Pparγ and Smad2 Mediate Ski Induced Energy Metabolism Shift and Oncogenic Transformation

PPARΓ AND SMAD2 MEDIATE SKI INDUCED ENERGY METABOLISM SHIFT AND ONCOGENIC TRANSFORMATION

FANG YE

Submitted in partial fulfillment of the requirements

For the degree of Doctor of Philosophy

Advisor: Dr. Edward Stavnezer

Department of Biochemistry

CASE WESTERN RESERVE UNIVERSITY

January, 2011

CASE WESTERN RESERVE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis/dissertation of

______

candidate for the ______degree *.

(signed)______(chair of the committee)

______

(date) ______

*We also certify that written approval has been obtained for any proprietary material contained therein. Table of Contents

List of Figures……………………………………………………………………….4

List of Tables ……………………………………………………………...... 6

Abstract ...... 8 Introduction ...... 10 Structure and Molecular Interactions of Ski Protein………………………….10 Biological Functions of Ski……………………… ……………………………….14 Tumorigenesis ...... 14 Hematopoietic transformation ...... 16 Knowledge from animal models ...... 17 In vitro biology ...... 19 Overview of the Role of PPARγ in Lipid Metabolism………………………...19 Smads and Ski…………………………………………………………………...23 Our Research Focus…………………………………………………………….28 Materials and Methods ...... 29 Materials…………………………………………………………………………..29 1. Plasmids ...... 29 2. Primers and synthetic oligos ...... 31 3. Cells ...... 35 4. Reagents ...... 35 Methods…………………………………………………………………………...45 1. DNA Cloning ...... 45 2. Tissue Culture and Transfection ...... 48 3. Retroviral infection ...... 49 4. Western blotting ...... 49 5. Quantitative Realtime-PCR ( Q-RTPCR) ...... 50 6. Reporter assays ...... 52 7. Genomic DNA preparation ...... 53

8. Immunoprecipitation ...... 54 Results and Discussion ...... 55 Chapter I PPARγ Mediates Ski Induced Energy Metabolism Shift…….55 Abstract ...... 55 Introduction…………………………………………………………………….....53 Materials and Methods…………………………………………………………..59 Establishment of PPARγKD-Ski-CEF Strains and Cell Culture ...... 59 Lactate and Glucose Assay ...... 60 Mitochondria related assays ...... 60 Flow cytometry ...... 61 Measurement of triglyceride synthesis by stable isotopes in cells ...... 62 Quantitative Realtime-PCR ...... 63 Western Blotting ...... 64 Metabolite Oxidation ...... 65 Reporter Assay ...... 65 Results…………………………………………………………………………….66 Glycolysis is suppressed in Ski-CEFs ...... 66 TCA cycle is enhanced in Ski-CEFs...... 72 Ski-CEFs exhibit enhanced mitochondrial respiration ...... 75 Ski Increases mitochondrial number and mass in CEFs ...... 85 Expression of PPARγ and lipid metabolic genesregulated by PPARγ are increased by Ski...... 94 Ski is a co-activator of PPARγ dependent transcription ...... 99 Ski increases β-oxidation of fatty acids ...... 103 PPARγ mediates the induction of lipid oxidation and mitochondrial mass by Ski .. 107 Discussion……………………………………………………………………….120 Chapter II Smad2 Mediates Ski Induced Transformation in CEFs………127

Abstract ...... 127 Introduction………………………………………………………………………….128 Experimental Methods………………………………………………………….132 Expression Vector Construction ...... 132 Tissue Culture and Cell Strain Establishment ...... 133 2

Western Blotting ...... 134 Immunostaining ...... 135 Antibodies ...... 136 Results…………………………………………………………………………...136 TGF-β independent nuclear localization and altered expression of Smads in Ski-CEFs ...... 136 Ski blocks TGF-β inhibition of CEF proliferation...... 143 Smad2 is upregulated and receptor-phosphorylated in the absence of added TGF-β in Ski-CEFs ...... 145 Smad2 phosphorylation in Ski-CEFs is not blocked by neutralizing antibodies toTGF-β...... 148 Smad2 phosphorylation in Ski-CEFs is not inhibited by blocking signaling through the activin or MAP kinase pathways ...... 151 Increased expression and nuclear localization of Smad2 correlate with transformation by Ski ...... 153 Re-establishing high Smad3 expression does not affect Smad2 localization or cellular transformation ...... 156 Knocking down Smad3 does not affect Ski induced transformation ...... 164 Constitutively active Smad2, does not complement transformation defective mutants of Ski ...... 167 Discussion……………………………………………………………………….172 Discussion and Future Directions ...... 176 Ski-Smad2 Function………………………………………………………………..176 PPARγ and Ski Interaction...………………………………….…………………..181 Glycolysis and PPARγ……………………………………………………………..183 Downstream mediators of Ski induced metabolic shift…………………………183 Energy production pathways……………………………………………………...184

List of Figures

Figure 1 Structure of Ski Protein ...... 13

Figure 2 A&B Lactate Production in Ski-CEFs ...... 67

Figure 2 C&D Glucose Utilization in Ski-CEFs ...... 69

Figure 2 E Alamar blue Reduction in Ski-CEFs ...... 71

Figure 3 A&B Oxidation of Acetate and Glucose in Ski-CEFs ...... 73

Figure 3 C Oxidation of Glutamate in Ski-CEFs ...... 74

Figure 4 A Heat Generation in Ski-CEFs ...... 76

Figure 4 B&C Oxygen Consumption of Intact Ski-CEFs ...... 77

Figure 4 D Oxygen Consumption of Permeabilized Ski-CEFs line graph ...... 81

Figure 4 E Oxygen Consumption of Permeabilized Ski-CEFs bar graph ...... 82

Figure 4 F UCR Ratios of Intact and Permeabilized Cell Respiration ...... 84

Figure 5 A Ski Increases Mitochondrial Protein Amount ...... 87

Figure 5 B Citrate Synthase Activity in Ski-CEF and PPARγKD Cells ...... 88

Figure 5 C Mitochondrial Mass of Ski-CEFs by Flow Cytometry ...... 89

Figure 5 D Electron Microscopic Iamges of Ski-CEFs ...... 91

Figure 5 E Measuring Mitochondrial DNA Amount in Ski-CEFs and CEFs ...... 92

Figure 6 A&B Ski Upregulates PPARγ at mRNA and Protein Level ...... 96

Figure 6 C Ski Increases the Expression of PPARγ Targets in Ski-CEFs ...... 98

Figure 6 D&E Ski Coactivates PPARγ Responsive Reporter ...... 100

Figure 6 F Ski Interacts with PPARγ ...... 102

Figure 7 A&B Ski Stimulates β-oxidation ...... 104

Figure 7 C&D Ski Reduces De Novo Synthesis of Palmitate and TG...... 106

Figure 8 A Screening PPARγKD-Ski-CEF Clones ...... 108

Figure 8 B Examine Alamar blue Reduction Activity of PPARγKD Clones...…….109

Figure 8 C The Expression of PPARγ Target Genes in PPARγKD Cells ...... 111

Figure 8 D&E Lacate Production and Glucose Utilization in PPARγKD Cells...... 112

Figure 8 F Intact Cell Repiration of PPARγKD Cells ...... 114

Figure 8 G Permeabilized Cell Respiration of PPARγKD Cells ...... 115

Figure 8 H Mitochondrial Protein Level of PPARγKD Cells ...... 117

Figure 8 I Mitochondrial Mass of PPARγKD Cells by Flow Cytometry ...... 118

Figure 8 J Electron Microscopic Images of PPARγKD Cells ...... 119

Figure 8 K Metabolism Related Features of ΔZ3/4 Ski Mutant ...... 125

Figure 9a Localization and Expression Patterns of Smad2 & 3 in Ski-CEFs ...... 138

Figure 9b TGF-β Independent Nuclear Localization of Smad2 in Ski-CEFs ...... 140

Figure 9c Ski is Resistant to TGF-β Induced Degradation ...... 142

Figure 10a Ski Reverses TGF-β Induced Growth Inhibition ...... 144

Figure 10b Smad2 is Phosphorylated in the Absence of Exogenous TGF-β ...... 147

Figure 11 Smad2 Phosphorylation is TGF-β Independent in Ski-CEFs ...... 149

Figure 12 Smad2 Phosphorylation is Activin or ERK1/2 Independent in Ski-CEFs .. 152

Figure 13a Schematic Structure of Transformation Defective Mutants of Ski ...... 154

Figure 13b Expression and Localization Patterns of Smad2/3 in Mutant-Ski-CEFs. 155

Figure 13c&d Increasing Smad3 Level Cannot Reverse Ski Induced Transformation.. 157

Figure 14a&b Screening Smad2 Knockdown Ski-CEF Clones ...... 160

Figure 14c&d Smad2 is Essential for Ski Induced Transformation ...... 162

Figure 15 Smad3 is not Required for Ski Induced Transformation ...... 165

Figure 16a Smad2E2 Cannot Rescue the Transformation Defective Mutants of Ski.169

Figure 16b Expression of Ski Mutants with Smad2E2 in CEFs ...... 170

List of Tables

Table 1 List of Lipid Metabolic Genes Regulated by Ski ...... 95

Acknowledgements

This dissertation could not have been written without supports from my advisor,

Dr. Ed Stavnezer, my thesis committee and my collaborators. Dr. Ed Stavnezer provided me the opportunity to do research in his laboratory. He encouraged and challenged me throughout my academic program. His knowledge, advice and the great patience are highly appreciated. It is worth to point out that he is the first pure scientist I have ever been in the close contact with. He enjoys the whole process of the research rather than pursuit of the outcome. Thank you, Ed, for leading me to the true spirit of science.

I would like to extend my gratitude to all the thesis committee members: Dr.

David Samols, Dr. Colleen Croniger, Dr. Hung-Ying Kao and Dr. Charles Hoppel,

for their valuable advice and criticisms which helped me pass numerous

checkpoints through my PhD study. A special thanks to Dr. Dave Samols for

being my first reader and all the revisions and corrections he made in a timely

fashion, which helps me significantly to complete my thesis.

I would like to thank all my collaborators, Dr. Richard Hanson and his group, Dr.

Charles Hoppel, Dr. Helene Lemieux, Dr. Colleen Croniger, Dr. Michelle

Puchowicz and Dr. Vernon Anderson, for their precious advice, skills and criticisms in the fields that are beyond our lab normal expertise. Without them, my research project could not have been as complete as it become.

PPARγ and Smad2 Mediate Ski Induced Metabolic Shift and

Oncogenic Transformation

Abstract

FANG YE

Ski is a versatile nuclear oncoprotein that regulates transcriptions of genes involved in multiple biological processes. Ski induces oncogenic transformation of chicken embryo fibroblasts (CEF). This biological function has been implicated with Ski’s ability to inhibit transforming growth factor β (TGF-β) signaling through interactions with Smad proteins via Dachshund homology domain (DHD) and

SAND (Sp100, AIRE-1, NucP41/75 and DEAF-1) domains by previous studies. In

Ski transformed CEFs (Ski-CEF), we discovered that Ski reversed the normal

Smad3 to Smad2 protein ratio, by increasing the Smad2 protein level while lowering the Smad3 level. Smad2 translocated from cytoplasm into nucleus in the absence of TGF-β stimulation. Moreover, Smad2 was constantly phosphorylated by some mechanism independent of TGF-β. Knocking down endogenous Smad2, but not Smad3 completely blocked the Ski induced oncogenic transformation judged by morphology and anchorage independent

Ski-CEFs display transformed cell phenotypes, but not the enhanced glycolysis in the presence of adequate oxygen: the Warburg effect that is the classical metabolic feature of tumor cells. On the contrary to making a lot of lactate from glucose, Ski-CEFs are slow to acidify culture medium. We unveiled that compared to CEFs; Ski-CEFs produced less lactate and utilized less glucose by suppressing glycolysis. In parallel, Ski-CEFs stimulated fatty acid oxidation, mitochondrial respiration and increased mitochondrial number and mass.

Interestingly, we found that PPARγ, a potent lipid metabolism regulator was dramatically elevated at protein and message level and activated as a transcription regulator in Ski-CEFs. PPARγ target genes involved in lipid transport, fatty acid oxidation and mitochondrial biogenesis were also upregulated by Ski. Knocking down PPARγ by RNAi in Ski-CEF reversed Ski induced fatty acid oxidation, mitochondrial respiration, mitochondrial biogenesis and restores glucose utilization. We conclude that PPARγ mediates the energy metabolism shift induced by Ski.

Key words: Ski, TGF-β, Smad2, Smad3, PPARγ, β-oxidation, transformation, metabolism, mitochondria, oxygen consumption

Introduction

Structure and Molecular Interactions of Ski Protein

Chicken c-Ski is a versatile oncogene. It was accidentally discovered when

chicken cells infected with a supposedly transformation defective retrovirus

displayed a transformed phenotype. An avian retrovirus, Sloan Kettering virus

(SKV) containing v-Ski was the cause of the transformation of these chicken cells

[1]. v-Ski is a truncated form of c-Ski. It is composed of 438 amino acids (aa)

corresponding to 20-458 region of the 750 aa long c-Ski [2, 3]. When expressed

in a retroviral vector, c-Ski also transforms chicken embryo fibroblasts (CEF) at a relatively higher level compared to v-Ski. v-Ski contains most of the N-terminal

304 residues encoded by the first coding exon of c-Ski. This exon 1 encoded sequence is both sufficient and essential for Ski’s transformation activity and other biological functions. Many deletions and point mutations in this region

completely abolish Ski’s biological activities [4-6].

c-Ski is universally expressed in all tissues of species from Drosophila to Homo sapien. The Exon 1 encoded region is the most evolutionarily conserved sequence and consists of a proline-rich region, several α-helixes and β-sheets

(residues 91-175) and several cysteine-histidine-rich regions that are features of zinc-finger like motif (residues 202-295) [7-9]. Mutational analysis of exon1 revealed that several of the above featured structural regions are important for

Ski’s biological functions, except for the proline-rich region [10]. Two distinctive

structural domains that are highly evolutionarily conserved were identified in exon

1 encoded region and they are characteristic of the Ski protein family which

consists of seven members, Ski, SnoN, Fussel-15, Fussel-18, Dach, COR1 and

LBXCOR1 [11] (Fig. 1). Both domains have been crystallized and the corresponding 3-D structures with associated motifs were resolved and defined by the X-ray crystallography [7, 12]. The more N-terminal domain is the dachshund homology domain (DHD) that contains several α-helixes and β- strands [7]. The helix-turn-helix motifs followed by the β- strands of the Ski DHD bear structural similarity with the forkhead/winged-helix family of DNA binding proteins [7, 13]. Deletions within the DHD domain completely eliminated Ski’s ability to repress transcription and transform CEFs [6, 14]. The second domain composed of a zinc-finger motif is highly homologous to the SAND (Sp100,

AIRE-1, NucP41/75 and DEAF-1) domain [10, 12] that mediates DNA binding by these proteins [15]. Point mutation or deletion in the SAND domain also causes the loss of transformation ability of Ski [12, 16]. Although the DHD and SAND domains are responsible for protein-DNA interactions in other proteins, in Ski, they mediate protein-protein interactions between Ski and other transcription factors.

The C-terminal region (residues 559-750) of c-Ski, which is truncated in v-Ski, is not essential for the biological functions of Ski although it is responsible for interactions with other proteins [17-19]. It is also responsible for high affinity

dimerization of Ski proteins [20]. Two structural motifs responsible for the

dimerization were discovered in full length Ski. The more N-terminal one consists

of five tandem repeats of 25 aa each, which is followed by the second motif, a

predicted leucine zipper that consists of six heptad repeats. The tandem repeat

motif is indispensable to mediate homo-dimerization of Ski and heterodimerization with the closely-related protein, SnoN [10, 20]. The leucine zipper motif increases the binding affinity for dimerization. Although lacking these C-

terminal motifs, v-Ski can still dimerize but with reduced efficiency. In fact, the potency of Ski’s transforming capacity has been attributed to this dimerization efficiency [10, 20].

Since Ski does not bind to DNA directly [21, 22], it interacts with several other

transcription factors via the DHD or SAND domains to regulate gene expression

as a transcription co-regulator. Ski binds to the nuclear receptor co-repressor (N-

CoR), the retinoic acid receptor (RAR), Ski interacting protein (SKIP), Smad2,

Smad3, GATA, PU.I and Six1 via interaction sites in the DHD domain [23-31].

Through the SAND domain, Ski interacts with Smad4, four and half LIM domain

protein 2 (FHL2), retinoblastoma protein (Rb) and methyl CpG binding protein 2

(MeCP2) [9, 17, 32, 33]. Ski also interacts with nuclear factor 1 (NF1),

MDS1/EVI1-LIKE GENE 1 (MEL1) and vitamin D receptor (VDR), however, the

specific interaction sites between Ski and these proteins has not yet been

mapped [22, 34-36]. In addition, Ski has been reported to bind to c-Jun via the

region between the SAND domain and the C-terminal region to suppress Smad2

Fig. 1 Structure of Ski Protein

Fig. 1 Structure of Ski protein.

A schematic cartoon of c-Ski and v-Ski protein structure. Regions of Ski known to participate in protein-protein interactions are indicated by dark boxes. Symbols indicate: Pro, proline-rich region; AH/BS, mixed α helixes and β strands; Zn, putative zinc fingers; TR, tandem repeat; LZ, leucine zipper motif. The locations of these conserved regions within Ski protein are indicated above the bar in number. Domains required for binding to Ski protein partners are also indicated in the figure. DHD: Dachshund homology domain, SAND: Sp100, AIRE-1,

NucP41/75 and DEAF-1 domain. Individual Protein interaction sites are indicated by bars.

transcriptional activity [37]. By these means, Ski acts as a transcription co- activator or a co-repressor. However, not all the Ski-protein interactions via

SAND domain are related to transcriptional regulation. Arkadia, an E3 ubiquitin ligase targets Ski and SnoN for degradation via binding to the SAND domain [38,

39].

Biological Functions of Ski

Tumorigenesis

Ski is known to block transforming growth factor-β (TGF-β) signaling via interactions with Smad proteins [4, 25, 27, 40, 41]. TGF-β is a potent inhibitor of epithelial cell growth and repression of TGF-β signaling results in tumor progression [42]. The inhibition of TGF-β signaling by Ski is relevant in several human cancers. In human pancreatic cancer cells, Panc-1, the expression of Ski is correlated with cancer cell proliferation. Silencing Ski by RNA interference

(RNAi) restores TGF-β signaling in vitro and decreases tumor growth in vivo [43].

Similar results were obtained from two other pancreatic cancer cell lines,

SW1990 and BxPC3 [44]. Surprisingly, Ski also acts as a metastasis suppressor in these pancreatic tumor cells since knocking down Ski lead to enhanced cell invasiveness in vitro and lung metastasis in vivo [44]. Interestingly, in the same tumor cells, Ski was shown to be the target of a therapeutic herbal drug that has anti-tumor effect by the same research group [45].

In a TGF-β responsive human gastric cancer cell line, OCUM-2MLN,

overexpression of Ski caused the inhibition of TGF-β signaling in vitro and

subcutaneous xenograft of Ski overexpressing OCUM-2MLN cells into

BALB/cnu/nu mice resulted in extensive acceleration of tumor growth and increased angiogenesis in vivo [46]. In MKN28 human gastric cancer cells, SKI and MEL1 were abnormally overexpressed as a result of 1p36.3 chromosome amplification that is a common cause of tumorigenesis. Surprisingly, Ski was found to directly interact with MEL1 to inhibit the TGF-β signaling. Knocking down both Ski and MEL1 decreased tumor growth in vivo and synergistically restored

TGF-β responses [35]. In human melanomas, overexpression of Ski is correlated with melanoma progression [47]. Again, knocking down Ski by RNAi generated minimal tumor xenograft nodules that grow slowly and activated TGF-β signaling indicated by high levels of nuclear Smad3 and P21 [48]. In this melanoma, Ski not only suppresses TGF-β signaling but also binds to FHL2 to promote the transcriptional activation of genes regulated by β-catenin, resulting in the activation of Wnt signaling. Knocking down of Ski inhibits melanoma growth in vivo [9, 47, 48]. In liver cancer cells, Ski forms a complex with MeCP2 and negatively regulates transcription of tristetrapolin (TTP), a TGF-β induced protein that down-regulates c-Myc at the posttranscriptional level [49]. In addition, the expression of Ski or SnoN as a TGF-β signaling repressor is correlated with colorectal carcinogenesis and esophageal squamous cell carcinoma progression

[50, 51]. Injection of Ski-CEFs into chicken results in squamous cell carcinomas

[52]. Fifty percent of the injected chicken develops palpable nodules at the subcutaneous site in the wing web within two weeks. These nodules progress slowly into open lesions. Examined by histopathological assays, these lesions are clearly moderately differentiated cutaneous squamous cell carcinomas [52].

Hematopoietic transformation

Hematopoietic progenitor cells normally undergo differentiations into the erythroid

and myeloid lineages [53]. Disrupting the transcriptional control of normal

hematopoietic differentiation results in the uncontrolled growth of myeloid

progenitor cells in bone marrow and an arrest in their maturation [54]. These

transformed multipotent myeloid cells cause a blood cell tumor, acute myeloid

leukemia (AML) [55]. RAR signaling is necessary for hematopoietic cells to

maintain their differentiation ability under physiological concentrations of retinoic

acid. Ski transforms bone marrow-derived myeloid/erythroid precursors [23] and

is involved in hematopoietic tumor formation by a mechanism unrelated to

interference with the TGF-β signaling [23, 56]. In these cells, Ski binds to RAR

and recruits N-CoR to repress transcription mediated by RAR responsive

element [23]. High dosage of retinoic acid reversed the Ski induced

hematopoietic transformation, indicating the requirement for RAR signaling.

Interestingly, SKI is expressed at high levels in a subset of human AML patients

that do not express PML-RAR, a RAR fusion protein that dominant-negatively

blocks RAR induced differentiation [54]. In these leukemias, Ski provides the

same function as the fusion partners by acting as a dominant co-repressor of

RAR regulated transcription [54]. Furthermore, Ski binds to GATA to block

erythroid differentiation [30] and binds to PU.I to negatively regulate myeloid

differentiation into macrophages [31]. These data suggest that Ski expression

contributes to erythroleukemia and myeloid leukemia. Interestingly although co-

repression of RAR by Ski is dependent on N-CoR, this is not the case for Smad2

co-repression [36]. Thus Ski promotes oncogenesis through interactions with different signaling pathways, involving different families of transcription factors

and functioning as a transcriptional repressor by different mechanisms.

Knowledge from animal models

As to animal models, Ski knockout mice on the 129 background show a high penetrance of exencephaly while knockout on the C57BL6/J background shows a high penetrance of severe midline facial clefting [57]. In addition, other craniofacial and skeletal defects including facial flattening, broad forehead, malformation of the eye and depressed nasal bridge show increased penetrance on the C57BL6/J background. These phenotypes resemble defects found in individuals affected by human 1p36 deletion syndrome, a disorder caused by the deletion of a region close to the tip of the short arm (1p) of human chromosome 1

[58-61]. Interestingly, the human SKI gene maps at chromosome 1p36.2 and is deleted in all of the patients with the syndrome [57]. It is likely that loss of SKI accounts for some of the observed craniofacial defects common in 1p36 deletion syndrome. Besides these craniofacial developmental defects, Ski knockout mice

exhibit a remarkable decrease of both skeletal muscle mass and brown adipose

tissue (BAT) [57, 62]. Transgenic mice overexpressing v-Ski or c-Ski cDNA show

selectively dramatic hypertrophy of type II fast skeletal muscle fibers but not type

I slow fibers [63, 64]. These findings from animal models are consistent with observations of muscle differentiation induction by Ski in vitro [29].

Recent studies of the hypertrophied type II fast muscles of Ski transgenic mice

revealed a mitochondrial abnormality [65] and enhanced lipid oxidation compared

to the muscles of wild type mice [66]. The expression of fatty acid biosynthesis

related genes such as FASN and SCD-1 are decreased as well as the

expression of a master lipid metabolism regulator, peroxisome proliferator-

activated receptor gamma (PPARγ). Interestingly, Ski +/+ mice run for longer

time on the treadmill than the Ski +/- mice do (Data by Stavnezer, E., Hanson, R.,

Colmenares, C., and Hakimi, P.). Exercise not only improves insulin sensitivity

and glucose transport into cells [67-69] but also increases adaptive oxidative

capacity and lipid metabolism of skeletal muscle [70, 71]. The induction of Ski

expression has been described in people who received endurance training,

compared to the control group [72]. Taken together, these data suggest that Ski

might play a role in regulating lipid metabolism perhaps mediated by affecting the

expression or the transcriptional activity of PPARγ.

In vitro biology

In vitro studies have shown that Ski also plays a role in muscle terminal

differentiation. Muscle differentiation can be induced by Ski in quail embryo

fibroblasts (QEF) [73] and in the mouse myoblast cell line (C2C12). Silencing Ski

in C2C12 cells, effectively blocked myotube formation by preventing the induction

of the key muscle regulatory gene, Myogenin [29]. The molecular mechanism

underlying this biological process is mediated by Ski binding to Six1 and Eya3 to

form a transcriptional activating complex on the Myogenin promoter region.

Therefore, Ski turns on the transcription of Myogenin that is a major regulator of

muscle differentiation [29]. Inhibition of TGF-β signaling suggests that Ski is also involved in inducing muscle differentiation, since Smad3 represses MyoD dependent myogenic transcription [74]. Besides connection with muscle differentiation and oncogenesis, Ski blocks bone morphogenic protein (BMP) signaling through interactions with Smad4, Smad1 and Smad5, [32, 75-78]. Thus

Ski is linked to mesoderm patterning development and bone morphogenesis [79].

Overview of the Role of PPARγ in Lipid Metabolism

The peroxisome proliferator-activated receptors (PPARs) belong to a subfamily

of nuclear receptors. Three members PPARα, PPARδ and PPARγ, encoded by

three separate genes, are activated by lipid ligands to regulate metabolic

homeostasis in different ways [80, 81]. PPARs heterodimerize with retinoid X

receptor (RXR) and form a complex that binds to peroxisome proliferator 19

response elements (PPRE) in the regulatory regions of target genes. Upon ligand

activation, the PPAR-RXR complex conformation is changed such that it binds to

coactivators with high affinity leading to histone acetylation and chromatin

remodeling. Thus, transcription of PPARs responsive genes is induced [82].

PPARα has a crucial role in controlling fatty acid oxidation in liver and muscle

[83], since PPARα null mice are incapable of meeting energy production demands during fasting and suffer from hypoglycemia, hyperlipidemia, hypoketonemia and fatty liver [84]. After an overnight or prolonged fast, PPARα is strongly induced in liver by fatty acids released from adipose tissue and transported into the liver. The activation of PPARα by fatty acids promotes

hepatic fatty acid oxidation to generate ketone bodies and to provide an energy

source for peripheral tissues [97].

PPARδ powerfully regulates fatty acid catabolism and energy homeostasis [85,

86]. PPARδ deficient mice fed with a high-fat diet are prone to obesity. In parallel,

tissue specific overexpression of an activated form of PPARδ in adipose tissue

result in lean mice which are resistant to obesity, hyperlipidemia and tissue

steatosis induced by a high-fat diet [87]. Interestingly, overexpression of a

constitutively active PPARδ in skeletal muscles of transgenic mice caused an

increase of oxidative muscle fiber and enhancement of exercise endurance by

nearly 100% in untrained adult mice [88]. Moreover, activation of PPARδ by

agonist with exercise training synergistically increased oxidative myofibers and

running endurance in adult mice [89].

However, unlike PPARδ and PPARα which are directly involved in fatty acid

oxidation, PPARγ induces adipogenesis, regulates whole body glucose

homeostasis and regulates lipid metabolism in two directions [90, 91]. Activation

of PPARγ induces adipocyte differentiation and lipid accumulation by modulating

genes regulating adipogenesis, lipid uptake and lipid metabolism. PPARγ

knockout mice fail to develop adipose tissue [92-94]. Consistent with these findings, humans with dominant-negative mutations in a single allele PPARG

have partial lipodystrophy and insulin resistance [95-97]. Many aspects of the

PPARγ‘s role in adipocyte differentiation and lipid metabolism actually followed

the discovery that an anti-diabetes drug, thiazolidinedione (TZD) is a high affinity

agonist of PPARγ [98]. TZD were found to be effective at treating type 2 diabetes.

The discovery that TZDs are PPARγ agonist connects PPARγ activity to insulin sensitivity and also poses a paradox because obesity is a major risk factor in type 2 diabetes and PPARγ is an adipogenic transcription factor. An unwanted side effect of TZD is that the modest body weight gain at least by increased adipogenesis and net fatty acid influx into adipose tissues. One plausible hypothesis provides an explanation to this paradox. In type 2 diabetes affected bodies, partitioning of free fatty acid (FFA) circulating among adipose tissue, liver and skeletal muscle is abnormally altered. The imbalance results in intracellular accumulation of triglycerides and fatty acids in these tissues and causes insulin

resistance [99]. TZDs reduce circulating FFA levels and improve insulin

sensitivity in muscle and liver by activating PPARγ to promote adipocyte

differentiation FFA uptake and lipid storage in adipose tissues. Genes functioning in lipid or FFA transport have been identified as PPARγ targets [100, 101].

Among PPARγ regulated genes, fatty acid binding protein 4 (FABP4), CD36, a receptor and transporter for lipids, lipoprotein lipase, an enzyme thst hydrolyses lipoproteins and fatty acid transporter 1(FATP-1) are involved with lipid uptake, transport and storage [102, 103]; Glycerol kinase (Gyk), sterol regulatory element binding protein 1 (SREBP-1) and stearoyl-Coenzyme A desaturase 1 (SCD-1), a regulator of unsaturated fatty acid synthesis, are PPARγ target genes involved with lipogenesis. On the other hand, activation of PPARγ stimulates the expression of adiponectin, adiponectin receptor and ACOX-1, all of which

promote fatty acid oxidation, though the latter two genes are common targets of

both PPARγ and PPARα [104-108].

PGC1-α is a transcriptional co-activator of PPARγ and itself is also transcriptionally activated by PPARγ [109]. Only by interacting with PGC1-α, can

PPARγ induce brown adipose tissue differentiation [110]. PGC1-α not only plays a critical role in promoting fatty acid oxidation in skeletal muscle but drives the formation of type I slow muscle fibers [111]. Besides stimulating lipid catabolism,

PGC1-α is sufficient to induce mitochondrial biogenesis [112].

Treating rats with a highly potent and selective PPARγ agonist (GW1929) resulted in increased expression of genes involved in lipogenesis, fatty acid transport, lipid storage and oxidation in both WAT and BAT [113]. By contrast, a marked decrease of expression was observed for a number of genes involved in fatty acid transport and oxidation in skeletal muscle, even though some of these

genes are verified PPARγ targets in other tissues. These findings are in concert with the hypothesis that PPARγ stimulates adipogenesis as well as the uptake

and oxidation of fatty acids in adipose tissue and suggest that PPARγ signaling

might result in different responses in regulating lipid metabolism in muscle and

adipose tissue.

Smads and Ski

Smad proteins have been well studied and all known to be signal transmitters of

TGF-β signaling pathway [114]. Smad2 and Smad3 are receptor-regulated Smad proteins (rSmads) that are cytoplasmically anchored by the Smad anchor for receptor activation (SARA) and activated by receptor phosphorylation [115-117].

Upon phosphorylation, rSmads form complex with the common mediator Smad,

Smad4 and enter the nucleus to turn on the transcription of TGF-β target genes or reporter constructs containing TGF-β response elements [115-117]. In the nucleus, Smad2 or Smad3 interact with more than 50 known proteins, including but not limited to transcriptional co-activators or co-repressors [114]. For example,

Ski binds to phosphorylated Smad2 and Smad3 and acts as a transcriptional co-

repressor [25, 40].

Although Smad2 and Smad3 are direct mediators of TGF-β signaling, they play distinct roles in different tissue or cell types in addition to overlapping roles in

TGF-β signaling [114]. Smad2 or Smad3 null mouse embryonic fibroblasts

(MEFs) show differential effects in the activation of Smad responsive reporters and of genes involved in the negative or positive feedback loop of TGF-β

signaling [118]. The activation of a Smad binding element (SBE) driven reporter

is independent of Smad2 but enhanced by Smad3, whereas the activation of an

activin response element (ARE) driven reporter is dependent on Smad2, and

enhanced in Smad3 null MEFs. Moreover, the induction of immediate-early TGF-

β responsive genes, c-Fos, Smad7 and

TGF-β1 auto-induction is Smad3 dependent but TGF-β induced

metalloproteinase-2 is selectively dependent on Smad2 [118]. Microarray

analysis of Smad2 and Smad3 null MEFs revealed that loss of Smad3 inhibited

the activation of immediate-early TGF-β responsive genes through

Smad3/Smad4 consensus DNA binding motifs whereas loss of Smad2 was

associated with hyperresponsiveness of these genes to TGF-β [119].

In HaCaT keratinocytes, silencing Smad3 by RNAi was sufficient to interfere with

TGF-beta-induced cell cycle arrest and to induce or suppress endogenous cell

cycle regulators, whereas knocking down Smad2 showed no effect [120].

Microarray analysis of mRNA from these HaCaT cells in the presence or

absence of TGF-β identified subsets of genes whose transcriptional regulation is

Smad3 dependent, Smad2 dependent or Smad2 and Smad3 dependent [120].

Furthermore, in a number of TGF-β sensitive cell lines, altering the endogenous

Smad3 to Smad2 ratio by knocking down Smad2 with RNAi remarkably enhanced the TGF-β cytostatic responses, but depleting Smad3 effectively inhibited the TGF-β mediated cell cycle arrest. Interestingly, silencing Smad2 facilitated the activation of Smad3 upon TGF-β treatment [121]. Taken together, these data above suggest that Smad2 and Smad3 do not have compensatory roles but distinct roles in mediating TGF-β signaling. Smad3 is an essential mediator of the TGF-β induced immediate-early response and cytostatic inhibition while Smad2 mediates transcription of a different set of TGF-β

responsive genes and probably negatively transmodulates Smad3 mediated

TGF-β signaling.

A study of the mechanism underlying the activation of the selective roles

between Smad2 and Smad3 in response to TGF-β has been carried out in a nontransformed human cell line transduced with retrovirus expressing Myc-His6

tagged Smad2 or Smad3. TGF-β activated Myc-His6 -Smad2 or Smad3

complexes were affinity purified and gel separated. The selected protein bands

were subjected to trypsin digestion followed by liquid chromatography and tandem mass spectrometric analysis. Subsets of interacting partners that uniquely associate with Smad2 or Smad3 and a subset of proteins that are

common binding partners to Smad2 and Smad3 were identified [122]. This

further suggests distinct roles of Smad2 and Smad3 in gene regulation in

response to TGF-β signaling. Consistent with previous studies, Ski was

discovered to associate with the Smad2 complex [122].

Biologically, both Smad2 and Smad3 are known as tumor suppressors. However,

knocking out Smad2 in mice resulted in early embryonic lethality at embryonic

day 7.5-12.5 [123-126], indicating a role of Smad2 in the embryo development.

By contrast Smad3 knockout mice are viable and fertile. Consistent with its tumor

suppressor role, one of the three Smad3 KO mice strains developed

adenocarcinoma [127-129]. Interestingly, Smad2 has also been implicated as an

important tumor suppressor in epithelia cell derived carcinomas. Silencing

Smad2 alone caused remarkable malignant transformation of a rat prostate

epithelial cell line and tumor growth in vivo, whereas knocking down Smad3

alone did not induce tumors [130]. In human skin squamous cell cancers, the frequency of the Smad2 loss is 14 times higher than that of Smad3 loss and 70%

of total tumor samples show the loss of Smad2 [131]. Compared to the WT, mice

with keratinocyte-specific Smad2 ablation exhibit accelerated formation and

malignant progression of skin tumors induced by carcinogen, whereas Smad3

knockout mice are resistant to the development of these carcinogen-induced skin

tumors [131]. Besides their differential roles in tumorigernesis, Smad2 cooperates with myocyte enhancer-binding factor 2 (MEF2), an essential

regulator of myogenesis [132] to enhance MEF2 mediated transcription [133],

indicating a positive effect of Smad2 on muscle differentiation. On the other hand,

Smad3 represses MyoD dependent myogenic transcription [74]. Since

overexpression of Ski causes squamous cell carcinoma and Ski promotes

muscle differentiation, these data indicate that Ski might have selective roles in

its interactions with Smad2 and Smad3.

On the molecular level, through interactions with Smad2, Smad3 and Smad4

proteins, Ski is known to block TGF-β signaling [4, 25, 27, 40, 41]. Previous

reports demonstrate that Ski forms a complex with activated Smad proteins to

bind the SBE and repress Smad mediated transactivation [40, 134]. In Ski-CEFs,

Ski represses Smad mediated transcription in the absence of exogenous TGF-β

[5, 135, 136]. In fact, a large fraction of the Ski protein in transformed CEFs

appears to be present in a complex with Smads [136]. In Ski protein there are

three independent sites responsible for Ski-Smad interactions. One is within the

first 35 aa and it is interrupted by the truncation of v-Ski and the second one is

positioned in the N-terminal region of DHD domain [4, 12, 25, 137, 138]. These

sites mediate Ski binding to Smad2 and 3.The third site is in the SAND domain

and is responsible for Ski binding to Smad 4 [12, 138]. Mutations in these

domains significantly reduce Ski’s potency as a Smad co-repressor and also reduce its transforming activity [5, 139]. Therefore, these observations suggest a

direct linkage between Ski-Smad interactions and transformation which involves

an alteration of TGF-β signaling by Ski.

Our Research Focus

The molecular mechanisms that explain how Ski executes its biological functions

involved in muscle differentiation, bone morphogenesis and oncogenesis have

been studied and partially clarified. However, although its affect on TGF-β

signaling is strongly implicated in the mechanism underlying the Ski induced transformation of avian cells [139], details of this mechanism remain unclear. In this study, I set out to clarify the roles of individual TGF-β activated Smad proteins and to identify genes affected by Ski to induce transformation, signaling pathways utilized in this process and genes essential for mediating Ski induced transformation.

Animal studies have suggested that Ski has novel functions related to metabolic regulation. Excitingly, the in vitro observation that Ski-transformed fibroblasts exhibit dramatically reduced culture medium acidification provided a clue for unraveling its affect on metabolism. Growing fibroblasts normally acidify culture medium due to lactate production via glycolysis. Based on this clue, I have tested the hypothesis that Ski causes fibroblasts to enhance utilization of other energy generation pathways in addition to glycolysis.

Materials and Methods

Materials

1. Plasmids

Plasmid Name Vector Insert

pcDNA3flagSmad2E2 pcDNAflagSmad2 (HpaI -XbaI) PCR product(HPaI-XbaI) see table Primer for cloning

RCASflagSmad2E2 RCASBPxs (BstBMI-SpeI) pcDNAflagSmad2E2 (ClaI-XbaI)

RdpInxEET7Ski-L19A RdpInxEESki (NcoI-XbaI) pcMVT7-Ski-L19A (NcoI-XbaI)

RdpInxEET7Ski-L110P RdpInxEESki (NcoI-XbaI) pcMVT7-Ski-L110A (NcoI-XbaI)

RdpInxEET7Ski-W255E RdpInxEESki (NcoI-XbaI) pcMVT7-Ski-W255E (NcoI-XbaI)

RdpInxEET7Ski-L19A,L110P RdpInxEESki (NcoI-XbaI) pcMVT7-Ski-L19A, L110P (NcoI-XbaI)

RdpInxEET7Ski-triple RdpInxEESki (NcoI-XbaI) pcMVT7-Ski-triple (NcoI-XbaI)

RdpInxEET7Ski-L19A,W255E RdpInxEESki (NcoI-XbaI) pcMVT7-Ski-L19A,W255E(NcoI-XbaI)

RdpInxEET7Ski-L110P,W255E RdpInxEESki (NcoI-XbaI) pcMVT7-Ski-L110P, W255E (NcoI-XbaI)

RdpuromiR30 RdpInx( KpnI-PmeI) PCR product( KpnI-PmeI)

see table for Primer

RdpuromiR30shSmad2-1 RdpuromiR30(XhoI -MfeI) PCR product shRNA Smad2-1(XhoI-EcoR I)

see table for Primer

RdpuromiR30shSmad2-2 RdpuromiR30(XhoI-MfeI) PCR product shRNA Smad2-2(XhoI-EcoR I)

see table for Primer

RdpuromiR30shSmad2-3 RdpuromiR30(XhoI-MfeI) PCR product shRNA Smad2-3(XhoI-EcoR I)

see table for Primer

RdpuromiR30shSmad2-4 RdpuromiR30(XhoI-MfeI) PCR product shRNA Smad2-4(XhoI-EcoR I)

see table for Primer

RdpuromiR30shSmad3-1 RdpuromiR30(XhoI-MfeI) PCR product shRNA Smad3-1(XhoI-EcoR I)

see table for Primer

RdpuromiR30shSmad3-2 RdpuromiR30(XhoI -MfeI) PCR product shRNA Smad3-2(XhoI-EcoR I)

see table for Primer

RdpuromiR30shPPARg-1 RdpuromiR30(XhoI-MfeI) PCR product shRNA PPARg-1(XhoI-EcoR I)

see table for Primer

RdpuromiR30shPPARg-2 RdpuromiR30(XhoI -MfeI) PCR product shRNA PPARg-2(XhoI-EcoR I)

see table for Primer

RdpuromiR30shPPARg-3 RdpuromiR30(XhoI-MfeI) PCR product shRNA PPARg-3(XhoI-EcoR I)

see table for Primer

RdpuromiR30shPPARg-4 RdpuromiR30(XhoI-MfeI) PCR product shRNA PPARg-4(XhoI-EcoR I)

RCASbpxs empty vector

RdpInx empty vector

RdpIEEski Ski expressing vector

RCASps27 encodes full length c-Ski pCMVT7Ski From Dr. Michael Hayman’s lab pGL3tkrenilla From Dr. Clemencia Colmenares's Lab pCMVPPRE luciferase From Dr. Hung-ying Kao lab pCMV RAR luciferase From Dr. Hung-ying Kao lab pCMVflagPPARg From Dr. Hung-ying Kao lab

TMP From Dr. Scott Lowe's lab

2. Primers and synthetic oligos

Cloning Primers

Name Sequence

5' Smad2E2 mutagenesis 5'-CTC CAG TTA ACC ATA GCC TGG ATT TGC- 3'

3' Smad2E2 mutagenesis 5'-GGCCCTCTAGACTTGAGTTATTCCATCTCGGAGCAGCGTACTG-3'

5' miR30 casette cloning primer 5'-CAGCGGTACCTCTCGACTAGGGATAACAGGGTA-3'

3' miR30 casette cloning primer 5'-CCGCCGTTTAAACGAATTGAAAAAAGTGATTTAA-3'

5' miR30 shRNA amplifying primer 5'-CAGAGGCTCGAGAAGGTATATTGCTGTTGACAGTGAGCG-3'

3' miR30 shRNA amplifying primer 5'-CGCGTCCAATTGTCCGAGGCAGTAGGCA-3'

Synthetic shRNA oligos

Name Starting base Sequence

Smad2-1 ShRNA 963 5'-TGCTGTTGACAGTGAGCGCGCCCGCATTCTGGTGTTCTATTAGTGAAG

CCACAGATGTAATAGAACACCAGAATGCGGGCTTGCCTACTGCCTCGGA-3'

Smad2 -2 ShRNA 566 5'-TGCTGTTGACAGTGAGCGAACAGCCACCATGAACTTAAAGTAGTGAAG

CCACAAGATGTACTTTAAGTTCATGGTGGCTGTGTGCCTACTGCCTCGG-3'

Smad2 -3 ShRNA 384 5'-TGCTGTTGACAGTGAGCGCACCAAGCACTTGCTCTGAAATTAGTGAAG

CCACAAGATGTAATTTCAGAGCAAGTGCTTGGTATGCCTACTGCCTCGG-3'

Smad2 -4 ShRNA 1087 5'-TGCTGTTGACAGTGAGCGCTGTCTAGGTCTGCTTTCCAACTAGTGAAG

CCACAGATGTAGTTGGAAAGCAGACCTAGACAATGCCTACTGCCTCGGA-3'

Smad3-1 ShRNA 809 5'-TGCTGTTGACAGTGAGCGATGCCTTGGTCTGCTCTCTAATTAGTGAAG

CCACAGAATGTAATTAGAGAGCAGACCAAGGCAGTGCCTACTGCCTCGG-3'

Smad3-2 ShRNA 359 5'-TGCTGTTGACAGTGAGCGATGTGTGAATCCCTATCATTACTAGTGAAG

CCACAGAATGTAGTAATGATAGGGATTCACACAGTGCCTACTGCCTCGG-3'

PPARg-1 ShRNA 986 5'-TGCTGTTGACAGTGAGCGACTGGCTTCTCTCATGAATAAATAGTGAAG

CCACAGATGTATTTATTCATGAGAGAAGCCAGGTGCCTACTGCCTCGGA-3'

PPARg -2 ShRNA 351 5'-TGCTGTTGACAGTGAGCGACCTCTGGATTTCATTATGGAGTAGTGAAG

CCACAGATGTACTCCATAATGAAATCCAGAGGCTGCCTACTGCCTCGGA-3'

PPARg -3 ShRNA 524 5'-TGCTGTTGACAGTGAGCGATCACATAATGCCATCAGGTTTTAGTGAAG

CCACAGATGTAAAACCTGATGGCATTATGTGACTGCCTACTGCCTCGGA-3'

PPARg -4 ShRNA 748 5'-TGCTGTTGACAGTGAGCGCTGACATGAACTCTTTAAGGATTAGTGAAG

CCACAGATGTAATCCTTAAAGAGTTCATGTCATTGCCTACTGCCTCGGA-3'

Q-RTPCR Primers

Gene Species Primers AACS Gallus F: 5'- ATCTCGGAATGGCTGTGGAAG

R: 5'- CCAGAAGTGCGTAGGCTGACA

ACOX -1 Gallus F: 5'- GACCTTGTGCGAGCATCTGAG

R: 5'- AGTCAAGACAGCGCAGACAGC

ANT2 Gallus F: 5'- TGACCTCCTATCCCTTCGATACC

R: 5'- CTTCCTCCAGCAGTCAATAGTG

ANT3 Gallus F: 5'- GGCTGGTGTGGTTTCCTATCC

R: 5'- AACGCCTTTCCTCCCTCATC

CD36 Gallus F: 5'- AGAACTGCACACTGGCTGGAG

R: 5'- CCTCCTCATTTGGGCTCAGAC

FABP -3 Gallus F: 5'- CGGTGAAGACCCATAGCACCT

R: 5'- ACCAGCTTGCCTCCATCTAGC

FABP -4 Gallus F: 5'- TGAGACCACAGCAGATGACAGA

R: 5'- GTTCCCATCCACCACTTTTCTC

FATP -1 Gallus F: 5'- GAGCTGATCCGTGATTCCAGA

R: 5'- CATTGACGTAGCCATCGAAGC

GyK Gallus F: 5'- ACCACGGTCTTCTAACCACGA

R: 5'- TAGATTGTCCCTCAGCCAACG

LPL Gallus F: 5'- TCAGCAGCACTACCCAGTGTCT

R: 5'- GTACCCCAGCAAGTGGACATT

PGC1 -α Gallus F: 5'- CGCAGACCCAGATATGACAGC

R: 5'- CTCTCCCTTTGTTTGGCCCT

PPARγ Gallus F: 5'- AGCTCCAGGATTGCCAAAGTG

R: 5'- CTTGTCCCCACACACACGACA

Smad2 Gallus F: 5'- GAATGCAACAGTGGAAATGACAAG

R: 5'- CTTTGATTACAGTTGGGGCTCTG

Smad3 Gallus F: 5'- CCAGGAATTCGCAGCTCTCT

R: 5'- ACCCTTGACGAAGCTCATCC

UCP -3 Gallus F: 5'- ACCTCATCGACACACTGCTGC

R: 5'- TACCGCGTCTTCACCACATC

Wnt 5A Gallus F: 5'- CAACAATGAGGCTGGGAGGA

R: 5'- CTTCAGGGCATCGCCTACC

ATPB3 Gallus F: 5'- ATTAAAGCCTGAAGGGGAACC

R: 5'- TACATCAAGTCAATCAAGCCTCCT

EEF1A1 Gallus F: 5'- TTCCGGCAAGAAGCTGGAG

R: 5'- TCATGTCACGCACAGCGAAA

16srRNA -1 Gallus F: 5'- CCCCCACACTAACAAGCAATAC

R: 5'- GTCTTTGGTTTGCCGAGTTC

16srRNA -2 Gallus F: 5'- AACCCACAGACCACAACTCTTC

R: 5'- GTCGTAAACCTCCTTGTCGATAT

Smad2 intron5 Gallus F: 5'- ATCACCTCCTCTCTTCCCTTGT

R: 5'- CTCTCGTCTCCTCAAGCTATGC

3. Cells

CEF (chicken embryo fibroblast)

Ski-CEF (Ski transformed CEF)

Δz3/4 CEF (CEFs expressing Δz3/4 mutant)

PPARγKD (Ski-CEFs with PPARγ knockdown)

4. Reagents

4.1 Common Reagents for RNA-related Experiments

DEPC-treated H2O

Add 0.5 ml diethylpyrocarbonate (DEPC) to 1000ml of Milli-Q H2O, mix well and let it sit at room temperature overnight. Next day, autoclave for 30min then store at room temperature.

DEPC-treated PBS

Add 1ml DEPC to 1000ml of 1×PBS, mix well and let it set at room temperature overnight. Next day, autoclave for 30min and store at room temperature.

Diethylpyrocarbonate (DEPC, Sigma)

RNase inhibitor: RNase ZAPTM (Ambion)

Trizol (Invitrogen)

RNase free DNase I (QIAGEN)

RNA Extraction Kit (RNeasy, QIAGEN)

4.2 Common DNA Manipulation Related reagents

Genomic DNA isolation Buffer

Tris-Cl (pH 7.5) 100mM

EDTA (pH 8.0) 5mM

SDS 0.5%

NaCl 200mM

Proteinase K 0.2μg/ml

DNA loading buffer (10X)

Glycerol 50%

EDTA (pH8.0) 100mM

SDS 1%

Bromophenol Blue 0.4%

Xylene cyanol 0.4%

TAE buffer (10X)

Tris 400mM

Acetate 400mM

EDTA (pH8.0) 10mM

TE buffer

Tris-Cl (pH8.0) 10mM

EDTA 1mM

SOB medium

BACTO-Tryptone 20g

BACTO-Yeast Extract 5.5g

1M NaCl 10ml

1M KCl 10ml

Add H2O to 1L, autoclave

Add 10ml of 2M Mg2+ (1M MgCl2 and 1M

MgSO4) and 10 ml 2M Glucose before use

LB medium

BACTO-Tryptone 10g

BACTO-Yeast Extract 5g

NaCl 10g

Add H2O to 1L, autoclave

LB-Ampr plate

BACTO-Tryptone 10g

BACTO-Yeast Extract 5g

NaCl 10g

Add H2O to 1L, adjust pH to 7.0 with 2N NaOH and add 20g agar

After autoclave, cool to 55°C and add Ampicillin to 100μg/ml

4.3 Whole cell lysate Extraction and Westernblot Related

RIPA lysis buffer

Tris-Cl (pH 8.0) 50mM

NaCl 100mM

EDTA (pH8.0) 1mM

Glycerol 10%

NP-40 0.2%

Sodium orthovanadate 0.1mM

Sodium pyrophosphate 5mM

NaF 1mM

Complete protease inhibitor cocktails (Roche)

Protein loading buffer (3X)

1M Tris-Cl (pH6.8) 1ml

Glycerol 2ml

10% SDS 4ml

1% Bromophenol blue 2ml

Add H2O to 10ml and add 2% β-Me before use

10% APS

0.1g ammonium persulfate (APS)

Add H2O to 1.0 ml; make fresh, store at 4°C

SDS-PAGE separating gel (6%, 20ml)

H2O 8.2ml

1M Tris-Cl (pH8.9) 7.5ml

10% SDS 0.2ml

29:1 A/B 4ml

10% APS 0.2ml

TEMED 20μl

SDS-PAGE stacking gel (3.5%, 11ml)

H2O 7.4ml

1M Tris-Cl (pH6.8) 1.3ml

10% SDS 0.1ml

29:1 A/B 1.3ml

10% APS 0.1ml

TEMED 10μl

TG-SDS (Gel running buffer, 10X)

Tris 120g

Glycine 576g

SDS 40g

Add H2O to 4L

Transfer buffer (10X)

Tris 60.6g (25mM)

Glycine 242.4g (192mM)

Add H2O to 4L

Before use, dilute 10X Transfer buffer to 1X and

add methanol to 12.5% final concentration

TBS (10X)

Tris 24.2g

NaCl 80g

Add H2O to 1L and adjust pH to 7.4

TBST

TBS 1X

Tween 20 0.1%

PBS (10X)

Na2HPO4 (anhydrous) 10.9g

NaH2PO4 (anhydrous) 3.2g

NaCl 90g

Add H2O to 1L and adjust pH to 7.4

PBST

PBS 1X

Tween 20 0.05%

Milk Blocking Buffer (general use)

PBST 1×

Tween 20 0.05%

Nonfat dry milk 5%

Casein Blocking Buffer (for anti-Ski G8 antibody)

PBS 1X

Casein (technical grade) 0.4%

PVP (MW 40,000) 1%

EDTA (pH8.0) 10mM

Tween 0.2%

Dissolve at 65°C for 2hr and cool to about 37°C

Add Sodium azide to 0.0.2% and adjust pH to 7.4

Assay Buffer

Diethanolamine (DEA) 100mM

MgCl2 1mM

Sodium azide 0.02%

Adjust pH to 10.0

Stripping buffer

Tris-Cl (pH 6.8) 62.5mM

SDS 2%

β-Me 100mM

4.4 Immunoprecipitation Related Reagents

NETN Buffer

Tris-Cl (pH8.0) 20mM

NaCl 100mM

EDTA (pH8.0) 1mM

NP-40 0.1%

Glycerol 10%

DTT 1mM

Complete protease inhibitors cocktails (Roche)

Protein A/G beads ( Roche) 50% slurry

Blocking buffer

Salmon testes DNA 1mg/ml

BSA 10mg/ml

Sodium azide 0.05%

TE 1X

4.5 Immunostaining Related Buffers

PBS (10×)

Na2HPO4 (anhydrous) 10.9g

NaH2PO4 (anhydrous) 3.2g

NaCl 90g

Add H2O to 1L and adjust pH to 7.2

Fixation buffer

PFA 3.7%

PBS 1X

Permeabilization buffer

PBS 1X

Goat serum 10%

Triton-X 100 1%

Blocking buffer

PBS 1X

Goat serum 10%

Tween 20 0.1%

4.6 Tissue Culture and Transfection Reagents

Medium

Dulbecco's Modified Eagle Medium (DMEM, Invitrogen)

FBS 10%, (Atlanta Biological)

Gentamycin (50 µg/ml)

Fungizone, (0.002%)

Others

Penicillin/Streptomycin (50 µg/ml)

PBS, PH=7.4

Chiken serum

Glutamine (100 µg/ml)

Puromycin (2ug/ml)

G418 (750 µg/ml)

Fugene 6 (Roche)

Fugene HD (Roche)

Methods

1. DNA Cloning

1.1 PCR amplification

To generate DNA fragments for subcloning, Pfx polymerase (Invitrogen) was

used

Reaction set up as follows:

Pfx PCR Buffer 1×

dNTPs 200μM each

MgSO4 1mM

Primers 100nM each

Plasmid DNA 5ng

Pfx DNA polymerase 1.25U

PCR enhancer 0.5×

Add sterile Milli-Q H2O to 50μl

PCR program was:

Step 1 (initial melting) 94°C =2min

Step 2 (amplification) 94°C =15sec

60°C =30sec or run gradient for optimal annealing temp

68°C=2min

Repeated for 30 cycles

Step 3 (final extension) 68°C =7min

1.2 Restriction enzyme digestion

Plasmid DNA/PCR product 1~5μg

NEB buffer 1×

BSA (if needed) 1×

Restriction enzyme (NEB) 5-25 Unit

Add sterile Milli-Q H2O to 30μl

The reaction was incubated at the appropriate temperature for 2 hours

1.3 Dephosphorylation

To avoid self-ligation of enzyme-digested vector with two compatible ends or

single cut vector ligating back due to the incomplete digestion of vectors,

Antarctic phosphatase (NEB) was used to treat the cut open vectors.

Enzyme-digested DNA (3~5kb) 1μg

Antarctic buffer 1×

Antarctic phosphatase 5U

The reaction was incubated at 37°C for 30min (sticky end) or 60 min (blunt end).

Afterward, the enzyme was heat-inactivated at 65°C for 20 min and the

dephosphorylated product was used directly for ligation.

1.4 Agarose gel separation and purification of DNA

DNA fragments were mixed with DNA loading buffer and separated on 0.8%

agarose gels (for fragments larger than 500bps) or 2% agarose gel (for

fragments smaller than 500bps). DNA fragments were visualized by ethidium

bromide-staining. DNA in the gel slices were recovered using a gel purification kit

(QIAGEN) following the manufacturer’s instructions. The concentration of

recovered DNA fragments was estimated by running an aliquot on a gel in

comparison with a Lamda HindIII cut Marker (Roche).

1.5 Ligation

In ligation reactions, 200~500ng vector DNA was used and the molar ratio of

insert to vector was around 3:1. The ligation reaction was carried out with T4

DNA ligase

(Roche) as follows:

Insert and Vector 8μl

T4 DNA ligation buffer 1μl

T4 DNA ligase 1μl

The reaction was incubated overnight at 4°C

1.6 Bacterial Transformation

50μl competent cells were transformed with 5μl of the ligation products.

Following steps applied.

30 min incubation on ice

2 min heat shock at 42°C

2 min incubation on ice

150 µl SOC medium added

Shaker at 220 rpm, 37°C for 45 min

Plate 50 µl of culture solution on LB-Ampr plates.

Keep plates in 37°C incubator for 12-16 hours.

The next day, DNA was extracted using a DNA Miniprep kit (Eppendorf) and

analyzed by restriction digestion. Correct clones were further verified by DNA

sequencing.

2. Tissue Culture and Transfection

Proliferating chicken embryo fibroblasts (CEFs) were maintained in medium

consisting of Dulbecco's modified Eagle's medium (DMEM) (Invitrogen)

supplemented with 10% fetal bovine serum (FBS, Atalantica Bioslogical),

50μg/ml penicillin, 50 units/ml streptomycin and 0.002% Fungizone (Invitrogen).

Cells were maintained at the confluence around 90% before split into the next passage. Transfection was done at the cell density around 50%~70% about

12~18 hours after initial plating.

3. Retroviral infection

Retrovirus virus constructs were initially introduced into cells by transfection of a

combination of two retroviral vectors. One retroviral vector is replication defective

but bearing a selection marker (puromycin resistant enzyme). The other vector is

replication potent but with no selection resistant capacity. Seventy-two hours

after transfection, puromycin (4 µg/ml, Sigma) was added to select the

transfected and the infected cell population.

4. Western blotting

Cells were harvested from 90~100 confluent 100 mm or 60mm culture dishes.

Operation steps as follows:

Cells were washed twice with PBS, scrapped into 500~200 μl of RIPA lysis buffer.

Incubate lysates for 10 min on ice. The suspension was subjected to three

freeze-thaw cycles. Cell debris was removed by centrifugation at 12,500 rpm for

10min at 4°C. Protein concentrations were determined by Bradford assay (Bio-

Rad). Equal amounts of proteins (20μg ~ 40μg) were boiled in protein loading

buffer. Samples were separated by 6% SDS–PAGE at 150V for 2.5 hours.

Samples were transferred to the PVDF membrane (0.45um, Millipore) at 100V for

3 hours. The blots were washed for 5min in PBST and blocked in 5% milk-PBST

for 2 hours R.T. After brief wash with PBST, primary antibodies were added in 5%

milk and rocked O.N. The next day, membranes were washed three times with

PBST for 5 min. Secondary antibodies was added at the dilution ratio of 1:20,000 and rocked for 2 hrs R.T. 5 min wash with PBST, trice. Alkaline phosphatase

(ALP) substrate, CDP-Star® (Tropix) was added at 1:50 dilution. After 15 min incubation, membrane was exposed to the HyBlot CL autoradiography film

(Denville Scientific).

When using anti-Ski G8 monoclonal antibody, the blots were pre-blocked overnight at 4°C in casein blocking buffer and then incubated with primary antibody in the same solution for 30min at room temperature. The membranes were washed twice with blocking buffer and further incubated with ALP- conjugated secondary antibodies for 30min at room temperature. Subsequently, the blots were washed with blocking buffer then with assay buffer and the signal was detected with CDP-Star®, an alkaline phosphatase chemiluminescent substrate as above.

5. Quantitative Realtime-PCR ( Q-RTPCR)

RNA was extracted by using the RNeasy kit (QIAGEN) according to manufacturer’s instructions. Cells from a 100mm dish (~90% confluence) were washed with DEPC-treated PBS and were quickly scrapped with 700μl RLT buffer. Lysates were homogenized by shredder column and quickly mixed with equal volume 70% ethanol. RNA was collected by RNeasy spin column where

DNAse treatment was done. RNA was eluted from the column by 30 µl RNase- free water. cDNA was generated using reverse transcriptase SuperScript™ III

(Invitrogen) with random hexamer primers according to the manufacturer’s instructions. Operation steps as following:

RNA was incubated with hexamer and dNTPs as

Total RNA 5μg

Hexamer 200ng

dNTP 1mM

Incubate at 65 °C for 5min and placed on ice for another 2min

This mixture was then incubate with

Reverse Transcription buffer 1X

MgCl2 2mM

DTT 10mM

RNase Inhibitor 40U

SuperScriptTM III RT 200U

The reverse transcription cycle was as follows:

25 °C =10min

50 °C =50min

85 °C =5min

Finally, the reaction was incubated with 1μl RNase H (2U/μl) at 37°C for 20min

Quantitative real-time-PCR was done with iCycler iQTM (Bio-Rad). Bio-Rad Sybr

green master mix was used to provide reaction system.

cDNA template 10 µl

Mastermix 2X 12.5 µl

Primer pair 10uM 0.5 µl

H2O 2 µl

The typical PCR program was listed below

Step 1 (initial melting) 94°C =15min

Step 2 (amplification) 94°C =15sec

60°C =30sec

72 °C=30sec

Repeated for 40 cycles

Step 3 (final extension) 72°C =10min

Step 4 (melting curve) starting from 72°C, the temperature was increased by

0.5°C each cycle until 94°C (totally 44 cycles)

6. Reporter assays

Cells (2×104 per well) were seeded in 96-well plates and 12 hours after plating, cells were transfected at 70% confluence using Fugene 6 (Invitrogen). Fugene 6 amount varies according to the DNA amount used. Normally, a 3:1 fugene6 to

DNA ratio (µl:µg) is obeyed. 60 hours after transfection, plates were harvested for luciferase assay. 30μl of 1× Reporter Lysis Buffer (Promega) with one round of freeze-thaw followed by incubation at room temperature for 20 min with shaking.

Cell debris was removed by centrifugation and the luciferase activity in the supernatant was determined by the dual luciferase reporter assay system

(Promega), according to the manufacturer’s instructions. A Renilla luciferase

vector pGL3-TK-Renillar was used as an internal control. Briefly, in a 96-well white plate, 70μl cell lysate was incubated with equal volume of Dual-GloTM

Luciferase Reagent for 10min and the Firefly luciferase activity was measured.

Afterwards, this mixture was incubated with 1 volume of the diluted Stop&Glo substrate (1/100 in Dual-GloTM Stop&Glo Reagent) for another 10min and the

Renilla luciferase activity was measured. Luciferase units were determined by using a MAXline microplate luminometer (Molecular Devices). The relative light units (RLUs) were generated by normalizing firefly luciferase units to Renilla luciferase units of the co-transfected pTK-Renilla-luc vector. Relative Response

Ratio was calculated by comparing RLUs of different samples. The experiments were done in 6 replicates and the reported results represent at least three independent experiments.

7. Genomic DNA preparation

Cells on 100mm plates were washed twice with PBS, scrapped into three

Eppendorf tubes and pelleted by centrifuge at 5,000rpm for 4min. Each tube of

cells was incubated with 1ml of lysis buffer at 55°C with occasional agitation for

3hr and then kept at 55°C without agitation overnight. After being mixed with an

equal volume of Phenol/CHCl3/isoamyl alcohol and centrifuging at 12,500rpm for

10min, the aqueous phase was mixed with 2.5 volume of cold 100% ethanol and

DNA was pelleted at 12500rpm for 10min. After washing with 70% ethanol, the

DNA pellet was air dry and dissolved in 250μl TE at 37°C overnight.

8. Immunoprecipitation

Cells (~ 50% confluent, 100mm dishes) were scrapped from the plates,

resuspended in 1ml of NETN buffer and sonicated 0.5sec on/0.5 sec off for a total of 1 min. Cell debris were removed by centrifugation for 15 min at 10,

000rpm at 4°C and lysate was preabsorbed with 30μl 50% slurry protein A beads for 1hr at 4 °C followed by centrifugation. After the protein concentration was determined with a Bradford protein assay (Bio-Rad), a fraction of the cell lysates were subjected to immunoblotting to detect the expression of the proteins of interest and, if needed, the amounts of lysates used for immunoprecipitation were adjusted accordingly. Immunoprecipitation was carried out as following:

~200μg precleared lysate was incubated in parallel with either 4μg rabbit polyclonal anti-Ski (H329, Santa Cruz) or normal rabbit IgG for 3 h at 4°C. The immunocomplexes were then incubated with 40μl 33% slurry protein A/G beads

(Santa Cruz) overnight at 4°C. Precipitates were washed 5 times in ice-cold

NETN buffer, resuspended and released from the beads by boiling in protein loading buffer for 10 min. The precipitated proteins were separated by 6% SDS–

PAGE and analyzed by Western blotting. If precipitating and primary Western blotting antibodies were from the same species, either HRP-conjugated Mouse

IgG TrueBlotTM or Rabbit IgG TrueBlot™ was used as the secondary antibody

accordingly.

Results and Discussion

Chapter I PPARγ Mediates Ski Induced Energy Metabolism Shift

Abstract Overexpression of Ski oncoprotein induces oncogenic transformation in chicken

embryo fibroblasts (CEFs). However, unlike most other transformed cells, Ski-

transformed CEFs (Ski-CEFs) do not display the classical Warburg effect. On the

contrary, Ski represses lactate production and glucose utilization in CEFs. We

found that compared to CEFs, Ski-CEFs exhibit stimulated fatty acid catabolism through β-oxidation, enhanced oxygen consumption and increased number and mass of mitochondria. Interestingly, the master adipogenesis and lipid metabolism regulator, PPARγ is dramatically elevated at both the mRNA and protein levels in Ski-CEFs. Moreover, PPARγ targets involved in lipid uptake, transport and oxidation are drastically upregulated by Ski. Knocking down

PPARγ by RNA interference reverses the elevated expression of these PPARγ targets, as well as the metabolic shift and the increased mitochondrial biogenesis in Ski-CEFs. Moreover, we found that Ski co-immunoprecipitates with PPARγ and co-activates a PPARγ-responsive reporter. Therefore, we conclude that

PPARγ mediates Ski induced metabolic shift and mitochondrial biogenesis in

CEFs.

Introduction

The nuclear oncoprotein Ski is a bifunctional transcription factor. Ski does not

bind DNA directly [21, 22], but interacts with several DNA-binding transcription

factors to regulate gene expression. The interaction of Ski with the TGF-β

activated Smad proteins results in transcriptional repression and is critical for its

oncogenic transformation of chicken embryo fibroblast (CEF) [1, 3, 4, 25, 27, 40,

41, 140]. In human pancreatic and gastric cancer cells, knocking down Ski by

RNA interference restores TGF-β signaling in vitro and decreases tumor growth

in vivo [35, 43]. In human melanomas, overexpression of Ski is correlated with

melanoma progression [47] and knocking down Ski inhibits melanoma growth in

vivo [9, 48]. In this case, Ski not only suppresses TGF-β signaling but also binds

to FHL2 to increase the transcriptional activation by β-catenin resulting in the activation of Wnt signaling. Ski transforms hematopoietic cells by binding RAR and recruiting N-CoR to repress transcription mediated by the RAR responsive element [23].

In apparent contrast to its oncogenic activity, Ski induces skeletal muscle differentiation in vitro and in vivo [64, 73, 141] It regulates skeletal muscle terminal differentiation though binding to Six1 and Eya3 to form a transcriptional activating complex on the Myogenin promoter region [29]. Ski knockout mice display a dramatic reduction of skeletal muscle development [57, 62], as well as the virtual absence of brown adipose tissue (BAT) (Zhang, PhD thesis).

Transgenic mice overexpressing c-Ski or v-Ski exhibit a remarkable increase of skeletal muscle mass but a decrease of adipose tissue [63, 142]. In these mice,

type IIb fast skeletal muscle fibers are selectively hypertrophied while the type I

slow muscle fibers are not affected. Interestingly, recent studies on Ski

transgenic mice revealed that lipid oxidation is increased in the affected muscle

despite decreased expression of a master regulator of lipogenesis, peroxisome

proliferator-activated receptor gamma (PPARγ) [66]. These findings suggested

the possibility that Ski might play a role in regulating lipid metabolism perhaps

mediated by an effect on the expression or transcriptional activity of PPARγ.

The peroxisome proliferator-activated receptors (PPARα, PPARγ and PPARδ)

are a group of nuclear receptors activated by lipid ligands. PPARα is a potent

fatty acid oxidation regulator in liver. PPARδ is a universally expressed regulator

of fatty acid catabolism [81]. PPARγ regulates adipogenesis and has a dual role

in lipid metabolism [90, 91]. Several genes involved in lipid uptake and

intracellular transport, (CD36, FATP-1, FABP4, FABP3, and LPL) are direct targets of agonist induced PPARγ transcriptional activation [106, 143-150].

PPARγ also regulates genes involved in fatty acid synthesis (FASN) and fatty

acid oxidation. ACOX-1, a potent β-oxidation regulatory enzyme and adiponectin

receptor that promotes fatty acid oxidation are common targets of both PPARγ

and PPARα [104-108]. PGC1-α is the coactivator of PPARγ and is also induced

by PPARγ [109]. PGC1-α stimulates fatty acid oxidation, mitochondrial

biogenesis and cell respiration in muscles and BAT [111, 112]. These data

indicate that PPARγ promotes not only lipid transportation, storage and lipogenesis but also lipid oxidation.

Normally fibroblasts generate energy through glycolysis and the activity of this pathway is increased by the action of oncogenes such as Src [151-153]. In

general, cancer cells exhibit the classical Warburg effect, of dramatically

accelerated glucose metabolism to lactate through glycolysis under aerobic

conditions [154]. Primary chicken embryo fibroblasts overexpressing Ski are transformed (Ski-CEF) and exhibit many features common to oncogenic

transformation including: altered morphology, focus formation in monolayer

culture and anchorage-independent growth in soft agar [5]. Surprisingly however, an unusual phenotype of Ski-CEFs is that they acidify culture media very slowly even after reaching confluence [52]. Medium acidification is due to the accumulation of lactate generated by glycolytic conversion of glucose. It therefore seemed likely that Ski-CEFs rely less on glycolysis and produce much less lactate than CEFs. This observation taken together with those obtained from analyses of Ski transgenic and Ski-/- mice, led us to investigate the possibility that Ski might regulate metabolic energy production, perhaps by increasing lipid metabolism in concert with PPARγ. The results of that investigation, reported here, demonstrate that Ski has a profound effect on cellular energetics, decreasing the reliance on glycolysis and stimulating biogenesis of mitochondria and lipid oxidation.

Materials and Methods

Establishment of PPARγKD-Ski-CEF Strains and Cell Culture

A microRNA (miR30) expression cassette was subcloned from the SIN-

TREmiR30-PIG (TMP) vector (kindly provided by Dr. Scott Lowe) [155-157] into a replication defective chicken retroviral vector, RdpI (modified from the

previously described RCAS vector [158]) to generate a chicken shRNA

expressing vector, RdpImiR30, using KpnI and PmeI. The original shRNA

insertion site (EcoRI) in the miR30 cassette was substituted with MfeI by

standard PCR mutagenesis. Based on the mRNA sequence of Gallus PPARγ

(GenBank NM_001001460), four 21mer-shRNAs (nucleotides 351-371, 524-544,

748-768 and 986-1086) targeting chicken PPARγ, were chosen by using RNAi

Codex (http://cancan.cshl.edu/cgi-bin/Codex/Codex.cgi). DNA forms of these

shRNA were generated by PCR amplification of 97-base synthetic

oligonucleotides using Pfu polymerase (Invitrogen) with a common set of primers

(F, 5’CAGAGGCTCGAGAAGGTATATTGCTGTTGACAGTGAGCG3’ and R,

5’CGCGGCGAATTCCGAGGCAGTAGGCA3’). PCR products were subsequently

digested by XhoI and EcoRI and inserted into RdpImiR30 between XhoI and MfeI

sites.

RdpImiR30shRNA was cotransfected with RCAS-Ski using Fugene6 (Roche

Applied Science) into CEFs at 60% confluence. Seventy-two hours after

transfection, cells were selected with puromycin (4 μg/ml, Sigma). Survivors were

cloned in 50% conditioned medium and colonies were screened by western 59

blotting to identify those with the most efficient PPARγ knock down (PPARγKD).

PPARγKD-Ski-CEFs, CEFs and Ski-CEFs were maintained in growth medium

consisting of Dulbecco’s modified Eagle’s medium (DMEM, Invitrogen) with 10%

fetal bovine serum (Atlanta Biological), glucose (4.5 g/L) and gentamycin (50

U/ml, Invitrogen).

Lactate and Glucose Assay

Medium was sampled every 4 hours or 24 hours from cultured CEFs, Ski-CEFS

or PPARγKD-Ski-CEFS. Harvested medium was diluted 50 times for lactate

measurement and 100 times for glucose measurement. The lactate assay was

done according to the protocol of a kit obtained from Dr. Te Lee at SUNY Buffalo.

The glucose assay was done according to the instructions of the Glucose Assay

kit (Sigma).

Mitochondria related assays

Oxygen consumption was measured with Oxgraphy-2K instruments from

OROBROS (www.oroboros.at). For intact cell respiration, cells were trypsinized and counted and resuspended in DMEM at a concentration of 1x106 cells/ml. For

permeabilized cell respiration, the same concentration of cells was resuspended

in Miro5 respiration medium (sucrose 110 mM, potassium lactobionate 60 mM,

EGTA 0.5 mM, MgCl2. 6H2O 3 mM, taurine 20 mM, KH2PO4 10 mM, HEPES 20

mM, and BSA 2 mg/ml, PH 7.1). The optimal concentration of digitonin (1.2 µg/ml)

to permeabilize cells was determined in a pilot experiment. The contribution of

individual electron transport chain (ETC) complex to respiration was examined using specific metabolites and inhibitors according to protocols previously described [159, 160].

Heat generation was measured with isothermal calorimetry (Microcal). CEFs and

Ski-CEFs (1X106) were suspended in growth medium and growth medium alone

was measured as background. The reference power value (30 µCal/s) and cell

number were optimized by pilot experiments.

Citrate synthase activity was measured as previously described [161]

Flow cytometry

Cells were trypsinized and resuspended in culture medium at a concentration of

2 x106/ml. Cells were stained with mitotracker Green (MTG, Invitrogen) at a

concentration of 80nM or Mitotracker Red (MTR, invitrogen) at concentration of

100nM for 30 min at 37 degree. Cells were then briefly washed with pre-warmed

medium and immediately analyzed on a BD SLR II flow cytometer (BD

Biosciences). Gain and amplifier settings were held constant during the

experiment.

Electron Microscopy

CEFs, Ski-CEFs and PPARγKD-Ski-CEFs (2x106 cells /ml) were suspended in

growth medium. Each cell suspension was mixed and fixed with an equal volume

of phosphate-buffered glutaraldehyde (2.5%). Mixtures were immediately spun

down and pellets were continued to be fixed in the same fixation solution for 2

hours at room temperature. Pellets then were rinsed with distilled water and

followed by a post-fixation in 1:1 unbuffered mixture (osmium tetroxide, 2% and

potassium ferrocyanide, 3%). Pellets then were rinsed with distilled water and

soaked in the acidified uranyl acetate (0.25%) solution at 4 c overnight. After another rinsing in distilled water, pellets were dehydrated in ascending concentrations of ethanol, passed through propylene oxide and embedded in the

Poly/Bed 812 embedding medium (Polysciences). Thin sections (70nm) were cut off from a RMC MT6000-XL ultramicrotome. Sections were mounted on a Gilder nickel grid (300 mesh) and then were sequentially stained with the acidified methanolic uranyl acetate and lead citrate. The sections were coated with carbon by a Denton DV-401 coater (Denton vacuum) and then examined by JEOL

1200EX electron microscope.

Measurement of triglyceride synthesis by stable isotopes in cells

To measure stable isotope incorporation, cells were grown to 80-90% confluence

2 and then refreshed with fresh media containing 10% H2O at 37ºC for 24h.

Medium was removed and saved for percent enrichment measurements and the cells were isolated and stored at -80 °C until analysis. The Case Mouse

Phenotyping Center (MMPC) measured triglyceride concentrations and de novo 62

lipogenesis as previously described [162]. Briefly, triglyceride from cells was

isolated, and labeled glycerol and palmitate were analyzed after derivitization by

mass spectrometry. The 2H label on triglyceride covalently linked to glycerol measures the amount of newly synthesized triglyceride, while the 2H label in

triglyceride covalently attached to palmitate indicates the amount of new

palmitate. The contribution of de novo lipogenesis to the pool of triglyceride

palmitate was calculated using the following equation: U% newly made palmitate

U= [total 2H-labeled palmitate • (2H-labeled water × n)]-1 × 100, where n is the

number of exchangeable hydrogens, assumed to equal 22 [163, 164]. The

percentage of total newly made triglyceride glycerol was calculated using the

following equation: U % total newly made triglyceride-glycerol U = [2H-labeled triglyceride-glycerol • (2H-labeled water × n)]-1 × 100, where 2H-labeled triglyceride-glycerol is the M1 isotopomer, 2H-labeled water is the average amount labeled in the media, and n is the exchange factor (experimentally determined from the M2/M1 ratio of triglyceride glycerol).

Quantitative Realtime-PCR

Total RNA was isolated by RNeasy Kit with on-column DNase treatment

(Qiagen). cDNA was prepared from 5 µg of RNA in 20 µl reactions using a first

strand DNA synthesis kit (Invitrogen). cDNA samples were diluted 1:20 and realtime PCR was performed on 1 µl samples using iQ SYBR Green master mix and

the iCycler Real-time PCR Detection System (Bio-Rad). Each primer pair was

The fold change in expression of each mRNA relative to ATPB3, was calculated

_Δ(ΔCt), as 2 where ΔCt = Cttarget - CtATPB3 and Δ(ΔCt) =ΔCtCEF - ΔCtSki-CEF or

ΔCtCEF--ΔCtPPARγKD

Western Blotting

Whole-cell extracts were prepared from 90% confluent cells on 100-mm culture

dishes as follows: cells were scraped into 400 μl of lysis buffer (50 mm Tris, pH

8.0, 100 mm NaCl, 1 mm EDTA, 10% glycerol, 0.2% Nonidet P-40, 0.1 mm

sodium orthovanadate, 5 mm sodium pyrophosphate, 1 mm NaF, and Complete

protease inhibitor mixture (Roche Applied Science)). After 10-min incubation on

ice, the suspension was subjected to three freeze-thaw cycles, and cell debris

was removed by centrifugation at 12,500 rpm for 10 min at 4 °C. Protein

concentrations were determined by Bradford assays (Bio-Rad), and equal amounts of proteins were boiled in freshly prepared protein loading dye (100 mm

Tris, pH 6.8, 20% glycerol, 4% SDS, 0.2% bromophenol blue, and 1.5% β-

mercaptoethanol), Proteins were separated by 6%~10% SDS-PAGE, and transferred to Immobilon-P membranes (0.45 μm, Millipore). Primary antibodies used for immunoblotting were: Ski (G8, monoclonal[10], Lerner Research

Institute Hybridoma Core Facility), PPARγ (E8, monoclonal, Santa Cruz), MFN2

(XX-1, monoclonal, Santa Cruz), CytC (6H2, monoclonal, Santa Cruz), β-actin

(monoclonal, Sigma), NDFSU3 (MS112), 70Kda Fp (MS204), PDHE1-α (MSP07),

Complex III subunit core1(MS303) and porin(MSA03). Antibodies for the last five mitochondrial proteins are all monoclonal from Mitosciences.

Metabolite Oxidation

10 x106 cells were resuspended in 3 ml DMEM medium in a flask and mixed

individually with 1-14C-sodium Acetate (5mM, 1µCi), 14C-U-glucose (30mM, 1

µCi), 1-14C-glutamate (1mM, 0.5 µCi), 1-14C-octanoate (0.1mM, 0.5µCi) or 1-14C-

palmitate (1mM, 0.5 µCi). The experimental protocol and devices used for

14 14 trapping released CO2 were previously described [165]. The C in the medium

14 and CO2 were obtained by scintillation counting (Beckman Coulter).

Reporter Assay

2x104 cells (CEF and Ski-CEF) were plated in one well of the 96-well-plate and transfected with 20 ng p-CMV-PPRE firefly luciferase reporter and 20 ng pGL3tkRenilla luciferase reporter in the presence or absence of Ski expressing plasmid. Twelve hours later cells were treated with or without 10 µM GW1929.

Luciferase activity was measured 48 hours after transfection according to the protocol for the Dual-Glo luciferase assay system (Promega). For each experiment, at least 6 replicates were performed.

Softwares

Graphs were drawn and related statistical analyses were performed using

GraphPad Prism5 (www.graphpad.com). Flow cytometry data were analyzed by

FlowJo (www.flowjo.com). Expression array data were normalized and analyzed

by Genespring (www.agilent.com) and followed by pathway analysis using

ingenuity pathway analysis (IPA, www.ingenuity.com).

Results

Glycolysis is suppressed in Ski-CEFs

To establish the basis for reduced medium acidification by Ski-CEFs compared

to that of CEFs, we measured lactate production by by both cell types during 24

hours after medium change and during a five-day time course (Fig. 2). The

resulting data show that Ski-CEFs produce about half as much lactate as CEFs

during the initial 24 hour period (Fig. 2 A) and accumulate less than one fifth the

lactate compared to CEFs in five days (Fig. 2 B). To confirm these findings, we

next examined the consumption of glucose, the major source of lactate in culture

medium. As expected, we detected that compared to CEFs, Ski-CEFs utilize 2/3

as much glucose at 24 hours and 1/5 as much in five days (Fig. 2 C and D).

Reduced lactate production and glucose utilization indicate that Ski causes a

decreasing the rate of glycolysis in CEFs. However, this decrease is apparently

not the result of an overall reduction in metabolic activity as Ski-CEFs proliferate at the same rate as CEFs [3, 73]. To assess the metabolic activity of these cells, we measured the reduction of Alamar blue, a non toxic fluorescent dye, which is often used to assess the activity of mitochondrial NADH dehydrogenase [166].

Because Alamar blue is also reduced by cellular dehydrogenases that utilize

FAD, NAD and NADP as coenzymes [166-168], its reduction reflects the overall metabolic state of cells [169]. Here we show that the rate of Alamar blue reduction by Ski-CEFs is at least 1.8 fold more than that of equal numbers of

CEFs (Fig. 2 E). 66

Fig. 2 A and B

Lactate production is repressed in Ski-CEFs

(A and B) Medium from cultured CEFs and Ski-CEFs was sampled every 4 hours or every 24 hours for lactate measurement as described in methods. Data were normalized by cell numbers. Mean±SD was plotted as bar graph. (A, B)

Differences between CEFs and Ski-CEFs were significant (p<0.0001) as determined by two- way analysis of variance (ANOVA), n=3.

Fig. 2 C and D

(C and D) Medium from cultured CEFs and Ski-CEFs was sampled every 4 hours or every 24 hours for glucose measurement as described in methods. Data were normalized by that cell number. Mean±SD was plotted as bar graph. (C) *: p<0.05 by Student’s t-test, n=3. (D) Differences between CEFs and Ski-CEFs were significant (p<0.0001) as determined by two-way analysis of variance

(ANOVA), n=3.

Fig. 2 E

(E) Kinetics of reduction of 10% Alamar Blue.

2x 104 of CEFs or Ski-CEF were plated in one well of a 96-well plate. Six hours after plating, cells were incubated with DMEM medium containing 10% Alamar blue (Invitrogen). Plates were read at 590 nm for fluorescence intensity every 60 min. Data were normalized by total protein amount. Linear regression analysis was performed. The difference between slopes of two lines are extremely significant, p<0.0001, n=8.

TCA cycle is enhanced in Ski-CEFs

The decreased rate of glycolysis and their increased overall metabolic activity

suggested that Ski-CEFs might maintain a more active TCA (tricarboxylic acid)

14 14 cycle. We examined this possibility by measuring CO2 production from 1- C

14 14 labeled acetate, C-U-Glucose and 1- C-glutamate (Fig. 3). The results demonstrated that flux through TCA cycle was accelerated in Ski-CEFs (Fig. 3 A,

B and C). Acetate, which enters the TCA cycle directly as acetyl CoA was

oxidized 4.9 times faster. Glucose, which enters the TCA cycle as acetyl-CoA via

pyruvate dehydrogenase, was oxidized 3.5 times faster. Although they

consumed much less glucose in 24 hours than CEFs, ,about 10% of total

glucose utilized by Ski-CEFs was oxidized by the TCA cycle , while this ratio was

about 2% in CEFs (Fig. 2 C and Fig. 3 B). Glutamate, which enters TCA cycle

via amino acid metabolism, was converted into CO2 6.7 fold faster. Thus the

overall rate of TCA cycle flux is about 5 times greater in Ski-CEFs than in CEFs.

Fig. 3 A and B

Fig. 3 C

Fig. 3 Ski Stimulates TCA cycle activity in CEFs

(A, B and C) TCA cycle activity was examined by measuring the proportion of 14C

labeled substrates (sodium acetate, glucose and sodium glutamate) oxidized into

14 6 s C-CO2. 10x10 CEF or Ski-CEFs were trypsinized and suspended with DMEM

14 in a flask. C-CO2 was captured and quantified as described in Material and

Methods. Mean±SD were plotted as bar graphs. p-values were calculated by

Student’s t-test, n=3. (A) **, p<0.001. (B) ***, p<0.0001. (C) *, p<0.05.

Ski-CEFs exhibit enhanced mitochondrial respiration

The observed increases in Alamar blue reduction and TCA cycle flux in Ski-CEFs

suggested that these cells should have a high level of mitochondrial respiration.

We used two methods to measure mitochondrial respiration: heat generation and

oxygen consumption. Firstly, we used isothermal calorimetry to measure heat

production, which is a side product of cytochrome oxidase activity at the end of

the electron transport chain. We found that Ski-CEFs generated 2.6 fold more

heat than CEFs, suggesting the possibility that a higher proportion of respiration

might be uncoupled from oxidative phosphorylation in Ski-CEFs (Fig. 4 A). To

examine this possibility and to assess mitochondrial function in more detail, we

measured oxygen consumption of intact Ski-CEFs and CEFs in the absence and presence of the ATP synthase inhibitor, oligomycin and the uncoupling agent,

FCCP. By measuring the intact cell respiration in standard tissue culture medium,

we found that constitutive or routine oxygen consumption of Ski-CEFs was about

2-fold more than CEFs (Fig. 4 B and C). Upon treatment with oligomycin, this

rate was reduced by an even greater extent in Ski-CEFs than in CEFs, indicating

that oxygen consumption was coupled to ATP generation. This conclusion is

consistent with the finding that oxygen consumption was strongly stimulated by

uncoupling with FCCP.. In fact the uncoupled respiration of Ski-CEFs was 3-fold

higher than that of CEFs when both were maximally uncoupled by treatment with

FCCP. Finally, to estimate the possible contribution of non-mitochondrial oxygen

consumption, we measured residual oxygen consumption (ROX) upon addition of

the Complex III inhibitor, Antimycin A. The trivial level of ROX in both cell types

shows that the oxygen consumption was almost completely due to the

mitochondrial respiration. These results confirm our suggestion that Ski-CEFs

have a high rate of respiration but indicate that it is not due to uncoupling.

Fig. 4 A

Fig. 4 Analysis of Mitochondrial Activity

(A) Heat production by 1x106 cells per ml standard growth medium was measure

by isothermal calorimetry as described in Methods. The background was determined by measurement with culture medium alone. Mean±SD is shown, n=8, **, p<0.01 by Student’s t-test.

Fig. 4 B and C

(B) Graph of oxygen consumption by intact CEFs and Ski-CEFs (1x106 cells/ml) in growth medium. Oligomycin (2 µg/ml) was added to inhibit ATP synthase.

FCCP was titrated in steps to uncouple the oxidative phosphorylation and obtain the maximum oxidation capacity. Antimycin-A (Ama, 2.5 µM) was added to inhibit the mitochondrial respiration in order to measure residual oxygen consumption (ROX). (C) Data for the maximum response to each treatment in (B) are presented as bar graphs. Initial respiration of untreated cells is labeled routine. Respiration in the absence of ATP synthesis (plus oligomycin) is labeled leak. The maximum respiration due to FCCP treatment is labeled uncoupled.

Mean±SD was plotted, n=12 *, p<0.05, ***, p<0.001 by Student’s t-test

In order to determine the basis of enhanced respiration, we examined the mitochondrial oxidation capacity more directly by measuring oxygen consumption

of cells permeabilized with digitonin (Fig. 4 D). This approach allows the use of

selective metabolites and inhibitors to assess the activity of the individual

electron transport chain complexes [160]. Glutamate and malate were added as

external energy substrates to specifically provide NADH to the electron transport

chain (ETC) Complex I. Upon permeabilization, the rate of oxygen consumption

gradually decreased due to depletion of endogenous ADP. The very low level of

this so-called leak indicates that oxygen consumption by mitochondria of both

cell types was tightly coupled to ATP generation. Addition of ADP allowed

assessment of mitochondrial coupling and Complex I-dependent respiration. This

treatment stimulated oxygen consumption by both cell types indicating tight

coupling of electron transport to ATP synthesis. But clearly, ADP addition

increased respiration of Ski-CEF mitochondria to a much higher level than that of

CEF mitochondria. A similar difference was seen upon addition of succinate which supplies FADH to ETC Complex II thereby providing a measure of the respiration contributed by both Complex I and II. Oxygen consumption in the maximally uncoupled state following FCCP titration was also much greater in Ski-

CEFs but the results demonstrated that respiration by both types of mitochondria

was tightly coupled to ATP production. Subsequently, inhibition of Complex I by

rotenone allowed measurement of Complex II-dependent respiration, which was

also greater in Ski-CEF mitochondria. Both types of mitochondria showed

virtually no ROX following inhibition of Complex III by Antimycin A. Finally,

exogenous ascorbate and TMPD served as artificial electron donors to

cytochrome c that was oxidized in Complex IV. However, the autoxidation of

TMPD causes strong chemical background oxygen consumption. Therefore, to specifically measure Complex IV respiration rate, we used sodium azide to inhibit

the Complex IV activity and obtain the background correction.

Fig. 4 D

(D) Line graph of the permeabilized cell respiration. 1 1x106 /ml cells were

suspended in Miro5 respiration medium. Digitonin (1.2 µg/ml) was added to permeabilize cell membranes. Glutamate (10 mM) and malate (2 mM), were used as external substrates NADH provider. Complex I respiration was

measured by adding ADP (2.5 mM). The sum effect due to Complex I and

Complex II respiration was detected by adding succinate (10 mM). Rotenone

(Rot, 0.5 µM) was injected to inhibit Complex I activity, thus Complex II

respiration was measured. Antimycin A (Ama, 2.5 µM) were used to inhibit

Complex III and measure the non mitochondrial respiration. Artificial substrates,

Ascorbate (As, 2 mM) and TMPD (0.5 mM) were used to provide electrons for

Complex IV. Sodium azide (Azd, 100 mM), a cytochrome oxidase inhibitor was

added finally in order to correct the chemical background oxygen consumption and calculate respiration specifically from Complex IV.

Fig. 4 E

(E) Data from panel D are summarized as bar graphs. Each bar value represents the equilibrium value after each treatment in the line graph, except for routine which represents the initial respiratory rate prior to treatment. Leak = respiration after permeabilization without adding ADP; Complex I = respiration after ADP addition. Complex I and Complex II = respiration after succinate addition,

Uncoupled = maximum effect of uncoupler, FCCP. Complex II = respiration after the inhibition of Complex I by Rot, the respiration was measured. Respiration of

Complex IV was calculated by subtracting Azd effect from As+TMPD effect. ROX

(residual oxygen consumption was subtracted from each individual bars.

Mean±SD were plotted, n=6, p<0.001 by two-way ANOVA.

Quantification of the oxygen consumption showed that the routine respiration of permeabilized cells exhibited the same pattern as the intact cells in growth medium with a 2-fold higher rate of oxygen consumption by the mitochondria of

Ski-CEFs compared to those of CEFs (Fig. 4 E). When maximally uncoupled by

FCCP treatment, permeabilized Ski-CEFs consumed 3 fold more oxygen than

CEFs, which was also consistent with the intact cell measurements. Respiration with substrates feeding electrons into either Complex I, Complex II or Complex IV was increased about 3 fold in Ski-CEFs compared to CEFs. Thus the enhanced oxygen consumption caused by Ski was not due to a selective effect on a specific electron transport complex (ETC). Finally, we calculated the uncoupling control ratio (UCR) by dividing the maximum uncoupled respiration by the routine respiration (Fig. 4 F). The UCR reflects the reserve capacity of the electron transport system [170]. This ratio was similar in CEFs and Ski-CEFs in both intact and permeabilized respiration experiments, indicating that the difference in mitochondrial respiration between Ski-CEFs and CEFs is due to the difference of total mitochondrial content between two kinds of cells. In combination, the results showed that Ski-CEFs had enhanced mitochondrial respiration that was due to a similar increase in all ETC complexes and was therefore the result of an increase in the number or mass of mitochondria.

Fig. 4 F

(F) UCR (uncoupling control ratio) of both the intact and the permeabilized cell respiration was shown as bars. UCR is defined as: UCR=Ju/Jr, where Ju and Jr are uncoupled and routine oxygen consumption.

Ski increases mitochondrial number and mass in CEFs

To determine whether Ski-CEFs have more mitochondria than CEFs we first quantified mitochondria by protein analysis. Immunoblotting revealed that Ski-

CEFs express higher levels of the mitochondrial outer membrane proteins (MFN2 and porin); mitochondrial intermembrane protein, cytochrome c (Cyt C); mitochondrial matrix proteins: pyruvate dehydrogenase subunit E1-α (PDH E1-

α); mitochondrial respiration chain complex proteins: NDFSU3, a Complex I

subunit, 70kDa Fp, a Complex II subunit and subunit core I of complex III than

CEFs (Fig. 5 A). In addition, assays of citrate synthase (CS) enzyme activities

also showed that the mitochondrial content of Ski-CEFs is greater than that of

CEFs (Fig. 5 B).This conclusion gained additional support from both flow

cytometric and electron microscopic analyses. We analyzed CEFs and Ski-CEFs by flow cytometry using both Mitotracker Green (MTG) and Mitotracker Red

(MTR), fluorescent dyes that covalently bind mitochondrial matrix proteins irrespective of mitochondrial membrane potential [171]. The observed increased fluorescence of Ski-CEFs compared to that of CEFs indicates a 2.2 fold greater mitochondrial mass, consistent with the 2 fold greater routine oxygen consumption caused by overexpression of Ski (Fig. 5 C). Forward scattering data

(not shown) indicated that the increased mitochondrial mass per cell was not due

to a difference in cell size between Ski-CEFs and CEFs. Finally, electron

microscopy (EM) revealed that both the mitochondrial number and size were

increased in Ski-CEFs. The 60K magnification images showed that Ski-CEFs

have different mitochondrial infrastructures than CEFs. Besides CEFs had

85 stronger mitochondrial matrix staining, mitochondrial cristae of Ski-CEFs were more compact thin layers, whereas cristae were more expanded in CEFs. (Fig. 5

D). We also attempted to assess the number of mitochondria per cell by quantifying mitochondrial genomic DNA by real-time Q-PCR [172, 173]. Our results, normalized to a single copy nuclear gene, show that these two cell types have the same amount of mitochondrial DNA per cell (Fig. 5 E). However, the results of four other methods presented in figure 4 clearly demonstrated that Ski-

CEFs contain about two-fold more mitochondria than CEFs. We believe the discrepancy can explained by the fact that mitochondrial DNA copy number can vary from 1-15 in each mitochondrion [174, 175]. Thus differences in the number of genomes per mitochondrion in the two cell types likely accounts for the failure of mtDNA quantification to accurately assess the number of mitochondria in Ski-

CEFs and CEFs.

Fig. 5 A

Fig. 5 Ski Increases Mitochondrial Mass and Numbers

(A) The expression of mitochondrial proteins in CEFs and Ski-CEFs was determined by immunoblotting as described in Methods. Porin and MFN2

(mitofusion-2) are mitochondrial outer membrane proteins. PDH E1α (pyruvate dehydrogenase subunit) is a matrix proteins, and CytC (cytochrome C) is an intermembrane protein. NDUFS3 (Complex I subunit), 70kDa Fp (Complex II subunit) and Complex III subunit core 1 are inner membrane electron transport chain proteins. β-actin serves as a loading control.

Fig. 5 B

(B) Activity of citrate synthase, an indicator of mitochondrial content, was measured with lysates from 1x106 cells/ml as described in Methods. Mean±SD

was plotted as bar graph, n=5, ***, p<0.001 by Student’s t-test.

Fig. 5 C

Fig. 5 C CEFs and Ski-CEFs at a concentration of 2x106 /ml were stained with 80

nM MTG (mitotracker green) or 100 nM MTR (mitotracker red) and analyzed by flow cytometry to measure the mitochondrial mass. Data was collected from

10,000 cells and subsequently analyzed by FlowJo software.

Fig. 5 D

(D) Electron micrographs of CEFs and Ski-CEFs obtained at the indicated magnifications show that Ski increases the number and size of mitochondria.

Scale bars: 2 µm at 3K magnification, 1 µm at 10K magnification and 0.5 µm at

60K magnification.

Fig. 5 E

Fig.5 E Mitochondrial genomic DNA measured by Q-RTPCR

Q-RTPCR was performed using total cellular DNA at three concentrations (0.32-

10ng/25ul reaction) as the template. Two different pairs of primers for amplifying mitochondrial 16S rDNA were used. Amplification using primers for intron five of

Smad 2, a single copy nuclear gene, served as an internal control. The data points in the figure are the threshold cycles of 16S rDNA normalized to that of

Smad2. The differences between the normalized threshold cycles of 16S rDNA from CEFs and Ski-CEFs (dCt) were independent of the amount of input DNA.

Ski increases expression of PPARγ and lipid metabolic genes regulated by

To determine the molecular basis for the increased mitochondrial function and

mass in Ski-CEFs, we explored global mRNA expression by Affymetrix micro-

array. Analysis of the results using GenespringTM and Ingenuity Pathway

Analysis software revealed a cluster of 182 upregulated genes that were involved

in lipid metabolism. Twenty-nine of these 182 genes were related to lipid and

fatty acid oxidation (Table.1). Importantly, these genes included the lipid

metabolism and mitochondrial biogenesis regulator, PPARγ, which is elevated by

4.4 fold in Ski-CEFs. Furthermore, several of these upregulated genes are known

to be directly regulated by PPARγ including CD36, ACOX1, FABP1, 3 and 4,

FATP-1, Gyk, LPL and Perilipin 1 [106, 107, 143-150]. We confirmed that PPARγ

was dramatically up-regulated at both the mRNA and protein levels in Ski-CEFs by Q-RTPCR and Immunoblotting (Fig. 6 A and B). Q-RTPCR data also verified the increased expression of PPARγ regulated genes involved in lipid transport and metabolism as well as the mitochondrial biogenesis related gene, PGC1-α

(Fig. 6 C). These gene expression data suggest a strong connection between

PPARγ activation and the alteration in cellular energy metabolism affected by Ski.

Table. 1

Lipid metabolic genes regulated in Ski-CEFs.

A partial list of lipid metabolic genes was generated by Ingenuity pathway analysis (IPA). Fold change calculated from the normalized microarray data show expression of genes in Ski-CEFs over that of CEFs. Gene symbols are those used by Genbank.

Fig. 6 A and B

Ski upregulates PPARγ at both mRNA and protein level

Differential PPARγ expression in Ski-CEFs and CEFs measured by Q-RTPCR (A) and by immunoblot assay (B). mRNA expression was first normalized to an internal control, ATPB3. The plotted expression value was normalized to that of

CEF, so that the expression in CEF was arbitrarily set to 1. Mean±SD is shown as a bar graph, n=6, ***, p<0.001 by Student’s t-test (A).

Fig. 6 C

(C) Expression of PPARγ targets by Q-RTPCR in Ski-CEFs relative to CEFs. The same normalization was performed as described in panel A. Threshold cycle values used to calculate mRNA expression were first normalized to an internal control, ATPB3. The plotted expression values were normalized to those of CEFs, so that the expression of each gene in CEF was arbitrarily set to 1. n=3, p-value by Student’s t-test. *, p<0.05, **, p<0.01 and ***, p<0.001.

Ski is a co-activator of PPARγ dependent transcription

In light of these findings, we examined the functional significance of elevated

PPARγ expression more directly by performing transient transcription assays with a luciferase reporter controlled by a PPARγ-responsive responsive element

(PPRE). The results obtained showed that under normal growth conditions with no added PPARγ ligand, the reporter was 5-fold more active in Ski-CEFs than in

CEFs (Fig. 6 D). Importantly, results of co-transfection reporter assays in CEFs revealed that Ski dramatically transactivated the PPRE reporter (Fig. 6 E).

Expression of the reporter was induced about five-fold by addition of a PPARγ

ligand and co-transfection of Ski increased ligand-activated expression twenty- fold. These results suggested that Ski might act as a PPARγ co-activator by direct protein/protein interaction. This suggestion was supported by the finding that endogenous PPARγ co-immunoprecipitated with Ski (Fig. 6 F). All of these results suggested that PPARγ signaling is strongly activated by Ski and might be responsible for the elevated expression of PPARγ targets leading to enhanced

lipid metabolism in Ski-CEFs.

Fig. 6 D and E

100

Fig. 6 (D and E) Ski co-activates PPARγ-dependent transcription.

CEFs and Ski-CEF were transfected with a PPRE driven reporter (D), or CEFs were transfected with PPRE driven reporter in the presence or absence of a Ski expression plasmid (E). Cells were untreated or treated with the PPARγ specific agonist, GW1929 at the concentration of 10 µM. Values were normalized to the internal control renilla luciferase activity and the Mean±SD plotted as bar graphs. p-value by Student’s t-test, n=4, each experiment with 6 replicates, *, p<0.05, **, p<0.01.

101

Fig. 6 F

(F) Ski interacts with PPARγ

Whole cell extracts from Ski-CEFs were immunoprecipitated and immunoblotted with anti-Ski or anti-PPARγ antibodies, to determine whether Ski interacts with

PPARγ. 10% of the lysates used for immunoprecipitation was loaded as the input control. Mouse IgG was used as the negative control. Anti-Ski antibody, H329

(Santa Cruz), anti-PPARγ antibody, E8 (Santa Cruz).

102

Ski increases β-oxidation of fatty acids

To determine whether the observed changes in PPARγ regulated gene

expression were reflected in alterations of lipid metabolism, we measured fatty

acid oxidation in Ski-CEFs and CEFs (Fig. 7 A and B). We found that oxidization

of both the medium-chain fatty acid, octanoate and the long-chain fatty acid,

palmitate, was about 2 times greater in Ski-CEFs than in CEFs. These results

suggest that the observed enhancement of TCA cycle flux in Ski-CEFs might be in part due to the increased fatty acid β-oxidation. The acetyl CoA produced enters TCA cycle for completely oxidation.. To determine whether increased lipid synthesis might provide fatty acids for enhanced β-oxidation in Ski-CEFs, we

2 measured the incorporation of H2O into palmitate and triglycerides by mass

spectrometry (Fig. 7 C). Surprisingly, we found that the contribution of de novo

lipogenesis to newly synthesized palmitate was 60% less in the Ski-CEFs compared to CEFs. In addition, by measuring 2H-labeled triglyceride-glycerol, we

found that the amount of newly made triglyceride was similar in CEF and Ski-

CEF cells. Given the observation of the decreased palmitate synthesis but the

increased palmitate oxidation in Ski-CEFs, we suggest that the increased utilization of fatty acids rather than de novo synthesis is responsible for the metabolic shift to more oxidative metabolism of lipids in Ski-CEFs. This conclusion is further supported by our micro-array and Q-RTPCR data showing an increased expression of genes involved in fatty acid import and intracellular transport in Ski-CEFs.

103

Fig. 7 A and B

104

Fig. 7 (A and B) β-oxidation was examined by measuring the proportion of 14C

14 14 14 14 labeled 1- C- octanoate (A) and 1- C- palmitate (B) oxidized into CO2. CO2 was captured and quantified as described in Material and Methods. 10x 107 cells were used for each experiment. Mean±SD was plotted as a bar graph. p-values were calculated by Student’s t-test, n=3. (A) **, p<0.01. (B) *, p<0.05.

105

Fig. 7 C and D

(C and D) Newly synthesized palmitate (C) and triglycerides (D) in CEF, Ski-CEF

2 and PPARγ KD cells cultured in the presence of H2O for 24h were measured by mass spectrometry. Values are the means ± SEM, n=3, *, P<0.05 by Student’s t-

test.

106

PPARγ mediates the induction of lipid oxidation and mitochondrial mass by Ski

To determine whether PPARγ is directly involved in the Ski induced enhancement of mitochondrial mass and lipid oxidation, we used RNAi to knockdown expression of the endogenous PPARγ in Ski-CEFs. We screened and picked Ski-CEF clones with optimal PPARγ knock-down by three different shRNAs targeting PPARγ (Fig. 8 A). Although PPARγ expression was not completely eliminated by any of these shRNAs, it was reduced to levels at or below that of CEFs. The metabolic activity of these clones was then compared with that of Ski-CEFs and CEFs by first measuring the reduction of Alamar blue

(Fig. 8 B). We found that reduction of Alamar blue was dependent on the level of

PPARγ expression. Even clone 2 which had a relatively high level of PPARγ showed about 75% reduction activity compared to Ski-CEF. The observation that three different shRNAs targeting PPARγ reversed the overall metabolic activity of

Ski-CEFs to that of CEFs indicated that the effect was not due to an off-target knock-down.

107

Fig. 8 A

(A) Screening of PPARγ knockdown Ski-CEF clones by immunoblotting. Band intensities were determined by densitometry using ImageJ software

(http://rsbweb.nih.gov/ij/). For each clone, intensities of PPARγ bands were

normalized to that of β-actin.

108

Fig. 8 B

(B) Alamar blue reduction was reduced in individual PPARγKD-Ski-CEF clones.

Experiments were done in the same way as described in Fig. 2 E. Data were normalized to the total DNA amount determined by Hochest 33322 fluorescence at 365 nm.. Mean±SD was plotted, p<0.001 by two way ANOVA, n=8.

Fluorescence of reduced Alamar blue at five hours by individual PPARγKD-Ski-

CEF clones is shown.

109

Those three clones with the lowest PPARγ protein levels and the lowest Alamar

Blue reduction activity (Clones 3, 9 and 10) were pooled and further analyzed.

The results showed that the induction of PPARγ target genes involved in lipid

oxidation and transport or mitochondrial biogenesis was drastically decreased

when the elevated PPARγ expression was reversed in Ski-CEFs (Fig. 8 C). In

PPARγ knockdown Ski-CEFs, expression of most of these genes was decreased

to about that observed in CEFs. The induction of FABP-3 and FABP-4 in PPARγ

knock-down cells was not completely blocked, but compared to CEFs, it was

reduced by 95% and 88%, respectively.

Analysis of respiration of the intact and the permeabilized cells revealed that the

PPARγ knock-down clones’ consumption of oxygen by the entire ETC and

individual mitochondrial ETC complexes was reduced to that of CEFs (Fig. 8 F

and G). Similarly, knocking down PPARγ expression in Ski-CEFs also resulted in

increased lactate production and glucose utilization, suggesting the entire shift in

energy metabolism by Ski was reversed (Fig. 8 D and E). Interestingly, mass

spectrometric analysis of 2H-labeled lipids revealed that both de novo palmitate synthesis and newly made triglyceride were reduced in PPARγ knockdown cells compared to CEFs and Ski-CEFs (Fig. 7 C and D). Thus although lipid synthesis was not enhanced in Ski-CEFs, the constitutive activity of PPARγ in promoting lipid synthesis was likely necessary for the metabolic shift induced by Ski.

110

Fig. 8 C

PPARγKD-Ski-CEFs. Relative mRNA levels were measured by Q-RTPCR and normalized to that of ATPB3 as in Fig. 6C. The expression values were normalized to that of CEF, so that the expression of each gene in CEF was arbitrarily set to 1. Mean±SD is shown, n=3, p-value by Student t-test. **, p<0.01 and ***, p<0.001. Each gene’s expression value of Ski-CEFs is significantly different with that of other two cell types.

111

Fig. 8 D and E

112

Fig. 8 (D) Cells at 80-90% confluence were incubated with fresh culture medium for 24 hours. Medium was sampled to measure the lactate production (D) and glucose utilization (E) as in Fig 1. Mean±SD is plotted as a bar graph. N=3, ***, p<0.0001 by one way ANOVA.

113

Fig. 8 F

(F) Oxygen consumption from intact PPARγKD-Ski-CEF cells.

Experiments were performed in the same way as described in Fig. 4 B.

Mean±SD is plotted, n=3 ***, p<0.001 by two-way ANOVA.

114

Fig. 8 G

(G) Permeabilized cell respiration in PPARγKD-Ski-CEFs.

Experiments were performed in the same way as described in Fig. 4 D.

Mean±SD were plotted, n=3, p<0.001 by two-way ANOVA.

115

Finally, we asked whether increased PPARγ expression was also necessary for the observed increase in mitochondrial mass in Ski-CEFs. Immnuoblot analysis of mitochondrial proteins revealed that by knocking-down PPARγ, the expression of mitochondrial ETC complex proteins, matrix protein and outer membrane proteins in Ski-CEFs was greatly reduced though not quite to the level of CEFs

(Fig. 8 I). Similarly, analysis of mitochondrial mass by flow cytometry of MTG and

MTR fluorescence showed that mitochondrial mass of PPARγ knockdown Ski-

CEFs was slightly more than in CEFs but much less than that of Ski-CEFs (Fig. 8

H). EM of these cells demonstrated that the mitochondrial number and size of

PPARγ knockdown Ski-CEFs was similar to that of CEFs but less than that of

Ski-CEFs (Fig. 8 J). The combined results of these PPARγ knockdown studies establish a central role for PPARγ as a mediator of the increased fatty acid oxidation and mitochondrial mass that underlie the enhanced respiration of Ski-

CEFs.

116

Fig. 8 H

Fig. 8 H mitochondrial content difference among CEFs, Ski-CEFs and

PPARγKD-Ski-CEFs is shown by immunoblot, Porin and MFN2 (mitofusion-2) are mitochondrial outer membrane proteins. PDH E1α (pyruvate dehydrogenase subunit) represents matrix proteins. NDUFS3 (Complex I subunit), 70 kDa Fp

(Complex II subunit) and Complex III subunit core 1 are electron transport chain proteins.

117

Fig. 8 I

Fig. 8 I Measurement of total mitochondrial mass in CEFs, Ski-CEFs and

PPARγKD-Ski-CEFs. Cells (2x106/ml) were stained with 80 nM MTG (mitotracker green) or 100 nM MTR (mitotracker red) and analyzed by flow cytometry. Data was collected from 10,000 cells and subsequently analyzed by FlowJo software.

118

Fig. 8 J

Fig. 8 J Electron micrographs show that the increase of mitochondrial size and number in Ski-CEFs is reversed when PPARγ is knocked down.

Scale bars: 2 µm at 3K magnification and 1 µm at 10K magnification.

119

Discussion

In this study, we delineated a molecular mechanism by which Ski shifts the

energy metabolism of CEFs from glycolysis to oxidative phosphorylation. In

concert with the decreased lactate production and glucose utilization (Fig. 2), we

found by Affymetrix microarray analysis that the expression of ADP-dependent

glucose kinase (ADPGK) and phosphofructokinase (PFK) was 3-fold lower by in

Ski-CEFs. Both kinases produce metabolic intermediates in glycolysis but PFK plays a key role by catalyzing the first regulated reaction of the glycolytic pathway

[176]. Furthermore, Ski elevated expression not only of PDH that catalyzes the conversion of pyruvate to acetyl CoA [177] (Fig. 5 A) but also of pantothenate

kinase that is a key component of the CoA synthetic pathway [178]. These

alterations in gene expression likely contributed to the metabolic shift to

increased TCA cycle flux that accompanies Ski transformation of fibroblasts

which normally generate energy through glycolysis.

The shift to oxidative metabolism in Ski-transformed cells was unexpected

because fibroblasts transformed by other oncogenes as well as most tumor cells

display dramatically elevated lactate production via glycolysis in the presence of

adequate oxygen [154, 179, 180]. This apparent inconsistency raised the

question of whether the shift to oxidative metabolism plays an obligatory role in

Ski transformation or is instead an unrelated activity of Ski. Our preliminary

attempts to answer this question have yielded inconclusive results but provide a

120

starting point for future experiments. We have found that CEFs expressing a

well-studied transformation defective Ski mutant∆Z3/ ( 4) [5] acidify culture medium and reduce Alamar blue at the same rate in normal CEFs. Furthermore, microarray and Q-RTPCR results showed that this mutant does not induce any of the lipid metabolic genes that are induced by wild-type Ski, including PPARγ and

PGC1-α (Fig. 8 s). These results suggested that Ski-induced transformation and

altered metabolism are highly correlated. On the other hand we have found that

knock-down of PPARγ in Ski-CEFs does not reverse their phenotypic

transformation. This result indicated that the correlation revealed by the Ski

mutant did not provide conclusive evidence for an obligatory role for enhanced

oxidative metabolism in cellular transformation by Ski.

Regardless of their possible connection to transformation, our results provided

insights into the functional interrelationship between Ski and PPARγ. We found

that instead of its well established role as a master regulator of adipogenesis,

under the influence of Ski, PPARγ mainly acts to promote oxidative metabolism

in general, and fatty acid β-oxidation specifically [81, 90, 91]. This might be an expected role for PPARα or PPARδ, which have been shown to regulate fatty acid oxidation and catabolism [83, 84]. However, neither of these genes were upregulated in Ski-CEFs so we focused on PPARγ whose expression was induced by Ski. This choice was also stimulated by the concordance of our gene expression data with published studies demonstrating that administration of the highly selective and potent PPARγ agonist, GW1929, to rats resulted in an

121

increased expression of PPARγ target genes involved in both lipid uptake and

lipid oxidation in BAT and WAT [113]. Among these PPARγ target genes, CD36

and FATP-1 were shown to enhance palmitate oxidation through their role in

delivering long chain fatty acids to carnitine palmitoyltransferase for transport into

mitochondria [102, 181-183]. These two and several additional PPARγ regulated

genes involved in the import and intracellular transport of fatty acids were also

found to be upregulated by Ski, including LPL, FABP3 and FABP4. Through its

interaction with CD36, LPL functions in the import of extracellular lipids into cells

[109, 184]. Interestingly, our microarray data also showed that fasting induced

adipose factor (FIAF), a known LPL inhibitor [185] was down regulated by 3 fold

and the lipid carrier protein ApoA1 was dramatically upregulated in Ski-CEFs. On the other hand, we did not detect increased expression of the PPARγ regulated genes involved in lipogenesis. This finding is consistent with our stable isotope incorporation studies which indicated that de novo fatty acid synthesis was decreased in response to Ski. Despite this, the level of newly synthesized triglycerides was maintained in Ski-CEFs, almost certainly a consequence of increased import of extracellular lipids.

In addition to the genes involved in lipid transport, we also detected increased expression of PPARγ regulated genes that play supporting roles in lipid metabolism. One of these, PGC1-α, which is also a transcriptional coactivator with PPARγ, is known to regulate genes that promote fatty acid oxidation, cell respiration and non-shivering thermogenesis [112]. In vivo, PGC1-α drives the

122

formation of oxidative slow type I muscle and PGC1-α transgenic mice display enhanced exercise ability and oxygen uptake, suggesting increased utilization of fatty acid [111, 186]. Pantothenate kinase is another PPARγ target that was found to be upregulated by Ski. By catalyzing the first rate limiting step in the synthesis of CoA [178, 187] it is a likely contributor to the enhanced β-oxidation we observed in Ski-CEFs. Taken together our results indicated that Ski forms a selective partnership with PPARγ to promote the import and intracellular transport of extracellular fatty acids to produce energy via enhanced β-oxidation.

Our results also clearly implicated PPARγ in the observed decreased utilization of glucose for energy production via glycolysis as this effect of Ski was reversed upon knocking down PPARγ expression in Ski-CEFs.

Transcription factors, c-Myc and PGC1-α have been shown to stimulate mitochondrial biogenesis [109, 188, 189]. HIF-1 inhibits mitochondrial biogenesis and cell respiration by promoting c-Myc degradation [172]. Interestingly, we found that the expression of c-Myc was increased but the expression of HIF-1 was decreased in Ski-CEFs (Affymetrix data). Since no direct interaction between PPARγ and HIF-1 has been reported, Ski-CEFs might utilize the Myc- regulated pathway to increase mitochondrial biogenesis in addition to the

PPARγ-PGC1α pathway. This speculative possibility is supported by our observation that mitochondrial biogenesis was not completely reversed in PPARγ

KD cells (Fig.7 H, I and J). Regardless of the possible involvement of Myc, it is clear that PPARγ is the major mediator of Ski-induced mitochondrial biogenesis

123

in CEFs. The resulting increase in mitochondrial mass provides the machinery for

the shift in energy production pathways from glycolysis to β-oxidation in Ski-

CEFs.

Studies of Ski transgenic mice show that Ski regulates body composition by

increasing type IIb muscle fibers while decreasing adipose tissue mass [63, 142].

A more recent analysis of the affected muscles of these mice revealed that

increased expression of Ski represses lipogenesis by downregulating the

expression of PPARγ, PGC1α and SREBP [66]. Surprisingly, these studies also

showed that Ski increases fatty acid oxidation capacity in skeletal muscle. Our

observation that Ski enhances β-oxidation of lipids in muscle agrees with our

findings in Ski-CEFs. But it is more difficult to reconcile Ski’s repression of

lipogenesis in Ski-CEFs where PPARγ expression is increased compared to that

of CEFs. However, it has been reported that rats treated with GW1929 have

different responses to this PPARγ agonist in different tissues [113]. In skeletal

muscle, an obvious decrease of the expression of genes involved in lipid uptake and oxidation was observed, even though some of these genes were known to

be induced by PPARγ in other tissue. The situation was completely the opposite

in adipose tissue. This discrepancy between observations is likely due to tissue

specificity and the dual role that PPARγ plays in lipid metabolism regulation. The

combined results and this duality of function suggest that PPARγ might induce

lipid synthesis but not fatty acid oxidation in skeletal muscle. However, in CEFs under the influence of Ski it has the opposite.

124

Fig. 8 K A

125

Fig. 8 K ΔZ3/4-CEFs do not display a metabolic shift as observed in Ski-CEF (A)

ΔZ3/4- CEF reduces Alamar blue at the similar level as CEFs. (B) The mRNA of

PPARγ and PGC1-α expression show no significant difference between ΔZ3/4-

CEFs and CEFs.

126

Chapter II Smad2 Mediates Ski Induced Oncogenic Transformation in CEFs

Abstract Overexpression of the Ski oncogene results in oncogenic transformation of avian fibroblasts. Previous results showed that Ski negatively regulates TGF-β signaling by binding Smad proteins and reversing their transcriptional activity. In several cell types this activity of Ski reverses the growth inhibitory action of TGF-

β and in chicken embryo fibroblasts it appears to be required for cellular transformation. TGF-β signaling results in the activation of both Smad2 and

Smad3 by phosphorylation, but these transcription factors activate distinct downstream pathways. In this study, we investigated the role of these two Smads in mediating transformation by Ski. We found that wild-type Ski, but not transformation defective Ski mutants, increases the Smad2/Smad3 ratio at both the mRNA and protein levels. In addition, the receptor-phosphorylated form of

Smad2 was found to be stabilized in Ski-transformed cells and insensitive to inhibition of TGF-β receptor activation. Furthermore, knocking down Smad2 but not Smad3 expression with shRNA blocked transformation by Ski. Our results establish the requirement for the TGF-β signaling pathway in Ski transformation and identify Smad2 as the essential mediator.

127

Introduction

The retroviral v-ski oncogene was discovered by virtue of its ability to transform

primary avian fibroblasts [1, 140]. Later it was found that the cellular homolog, c- ski, when experimentally overexpressed, was even more potent in transformation than the original retroviral oncogene [64]. Both in vitro and in vivo studies revealed that ski could also function as a co-operating oncogene to transform bone marrow cells and induce erythroid/myeloid leukemias [56, 190]. Surprisingly, analysis of the same chemically induced tumor in mice suggested that Ski acted as a tumor suppressor [191] and this observation has gained some support in studies of selected human tumors [192]. However, there is also convincing data that SKI functions as an oncogene in several human tumors, including melanoma,

AML, esophageal, gastric, pancreatic and colon carcinomas [35, 48, 54].

The oncogenic action of Ski has been linked to its function as a potent transcriptional co-regulator. Ski does not bind DNA but rather exerts its regulatory activity by protein-protein interactions with several other co-regulators and DNA-binding transcription factors [21, 22, 29, 51, 193]. These interactions are mediated, for the most part, by two conserved functional domains in the N- terminal half of Ski whose structures have been determined by X-ray crystallography [7]. The more N-terminal of these is the Dachshund homology domain (DHD) through which Ski interacts with N-CoR, RAR, Smad2/3, Six1,

GATA1 and PU.1 [23, 25, 27, 29-31, 36, 40, 138]. The second domain, SAND

(Sp100, AIRE, NucP41/75 and DEAF1), is responsible for the interaction

128 between Ski and Rb, Smad4, FHL2 and MeCP2 [12, 17, 18, 32, 33, 194].

Through these interactions Ski regulates transcription either positively or negatively [9, 17, 18, 23, 32, 33, 40].

Only two sets of transcriptional partners have been directly linked to the oncogenic activity of Ski. The first of these was discovered by Hayman and coworkers who showed that Ski binds to retinoic acid receptor alpha (RARα) and stabilizes it as a transcriptional repressor, even in the ligand-bound form [36].

Through this activity, Ski suppresses differentiation of both chicken and human myeloid cells and promotes myeloid leukemogenesis [23, 54]. The second set of interactions shown to play an important role in transformation by Ski centers on the TGF-β pathway and includes Smads 2,3 and 4 and MEL1 [4, 9, 17, 18, 23,

25, 27, 32, 33, 36, 40, 41]. Through these interactions, Ski acts as a Smad co- repressor thereby suppressing the cytostatic and metastatic effects of TGF-β signaling to promote the development of human melanoma, and gastric and pancreatic carcinomas [9, 35, 44-46, 48] . Interaction with the Smad proteins has also been directly connected to the capacity of Ski to transform primary chicken embryo fibroblasts (CEFs) thereby implicating the TGF-β pathway in this process as well [5, 139].

Because TGF-β, acts as a potent inhibitor of proliferation, the members of the pathway have been classified as tumor suppressors {[42]}. TGF-β signaling is generated from ligand-binding transmembrane receptors that phosphorylate

129

cytoplasmically anchored Smad2 and Smad3 [42, 115]. As a result of

phosphorylation, these receptor-activated rSmads form complexes with the

common mediator Smad4 that enter the nucleus and activate the transcription of

target genes [42, 115]. Transcriptional activation by Smads is mediated by direct

DNA binding to Smad binding elements (SBE) or by interaction with other DNA binding transcription factors [195]. The latter mechanism is the predominant mechanism for the most highly expressed form of Smad2 which, unlike Smad3, does not possess an intact DNA binding domain [196]. In addition to this obvious difference between Smad2 and Smad3, it is clear that, although they are both activated by the same TGF-β receptor, the two proteins activate distinct downstream responses [114] .

In the nuclei of transformed CEFs, Ski forms complexes with rSmads and Smad4 that bind the SBE and represses rather than activate transcription [40, 134].

Surprisingly, this nuclear association and activity occurs in the absence of added

TGF-β which is thought to be required for nuclear localization of rSmads [5, 135,

136]. It seems likely that binding by Ski is responsible for nuclear localization or retention of the Smads under these conditions. Three separate regions within the minimal transformation domain in the Ski protein are responsible for interaction with the Smads [4, 12, 25, 137, 138]. Two N-terminal sites mediate Ski binding to

Smad2 and 3. The more N-terminal site at amino acids (aa) 18-35 aa is interrupted by the truncation of v-Ski and inactivated by a L19A substitution mutation. The second site is positioned in the N-terminal region of DHD domain

130 and its binding of Smad2/3 is abolished by a L110P substitution mutation [4, 12,

25, 137]. The third site is in the SAND domain and is responsible for Ski binding to Smad 4 and is disrupted by deletion of a zinc finger in the ΔZ3/4 mutant or by the W255E substitution [12, 138]. These and other mutations in these domains that significantly reduce the binding of Ski to Smads and its potency as a Smad co-repressor also reduce its transforming activity [5, 139]. These observations suggest that the direct linkage between the Ski-Smad interaction and transformation involves an alteration of TGF-β signaling by Ski.

The mutations that disrupted Ski/Smad interactions were in domains that mediate other protein interactions of Ski and are likely to abolish those interactions as well.

Therefore in the present report we further investigated the requirement for Ski’s influence on the TGF-β signaling pathway in its ability to transform CEFs by modifying Smad expression or activity. In light of the known differences between

Smads 2 & 3, we examined the requirement for specific Ski/Smad interactions in transformation. Our results show that transformation of CEFs by Ski is accompanied by a switch in the relative levels of expression of Smad2 and

Smad3 proteins. We found that a large fraction of the Smad2 protein in these cells is nuclear and is receptor phosphorylated in the absence of exogenous

TGF-β. We determined that re-establishing the normal Smad3/Smad2 expression ratio did not affect Smad2 nuclear localization or Ski induced transformation. By knocking down Smad2 or Smad3 expression, we demonstrate that the transforming capacity of Ski is dependent on Smad2. Our results provide

131

evidence for an essential link between Ski’s transforming activity, its interaction

with TGF-β signaling and its selective influence on constitutive Smad2 activation.

Experimental Methods

Expression Vector Construction

The flagSmad2E2 mutant was generated by PCR mutagenesis. A primer set to

amplify Smad2 (5’-ctccagttaaccatagcctggatttgc-3’ and

5’ggccctctagacttgagttattccatctcggagcagcgtactg-3’) was used and the PCR

product was subcloned at HpaI and XhoI sites to generate pcDNA3flagSmad2 .

FlagSmad2E2 was transferred into a chicken retroviral vector RCAS that was

previously described [158]. Ski mutants (SkiL19A, SkiL110P, SkiW255E, Ski (L19A, L110P),

Ski (L19A, L110P, W255E) were kindly provided by Dr. Michael Hayman and sub-cloned

downstream of the puromycin resistance gene in the replication defective

retroviral vector RdpI (modified from RCAS).

A microRNA-30 (miR30) expression cassette was subcloned from the SIN-

TREmiR30-PIG (TMP) vector (kindly provided by Dr. Scott Lowe) [155-157] into

RdpI, to generate a chicken shRNA expressing vector, RdpImiR30, using KpnI

and PmeI. The original shRNA insertion sites (XhoI and EcoRI) in miR30

cassette was substituted with XhoI and MfeI by PCR (primer set: F,

5’AGAAGGCTCGAGCAACCACAATTGAAGGGG3’ and R, 5’

GCGCCTACCGGTGGATGTGGAATGTGT3’). Based on the mRNA sequence of

132

Gallus Smad2 (GenBank NM_204561) and Smad3 (Genbank NM_204475), four

21mer-shRNAs (nucleotides 566-586, 963-983, 384-404 and 1087-1107)

targeting chicken Smad2 and two 21mer shRNAs (359-379 and 809-829)

targeting Smad3 were chosen by using RNAi Codex (http://cancan.cshl.edu/cgi-

bin/Codex/Codex.cgi). DNA forms of these shRNAs were generated by PCR

amplification of 97-base synthetic oligos using Pfu polymerase (Invitrogen) with a

common set of primers (F,

5’CAGAGGCTCGAGAAGGTATATTGCTGTTGACAGTGAGCG3’ and R,

5’CGCGGCGAATTCCGAGGCAGTAGGCA3’). PCR products were subsequently

digested by XhoI and EcoRI and inserted into RdpImiR30 between XhoI and MfeI

sites.

Tissue Culture and Cell Strain Establishment

CEFs, Ski-CEFs and derivatives were cultured in growth medium consisting of

Dulbecco’s modified Eagle’s medium (DMEM, Invitrogen) with 10% fetal bovine

serum (FBS, Atlanta Biological) and 50 U/ml gentamycin. To generate cell strains

stably expressing Smad2E2 and Ski mutants, RCASSmad2E2 and RdpISki were

cotransfected into CEFs using FuGENE6 (Roche Applied Science) according to

the manufacturer’s instructions. Puromycin (Sigma, 4 µg/ml) selection was started 3 days after transfection. The expression of Smad2E2 and Ski was

assessed by Western blot and immunostaining. To generate cell strains stably

expressing Smad3 with or without Ski, Smad3 expressed in a retroviral vector,

133

pOG1 (Obtained from Dr. Hartmut Beug) and RCASSki were cotransfected into

CEFs and selected by hygromycin (75 µg/ml) three days after transfection.

RdpImiR30shRNA was cotransfected with RCAS-Ski by Fugene6 (Roche

Applied Science) into 60% confluent CEFs. 72 hours after transfection, cells were

selected with puromycin (4 μg/ml). Survivors were cloned in 50% conditioned

medium and colonies were screened by Western blotting for the most efficient

knock down of Smad2 or Smad3.

Anchorage independent growth assay

Cells (2 x 104) were plated in triplicate into 35mm wells of 6-well dishes as

suspensions in 0.35% agar in HEPES buffered 199 medium (10% FBS, 1%

chicken serum, 25mg/ml glutamate, 0.8% sodium bicarbonate, 50ug/ml

gentamycin). These cells were fed every week with 2 ml of the same medium in

0.7% agar. The number of colonies formed in agar was counted after three

weeks.

Western Blotting

Cell lysates were prepared in lysis buffer (50mM Tris, pH8.0, 100mM NaCl, 1mM

EDTA, 10% glycerol and 0.2% Nonidet P-40) with protease inhibitor cocktail

(Roch Applied Science). Proteins were separated by 6% SDS-PAGE and

transferred to Immobilon-P membrane (0.45uM Millipore). Blots were preblocked

for 1h with 5% milk-TBST (TBST+10% (w/v) nonfat dry milk) for all antibodies

134

used for this study except for anti-Ski and anti-β-actin which were pre-blocked by overnight treatment with casein based blocking buffer (PBS (7.7mM Na2HPO4,

2.7mM NaH2PO4 and 150mM NaCl), 0.4% casein, 1% PVP (40,000), 10mM

EDTA and 0.2% Tween20, pH 7.2). After blocking, blots were incubated

overnight at 4 °C with primary antibodies in 5% milk-TBST or in blocking buffer for Ski and β-actin antibodies. Membranes were then washed with TBST and incubated for 1h at room temperature (RT) with alkaline phosphatase conjugated

secondary antibody. After washing with TBST, blots were incubated with CDP-

Star® alkaline phosphatase substrate (Tropix), washed again and exposed to

HyBlotCL autoradiography film (Denville Scientific) for signal detection.

Immunostaining

Cells were cultured in 12-well-plates and fixed with 3.7% paraformaldehyde at 60%

confluence. After three washes with PBSTX (PBS+0.2% TtritonX-100), cells were

blocked with 2% goat serum in PBS TX at RT for 1h and incubated with a

primary antibody overnight at 4 °C. Plates were washed 6 times with PBSTX

before 1h R.T incubation with biotinylated secondary antibody. Signal was

detected by Vectastain ABC kit (VectorLab). Images were obtained with Zeiss

inverted microscope equipped with a Spotlight digital camera.

135

Antibodies

The antibodies used for Western blot and immunostaining were as follows: anti-

Ski: (our G8 mouse monoclonal antibody (mAb), produced by the Lerner

Research Institute Hybridoma Core Facility); anti-Smad2: (rabbit anti-Smad2,

Zymed); anti-Smad2/3 (N-19, mouse mAb, ,Santa Cruz); rabbit anti-

phosphoSmad2, (Cell Signaling); rabbit anti-Smad3: (Zymed); rabbit anti-TBP:

(N-12, Santa Cruz); anti-β-actin: (mouse mAb, Sigma); and anti-Flag: (M2, mouse mAb, Sigma). Secondary antibodies used are AP conjugated anti-mouse

IgG (Fc specific) or anti-rabbit IgG (Sigma).

Results

TGF-β independent nuclear localization and altered expression of Smads in Ski-CEFs

We had previously shown that Ski and Smad proteins could be co-

immunoprecipitated from extracts of Ski-transformed fibroblasts [40]. Those

experiments were performed in the absence of added TGF-β where the rSmads

would be expected to reside predominantly in the cytoplasm and we did not

directly determine the intracellular location of the protein complexes. We have

therefore re-analyzed the intracellular location of Ski and Smad2/3 in normal and

ski-transformed fibroblasts in the presence or absence of TGF-β (Fig. 9a). As

expected, we found that in normal fibroblasts the rSmad proteins were almost

exclusively cytoplasmical in the absence of TGF-β and partly translocated to the

136

nucleus upon TGF-β treatment (Fig. 9a, top panel). The picture was quite different in Ski transformed fibroblasts. In these cells, the rSmad proteins were

largely nuclear in both the absence and presence of TGF-β (Fig. 9a, middle panel). Because TGF-β treatment was done in low serum and our cells are normally grown in 10% serum, we asked whether serum alone might be responsible for the difference between rSmad localization in normal and transformed cells. This was not the case since Smad2/3 localization was found to be the same in the presence of 10% serum (Fig. 9a, lanes 5 & 6) as it was in 0.2 % serum (Fig. 9a, lanes 1 & 2). The results also show that in these cells, Ski was located in the nucleus under all conditions (Fig. 9a, bottom panel).

137

Fig. 9a

Fig. 9a Nuclear localization and expression pattern of Smad2/Smad3 in Ski-

CEFs. Nuclear extracts from CEFs and Ski-CEFs treated with or without TGF-β

were assayed by immunoblot. The serum concentration in the medium was either

10% (+) or 2 % (-). Smad2 and Smad3 were detected by a single primary

antibody, N-19, monoclonal from Santa Cruz. Nuc: Nuclear; Cyto: cytoplasm

138

These results were confirmed by immunostaining. We found that in CEFs Smads

were present in both the cytoplasm and nucleus in the absence of TGF-β. Upon addition of TGF-β to these cells, rSmads localized to the nucleus (Fig. 9b, upper panels). When Ski was overexpressed in CEFs, Smads were located in the nucleus in both the absence and presence of TGF-β (Fig. 9b, lower panels).

TGF-β-independent accumulation of Smads in the nucleus of CEFs upon overexpression of Ski contrasts with findings by others and by us in mammalian cells where Smad nuclear localization and association with Ski were shown to be

TGF-β dependent [4, 18, 25, 27, 28, 40, 41, 61, 197-201]. A possible explanation

of these differences is that the mammalian cells were established cell lines

whereas the CEFs were primary cultures. To test this possibility, we asked

whether Ski caused TGF-β independent nuclear localization of Smads in an

established chicken fibroblast cell line, UMN-SAH/DF#1 (DF1). The results

showed that Ski overexpression had the same effect on Smad localization in

these cells as in primary CEFs (data not shown), suggesting that this response,

like transformation, is different in avian and mammalian cells.

139

Fig. 9b

Fig. 9b Immunofluorescence shows that Smads nuclear localization in CEFs is

TGF-β independent in the presence of overexpressed Ski. Cells were treated with anti-Smad2/3 or non-immune IgG (control).

140

Published data has demonstrated that protein turnover plays an important role in the TGF-β signaling pathway. Treatment of mammalian cells with TGF-β was shown to target Ski, SnoN and associated Smad proteins for degradation by a proteasome-mediated mechanism [27, 198, 202]. Other reports demonstrated down regulation of Smad3 expression in primary human fibroblasts and lung epithelial cells upon TGF-β treatment [28, 41, 203-208]. It therefore seemed possible that differences in the stability of Ski and Smad proteins between CEFs and mammalian cells could underlie the observed differences in their expression and accumulation in the nucleus. We therefore examined the steady state level of Ski and Smad2/3 proteins in CEFs and Ski-CEFs at various times of TGF-β treatment (Fig. 9c). We found that the levels of both endogenous c-Ski (Fig. 9c, lanes 1-6) and retrovirally expressed c-Ski (Fig. 9c lanes 7-12) remained unchanged during 24 hours of TGF-β treatment. The same was true of Smad2 and Smad3 although the ratio of their expression was very different in the normal and Ski-transformed cells (Fig. 9c, middle panels). These results suggest that Ski may be more stable in TGF-β treated chicken cells than it is in mammalian cells but this difference does not explain the nuclear accumulation of Smads in Ski-

CEFs because both Ski and Smads are equally stable in the normal and transformed cells.

141

Fig. 9c

Fig. 9c Both endogenous (Ski and Smad2/3) and exogenous Ski are resistant to

TGF-β induced degradation in CEFs. A single Western blot was cut into strips which were separately probed with antibodies to the indicated proteins.

142

Ski blocks TGF-β inhibition of CEF proliferation

In mammalian epithelial and pro-B cell lines, overexpression of Ski was sufficient to abrogate the inhibition of proliferation caused by TGF-β treatment [18, 25, 27,

28, 40, 41, 61, 197-201]. It seemed possible that this response to Ski might also be different in CEFs and mammalian cells in light of the high constitutive levels of

Smad proteins in the nucleus of Ski-CEFs and their stability upon TGF-β treatment. We therefore examined the effect of TGF-β on cell proliferation in

CEFs and Ski-CEFs (Fig. 10a). We found that, after a two day lag, TGF-β treatment was highly inhibitory of CEF proliferation whereas Ski-CEFs continued to proliferate at about the same rate as their untreated counterparts. Thus Ski’s ability to overcome inhibition of cellular proliferation by TGF-β is similar in CEFs and mammalian cells.

143

Fig. 10a

Fig. 10a TGF-β induced growth inhibition is reversed in Ski-CEFs. Cells were

plated in triplicate in 6-well dishes in the presence or absence of 2 ng/ ml of TGF-

β. Two wells of cells were sacrificed for cell number determination by counting in a hemocytometer every two days.

144

Smad2 is upregulated and receptor-phosphoryated in the absence of added TGF-β in Ski-CEFs

Our results also indicated that the relative levels of the two Smad proteins detected in Fig. 9a and 9c, were very different in CEFs and Ski-CEFs. The antibody used in these blots recognizes both Smad2 and Smad3 but based on their relative sizes. It appeared that normal cells expressed more Smad3 than

Smad2 while the ratio was reversed in the ski transformed cells. We have examined this change in expression further because of its possible importance in

Ski transformation. With the Smad2/3 antibody we again observed the altered expression ratio of the two proteins in the transformed cells compared to the nontransformed cells expressing a Ski mutant, SkiΔ3 (Fig. 10b i). Use of a specific antibody confirmed that the major species expressed in the transformed cells was Smad2 (Fig. 10b ii). The change in the ratio of Smad proteins upon Ski transformation appears to be due to a change in the mRNA levels because, using real-time RT-PCR, we have found that the relative accumulation of Smad2 and Smad3 mRNAs mirrors that of the two proteins (data not shown).

There are two forms of Smad2 that differ as a result of alternative splicing of exon 3 [209]. We have consistently detected Smad2 as a tightly spaced doublet with a major upper band and a minor faster migrating species. To determine whether these were the products of the alternatively spliced Smad2 mRNAs, we used an antibody specific for the alternatively spliced third exon (Fig. 10b iii). This antibody detected only the larger molecular weight form indicating that the majority of the Smad2 protein expressed is the non-DNA binding isoform of

145

Smad2 [210]. As shown in Figs.8a and 1b, most of the Smad2 protein in Ski- transformed cells was found in the nucleus even though those cells had not been treated with TGF-β. It therefore appeared possible that, in the absence of receptor activation by TGF-β, nuclear translocation of Smad2 might be independent of receptor phosphorylation. To determine whether this was the case, we used an antibody that specifically recognizes phosphoserines 465 and

467 of Smad2, which are produced by the activated TGF-β receptor (Fig. 10b iv).

We found that both forms of Smad2 were apparently receptor-phosphorylated in the ski-transformed cells but not in the normal cells.

146

Fig. 10b

Fig. 10b Smad2 is highly expressed and phosphorylated in cells overexpressing wild-type Ski. Whole cells extracts from cells overexpressing chicken c-Ski and a transformation defective mutant (c-SkiΔ3) were examined using antibodies to proteins indicated in the left of each panel.

147

Smad2 phosphorylation in Ski-CEFs is not blocked by neutralizing antibodies to TGF-β

In the absence of added TGF-β, it seemed possible that autocrine activation of the TGF-β receptor was responsible for Smad2 phosphorylation. In fact, we have

found that CEFs produce low levels of TGF-β (data not shown). To determine

whether this might be the case, we attempted to block phosphorylation of Smad2

by intercepting possible autocrine signaling using a neutralizing antibody that

reacts with all three TGF-β isoforms. To assess the effectiveness of the antibody,

we first examined Smad2 phosphorylation in CEFs in response to different doses

of TGF-β and then assessed the effectiveness of the antibody to block the

response. We found that Smad2 phosphorylation was completely dependent on

TGF-β addition in these cells and reached a maximum at the concentration of 2

ng/ml (Fig. 11a). We next asked whether the neutralizing antibody could block

Smad2 phosphorylation in response to this dose of TGF-β (Fig. 11b). The results demonstrated that addition of the TGF-β antibody to the culture medium completely blocked Smad2 phosphorylation whereas the anti-Myc antibody control had no effect (Fig. 11b, lanes 3 & 5). In fact, the inhibition was as great as that observed in the positive control in which the TGF-β was preincubated with the neutralizing antibody before adding it to the cells (Fig. 11b, lane 4). Having

established the effectiveness of the neutralizing antibody, we asked whether it

could block Smad2 phosphorylation in ski-transformed cells (Fig. 11c). In contrast to the results obtained with normal cells, we found that the neutralizing antibody did not reduce Smad2 phosphorylation, even at 24 hours after addition to the transformed cells.

148

Fig. 11

149

Fig. 11 Smad2 phosphorylation in Ski-CEFs is not TGF-β dependent. (a) Dose response of Smad2 phosphorylation by TGF-β in CEFs. (b) anti-TGF-β antibody is capable of blocking the Smad2 phosphorylation induced by 2 ng/ml of TGF-β in CEFs. Anti-Myc antibody served as a negative control. (C) anti-TGF-β does not block Smad2 phosphorylation in Ski-CEFs. TBP was used as loading control.

150

Smad2 phosphorylation in Ski-CEFs is not inhibited by blocking signaling through the activin or MAP kinase pathways

Smad2 phosphorylation and nuclear localization are also elicited by another

TGF-β family member, activin. It was therefore possible that Ski-activated autocrine signaling through the activin receptor was responsible for nuclear

accumulation of phospho-Smad2. To test this possibility, we asked whether an

activin-neutralizing antibody could block Smad2 phosphorylation in Ski-CEFs.

We found that the antibody completely blocked activin-stimulated

phosphorylation of Smad2 in normal CEFs (Fig. 12a lanes 1 & 2) but had no

effect in Ski-transformed CEFs (Fig. 12a, lanes 3 & 4). Because Smad activation

has also been elicited by activation of the MAPK pathway, we tested whether

inhibition of this pathway might block Smad2 activation. We used the well-studied

MAPK inhibitor PD98059 and found that at a concentration that drastically inhibited phosphorylation of ERK1/2, Smad2 phosphorylation was unaffected

(Fig. 12b, lane 3).

151

Fig. 12

Fig. 12 Smad2 phosphorylation in Ski-CEFs is neither activin nor ERK1/2 dependent. Anti-activin effectively blocked Smad2 phosphorylation in CEFs (a) but had no effect on Smad2 phosphorylation in Ski-CEFs. (b) Erk1/2 inhibitor,

PD98059 did not block Smad2 phosphorylation in Ski-CEFs.

152

Increased expression and nuclear localization of Smad2 correlate with transformation by Ski

Overexpression of Ski in CEFs leads to cellular transformation and is

accompanied by increased expression, phosphorylation and nuclear

accumulation of Smad2. To determine whether the effects on Smad2 correlate

with transformation by Ski, we asked whether they were observed in cells

expressing previously described transformation-defective ski mutants. We chose two mutants that eliminate Ski/Smad4 interaction, Δ3 (d230-324) and MT3

(C229S), and one that allows Smad4 interaction but is defective in transcriptional repression, ΔH4 (d190-204) [5, 12, 14, 111, 211] (Fig. 13a). We found that the intracellular location of Smads in cells expressing the three mutant forms of Ski was identical to that of normal CEFs. These cells express both Smad2 and

Smad3 and virtually all of the Smad proteins in these cells were cytoplasmical in the absence of TGF-β but translocated to the nucleus in its presence (Fig. 13b lanes 1, 2 & 5-10). As noted above, a large fraction of Smad2 in cells overexpressing the wild-type Ski was nuclear in both the absence and presence of TGF-β (Fig. 13b, lanes 3 & 4). Interestingly, the effect of the mutants on intracellular location was not dependent on their effects on the relative expression of the two Smad proteins. The Δ3 mutant did not alter expression from that seen in the normal CEFs, while the other two (MT3 and ΔH4) caused a mild increase in the Smad2/Smad3 ratio (Fig. 13b lanes 6, 8 & 10). Real-time

RT-PCR analysis of Smad2 and Smad3 mRNAs gave essentially the same relative expression as that of the Smad proteins indicating that the differences seen are largely due to regulation of mRNA abundance (data not shown). Our

153 results established a strong correlation between Ski’s ability to cause increased expression and nuclear accumulation of Smad2 and its ability to transform CEFs.

This correlation holds for v-Ski as well (data not shown).

Fig. 13a

Fig. 13a Schematic drawing of the Ski protein structure with positions of transformation defective mutations. ΔAH4, deletion of (190-204); Δ3, deletion of

(230-324): MT3, a point mutation C229S.

154

Fig. 13b

Fig. 13b Nuclear and cytoplasm distribution of Smad 2 and Smad3 in the absence (Ctrl) or presence of TGF-β in CEFs expressing transformation defective mutants was examined by immunoblot.

155

Re-establishing high Smad3 expression does not affect Smad2 localization or cellular transformation

To determine whether the changes in Smad expression and nuclear localization

of Smad2 are essential for transformation, we first asked whether re-establishing

the normal Smad2/Smad3 ratio by overexpressing Smad3 affected Smad2

localization and cellular transformation. We found that overexpression of Smad3

(Fig. 13c lanes 1, 2, 7 & 8 and Fig. 13d) had no obvious effect on the morphology

or growth of normal CEFs and it did not reverse transformation in Ski-CEFs. A

significant fraction of the overexpressed Smad3 was nuclear in both CEFs and

Ski-CEFs (Fig. 13c, lanes 1 & 7) but this did not affect the expression or nuclear localization of Smad2 (Fig. 13c, compare lanes 5 & 7). Thus it appears that the

reduction of Smad3 expression observed in Ski-transformed cells is not required

for either transformation of CEFs or nuclear localization of Smad2.

156

Fig. 13c and d

157

Fig. 13c&d Altering the Smad2/Smad3 ratio in Ski-CEFs does not reverse Smad2 expression or localization. Exogenous Smad3 was introduced to CEFs by retrovirus (pOG1) with or without Ski expressing virus (RCASSki). (c) Samples of the indicated whole cellular lysates (WC) and nuclear fractions (N) were immunoblotted with anti-Ski and anti-Smad2/3. (d) Overexpressing of Smad3 in

CEFs did not results in morphology change of CEFs and reverse the transformation induced by Ski based on the morphology.

158

Ski induced transformation is dependent on Smad2

To provide direct evidence for the importance of Smad2 in transformation by Ski,

we used shRNA technology to knock down Smad2 expression. Four different

shRNAs targeting chicken Smad2 were introduced into CEFs and Ski-CEFs

using the retroviral vectors shown in Fig. 14a. Three out of four shRNAs against

Smad2 effectively knocked down endogenous Smad2 in both CEFs and Ski-

CEFs (Fig. 14b). The knock-down of Smad2 was clearly detectable in cloned Ski-

CEFs expressing these three shRNAs (sh-1, sh-2 and sh-4). We next analyzed these clones for morphological transformation and anchorage-independent growth. The results show that loss of Smad2 expression blocked the ability of Ski to induce either of these hallmarks of cellular transformation but had little effect on CEF morphology and anchorage independent growth (Fig. 14c & 14d i-ix).

These results suggest that Ski induced transformation is dependent on Smad2 protein. Moreover, the correlation between the extent of knockdown in individual clones and the capacity of Ski-CEFs to clone in soft agar indicates that transformation by Ski is dependent upon the amount of Smad2. The fact that Ski transformation was blocked by three different Smad2-shRNAs strongly argues that the results are not due to a possible off-target effect.

159

Fig. 14a and b

160

Fig. 14a & b Smad2 is required to induce transformation by Ski (a) Diagrams of retroviral vectors expressing shRNA targeting Smad2 and expressing Ski. RCAS-

Ski or RCAS plus the Smad2 shRNA retroviral vector were introduced into CEFs simultaneously by transfection. After passage of the cells to allow virus spread only the CEFs infected by both retroviral vectors survived puromycin selection and were cloned. (b) Immunoblot of clones with anti-Smad2 to assess knockdown by different Smad2 shRNAs (sh-1, 3, & 4). The immunoblot with anti-Ski showed that the appropriate clones were co-infected by RCASSki. Numbers represent the Smad2 shRNAs and lower case letters represent the clones for that specific shRNA.

161

Fig. 14c and d

162

Fig. 14c and d

Knock-down of Smad2 blocks cellular transformation by Ski. (c) 20,000 cells from each clone were assayed for growth in soft agar as described in Methods.

Cloning efficiency = % of cells forming macroscopic colonies in three weeks. (d).

Knocking down Smad2 blocks Ski induced morphological transformation. i: vector alone, ii: clone sh1, iii: clone sh3, iv: Ski alone, v: Ski+ sh3a, vi: Ski+sh3b, vii: Ski+sh1, viii: Ski+sh4a, ix: Ski+sh4b. Only panel iv (Ski alone) shows the transformed morphology.

163

Knocking down Smad3 does not affect Ski induced transformation

Although our data showing the decreased expression of Smad3 in Ski-CEFs rendered it unlikely, we asked whether knocking down Smad3 would also inhibit transformation by Ski. Using the same vectors as described for Smad2, two

Smad3 shRNAs were found to effectively knock down endogenous Smad3 (Fig.

15a). However, knocking down Smad3 neither reversed morphological transformation by Ski (Fig. 15c), nor affected soft agar cloning o are both R-

Smad2s that directly mediate TGF-β signaling from the cytoplasm into the nucleus, the lack of an effect of Smad3 knock-down underscored our conclusion that Smad2 is selectively involved in Ski induced transformation.

164

Fig. 15

165

Fig. 15 Smad3 is not required for Ski induced transformation

(a) Ski and Smad3 shRNA were introduced into CEFs by the same strategy used

in Fig. 14b. Resulting clones of CEFs and Ski-CEFs were examined for Smad3 knock-down by immunoblotting with anti-Smad3, polyclonal Ski was detected by

G8, monoclonal. Number represents the Smad3 shRNA type. (b) Anchorage independent growth assay of Smad3 knockdown clones. (c) Morphological examination of Smad3 knockdown clones

166

Constitutively active Smad2, does not complement transformation defective mutants of Ski

Several transformation defective mutants of Ski are deletions and missense

mutations that eliminate binding to Smad proteins [5]. As shown in Fig. 13a,

mutants of this type also fail to induce Smad2 expression, phosphorylation and

nuclear translocation. Since Ski induced transformation is dependent on Smad2,

we tested whether a constitutively active form of Smad2 would complement the

transformation defectiveness of such mutants. Phosphorylation of serines 465

and 467 by the TGF-β receptor is necessary to activate Smad2 and mutation of

these residues to glutamates in Samd2-E2 (Ser465/467E) has been shown to

produce a constitutively active protein [212]. Because the deletion and

substitution mutations studied previously, are likely cause structural alterations in

Ski, we chose a well characterized set of Ski missense mutants that eliminate

binding to individual Smad proteins [138]. Hayman and coworkers have shown that Ski contains two binding sites for Smads2 and 3 and that mutation of both sites in the L19A/L110P mutant eliminates Smad2/3 binding. They and others also showed that the W255E mutation eliminates binding to Smad4 [12]. The

triple mutant (L19A, L110P and W255E) does not bind any of the Smad proteins

and a previous study showed that an analogous mutant was completely

transformation defective [139].

We first determined the transforming activity of the mutants, individually and in

combination. In agreement with the earlier study, we found that only the triple

mutant of Ski was completely defective in transformation (Fig. 16a). The other

167

mutants exhibited partial transforming activity as measured by growth in soft agar.

To test whether Smad2-E2 could salvage the reduced transforming ability of

these Ski mutants, they were introduced into CEFs with Smad2-E2 by retroviral

infection. Immunostaining shows both the Ski mutant proteins and Smad2-E2

were stably overexpressed and localized in the nuclei of the cells (Fig. 16b).

However, Smad2-E2 failed to complement the transformation defectiveness of any of the mutants (Fig. 16a). We therefore conclude that direct interaction with

Ski, not merely overexpression and constitutive activation of Smad2 is necessary for transformation of CEFs.

168

Fig. 16a

Fig. 16a Smad2E2 did not complement did not alter the ability of a series of Ski mutants to promote anchorage-independent growth in soft agar. Triplicate samples of 2x 104 CEFs expressing the indicated proteins were assayed for

colony formation after three weeks of culture in soft agar. SkiW255E, does not bind

Smad4; SkiL110P disrupts one of the two Smad2 binding sites; SkiL19A+L110P does

not bind Smad2 binding; and SkiL19A+L110P+W255E does not bind either Smad protein.

169

Fig. 16b

170

Fig. 16b Immunostaining with anti-Ski G8 and anti-Smad2 shows the expression of the Ski mutants and Smad2E2 in CEFs expressing the proteins indicated on the left of each panel. Ski (A) = SkiL19A, Ski (P) = SkiL110P,Ski (E) = SkiW255E,

Ski(A,P)=Ski(L19A, L110P), Ski (A,P,E)=Ski(L19A,L110P, W255E)

171

Discussion

The present results demonstrate that Ski has several effects on the TGF-β

signaling pathway in primary chicken embryo fibroblasts. One of the most

dramatic effects is the induction of Smad2 and repression of Smad3. In normal

CEFs, Smad3 is expressed at a higher level than is Smad2 but in Ski-CEFs this

ratio is reversed with respect to both the proteins and their mRNAs. Smad2 and

Smad3 are closely related proteins and are both activated by phosphorylation by

the TGF-β receptor in response to ligand binding. However, a large body of work

has shown that they have clearly distinct activities in vitro and in vivo [15, 125,

127, 213-217]. Therefore the change in the expression ratio of these rSmads that

we detected upon Ski transformation of CEFs would be expected to have

profound effects on the response of these cells to signaling by TGF-β.

Because Ski can function as a co-repressor with either Smad2 or Smad3, we had

expected that the down regulation of Smad3 expression would be an important

determinant of Ski’s activity. For example, this reduction in Smad3 expression might underlie the resistance of Ski-CEFs to the cytostatic effect of TGF-β which has been shown to be largely mediated by activation of Smad3 [120, 121].

Reduced expression of Smad3 may turn out to be extremely important in the context of Ski’s ability to induce muscle differentiation in avian fibroblast because

Smad3 has been shown to mediate inhibition of muscle differentiation by TGF-β

and Smad2 was not able to perform this role [15]. However, it does not appear to

172

play an important role in transformation because re-establishing the normal

Smad3 to Smad2 expression ratio in Ski-CEFs did not affect the transformed phenotype, although it did result in constitutive nuclear accumulation of both

Smad proteins. Moreover, knocking down Smad3 did not block the transformation of CEFs by Ski.

It is clear from previous studies that the ability of Ski to interact with Smad proteins correlates with its transforming ability [14, 218]. However the mutations in Ski that established that correlation were in conserved domains that are known to mediate interactions with several other transformation-related proteins. It was therefore not clear that interaction with the Smad proteins was the key to Ski’s transforming ability. Our present data showing that knocking down endogenous

Smad2 expression completely blocked Ski induced transformation of CEFs provides direct evidence that Smad2 is indispensable for Ski’s transforming process. But increased expression and nuclear localization of Smad2 are not sufficient for transformation because overexpression of constitutively active

Smad2 did not transform CEFs or complement the transformation defectiveness of Ski mutants that fail to bind Smad proteins. Those results suggest the requirement of not only Smad2 expression and activation but also its direct interaction with overexpressed Ski in the transformation of CEFs. Interestingly, although Ski stabilizes the nuclear phosphorylated form of Smad2 in CEFs, it has been shown to block phosphorylation of Smad2 in mammalian cells [219]. This

173

difference may not only underlie Ski’s mechanism of transformation of avian cells

but also its inability to transform mammalian cells.

It is likely that the keys to Ski’s transforming ability are among the specific

targets of Smad2 regulation. These targets will certainly be governed by the

observation that the major form of Smad2 expressed in CEFs and Ski-CEFS

(and most other cells) contains the segment encoded by the alternatively spliced third exon which disrupts the DNA binding domain and renders the protein incapable of binding the Smad binding element (SBE) [209, 220, 221]. This form of Smad2 functions as a co-activator; relying on the DNA binding activity of

Smad4 or other associated transcription factors, such as Fast-1, to determine the targets of its action [222-224]. This feature certainly limits the targets of

Ski/Smad2 transcriptional regulation relative to those of Smad3, which has an intact DNA binding domain that recognizes the same DNA elements as Smad4

[220]. We had previously shown that Ski-CEFs contained a multi-protein complex, subsequently identified as Ski, Smad2 and Smad4, that specifically bound the

GTCTAGAC form of the SBE but with greatly reduced affinity to half sites of the

SBE [14]. Thus, we concluded that binding involved two co-operating Ski/Smad complexes and Smad4 alone was responsible for the binding of the SBE. Those results support our conclusion that the likely targets of Ski’s action as a Smad co- repressor in these cells to genes with regulatory regions containing complete

SBEs or elements with binding sites for Smad-interacting transcription factors

174

such as Fast-1 and c-Jun [37, 222, 225]. Therefore, we suggest that Ski

transforms CEFs, by subverting the Smad2 arm of the TGF-β pathway. By this action Ski/Smad2 represses genes whose normal role in these cells is to suppress transformation/cell growth through the action of other signaling pathways that dominate in the absence of TGF-β.

175

Discussion and Future Directions

Ski-Smad2 Function

In this study, we discovered that Ski selectively upregulates Smad2 and down

regulates Smad3 in Ski-CEFs. In these cells, Smad2 is phosphorylated as the

active form (p-Smad2) and the activation of Smad2 is independent of exogenous

TGF-β signaling. Although both Smad2 and Smad3 respond to TGF-β via

receptor phosphorylation, the cytostatic response of TGF-β signaling relies on

Smad3 instead of Smad2 [121]. These data suggest that Ski blocks TGF-β

signaling by changing the signaling strength and shifting the signal transmission

from Smad3 to Smad2 in addition to negatively regulating the transcriptional

activity of Smad complexes as a transcription corepressor. Interestingly, Smad2

has been shown to be a more potent tumor suppressor than Smad3 in the

squamous cell cancer carcinogenesis [130, 131]. Overexpressing Ski transforms

CEFs and injection of Ski-CEFs caused squamous cell carcinomas in chicken

[52]. In vivo, Ski -/- mice display severe lip clefting. Overexpressing p-Smad2 rescues cleft palate in TGF-β 3 null mice [226]. Although cleft lip is different from cleft palate, they are both types of craniofacial clefting. About 15% of the patients with 1p36 syndrome in which SKI is deleted, exhibit facial or palatal clefting [227].

These data suggest a Smad2-Ski connection during craniofacial development.

Taken together, these findings set up a strong Ski-Smad2 connection rather than

Ski-Smad3 and also suggest that 1) Ski, as an oncoprotein overcomes the tumor suppressing ability of Smad2; 2) Ski-p-Smad2 interaction endows Ski negative regulation of a Smad2-activated subset of TGF-β responsive genes. 176

Gene expression profiling studies identified groups of genes whose expression

are either dependent on Smad2 regulation or dependent on Smad3 regulation in

response to TGF-β, [119, 120]. Comparing the Affymetrix gene expression

profile of Ski-CEF and CEF, we found that Ski down regulates connective tissue

growth factor (CTGF), an immediate-early TGF-β responsive gene and a Smad3

dependent gene [228] by at least 3 fold. Ski also down regulates

metalloproteinase-2, a Smad2 dependent gene [118] by more than1.5 fold

(Affymetrix data). The expression data of these two genes in Ski-CEFs, not only

confirm that Ski inhibits the canonical TGF-β signaling pathway by lowering the

Smad3 amount and likely still acting as a Smad3 co-repressor but also provide

evidence that Ski alters the normal transcription control of Smad2 containing

complexes on Smad2 dependent genes even in the presence of excess of p-

Smad2.

Previous results from our lab and others on several Ski mutants with mutations in

DHD and SAND domains show that the cellular transforming capacity of Ski is

tightly correlated to Ski’s ability to repress Smad activity [5, 139]. These mutants

are ΔAH4 (with a deletion of the fourth α-helix in the DHD), Δz3/4 (with a deletion

preceding the zinc finger motif in the SAND domain), MT3 (with a Cys to Ser

point mutation in the SAND domain that causes the loss of the Smad4 binding)

and mS4HE, (with a double point mutations in the SAND domain that causes the

loss off the Smad4 binding). In general, point mutations results in the minimal conformational changes to proteins whereas deletion a region almost certainly changes the conformation. We also found that another point mutant of Ski,

177

W255E, does not bind to Smad4, but still partially transforms CEFs. These MT3

and W255E have lost the same function but cause different biological outcomes,

indicating that even a point mutation might affect Ski’s interaction with proteins

other than Smad4. In fact, to the best knowledge, Ski interacts with more than 13

proteins via these two domains. It is arguable that these protein–protein

interaction sites in the DHD and the SAND domain might overlap with each other.

Mutations of Smads binding sites might affect Ski binding to other proteins. To

determine directly whether Ski-Smads interaction is required for Ski induced cellular transformation, we knocked down endogenous Smad2 with RNAi and proved that Smad2 is essential for Ski-induced transformation of CEFs. Based on above results, we hypothesize that when transforming CEFs, Ski does not simply inhibit TGF-β signaling but empowers new functions though the Ski-pSmad2 complex to reverse the regulation of Smad2 on genes that normally act as secondary regulators or effectors of tumor suppressing pathways in the absence of TGF-β stimulation. However, we have not identified the downstream targets of

Ski-pSmad2 complex that are involved in transformation. One ideal way to discover these downstream targets is to use the ChIP-on ChIP technique to identify promoters occupied by Ski and Smad2 during Ski induced transformation.

Besides the distinct group of target genes regulated by Smad2 and Smad3, both

Smads have distinct groups of interacting partners. These results were obtained by liquid chromatography–mass spectrometry analysis of affinity purified tagged

Smad2 or Smad3 containing protein complexes [122]. In addition to the known

178

binding partners, 13 proteins were discovered to associate with Smad2 uniquely

and 20 proteins were identified as Smad3 unique partners. 13 proteins were

shown as common partners for both Smad2 and Smad3. This discovery

suggests that Ski-pSmad2 targets could be identified by studying the relationship

between Ski and these Smad2 partners and expands the knowledge of the

molecular biology of Ski. Interestingly, two of them, ANT2/3 (two isoforms cannot

be resolved by peptide analysis) and MFN2 are both mitochondrial proteins.

ANT2/3 is an ADP/ATP transporter located in the mitochondrial inner membrane

and has been reported to contribute to the cold induced mitochondrial uncoupling

in chicken skeletal muscle [229]. MFN2 is a mitochondrial outer membrane

protein that regulates mitochondrial fusion [230]. In this thesis, we showed that

Ski increases mitochondrial number, size, content and respiration in CEFs.

Consistently, these two proteins are elevated in Ski-CEFs suggesting a role for

Smad2 linked to mitochondrial activity.

Ski transformed CEFs show typical features of oncogenic transformation. Ski-

CEFs are apoptosis-resistant, morphologically different from CEFs and capable of forming colonies in soft agar. However, unlike most other transformed cells,

Ski-CEFs do not display the classical Warburg effect: enhanced glycolysis in the presence of adequate oxygen. On the contrary, Ski suppresses glucose utilization and glycolysis but stimulated oxygen consumption. Previous studies revealed that TGF-β stimulated glycolysis and even repressed oxygen consumption in rat kidney fibroblasts, rat renal proximal tubular cells and Hek293

179

cells [231-233]. The opposite effect of TGF-β and Ski on glycolysis and oxygen consumption suggested that the alteration of energy metabolism in Ski-CEFs might be linked to the inhibition of TGF-β signaling. Moreover, CEFs, expressing a transformation defective mutant, Δz3/4 do not suppress glycolysis and lactate

production. Δz3/4-CEFs also do not increase the NADH, FADH coupled enzyme

activities assayed by Alamar blue reduction. We show that Ski-CEFs stimulate

fatty acid oxidation and mitochondrial respiration though PPARγ mediated

transcription activation. Microarray and Q-RTPCR analysis revealed that Δz3/4

did not induce PPARγ as well as lipid metabolism related PPARγ target genes.

This suggests that in Ski-CEFs, the PPARγ mediated energy metabolic shift is

seemly linked to the Ski induced transformation. However, knocking down

PPARγ in Ski-CEFs, reversed the energy metabolic shift as well as the induction

of mitochondrial biogenesis but these cells still kept their transformed cell

phenotype judged by the morphological change. Therefore, we conclude that the

energy metabolic shift is not correlated to the Ski induced transformation. On the

other hand, Δz3/4 mutant also fails to change the Smad2/Smad3 ratio, up-

regulate the p-Smad2 level and inhibit TGF-β signaling, suggesting that the

inhibition of TGF-β signaling might be involved in the metabolic shift via PPARγ.

Recently, PPARγ was reported to repress the Smad2/3 mediated induction of

collagen1 A2 (COL1A2) by preventing the p300 recruitment and histone H4

hyperacetylation [234]. Interestingly, we found that Ski is a binding partner of

PPARγ and our expression array data showed that COLA2 was down-regulated

in Ski-CEFs. These results suggest not only the possible involvement of PPARγ

180

in Ski directed TGF-β signaling but also that in some contexts, Ski might function

as a PPARγ co-repressor. Taken together, we postulate that the up-regulation of

PPARγ is downstream of the Ski-pSmad2 interaction.

PPARγ and Ski Interaction

Our data also show that PPARγ is up-regulated and activated in Ski-CEFs

PPARγ is known as a master regulator of adipogensis and lipogenesis (175). Our

results apparently contradict to these known PPARγ activity because Ski-CEFs

produce fewer lipids and do not store fatty acids. A plausible explanation is that

different tissues have different responses to the activation of the PPARγ pathway.

In both WAT and BAT, PPARγ strongly induces target genes functioning in lipid

synthesis, transport and storage as well as fatty acid oxidation upon ligand

activation. However, the same agonist of PPARγ did not induce the expression of

the same set of target genes in skeletal muscle [113]. In the muscle of Ski

transgenic mice, PPARγ expression is lower than in wild type mouse muscle.

Although we cannot classify Ski-CEFs with one of these tissue types, Ski-CEFs are distant from adipocytes, since a series of typical PPARγ induced adipocyte markers is not co-upregulated with adipogenic markers by Ski in Ski-CEFs.

Furthermore, Wnt signaling is activated in Ski-CEFs (my microarray and reporter assay data), while it was reported that PPARγ represses Wnt/β-catenin signaling during adipogenesis by targeting β-catenin to degradation [235]. Thus the

PPARγ-dependent pathways affected in Ski-CEFs appear to mixture of its activities in different mammalian tissues. 181

Previously, Ski was connected to other nuclear hormone receptor family

members, RAR and vitamin D receptor (VDR). Ski binds to RAR and recruits the

corepressor, N-CoR to repress the RAR signaling when transforming

hematopoietic cells. Ski also directly binds to VDR and recruits N-CoR to negatively regulate VDR mediated transcription [36]. Interestingly, we showed that by co-immnuoprecipitation, Ski not only physically interacts with PPARγ-

RXR heterodimer but transactivates transcription from PPRE containing reporters.

Still unanswered is that whether Ski is able to recruit transcriptional co-

repressors that negatively regulate transcription of PPARγ target genes involved

in lipogenesis while recruiting transcriptional co-activators to PPARγ-RXR

complexes that upregulate genes involved in lipid transport and oxidation. Genes

upregulated by PPARγ in Ski-CEFs such as FABP4, LPL, ACOX1 are good

candidates for this examination. These possibilities can be tested by serial

chromatin co-immnuoprecipitation (ChIP) or ChIP-Western blot to examine

whether Ski is bound to promoters of PPARγ target genes with PPARγ in Ski-

CEFs or in CEFs in response to the ligand activation. Next, we could also combine immuno-purification techniques with mass spectrometry analysis to generate profiles of Ski-PPARγ interacting proteins in order to identify other co-

factors with activating or repressing functions in the Ski-PPARγ complex.

182

Glycolysis and PPARγ

Activation of PPARγ by ligands increases insulin sensitivity and improves glucose homeostasis [236]. However, in Ski-CEFs, we observed a decrease of glucose utilization and we have gene expression evidence that some of the glycolysis regulatory enzymes, such as phosphofructokinase (PFK), ADP dependent glucose-6-kinase (ADPGK) and (pyruvate dehydrogenase kinase)

PDK are down regulated in Ski-CEFs. Interestingly, microarray data show that the potent glycolysis regulator, HIF-1, is down regulated in Ski-CEFs. Knocking down PPARγ restores the glucose utilization in Ski-CEFs to the level observed in

CEFs. Little is known about direct connection between PPARγ and glycolytic regulation. We can test the possibility in Ski-CEFs that PPARγ directly affects regulatory enzymes of glycolysis by mRNA analysis, ChIP or indirectly though

HIF-1 activity regulation by reporter, immunoblot or ChIP. Since TGF-β stimulation enhances glycolysis, we can also test whether Smads are involved in this regulation.

Downstream mediators of Ski induced metabolic shift

Our results show that PPARγ mediates the Ski induced metabolic shift and mitochondrial biogenesis. However, PPARγ, as a master regulator, functions on the top of a transcriptional regulatory cascade. The importance of downstream target genes is not known from our current experimental results. It would be very interesting to identify the intermediate downstream regulators that mediate

PPARγ regulation on effectors in Ski-CEFs. For example, PGC1-α, a known

183

target gene induced by PPARγ is a good candidate, since it promotes mitochondrial biogenesis, fatty acid oxidation and slow type I muscle formation.

By knocking down PGC1-α in Ski-CEFs, we can ask whether Ski-CEFs are still capable of stimulating oxygen consumption, fatty acid oxidation, and mitochondrial biogenesis without disturbing the ability of importing massive amount of lipids into cells.

Energy production pathways

In this study, we demonstrated that Ski-CEFs stimulate fatty acid oxidation and suppress glucose utilization. However, although Ski-CEFs consume 1/5th the

amount of glucose as CEFs they metabolize four times as much glucose via the

TCA cycle and oxidative phosphorylation. Therefore, without accurately measuring the amount of ATP produced by glucose and fatty acid oxidation respectively in Ski-CEFs, we do not know which energy production pathway is the main pathway utilized by Ski-CEF with respect to energy production. To

answer this question, we need to determine the steady state pool size of ATP and the new ATP synthesis rate in Ski-CEFs. We can use 18O- phosphate to label the phosphate-phosphate exchange in order to monitor the new ATP synthesis [237]. We need to also measure the daily usage of isotope labeled glucose and palmitate in the medium by mass spectrometry and newly synthesized glucose and palmitate in cells using deuterium water labeled medium. By calculation of the net ATP production with the net consumption of

184 glucose and palmitate daily, we could figure out the net energy production generated from these specific sources.

185

References

1. Stavnezer, E., et al., Generation of transforming viruses in cultures of chicken fibroblasts infected with an avian leukosis virus. J Virol, 1981. 39(3): p. 920-34. 2. Stavnezer, E., D. Brodeur, and L.A. Brennan, The v-ski oncogene encodes a truncated set of c-ski coding exons with limited sequence and structural relatedness to v-myc. Mol Cell Biol, 1989. 9(9): p. 4038-45. 3. Colmenares, C., et al., Activation of the c-ski oncogene by overexpression. J Virol, 1991. 65(9): p. 4929-35. 4. Akiyoshi, S., et al., c-Ski acts as a transcriptional co-repressor in transforming growth factor-beta signaling through interaction with smads. J Biol Chem, 1999. 274(49): p. 35269-77. 5. Nicol, R., et al., Association of specific DNA binding and transcriptional repression with the transforming and myogenic activities of c-Ski. Cell Growth Differ, 1999. 10(4): p. 243-54. 6. Colmenares, C., J.K. Teumer, and E. Stavnezer, Transformation-defective v-ski induces MyoD and myogenin expression but not myotube formation. Mol Cell Biol, 1991. 11(2): p. 1167-70. 7. Wilson, J.J., et al., Crystal structure of the dachshund homology domain of human SKI. Structure, 2004. 12(5): p. 785-92. 8. Zheng, G., et al., High affinity dimerization by Ski involves parallel pairing of a novel bipartite alpha-helical domain. J Biol Chem, 1997. 272(50): p. 31855-64. 9. Chen, D., et al., SKI activates Wnt/beta-catenin signaling in human melanoma. Cancer Res, 2003. 63(20): p. 6626-34. 10. Zheng, G., et al., Identification of a core functional and structural domain of the v-Ski oncoprotein responsible for both transformation and myogenesis. Oncogene, 1997. 15(4): p. 459-71. 11. Arndt, S., et al., Cloning and functional characterization of a new Ski homolog, Fussel-18, specifically expressed in neuronal tissues. Lab Invest, 2005. 85(11): p. 1330-41. 12. Wu, J.W., et al., Structural mechanism of Smad4 recognition by the nuclear oncoprotein Ski: insights on Ski-mediated repression of TGF-beta signaling. Cell, 2002. 111(3): p. 357-67. 13. Kim, S.S., et al., Structure of the retinal determination protein Dachshund reveals a DNA binding motif. Structure, 2002. 10(6): p. 787-95. 14. Nicol, R., et al., Association of specific DNA binding and transcriptional repression with the transforming and myogenic activities of c-Ski [In Process Citation]. Cell Growth Differ, 1999. 10(4): p. 243-54. 15. Bottomley, M.J., et al., The SAND domain structure defines a novel DNA-binding fold in transcriptional regulation. Nat Struct Biol, 2001. 8(7): p. 626-33. 16. Suzuki, H., et al., c-Ski inhibits the TGF-beta signaling pathway through stabilization of inactive Smad complexes on Smad-binding elements. Oncogene, 2004. 23(29): p. 5068-76. 17. Kokura, K., et al., The Ski protein family is required for MeCP2-mediated transcriptional repression. J Biol Chem, 2001. 276(36): p. 34115-21.

186

18. Nomura, T., et al., Ski is a component of the histone deacetylase complex required for transcriptional repression by Mad and thyroid hormone receptor. Genes Dev, 1999. 13(4): p. 412-23. 19. Kokura, K., et al., The Ski-binding protein C184M negatively regulates tumor growth factor-beta signaling by sequestering the Smad proteins in the cytoplasm. J Biol Chem, 2003. 278(22): p. 20133-9. 20. Heyman, H.C. and E. Stavnezer, A carboxyl-terminal region of the ski oncoprotein mediates homodimerization as well as heterodimerization with the related protein SnoN. J Biol Chem, 1994. 269(43): p. 26996-7003. 21. Nagase, T., et al., Requirement of protein co-factor for the DNA-binding function of the human ski proto-oncogene product. Nucleic Acids Res, 1990. 18(2): p. 337- 43. 22. Tarapore, P., et al., DNA binding and transcriptional activation by the Ski oncoprotein mediated by interaction with NFI. Nucleic Acids Res, 1997. 25(19): p. 3895-903. 23. Dahl, R., et al., Transformation of hematopoietic cells by the Ski oncoprotein involves repression of retinoic acid receptor signaling. Proc Natl Acad Sci U S A, 1998. 95(19): p. 11187-92. 24. Dahl, R., B. Wani, and M.J. Hayman, The Ski oncoprotein interacts with Skip, the human homolog of Drosophila Bx42. Oncogene, 1998. 16(12): p. 1579-86. 25. Luo, K., et al., The Ski oncoprotein interacts with the Smad proteins to repress TGFbeta signaling. Genes Dev, 1999. 13(17): p. 2196-206. 26. Nomura, T., et al., Oncogenic activation of c-Myb correlates with a loss of negative regulation by TIF1beta and Ski. J Biol Chem, 2004. 279(16): p. 16715- 26. 27. Sun, Y., et al., Interaction of the Ski oncoprotein with Smad3 regulates TGF-beta signaling. Mol Cell, 1999. 4(4): p. 499-509. 28. Xu, J. and L. Attisano, Mutations in the tumor suppressors Smad2 and Smad4 inactivate transforming growth factor beta signaling by targeting Smads to the ubiquitin-proteasome pathway. Proc Natl Acad Sci U S A, 2000. 97(9): p. 4820-5. 29. Zhang, H. and E. Stavnezer, Ski regulates muscle terminal differentiation by transcriptional activation of Myog in a complex with Six1 and Eya3. J Biol Chem, 2009. 284(5): p. 2867-79. 30. Ueki, N., L. Zhang, and M.J. Hayman, Ski negatively regulates erythroid differentiation through its interaction with GATA1. Mol Cell Biol, 2004. 24(23): p. 10118-25. 31. Ueki, N., L. Zhang, and M.J. Hayman, Ski can negatively regulates macrophage differentiation through its interaction with PU.1. Oncogene, 2008. 27(3): p. 300-7. 32. Takeda, M., et al., Interaction with Smad4 is indispensable for suppression of BMP signaling by c-Ski. Mol Biol Cell, 2004. 15(3): p. 963-72. 33. Tokitou, F., et al., Viral ski inhibits retinoblastoma protein (Rb)-mediated transcriptional repression in a dominant negative fashion. J Biol Chem, 1999. 274(8): p. 4485-8. 34. Baldwin, A., L. Pirisi, and K.E. Creek, NFI-Ski interactions mediate transforming growth factor beta modulation of human papillomavirus type 16 early gene expression. J Virol, 2004. 78(8): p. 3953-64.

187

35. Takahata, M., et al., SKI and MEL1 cooperate to inhibit transforming growth factor-beta signal in gastric cancer cells. J Biol Chem, 2009. 284(5): p. 3334-44. 36. Ueki, N. and M.J. Hayman, Signal-dependent N-CoR requirement for repression by the Ski oncoprotein. J Biol Chem, 2003. 278(27): p. 24858-64. 37. Pessah, M., et al., c-Jun associates with the oncoprotein Ski and suppresses Smad2 transcriptional activity. J Biol Chem, 2002. 277(32): p. 29094-100. 38. Nagano, Y., et al., Arkadia induces degradation of SnoN and c-Ski to enhance transforming growth factor-beta signaling. J Biol Chem, 2007. 282(28): p. 20492-501. 39. Yuzawa, H., et al., Arkadia represses the expression of myoblast differentiation markers through degradation of Ski and the Ski-bound Smad complex in C2C12 myoblasts. Bone, 2009. 44(1): p. 53-60. 40. Xu, W., et al., Ski acts as a co-repressor with Smad2 and Smad3 to regulate the response to type beta transforming growth factor. Proc Natl Acad Sci U S A, 2000. 97(11): p. 5924-9. 41. Shinagawa, T., et al., The sno gene, which encodes a component of the histone deacetylase complex, acts as a tumor suppressor in mice. EMBO J, 2000. 19(10): p. 2280-91. 42. Massague, J., TGFbeta in Cancer. Cell, 2008. 134(2): p. 215-30. 43. Heider, T.R., et al., Ski promotes tumor growth through abrogation of transforming growth factor-beta signaling in pancreatic cancer. Ann Surg, 2007. 246(1): p. 61-8. 44. Wang, P., et al., Dual role of Ski in pancreatic cancer cells: Tumor-promoting versus metastasis-suppressive function. Carcinogenesis, 2009. 45. Wang, P., et al., Ski acts as therapeutic target of qingyihuaji formula in the treatment of SW1990 pancreatic cancer. Integr Cancer Ther. 9(1): p. 50-8. 46. Kiyono, K., et al., c-Ski overexpression promotes tumor growth and angiogenesis through inhibition of transforming growth factor-beta signaling in diffuse-type gastric carcinoma. Cancer Sci, 2009. 100(10): p. 1809-16. 47. Fumagalli, S., et al., Expression of the c-ski proto-oncogene in human melanoma cell lines. Melanoma Res, 1993. 3(1): p. 23-7. 48. Chen, D., et al., SKI knockdown inhibits human melanoma tumor growth in vivo. Pigment Cell Melanoma Res, 2009. 22(6): p. 761-72. 49. Sohn, B.H., et al., Functional switching of TGF-beta1 signaling in liver cancer via epigenetic modulation of a single CpG site in TTP promoter. Gastroenterology. 138(5): p. 1898-908. 50. Bravou, V., et al., TGF-beta repressors SnoN and Ski are implicated in human colorectal carcinogenesis. Cell Oncol, 2009. 31(1): p. 41-51. 51. Fukuchi, M., et al., Increased expression of c-Ski as a co-repressor in transforming growth factor-beta signaling correlates with progression of esophageal squamous cell carcinoma. Int J Cancer, 2004. 108(6): p. 818-24. 52. Stavnezer, E., The Ski Oncogene, in The Oncogene Handbook. 1988, Elsevier Science Publishers B. V. p. 393. 53. Lowenberg, B., J.R. Downing, and A. Burnett, Acute myeloid leukemia. N Engl J Med, 1999. 341(14): p. 1051-62.

188

54. Ritter, M., et al., Inhibition of retinoic acid receptor signaling by Ski in acute myeloid leukemia. Leukemia, 2006. 20(3): p. 437-43. 55. Pandolfi, P.P., In vivo analysis of the molecular genetics of acute promyelocytic leukemia. Oncogene, 2001. 20(40): p. 5726-35. 56. Larsen, J., et al., Transformation of chicken bone marrow cells by the v-ski oncogene. Oncogene, 1993. 8(12): p. 3221-8. 57. Colmenares, C., et al., Loss of the SKI proto-oncogene in individuals affected with 1p36 deletion syndrome is predicted by strain-dependent defects in Ski-/- mice. Nat Genet, 2002. 30(1): p. 106-9. 58. Keppler-Noreuil, K.M., et al., Chromosome 1p terminal deletion: report of new findings and confirmation of two characteristic phenotypes. J Med Genet, 1995. 32(8): p. 619-22. 59. Reish, O., S.A. Berry, and B. Hirsch, Partial monosomy of chromosome 1p36.3: characterization of the critical region and delineation of a syndrome. Am J Med Genet, 1995. 59(4): p. 467-75. 60. Shapira, S.K., et al., Chromosome 1p36 deletions: the clinical phenotype and molecular characterization of a common newly delineated syndrome. Am J Hum Genet, 1997. 61(3): p. 642-50. 61. Slavotinek, A., L.G. Shaffer, and S.K. Shapira, Monosomy 1p36. J Med Genet, 1999. 36(9): p. 657-63. 62. Berk, M., et al., Mice lacking the ski proto-oncogene have defects in neurulation, craniofacial, patterning, and skeletal muscle development. Genes Dev, 1997. 11(16): p. 2029-39. 63. Lana, D.P., et al., Selective expression of a ski transgene affects IIb fast muscles and skeletal structure. Dev Dyn, 1996. 205(1): p. 13-23. 64. Sutrave, P., A.M. Kelly, and S.H. Hughes, ski can cause selective growth of skeletal muscle in transgenic mice. Genes Dev, 1990. 4(9): p. 1462-72. 65. Bruusgaard, J.C., et al., Muscle hypertrophy induced by the Ski protein: cyto- architecture and ultrastructure. Acta Physiol Scand, 2005. 185(2): p. 141-9. 66. Leong, G.M., et al., The Ski proto-oncogene regulates body composition and suppresses lipogenesis. Int J Obes (Lond). 34(3): p. 524-36. 67. Goodyear, L.J. and B.B. Kahn, Exercise, glucose transport, and insulin sensitivity. Annu Rev Med, 1998. 49: p. 235-61. 68. Hansen, P.A., et al., Increased GLUT-4 translocation mediates enhanced insulin sensitivity of muscle glucose transport after exercise. J Appl Physiol, 1998. 85(4): p. 1218-22. 69. Helge, J.W. and F. Dela, Effect of training on muscle triacylglycerol and structural lipids: a relation to insulin sensitivity? Diabetes, 2003. 52(8): p. 1881-7. 70. Phillips, S.M., et al., Progressive effect of endurance training on metabolic adaptations in working skeletal muscle. Am J Physiol, 1996. 270(2 Pt 1): p. E265- 72. 71. Schmitt, B., et al., Transcriptional adaptations of lipid metabolism in tibialis anterior muscle of endurance-trained athletes. Physiol Genomics, 2003. 15(2): p. 148-57.

189

72. Teran-Garcia, M., et al., Endurance training-induced changes in insulin sensitivity and gene expression. Am J Physiol Endocrinol Metab, 2005. 288(6): p. E1168-78. 73. Colmenares, C. and E. Stavnezer, The ski oncogene induces muscle differentiation in quail embryo cells. Cell, 1989. 59(2): p. 293-303. 74. Liu, D., B.L. Black, and R. Derynck, TGF-beta inhibits muscle differentiation through functional repression of myogenic transcription factors by Smad3. Genes Dev, 2001. 15(22): p. 2950-66. 75. Luo, K., Negative regulation of BMP signaling by the ski oncoprotein. J Bone Joint Surg Am, 2003. 85-A Suppl 3: p. 39-43. 76. Arndt, S., et al., Fussel-15, a novel Ski/Sno homolog protein, antagonizes BMP signaling. Mol Cell Neurosci, 2007. 34(4): p. 603-11. 77. Paez-Pereda, M., et al., Involvement of bone morphogenetic protein 4 (BMP-4) in pituitary prolactinoma pathogenesis through a Smad/estrogen receptor crosstalk. Proc Natl Acad Sci U S A, 2003. 100(3): p. 1034-9. 78. Xu, X., et al., Developmental expression of Smad1-7 suggests critical function of TGF-beta/BMP signaling in regulating epithelial-mesenchymal interaction during tooth morphogenesis. Int J Dev Biol, 2003. 47(1): p. 31-9. 79. Waite, K.A. and C. Eng, From developmental disorder to heritable cancer: it's all in the BMP/TGF-beta family. Nat Rev Genet, 2003. 4(10): p. 763-73. 80. Berger, J. and D.E. Moller, The mechanisms of action of PPARs. Annu Rev Med, 2002. 53: p. 409-35. 81. Evans, R.M., G.D. Barish, and Y.X. Wang, PPARs and the complex journey to obesity. Nat Med, 2004. 10(4): p. 355-61. 82. Gampe, R.T., Jr., et al., Asymmetry in the PPARgamma/RXRalpha crystal structure reveals the molecular basis of heterodimerization among nuclear receptors. Mol Cell, 2000. 5(3): p. 545-55. 83. Reddy, J.K. and T. Hashimoto, Peroxisomal beta-oxidation and peroxisome proliferator-activated receptor alpha: an adaptive metabolic system. Annu Rev Nutr, 2001. 21: p. 193-230. 84. Kersten, S., et al., Peroxisome proliferator-activated receptor alpha mediates the adaptive response to fasting. J Clin Invest, 1999. 103(11): p. 1489-98. 85. Barak, Y., et al., Effects of peroxisome proliferator-activated receptor delta on placentation, adiposity, and colorectal cancer. Proc Natl Acad Sci U S A, 2002. 99(1): p. 303-8. 86. Peters, J.M., et al., Growth, adipose, brain, and skin alterations resulting from targeted disruption of the mouse peroxisome proliferator-activated receptor beta(delta). Mol Cell Biol, 2000. 20(14): p. 5119-28. 87. Wang, Y.X., et al., Peroxisome-proliferator-activated receptor delta activates fat metabolism to prevent obesity. Cell, 2003. 113(2): p. 159-70. 88. Wang, Y.X., et al., Regulation of muscle fiber type and running endurance by PPARdelta. PLoS Biol, 2004. 2(10): p. e294. 89. Narkar, V.A., et al., AMPK and PPARdelta agonists are exercise mimetics. Cell, 2008. 134(3): p. 405-15. 90. Lehrke, M. and M.A. Lazar, The many faces of PPARgamma. Cell, 2005. 123(6): p. 993-9.

190

91. Walczak, R. and P. Tontonoz, PPARadigms and PPARadoxes: expanding roles for PPARgamma in the control of lipid metabolism. J Lipid Res, 2002. 43(2): p. 177-86. 92. Barak, Y., et al., PPAR gamma is required for placental, cardiac, and adipose tissue development. Mol Cell, 1999. 4(4): p. 585-95. 93. Kubota, N., et al., PPAR gamma mediates high-fat diet-induced adipocyte hypertrophy and insulin resistance. Mol Cell, 1999. 4(4): p. 597-609. 94. Rosen, E.D., et al., PPAR gamma is required for the differentiation of adipose tissue in vivo and in vitro. Mol Cell, 1999. 4(4): p. 611-7. 95. Agarwal, A.K. and A. Garg, A novel heterozygous mutation in peroxisome proliferator-activated receptor-gamma gene in a patient with familial partial lipodystrophy. J Clin Endocrinol Metab, 2002. 87(1): p. 408-11. 96. Hegele, R.A., et al., PPARG F388L, a transactivation-deficient mutant, in familial partial lipodystrophy. Diabetes, 2002. 51(12): p. 3586-90. 97. Savage, D.B., et al., Human metabolic syndrome resulting from dominant- negative mutations in the nuclear receptor peroxisome proliferator-activated receptor-gamma. Diabetes, 2003. 52(4): p. 910-7. 98. Lehmann, J.M., et al., An antidiabetic thiazolidinedione is a high affinity ligand for peroxisome proliferator-activated receptor gamma (PPAR gamma). J Biol Chem, 1995. 270(22): p. 12953-6. 99. Randle, P.J., Regulatory interactions between lipids and carbohydrates: the glucose fatty acid cycle after 35 years. Diabetes Metab Rev, 1998. 14(4): p. 263- 83. 100. Singh Ahuja, H., et al., Differential effects of rexinoids and thiazolidinediones on metabolic gene expression in diabetic rodents. Mol Pharmacol, 2001. 59(4): p. 765-73. 101. Martin, G., et al., Coordinate regulation of the expression of the fatty acid transport protein and acyl-CoA synthetase genes by PPARalpha and PPARgamma activators. J Biol Chem, 1997. 272(45): p. 28210-7. 102. Guitart, M., et al., FATP1 localizes to mitochondria and enhances pyruvate dehydrogenase activity in skeletal myotubes. Mitochondrion, 2009. 9(4): p. 266- 72. 103. Schoonjans, K., et al., PPARalpha and PPARgamma activators direct a distinct tissue-specific transcriptional response via a PPRE in the lipoprotein lipase gene. Embo J, 1996. 15(19): p. 5336-48. 104. Chinetti, G., et al., Expression of adiponectin receptors in human macrophages and regulation by agonists of the nuclear receptors PPARalpha, PPARgamma, and LXR. Biochem Biophys Res Commun, 2004. 314(1): p. 151-8. 105. Tsuchida, A., et al., Peroxisome proliferator-activated receptor (PPAR)alpha activation increases adiponectin receptors and reduces obesity-related inflammation in adipose tissue: comparison of activation of PPARalpha, PPARgamma, and their combination. Diabetes, 2005. 54(12): p. 3358-70. 106. Chawla, A., et al., A PPAR gamma-LXR-ABCA1 pathway in macrophages is involved in cholesterol efflux and atherogenesis. Mol Cell, 2001. 7(1): p. 161-71. 107. Perera, R.J., et al., Identification of novel PPARgamma target genes in primary human adipocytes. Gene, 2006. 369: p. 90-9.

191

108. Wan, Y. and R.M. Evans, Rosiglitazone activation of PPARgamma suppresses fractalkine signaling. J Mol Endocrinol. 44(2): p. 135-42. 109. Wilson-Fritch, L., et al., Mitochondrial remodeling in adipose tissue associated with obesity and treatment with rosiglitazone. J Clin Invest, 2004. 114(9): p. 1281-9. 110. Puigserver, P., et al., A cold-inducible coactivator of nuclear receptors linked to adaptive thermogenesis. Cell, 1998. 92(6): p. 829-39. 111. Lin, J., et al., Transcriptional co-activator PGC-1 alpha drives the formation of slow-twitch muscle fibres. Nature, 2002. 418(6899): p. 797-801. 112. Wu, Z., et al., Mechanisms controlling mitochondrial biogenesis and respiration through the thermogenic coactivator PGC-1. Cell, 1999. 98(1): p. 115-24. 113. Way, J.M., et al., Comprehensive messenger ribonucleic acid profiling reveals that peroxisome proliferator-activated receptor gamma activation has coordinate effects on gene expression in multiple insulin-sensitive tissues. Endocrinology, 2001. 142(3): p. 1269-77. 114. Brown, K.A., J.A. Pietenpol, and H.L. Moses, A tale of two proteins: differential roles and regulation of Smad2 and Smad3 in TGF-beta signaling. J Cell Biochem, 2007. 101(1): p. 9-33. 115. Derynck, R., Y. Zhang, and X.H. Feng, Smads: transcriptional activators of TGF- beta responses. Cell, 1998. 95(6): p. 737-40. 116. Massague, J., TGF-beta signal transduction. Annu Rev Biochem, 1998. 67: p. 753-91. 117. Attisano, L. and J.L. Wrana, Signal transduction by the TGF-beta superfamily. Science, 2002. 296(5573): p. 1646-7. 118. Piek, E., et al., Functional characterization of transforming growth factor beta signaling in Smad2- and Smad3-deficient fibroblasts. J Biol Chem, 2001. 276(23): p. 19945-53. 119. Yang, Y.C., et al., Hierarchical model of gene regulation by transforming growth factor beta. Proc Natl Acad Sci U S A, 2003. 100(18): p. 10269-74. 120. Kretschmer, A., et al., Differential regulation of TGF-beta signaling through Smad2, Smad3 and Smad4. Oncogene, 2003. 22(43): p. 6748-63. 121. Kim, S.G., et al., The endogenous ratio of smad2 and smad3 influences the cytostatic function of smad3. Mol Biol Cell, 2005. 16(10): p. 4672-83. 122. Brown, K.A., et al., Identification of novel Smad2 and Smad3 associated proteins in response to TGF-beta1. J Cell Biochem, 2008. 105(2): p. 596-611. 123. Heyer, J., et al., Postgastrulation Smad2-deficient embryos show defects in embryo turning and anterior morphogenesis. Proc Natl Acad Sci U S A, 1999. 96(22): p. 12595-600. 124. Nomura, M. and E. Li, Smad2 role in mesoderm formation, left-right patterning and craniofacial development. Nature, 1998. 393(6687): p. 786-90. 125. Waldrip, W.R., et al., Smad2 signaling in extraembryonic tissues determines anterior-posterior polarity of the early mouse embryo. Cell, 1998. 92(6): p. 797- 808. 126. Weinstein, M., et al., Failure of egg cylinder elongation and mesoderm induction in mouse embryos lacking the tumor suppressor smad2. Proc Natl Acad Sci U S A, 1998. 95(16): p. 9378-83.

192

127. Datto, M.B., et al., Targeted disruption of Smad3 reveals an essential role in transforming growth factor beta-mediated signal transduction. Mol Cell Biol, 1999. 19(4): p. 2495-504. 128. Yang, X., et al., Targeted disruption of SMAD3 results in impaired mucosal immunity and diminished T cell responsiveness to TGF-beta. EMBO J, 1999. 18(5): p. 1280-91. 129. Zhu, Y., et al., Smad3 mutant mice develop metastatic colorectal cancer. Cell, 1998. 94(6): p. 703-14. 130. Yang, J., R. Wahdan-Alaswad, and D. Danielpour, Critical role of Smad2 in tumor suppression and transforming growth factor-beta-induced apoptosis of prostate epithelial cells. Cancer Res, 2009. 69(6): p. 2185-90. 131. Hoot, K.E., et al., Keratinocyte-specific Smad2 ablation results in increased epithelial-mesenchymal transition during skin cancer formation and progression. J Clin Invest, 2008. 118(8): p. 2722-32. 132. Bour, B.A., et al., Drosophila MEF2, a transcription factor that is essential for myogenesis. Genes Dev, 1995. 9(6): p. 730-41. 133. Quinn, Z.A., et al., Smad proteins function as co-modulators for MEF2 transcriptional regulatory proteins. Nucleic Acids Res, 2001. 29(3): p. 732-742. 134. Nicol, R. and E. Stavnezer, Transcriptional repression by v-Ski and c-Ski mediated by a specific DNA binding site. J Biol Chem, 1998. 273(6): p. 3588-97. 135. Cohen, S.B., R. Nicol, and E. Stavnezer, A domain necessary for the transforming activity of SnoN is required for specific DNA binding, transcriptional repression and interaction with TAF(II)110. Oncogene, 1998. 17(19): p. 2505-13. 136. Lu, Z.H. and J. Chen, [Role of Ski/SnoN protein in the regulation of TGF-beta signal pathway]. Zhongguo Yi Xue Ke Xue Yuan Xue Bao, 2003. 25(2): p. 233-6. 137. Mizuide, M., et al., Two short segments of Smad3 are important for specific interaction of Smad3 with c-Ski and SnoN. J Biol Chem, 2003. 278(1): p. 531-6. 138. Ueki, N. and M.J. Hayman, Direct interaction of Ski with either Smad3 or Smad4 is necessary and sufficient for Ski-mediated repression of transforming growth factor-beta signaling. J Biol Chem, 2003. 278(35): p. 32489-92. 139. He, J., et al., The transforming activity of Ski and SnoN is dependent on their ability to repress the activity of Smad proteins. J Biol Chem, 2003. 278(33): p. 30540-7. 140. Li, Y., et al., Unique sequence, ski, in Sloan-Kettering avian retroviruses with properties of a new cell-derived oncogene. J Virol, 1986. 57(3): p. 1065-72. 141. Namciu, S., et al., Enhanced expression of mouse c-ski accompanies terminal skeletal muscle differentiation in vivo and in vitro. Dev Dyn, 1995. 204(3): p. 291-300. 142. Sutrave, P., et al., The induction of skeletal muscle hypertrophy by a ski transgene is promoter-dependent. Gene, 2000. 241(1): p. 107-116. 143. Chawla, A., et al., PPAR-gamma dependent and independent effects on macrophage-gene expression in lipid metabolism and inflammation. Nat Med, 2001. 7(1): p. 48-52. 144. Rageul, J., et al., KLF4-dependent, PPARgamma-induced expression of GPA33 in colon cancer cell lines. Int J Cancer, 2009. 125(12): p. 2802-9.

193

145. Savage, D.B., PPAR gamma as a metabolic regulator: insights from genomics and pharmacology. Expert Rev Mol Med, 2005. 7(1): p. 1-16. 146. Sutinen, J., et al., Effects of rosiglitazone on gene expression in subcutaneous adipose tissue in highly active antiretroviral therapy-associated lipodystrophy. Am J Physiol Endocrinol Metab, 2004. 286(6): p. E941-9. 147. Tachibana, K., et al., Gene expression profiling of potential peroxisome proliferator-activated receptor (PPAR) target genes in human hepatoblastoma cell lines inducibly expressing different PPAR isoforms. Nucl Recept, 2005. 3: p. 3. 148. Wiczer, B.M., et al., FATP1 mediates fatty acid-induced activation of AMPK in 3T3-L1 adipocytes. Biochem Biophys Res Commun, 2009. 387(2): p. 234-8. 149. Wu, Q., et al., FATP1 is an insulin-sensitive fatty acid transporter involved in diet-induced obesity. Mol Cell Biol, 2006. 26(9): p. 3455-67. 150. Zaripheh, S., et al., Dietary lycopene downregulates carotenoid 15,15'- monooxygenase and PPAR-gamma in selected rat tissues. J Nutr, 2006. 136(4): p. 932-8. 151. Eagle, H., et al., The utilization of carbohydrates by human cell cultures. J Biol Chem, 1958. 233(3): p. 551-8. 152. L Levintow, a.H.E., Biochemistry of Cultured Mammalian Cells. Annual Review of Biochemistry, 1961. 30: p. 605-640. 153. Carroll, R.C., et al., Reversion of transformed glycolysis to normal by inhibition of protein synthesis in rat kidney cells infected with temperature-sensitive mutant of Rous sarcoma virus. Proc Natl Acad Sci U S A, 1978. 75(10): p. 5015-9. 154. Warburg, O., Posener, K., and Negelein, E. , U¨ ber den Stoffwechsel der Tumoren. Biochemische Zeitschrift,, 1924. 152: p. 319-344. 155. Dickins, R.A., et al., Probing tumor phenotypes using stable and regulated synthetic microRNA precursors. Nat Genet, 2005. 37(11): p. 1289-95. 156. Silva, J.M., et al., Second-generation shRNA libraries covering the mouse and human genomes. Nat Genet, 2005. 37(11): p. 1281-8. 157. Stegmeier, F., et al., A lentiviral microRNA-based system for single-copy polymerase II-regulated RNA interference in mammalian cells. Proc Natl Acad Sci U S A, 2005. 102(37): p. 13212-7. 158. Hughes, S.H., et al., Adaptor plasmids simplify the insertion of foreign DNA into helper-independent retroviral vectors. J Virol, 1987. 61(10): p. 3004-12. 159. Gnaiger, E., MitoPathways - the blue book: Mitochondrial Pathways and Respiratory Control. OROBOROS MiPNet Publications, 2007. 160. Gnaiger, E., MitoPathways - the blue book: Mitochondrial Pathways and Respiratory Control. 2nd Edition. Mitochondrial Physiology Network 2009. 161. Hoppel, C.L., et al., The malonyl-CoA-sensitive form of carnitine palmitoyltransferase is not localized exclusively in the outer membrane of rat liver mitochondria. J Biol Chem, 1998. 273(36): p. 23495-503. 162. Bederman, I.R., et al., Triglyceride Synthesis in Epididymal Adipose Tissue: CONTRIBUTION OF GLUCOSE AND NON-GLUCOSE CARBON SOURCES. J Biol Chem, 2009. 284(10): p. 6101-8.

194

163. Lee, W.N., et al., Measurement of fractional lipid synthesis using deuterated water (2H2O) and mass isotopomer analysis. Am J Physiol, 1994. 266(3 Pt 1): p. E372-83. 164. Diraison, F., C. Pachiaudi, and M. Beylot, Measuring lipogenesis and cholesterol synthesis in humans with deuterated water: use of simple gas chromatographic/mass spectrometric techniques. J Mass Spectrom, 1997. 32(1): p. 81-6. 165. Krieg, R.C., et al., Trapping radioactive carbon dioxide during cellular metabolic assays under standard culture conditions: description of a unique gas-capturing device. J Biochem Biophys Methods, 2004. 58(2): p. 119-24. 166. de Fries, R. and M. Mitsuhashi, Quantification of mitogen induced human lymphocyte proliferation: comparison of alamarBlue assay to 3H-thymidine incorporation assay. J Clin Lab Anal, 1995. 9(2): p. 89-95. 167. Gonzalez, R.J. and J.B. Tarloff, Evaluation of hepatic subcellular fractions for Alamar blue and MTT reductase activity. Toxicol In Vitro, 2001. 15(3): p. 257-9. 168. O'Brien, J., et al., Investigation of the Alamar Blue (resazurin) fluorescent dye for the assessment of mammalian cell cytotoxicity. Eur J Biochem, 2000. 267(17): p. 5421-6. 169. Al-Nasiry, S., et al., The use of Alamar Blue assay for quantitative analysis of viability, migration and invasion of choriocarcinoma cells. Hum Reprod, 2007. 22(5): p. 1304-9. 170. Renner, K., et al., Changes of mitochondrial respiration, mitochondrial content and cell size after induction of apoptosis in leukemia cells. Biochim Biophys Acta, 2003. 1642(1-2): p. 115-23. 171. Presley, A.D., K.M. Fuller, and E.A. Arriaga, MitoTracker Green labeling of mitochondrial proteins and their subsequent analysis by capillary electrophoresis with laser-induced fluorescence detection. J Chromatogr B Analyt Technol Biomed Life Sci, 2003. 793(1): p. 141-50. 172. Zhang, H., et al., HIF-1 inhibits mitochondrial biogenesis and cellular respiration in VHL-deficient renal cell carcinoma by repression of C-MYC activity. Cancer Cell, 2007. 11(5): p. 407-20. 173. Reynier, P., et al., Mitochondrial DNA content affects the fertilizability of human oocytes. Mol Hum Reprod, 2001. 7(5): p. 425-9. 174. Satoh, M. and T. Kuroiwa, Organization of multiple nucleoids and DNA molecules in mitochondria of a human cell. Exp Cell Res, 1991. 196(1): p. 137-40. 175. Wiesner, R.J., J.C. Ruegg, and I. Morano, Counting target molecules by exponential polymerase chain reaction: copy number of mitochondrial DNA in rat tissues. Biochem Biophys Res Commun, 1992. 183(2): p. 553-9. 176. Evans, P.R., G.W. Farrants, and P.J. Hudson, Phosphofructokinase: structure and control. Philos Trans R Soc Lond B Biol Sci, 1981. 293(1063): p. 53-62. 177. Roche, T.E. and Y. Hiromasa, Pyruvate dehydrogenase kinase regulatory mechanisms and inhibition in treating diabetes, heart ischemia, and cancer. Cell Mol Life Sci, 2007. 64(7-8): p. 830-49. 178. Leonardi, R., et al., Coenzyme A: back in action. Prog Lipid Res, 2005. 44(2-3): p. 125-53.

195

179. Shaw, R.J., Glucose metabolism and cancer. Curr Opin Cell Biol, 2006. 18(6): p. 598-608. 180. Kim, J.W. and C.V. Dang, Cancer's molecular sweet tooth and the Warburg effect. Cancer Res, 2006. 66(18): p. 8927-30. 181. Campbell, S.E., et al., A novel function for fatty acid translocase (FAT)/CD36: involvement in long chain fatty acid transfer into the mitochondria. J Biol Chem, 2004. 279(35): p. 36235-41. 182. Sebastian, D., et al., Novel role of FATP1 in mitochondrial fatty acid oxidation in skeletal muscle cells. J Lipid Res, 2009. 50(9): p. 1789-99. 183. Holloway, G.P., et al., Skeletal muscle mitochondrial FAT/CD36 content and palmitate oxidation are not decreased in obese women. Am J Physiol Endocrinol Metab, 2007. 292(6): p. E1782-9. 184. Vluggens, A., et al., Functional significance of the two ACOX1 isoforms and their crosstalks with PPARalpha and RXRalpha. Lab Invest. 90(5): p. 696-708. 185. Fugier, C., et al., The lipoprotein lipase inhibitor ANGPTL3 is negatively regulated by thyroid hormone. J Biol Chem, 2006. 281(17): p. 11553-9. 186. Calvo, J.A., et al., Muscle-specific expression of PPARgamma coactivator-1alpha improves exercise performance and increases peak oxygen uptake. J Appl Physiol, 2008. 104(5): p. 1304-12. 187. Leonardi, R., et al., Activation of human mitochondrial pantothenate kinase 2 by palmitoylcarnitine. Proc Natl Acad Sci U S A, 2007. 104(5): p. 1494-9. 188. Bogacka, I., et al., Pioglitazone induces mitochondrial biogenesis in human subcutaneous adipose tissue in vivo. Diabetes, 2005. 54(5): p. 1392-9. 189. Kelly, D.P. and R.C. Scarpulla, Transcriptional regulatory circuits controlling mitochondrial biogenesis and function. Genes Dev, 2004. 18(4): p. 357-68. 190. Larsen, J., H. Beug, and M.J. Hayman, The v-ski oncogene cooperates with the v- sea oncogene in erythroid transformation by blocking erythroid differentiation. Oncogene, 1992. 7(10): p. 1903-11. 191. Shinagawa, T., et al., Increased susceptibility to tumorigenesis of ski-deficient heterozygous mice. Oncogene, 2001. 20(56): p. 8100-8. 192. Le Scolan, E., et al., Transforming growth factor-beta suppresses the ability of Ski to inhibit tumor metastasis by inducing its degradation. Cancer Res, 2008. 68(9): p. 3277-85. 193. Dai, P., et al., Ski is involved in transcriptional regulation by the repressor and full-length forms of Gli3. Genes Dev, 2002. 16(22): p. 2843-8. 194. Xu, J., X. Luo, and N. Chegini, Differential expression, regulation, and induction of Smads, transforming growth factor-beta signal transduction pathway in leiomyoma, and myometrial smooth muscle cells and alteration by gonadotropin- releasing hormone analog. J Clin Endocrinol Metab, 2003. 88(3): p. 1350-61. 195. Liu, X., et al., Ski/Sno and TGF-beta signaling. Cytokine Growth Factor Rev, 2001. 12(1): p. 1-8. 196. Dennler, S., et al., Direct binding of Smad3 and Smad4 to critical TGF beta- inducible elements in the promoter of human plasminogen activator inhibitor-type 1 gene. Embo J, 1998. 17(11): p. 3091-100. 197. Stroschein, S.L., et al., Negative feedback regulation of TGF-beta signaling by the SnoN oncoprotein. Science, 1999. 286(5440): p. 771-4.

196

198. Sun, Y., et al., SnoN and Ski protooncoproteins are rapidly degraded in response to transforming growth factor beta signaling. Proc Natl Acad Sci U S A, 1999. 96(22): p. 12442-7. 199. Dou, C., et al., BF-1 interferes with transforming growth factor beta signaling by associating with Smad partners. Mol Cell Biol, 2000. 20(17): p. 6201-11. 200. Xu, L., Y.G. Chen, and J. Massague, The nuclear import function of Smad2 is masked by SARA and unmasked by TGFbeta-dependent phosphorylation. Nat Cell Biol, 2000. 2(8): p. 559-62. 201. Xu, X.L., K.R. Dai, and T.T. Tang, [The role of Smads and related transcription factors in the signal transduction of bone morphogenetic protein inducing bone formation]. Zhongguo Xiu Fu Chong Jian Wai Ke Za Zhi, 2003. 17(5): p. 359-62. 202. Stroschein, S.L., et al., Smad3 recruits the anaphase-promoting complex for ubiquitination and degradation of SnoN. Genes Dev, 2001. 15(21): p. 2822-36. 203. Ghosh, A.K., et al., Smad-dependent stimulation of type I collagen gene expression in human skin fibroblasts by TGF-beta involves functional cooperation with p300/CBP transcriptional coactivators. Oncogene, 2000. 19(31): p. 3546-55. 204. Lee, K.S., et al., Runx2 is a common target of transforming growth factor beta1 and bone morphogenetic protein 2, and cooperation between Runx2 and Smad5 induces osteoblast-specific gene expression in the pluripotent mesenchymal precursor cell line C2C12. Mol Cell Biol, 2000. 20(23): p. 8783-92. 205. Mori, Y., S.J. Chen, and J. Varga, Modulation of endogenous Smad expression in normal skin fibroblasts by transforming growth factor-beta. Exp Cell Res, 2000. 258(2): p. 374-83. 206. Vignais, M.L., [Ski and SnoN: antagonistic proteins of TGFbeta signaling]. Bull Cancer, 2000. 87(2): p. 135-7. 207. Yanagisawa, K., et al., Heterogeneities in the biological and biochemical functions of Smad2 and Smad4 mutants naturally occurring in human lung cancers. Oncogene, 2000. 19(19): p. 2305-11. 208. Yanagisawa, K., et al., Induction of apoptosis by Smad3 and down-regulation of Smad3 expression in response to TGF-beta in human normal lung epithelial cells. Oncogene, 1998. 17(13): p. 1743-7. 209. Yagi, K., et al., Alternatively spliced variant of Smad2 lacking exon 3. Comparison with wild-type Smad2 and Smad3. J Biol Chem, 1999. 274(2): p. 703-9. 210. Takenoshita, S., et al., Characterization of the MADH2/Smad2 gene, a human Mad homolog responsible for the transforming growth factor-beta and activin signal transduction pathway. Genomics, 1998. 48(1): p. 1-11. 211. Wan, M., et al., Jab1 antagonizes TGF-beta signaling by inducing Smad4 degradation. EMBO Rep, 2002. 3(2): p. 171-6. 212. Funaba, M. and L.S. Mathews, Identification and characterization of constitutively active Smad2 mutants: evaluation of formation of Smad complex and subcellular distribution [In Process Citation]. Mol Endocrinol, 2000. 14(10): p. 1583-91. 213. Vijayachandra, K., J. Lee, and A.B. Glick, Smad3 regulates senescence and malignant conversion in a mouse multistage skin carcinogenesis model. Cancer Res, 2003. 63(13): p. 3447-52.

197

214. Tremblay, K.D., et al., Formation of the definitive endoderm in mouse is a Smad2-dependent process. Development, 2000. 127(14): p. 3079-90. 215. Goumans, M.J. and C. Mummery, Functional analysis of the TGFbeta receptor/Smad pathway through gene ablation in mice. Int J Dev Biol, 2000. 44(3): p. 253-65. 216. Itoh, S., et al., The transcriptional co-activator P/CAF potentiates TGF- beta/Smad signaling [In Process Citation]. Nucleic Acids Res, 2000. 28(21): p. 4291-8. 217. Weinstein, D.C., et al., FGF-mediated mesoderm induction involves the Src- family kinase Laloo. Nature, 1998. 394(6696): p. 904-8. 218. Abraham, S., et al., Regulation of MCP-1 gene transcription by Smads and HIV-1 Tat in human glial cells. Virology, 2003. 309(2): p. 196-202. 219. Dumont, E., et al., Evidence for a role of Smad3 and Smad2 in stabilization of the tumor-derived mutant Smad2.Q407R. J Biol Chem, 2003. 278(27): p. 24881-7. 220. Zawel, L., et al., Human Smad3 and Smad4 are sequence-specific transcription activators. Mol Cell, 1998. 1(4): p. 611-7. 221. Dennler, S., S. Huet, and J.M. Gauthier, A short amino-acid sequence in MH1 domain is responsible for functional differences between Smad2 and Smad3. Oncogene, 1999. 18(8): p. 1643-8. 222. Chen, Y., A. Bhushan, and W. Vale, Smad8 mediates the signaling of the ALK-2 [corrected] receptor serine kinase. Proc Natl Acad Sci U S A, 1997. 94(24): p. 12938-43. 223. Randall, R.A., et al., Different Smad2 partners bind a common hydrophobic pocket in Smad2 via a defined proline-rich motif. Embo J, 2002. 21(1-2): p. 145- 56. 224. Germain, S., et al., Homeodomain and winged-helix transcription factors recruit activated Smads to distinct promoter elements via a common Smad interaction motif. Genes Dev, 2000. 14(4): p. 435-51. 225. Rodriguez-Pascual, F., M. Redondo-Horcajo, and S. Lamas, Functional cooperation between Smad proteins and activator protein-1 regulates transforming growth factor-beta-mediated induction of endothelin-1 expression. Circ Res, 2003. 92(12): p. 1288-95. 226. Cui, X.M., et al., Overexpression of Smad2 in Tgf-beta3-null mutant mice rescues cleft palate. Dev Biol, 2005. 278(1): p. 193-202. 227. Faivre, L., et al., Prenatal detection of a 1p36 deletion in a fetus with multiple malformations and a review of the literature. Prenat Diagn, 1999. 19(1): p. 49-53. 228. Leask, A., et al., The control of ccn2 (ctgf) gene expression in normal and scleroderma fibroblasts. Mol Pathol, 2001. 54(3): p. 180-3. 229. Toyomizu, M., et al., Cold-induced mitochondrial uncoupling and expression of chicken UCP and ANT mRNA in chicken skeletal muscle. FEBS Lett, 2002. 529(2-3): p. 313-8. 230. Chen, H., et al., Mitofusins Mfn1 and Mfn2 coordinately regulate mitochondrial fusion and are essential for embryonic development. J Cell Biol, 2003. 160(2): p. 189-200.

198

231. Boerner, P., R.J. Resnick, and E. Racker, Stimulation of glycolysis and amino acid uptake in NRK-49F cells by transforming growth factor beta and epidermal growth factor. Proc Natl Acad Sci U S A, 1985. 82(5): p. 1350-3. 232. Arrick, B.A., et al., Altered metabolic and adhesive properties and increased tumorigenesis associated with increased expression of transforming growth factor beta 1. J Cell Biol, 1992. 118(3): p. 715-26. 233. Nowak, G. and R.G. Schnellmann, Autocrine production and TGF-beta 1- mediated effects on metabolism and viability in renal cells. Am J Physiol, 1996. 271(3 Pt 2): p. F689-97. 234. Ghosh, A.K., et al., Peroxisome proliferator-activated receptor-gamma abrogates Smad-dependent collagen stimulation by targeting the p300 transcriptional coactivator. FASEB J, 2009. 23(9): p. 2968-77. 235. Moldes, M., et al., Peroxisome-proliferator-activated receptor gamma suppresses Wnt/beta-catenin signalling during adipogenesis. Biochem J, 2003. 376(Pt 3): p. 607-13. 236. Sharma, A.M. and B. Staels, Review: Peroxisome proliferator-activated receptor gamma and adipose tissue--understanding obesity-related changes in regulation of lipid and glucose metabolism. J Clin Endocrinol Metab, 2007. 92(2): p. 386-95. 237. Cohn, M. and A. Hu, Isotopic (18O) shift in 31P nuclear magnetic resonance applied to a study of enzyme-catalyzed phosphate--phosphate exchange and phosphate (oxygen)--water exchange reactions. Proc Natl Acad Sci U S A, 1978. 75(1): p. 200-3.

199