transporters contain a small pore that facilitates cellular adenosine up‐take. Adenine nucleotides or other adenosine derivatives are larger, often contain charged group(s) and may not be taken up via adenosine transporters. As earlier discussed, adenosine‐containing molecules are, most‐likely, required to remain extracellular to induce Txnip expression, and the up‐take of adenosine via adenosine transporters can actually diminish the effect of adenosine in a long‐term. Therefore, the
84 Chapter 3 transporter activity of adenosine transporters itself may not participate in inducing
Txnip expression by adenosine‐containing molecules; but these transporters or other structurally similar membrane targets may serve as anchors for these molecules.
Their engagement with the membrane targets does not necessarily facilitate cellular transport but likely transmits extracellular signals into cells.
As adenosine transporter inhibitors can effectively block the induction of
Txnip expression (Figure 28), at this stage I do not absolutely rule out the possibility that the up‐taken adenosine, be it native or derived extracellularly from adenosine‐ containing molecules, may exert an effect in inducing Txnip expression. Further studies will be carried out to clarify this issue.
Interestingly, the negative charges around the ribose moiety of these molecules affected their efficacies in a dose‐dependent manner. NAD(H) and AMP contain fewer formal charges around the ribose than NADP(H), ADP and ATP; and the former, but not the latter and uncharged adenosine, sustained the stimulatory effects over a wide titration (Figure 12). Thus, adenosine‐containing molecules with intermediate levels of negative charges might be optimally configured to bind to plasma membrane targets and induce Txnip expression over a wide range of doses; the rest of molecules at high dosages might overwhelm the signaling system involving the membrane targets. Hence, the homeostasis of various adenosine‐ containing molecules might be critical for regulating and fine‐tuning Txnip expression.
3.3.3 Signaling Pathway(s) Involved in the Induction of Txnip Expression by
Adenosine‐containing Molecules
It was reported that adenosine, ATP, ADP, AMP and ATPγS, but not dATP,
85 Chapter 3
UTP and GTP could up‐regulate the cellular cAMP levels (Matsuoka et al., 1995). The effects of these molecules on cellular cAMP levels correlated with their effects on
Txnip expression (Table 3). However, in this study, the involvement of cAMP related signaling pathways in the induction of Txnip expression has been excluded
(see section 3.2.1.13.2).
If adenosine‐containing molecules exerted their effects by converting to
adenosine, AMPK would be a potential downstream signaling molecule, as
adenosine uptake can induce AMPK activity (da Silva et al., 2006; Aymerich et al.,
2006). However, AMPK inhibitor failed to abolish the effect of adenosine‐containing molecules on Txnip expression (Figure 32). This suggests that AMPK is not critical
for mediating the signaling evoked by adenosine‐containing molecules. It has been
shown that AMPK can modulate activity of ChREBP, a homolog of MondoA, and
regulate expression of several glucose‐responsive genes (Kawaguchi et al., 2002). But the targeting sites in ChREBP are not conserved in MondoA.
MAPKs are also potential molecules transmitting signals of extracellular adenosine‐containing molecules to MondoA and/or MLX transcription factors for
Txnip expression, as a role of p38 MAPK in glucose induced Txnip expression has been suggested (Schulze et al., 2004). However, the role of MAPK in the induction of
Txnip expression by adenosine‐containing molecules is controversial, different inhibitors against MAPK exhibited inconsistent results (Figures 30 and 31). Among these inhibitors, SB202190 and PD169316 are structural analogs and both can target p38 MAPK, but their effects on Txnip expression in the presence of NAD+ or ATP were different. Therefore, it is hard to claim that p38 MAPK is a signaling molecule mediating the induction of Txnip expression by adenosine‐containing molecules and,
86 Chapter 3 probably, glucose. PD169316 abolished the effect of both adenosine‐containing molecules (Figures 30 and 31) and glucose (partially, Figure 39), thus it could function as an important tool to regulate Txnip expression under different environmental conditions, although it may not function through p38 MAPK. The
real relevant target(s) of PD159316 may play a critical role in the regulation of Txnip expression.
In the presence of a cell permeable Ca2+ chelator, adenosine‐containing molecules failed to induce the expression of Txnip (Figure 28), indicating that the intracellular Ca2+ signaling (or homeostasis) is critical for this induction. At this stage,
how the cross‐plasma membrane signaling evoked by extracellular adenosine‐
containing molecules is transmitted to Ca2+, and then to the MondoA/MLX complex to activate the Txnip transcription is unclear; the involved signaling steps deserve further investigation. BAPTA‐AM not only abolished the effect of adenosine‐ containing molecules on Txnip expression (Figure 28), but also impeded the effect of glucose (Figure 40) and oxidative phosphorylation inhibitors (Figure 46). Thus Ca2+ may function as a master regulator for Txnip expression. A recent study indicates
that Txnip expression is also reduced by calcium channel blockers (Chen et al., 2009), suggesting that intracellular Ca2+ homoeostasis is critical for Txnip expression.
3.3.4 The MondoA/MLX Complex Mediates Txnip Expression
MondoA or ChREBP in association with MLX was implicated in enhancing the expression of many glucose responsive genes, including those encoding L‐ pyruvate kinase, fatty acid synthase, and lactate dehydrogenase‐A (Ma et al., 2006;
Sans et al., 2006; Ma et al., 2005; Dentin et al., 2004; Ishii et al., 2004; Stoeckman et al.,
2004; Yamashita et al., 2001). While MLX is essential for mediating the induction of
87 Chapter 3
Txnip expression by adenosine‐containing molecules (Figure 17), which factor
(MondoA or ChREBP) is the genuine partner requires further resolution.
The ChREBP expression is restricted to tissues such as liver, kidney, pancreas and intestine (de Luis et al., 2000; Cairo et al., 2001), and is extremely low in other cell types, as opposed to the MondoA expression that maintains measurable levels in a variety of cell types (not shown). Ectopically expressed MondoA (Figure 18) or
ChREBP (not shown) exhibited massive nuclear localization in 2DG‐treated L6 cells.
This indicates that both over‐expressed factors are potent in receiving glucose‐ mediated signal(s). In our assay system, MondoA, despite a less‐efficient ectopic expression than ChREBP, functioned much more potently than ChREBP; in
MondoA‐deficient cells, the ectopic RNAi‐resistant MondoAr, but not ectopic
ChREBP, restored the basal Txnip expression and its induction by adenosine‐
containing molecules (Figure 19). In addition, the transcription capability of ChREBP
is activated by glucose through the pentose pathway in conjunction with protein phosphatase 2A (PP2A) (Kabashima et al., 2003); and I found that okadaic acid, a potent PP2A inhibitor, did not affect the induction of Txnip expression by adenosine‐ containing molecules (not shown). Finally, MondoA occupies the native Txnip promoter (Stoltzman et al., 2008).
Taken together, the MondoA/MLX complex is a major player that supports the ChoRE‐dependent Txnip expression and its induction by adenosine‐containing molecules, which most likely takes place ubiquitously. A recent study, however, shows that glucose leads to a time‐ and does‐dependent recruitment of ChREBP onto the Txnip promoter in INS‐1 beta cells (cha‐molstad et al., 2009); this might be due to
88 Chapter 3 high expression level of ChREBP in pancreatic beta cells, and suggests that the Txnip gene may also employ tissue‐specific transcription factors to regulate its expression.
Adenosine‐containing molecules might elevate G6P levels by modulating the activities of enzymes that produce or utilize G6P; this in turn may enhance an accumulation of MondoA/MLX in the nuclei and stimulate Txnip expression. It is equally plausible that these molecules modulate Txnip expression at other levels.
For instance, in a G6P dependent manner, the signaling evoked by adenosine‐ containing‐molecules may accelerate the release of MondoA/MLX from mitochondria or nuclear transport of this complex. In this scenario, the glucose signaling alone might trigger the MondoA/MLX release and nuclear translocation with a limited scale, and this process can be amplified by adenosine‐containing molecules to a larger scale to enhance Txnip expression (Figure 21).
3.3.5 Physiological Significance
Many adenosine‐containing molecules are readily derivable from foods, and might be natural metabolites used by our body to regulate Txnip expression. Some adenosine‐containing molecules can also be released by certain tissues under physiological or pathological conditions (Han et al., 1998; Yin et al., 2007), hence implying local functions. The expression of Txnip is induced by glucose, and Txnip can inhibit glucose uptake, thus forming a negative feedback loop for the regulation of the glucose utilization (Parikh et al., 2007; Yoshioka et al., 2007; Minn et al., 2005;
Schulze et al., 2004). The induction of Txnip expression by adenosine‐containing molecules is glucose‐dependent, suggesting an (accelerating) impact on the kinetics of this negative feedback loop (Figure 47).
89 Chapter 3
Figure 47. A negative feed‐back loop for glucose uptake. Elevated extracellular glucose concentration will increase glucose uptake, which in turn will induced Txnip expression via ChoRE and MondoA/MLX; Txnip accumulated will inhibit the uptake of glucose. The kinetics of this negative feed‐back loop is speeded up in the presence of adenosine‐containing molecules. MdA: MondoA.
Glucose uptake and utilization are highly regulated to maintain normal physiology; eukaryotic cells employ diverse mechanisms to sense glucose and regulate the expression of genes involved in metabolic control. Organisms often develop metabolic diseases, e.g., diabetes mellitus and hyperlipidemia, attributed to abnormal glucose homeostasis and/or metabolism. The revelation that adenosine‐
containing molecules stimulate Txnip expression, most likely by amplifying the glucose signaling, may set the stage for chemical mimicries or antagonisms for intervention of metabolic disorders resulting from abnormal glucose homeostasis.
In this study, I also found that extracellular NAD+ could induce cell cycle
arrest and up‐regulate protein level of p21cip1; this is consistent with the anti‐ proliferative function of Txnip (Joguchi et al., 2002; Schulze et al., 2002; Han et al.,
2003; Nishinaka et al., 2004b; Jeon et al., 2005). Extracellular adenosine‐containing
90 Chapter 3 molecules, such as ATP and adenosine, can also act as signaling molecules regulating cell death (Wen and Knowles, 2003). Txnip can also induce apoptosis in certain cells
(Wang et al., 2002; Wang et al., 2006), suggesting that Txnip is a potential mediator in the process of cell death induced by adenosine‐containing molecules.
3.3.6 Two Signaling Pathways Evoked by Glucose for Inducing Txnip Expression
How a cell can sense the concentration of extracellular glucose (or other sugars) is a fundamental problem in biology. In yeast, cells can use two plasma membrane proteins, Snf3 and Rgt2 (homologous to mammalian glucose transporter), as glucose sensor (Ozcan et al., 1998). Snf3 and Rgt2 can, without glucose transport activity, sense glucose availability and activate signaling pathways to regulate expression of glucose sensitive genes.
In mammalian systems, ortholog of yeast Snf3 or Rgt2 has not been identified.
Therefore, mammalian cells do not have known plasma membrane sensor(s) for extracellular glucose. To my knowledge, only sporadic reports have shown existence of plasma membrane glucose sensors in mammalian cells. Glut1 has been proposed as a sensor, transducer and amplifier for glucose signaling, and upon binding of glucose and some disaccharides, Glut1 is able to initiate a signaling pathway to activate ERK (Bandyopadhyay et al., 2000). In intestine, SGLT3, a member of sodium/glucose co‐transporters, cannot transport glucose but may function as a tissue specific glucose sensor (Diez‐Sampedro et al., 2003). However, in the conventional concept, glucose sensing in mammals requires glucose to be taken up by cells and phosphorylated by hexokinases. Thus the flow of glucose and subsequent phosphorylation play important roles in glucose sensing and modulation of intracellular signaling and gene expression network.
91 Chapter 3
It has been proposed that glucose can stimulate two signaling pathways to induce Txnip expression in mammalian cells, one pathway is metabolism (glucose phosphorylation, G6P) dependent (Stoltzman et al., 2008), and the other is metabolism independent as the Txnip expression is induced by 3OMG, which cannot be phosphorylated (Minn et al., 2006). In both models, the flow of glucose into cells is required.
In this study, I found that the flow of glucose into cells may not be required to induce the Txnip expression. G6P, which is cell impermeable, can also induce Txnip expression at high concentrations (e.g., 25 mM) (Figure 38). Maltose (di‐glucose) is also cell impermeable and can only marginally induce Txnip expression; however, the marginally‐induced Txnip expression can be dramatically boosted by adenosine‐ containing molecules (Figure 38). Adenosine‐containing molecules may facilitate the binding of maltose to glucose transporters, and the engagement of maltose onto the glucose transporters, without being up‐taken, might induce a signaling pathway to activate the Txnip expression. Indeed, an adenine nucleotide binding site on the extracellular region of Glut1 has been predicated (Liu et al., 2001), and adenosine and
ATP has been shown to modulate glucose transport in vitro (Lachaal et al., 2001).
3OMG has a lower affinity to glucose transporters; its effect on Txnip expression was significant only when at high concentrations (Figure 38). The effect of 3OMG on the
Txnip expression may not require the transport activity of glucose transporters, the engagement of 3OMG onto glucose transporters might suffice.
If the above explanations are legitimate, the signaling pathways mediating glucose induced Txnip expression might be segregated into two pathways. One pathway is metabolism dependent, which requires flow and phosphorylation of
92 Chapter 3 glucose, thus correlated to cellular G6P levels. The other pathway is glucose
transporter engagement dependent, which does not require glucose flow, and may
function as a glucose sensor at plasma membrane. The transporter engagement
dependent pathway might be amplified by adenosine‐containing molecules.
3.3.7 Txnip Expression Is a Sensor of Oxidative Phosphorylation Status
Other than extracellular glucose or adenosine‐containing molecules, Txnip expression was also modulated by inhibitors of oxidative phosphorylation (Figure
45); this suggests that the Txnip expression can sense the status of mitochondrial oxidative phosphorylation. One common feature of these inhibitors is that all of them can inhibit ATP synthesis in mitochondria; it is possible that some ATP sensing molecules can be modulated to repress Txnip expression, and further investigations are required to test this possibility. Some hexokinases are associated with the outer membrane of mitochondria and preferred to access mitochondria‐generated ATP and convert glucose to G6P (Shinohara et al., 1997)). Cellular G6P level is critical for
Txnip expression (Stoltzman et al., 2008), thus the inhibition of mitochondrial ATP synthesis by inhibitors of oxidative phosphorylation may affect the intracellular G6P levels and G6P‐dependent Txnip expression.
Another possible mechanism for the repression of Txnip expression by inhibitors of oxidative phosphorylation is related to HIF‐1α. Hypoxia is able to induce Txnip expression (Xiang et al., 2005; Le Jan et al., 2006; Karar et al., 2007; Baker et al., 2008), and the stability of HIF‐1α is down‐regulated by inhibitors of oxidative phosphorylation under hypoxic conditions (Gong and Agani, 2005), the latter might be due to redistribution of oxygen toward non‐respiratory oxygen‐dependent targets
(Hagen et al., 2003). Therefore, when oxidative phosphorylation is inhibited, the
93 Chapter 3 destruction of HIF‐1α will be enhanced, and this may contribute to reduced Txnip expression. However, role(s) of these inhibitors targeting oxidative phosphorylation on HIF‐1α is not clear in cells maintained under normal oxygen level; therefore, further studies should be carried out to test the involvement of HIF‐1α in Txnip
expression under disrupted oxidative phosphorylation.
3.4 Conclusion and Perspectives
In the process of identifying small molecules modulating Txnip expression, I
discovered that the transcription of Txnip was induced by an array of adenosine‐ containing molecules, of which an intact adenosine moiety is necessary and sufficient.
The induction of Txnip expression by these molecules was glucose‐dependent and was mediated by ChoRE and its associated transcription factors (MondoA and MLX)
that also mediate the glucose‐induced Txnip expression. Hence the regulatory role
of these small molecules on Txnip expression may be exerted via amplifying glucose
signaling.
Although the exact plasma membrane target(s) of extracellular adenosine‐
containing molecules has not been identified, the results obtained so far suggest that
adenosine transporters or membrane proteins that are functionally analogous to the
adenosine transporters are responsible to transmit these extracellular signals into
cells. The signaling pathway(s) evoked by adenosine‐containing molecules was not
clear at this stage; I had shown that intracellular Ca2+ signaling is crucial for the
induction of Txnip expression by adenosine‐containing molecules.
In this study, I also investigated the mechanism underlying glucose induced
Txnip expression, which supports a two‐pathway model for the induction of Txnip
94 Chapter 3 expression by glucose. In this model, one pathway is glucose metabolism (G6P)
dependent, and the other pathway is glucose transporter engagement dependent.
The transporter engagement dependent pathway suggests a direct glucose sensing mechanism at plasma membrane of mammalian cells.
The expression level of Txnip was also tightly correlated with the status of mitochondrial oxidative phosphorylation. This suggests that the Txnip expression is not only regulated by glucose availability, but also by intracellular metabolic competencies, and supports a pivotal role of Txnip expression in response to cellular metabolic status.
In this study, I provide the first description of the adenosine‐containing molecules‐induced Txnip expression, reveal a glucose transporter engagement dependent pathway for Txnip expression, and show that the Txnip expression is correlated with the oxidative phosphorylation status. These achievements place the
Txnip expression in an important position in the field of nutritional/metabolic status sensing. A lot of basic biological problems related to these concepts should be investigated in the future, which includes identification of the membrane target(s) of adenosine‐containing molecules and signaling pathway(s) induced by glucose and adenosine‐containing molecules, study on the synergistic mechanism of glucose and adenosine‐containing molecules on Txnip expression, and dissecting the retrograde signaling (mitochondria to nuclear) induced by oxidative phosphorylation inhibitors.
The in vivo functions (e.g., glucose utilization and cell cycle progression) of Txnip expression inducible by adenosine‐containing molecules‐induced can be investigated as well. All the above endeavors may eventually shed some light on the development of drugs targeting some metabolic diseases and cancer.
95
Chapter 4
Regulatory Mechanisms Underlying the Induction of Txnip
Expression by Glucose or Adenosine‐containing Molecules Chapter 4
4.1 Preface
The expression of Txnip gene is regulated by a diverse array of factors
(section 1.5). A number of physiological cues can dictate the efficacy of Txnip
expression, which is inhibited by insulin (Parikh et al., 2007), stimulated by
glucocorticoid (Wang et al., 2006; Kolbus et al., 2003), vitamin D (Chen and DeLuca,
1994), peroxisome proliferator‐activated receptor (PPAR) agonists (Rakhshandehroo
et al., 2007; Billiet et al., 2008a,b; Oka et al., 2006a), transforming growth factor beta
(TGF‐β) (Han et al., 2003), inhibitors of HDAC (Butler et al., 2002; Huang and Pardee,
2000) and certain stress signals (Kim et al., 2004). Most interestingly, the expression level of Txnip is tightly correlated with the extracellular glucose levels (Minn et al.,
2005; Schulze et al., 2004; Hirota et al., 2002; Shalev et al., 2002), and this glucose‐ induced Txnip expression negatively feeds back to the cellular glucose uptake system (chapter 3; Chutkow et al., 2008; Parikh et al., 2007). I have also identified a large class of adenosine‐containing molecules that are capable of inducing the Txnip expression in a glucose‐dependent manner (chapter 3).
While majority of Txnip‐related research focuses on the Txnip function, the underlying mechanisms that govern Txnip gene transcription are largely unknown despite the identification of a lot of modulators for Txnip expression. For the glucose or adenosine‐containing molecules induced Txnip expression, MondoA and MLX function through ChoRE and are crucial for transmitting signals into the nucleus to activate the Txnip promoter (chapter 3; Stoltzman et al., 2008); however, a detailed
description of this signaling pathway and the regulatory mechanisms at the
promoter level remains elusive.
The aim of this study is to dissect the Txnip promoter in a detailed manner
96 Chapter 4 and identify cis‐regulatory elements required for mediating the induction of Txnip
expression by glucose or adenosine‐containing molecules. In addition, the dynamic
recruitment of trans‐factors, e.g., MondoA and MLX, onto Txnip promoters in
response to glucose with or without adenosine‐containing molecules, also deserve a detailed investigation. These studies will help us understand the underlying mechanism of the expression of genes in response to nutritional factors in mammals.
Txnip plays important roles in metabolism. Txnip knock‐out mice have exhibited phenotypes such as hyperlipidemia, hypoglycemia and high plasma ketone bodies (Oka et al., 2006b). Txnip has been genetically linked to abnormal blood pressure in diabetes mellitus and hypertriglyceridemia (van Greevenbroek et al., 2007). A nonsense mutation in Txnip gene, encoding a truncated Txnip protein, was implicated in an FCHL‐like phenotype in HcB‐19 mice (Bodnar et al., 2002). In some other studies, however, the Txnip gene was not linked to FCHL in HcB‐19 mice or in humans; instead, the causation for the FCHL‐like phenotype in HcB‐19 mice seemed to be defect(s) in the gene encoding USF1 (Pajukanta et al., 2004; Coon et al.,
2004; van der Vleuten et al., 2004). USF1 can regulate gene expression by binding to
E‐box as homo‐dimer or hetero‐dimer with USF2, and multiple E‐boxes (e.g., ChoRE contains two E‐boxes) can be identified on Txnip promoter. Hence, USFs might be transcription factors regulating Txnip expression. In this study, investigating the role of USFs in Txnip expression, including involvement of USFs in ChoRE‐mediated induction of Txnip expression by adenosine‐containing molecules or glucose, is also an aim. Results obtained from these studies may help us designate causation of the development of FCHL and better understand other genetic diseases that have been causatively linked to abnormal glucose or lipid metabolism.
97 Chapter 4
4.2 Results
4.2.1 Regulatory Mechanisms at the Promoter Level Governing Glucose or
Adenosine‐containing Molecules Induced Txnip Expression
4.2.1.1 Txnip Promoter Regions Critical for Expression Induction by NAD+ or
Glucose
A ChoRE site (at ~80 bp upstream of the transcriptional start site) is essential to support the Txnip expression induction by glucose or adenosine‐containing molecules (chapter 3; Minn et al., 2005b); however, the minimal promoter that remains inducible to these molecules was not yet defined. To this end, I performed a more detailed promoter analysis using Txnip promoters with different lengths fused to a luciferase reporter gene (Figure 48). When ectopically introduced into
U2OS cells, reporters with 269 bp or longer Txnip promoter sequences exhibited ~6‐ fold stimulation by NAD+; the promoter with 73 bp promoter sequence or empty pGL‐3 vector were not induced by NAD+ (Figure 49). This indicates that the cis‐ regulatory element(s) responsible for NAD+ induced Txnip promoter activity is located somewhere between ‐269 and ‐73.
Within this region (‐269 to ‐73), I generated more truncated Txnip promoters
(Figure 48), and their responses to NAD+ were tested. As shown in Figure 50A and
50B, in U2OS cells, reporters with less than 169 bp Txnip promoter sequences exhibited marginal response to NAD+ treatment. I also tested the responses of these promoters to NAD+ in HeLa cells, reporters with 142 bp or longer promoter sequences exhibited ~4‐fold, and a reporter with 111 bp sequences showed ~2‐fold, stimulation by NAD+; however, reporters with shorter sequences were not responsive to NAD+ (Figure 50C and 50D). Although some differences exist between
98 Chapter 4 results obtained from HeLa and U2OS cells, the previously identified ChoRE, which is located at ~80 bp upstream of the transcriptional start site, is apparently not
sufficient to support optimal induction of Txnip expression by NAD+.
Figure 48. A schematic representation of truncated Txnip promoters. The length (to the transcription start site) of Txnip promoters was shown by numbers on the right. The position of an earlier defined ChoRE was indicated by the blue line.
A B 1.00 8 NAD+ - 7 ** ** * 0.75 6 ** ** ** 5 0.50 4 3 0.25 2
Promoter Activity 1 Fold(s) of Induction of Fold(s) 0.00 0 3 3 3 33 7 99 41 73 760 65 6 441 269 L3 760 653 63 4 269 L3 1299 pG 12 pG
Figure 49. Responses of Txnip promoters to NAD+. Txnip promoters (serials deletion) were transfected into U2OS cells, and cells were treated with or without 0.2 mM of NAD+ for 16 hrs. Luciferase activities were measured (A) and fold(s) of induction by NAD+ were calculated (B).
99 Chapter 4
A 0.7 B 5 + 0.6 - * NAD + 4 * 0.5 NAD 0.4 3 (U2OS) 0.3 (U2OS) 2 0.2 1 Activity Promoter 0.1 Fold(s) ofInduciton 0.0 0 9 2 1 73 69 63 73 L3 269 169 163 142 111 102 269 1 1 14 11 102 G 1299 pGL3 129 p
C 1.5 D 5 * + - ** * * NAD 4 * NAD+ 1.0 (HeLa) 3 (HeLa) 2 0.5 1
Activity Promoter Fold(s) ofInduction 0.0 0 9 9 3 1 2 3 9 9 3 2 3 3 6 6 1 0 73 6 6 6 0 7 269 1 1 142 1 1 L 2 1 1 142 111 1 L 129 pG 1299 pG E F 12 12 - Glc 10 10 * glc * * 8 8 6 6 * 4 4 * 2 Promoter Activity 2 Fold(s) ofInduction 0 0 9 9 3 1 2 9 6 6 1 0 73 69 42 73 269 1 1 142 1 1 2 169 163 1 111 102 12 pGL3 1299 pGL3
Figure 50. Responses of Txnip promoters to NAD+ or Glucose. Txnip promoters (serials deletion) were transfected into U2OS cells (A and B), HeLa cells (C and D) or L6 cells (E and F), and cells were treated with or without 0.2 mM of NAD+ (A‐D) or 10 mM of glucose (E and F) for 16 hrs. Luciferase activities were measured (A, C and E) and fold(s) of induction by NAD+ or glucose were calculated (B, D, F). In (E) and (F), cells were incubated in glucose‐free medium before induction.
I have also tested the effect of glucose on these Txnip promoters. In L6 cells, ectopic promoters with 111 bp or less sequences were not responsive to glucose; however, promoters with 142 bp or longer sequences showed 4‐8‐fold stimulation by
100 Chapter 4 glucose (Figure 50E and 50F). This suggests that the earlier defined ChoRE alone cannot support optimal induction of Txnip expression by glucose, and that some nucleotide sequences upstream of this ChoRE is also critical for the responsiveness of
Txnip promoter to glucose or adenosine‐containing molecules.
A B 5 * * 4
3
2
1 Fold(s) ofInduction 0
95 70 63 TA .1 69- el TA 269-632 145-63D 184-
Figure 51. The minimal Txnip promoter sequence required for mediating the stimulatory effect of NAD+. (A) Txnip promoter fragments were fused with a TATA box driven luciferase reporter. (B) These hybrid promoters were transfected into U2OS cells and their responses to NAD+ were tested (0.2 mM NAD+, 16 hrs).
To identify minimal promoter sequences required for the induction of Txnip expression by adenosine‐containing molecules, various Txnip promoter fragments were fused to a TATA‐only core promoter‐luciferase reporter gene (Figure 51A), and
their responses to NAD+ were tested in U2OS cells. As shown (Figure 51B), a promoter fragment (184‐63 bp upstream of the transcription start site) was sufficient to support the induction by NAD+, while a fragment (269‐95) lacking an earlier
defined ChoRE or fragments (145‐63 or d170) lacking sequences surrounding the ‐170
positions did not respond to NAD+. I conclude that both the earlier defined ChoRE and nucleotide sequences near the ‐170 positions are critical for the induction of
Txnip expression by adenosine‐containing molecules.
101 Chapter 4
The responses of Txnip expression to extracellular glucose or adenosine‐ containing molecules are general phenomena, which have been observed in diverse mammalian cell lines from different tissue origins (Chapter 3 and data not shown).
However, I found that the response to adenosine‐containing molecules was most
prominent in U2OS cells, and the response to glucose is most dramatic in L6 cells.
Therefore, in most experiments, I respectively used U2OS or L6 cells to test the Txnip
promoter activities in cells treated with adenosine‐containing molecules or glucose.
Figure 52. cis‐regulatory elements on Txnip promoter. ChoREs, CCAAT boxes and FOXO‐binding site on Txnip promoter were shown (underlined nucleotides); numbers indicate the distance to the transcription start site. Modified nucleotides introduced into mutant Txnip promoters are indicated by lower case letters (in red).
4.2.1.2 Tandem ChoREs on Txnip Promoters
I examined the Txnip promoter sequences (184‐63 bp) and found that, apart from the earlier known ChoRE (CACGAGggcagCACGAG; ~80 bp upstream of the transcription start site), the nucleotide sequences around the ‐170 region
(CACACCgtgtcCACGCG; Figure 52) mimicked a degenerate ChoRE, which is defined as two E‐boxes (CACGTG) separated by 5 nucleotides (Ma et al., 2006). Thus, an additional candidate ChoRE sequence on the Txnip gene promoter was identified; for convenience, I dubbed the previously identified ChoRE (~80 bp) as ChoRE‐a and
102 Chapter 4 -elements (ChoREs, FOXO-binding site, CCAAT or -elements (ChoREs, Sequence alignment of Txnip promot ers. promoter sequences ing CLUSTAL W program. The conserved cis oters from different species. Figure 53. Alignment of Txnip prom oters from of different species were aligned us inverted CCAAT) were hi ghlighted using boxes.
103 Chapter 4 the newly identified ChoRE (~170 bp) as ChoRE‐b (Figure 52, highlighted using
purple color). When ChoRE‐b was deleted, the promoter was not optimally induced
by glucose or NAD+ (Figures 50 and 51). This indicates that both ChoRE‐a and
ChoRE‐b are responsible for optimal induction of Txnip expression by glucose or
adenosine‐containing molecules.
Figure 54. Phylogenetic tree of Txnip promoters. The phylogenetic tree built from Txnip promoters of different species using Neighbor‐Joining (NJ) method.
Bioinformatics efforts allowed us to conclude that Txnip is vertebrate‐specific, and obtain genomic sequences of Txnip promoters of diverse species from fishes to human. I aligned the Txnip promoter sequences and found that the two ChoRE sequences, a CCAAT box, an inverted CCAAT box and a forkhead box O (FOXO)
104 Chapter 4 binding site were all well conserved among these species (Figure 53). I built a
phylogenic tree based on the similarity of sequences covering the two ChoREs, and
the tree fitted well with the evolutionary tree (fishÆamphibianÆmammals; Figure
54).
In fish (fugu, tetraodon, zebrafish and medaka) Txnip promoters, the
nucleotide sequences at the ChoRE‐a position actually deviate from a canonical
ChoRE; on the contrary, the fish ChoRE‐b better mimics a canonical ChoRE sequence
than does human ChoRE‐b (Figure 55). The sequences of both the ChoRE sites on the frog Txnip promoter are moderately degenerate thus lying between fishes and mammals (Figures 53 and 55).
Figure 55. Sequence alignment of fish and frog Txnip promoters with the human Txnip promoter. Sequences corresponding to ChoRE‐a in fish Txnip promoters are not a good ChoRE (as indicated by red boxes). Sequences corresponding to ChoRE‐b in fish Txnip promoters are more similar to the canonical ChoRE (green box with dotted lines).
4.2.1.3 The MondoA/MLX Complex Binds to Both ChoREs in vitro
It was shown previously that the earlier defined ChoRE (i.e., ChoRE‐a) recruited a protein complex of MondoA and MLX; thus I sought to know whether
the newly identified ChoRE (ChoRE‐b) could also function as an anchor for MondoA
and MLX.
105 Chapter 4
A
B
Figure 56. EMSAs using two Txnip promoter segments containing ChoRE‐a or ChoRE‐b. (A) The formation of DNA‐protein complexes were competed with different cold probes. (B) The effect of different antibodies on the migration of DNA‐ protein complexes. Refer to text for details. N.S., non‐specific.
106 Chapter 4
I performed electrophoresis mobility shift assays (EMSAs; Figure 56) using
33P‐end‐labeled probes and whole cell extract prepared from cells over‐expressing
ectopic HA‐MondoA and Myc‐MLX. Similar to a probe containing the ChoRE‐a and
CCAAT box (Figure 56A, lane 2), a probe containing the ChoRE‐b and inverted
CCAAT box formed multiple DNA‐protein complexes (lane 10); the formation of these complexes were significantly reduced by molar excess of unlabeled probes (WT; lanes 3 and 11). When molar excess of unlabeled probes of mutated ChoRE sites
were used, two bands were preserved (lanes 4 and 12); the higher band (star) contained MondoA and MLX (see below), and the lower one might be due to an unknown ChoRE‐ or E‐box‐binding protein(s). The two bands were competed by shorter cold probes containing ChoREs but lacking the CCAAT boxes (sCho‐a or sCho‐b; lanes 7 and 15) but not by similar probes containing mutated ChoREs
(msCho‐a or msCho‐b; lanes 8 and 16). When cold probes with mutations at the
CCAAT box (mCAT) or the inverted CCAAT box (miCAT) were used in competition assays, two bands were preserved (arrows, lanes 5 and 13); these two bands were significantly competed by a short cold probe containing ATTGG sequences (sNFY, lanes 6 and 14). These results suggest that both labeled probes (CAT/ChoRE‐a and iCAT/ChoRE‐b) can at least form three major DNA‐protein complexes, one (star) is formed between ChoRE and the MondoA/MLX complex, and the other two (arrows) are formed between CCAAT boxes (CCAAT or inverted CCAAT) and their associated factors.
To obtain identities of the protein(s) present in these complexes, I included antibodies in EMSAs (Figure 56B). The band corresponding to ChoRE (single star) was abolished by anti‐HA antibodies (lanes 4 and 12), and super‐shifted by anti‐Myc
107 Chapter 4
(lanes 5 and 13; double stars) or anti‐MLX (lanes 8 and 16; triple stars) antibodies.
The bands corresponding to CCAAT or inverted CCAAT boxes (arrows) were super‐
shifted by anti‐NF‐YA antibodies (lanes 6 and 14; double head arrow). Naive mouse
or rabbit IgG (mIgG or rIgG) as a control did not change the EMSA patterns. I
conclude that HA‐MondoA and Myc‐MLX are able to interact with both ChoRE‐a
and ChoRE‐b, and that NF‐Y can bind to both the CCAAT and inverted CCAAT
boxes, on the Txnip promoter.
I emphasize that the expression of endogenous MondoA and MLX is low. To
increase the EMSA sensitivity, I used extracts from cells over‐expressing epitope‐
tagged HA‐ModoA and Myc‐MLX for detection by epitope‐specific antibodies.
4.2.1.4 Both ChoREs Are Required for Optimal Txnip Promoter Activity
To confirm whether both ChoREs are involved in the induction of Txnip
expression by glucose or adenosine‐containing molecules, I have generated reporter
genes driven by Txnip promoters containing mutations at ChoRE‐a (chapter 3),
ChoRE‐b or both (dmChoRE, refer to Figure 52 for mutated nucleotide sequences),
and tested their response to adenosine‐containing molecules or glucose. As shown
(Figure 57A and 57B), when either ChoRE or both ChoREs were mutated, the
reporter genes were no longer induced by NAD+. For the glucose response test, the activities of promoters with mutated ChoRE‐a or with double mutations at both
ChoREs were not stimulated by glucose; and as compared with the wild type promoter, the response of the ChoRE‐b mutant promoter was significantly reduced
(Figure 57C and 57D). These results suggest that both ChoREs are required for optimal induction of Txnip expression by adenosine‐containing molecules or glucose.
108 Chapter 4
A B 0.2 4 ** - + + NAD NAD 3
0.1 2 * ** * ** ** 1 Promoter Activity Promoter
Fold(s) ofInduction 0.0 0 E T WT R W ho oRE-b hoRE h C ChoRE-a m mChoRE-bdmC mChoRE-amC dm
C D 10 10 ** - Glc 8 Glc 8 6 6 ** 4 4 * 2 2 Promoter Activity * ** ** Fold(s) ofInduction 0 0
-b E WT E WT R R ho ho ChoRE-a ChoRE-a m mC dmChoRE m mChoRE-bdmC
Figure 57. Responses of wild‐type (WT) or ChoRE mutant Txnip promoters to NAD+ or glucose. Response to NAD+ (in U2OS cells, A and B); response to glucose (in L6 cells, C and D). *: significantly induced by NAD+ (A) or glucose (C); significantly different from wild‐type promoter (B and D).
Given that the mChoRE‐a mutant Txnip promoter possesses a more dramatically reduced response to glucose than the mChoRE‐b mutant counterpart,
ChoRE‐a could play a major role in the glucose sensing (also see discussion). On the other hand, responses of these two promoters to NAD+ were similar in U2OS cells.
This may be due to a unique feature of the cell line used in the assay system. Indeed, for NAD+ response, the reduced response of the mChoRE‐a mutant Txnip promoter was to a larger degree than that of the mChoRE‐b mutant counterpart in HeLa cells
109 Chapter 4
(data not shown; also refer to Figure 50). Thus, in general, ChoRE‐a may well have higher efficacies than ChoRE‐b in sensing extracellular glucose or adenosine‐ containing molecules.
4 * * - NAD+ 3
2
Promoter Activity 1
0
R A R A R A R A do do do do MLXn MLXn MLXn MLXn Mo Mo Mo Mo WT mChoRE-b mChoRE-a dmChoRE
Figure 58. Effect of siRNAs against MLX or MondoA on Txnip promoter activities. siRNAs against MLX or MondoA down‐regulated the wild‐type but not the ChoRE mutant Txnip promoters and abolished the stimulatory effect of NAD+ on Txnip promoters in U2OS cells. Expression levels of MondoA or MLX were knocked down by specific siRNAs. R: random (control) siRNA. Asterisk: comparison between NAD+ treated and untreated samples; diamond: comparison between basal activities (without NAD+); n=2.
Previously, I have shown that knock‐down of MondoA or MLX expression by siRNAs in U2OS cells reduced the basal activity of the Txnip promoter and abolished the induction by adenosine‐containing molecules; for a promoter with mutation at
ChoRE‐a, the effect of MondoA or MLX siRNAs was diminished (Chapter 3, Figure
19). In this study, I found that the activities of promoters with mutations at either
110 Chapter 4
ChoRE or both ChoREs were largely unaffected by siRNAs targeting MondoA or
MLX (Figure 58). Thus, both the ChoRE motifs are required for induction of Txnip
promoter by adenosine‐containing molecules, which act most likely in a
MondoA/MLX dependent manner.
4.2.1.5 ChoREs Are Not Sufficient for the Induction of Txnip Expression
Both ChoREs are required for optimal Txnip induction by adenosine‐
containing molecules or glucose (Figures 57 and 58), and the promoter sequences
spanning from ‐184 to ‐63 appear to contain all the information necessary to mediate
the induction (Figure 51); however, contribution(s) of other sequences lying between
the two ChoREs could not be excluded, thus I asked whether the two motifs are
sufficient for this stimulatory process. When nucleotide sequences between the two
ChoREs were shuffled (Shuffle, Figure 59A); the basal activity of this mutant
promoter was dramatically decreased, and the promoter no longer responded to
NAD+ (Figure 59C) or glucose (Figure 59D). I also engineered tandem ChoRE‐a or
ChoRE‐b (dChoRE‐a or dChoRE‐b), or linked ChoRE‐a and ChoRE‐b (ChoRE‐ab), into the TATA box only luciferase reporter construct (Figure 59A); the activities of these chimerical promoters were not induced by NAD+ (Figure 59B). Hence, the two
ChoREs are not sufficient for optimal induction of the Txnip expression by adenosine‐containing molecules or glucose, and additional regulatory elements are
required.
4.2.1.6 Tandem NF‐Y Binding Sites Are Required for the Induction of Txnip
Expression
The nucleotides between two ChoREs contain a CCAAT box, an inverted
CCAAT box and a FOXO binding site (Figure 52); these three sites are highly
111 Chapter 4 conserved in Txnip promoters from diverse species (Figure 53). In EMSAs, bands corresponding to the CCAAT box or inverted CCAAT box were super‐shifted by NF‐
YA antibodies (Figure 56), suggesting that both the CCAAT and inverted CCAAT boxes are potential binding sites for NF‐Y.
A C
4 + - NAD ** 3
2 1 Activity Promoter 0 T e fl W Shuf
B D 8 4 - Glc 7 ** - 6 ** 3 NAD+ 5 2 4 3 1 2 Promoter Activity Promoter Promoter Activity Promoter 1 0 0 -a WT E tor ec WT hoR V uffle C hoRE-ab d dChoRE-bC Sh
Figure 59. Response of hybrid Txnip promoters to NAD+ or glucose. (A) Fusion of a TATA‐driven luciferase reporter with synthetic Txnip promoters. Shuffle, nucleotides between two ChoREs were scrambled; other promoters contain one or two ChoREs. (B) The activity of the wild‐type Txnip promoter was, but Txnip promoters containing tandem ChoRE‐a, ‐b, or linked ChoRE‐ab were not, induced by NAD+. When nucleotides between the two ChoREs were shuffled, the promoter (Shuffle) was not induced by NAD+ (C) or glucose (D).
112 Chapter 4
A B 1.0 6 - + * 5 NAD 0.8 NAD+ 4 0.6 * 3 0.4 2 * 0.2 1 Promoter Activity Promoter *
Fold(s) ofInduction ** 0.0 0 T T T T T W W iCA CA CA m mCAT C m dm D miCAT dmCAT
7 ** 8 6 - Glc 5 Glc 6 4 4 3 * 2 * 2 * Promoter Activity Promoter 1 ** Induction of Fold(s) 0 0 T WT AT W C CAT CAT mCAT m m mi d miCAT dmCAT
Figure 60. Responses of wild‐type (WT) or CCAAT box mutant Txnip promoters to NAD+ or glucose. Response to NAD+(in U2OS cells, A and B); response to glucose (in L6 cells, C and D). Asterisk: significantly induced by NAD+ (A) or glucose (C); significantly different from wild‐type promoter (B and D); diamond: basal promoter activities were significantly different from WT promoter; n=2.
It was previously shown that mutated FOXO binding site did not have significant effect on basal or NAD+‐induced Txnip promoter activity (Chapter 3).
Here, I introduced mutations into the CCAAT box, inverted CCAAT box or both
(refer to Figure 52), and their response to NAD+ or glucose was tested. When either
CCAAT box was mutated, the basal activity of the mutant Txnip promoter was decreased, and the response of the mutant promoter to NAD+ was only slightly repressed; the promoter with mutations at both CCAAT boxes was found to have a
113 Chapter 4 diminished basal activity as well as an impeded induction by NAD+ (Figure 60A and
60B). I also examined effects of glucose on these promoters; promoters with single mutations showed retarded glucose response, and the double mutant promoter exhibited marginal response (Figure 60C and 60D). Thus, CCAAT boxes are not only
important for maintaining a basal level transcription, but also required for the induction of Txnip expression by adenosine‐containing molecules or glucose.
4.2.1.7 NF‐Y Mediated Induction of Txnip Expression by SAHA Requires
MondoA/MLX
Histone acetyltransferases usually positively, while histone deacetylases
negatively, impact gene transcription (reviewed in Kouzarides, 2007). SAHA, an
HDAC inhibitor, was earlier shown to induce Txnip expression, and this induction is
mediated by NF‐Y and the inverted CCAAT box (Butler et al., 2002). Given that the
CCAAT box is bound essentially by the same factor(s) as the inverted CCAAT box, at
least in vitro (Figure 56), I asked whether this motif was also important for the Txnip expression induction by SAHA. As seen (Figure 61A and 61B), the mutated CCAAT
box rendered the promoter to respond to SAHA in a significantly compromised manner, as the promoter with a mutated inverted CCAAT box; when both CCAAT boxes were mutated, the promoter was no longer sensitive to SAHA. Thus, both
CCAAT boxes and NF‐Y are important for mediating the induction of Txnip expression by SAHA. I also tested the effect of SAHA on Txnip promoters with mutated ChoREs, and observed similar phenotypes as I did with the CAATT box mutants (Figure 61A and 61B). These observations suggest that both ChoREs and the
CCAAT boxes are critical for the induction of Txnip expression by SAHA.
114 Chapter 4
A 20 - 15 SAHA
10
5
Promoter Activity 0 T T E ec WT A AT A V oR iC C h m m dmC ChoRE-a mChoRE-bm dmC
B 6 5 ** SAHA
4 ** ** ** ** 3 2 * 1 Fold(s) ofIndcution 0 T T E ec WT A AT A V oR iC C h m m dmC ChoRE-a mChoRE-bm dmC
C 125 l * 100
75
50
25
Txnip mRNA Leve mRNA Txnip * 0 - Glc SAHA Glc SAHA
Figure 61. The effect of SAHA. (A) responses of Txnip promoters to SAHA. (B) induction folds; asterisk: significantly different from empty vectors. (C) Txnip mRNA levels in HeLa cells that were incubated in glucose‐free medium overnight, and then treated with 10 mM glucose, or 5 μM of SAHA, or both, for 4 hrs. Asterisk: significantly different from the untreated sample; diamond: significantly induced by SAHA; n=2.
115 Chapter 4
As ChoRE is a glucose responding element, I tested if the Txnip expression
induction by SAHA was glucose dependent. In the absence of glucose, SAHA alone
exhibited little effect on Txnip mRNA level; when glucose was added back to the
culture medium, Txnip expression was induced by glucose, which was further
induced by SAHA dramatically (Figure 61C). Another HDAC inhibitor, trichostatin
A (TSA), exhibited similar glucose requirement in inducing Txnip expression (data
not shown). Therefore, a glucose induced signaling is required for the induction of
Txnip expression by HDAC inhibitors; and HDAC inhibitors and glucose may induce Txnip expression in a synergistic manner. It seems that the tandem CCAAT boxes and ChoREs on Txnip promoter are mutually required to optimally induce the
Txnip expression in response to glucose, adenosine‐containing molecules or HDAC
inhibitors.
4.2.1.8 Txnip Promoter Recruits MondoA/MLX Complex in an NF‐Y Dependent
Manner
Results from promoter activity assays suggest that both NF‐Y and
MondoA/MLX complex are required for optimal induction of Txnip expression by
glucose or adenosine‐containing molecules. To confirm this, I performed chromatin
immuno‐precipitation (ChIP) assays to determine the occupancy of NF‐Y and MLX
on wild‐type or mutant Txnip promoters that were parts of integrated trans‐genes in stable cell lines (see “Materials and Methods” and Figure 62A for details). The integration of foreign Txnip promoters into chromosome were confirmed by PCR
using the targeting vector (pEGFP‐1) specific primers a and b (Figure 62B) and
corresponding genomic DNA. A very critical step in ChIP assays is the preparation
of chromatin DNA fragments; in this study, ~500 bp fragments were prepared by
116 Chapter 4 sonication from different cell lines, Figure 62C illustrates a typical example.
Figure 62. Stable cell lines for ChIP assays. (A) Txnip promoters (wild‐type or mutant) were inserted into pEGFP‐1 vectors (Xho I and BamH I sites); these constructs were used to generate stable cell lines. Primers a and b (or d) amplify the chromosomally‐integrated foreign Txnip promoters, and primers c and d amplify the endogenous Txnip promoter. (B) Confirmation of chromosomal integration of foreign Txnip promoters. Genomic DNA was prepared from indicated stable cell lines and PCR‐amplified using primers a and b. (C) Total lysates of indicated cell lines were sheered by sonication to generate chromatic DNA fragments at ~500 bp, and 100 bp DNA ladders were used as markers.
In ChIP assays, MLX was recruited normally to the wild type promoter,
efficiently recruited (albeit with a slightly reduced efficacy) to the ChoRE‐b mutant promoter, and recruited to the ChoRE‐a mutant promoter with a reduced efficacy; when both ChoREs were mutated, the MLX recruitment was completely impeded
(Figure 63A). These results suggest that both ChoREs are able to recruit the
MondoA/MLX complex, with the proximal ChoRE (ChoRE‐a) being a predominant binding site. This might be in line with the severities of the mutations in the promoter functional analyses, especially regarding the glucose response (Figure 57).
117 Chapter 4
Figure 63. Recruitment of MLX and NF‐Y to respective targets (ChoREs and CCAAT boxes) assessed by ChIP assays. Stable cell lines with integrated WT or mutant Txnip promoters were subjected to ChIP assays using different antibodies. Recovered pulled‐down DNA was analyzed by primers a and d (for foreign Txnip promoters), c and d (for the endogenous Txnip promoter), or primers for a negative control chromosomal region.
The wild‐type Txnip promoter was precipitated by either NF‐YA or NF‐YB antibodies; Txnip promoters with single mutation at the CCAAT box or the inverted
CCAAT box were also precipitated by NF‐Y antibodies although the signals were reduced; however, the Txnip promoter with mutations at both CCAAT boxes was not pulled‐down by NF‐Y antibodies (Figure 63B). This confirms that both CCAAT boxes on Txnip promoter are targets of NF‐Y, and both sites are required for optimal recruitment of NF‐Y onto the Txnip promoter.
In all cell lines tested, the endogenous Txnip promoters were similarly recruited by MLX or NF‐Y antibodies but not naive IgGs (Figure 63A and 63B). A control region on chromosome 12 was not precipitated by every tested batch of
118 Chapter 4 antibodies (Figure 63A and 63B). This suggests that the precipitation of protein‐
DNA complexes by these antibodies in ChIP assays is efficient and specific.
Figure 64. MLX is not recruited onto CCAAT boxes‐mutated Txnip promoter. Stable cell lines with integrated WT or mutant Txnip promoters were subjected to ChIP assays using indicated antibodies. Recovered DNA was analyzed by primers targeting foreign Txnip promoters (A), the endogenous Txnip promoter (B) or a negative control region (C). The presence of the empty vector (pEGFP1) sequence was confirmed using primers a and b.
NF‐YA and NF‐YB were recruited onto the double ChoRE mutant (dmChoRE)
promoters with similar efficacy as the wild‐type promoter (Figure 64A). Thus the
interaction of NF‐Y with CCAAT boxes is not dependent on ChoREs and their
associated factors. Interestingly, the interaction of MLX with the double CCAAT
boxes mutated promoter (dmCAT) was dramatically reduced (Figure 64A). As
controls, the occupancy of MLX or NF‐Y on endogenous Txnip promoters was
similar in these cell lines (Figure 64B); and MLX and NF‐Y did not interact DNA
sequences located on a chromosome separated from the endogenous Txnip gene
(Figure 64C).
119 Chapter 4
The above experiments were performed using cells maintained in complete
DMEM with 1 g/L glucose; we also tested the Txnip promoter occupancies by MLX or NF‐Y in cells incubated in glucose‐free medium supplemented without or with 2‐ deoxy‐D‐glucose (2DG), which sustains a robust glucose signaling (Figure 38). As shown (Figure 65), NF‐YA was recruited to the integrated (i) wild type or the endogenous (e) Txnip promoters at similar efficacies under both the glucose‐free and
2DG‐treated conditions, but not to the double CCAAT box‐mutated (dmCAT) Txnip promoter, suggesting that the CCAAT box binding by NF‐Y was not sensitive to glucose. On the other hand, under glucose‐free condition, only residue amount of
MLX was bound with the Txnip promoters (WT or dmCAT) and, in cells treated with
2DG, the MLX recruitment to the integrated wild type and the endogenous Txnip promoters was dramatically increased; however, the MLX recruitment to the dmCAT promoter was significantly repressed. Thus, mutations at the CCAAT boxes can effectively disrupt the interaction of cognate factors with the Txnip promoter, and the recruitment of NF‐Y to Txnip promoter is probably a prerequisite for the recruitment of MondoA/MLX complex.
Suggesting that a chromosomally‐integrated transfected promoter can function similarly as the endogenous counterpart (Txnip promoter in response to glucose), the occupancies of the integrated wild type promoter by MLX or NF‐Y were similar to that of the endogenous Txnip promoter in stable cell lines (Figure 65A), and similar to that of the endogenous Txnip promoter in normal HeLa cells that have undergone no transfection and G418 selection (Figure 65B).
120 Chapter 4
Figure 65. The interaction between MLX and ChoRE is glucose dependent. Stable cell lines (A; WT or dmCAT) or normal HeLa cells (B) were incubated in glucose‐free medium overnight, and treated with or without 10 mM 2DG for 2 hrs. Cells were then subjected to ChIP assays. The recovered DNA was analyzed by primers targeting integrated (i) WT or dmCAT Txnip promoter, endogenous (e) promoter or negative control region (Neg.).
4.3 The Role of USFs in Txnip Expression
4.3.1 Expression and Purification of His‐tagged USF1
Figure 66. Bacterial expression of USF1. Lane 1, size markers; lanes 2‐3, lysates from cells without (‐) or with IPTG (+) induction; lanes 4‐9, a test for conditions of elution with different imidazole levels; pooled proteins (lanes 7‐9) were re‐purified (lanes 10‐ 13) using empirically determined imidazole levels for the wash and elution to obtain homogenous His‐tagged USF 1 (pooled protein in lanes 12‐13).
ChoRE is defined as two E‐boxes (CACGTG) separated by five nucleotides
(Ma et al., 2006). E‐box can recruit several transcription factors including USF1 and
121 Chapter 4
USF2. To study the role of USFs in Txnip expression, I firstly tested if USFs could
bind to ChoREs on the Txnip promoter in vitro. To this end, the human USF1 coding
sequence (long isoform) was inserted into pET vector for expressing Histine (His) tagged recombinant USF1. Bacterially expressed USF1 was affinity‐purified using
Ni‐NTA agarose beads; pooled fractions (lanes 12 and 13; Figure 66) were used for
EMSAs with labeled ChoRE‐containing Txnip promoter DNA fragments as probes.
4.3.2 USF1 Interacts with ChoRE Sites in vitro
Figure 67. His‐USF1 interacts with ChoRE‐a or ChoRE‐b. Purified His‐USF1 and labeled EM‐a, EM‐b or E‐box probes were mixed; and BSA was used as negative control (lanes 1, 8 and 15). Competition was performed by inclusion of 100‐fold molar excess cold probes in reactions. The position of the major USF1‐DNA complexes was indicated by an arrow.
ChoRE oligonucleotides (EM‐a or EM‐b, respectively containing either
ChoRE‐a or ChoR‐b; see Table 1 in Chapter 2 for sequence information) were labeled
and used as probes with a perfect E‐box‐containing probe included as a control in
122 Chapter 4
EMSAs. His‐USF1 was incubated with labeled probes without or with competing cold probes. As shown in Figure 67, in the presence of His‐USF1, a major protein‐
DNA complex was formed (lanes 2 and 9). The His‐USF1/EM‐a complex was competed by the cold probes containing a perfect E‐box (Ebox) or wild type ChoRE‐a site (EM‐a), partially competed by a cold ChoRE‐b site (EM‐b) probe, but not by the cold probe containing a mutated ChoRE‐a site (mEM‐a). Similarly, His‐USF1/EM‐b complex was competed by cold probes containing a perfect E‐box, ChoRE‐a site or
ChoRE‐b site, but not by a cold probe containing the mutated ChoRE‐b site (mEM‐b).
These results indicate that both ChoRE sites can be bound by USF1, although the affinity seems to be weaker as compared to the interaction between USF1 and a perfect E‐box (lane 16).
I also performed EMSAs using nuclear extracts (NE). NE was prepared from cells grown in glucose free medium or cells treated with 10 mM 2DG or NAD+.
Labeled probes (EM‐a or EM‐b) were mixed with NE or His‐USF1 with or without
cold probes to analyzed the protein‐DNA complex formation (Figure 68). With NE, a
dominant complex (black arrow) that migrated slightly slower than the His‐USF1
containing complex was detected (lanes 3 and 10), which was competed by cold
probes containing the wild type but not the mutant ChoREs (a or b; lanes 6, 7, 13, 14).
These results suggest that there is a protein(s) in NE that interacts specifically
with the Txnip promoter ChoRE sequences. The unknown protein could be a hetero‐ dimmer of USF1 and USF2. USF2 is larger that USF1, so the complex migrated
slower than the complex formed by His‐USF1 homo‐dimmer. USF1 and USF2
undergo post‐translational modifications in mammalian cells; this may also
contribute to the slow mobility in EMSA. In EMSA, the formation of DNA‐protein
123 Chapter 4 complexes was similar when using NE prepared from cells treated with 2DG or
NAD+ (Figure 68, lanes 3‐5 and 10‐12), suggesting that the interaction between USF1 and ChoREs was not modulated by 2DG or NAD+.
Figure 68. ChoRE‐containing probes interact with a protein(s) in HeLa nuclear extract (NE). Purified His‐USF1 (20 ng) or HeLa NE (5 μg) was used in EMSA; BSA was used in the negative control lanes (1 and 8). The positions of USF‐DNA complexes were indicated by arrows.
To test whether USF1 and USF2 are proteins in NE that interact with ChoRE‐ containing probes, EMSAs were carried out in the presence of naive IgGs or anti‐
USF1 or ‐USF2 antibodies. USF1 antibodies super‐shifted the band corresponding to the ChoRE‐a/His‐USF1 complex (lane 4, the band at the position of the open circle versus that of the dotted arrow), and naive IgG or USF2 antibodies did not show any effect on this complex; on the other hand, in EMSA performed with NE, the complex
124 Chapter 4
(black arrow) was super‐shifted (lanes 9 and 10, positions indicated by the black circle and diamond) by either USF1 or USF2 antibodies but not the control IgG
(Figure 69). These results indicate that both USF1 and USF2, present in NE, can interact with ChoREs on the Txnip promoter. Therefore, at least in vitro, both
ChoREs are able to interact with USFs.
Figure 69. Bands shift by USF antibodies. Labeled probe (EM‐a) was used in EMSA with BSA, purified His‐USF1 (20 ng) or HeLa NE (5 μg). The effects of control IgGs, anti‐USF1 or ‐USF2 antibodies on the mobility of USF‐DNA complexes (arrows) were tested. The open circle, dot or diamond indicates positions of super‐shifted bands.
4.3.3 Down‐regulation of USFs Reduces Txnip Expression
To explore potential functions of USFs on Txnip expression, I tested the effects of their expression silencing by way of RNA interference using siRNAs specifically targeting USF1 or USF2. As shown in Figure 70, in U2OS cells, the protein levels of USF1 or USF2 were effectively down‐regulated by siRNAs against,
125 Chapter 4 respectively, USF1 or USF2. When the USF1 expression was silenced, the Txnip
mRNA level was reduced significantly; however, the effect of USF2 siRNA was not
significant (Figure 70A). The induction fold of Txnip expression by NAD+ or ATP
was affected by neither siRNA (Figure 70B and 70C).
A
B 5 C 6 - ATP 4 * 5 + * ATP NAD + 4 3 NAD * 3 2 2 1 * 1 Txnip mRNA Level mRNA Txnip * Fold(s) of Induction 0 0 Random siUSF1 siUSF2 Random siUSF1 siUSF2
D 2.0 E 6 - * 5 1.5 NAD+ 4 1.0 * 3 2 0.5 Promoter Activity Promoter 1 * Induction of Fold(s) 0.0 0 Random siUSF1 siUSF2 Random siUSF1 siUSF2
Figure 70. The effects of USF1‐ or USF2‐specific siRNAs on Txnip expression. (A) down‐regulation of USF1 or USF2 protein level by siRNAs against USF1 or USF2 respectively. (B) and (D), the effects of the indicated siRNAs on the mRNA level or promoter activity of Txnip, with or without adenosine‐containing molecules. (C) and (E), induction fold(s) by adenosine‐containing molecules, respectively corresponding to (B) and (D). Asterisk: significantly different from untreated samples; Diamond: Significantly different from random siRNA treated samples.
126 Chapter 4
The effects of USF siRNAs on the activity of Txnip promoter were also tested.
The siRNA against USF1 dramatically reduced Txnip promoter activity, and siRNA against USF2 also reduced the activity, but less potently; the Txnip promoter activity remains inducible by NAD+ under either siRNA treatment (Figure 70D and 70E).
Taken together, an overall USF level is important for basal level Txnip expression;
however, neither USF1 nor USF2 is critical for the induction by adenosine‐containing
molecules.
4.3.4 Over‐expression of USFs Induces Txnip Promoter Activity
The effect of over‐expression of USF on Txnip expression has also been
studied. When HA‐tagged USF1 or USF2 was over‐expressed in U2OS cells, the
Txnip promoter activity was induced in a dose‐dependent manner; either USF when
over‐expressed can induce the Txnip promoter activity up to >20‐fold (Figure 71).
Figure 71. The induction of Txnip promoter activity by over‐expression of USFs. The amount of HA‐tagged USF expression vector used were 0, 1/16, 1/4 or 1 μg per 6‐ well. The expression of HA‐USF was shown by immuno‐blots using anti‐HA antibodies; and GAPDH levels were used as controls.
127 Chapter 4
Next, I tested the effect of USF over‐expression on the responsiveness of
Txnip promoter to adenosine‐containing molecules. When HA‐tagged USF1 or USF2 was over‐expressed alone or in combination, the activity of Txnip promoter was dramatically induced; however, the Txnip promoter was not responsive to NAD+ or
ATP (Figure 72A and 72B). It is possible that the Txnip promoter has already been
maximally configured for transcription when USFs were over‐expression, and the expression is quite high such that there is no space for further induction by
adenosine‐containing molecules.
A
B 30 - + ATP NAD 25 20 15 10
Promoter Activity 5 * * 0 - USF1 USF2 USF1/2
Figure 72. The response of Txnip promoter to NAD+ or ATP under USF over‐ expression. (A) Over‐expression of HA‐tagged USF1 and/or USF2. (B) the effect of USF over‐expression on Txnip promoter activity, with or without of NAD+ or ATP. Asterisk: significantly different from untreated samples; Diamond: Significantly different from vector transfected samples.
128 Chapter 4
4.3.5 USF Is Not Involved in the Txnip Induction by Adenosine‐containing
Molecules
USFs could interact with ChoREs in vitro, and Txnip expression was affected by modulation of USFs levels. However, it is not clear whether the effect of USF on
Txnip expression is mediated by ChoREs. To address this issue, the effect of USF1 over‐expression on activities of truncated Txnip promoters was examined. As shown
(Figure 73A and 73B), all the tested Txnip promoters were dramatically induced by
USF1 over‐expression, even for the 73 bp long Txnip promoter that does not contain a ChoRE; on the other hand, a TATA only promoter was not significantly induced.
Thus, the ChoREs are not required for the induction of Txnip promoter activity by over‐expression of USFs.
I also tested the effects of USF over‐expression on the activity of the double
ChoRE mutant Txnip promoter. Although this promoter was not sensitive to NAD+ or ATP, it was effectively induced by over‐expression of HA‐USF1, HA‐USF2 or both
(Figure 73C), and the induction was comparable to the wild type Txnip promoter
(Figure 72B). Again, this suggests that the induction of Txnip promoter activity by
USF over‐expression is not dependent on ChoREs, and other element(s) may mediate this induction. Indeed, several loosely defined (cryptic) E‐boxes (CANNTG) on the
Txnip promoter, even between the TATA box and ‐73 positions, were identified; they may function as docking sites for USFs to mediate their stimulatory effect on Txnip expression.
129 Chapter 4
A 30 - USF1 20
10 Luciferase Activity Luciferase 0 269 169 163 122 111 73 TATA Txnip Promoter(s)
B 35 30 25 20 15 10 5 Fold(s) of Induction of Fold(s) 0 269 169 163 122 111 73 TATA Txnip Promoter(s)
30 C - NAD+ ATP 25 20 15 10
Promoter Activity 5 0 - USF1 USF2 USF1/2
Figure 73. Effects of USF over‐expression on the activities of truncated or mutant Txnip promoters. (A) Effects of ectopic expression of HA‐USF1 on activities of truncated Txnip promoters or a TATA only promoter. (B) induction fold(s) corresponding to (A). (C) Effects of over‐expression, of USF1, USF2 or both, on the activity of the double ChoRE mutant Txnip promoter, with or without NAD+ or ATP.
A‐USF (provided by Dr. Charles Vinson at National Cancer Institute,
Maryland) is a dominant negative USF1 mutant; it can form a dimer with either
130 Chapter 4
USF1 or USF2, but the dimer is unable to interact with E‐box thus cannot function as a transcription activator; the effect of A‐USF on Txnip expression has been tested. As shown (Figure 74), over‐expression of HA‐USF1 dramatically induced Txnip expression (lanes 4‐6); however, when A‐USF was co‐expressed, the stimulatory effect of HA‐USF1 was abolished (lane 7). When cells co‐transfected with HA‐USF1 and A‐USF and control cells (transfected with empty vectors) were compared, the basal activity was similar (lanes 1 and 7) and the promoter activity was also similarly induced by NAD+ or ATP (lanes 2 and 3; lanes 8 and 9). Over‐expression of A‐USF alone did not show significant effect on the basal promoter activity, the responsiveness to adenosine‐containing molecules was also intact (lanes 10‐12).
These results indicate that A‐USF was sufficient to abolish the effect of USF over‐ expression on Txnip basal promoter activity, but it cannot abolish the stimulatory effect of adenosine‐containing molecules.
Figure 74. A‐USF did not repress the induction of Txnip promoter activity by NAD+ or ATP. HA‐USF1 (lanes 4‐6) A‐USF1 (lanes 10‐12) or both (lane 7‐9) were over‐expressed in U2OS cells, the basal activity of Txnip promoter and its response to NAD+ or ATP was examined.
131 Chapter 4
4.3.6 USF Is Not Involved in the Txnip Induction by Glucose
The effect of A‐USF on glucose induced Txnip promoter activity was also tested. As shown in Figure 75, glucose induced Txnip promoter activity in L6 cells
(lane 2), and over‐expression of HA‐USF1 also induced Txnip promoter activity (lane
3). When A‐USF was over‐expressed, or when HA‐USF1 and A‐USF were co‐
expressed, the Txnip promoter remained similar basal activity (lanes 5 and 7), but it
was induced by glucose treatment (lanes 6 and 8). These results indicate that A‐USF
can effectively abolished the stimulatory effect of USF1, but not the induction of
Txnip promoter activity by glucose; therefore, I conclude that USF1 is not involved in
the induction of Txnip expression by glucose. I propose that over‐expression of
USF1 may artificially increase the recruitment of USF1 to the Txnip promoter via the
above mentioned cryptic E boxes, hence artificially enhancing the Txnip promoter
activity; however, the occupancy of USFs on the Txnip promoter is apparently not
affected by signaling pathway evoked by adenosine‐containing molecules.
Figure 75. A‐USF did not repress the induction of Txnip promoter activity by glucose. HA‐USF1,A‐USF1 or both were over‐expressed in L6 cells, the basal activity of Txnip promoter and its response to glucose (10 mM) was examined.
132 Chapter 4
4.4 Discussion
4.4.1 Tandem ChoREs on the Txnip Gene Promoter
It has been shown that the induction of Txnip expression by glucose or adenosine‐containing molecules was mediated by the proximal ChoRE and its associated transcription factors MondoA/MLX (Chapter 3; Stoltzman et al., 2008;
Minn et al., 2005b). When testing responses of a series of truncated or mutated Txnip promoters to NAD+ or glucose, I found that the sequence around 170 bp upstream of
the transcription start site was also critical for the induction (Figure 50); this sequence is highly conserved in the Txnip promoters of diverse species and resembles a degenerate ChoRE (Figures 52‐54). In EMSA and reporter assays, this newly identified ChoRE, designated ChoRE‐b, functioned similarly as the previously
reported ChoRE (ChoRE‐a; Figures 56 and 57). In ChIP assays, when ChoRE‐a was mutated, MLX could be recruited to the Txnip promoter presumably via a weak interaction with ChoRE‐b (Figure 63A). Thus, I conclude that there are two ChoREs on the Txnip promoter, in which the ChoRE‐b is also a significant contributor for the induction of Txnip expression by glucose or adenosine‐containing molecules in a
MondoA/MLX dependent fashion.
A study by other colleagues paid attention to a Txnip promoter region around ‐170 and concluded that a potential ChoRE (GAGCACACCGTGTCCACGCG)
was not critical for the stimulatory effect of glucose on Txnip expression (Pang et al.,
2009). Our prevailing view is that the ChoRE‐b sequence around ‐170 should be
GAGCACACCGTGTCCACGCG, especially when the fish Txnip promoters were considered (Figure 55). Therefore, the results and the conclusion in their study might be due to site‐directed mutagenesis that was off‐target, or due to employment of cells
133 Chapter 4 that were less sensitive to a missing function of ChoRE‐b and the associated factors.
In our hands, the response of Txnip promoters lacking a functional ChoRE‐b to glucose was significantly reduced, which was observed in various tested cell lines such as L6 (Figures 50 and 57), HepG2, HeLa and C2C12 cells (data not shown).
I realized, however, that the activities of ectopic Txnip promoters lacking a functional ChoRE‐b were induced by NAD+ in HeLa cells but not in U2OS cells
(Figure 50); this may suggest that the epigenetic statuses of the Txnip promoter may vary in different cell types. The action of an array of adenosine‐containing molecules is on top of the glucose‐induced Txnip expression (Chapter 3), i.e., amplification of glucose signaling. Based on the integrity of glucose signaling, the Txnip promoter in
U2OS cells may thus have a more open chromatin structure than that in HeLa cells.
This may explain a higher potential of the Txnip expression in U2OS cells than that in HeLa and other cells in responding to NAD+ or ATP (Chapter 3, Figure 9). In this scenario, the Txnip promoter in U2OS cells could be more sensitive to missing
functions of the ChoRE‐b and associated MondoA/MLX. Alternatively, signaling
pathway(s), e.g., MAP kinases, evoked by adenosine‐containing molecules in U2OS cells may be different from those evoked in HeLa or other cell types (Chapter 3,
Figures 30 and 31); this may also contribute to the different responses to NAD+ in
U2OS and HeLa cells.
4.4.2 Full Induction of Txnip Expression Requires Both ChoREs and CCAAT Boxes
Within the minimal promoter (184‐63) that supported full induction (Figure
51), two ChoREs, two CCAAT boxes and one FOXO‐binding site have been identified (Figure 52). Among these motifs, the FOXO‐binding site is probably required for mediating the inhibition of Txnip expression by insulin (de Candia et al.,
134 Chapter 4
2008; Parikh et al., 2007; Schulze et al., 2004); the proximal ChoRE (ChoRE‐a) has been shown to be critical for the induction of Txnip expression by glucose or adenosine‐ containing molecules (Chapter 3; Minn et al., 2005b); and the inverted CCAAT box was identified as an element responsible for mediating the stimulatory effect of
HDAC inhibitors, e.g., SAHA (Butler et al., 2002).
Here, however, I have found that both the distal (ChoRE‐b) and the proximal
(ChoRE‐a) ChoRE on Txnip promoter are required for the induction of Txnip
expression by glucose or adenosine‐containing molecules (see above). It is possible
that, for optimal induction of Txnip expression, the two ChoREs might well be
organized in proximity by the interplay of their associated transcription factors
(MLX/MondoA) and cognate co‐activators to stimulate Txnip expression (Figure 76).
This is similar to a model proposed for transcription stimulation of the MCP‐1 gene,
in which two tandem nuclear factor κB (NF‐κB) binding sites recruit NF‐κB proteins and co‐activators p300/CBP upon tumor necrosis factor or lipopolysaccharides stimulation (Leung et al., 2004).
Two tandem ChoREs alone, however, are not sufficient, and two tandem NF‐
Y‐binding sites, i.e., the CCAAT and the inverted CCAAT motifs, are additionally required for optimal Txnip expression induction by NAD+ or glucose (Figures 59 and
60). Txnip promoters with a singly mutated CCAAT box exhibited much reduced
activities but nevertheless retained the induction potentials by NAD+ or glucose; however, when both CCAAT motifs were mutated, the induction potentials were completely abolished (Figure 60). I propose that the NF‐Y‐binding sites are not only critical for maintaining a basal Txnip promoter activity but also involved in the induction of Txnip expression by glucose or adenosine‐containing molecules,
135 Chapter 4 presumably through sustaining certain chromatin status that helps optimize the
induction (Figure 76; and also see below).
In this study, I have identified that the inverted CCAAT box was also
important for the induction of Txnip expression by SAHA; more interestingly, the
response of Txnip promoters with mutations at ChoREs to SAHA was also reduced
(Figure 61). These results taken together suggest that glucose and SAHA may induce
Txnip expression in a synergistic manner, and both CCAAT boxes and ChoREs are
required for optimal induction of the Txnip expression by SAHA.
4.4.3 NF‐Y and MondoA/MLX Cooperate to Stimulate Txnip Expression
The NF‐Y is a complex containing NF‐YA, NF‐YB and NF‐YC subunits, all of
which are required for interaction with the CCAAT motif that is a widely distributed
promoter element in human genome (Ceribelli et al., 2008). NF‐Y can recruit certain
histone acetyltransferases (HATs) such as GCN5 and p300/CBP‐associated factor
(PCAF) onto a target promoter, adding acetylation marks onto histones and
transcription factors (Currie, 1998). These epigenetic marks usually positively
impact target gene transcription (reviewed in Kouzarides, 2007). Consistent with
this notion, the inverted CCAAT box is known to mediate the Txnip expression
induction by SAHA (Butler et al., 2002), which as a histone deacetylase inhibitor can
in principle counter‐act the removal of the above epigenetic marks hence sustaining gene activation.
ChIP results indicate that NF‐Y was recruited onto the Txnip promoter via both CCAAT boxes (Figure 63); this promoter occupation seems to be a prerequisite for the recruitment of the MondoA/MLX complex (Figures 64 and 65). Hence, NF‐Y and MondoA/MLX might coordinate to maximize the Txnip expression. NF‐Y may
136 Chapter 4 occupy the Txnip promoter constitutively, which facilitates the occupancy of
MondoA/MLX in a glucose‐dependent manner. These promoter‐bound factors in
turn recruit co‐activator(s) which set up local epigenetic environment that favors the induction by other signals from, e.g., adenosine‐containing molecules.
Figure 76. A model for the transcriptional regulation of the Txnip gene promoter by NF‐Y, MondoA/MLX and other (co)factors.
In the above scenario, and given that the epigenetic modifications are dynamic, HDAC inhibitors such as SAHA may tip the balance towards histone hyper‐acetylation, which allows the Txnip promoter to be configured for transcription enhancement (Figure 76). SAHA alone is unable to induce Txnip transcription in that its effect on the induction of Txnip expression is abolished in the absence of glucose (Figure 61), suggesting a requirement for additional signals for optimal Txnip expression induction. When glucose is present, the MondoA/MLX complex can be mobilized and bind to ChoREs on the Txnip promoter, which cooperates with the promoter‐bound NF‐Y to create an optimally configured Txnip promoter that dramatically induces Txnip transcription (Figures 61 and 76).
It has been shown that NF‐Y can also cooperate with several other transcription factors, such as USF1/2 (Zhu et al., 2003), Oct1 (Kam et al., 2005), GATA‐
137 Chapter 4
1 (Huang et al., 2004), SP‐1 (Yamada et al., 2000), ATF‐2 (Alonso et al., 1996),
Regulatory factor X (RFX) (Villard et al., 2000) and sterol regulatory element binding protein‐1 (SREBP1) (Ericsson et al., 1996) to regulate transcription of target genes under nutritional, hormonal or immunological challenges. For instance, the promoter of another metabolic related gene, which encodes the type II Hexokinase
(HKII), contains a CCAAT box, an inverted CCAAT box and multiple GC boxes that are bound by the SP family transcription factors (Lee and Pedersen, 2003). CCAAT box‐containing promoters might employ a common regulatory mechanism in which
NF‐Y might function as a regulator that facilitates the function of other gene specific transcription factors to induce gene transcription as illustrated in Figure 76. On the more distal Txnip promoter region, there is a vitamin D response element and a glucocorticoid response element (Wang et al., 2006; Butler et al., 2002); it would be of high interest to study if NF‐Y can also cooperate with vitamin D or glucocorticoid receptor to mediate vitamin D or glucocorticoid induced Txnip expression.
4.4.4 The Role of USFs on Txnip Expression
Recently, Txnip or USF1 gene has been linked to FCHL‐like phenotype in mice or human, and it seems that USF1 is most likely the gene associated with FCHL
(Bodnar et al., 2002; Coon et al., 2004; Pajukanta et al., 2004; Coon et al., 2005; Huertas‐
Vazquez et al., 2005; Lee et al., 2007). However, there are multiple E‐boxes on Txnip promoter, and these E‐boxes are candidate USF‐binding sites, suggesting a potential regulatory role of USF on Txnip expression. In this case, the phenotype caused by
USF1 (deficiency) may partially contribute to deregulated Txnip expression.
In this study, I have investigated the role of USFs on Txnip expression.
Clearly, Txnip expression was affected by modulation of USF levels; siRNAs against
138 Chapter 4
USF1 or USF2 repressed the mRNA level and Txnip promoter activity (Figure 70);
over expression of ectopic USF1 or USF2 can dramatically induce Txnip expression
(Figures 71‐73). As USFs interact with DNA via E‐box that is a component of ChoRE,
the stimulatory effect of USF on Txnip expression might be mediated by ChoREs. In
EMSA, both USF1 and USF2 interact with ChoREs on Txnip promoter (Figures 67‐69),
suggesting that the ChoREs are potential cis‐regulatory elements for USF proteins.
However, I found that the stimulatory effect of USFs on Txnip promoter activity
persisted even when ChoREs were mutated or deleted (Figure 73). Hence, USFs
probably regulates the Txnip expression through some other (cryptic) E‐boxes on the
Txnip promoter. A dominant negative USF1, A‐USF, completely abolished the
induction of Txnip promoter activity by over‐expression of USF1 or USF2; however,
it did not repress the stimulatory effect of adenosine‐containing molecules or glucose
on Txnip expression. Thus, USF proteins are not likely involved in the induction of
Txnip expression by adenosine‐containing molecules or glucose.
USF proteins are widely expressed in different tissues, and they can regulate
expression of many genes as E‐box is an element that frequently exists in human
genome (Rada‐Iglesias et al., 2008). Due to the complexity of USF‐related gene
regulatory network, it is hard to pinpoint downstream gene(s) responsible for
mediating the role of USF1 in the development of FCHL. There are multiple E‐boxes,
other than those in ChoREs, on the Txnip promoter; possibly USF proteins can be
recruited to the Txnip promoter and maintain a basal level expression of Txnip.
However, the regulatory role of USF proteins on the Txnip expression is not
responsive to some nutritional factors, e.g., glucose and adenosine‐containing
molecules. Thus, Txnip may not be involved in USF1‐linked FCHL development,
139 Chapter 4 despite that Txnip plays an important role in glucose or lipid metabolism.
4.5 Conclusion and Perspectives
In the detailed promoter analysis, I have determined the minimal promoter
(nucleotide sequences between ‐184 to ‐63 [bps to transcription start site]) required to
mediating the stimulatory effect of adenosine‐containing molecules or glucose on
Txnip expression. Within the minimal promoter, an additional ChoRE has been
identified; this newly identified ChoRE (ChoRE‐b) was able to recruit MondoA/MLX both in vitro (EMSA) and in vivo (ChIP), and was required along with the previously identified ChoRE (ChoRE‐a) to support optimal induction of Txnip expression by
adenosine‐containing molecules or glucose.
In majority mammalian cells, two ChoREs function cooperatively to mediate
optimal stimulation of Txnip expression by glucose or adenosine‐containing
molecules. In this stimulatory process, ChoRE‐a may play a predominant role in response to glucose or adenosine‐containing molecules, while ChoRE‐b exhibits a lower efficacy; however, in fishes, ChoRE‐b better resembles a canonical ChoRE, and might be the major element in sensing glucose and adenosine‐containing molecules to induce Txnip expression. The sequences of ChoRE‐a in fishes, on the other hand, drastically deviate from the canonical ChoRE, and may not be able to recruit
MondoA/MLX and induce Txnip expression. This also implies that, from an evolutionary point of view, mammals have adapted to use ChoRE‐a, and lower vertebrates prefer ChoRE‐b, as the predominant regulatory element.
Two ChoREs were required, but not sufficient, for mediating the stimulatory effect of glucose or adenosine‐containing molecules on Txnip expression, and two
NF‐Y‐binding sites (the CCAAT box and the inverted CCAAT box) were also
140 Chapter 4 required for this stimulatory process. I also found that the function of ChoREs and associated factors is contingent on tandem CCAAT motifs, in that the occupancy of the Txnip promoter by the NF‐Y (via CCAAT boxes) is a prerequisite for efficacious recruitment of MondoA/MLX to ChoREs. Such a strategy suggests a synergy between NF‐Y and MondoA/MLX in enhancing Txnip expression presumably through inducing dynamic chromatin changes in response to diverse physiological inducers.
When cells were incubated with glucose‐free medium, NF‐Y was able to occupy Txnip promoter, and this occupancy was not affected by supplying glucose
to cells. Thus, NF‐Y may constitutively occupy Txnip promoter via interaction with
CCAAT boxes. But MondoA/MLX occupy the Txnip promoter in a glucose‐sensitive manner: without glucose, only residual amount MLX was bound with the Txnip
promoter, and when glucose was supplied, MLX was massively enriched on Txnip
promoter via interaction with ChoREs. Therefore, NF‐Y can bind with Txnip
promoter and maintain a basal level transcription, and the binding of NF‐Y on Txnip
promoter also make the promoter sensitive to glucose and competent for recruitment
of MondoA/MLX under glucose treatment.
In this study, I also discovered that SAHA‐induced Txnip expression was
glucose‐dependent; both CCAAT boxes and ChoREs were required for the induction
of Txnip expression by SAHA. SAHA is a histone deacetylase (HDAC) inhibitor, and
may exert its effect by modulating interactions between NF‐Y and other cofactors;
this in turn might change the local environment (e.g., histone modifications) on the
Txnip promoter thus making the promoter more sensitive to glucose signaling.
Hence, SAHA and glucose may synergistically induce the Txnip expression.
141 Chapter 4
In the future, it is of high interest to determine the binding partners of NF‐Y or MondoA/MLX on Txnip promoter; these proteins could be histone modifiers, chromatin remodelers and other cofactors involved in transcriptional control. These results will be very helpful for dissecting the dynamic changes on Txnip promoter in response to glucose or adenosine‐containing molecules. Moreover, there is an apparent synergy between SAHA and glucose on Txnip expression; this is also an interesting discovery which deserves further investigation.
142 References
References
Ahsan,M.K., Masutani,H., Yamaguchi,Y., Kim,Y.C., Nosaka,K., Matsuoka,M., Nishinaka,Y., Maeda,M., and Yodoi,J. (2006) Loss of interleukin‐2‐dependency in HTLV‐I‐infected T cells on gene silencing of thioredoxin‐binding protein‐2 Oncogene 25: 2181‐2191.
Aitken,C.J., Hodge,J.M., Nishinaka,Y., Vaughan,T., Yodoi,J., Day,C.J., Morrison,N.A., and Nicholson,G.C. (2004) Regulation of human osteoclast differentiation by thioredoxin binding protein‐2 and redox‐sensitive signaling J Bone Miner.Res 19: 2057‐2064.
Alberts,B., Johnson,A., Lewis,J., Raff,M., Roberts,K., and Walter,P. (2002) Molecular biology of the cell (4th) New York: Garland Science, pp. 375‐391.
Alonso,C.R., Pesce,C.G., and Kornblihtt,A.R. (1996) The CCAAT‐binding proteins CP1 and NF‐I cooperate with ATF‐2 in the transcription of the fibronectin gene J Biol Chem 271: 22271‐22279.
Aparicio,O., Geisberg,J.V., Sekinger,E., Yang,A., Moqtaderi,Z., and Struhl,K. (2005) Chromatin immunoprecipitation for determining the association of proteins with specific genomic sequences in vivo Curr Protoc.Mol Biol Chapter 21: Unit 21.3.
Arner S.J. and Holmgren A. Measurement of thioredoxin and thioredoxin reductase. Kathy Morgan. Current Protocols in Toxicology. 7.4.1‐7.4.14. 2000. New York, Wiley & Sons, Inc. Current Protocols in Toxicology. Kathy Morgan. Ref Type: Generic
Aymerich,I., Foufelle,F., Ferre,P., Casado,F.J., and Pastor‐Anglada,M. (2006) Extracellular adenosine activates AMP‐dependent protein kinase (AMPK) J Cell Sci 119: 1612‐1621.
Baeuerle,P.A., Baltimore,D. (1996) NF‐kappa B: ten years after Cell 87: 13‐20.
Baker,A.F., Koh,M.Y., Williams,R.R., James,B., Wang,H., Tate,W.R., Gallegos,A., Von Hoff,D.D., Han,H., and Powis,G. (2008) Identification of thioredoxin‐interacting protein 1 as a hypoxia‐inducible factor 1alpha‐induced gene in pancreatic cancer Pancreas 36: 178‐186.
Bandyopadhyay,G., Sajan,M.P., Kanoh,Y., Standaert,M.L., Burke,T.R., Jr., Quon,M.J., Reed,B.C., Dikic,I., Noel,L.E., Newgard,C.B., and Farese,R. (2000) Glucose activates mitogen‐activated protein kinase (extracellular signal‐regulated kinase) through proline‐rich tyrosine kinase‐2 and the Glut1 glucose transporter J Biol Chem 275: 40817‐40826.
Berger,J., Moller,D.E. (2002) The mechanisms of action of PPARs Annu Rev Med. 53: 409‐435.
Berk,A.J. (2000) TBP‐like factors come into focus Cell 103: 5‐8.
143 References
Beukers,M.W., Kerkhof,C.J., van Rhee,M.A., Ardanuy,U., Gurgel,C., Widjaja,H., Nickel,P., IJzerman,A.P., and Soudijn,W. (1995) Suramin analogs, divalent cations and ATP gamma S as inhibitors of ecto‐ATPase Naunyn Schmiedebergs Arch.Pharmacol. 351: 523‐528.
Billiet,L., Furman,C., Larigauderie,G., Copin,C., Page,S., Fruchart,J.C., Brand,K., and Rouis,M. (2008a) Enhanced VDUP‐1 gene expression by PPARgamma agonist induces apoptosis in human macrophage J Cell Physiol 214: 183‐191.
Billiet,L., Furman,C., Cuaz‐Perolin,C., Paumelle,R., Raymondjean,M., Simmet,T., and Rouis,M. (2008b) Thioredoxin‐1 and its natural inhibitor, vitamin D3 up‐regulated protein 1, are differentially regulated by PPARalpha in human macrophages J Mol Biol 384: 564‐576.
Billin,A.N., Ayer,D.E. (2006) The Mlx network: evidence for a parallel Max‐like transcriptional network that regulates energy metabolism Curr Top.Microbiol Immunol 302: 255‐278.
Blackwood,E.M., Eisenman,R.N. (1991) Max: a helix‐loop‐helix zipper protein that forms a sequence‐specific DNA‐binding complex with Myc Science 251: 1211‐1217.
Bodnar,J.S., Chatterjee,A., Castellani,L.W., Ross,D.A., Ohmen,J., Cavalcoli,J., Wu,C., Dains,K.M., Catanese,J., Chu,M., Sheth,S.S., Charugundla,K., Demant,P., West,D.B., de Jong,P., and Lusis,A.J. (2002) Positional cloning of the combined hyperlipidemia gene Hyplip1 Nat Genet. 30: 110‐116.
Brivanlou,A.H., Darnell,J.E., Jr. (2002) Signal transduction and the control of gene expression Science 295: 813‐818.
Bruegge,K., Jelkmann,W., and Metzen,E. (2007) Hydroxylation of hypoxia‐inducible transcription factors and chemical compounds targeting the HIF‐alpha hydroxylases Curr Med.Chem 14: 1853‐1862.
Bruzzone,S., Moreschi,I., Guida,L., Usai,C., Zocchi,E., and De Flora,A. (2006) Extracellular NAD+ regulates intracellular calcium levels and induces activation of human granulocytes Biochem J 393: 697‐704.
Buratowski,S., Hahn,S., Guarente,L., and Sharp,P.A. (1989) Five intermediate complexes in transcription initiation by RNA polymerase II Cell 56: 549‐561.
Buratowski,S. (2005) Connections between mRNA 3ʹ end processing and transcription termination Curr Opin Cell Biol 17: 257‐261.
Burnstock,G. (2008) Purinergic signalling and disorders of the central nervous system Nat Rev Drug Discov. 7: 575‐590.
Butler,J.E., Kadonaga,J.T. (2002) The RNA polymerase II core promoter: a key component in the regulation of gene expression Genes Dev 16: 2583‐2592.
144 References
Butler,L.M., Zhou,X., Xu,W.S., Scher,H.I., Rifkind,R.A., Marks,P.A., and Richon,V.M. (2002) The histone deacetylase inhibitor SAHA arrests cancer cell growth, up‐ regulates thioredoxin‐binding protein‐2, and down‐regulates thioredoxin Proc Natl Acad Sci U S A 99: 11700‐11705.
Butow,R.A., Avadhani,N.G. (2004) Mitochondrial signaling: the retrograde response Mol Cell 14: 1‐15.
Cairo,S., Merla,G., Urbinati,F., Ballabio,A., and Reymond,A. (2001) WBSCR14, a gene mapping to the Williams‐‐Beuren syndrome deleted region, is a new member of the Mlx transcription factor network Hum.Mol Genet. 10: 617‐627.
Ceribelli,M., Dolfini,D., Merico,D., Gatta,R., Vigano,A.M., Pavesi,G., and Mantovani,R. (2008) The histone‐like NF‐Y is a bifunctional transcription factor Mol Cell Biol 28: 2047‐2058.
Cha‐Molstad,H., Saxena,G., Chen,J., and Shalev,A. (2009) Glucose‐stimulated expression of TXNIP is mediated by CHREBP, p300 and histone H4 acetylation in pancreatic beta cells J Biol Chem.
Champliaud,M.F., Viel,A., and Baden,H.P. (2003) The expression of vitamin D‐ upregulated protein 1 in skin and its interaction with sciellin in cultured keratinocytes J Invest Dermatol. 121: 781‐785.
Chen,C.L., Lin,C.F., Chang,W.T., Huang,W.C., Teng,C.F., and Lin,Y.S. (2008c) Ceramide induces p38 MAPK and JNK activation through a mechanism involving a thioredoxin‐interacting protein‐mediated pathway Blood 111: 4365‐4374.
Chen,J., Couto,F.M., Minn,A.H., and Shalev,A. (2006) Exenatide inhibits beta‐cell apoptosis by decreasing thioredoxin‐interacting protein Biochem Biophys.Res Commun. 346: 1067‐1074.
Chen,J., Hui,S.T., Couto,F.M., Mungrue,I.N., Davis,D.B., Attie,A.D., Lusis,A.J., Davis,R.A., and Shalev,A. (2008a) Thioredoxin‐interacting protein deficiency induces Akt/Bcl‐xL signaling and pancreatic beta‐cell mass and protects against diabetes FASEB J 22: 3581‐3594.
Chen,J., Saxena,G., Mungrue,I.N., Lusis,A.J., and Shalev,A. (2008b) Thioredoxin‐ interacting protein: a critical link between glucose toxicity and beta‐cell apoptosis Diabetes 57: 938‐944.
Chen,J., Cha‐Molstad,H., Szabo,A., and Shalev,A. (2009) Diabetes induces and calcium channel blockers prevent cardiac expression of proapoptotic thioredoxin‐ interacting protein Am.J Physiol Endocrinol Metab 296: E1133‐E1139.
Chen,K.S., DeLuca,H.F. (1994) Isolation and characterization of a novel cDNA from HL‐60 cells treated with 1,25‐dihydroxyvitamin D‐3 Biochim.Biophys.Acta 1219: 26‐32.
145 References
Cheng,D.W., Jiang,Y., Shalev,A., Kowluru,R., Crook,E.D., and Singh,L.P. (2006) An analysis of high glucose and glucosamine‐induced gene expression and oxidative stress in renal mesangial cells Arch.Physiol Biochem 112: 189‐218.
Cheng,G.C., Schulze,P.C., Lee,R.T., Sylvan,J., Zetter,B.R., and Huang,H. (2004) Oxidative stress and thioredoxin‐interacting protein promote intravasation of melanoma cells Exp Cell Res 300: 297‐307.
Chung,J.W., Jeon,J.H., Yoon,S.R., and Choi,I. (2006) Vitamin D3 upregulated protein 1 (VDUP1) is a regulator for redox signaling and stress‐mediated diseases J Dermatol. 33: 662‐669.
Chutkow,W.A., Patwari,P., Yoshioka,J., and Lee,R.T. (2008) Thioredoxin‐interacting protein (Txnip) is a critical regulator of hepatic glucose production J Biol Chem 283: 2397‐2406.
Clapier,C.R., Cairns,B.R. (2009) The biology of chromatin remodeling complexes Annu Rev Biochem 78: 273‐304.
Conaway,J.W., Shilatifard,A., Dvir,A., and Conaway,R.C. (2000) Control of elongation by RNA polymerase II Trends Biochem Sci 25: 375‐380.
Conaway,R.C., Brower,C.S., and Conaway,J.W. (2002) Emerging roles of ubiquitin in transcription regulation Science 296: 1254‐1258.
Coon,H., Singh,N., Dunn,D., Eckfeldt,J.H., Province,M.A., Hopkins,P.N., Weiss,R., Hunt,S.C., and Leppert,M.F. (2004) TXNIP gene not associated with familial combined hyperlipidemia in the NHLBI Family Heart Study Atherosclerosis 174: 357‐ 362.
Cooper,S.J., Trinklein,N.D., Anton,E.D., Nguyen,L., and Myers,R.M. (2006) Comprehensive analysis of transcriptional promoter structure and function in 1% of the human genome Genome Res 16: 1‐10.
Currie,R.A. (1998) NF‐Y is associated with the histone acetyltransferases GCN5 and P/CAF J Biol Chem 273: 1430‐1434. da Silva,C.G., Jarzyna,R., Specht,A., and Kaczmarek,E. (2006) Extracellular nucleotides and adenosine independently activate AMP‐activated protein kinase in endothelial cells: involvement of P2 receptors and adenosine transporters Circ Res 98: e39‐e47.
Dai,R.P., Yu,F.X., Goh,S.R., Chng,H.W., Tan,Y.L., Fu,J.L., Zheng,L., and Luo,Y. (2008) Histone 2B (H2B) expression is confined to a proper NAD+/NADH redox status J Biol Chem 283: 26894‐26901.
Darnell,J.E., Jr. (1997) STATs and gene regulation Science 277: 1630‐1635.
146 References
de Candia,P., Blekhman,R., Chabot,A.E., Oshlack,A., and Gilad,Y. (2008) A combination of genomic approaches reveals the role of FOXO1a in regulating an oxidative stress response pathway PLoS.ONE. 3: e1670. de Luis,O., Valero,M.C., and Jurado,L.A. (2000) WBSCR14, a putative transcription factor gene deleted in Williams‐Beuren syndrome: complete characterisation of the human gene and the mouse ortholog Eur.J Hum.Genet. 8: 215‐222.
Dennis,P.B., Jaeschke,A., Saitoh,M., Fowler,B., Kozma,S.C., and Thomas,G. (2001) Mammalian TOR: a homeostatic ATP sensor Science 294: 1102‐1105.
Dentin,R., Pegorier,J.P., Benhamed,F., Foufelle,F., Ferre,P., Fauveau,V., Magnuson,M.A., Girard,J., and Postic,C. (2004) Hepatic glucokinase is required for the synergistic action of ChREBP and SREBP‐1c on glycolytic and lipogenic gene expression J Biol Chem 279: 20314‐20326.
Deroo,B.J., Hewitt,S.C., Peddada,S.D., and Korach,K.S. (2004) Estradiol regulates the thioredoxin antioxidant system in the mouse uterus Endocrinology 145: 5485‐5492.
Diez‐Sampedro,A., Hirayama,B.A., Osswald,C., Gorboulev,V., Baumgarten,K., Volk,C., Wright,E.M., and Koepsell,H. (2003) A glucose sensor hiding in a family of transporters Proc Natl Acad Sci U S A 100: 11753‐11758.
Donnelly,K.L., Margosian,M.R., Sheth,S.S., Lusis,A.J., and Parks,E.J. (2004) Increased lipogenesis and fatty acid reesterification contribute to hepatic triacylglycerol stores in hyperlipidemic Txnip‐/‐ mice J Nutr. 134: 1475‐1480.
Dutta,K.K., Nishinaka,Y., Masutani,H., Akatsuka,S., Aung,T.T., Shirase,T., Lee,W.H., Yamada,Y., Hiai,H., Yodoi,J., and Toyokuni,S. (2005) Two distinct mechanisms for loss of thioredoxin‐binding protein‐2 in oxidative stress‐induced renal carcinogenesis Lab Invest 85: 798‐807.
Ericsson,J., Jackson,S.M., and Edwards,P.A. (1996) Synergistic binding of sterol regulatory element‐binding protein and NF‐Y to the farnesyl diphosphate synthase promoter is critical for sterol‐regulated expression of the gene J Biol Chem 271: 24359‐ 24364.
Filby,C.E., Hooper,S.B., Sozo,F., Zahra,V.A., Flecknoe,S.J., and Wallace,M.J. (2006) VDUP1: a potential mediator of expansion‐induced lung growth and epithelial cell differentiation in the ovine fetus Am.J Physiol Lung Cell Mol Physiol 290: L250‐L258.
Flores,O., Ha,I., and Reinberg,D. (1990) Factors involved in specific transcription by mammalian RNA polymerase II. Purification and subunit composition of transcription factor IIF J Biol Chem 265: 5629‐5634.
Flores,O., Lu,H., and Reinberg,D. (1992) Factors involved in specific transcription by mammalian RNA polymerase II. Identification and characterization of factor IIH J Biol Chem 267: 2786‐2793.
147 References
Ge,K., Guermah,M., Yuan,C.X., Ito,M., Wallberg,A.E., Spiegelman,B.M., and Roeder,R.G. (2002) Transcription coactivator TRAP220 is required for PPAR gamma 2‐stimulated adipogenesis Nature 417: 563‐567.
Gill,G. (2005) Something about SUMO inhibits transcription Curr Opin Genet.Dev 15: 536‐541.
Glover,J.N., Harrison,S.C. (1995) Crystal structure of the heterodimeric bZIP transcription factor c‐Fos‐c‐Jun bound to DNA Nature 373: 257‐261.
Goldberg,S.F., Miele,M.E., Hatta,N., Takata,M., Paquette‐Straub,C., Freedman,L.P., and Welch,D.R. (2003) Melanoma metastasis suppression by chromosome 6: evidence for a pathway regulated by CRSP3 and TXNIP Cancer Res 63: 432‐440.
Gong,Y., Agani,F.H. (2005) Oligomycin inhibits HIF‐1alpha expression in hypoxic tumor cells Am.J Physiol Cell Physiol 288: C1023‐C1029.
Green,M.R. (2000) TBP‐associated factors (TAFIIs): multiple, selective transcriptional mediators in common complexes Trends Biochem Sci 25: 59‐63.
Gregor,P.D., Sawadogo,M., and Roeder,R.G. (1990) The adenovirus major late transcription factor USF is a member of the helix‐loop‐helix group of regulatory proteins and binds to DNA as a dimer Genes Dev 4: 1730‐1740.
Gu,M., Lima,C.D. (2005) Processing the message: structural insights into capping and decapping mRNA Curr Opin Struct.Biol 15: 99‐106.
Guo,Y., Einhorn,L., Kelley,M., Hirota,K., Yodoi,J., Reinbold,R., Scholer,H., Ramsey,H., and Hromas,R. (2004) Redox regulation of the embryonic stem cell transcription factor oct‐4 by thioredoxin Stem Cells 22: 259‐264.
Haag,F., Adriouch,S., Brass,A., Jung,C., Moller,S., Scheuplein,F., Bannas,P., Seman,M., and Koch‐Nolte,F. (2007) Extracellular NAD and ATP: Partners in immune cell modulation Purinergic.Signal. 3: 71‐81.
Hagen,T., Taylor,C.T., Lam,F., and Moncada,S. (2003) Redistribution of intracellular oxygen in hypoxia by nitric oxide: effect on HIF1alpha Science 302: 1975‐1978.
Han,D.H., Hansen,P.A., Nolte,L.A., and Holloszy,J.O. (1998) Removal of adenosine decreases the responsiveness of muscle glucose transport to insulin and contractions Diabetes 47: 1671‐1675.
Han,S.H., Jeon,J.H., Ju,H.R., Jung,U., Kim,K.Y., Yoo,H.S., Lee,Y.H., Song,K.S., Hwang,H.M., Na,Y.S., Yang,Y., Lee,K.N., and Choi,I. (2003) VDUP1 upregulated by TGF‐beta1 and 1,25‐dihydorxyvitamin D3 inhibits tumor cell growth by blocking cell‐cycle progression Oncogene 22: 4035‐4046.
Hanley,P.J., Musset,B., Renigunta,V., Limberg,S.H., Dalpke,A.H., Sus,R., Heeg,K.M., Preisig‐Muller,R., and Daut,J. (2004) Extracellular ATP induces oscillations of
148 References
intracellular Ca2+ and membrane potential and promotes transcription of IL‐6 in macrophages Proc Natl Acad Sci U S A 101: 9479‐9484.
Harrison,D.G. (2005) The shear stress of keeping arteries clear Nat Med. 11: 375‐376.
Hay,R.T. (2005) SUMO: a history of modification Mol Cell 18: 1‐12.
Herr,A.J., Jensen,M.B., Dalmay,T., and Baulcombe,D.C. (2005) RNA polymerase IV directs silencing of endogenous DNA Science 308: 118‐120.
Hirota,T., Okano,T., Kokame,K., Shirotani‐Ikejima,H., Miyata,T., and Fukada,Y. (2002) Glucose down‐regulates Per1 and Per2 mRNA levels and induces circadian gene expression in cultured Rat‐1 fibroblasts J Biol Chem 277: 44244‐44251.
Huang,D.Y., Kuo,Y.Y., Lai,J.S., Suzuki,Y., Sugano,S., and Chang,Z.F. (2004) GATA‐1 and NF‐Y cooperate to mediate erythroid‐specific transcription of Gfi‐1B gene Nucleic Acids Res 32: 3935‐3946.
Huang,L., Pardee,A.B. (2000) Suberoylanilide hydroxamic acid as a potential therapeutic agent for human breast cancer treatment Mol Med. 6: 849‐866.
Hui,S.T., Andres,A.M., Miller,A.K., Spann,N.J., Potter,D.W., Post,N.M., Chen,A.Z., Sachithanantham,S., Jung,D.Y., Kim,J.K., and Davis,R.A. (2008) Txnip balances metabolic and growth signaling via PTEN disulfide reduction Proc Natl Acad Sci U S A 105: 3921‐3926.
Hui,T.Y., Sheth,S.S., Diffley,J.M., Potter,D.W., Lusis,A.J., Attie,A.D., and Davis,R.A. (2004) Mice lacking thioredoxin‐interacting protein provide evidence linking cellular redox state to appropriate response to nutritional signals J Biol Chem 279: 24387‐24393.
Hwang,C.Y., Ryu,Y.S., Chung,M.S., Kim,K.D., Park,S.S., Chae,S.K., Chae,H.Z., and Kwon,K.S. (2004) Thioredoxin modulates activator protein 1 (AP‐1) activity and p27Kip1 degradation through direct interaction with Jab1 Oncogene 23: 8868‐8875.
Ichijo,H., Nishida,E., Irie,K., ten Dijke,P., Saitoh,M., Moriguchi,T., Takagi,M., Matsumoto,K., Miyazono,K., and Gotoh,Y. (1997) Induction of apoptosis by ASK1, a mammalian MAPKKK that activates SAPK/JNK and p38 signaling pathways Science 275: 90‐94.
Ikarashi,M., Takahashi,Y., Ishii,Y., Nagata,T., Asai,S., and Ishikawa,K. (2002) Vitamin D3 up‐regulated protein 1 (VDUP1) expression in gastrointestinal cancer and its relation to stage of disease Anticancer Res 22: 4045‐4048.
Iraqui,I., Vissers,S., Bernard,F., de Craene,J.O., Boles,E., Urrestarazu,A., and Andre,B. (1999) Amino acid signaling in Saccharomyces cerevisiae: a permease‐like sensor of external amino acids and F‐Box protein Grr1p are required for transcriptional induction of the AGP1 gene, which encodes a broad‐specificity amino acid permease Mol Cell Biol 19: 989‐1001.
149 References
Ishii,S., Iizuka,K., Miller,B.C., and Uyeda,K. (2004) Carbohydrate response element binding protein directly promotes lipogenic enzyme gene transcription Proc Natl Acad Sci U S A 101: 15597‐15602.
Isogai,Y., Keles,S., Prestel,M., Hochheimer,A., and Tjian,R. (2007) Transcription of histone gene cluster by differential core‐promoter factors Genes Dev 21: 2936‐2949.
Jeon,J.H., Lee,K.N., Hwang,C.Y., Kwon,K.S., You,K.H., and Choi,I. (2005) Tumor suppressor VDUP1 increases p27(kip1) stability by inhibiting JAB1 Cancer Res 65: 4485‐4489.
Joguchi,A., Otsuka,I., Minagawa,S., Suzuki,T., Fujii,M., and Ayusawa,D. (2002) Overexpression of VDUP1 mRNA sensitizes HeLa cells to paraquat Biochem Biophys.Res Commun. 293: 293‐297.
Junn,E., Han,S.H., Im,J.Y., Yang,Y., Cho,E.W., Um,H.D., Kim,D.K., Lee,K.W., Han,P.L., Rhee,S.G., and Choi,I. (2000) Vitamin D3 up‐regulated protein 1 mediates oxidative stress via suppressing the thioredoxin function J Immunol 164: 6287‐6295.
Kabashima,T., Kawaguchi,T., Wadzinski,B.E., and Uyeda,K. (2003) Xylulose 5‐ phosphate mediates glucose‐induced lipogenesis by xylulose 5‐phosphate‐activated protein phosphatase in rat liver Proc Natl Acad Sci U S A 100: 5107‐5112.
Kaimul,A.M., Nakamura,H., Masutani,H., and Yodoi,J. (2007) Thioredoxin and thioredoxin‐binding protein‐2 in cancer and metabolic syndrome Free Radic.Biol Med. 43: 861‐868.
Kam,K.Y., Jeong,K.H., Norwitz,E.R., Jorgensen,E.M., and Kaiser,U.B. (2005) Oct‐1 and nuclear factor Y bind to the SURG‐1 element to direct basal and gonadotropin‐ releasing hormone (GnRH)‐stimulated mouse GnRH receptor gene transcription Mol Endocrinol 19: 148‐162.
Karar,J., Dolt,K.S., Mishra,M.K., Arif,E., Javed,S., and Pasha,M.A. (2007) Expression and functional activity of pro‐oxidants and antioxidants in murine heart exposed to acute hypobaric hypoxia FEBS Lett. 581: 4577‐4582.
Karin,M., Liu,Z., and Zandi,E. (1997) AP‐1 function and regulation Curr Opin Cell Biol 9: 240‐246.
Kawaguchi,T., Osatomi,K., Yamashita,H., Kabashima,T., and Uyeda,K. (2002) Mechanism for fatty acid ʺsparingʺ effect on glucose‐induced transcription: regulation of carbohydrate‐responsive element‐binding protein by AMP‐activated protein kinase J Biol Chem 277: 3829‐3835.
Keung,Y.K., Yung,C., Wong,J.W., Shah,F., Cobos,E., and Tonk,V. (1998) Unusual presentation of multiple myeloma with ʺjumping translocationʺ involving 1q21. A case report and review of the literature Cancer Genet.Cytogenet. 106: 135‐139.
150 References
Kim,K.Y., Shin,S.M., Kim,J.K., Paik,S.G., Yang,Y., and Choi,I. (2004) Heat shock factor regulates VDUP1 gene expression Biochem Biophys.Res Commun. 315: 369‐375.
Kim,S.Y., Suh,H.W., Chung,J.W., Yoon,S.R., and Choi,I. (2007) Diverse functions of VDUP1 in cell proliferation, differentiation, and diseases Cell Mol Immunol 4: 345‐351.
Klemm,J.D., Rould,M.A., Aurora,R., Herr,W., and Pabo,C.O. (1994) Crystal structure of the Oct‐1 POU domain bound to an octamer site: DNA recognition with tethered DNA‐binding modules Cell 77: 21‐32.
Kobayashi,T., Uehara,S., Ikeda,T., Itadani,H., and Kotani,H. (2003) Vitamin D3 up‐ regulated protein‐1 regulates collagen expression in mesangial cells Kidney Int. 64: 1632‐1642.
Kolbus,A., Blazquez‐Domingo,M., Carotta,S., Bakker,W., Luedemann,S., von Lindern,M., Steinlein,P., and Beug,H. (2003) Cooperative signaling between cytokine receptors and the glucocorticoid receptor in the expansion of erythroid progenitors: molecular analysis by expression profiling Blood 102: 3136‐3146.
Kornberg,R.D. (2005) Mediator and the mechanism of transcriptional activation Trends Biochem Sci 30: 235‐239.
Kornberg,R.D. (2007) The molecular basis of eukaryotic transcription Proc Natl Acad Sci U S A 104: 12955‐12961.
Kouzarides,T. (2007) Chromatin modifications and their function Cell 128: 693‐705.
Lachaal,M., Spangler,R.A., and Jung,C.Y. (2001) Adenosine and adenosine triphosphate modulate the substrate binding affinity of glucose transporter GLUT1 in vitro Biochim.Biophys.Acta 1511: 123‐133.
Le Jan,S., Le Meur,N., Cazes,A., Philippe,J., Le Cunff,M., Leger,J., Corvol,P., and Germain,S. (2006) Characterization of the expression of the hypoxia‐induced genes neuritin, TXNIP and IGFBP3 in cancer FEBS Lett. 580: 3395‐3400.
Lee,K.N., Kang,H.S., Jeon,J.H., Kim,E.M., Yoon,S.R., Song,H., Lyu,C.Y., Piao,Z.H., Kim,S.U., Han,Y.H., Song,S.S., Lee,Y.H., Song,K.S., Kim,Y.M., Yu,D.Y., and Choi,I. (2005) VDUP1 is required for the development of natural killer cells Immunity. 22: 195‐208.
Lee,M.G., Pedersen,P.L. (2003) Glucose metabolism in cancer: importance of transcription factor‐DNA interactions within a short segment of the proximal region og the type II hexokinase promoter J Biol Chem 278: 41047‐41058.
Lee,Y., Kim,M., Han,J., Yeom,K.H., Lee,S., Baek,S.H., and Kim,V.N. (2004) MicroRNA genes are transcribed by RNA polymerase II EMBO J 23: 4051‐4060.
Leung,T.H., Hoffmann,A., and Baltimore,D. (2004) One nucleotide in a kappaB site can determine cofactor specificity for NF‐kappaB dimers Cell 118: 453‐464.
151 References
Levine,M., Tjian,R. (2003) Transcription regulation and animal diversity Nature 424: 147‐151.
Li,B., Carey,M., and Workman,J.L. (2007a) The role of chromatin during transcription Cell 128: 707‐719.
Li,X., Rong,Y., Zhang,M., Wang,X.L., LeMaire,S.A., Coselli,J.S., Zhang,Y., and Shen,Y.H. (2009) Up‐regulation of thioredoxin interacting protein (Txnip) by p38 MAPK and FOXO1 contributes to the impaired thioredoxin activity and increased ROS in glucose‐treated endothelial cells Biochem Biophys.Res Commun. 381: 660‐665.
Li,Y.M., Schilling,T., Benisch,P., Zeck,S., Meissner‐Weigl,J., Schneider,D., Limbert,C., Seufert,J., Kassem,M., Schutze,N., Jakob,F., and Ebert,R. (2007b) Effects of high glucose on mesenchymal stem cell proliferation and differentiation Biochem Biophys.Res Commun. 363: 209‐215.
Liu,Q., Vera,J.C., Peng,H., and Golde,D.W. (2001) The predicted ATP‐binding domains in the hexose transporter GLUT1 critically affect transporter activity Biochemistry 40: 7874‐7881.
Liu,Y., Min,W. (2002) Thioredoxin promotes ASK1 ubiquitination and degradation to inhibit ASK1‐mediated apoptosis in a redox activity‐independent manner Circ Res 90: 1259‐1266.
Ludwig,D.L., Kotanides,H., Le,T., Chavkin,D., Bohlen,P., and Witte,L. (2001) Cloning, genetic characterization, and chromosomal mapping of the mouse VDUP1 gene Gene 269: 103‐112.
Luo,Y., Roeder,R.G. (1999) B‐cell‐specific coactivator OCA‐B: biochemical aspects, role in B‐cell development and beyond Cold Spring Harb.Symp.Quant.Biol 64: 119‐131.
Ma,L., Tsatsos,N.G., and Towle,H.C. (2005) Direct role of ChREBP.Mlx in regulating hepatic glucose‐responsive genes J Biol Chem 280: 12019‐12027.
Ma,L., Robinson,L.N., and Towle,H.C. (2006) ChREBP*Mlx is the principal mediator of glucose‐induced gene expression in the liver J Biol Chem 281: 28721‐28730.
Makino,Y., Yoshikawa,N., Okamoto,K., Hirota,K., Yodoi,J., Makino,I., and Tanaka,H. (1999) Direct association with thioredoxin allows redox regulation of glucocorticoid receptor function J Biol Chem 274: 3182‐3188.
Malik,S., Roeder,R.G. (2005) Dynamic regulation of pol II transcription by the mammalian Mediator complex Trends Biochem Sci 30: 256‐263.
Mandalaywala,N.V., Chang,S., Snyder,R.G., Levendusky,M.C., Voigt,J.M., and Dearborn,R.E., Jr. (2008) The tumor suppressor, vitamin D3 up‐regulated protein 1 (VDUP1), functions downstream of REPO during Drosophila gliogenesis Dev Biol 315: 489‐504.
152 References
Mangus,D.A., Evans,M.C., and Jacobson,A. (2003) Poly(A)‐binding proteins: multifunctional scaffolds for the post‐transcriptional control of gene expression Genome Biol 4: 223.
Martin,K.C., Ephrussi,A. (2009) mRNA localization: gene expression in the spatial dimension Cell 136: 719‐730.
Masson,E., Koren,S., Razik,F., Goldberg,H.J., Kwan,E.P., Sheu,L., Gaisano,H., and Fantus,I.G. (2009) HIGH BETA‐CELL MASS PREVENTS STREPTOZOTOCIN‐ INDUCED DIABETES IN THIOREDOXIN‐INTERACTING PROTEIN‐DEFICIENT MICE Am.J Physiol Endocrinol Metab.
Matera,A.G. (1999) Nuclear bodies: multifaceted subdomains of the interchromatin space Trends Cell Biol 9: 302‐309.
Matsui,T., Segall,J., Weil,P.A., and Roeder,R.G. (1980) Multiple factors required for accurate initiation of transcription by purified RNA polymerase II J Biol Chem 255: 11992‐11996.
Matsuoka,I., Zhou,Q., Ishimoto,H., and Nakanishi,H. (1995) Extracellular ATP stimulates adenylyl cyclase and phospholipase C through distinct purinoceptors in NG108‐15 cells Mol Pharmacol. 47: 855‐862.
Mayr,B., Montminy,M. (2001) Transcriptional regulation by the phosphorylation‐ dependent factor CREB Nat Rev Mol Cell Biol 2: 599‐609.
McGee,S.L., Hargreaves,M. (2008) AMPK and transcriptional regulation Front Biosci. 13: 3022‐3033.
Minn,A.H., Pise‐Masison,C.A., Radonovich,M., Brady,J.N., Wang,P., Kendziorski,C., and Shalev,A. (2005a) Gene expression profiling in INS‐1 cells overexpressing thioredoxin‐interacting protein Biochem Biophys.Res Commun. 336: 770‐778.
Minn,A.H., Hafele,C., and Shalev,A. (2005b) Thioredoxin‐interacting protein is stimulated by glucose through a carbohydrate response element and induces beta‐ cell apoptosis Endocrinology 146: 2397‐2405.
Minn,A.H., Couto,F.M., and Shalev,A. (2006) Metabolism‐independent sugar effects on gene transcription: the role of 3‐O‐methylglucose Biochemistry 45: 11047‐11051.
Muoio,D.M. (2007) TXNIP links redox circuitry to glucose control Cell Metab 5: 412‐ 414.
Ng,M.K., Wu,J., Chang,E., Wang,B.Y., Katzenberg‐Clark,R., Ishii‐Watabe,A., and Cooke,J.P. (2007) A central role for nicotinic cholinergic regulation of growth factor‐ induced endothelial cell migration Arterioscler Thromb Vasc Biol 27: 106‐112.
Nishinaka,Y., Masutani,H., Oka,S., Matsuo,Y., Yamaguchi,Y., Nishio,K., Ishii,Y., and Yodoi,J. (2004a) Importin alpha1 (Rch1) mediates nuclear translocation of
153 References
thioredoxin‐binding protein‐2/vitamin D(3)‐up‐regulated protein 1 J Biol Chem 279: 37559‐37565.
Nishinaka,Y., Nishiyama,A., Masutani,H., Oka,S., Ahsan,K.M., Nakayama,Y., Ishii,Y., Nakamura,H., Maeda,M., and Yodoi,J. (2004b) Loss of thioredoxin‐binding protein‐ 2/vitamin D3 up‐regulated protein 1 in human T‐cell leukemia virus type I‐ dependent T‐cell transformation: implications for adult T‐cell leukemia leukemogenesis Cancer Res 64: 1287‐1292.
Nishiyama,A., Matsui,M., Iwata,S., Hirota,K., Masutani,H., Nakamura,H., Takagi,Y., Sono,H., Gon,Y., and Yodoi,J. (1999) Identification of thioredoxin‐binding protein‐ 2/vitamin D(3) up‐regulated protein 1 as a negative regulator of thioredoxin function and expression J Biol Chem 274: 21645‐21650.
Nishiyama,A., Masutani,H., Nakamura,H., Nishinaka,Y., and Yodoi,J. (2001) Redox regulation by thioredoxin and thioredoxin‐binding proteins IUBMB.Life 52: 29‐33.
Ohta,S., Lai,E.W., Pang,A.L., Brouwers,F.M., Chan,W.Y., Eisenhofer,G., de Krijger,R., Ksinantova,L., Breza,J., Blazicek,P., Kvetnansky,R., Wesley,R.A., and Pacak,K. (2005) Downregulation of metastasis suppressor genes in malignant pheochromocytoma Int.J Cancer 114: 139‐143.
Oka,S., Masutani,H., Liu,W., Horita,H., Wang,D., Kizaka‐Kondoh,S., and Yodoi,J. (2006a) Thioredoxin‐binding protein‐2‐like inducible membrane protein is a novel vitamin D3 and peroxisome proliferator‐activated receptor (PPAR)gamma ligand target protein that regulates PPARgamma signaling Endocrinology 147: 733‐743.
Oka,S., Liu,W., Masutani,H., Hirata,H., Shinkai,Y., Yamada,S., Yoshida,T., Nakamura,H., and Yodoi,J. (2006b) Impaired fatty acid utilization in thioredoxin binding protein‐2 (TBP‐2)‐deficient mice: a unique animal model of Reye syndrome FASEB J 20: 121‐123.
Oka,S., Yoshihara,E., Bizen‐Abe,A., Liu,W., Watanabe,M., Yodoi,J., and Masutani,H. (2009) Thioredoxin binding protein‐2/thioredoxin‐interacting protein is a critical regulator of insulin secretion and peroxisome proliferator‐activated receptor function Endocrinology 150: 1225‐1234.
Ozcan,S., Dover,J., and Johnston,M. (1998) Glucose sensing and signaling by two glucose receptors in the yeast Saccharomyces cerevisiae EMBO J 17: 2566‐2573.
Pajukanta,P., Lilja,H.E., Sinsheimer,J.S., Cantor,R.M., Lusis,A.J., Gentile,M., Duan,X.J., Soro‐Paavonen,A., Naukkarinen,J., Saarela,J., Laakso,M., Ehnholm,C., Taskinen,M.R., and Peltonen,L. (2004) Familial combined hyperlipidemia is associated with upstream transcription factor 1 (USF1) Nat Genet. 36: 371‐376.
Pang,S.T., Hsieh,W.C., Chuang,C.K., Chao,C.H., Weng,W.H., and Juang,H.H. (2009) Thioredoxin‐interacting protein: an oxidative stress‐related gene is upregulated by glucose in human prostate carcinoma cells J Mol Endocrinol 42: 205‐214.
154 References
Panne,D., Maniatis,T., and Harrison,S.C. (2007) An atomic model of the interferon‐ beta enhanceosome Cell 129: 1111‐1123.
Papadia,S., Soriano,F.X., Leveille,F., Martel,M.A., Dakin,K.A., Hansen,H.H., Kaindl,A., Sifringer,M., Fowler,J., Stefovska,V., McKenzie,G., Craigon,M., Corriveau,R., Ghazal,P., Horsburgh,K., Yankner,B.A., Wyllie,D.J., Ikonomidou,C., and Hardingham,G.E. (2008) Synaptic NMDA receptor activity boosts intrinsic antioxidant defenses Nat Neurosci 11: 476‐487.
Parikh,H., Carlsson,E., Chutkow,W.A., Johansson,L.E., Storgaard,H., Poulsen,P., Saxena,R., Ladd,C., Schulze,P.C., Mazzini,M.J., Jensen,C.B., Krook,A., Bjornholm,M., Tornqvist,H., Zierath,J.R., Ridderstrale,M., Altshuler,D., Lee,R.T., Vaag,A., Groop,L.C., and Mootha,V.K. (2007) TXNIP regulates peripheral glucose metabolism in humans PLoS.Med. 4: e158.
Patwari,P., Higgins,L.J., Chutkow,W.A., Yoshioka,J., and Lee,R.T. (2006) The interaction of thioredoxin with Txnip. Evidence for formation of a mixed disulfide by disulfide exchange J Biol Chem 281: 21884‐21891.
Perissi,V., Rosenfeld,M.G. (2005) Controlling nuclear receptors: the circular logic of cofactor cycles Nat Rev Mol Cell Biol 6: 542‐554.
Perkins,N.D. (2000) The Rel/NF‐kappa B family: friend and foe Trends Biochem Sci 25: 434‐440.
Powis,G., Montfort,W.R. (2001) Properties and biological activities of thioredoxins Annu Rev Biophys.Biomol.Struct. 30: 421‐455.
Prahlad,V., Morimoto,R.I. (2009) Integrating the stress response: lessons for neurodegenerative diseases from C. elegans Trends Cell Biol 19: 52‐61.
Price,S.A., Gardiner,N.J., Duran‐Jimenez,B., Zeef,L.A., Obrosova,I.G., and Tomlinson,D.R. (2006) Thioredoxin interacting protein is increased in sensory neurons in experimental diabetes Brain Res 1116: 206‐214.
Qin,J., Clore,G.M., Kennedy,W.M., Huth,J.R., and Gronenborn,A.M. (1995) Solution structure of human thioredoxin in a mixed disulfide intermediate complex with its target peptide from the transcription factor NF kappa B Structure. 3: 289‐297.
Qin,J., Clore,G.M., Kennedy,W.P., Kuszewski,J., and Gronenborn,A.M. (1996) The solution structure of human thioredoxin complexed with its target from Ref‐1 reveals peptide chain reversal Structure. 4: 613‐620.
Rada‐Iglesias,A., Ameur,A., Kapranov,P., Enroth,S., Komorowski,J., Gingeras,T.R., and Wadelius,C. (2008) Whole‐genome maps of USF1 and USF2 binding and histone H3 acetylation reveal new aspects of promoter structure and candidate genes for common human disorders Genome Res 18: 380‐392.
155 References
Rakhshandehroo,M., Sanderson,L.M., Matilainen,M., Stienstra,R., Carlberg,C., de Groot,P.J., Muller,M., and Kersten,S. (2007) Comprehensive Analysis of PPARalpha‐ Dependent Regulation of Hepatic Lipid Metabolism by Expression Profiling PPAR.Res 2007: 26839.
Reinberg,D., Roeder,R.G. (1987a) Factors involved in specific transcription by mammalian RNA polymerase II. Transcription factor IIS stimulates elongation of RNA chains J Biol Chem 262 : 3331‐3337.
Reinberg,D., Roeder,R.G. (1987b) Factors involved in specific transcription by mammalian RNA polymerase II. Purification and functional analysis of initiation factors IIB and IIE J Biol Chem 262: 3310‐3321.
Roeder,R.G., Rutter,W.J. (1969) Multiple forms of DNA‐dependent RNA polymerase in eukaryotic organisms Nature 224: 234‐237.
Roeder,R.G., Rutter,W.J. (1970) Multiple ribonucleic acid polymerases and ribonucleic acid synthesis during sea urchin development Biochemistry 9: 2543‐2553.
Roeder,R.G. (2003) Lasker Basic Medical Research Award. The eukaryotic transcriptional machinery: complexities and mechanisms unforeseen Nat Med. 9: 1239‐1244.
Roeder,R.G. (2005) Transcriptional regulation and the role of diverse coactivators in animal cells FEBS Lett. 579: 909‐915.
Roy,A.L., Meisterernst,M., Pognonec,P., and Roeder,R.G. (1991) Cooperative interaction of an initiator‐binding transcription initiation factor and the helix‐loop‐ helix activator USF Nature 354: 245‐248.
Saitoh,M., Nishitoh,H., Fujii,M., Takeda,K., Tobiume,K., Sawada,Y., Kawabata,M., Miyazono,K., and Ichijo,H. (1998) Mammalian thioredoxin is a direct inhibitor of apoptosis signal‐regulating kinase (ASK) 1 EMBO J 17: 2596‐2606.
Saitoh,T., Tanaka,S., and Koike,T. (2001) Rapid induction and Ca(2+) influx‐mediated suppression of vitamin D3 up‐regulated protein 1 (VDUP1) mRNA in cerebellar granule neurons undergoing apoptosis J Neurochem. 78: 1267‐1276.
Samuels,M., Fire,A., and Sharp,P.A. (1982) Separation and characterization of factors mediating accurate transcription by RNA polymerase II J Biol Chem 257: 14419‐14427.
Sans,C.L., Satterwhite,D.J., Stoltzman,C.A., Breen,K.T., and Ayer,D.E. (2006) MondoA‐Mlx heterodimers are candidate sensors of cellular energy status: mitochondrial localization and direct regulation of glycolysis Mol Cell Biol 26: 4863‐ 4871.
Schulze,P.C., De Keulenaer,G.W., Yoshioka,J., Kassik,K.A., and Lee,R.T. (2002) Vitamin D3‐upregulated protein‐1 (VDUP‐1) regulates redox‐dependent vascular
156 References
smooth muscle cell proliferation through interaction with thioredoxin Circ Res 91: 689‐695.
Schulze,P.C., Yoshioka,J., Takahashi,T., He,Z., King,G.L., and Lee,R.T. (2004) Hyperglycemia promotes oxidative stress through inhibition of thioredoxin function by thioredoxin‐interacting protein J Biol Chem 279: 30369‐30374.
Schulze,P.C., Liu,H., Choe,E., Yoshioka,J., Shalev,A., Bloch,K.D., and Lee,R.T. (2006) Nitric oxide‐dependent suppression of thioredoxin‐interacting protein expression enhances thioredoxin activity Arterioscler Thromb Vasc Biol 26: 2666‐2672.
Scott,J.W., Hawley,S.A., Green,K.A., Anis,M., Stewart,G., Scullion,G.A., Norman,D.G., and Hardie,D.G. (2004) CBS domains form energy‐sensing modules whose binding of adenosine ligands is disrupted by disease mutations J Clin Invest 113: 274‐284.
Semenza,G.L. (2004) Hydroxylation of HIF‐1: oxygen sensing at the molecular level Physiology.(Bethesda.) 19: 176‐182.
Sergeant,S., Kim,H.D. (1985) Inhibition of 3‐O‐methylglucose transport in human erythrocytes by forskolin J Biol Chem 260: 14677‐14682.
Shaked,M., Ketzinel‐Gilad,M., Ariav,Y., Cerasi,E., Kaiser,N., and Leibowitz,G. (2009) Insulin counteracts glucotoxic effects by suppressing thioredoxin‐interacting protein production in INS‐1E beta cells and in Psammomys obesus pancreatic islets Diabetologia 52: 636‐644.
Shalev,A., Pise‐Masison,C.A., Radonovich,M., Hoffmann,S.C., Hirshberg,B., Brady,J.N., and Harlan,D.M. (2002) Oligonucleotide microarray analysis of intact human pancreatic islets: identification of glucose‐responsive genes and a highly regulated TGFbeta signaling pathway Endocrinology 143: 3695‐3698.
Sharp,P.A. (2005) The discovery of split genes and RNA splicing Trends Biochem Sci 30: 279‐281.
Sheth,S.S., Castellani,L.W., Chari,S., Wagg,C., Thipphavong,C.K., Bodnar,J.S., Tontonoz,P., Attie,A.D., Lopaschuk,G.D., and Lusis,A.J. (2005) Thioredoxin‐ interacting protein deficiency disrupts the fasting‐feeding metabolic transition J Lipid Res 46: 123‐134.
Sheth,S.S., Bodnar,J.S., Ghazalpour,A., Thipphavong,C.K., Tsutsumi,S., Tward,A.D., Demant,P., Kodama,T., Aburatani,H., and Lusis,A.J. (2006) Hepatocellular carcinoma in Txnip‐deficient mice Oncogene 25: 3528‐3536.
Shinohara,Y., Sagawa,I., Ichihara,J., Yamamoto,K., Terao,K., and Terada,H. (1997) Source of ATP for hexokinase‐catalyzed glucose phosphorylation in tumor cells: dependence on the rate of oxidative phosphorylation relative to that of extramitochondrial ATP generation Biochim.Biophys.Acta 1319: 319‐330.
157 References
Simmons,D.G., Kennedy,T.G. (2004) Rat endometrial Vdup1 expression: changes related to sensitization for the decidual cell reaction and hormonal control Reproduction. 127: 475‐482.
Sklar,V.E., Schwartz,L.B., and Roeder,R.G. (1975) Distinct molecular structures of nuclear class I, II, and III DNA‐dependent RNA polymerases Proc Natl Acad Sci U S A 72: 348‐352.
Son,A., Nakamura,H., Okuyama,H., Oka,S., Yoshihara,E., Liu,W., Matsuo,Y., Kondo,N., Masutani,H., Ishii,Y., Iyoda,T., Inaba,K., and Yodoi,J. (2008) Dendritic cells derived from TBP‐2‐deficient mice are defective in inducing T cell responses Eur.J Immunol 38: 1358‐1367.
Song,H., Cho,D., Jeon,J.H., Han,S.H., Hur,D.Y., Kim,Y.S., and Choi,I. (2003) Vitamin D(3) up‐regulating protein 1 (VDUP1) antisense DNA regulates tumorigenicity and melanogenesis of murine melanoma cells via regulating the expression of fas ligand and reactive oxygen species Immunol Lett. 86: 235‐247.
Stark,G.R., Kerr,I.M., Williams,B.R., Silverman,R.H., and Schreiber,R.D. (1998) How cells respond to interferons Annu Rev Biochem 67: 227‐264.
Stoeckman,A.K., Ma,L., and Towle,H.C. (2004) Mlx is the functional heteromeric partner of the carbohydrate response element‐binding protein in glucose regulation of lipogenic enzyme genes J Biol Chem 279: 15662‐15669.
Stoltzman,C.A., Peterson,C.W., Breen,K.T., Muoio,D.M., Billin,A.N., and Ayer,D.E. (2008) Glucose sensing by MondoA:Mlx complexes: a role for hexokinases and direct regulation of thioredoxin‐interacting protein expression Proc Natl Acad Sci U S A 105: 6912‐6917.
Taatjes,D.J., Marr,M.T., and Tjian,R. (2004) Regulatory diversity among metazoan co‐ activator complexes Nat Rev Mol Cell Biol 5: 403‐410.
Takahashi,Y., Nagata,T., Ishii,Y., Ikarashi,M., Ishikawa,K., and Asai,S. (2002) Up‐ regulation of vitamin D3 up‐regulated protein 1 gene in response to 5‐fluorouracil in colon carcinoma SW620 Oncol.Rep. 9: 75‐79.
Taparia,S., Fleet,J.C., Peng,J.B., Wang,X.D., and Wood,R.J. (2006) 1,25‐ Dihydroxyvitamin D and 25‐hydroxyvitamin D‐‐mediated regulation of TRPV6 (a putative epithelial calcium channel) mRNA expression in Caco‐2 cells Eur.J Nutr. 45: 196‐204.
Thorn,J.A., Jarvis,S.M. (1996) Adenosine transporters Gen.Pharmacol. 27: 613‐620.
Tissing,W.J., den Boer,M.L., Meijerink,J.P., Menezes,R.X., Swagemakers,S., van der Spek,P.J., Sallan,S.E., Armstrong,S.A., and Pieters,R. (2007) Genomewide identification of prednisolone‐responsive genes in acute lymphoblastic leukemia cells Blood 109: 3929‐3935.
158 References
Tomoda,K., Kubota,Y., and Kato,J. (1999) Degradation of the cyclin‐dependent‐ kinase inhibitor p27Kip1 is instigated by Jab1 Nature 398: 160‐165.
Turturro,F., Friday,E., and Welbourne,T. (2007) Hyperglycemia regulates thioredoxin‐ROS activity through induction of thioredoxin‐interacting protein (TXNIP) in metastatic breast cancer‐derived cells MDA‐MB‐231 BMC.Cancer 7: 96. van der Vleuten,G.M., Hijmans,A., Kluijtmans,L.A., Blom,H.J., Stalenhoef,A.F., and de Graaf,J. (2004) Thioredoxin interacting protein in Dutch families with familial combined hyperlipidemia Am.J Med.Genet.A 130A: 73‐75.
Van Dyke,M.W., Roeder,R.G., and Sawadogo,M. (1988) Physical analysis of transcription preinitiation complex assembly on a class II gene promoter Science 241: 1335‐1338. van Greevenbroek,M.M., Vermeulen,V.M., Feskens,E.J., Evelo,C.T., Kruijshoop,M., Hoebee,B., van der Kallen,C.J., and de Bruin,T.W. (2007) Genetic variation in thioredoxin interacting protein (TXNIP) is associated with hypertriglyceridaemia and blood pressure in diabetes mellitus Diabet.Med. 24: 498‐504.
Villalobo,A. (2006) Nitric oxide and cell proliferation FEBS J 273: 2329‐2344.
Villard,J., Peretti,M., Masternak,K., Barras,E., Caretti,G., Mantovani,R., and Reith,W. (2000) A functionally essential domain of RFX5 mediates activation of major histocompatibility complex class II promoters by promoting cooperative binding between RFX and NF‐Y Mol Cell Biol 20: 3364‐3376.
Wang,Y., De Keulenaer,G.W., and Lee,R.T. (2002) Vitamin D(3)‐up‐regulated protein‐1 is a stress‐responsive gene that regulates cardiomyocyte viability through interaction with thioredoxin J Biol Chem 277: 26496‐26500.
Wang,Z., Rong,Y.P., Malone,M.H., Davis,M.C., Zhong,F., and Distelhorst,C.W. (2006) Thioredoxin‐interacting protein (txnip) is a glucocorticoid‐regulated primary response gene involved in mediating glucocorticoid‐induced apoptosis Oncogene 25: 1903‐1913.
Weil,P.A., Luse,D.S., Segall,J., and Roeder,R.G. (1979) Selective and accurate initiation of transcription at the Ad2 major late promotor in a soluble system dependent on purified RNA polymerase II and DNA Cell 18: 469‐484.
Wen,L.T., Knowles,A.F. (2003) Extracellular ATP and adenosine induce cell apoptosis of human hepatoma Li‐7A cells via the A3 adenosine receptor Br.J Pharmacol. 140: 1009‐1018.
Wu,C. (1995) Heat shock transcription factors: structure and regulation Annu Rev Cell Dev Biol 11: 441‐469.
Xiang,G., Seki,T., Schuster,M.D., Witkowski,P., Boyle,A.J., See,F., Martens,T.P., Kocher,A., Sondermeijer,H., Krum,H., and Itescu,S. (2005) Catalytic degradation of
159 References
vitamin D up‐regulated protein 1 mRNA enhances cardiomyocyte survival and prevents left ventricular remodeling after myocardial ischemia J Biol Chem 280: 39394‐39402.
Xu,W., Ngo,L., Perez,G., Dokmanovic,M., and Marks,P.A. (2006) Intrinsic apoptotic and thioredoxin pathways in human prostate cancer cell response to histone deacetylase inhibitor Proc Natl Acad Sci U S A 103: 15540‐15545.
Yamada,K., Tanaka,T., Miyamoto,K., and Noguchi,T. (2000) Sp family members and nuclear factor‐Y cooperatively stimulate transcription from the rat pyruvate kinase M gene distal promoter region via their direct interactions J Biol Chem 275: 18129‐18137.
Yamaguchi,F., Takata,M., Kamitori,K., Nonaka,M., Dong,Y., Sui,L., and Tokuda,M. (2008) Rare sugar D‐allose induces specific up‐regulation of TXNIP and subsequent G1 cell cycle arrest in hepatocellular carcinoma cells by stabilization of p27kip1 Int.J Oncol. 32: 377‐385.
Yamamoto,Y., Sakamoto,M., Fujii,G., Kanetaka,K., Asaka,M., and Hirohashi,S. (2001) Cloning and characterization of a novel gene, DRH1, down‐regulated in advanced human hepatocellular carcinoma Clin Cancer Res 7: 297‐303.
Yamanaka,H., Maehira,F., Oshiro,M., Asato,T., Yanagawa,Y., Takei,H., and Nakashima,Y. (2000) A possible interaction of thioredoxin with VDUP1 in HeLa cells detected in a yeast two‐hybrid system Biochem Biophys.Res Commun. 271: 796‐800.
Yamashita,H., Takenoshita,M., Sakurai,M., Bruick,R.K., Henzel,W.J., Shillinglaw,W., Arnot,D., and Uyeda,K. (2001) A glucose‐responsive transcription factor that regulates carbohydrate metabolism in the liver Proc Natl Acad Sci U S A 98: 9116‐9121.
Yamawaki,H., Pan,S., Lee,R.T., and Berk,B.C. (2005) Fluid shear stress inhibits vascular inflammation by decreasing thioredoxin‐interacting protein in endothelial cells J Clin Invest 115: 733‐738.
Yin,J., Xu,K., Zhang,J., Kumar,A., and Yu,F.S. (2007) Wound‐induced ATP release and EGF receptor activation in epithelial cells J Cell Sci 120: 815‐825.
Yoshida,T., Nakamura,H., Masutani,H., and Yodoi,J. (2005) The involvement of thioredoxin and thioredoxin binding protein‐2 on cellular proliferation and aging process Ann.N.Y.Acad Sci 1055: 1‐12.
Yoshida,T., Kondo,N., Oka,S., Ahsan,M.K., Hara,T., Masutani,H., Nakamura,H., and Yodoi,J. (2006) Thioredoxin‐binding protein‐2 (TBP‐2): its potential roles in the aging process Biofactors 27: 47‐51.
Yoshioka,J., Imahashi,K., Gabel,S.A., Chutkow,W.A., Burds,A.A., Gannon,J., Schulze,P.C., MacGillivray,C., London,R.E., Murphy,E., and Lee,R.T. (2007) Targeted deletion of thioredoxin‐interacting protein regulates cardiac dysfunction in response to pressure overload Circ Res 101: 1328‐1338.
160 References
Zhou,J., Cidlowski,J.A. (2005) The human glucocorticoid receptor: one gene, multiple proteins and diverse responses Steroids 70: 407‐417.
Zhu,J., Giannola,D.M., Zhang,Y., Rivera,A.J., and Emerson,S.G. (2003) NF‐Y cooperates with USF1/2 to induce the hematopoietic expression of HOXB4 Blood 102: 2420‐2427.
Zimmermann,H. (2000) Extracellular metabolism of ATP and other nucleotides Naunyn Schmiedebergs Arch.Pharmacol. 362: 299‐309.
161 Appendices
Appendix I
Buffers/Gels Used in This Study
SDS‐PAGE sample buffer (5×) 10% w/v SDS 10 mM Dithiothreitol (DTT) 10% v/v Glycerol 0.2 M Tris‐HCl, pH6.8 0.05% w/v Bromophenolblue
SDS‐PAGE running buffer (1×) for 10 L Tris base 30 g Glycine 144 g SDS 10 g It is not necessary to adjust pH.
Coomassie blue staining solution 50% (v/v) methonal 0.05% coomassie brilliantblue R‐250 10% (v/v) acetic acid 40% sterile water.
Destain solution 50% v/v methanol 10% v/v acetic acid 40% sterile water.
Transfer buffer (Electro‐blotting) for 4L 20mM Tris Base (pH8.6) 9.7g 150mM Glycine 45.0g 10% Methanol 400.0mL
TBST 100 mM Tris base, pH7.6 150 mM NaCl 0.1% v/v Tween‐20
PBS‐TX PBS (1×) 0.1% v/v Triton X‐100
START SOLUTION (10×) 1 mM 2‐deoxyglucose 5 μCi/ml [3H]‐2‐deoxyglucose
162 Appendices
KRH Buffer for 1L 50mM HEPES, pH 7.4 5.2g (or 20mL of 1M stock) 136mM NaCl 7.9g (or 27.2 ml of 5 stock) 4.7mM KCl 0.35g (or 4.7 ml of 1M stock) 1.25mM MgSO4 0.3g (or 1.25 ml of 1M stock) 1.25mM CaCl2 0.5ml of 2.5 M stock
Bacterial lysis buffer 20 mM Tris pH 8.0 500 mM NaCl 0.5 mg/ml Lysozyme 0.5% Triton 1 mM PMSF 1mM Beta‐Maecaptoethanol or DTT
MC buffer (for ChIP) 10 mM Tris‐HCl pH 7.5 10 mM NaCl 3 mMMgCl2 0.5% (v/v) NP‐40 Store up to 1 year at 4◦C
1× GS Buffer (for 1ml) 0.2ml PolydI‐dC‐PolydI‐dC (1mg/ml in H2O) 0.3ml H2O 0.5ml 2× GS 1μl 1M DTT Aliquot and store at ‐20oC
2× GS Buffer 25mM Hepes pH8.4 62.5mM KCl 0.05% NP‐40 or IGEPAL 2mM MgCl2 8% Ficoll‐400 500μg/ml BSA 0.5mM PMSF 2X Protease Inhibitor cocktail
10× TGEMN Running Buffer 62.5mM Tris pH8.2 500mM Glycine 1mM EDTA 10mM MgCl2 0.25% NP40 or IGEPAL Store at RT; 1X TGEMN can be stored at 4°C, add 0.5 mM DTT freshly.
163 Appendices
BC100 20 mM Tris‐HCl, pH 7.9 at 4oC 20 % glycerol 0.25 mM EDTA 0.125 mM EGTA 0.025% Triton X‐100 100 mM KCl Before use, freshly add 0.5 mM DTT and 0.25 mM PMSF
Gel for EMSA (4%, for 35 ml) 7ml 20% Acrylamide/Bis solution (59:1 linkage) 7ml 20% Glycerol 3.5ml 10X TGEMN 17.3ml H2O 175μl 10% APS 17.5μl TEMED 17.5μl 1M DTT
SDS‐PAGE Gel composition
Lower Gel (10 ml) 5% 8% 10% 12% 15% 30% Acry‐Bis (29 :1) 1.665 2.665 3.335 4 5 65% Sucrose 0.385 0.59 0.78 1.25 1.565 10× Lower Buffer 1 1 1 1 1 dH2O 6.9 5.74 4.875 3.75 2.44 10% APS 100 100 100 100 100 TEMED 10 10 10 10 10 Stacking gel (5 ml) 4% 30% Acryl‐Bis (29 :1) 0.67 4× Stacking Buffer 1.25 dH2O 3.05 10% APS 50 TEMED 10
164 Appendices
Appendix II
RNA Integrity
RNA (2 μg) prepared using RNeasy Mini Kit was separated and visualized using ethidium bromide agarose gel. The band corresponding to 28S or 18S rRNA was strong and sharp (lanes 2‐6), indicating high‐quality RNA. Lane 1, molecular weight marker (100 bp ladder).
1 2 3 4 5 6
165 Appendices
Appendix III
Primer Specificity
Following Real‐Time PCR (primers against human Txnip or β‐actin), the dissociation curve of PCR products were plotted. As shown below, unique and sharp peak was revealed in dissociation curve for each primer pair, indicating that primers used were specific to their targeting gene. Dissociation curves for primers against other genes were similar (single peak).
Txnip
β-actin
166 Appendices
Appendix IV
Primers Used in This Study
Primers Remarks GCCTCGAGCTCCAAATCGAGGAAACCG-reverse CCGGCTAGCCCAACAAGAATGAAGAGAG-1299 CCGGCTAGCGGGTACAAGCTGGGGG-760 CCGGCTAGCGAGGCCTGAAAGTTCTC-653 Primers for CCGGCTAGCCCTCAGAGACGGTGG-633 generating serial CCGGCTAGCGCACTGGCTAAGACTAG-441 Txnip promoter CCGGCTAGCACAGCCCCCTCCTTCCC-269 deletions. The reverse primer is a CCGGCTAGCGTGTCCACGCGCCACAGC-169 common. Numbers CCGGCTAGCACGCGCCACAGCGATCTC-163 indicate the base pairs upstream of CCGGCTAGCACAGCGATCTCACTGATTG-157 the transcription CCGGCTAGCTCACTGATTGGTCGGGCTC-147 staring site. Restriction sites are CCGGCTAGCGATTGGTCGGGCTCCTGG-142 underlined. CCGGCTAGCAGCCAATGGGAGGGATG-111 CCGGCTAGCGGGAGGGATGTGCACGAGGG-102 CCGGCTAGCCCTCCGGGCCAGCGCTCG-73 CCGGCTAGCCAATCCTCCGGGCCAGCGCTCG-73cat CCGGCTAGCCCCTCCCATTGGCTGCCCG-Rev95 Primers for generating Txnip‐ CCGGCTAGCGGTCCGGAGGCTCGTGCTGC-Rev63 TATA fusion CCGGCTAGCGGTCCGGAGGCTCATACTGC-Rev63mut promoters. 184mut and Rev63mut were CCGACGCGTACAGCCCCCTCCTTCCC-269 used to generate CCGACGCGTAGCCAGGAGCACACCGTGTC-184 mutations at CCGACGCGTAGCCAGGAGTATACCGTGTC-184mut ChoREs (one E‐box mutated). Primers CCGACGCGTCTGATTGGTCGGGCTCCTGG-145 with d170 label CGTGTCCACGCGCCACAGCGAT-d170fwd were used to delete sequences around ‐ CTGGCTGGGAAAATGGTTGTTGCG-d170rev 170. TGTCTATGCGCCACAGCGATCTCAC Primers for ChoRE‐ CGGTATACTCCTGGCTGGGAAAATG b mutation
167 Appendices
Primers Remarks
GCAGTATGAGCCTCCGGGCCAGCG Primers for ChoRE‐ CCTCATACACATCCCTCCCATTGGC a mutation ACTCAGTGAGATCGCTGTGGCG Primers for inverted CCAAT box AGTCGGGCTCCTGGTAAACAAG mutation GATGGGAGGGATGTGCACGAGGGC Primers for CCAAT AGCTGCCCGGTCCTTGTTTACCAG box mutation
CATAGAGACGTTTCCGCCTCCTGC Primers for TATA TCGTAGCCCGGGCCAGAAGAGC box mutation Shuffle CGCGTAGCCAGGAGCACACCGTGTCCACGCGCCACCATACC TAGGACTTGGCTGGTCTGGGTGCCAGCGTAGCAAAGGACGG AGGCCACATGATAGGGTGCACGAGGGCAGCACGAGCCTCCG GACCG-fwd CTAGCGGTCCGGAGGCTCGTGCTGCCCTCGTGCACCCTATC ATGTGGCCTCCGTCCTTTGCTACGCTGGCACCCAGACCAGC CAAGTCCTAGGTATGGTGGCGCGTGGACACGGTGTGCTCCT GGCTA-rev dChoRE-b CGCGAGCACACCGTGTCCACGCGCCTGAGCACACCGTGTCC ACGCGCC-fwd CTAGGGCGCGTGGACACGGTGTGCTCAGGCGCGTGGACACG Synthesized GTGTGCT-rev oligonucleotides for dChoRE-a making Txnip‐ CGCGTGCACGAGGGCAGCACGAGCCTGTGCACGAGGGCAGC ACGAGCC-fwd TATA fusion CTAGGGCTCGTGCTGCCCTCGTGCACAGGCTCGTGCTGCCC promoters TCGTGCA-rev ChoRE-ab CGCGAGCACACCGTGTCCACGCGCCTGTGCACGAGGGCAGC ACGAGCC-fwd CTAGGGCTCGTGCTGCCCTCGTGCACAGGCGCGTGGACACG GTGTGCT-rev ChoRE-b CGCGAGCACACCGTGTCCACGCGCC-fwd CTAGGGCGCGTGGACACGGTGTGCT-rev ChoRE-a CGCGTGCACGAGGGCAGCACGAGCC-fwd CTAGGGCTCGTGCTGCCCTCGTGCA-rev pTxnip269_xho1 Transfer promoters CCGCTCGAGACAGCCCCCTCCTTCCC from pGL3‐Txnip pTxnipR_BamH1 (WT or Mut) to GATGGATCCCTCCAAATCGAGGAAACCG pEGFP1 mChREBP-EcoRI-F Primers for cloning TCTGAATTCCTTCTTCCTGAAGACCCTA of mouse ChREBPζ
168 Appendices
Primers Remarks mChREBP-EcoRI-R Primers for cloning TAGGGTCTTCAGGAAGAAGGAATTCAGA of mouse ChREBPζ mChREBP-Noc1-F CTGCCCCATGGGTACCTGGAACCCGTCT mChREBP-Noc1-R AGACGGGTTCCAGGTACCCATGGGGCAG mChREBP-XhoI CACCTCGAGATGGCGCGCGCGCTGGCGGATCTATCCGTGAA C mChREBP-Mlu1 TACACGCGTTATAATGGTCTCCCCAGGGTGCCCTCTGTGAC TGC mChREBP_seq680 TGCTGCTTGGGGGCTCCGA mChREBP_seq1170 Primers for ChREBP CCAAGATCCCACCTGCTCC sequencing mChREBP_seq1510 CAAGCCCTCCTCCCCATCC mMlx-xho1 CACCTCGAGATGACGGAGCCGGGCGCCTC Primers for cloning mMlx-xba1 of mouse MLXβ CGCTCTAGATCAGTAGAGTTGGTTTTTCAACTG mMLX_R88S Primers for making AGTGATGCTATTAAGAGAGGCTAT mMLX_R87A dominant negative TGCCTTCTGTTCAGCCTGAGTGTG mouse MLX Mlx-Xho1 TCGCTCGAGATGACGGAGCCGGGCGCCTCTC Mlx-Stop-Not1 GTAGCGGCCGCTCAGTAAAGCTGGTTTTTCAATTGGTGCAG Primers for cloning Mlx-EcoRVf(human) of human MLX TCTGATATCATGACGGAGCCGGGCGCCTCTC Mlx-not1r(human) GTGGCGGCCGCACGTAAAGCTGGTTTTTCAATTGGTGCAG mUSF1f-mlu1 Primers for cloning TACACGCGTAATGAAGGGGCAGCAGAAAACAGC mUSF1r-not1 of human or mouse TACGCGGCCGCTTAGTTGCTGTCATTCTTGATGACGAC USF (USF1 or mUSF2f-mlu1 USF2). TACACGCGTAATGGACATGCTGGACCCGGGTC mUSF2r-not1 TACGCGGCCGCTCACTGCCGGGTACTCTCGCC USF1f-sMlu1 TACACGCGTAATGTACAGGGTGATCCAGGTGTCTG USF1r-not1 Primers for cloning TACGCGGCCGCTTAGTTGCTGTCATTCTTGATGACGAC of human or mouse USF1f-lMlu1 USF (USF1 or TACACGCGTAATGAAGGGGCAGCAGAAAACAGC USF2f-mlu1 USF2).
169 Appendices
Primers Remarks TACACGCGTAATGGACATGCTGGACCCGGGTC USF2r-not1 TACGCGGCCGCTCACTGCCGGGTGCCCTCGC USF1f-lNde1 Primers for making TACCATATGAAGGGGCAGCAGAAAACAGC His‐USF1 USF1r-BamH1 TACGGATCCTTAGTTGCTGTCATTCTTGATGACGAC (pET vector)
MondoA-EcoRVf TCTGATATCATGGCCGCCGACGTCTTCATGTGCTC MondoA-not1r GTGGCGGCCGCACGGACTCTCCCAATCTCTTGCCAATCC MondoA-EcoRIf ATGTCGGAATTCAGCGACACCCTCTT MondoA-EcoRIr Primers for cloning AAGAGGGTGTCGCTGAATTCCGACATMondoA-Bgl2f of human MondoA GCATCCTGGTGACAGATCTCGGCCAT MondoA-Bgl2r ATGGCCGAGATCTGTCACCAGGATGC MondoA-Xho1 TCGCTCGAGATGGCCGCCGACGTCTTCATGTGCTC MondoA-Stop-Not1 GTAGCGGCCGCCTAGGACTCTCCCAATCTCTTGCCAATCC MondoA_Seq1 CTGGACGGCTCTGTGGACG MondoA_Seq2 ATCCCAACAACCCACCTGC MondoA_seq3 GCTGGCACTGTCTCCTGTC Primers for MondoA_seq4 MondoA ATCGGAGCAGAGCCCCAGT MondoAseqR1 sequencing CCATCCATCCCCGTGCTTGT MondoAseqR2 AAGGGCTGCGGGACACTCA MondoAsewR3 TTGGACTGGGACACGGTGG MondoA-Not1r Insert MondoA GTAGCGGCCGCCCGGACTCTCCCAATCTCTTGCCAATCC Δ237 into cas12 Δ237MondoA-not1 CATGCGGCCGCCCGATGAGGACCTCTCCAGCCTGGTC vector CTGGCAAGTCCGACCCCAAAA Primers for making AACAATTATTCTGGGGAGATTGAG the MondoAr gstMdaEcR1 Primer for making TACGAATTCACGGTCAACAAACAGACGTGCC GST‐MondoA gstMlxEcr1 Primer for making TACGAATTCCAGAAGAGGAGGGACGCCATC GST‐MLX
170 Appendices
Primers Remarks NFYA-xho1 TACCTCGAGATGGAGCAGTATACAGCAAACAGCA Primers for cloing of NFYA-not1 human NF‐YA TACGCGGCCGCTTAGGACACTCGGATGATCTGTGT pCI-rev (seq) gtatcttatcatgtctgctcg Sequencing primers pCI-rev-2 (seq) for pCI‐neo vector CCCTGAACCTGAAACATAAA (pdHA, pdFLAG or pCI-SEQfar pdMyc) CTTACTGACATCCACTTTGC GCCGACAGGATGCAGAAGGAGATCA β‐actin AAGCATTTGCGGTGGACGATGGA RT‐PCR primers
GGCGGGTGTCTGTCTCTGCT Human Txnip GGCAAGGTAAGTGTGGCGGG RT‐PCR primers
GACCCCGATGATGAGGACAG MLX TGTAGAACGATGGCTTTGCTG RT‐PCR primers
CCTCCACCGTGTCCCAGTC MondoA CTGGGGCTCTGCTCCGATG RT‐PCR primers
TCGGACACAGACTCGGAGG ChREBP AGAGGCGTGTGAGTGTGGG RT‐PCR primers
GAAGTTACCCGAGTCAAAGC Mouse Txnip CGCAAGTAGTCCAAAGTCTG RT‐PCR primers
ATGGCACTGGCGGCAGGTCC Mouse cyclophilin TTGCCATTCCTGGACCCAAA RT‐PCR primers
171 Appendices
Appendix V, Paper 1 (Abstract)
Yu FX, Goh SR, Dai RP, Luo Y. Mol Endocrinol. 2009 Jun;23(6):932‐42 Adenosine‐containing molecules amplify glucose signaling and enhance txnip expression.
Eukaryotic cells sense extracellular glucose concentrations via diverse mechanisms to regulate the expression of genes involved in metabolic control. One such example is the tight correlation between the expression of thioredoxin‐interacting protein (Txnip) and extracellular glucose levels. In this report, we show that the transcription of the Txnip gene is induced by adenosine‐containing molecules, of which an intact adenosine moiety is necessary and sufficient. Txnip promoter contains a carbohydrate response element, which mediates the induction of Txnip expression by these molecules in a glucose‐dependent manner. Max‐like protein X and MondoA are transcription factors previously shown to stimulate glucose‐dependent Txnip expression and are shown here to convey stimulatory signals from extracellular adenosine‐containing molecules to the Txnip promoter. The regulatory role of these molecules may be exerted via amplifying glucose signaling. Hence, this revelation may pave the way for interventions aimed toward metabolic disorders resulting from abnormal glucose homeostasis.
172 Appendices
Appendix VI, Paper 2 (Abstract)
Yu FX and Luo Y. PLoS ONE. 2009, Dec; 4(12): e8397 Tandem ChoRE and CCAAT motifs and associated factors regulate Txnip expression in response to glucose or adenosine‐containing molecules.
Thioredoxin interacting protein (Txnip) is a multifunctional protein involved in regulation of cell cycle events and cellular metabolism. The expression of Txnip is known to be induced by glucose, adenosine‐containing molecules, HDAC inhibitors and other physiological cues; however, the underlying regulatory mechanisms remain elusive. Here, we show identification of an additional carbohydrate response element (ChoRE) on the promoter of Txnip gene, which functions cooperatively with the earlier identified ChoRE to mediate optimal Txnip expression. We also show that the function of ChoREs and associated factors is contingent on tandem CCAAT motifs, in that the occupancy of the Txnip promoter by the CCAAT‐box‐associated nuclear factor Y (NF‐Y) is a prerequisite for efficacious recruitment of MondoA/MLX to ChoREs. Such a strategy suggests a synergy between NF‐Y and MondoA/MLX in enhancing Txnip expression presumably through inducing dynamic chromatin changes in response to diverse physiological inducers.
173