digital.library.adelaide.edu.au · v abstract the hypoxia inducible factors (hifs) are widely...
TRANSCRIPT
i
INVESTIGATING POTENTIAL POST‐
TRANSLATIONAL MODIFICATION OF
FACTOR‐INHIBITING HIF (FIH‐1)
iii
INVESTIGATING POTENTIAL POST‐
TRANSLATIONAL MODIFICATION OF
FACTOR‐INHIBITING HIF (FIH‐1)
Submitted for the degree of Doctor of Philosophy
Karolina Lisy BSc (Hons) Biotechnology
School of Molecular and Biomedical Science, University of Adelaide
June, 2011
v
ABSTRACT
The Hypoxia Inducible Factors (HIFs) are widely expressed transcription factors critical
for altering gene expression in hypoxic cells and enabling cellular adaptation to
conditions of limited oxygen availability. The HIFs are labile and inactive when oxygen
levels are sufficient to meet cellular oxygen demand, but become stabilised and
transcriptionally active when oxygen levels decrease. Factor Inhibiting HIF‐1α (FIH‐1) is
an asparaginyl hydroxylase that was first identified via its interaction with the HIF‐α
subunit. The canonical role for the enzyme involves the oxygen‐dependent regulation of
HIF transcriptional activity. In normoxia, FIH‐1 hydroxylates an asparaginyl residue in
the C‐terminal transactivation domain of HIF‐α, blocking interaction with vital
transcriptional coactivators and abrogating HIF transcriptional activity. As FIH‐1 requires
oxygen for hydroxylation, activity of the enzyme decreases with decreasing oxygen
levels, allowing HIF‐α to escape hydroxylation and consequently activate target gene
expression during periods of insufficient oxygen tension.
More recently, FIH‐1 has been found to bind and hydroxylate a number of proteins
containing ankyrin repeat domains (ARDs). However, despite the prevalence of ARD
hydroxylation, there is, as yet, no established role attributed to these modification
events. Additionally, FIH‐1 knockout mice have revealed a surprising role for FIH‐1 as a
neuronal regulator of metabolism, suggesting a novel, cell‐specific role for the enzyme.
As FIH‐1 requires O2 for catalysis, the availability of intracellular oxygen is thought to
determine activity of the enzyme, branding FIH‐1 as a putative cellular oxygen sensor.
Aside from modulation of enzyme activity by oxygen levels, little is known about the
regulation of FIH‐1. Several lines of evidence have suggested that FIH‐1 exhibits cell
type‐specific differences in activity toward HIF‐α substrates that may act in addition to
or separately from the regulation of enzyme activity by levels of available oxygen.
vi
The primary aim of this work was to investigate the post‐translational modification
(PTM) of FIH‐1 in order to uncover any additional regulatory mechanisms that may
exist. Two‐dimensional electrophoresis (2‐DE) experiments from a number of cancer cell
lines and mouse embryonic fibroblasts revealed heterogeneity in the isoelectric point of
FIH‐1, suggesting the existence of multiple post‐translationally‐modified forms of the
enzyme. Overexpressed FIH‐1 was affinity purified for mass spectrometric (MS)
identification of PTMs. MS analysis was able to demonstrate asparaginyl deamidation
and methionine oxidation. It is unclear, however, whether these modifications
represent modifications occurring in the cellular environment or during sample
processing.
Due to inefficient purification of FIH‐1 from cells, it could not be ascertained if
phosphorylation was present. However, phosphatase treatment of cell lysate followed
by 2‐DE consistently showed a decrease in spot number and shift of FIH‐1 spots to a
more basic position on a 2‐D field, suggesting that the isoelectric point differences of
FIH‐1 could be attributed, in part, to phosphorylation. Furthermore, in vitro
phosphorylation assays indicated that recombinant FIH‐1 was able to be
phosphorylated by kinases supplied by cell lysate.
In summary, the work presented here provides evidence for the existence of novel
PTMs of FIH‐1, and suggests that FIH‐1 may be a kinase substrate.
vii
CANDIDATES DECLARATION
This work contains no material which has been accepted for the award of any other
degree or diploma in any university or other tertiary institution to Karolina Lisy and, to
the best of my knowledge and belief, contains no material previously published or
written by another person, except where due reference has been made in the text.
I give consent to this copy of my thesis, when deposited in the University Library, being
made available for loan and photocopying, subject to the provisions of the Copyright
Act 1968.
I also give permission for the digital version of my thesis to be made available on the
web, via the University’s digital research repository, the Library catalogue, the
Australasian Digital Theses Program (ADTP) and also through web search engines,
unless permission has been granted by the University to restrict access for a period of
time.
Signature: ……………………………………………….
Date: ………………..
viii
ACKNOWLEDGMENTS
First and foremost I’d like to thank my supervisor Dr. Daniel Peet for inviting me into the
lab to do my PhD. The guidance and support I’ve received over the years have been
invaluable and I sincerely appreciate the time and input invested into the project.
Thank you also to Murray Whitelaw
To all the members of the Peet lab, both past and present, thanks for making the lab as
fun as it was! To darling Rachel, a wonderful person with a heart of gold and limitless
patience, we started out together as two little Honours students who had no idea what
we were doing, and it has been a privilege to work with you and thanks so much for
your friendship over the years. To Sarah Linke, hands down the most brilliant person I
know, I wish all the best for what’s coming next, and to the delightful Sarah Wilkins, all
the best for the future and for Oxford. Anne Chapman‐Smith, thanks for all the talks and
advice on both science and personal matters, and Colleen Bindloss I feel so fortunate to
have had the opportunity to meet and work with such a wonderful group of people, so
thanks Anne Raimondo, Briony Davenport, Sam Olechnowicz, Cameron Bracken,
Anthony Fedele, Teresa Otto, Rebecca Bilton, Natalia Martin, Max Tollenaere and
Lauren Watkins.
Mum and Dad, your unwavering support and help have been so important to me and I
would like to thank you for Thanks to my beautiful grandparents, who have always
taken an interest in what I do, and in particular to Dĕdeček who will probably try to read
this thesis even though he doesn’t speak English! And lastly to Patrick, who has been
witness to all the horrors involved in finishing this PhD
ix
xi
ABBREVIATIONS
2‐DE Two‐Dimensional Electrophoresis
2‐OG 2‐oxoglutarate
3’UTR 3’ untranslated region
6His 6x Histidine
A280 Absorbance at 280 nm
Ank Ankyrin
Amp Ampicillin
ARD Ankyrin repeat domain
Arnt Aryl hydrocarbon nuclear translocator
APS Ammonium persulphate
ATP Adenosine triphosphate
bHLH Basic helix‐loop‐helix
BME Beta mercaptoethanol
bp base pairs
BSA Bovine serum albumin
CAD Carboxy‐terminal transactivation domain
CA9 Carbonic Anhydrase 9
CBP CREB‐binding protein
CCRCC Clear cell renal cell carcinoma
CD Deamidation coefficient
CDP CCAAT‐displacement protein
CHAPS 3‐[(3‐Cholamidopropyl)dimethylammonio]‐1‐
propanesulfonate
CNBr Cyanogen Bromide
DAPI 4',6‐diamidino‐2‐phenylindole
xii
DBD DNA binding domain
DMEM Dulbecco’s modified eagle medium
DMSO Dimethylsulphoxide
DNA Deoxyribonucleic acid
DP 2,2’ dipyridyl
DTT Dithiothreitol
ECL Enhanced chemiluminescence
EDTA Ethylene diamide tetra‐acetic acid
ELISA Enzyme‐linked immunosorbent assay
EPO Erythropoetin
EtBr Ethidium bromide
FCS Fetal calf serum
FIH‐1 Factor Inhibiting HIF
GFP Green fluorescent protein
Glut Glucose transporter
GSV Glut4 storage vesicle
GTS Glycine/trizma base/SDS
HDAC Histone deacetylase
HEPES 4‐(2‐Hydroxyethyl)piperazine‐1‐ethanesulfonic acid
HIF‐ Hypoxia inducible factor subunit
HIF‐β Hypoxia inducible factor β subunit
HIF Hypoxia inducible factor (heterodimer)
hnRNP B1 Heterogeneous nuclear ribonucleoprotein B1
HNSCC Head and neck squamous cell carcinoma
HRE Hypoxic response element
HRP Horseradish peroxidise
ICD Intracellular domain
IEF Isoelectric focusing
xiii
IF Immunofluorescence
Ig Immunoglobulin
ILK‐1 Integrin‐linked kinase 1
IMAC Immobilised metal affinity chromatography
IP Immunoprecipitation
IPG Immobilised pH gradient
IPTG Isopropyl‐‐D‐thiogalactopyranoside
IRAP Insulin‐regulated aminopeptidase
IRES Internal ribosome entry site
Kb Kilobase
KDa Kilodalton
Km Michaelis constant
Ko knockout
LB Luria Broth
LDHA Lactate dehydrogenase A
Luc Luciferase
MALDI Matrix‐assisted laser desorption/ionisation
MAPK Mitogen‐activated protein kinase
MBP Maltose binding protein
MEF Mouse Embryonic Fibroblast
miRNA Micro RNA
MQ Milli‐Q
mRNA Messenger ribonucleic acid
MS Mass Spectrometry
MT1‐MMP Membrane type 1 matrix metalloprotease
MW Molecular weight
NAD N‐terminal transactivation domain
Ni‐IDA Nickel nitrilotriacetic acid
xiv
NO Nitric oxide
OD600 Optical density at 600 nm
ODDD Oxygen‐Dependent Degradation Domain
O/N Overnight
PAGE Polyacrylamide gel electrophoresis
PAS Per‐Arnt‐SIM homology domain
PBS Phosphate buffered saline
PBT Phosphate buffered saline with 0.1% Tween‐20
PCR Polymerase chain reaction
PI3K phosphatidylinositol 3 kinase
Pen/Strep Penicillin/streptomycin
PGK1 Phosphoglycerate kinase 1
PHD Prolyl hydroxylase domain‐containing protein
pI Isoelectric point
PKC Protein kinase C zeta
PM Plasma membrane
PMSF Phenylmethyl sulfonyl fluoride
PoAb Polyclonal antibody
PP1R12C Protein phosphatase 1 regulatory subunit 12 C
PPAse Phosphatase
PTM Post‐translational modification
pVHL Von Hippel Lindau protein
RCC Renal cell carcinoma
RNA Ribonucleic acid
ROS Reactive oxygen species
RT Room temperature
SB3‐10 N‐decyl‐N,N‐dimethyl‐3‐ammonio‐l‐propane‐sulfonate
SDS Sodium dodecyl sulfate
xv
SDS PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis
shRNA short hairpin RNA
siRNA Small interfering RNA
TAD Transactivation domain
TCA cycle Tricarboxylic acid cycle
TE Tris/EDTA
TEMED N,N,N1,N1‐teramethyl‐ethylenediamide
Tween‐20 polyoxyethylene‐sorbitan monolaurate
Tris Tris (hydroxymethyl) aminomethane
Trx Thioredoxin
VEGF Vascular endothelial growth factor
VHr Volt hour
Vmax Maximum rate
WB Western blot
WCE Whole cell extract
WCEB Whole cell extract buffer
Wt Wildtype
Y2H Yeast 2 hybrid
xvii
TABLE OF CONTENTS
1 INTRODUCTION .................................................................................................................. 3
1.1 REGULATION OF THE HYPOXIA INDUCIBLE FACTORS 3
1.1.1 THE NECESSITY OF OXYGEN ........................................................................................................... 3 1.1.2 DEFINITIONS OF NORMOXIA AND HYPOXIA ...................................................................................... 4 1.1.3 PHYSIOLOGICAL AND PATHOPHYSIOLOGICAL CAUSES OF HYPOXIA ........................................................ 4 1.1.4 THE HYPOXIA INDUCIBLE FACTORS ................................................................................................. 5 1.1.5 HIF TARGET GENES ..................................................................................................................... 7 1.1.6 OXYGEN‐DEPENDENT REGULATION OF HIF‐Α ................................................................................. 11 1.1.7 OXYGEN‐DEPENDENT REGULATION OF HIF‐Α STABILITY ................................................................... 12 1.1.8 OXYGEN‐DEPENDENT REGULATION OF THE HIF‐Α‐CAD TRANSCRIPTIONAL ACTIVITY ............................. 15
1.2 FACTOR‐INHIBITING HIF 19
1.2.1 FIH‐1 EXPRESSION AND SUBCELLULAR LOCALISATION ...................................................................... 19 1.2.2 FIH‐1 STRUCTURE .................................................................................................................... 19 1.2.3 REGULATION OF FIH‐1 ACTIVITY BY OXYGEN LEVELS: IN VITRO DATA .................................................. 20 1.2.4 REGULATION OF FIH‐1 ACTIVITY BY OXYGEN LEVELS: DATA FROM CELL‐BASED EXPERIMENTS ................. 21
1.3 ARD‐CONTAINING SUBSTRATES 29
1.3.1 DOES FIH‐1 POST‐TRANSLATIONALLY MODIFY OTHER PROTEINS? ...................................................... 29 1.3.2 IDENTIFICATION OF NOVEL FIH‐1 SUBSTRATES ............................................................................... 29 1.3.3 NOTCH IS AN FIH‐1 SUBSTRATE .................................................................................................. 30 1.3.4 FUNCTION OF NOTCH HYDROXYLATION......................................................................................... 31 1.3.5 ANKYRIN REPEAT DOMAIN‐CONTAINING SUBSTRATES ..................................................................... 32 1.3.6 FUNCTION OF ARD HYDROXYLATION ............................................................................................ 33
1.4 WHAT IS THE ROLE OF FIH‐1? 36
1.4.1 FIH‐1 KNOCKOUT MOUSE PHENOTYPE .......................................................................................... 36 1.4.2 A CELL‐TYPE SPECIFIC ROLE FOR FIH‐1? ........................................................................................ 38
1.5 REGULATION OF FIH‐1 39
1.5.1 REGULATION AT THE FIH‐1 PROMOTER BY PROTEIN KINASE C ....................................................... 39 1.5.2 REGULATION OF FIH‐1 TRANSLATION BY MICRORNA 31 ................................................................. 40 1.5.3 REGULATION OF HIF TRANSCRIPTIONAL ACTIVITY BY NITRIC OXIDE ..................................................... 41 1.5.4 FIH‐1 IS NOT REGULATED BY TRICARBOXYLIC ACID CYCLE INTERMEDIATES ........................................... 42 1.5.5 REGULATION OF FIH‐1 SUBCELLULAR DISTRIBUTION BY MT1‐MMP AND MINT3 ............................... 43 1.5.7 IS FIH‐1 PHOSPHORYLATED? ...................................................................................................... 44
xviii
1.6 FIH‐1 AND DISEASE 49
1.6.1 HYPOXIA, THE HIF PATHWAY AND SOLID TUMOUR GROWTH ............................................................. 49
1.7 THESIS OBJECTIVES 52
2 MATERIALS AND METHODS ............................................................................................. 57
2.1 CHEMICALS AND REAGENTS 57
2.2 RADIOCHEMICALS 58
2.2 COMMERCIAL KITS 58
2.3 ENZYMES 58
2.4 ANTIBODIES 59
2.4.1 PRIMARY ANTIBODIES ................................................................................................................ 59 2.4.3 SECONDARY ANTIBODIES ............................................................................................................ 60
2.5 BUFFERS AND SOLUTIONS 60
2.7 BACTERIAL STRAINS AND GROWTH MEDIA 62
2.8 PLASMIDS 63
2.8.1 BACTERIAL EXPRESSION PLASMIDS ............................................................................................... 63 2.8.2 MAMMALIAN EXPRESSION PLASMIDS ........................................................................................... 64
2.9 SIRNA‐MEDIATED KNOCKDOWN OF FIH‐1 65
2.9.1 PLASMID‐BASED SYSTEM ............................................................................................................ 65 2.9.2 SIRNA OLIGONUCLEOTIDES ......................................................................................................... 66
2.10 GENERAL DNA METHODS 66
2.10.1 TRANSFORMATIONS ................................................................................................................. 66 2.10.2 DNA PREPARATION ................................................................................................................. 67 2.10.3 AGAROSE GEL ELECTROPHORESIS ............................................................................................... 67 2.10.4 GEL‐PURIFICATION OF DNA ...................................................................................................... 68 2.10.5 RESTRICTION DIGESTS .............................................................................................................. 68 2.10.6 LIGATIONS ............................................................................................................................. 68 2.10.7 SEQUENCING .......................................................................................................................... 68
2.11 RECOMBINANT PROTEIN PURIFICATION METHODS 69
2.11.1 NI2+‐AFFINITY PURIFICATION OF RECOMBINANT HIS‐TAGGED PROTEINS ............................................ 69 2.11.2 AMYLOSE‐AFFINITY PURIFICATION OF RECOMBINANT MBP‐TAGGED PROTEINS .................................. 70
xix
2.12 GENERAL PROTEIN METHODS 70
2.12.1 PREPARATION OF CELL LYSATES ................................................................................................. 70 2.12.2 PREPARATION OF NUCLEAR AND CYTOSOLIC EXTRACTS .................................................................. 71 2.12.3 PROTEIN QUANTIFICATION ........................................................................................................ 71 2.12.4 SODIUM DODECYL SULFATE POLYACRYLAMIDE GEL ELECTROPHORESIS .............................................. 71 2.12.5 PROTEIN STAINING .................................................................................................................. 72 2.12.6 WESTERN BLOTTING ................................................................................................................ 72 2.12.7 STRIPPING AND RE‐PROBING WESTERN BLOTS .............................................................................. 72 2.12.8 IMMUNOFLUORESCENCE .......................................................................................................... 73
2.13 IGM COLUMN PURIFICATION 73
2.14 CYANOGEN BROMIDE‐ACTIVATED SEPHAROSE PURIFICATION OF IGM 74
2.14.1 PREPARATION OF CYANOGEN BROMIDE‐ACTIVATED SEPHAROSE ...................................................... 74 2.12.2 9F6 DIALYSIS .......................................................................................................................... 74 2.12.3 COUPLING 9F6 TO CYANOGEN BROMIDE‐ACTIVATED SEPHAROSE .................................................... 74 2.12.4 CAPTURING THE 9F6 ANTIGEN FROM CELL LYSATE ........................................................................ 75 2.14.5 ELUTION OF CAPTURED ANTIGEN ............................................................................................... 75
2.15 GENERAL MAMMALIAN CELL CULTURE METHODS 75
2.15.1 MAMMALIAN CELL LINES AND MEDIA ......................................................................................... 75 2.15.2 TRANSFECTION ....................................................................................................................... 76 2.15.3 GENERATION OF STABLE CELL LINES ............................................................................................ 76
2.16 TWO‐DIMENSIONAL ELECTROPHORESIS 76
2.16.1 SAMPLE PREPARATION: METHOD 1 ............................................................................................ 76 2.16.2 SAMPLE PREPARATION: METHOD 2 ............................................................................................ 77 2.16.3 STRIP REHYDRATION AND ISOELECTRIC FOCUSING ......................................................................... 77 2.16.4 STRIP EQUILIBRATION .............................................................................................................. 78 2.16.5 SECOND DIMENSION SDS‐PAGE ............................................................................................... 78 2.16.6 VISUALISATION ....................................................................................................................... 79
2.17 IMMUNOPRECIPITATION 79
2.17.1 CELL LYSATE PREPARATION ....................................................................................................... 79 2.17.2 PRECLEARING ......................................................................................................................... 79 2.17.3 ANTIGEN‐ANTIBODY COMPLEX FORMATION ................................................................................. 80 2.17.4 BEAD REHYDRATION AND BLOCKING ........................................................................................... 80 2.17.5 IMMUNE COMPLEX BINDING TO RESIN ........................................................................................ 80 2.17.6 ELUTION ................................................................................................................................ 80
2.18 NOTCH‐AFFINITY PULLDOWNS 81
2.18.1 NOTCH CONSTRUCT EXPRESSION ............................................................................................... 81
xx
2.18.2 NOTCH CONSTRUCT PURIFICATION ............................................................................................. 81 2.18.3 NOTCH‐AFFINITY PURIFICATION ................................................................................................. 81
2.19 NI2+‐AFFINITY PURIFICATION OF FIH‐1 FROM HELA CELLS 82
2.20 PHOSPHATASE TREATMENTS 83
2.20.1 CELL LYSIS .............................................................................................................................. 83 2.20.2 PHOSPHATASE TREATMENTS ..................................................................................................... 83 2.20.3 POSITIVE CONTROL .................................................................................................................. 83 2.20.4 2‐DE .................................................................................................................................... 84 2.20.5 INTERNAL REFERENCE PROTEIN .................................................................................................. 84
2.21 IN VITRO PHOSPHORYLATION ASSAY 84
2.21.1 PROTEIN EXPRESSION AND PURIFICATION .................................................................................... 84 2.21.2 PREPARATION OF CELL LYSATE ................................................................................................... 85 2.21.3 PHOSPHORYLATION ASSAY ........................................................................................................ 85 2.21.4 PHOSPHATASE TREATMENT ....................................................................................................... 85 2.21.5 TEV CLEAVAGE ....................................................................................................................... 86
3 CHARACTERISING FIH‐1 EXPRESSION ............................................................................... 89
3. 1 INTRODUCTION 89
3.1.1 EXPRESSION AND IMPORTANCE OF HIF‐1Α AND HIF‐2Α IN CANCER PROGRESSION ............................... 89 3.1.2 WHAT WAS KNOWN ABOUT FIH‐1 EXPRESSION? ............................................................................ 90
3.2 HONOURS RESULTS 91
3.2.1 GENERATION OF FIH‐1 MONOCLONAL ANTIBODY (UNDERGRADUATE RESEARCH YEAR, 2004) ............... 91 3.2.2 GENERATION OF AN ANTI‐FIH‐1 MONOCLONAL ANTIBODY ............................................................... 92 3.2.3 PRELIMINARY USE OF 9F6 ........................................................................................................... 92 3.2.4 CONCLUSIONS FROM UNDERGRADUATE RESEARCH .......................................................................... 95
3.3 FURTHER CHARACTERISATION OF 9F6 ANTIGEN 99
3.3.1 9F6 CAN DETECT PURIFIED AND OVEREXPRESSED FIH‐1 ................................................................... 99 3.3.2 SIRNA‐MEDIATED KNOCKDOWN OF FIH‐1 .................................................................................. 103 3.3.3 9F6 IS AN IGM MONOCLONAL ANTIBODY .................................................................................... 107 3.3.4 IMMUNOPRECIPITATION OF 9F6 ANTIGEN BY THIOPHILIC INTERACTION CHROMATOGRAPHY ................ 108 3.3.5 IMMUNOPRECIPITATION OF 9F6 ANTIGEN USING CNBR‐ACTIVATED SEPHAROSE ................................ 113 3.3.6 9F6 ANTIGEN IS NOT FIH‐1 ...................................................................................................... 118 3.3.7 USE OF POLYCLONAL ANTI‐FIH‐1 ANTIBODY TO INVESTIGATE FIH‐1 EXPRESSION ............................... 121
3.4 SUMMARY AND DISCUSSION 121
3.4.1 EXPRESSION OF FIH‐1 IN CANCER: PAPER PUBLISHED .................................................................... 125
xxi
3.4.2 EXPRESSION OF FIH‐1 IN BREAST CANCER: IS SUBCELLULAR LOCALISATION OF FIH‐1 IMPORTANT? ....... 126 3.4.3 OTHER STUDIES EXAMINING FIH‐1 EXPRESSION IN CANCER ............................................................ 127
4 TWO‐DIMENSIONAL ELECTROPHORESIS ......................................................................... 131
4.1 INTRODUCTION 131
4.1.1 EVIDENCE OF CELL‐SPECIFIC DIFFERENCES IN FIH‐1 ACTIVITY .......................................................... 131 4.1.2 REGULATION OF FIH‐1 IN CANCER? ........................................................................................... 132 4.1.3 IS FIH‐1 POST‐TRANSLATIONALLY MODIFIED? .............................................................................. 133
4.2 METHODS EMPLOYED FOR TWO‐DIMENSIONAL ELECTROPHORESIS 135
4.2.1 OVERVIEW ............................................................................................................................. 135 4.2.2 2‐DE SAMPLE PREPARATION ..................................................................................................... 139 4.2.3 METHODS OF SAMPLE PREPARATION .......................................................................................... 140 4.2.4 PROTEIN QUANTIFICATION ........................................................................................................ 141 4.2.5 ISOELECTRIC FOCUSING ............................................................................................................ 145 4.2.6 EQUILIBRATION ...................................................................................................................... 146 4.2.7 SECOND DIMENSION SEPARATION .............................................................................................. 146 4.2.8 VISUALISATION ....................................................................................................................... 146
4.3 OPTIMISATION OF TWO‐DIMENSIONAL ELECTROPHORESIS 147
4.3.1 SAMPLE PREPARATION METHOD 1 ............................................................................................. 147 4.3.2 SAMPLE PREPARATION METHOD 2 AND PRECIPITATION ................................................................. 151 4.3.3 SAMPLE PREPARATION USING METHOD 2 AND OPTIMISED PRECIPITATION PROTOCOL ......................... 155 4.3.4 DIFFERENT SPOT PROFILES AN ARTEFACT OF SAMPLE PREPARATION ................................................. 161 4.3.4 SUMMARY OF PRELIMINARY HELA EXPERIMENTS .......................................................................... 162
4.4 TWO‐DIMENSIONAL ELECTROPHORESIS RESULTS 163
4.4.1 2‐DE OF MEF LYSATES ............................................................................................................ 164 4.4.2 2‐DE OF COS‐1, 293T AND HELA CELL LYSATES .......................................................................... 169 4.4.3 2‐DE OF 293T, COS‐1, CACO‐2 AND HEPG2 CELL LYSATES ........................................................... 175
4.5 SUMMARY AND DISCUSSION 185
5 PURIFICATION OF FIH‐1 .................................................................................................. 191
5.1 OVERVIEW 191
5.2 IMMUNOPRECIPITATION OF ENDOGENOUS FIH‐1 192
5.3 AFFINITY PULLDOWNS OF FIH‐1 USING NOTCH1 ANKYRIN REPEATS 1‐4.5 198
xxii
5.3.1 STRATEGY OF NOTCH1 ANK1‐4.5‐AFFINITY PURIFICATION ............................................................. 198 5.3.2 RESULTS OF NOTCH1 ANK1‐4.5‐AFFINITY PURIFICATION ............................................................... 201 5.3.3 OPTIMISING ELUTION OF FIH‐1 ................................................................................................. 201 5.3.4 SUMMARY OF NOTCH ANK1‐4.5‐AFFINITY PURIFICATION OF FIH‐1 ................................................. 213
5.4 GENERATION OF STABLE MYC‐6HIS‐HFIH‐1 HELA CELL LINE 213
5.4.1 STRATEGY .............................................................................................................................. 213 5.4.2 GENERATION OF STABLE MYC‐6HIS‐HFIH‐1 HELA POLYCLONAL CELL LINE ........................................ 214 5.4.3 GENERATION OF MYC‐6HIS‐FIH‐1 MONOCLONAL HELA CELL LINES ................................................ 217
5.5 NI2+‐AFFINITY CHROMATOGRAPHY 227
5.5.1 SMALL SCALE NI2+‐AFFINITY PURIFICATIONS ................................................................................. 227 5.5.2 SCALE‐UP OF NI2+‐AFFINITY PURIFICATIONS ................................................................................. 231 5.5.3 OPTIMISATION OF NI2+‐AFFINITY CHROMATOGRAPHY .................................................................... 235 5.5.4 FURTHER OPTIMISATION OF NI2+‐AFFINITY PURIFICATION ............................................................... 240 5.5.5 MASS SPECTROMETRY RESULTS ................................................................................................. 243 5.5.6 CONTINUATION OF FIH‐1 PURIFICATION ..................................................................................... 249 5.5.7 SUMMARY OF NI2+‐AFFINITY CHROMATOGRAPHY ......................................................................... 249
5.7 SUMMARY AND DISCUSSION 250
5.7.1 METHIONINE OXIDATION .......................................................................................................... 251 5.7.2 ASPARAGINYL DEAMIDATION ..................................................................................................... 251 5.7.3 PHOSPHORYLATION ................................................................................................................. 258 5.7.4 FINAL SUMMARY OF FIH‐1 PURIFICATION AND MS RESULTS ........................................................... 261
6 INVESTIGATING POTENTIAL PHOSPHORYLATION OF FIH‐1 ............................................. 265
6.1 OVERVIEW 265
6.2 PHOSPHATASE TREATMENTS 267
6.2.1 PHOSPHATASE TREATMENTS ..................................................................................................... 267 6.2.2 PHOSPHATASE TREATMENT WITH INTERNAL REFERENCE PROTEIN .................................................... 271 6.2.3 SUMMARY OF PHOSPHATASE TREATMENTS .................................................................................. 283
6.3 PHOSPHORYLATION ASSAY 283
6.3.1 STRATEGY FOR IN VITRO PHOSPHORYLATION ASSAY ....................................................................... 283 6.3.2 IN VITRO PHOSPHORYLATION OF TRX‐6HIS‐FIH‐1 ........................................................................ 284 6.3.3 TEV CLEAVAGE OF TRX‐6HIS‐FIH‐1 ........................................................................................... 285 6.3.4 IN SILICO PREDICTION OF PHOSPHORYLATION SITES........................................................................ 295 6.3.5 IN VITRO PHOSPHORYLATION ASSAY WITH SER36 FIH‐1 MUTANTS .................................................. 296 6.4 SUMMARY AND DISCUSSION 299
6.4.1 EVIDENCE IN SUPPORT OF FIH‐1 PHOSPHORYLATION ..................................................................... 299 6.4.2 POSSIBLE SITE OF PHOSPHORYLATION AND INVOLVEMENT OF AKT ................................................... 302
xxiii
7 DISCUSSION AND CONCLUDING REMARKS ..................................................................... 309
7. 1 EXPRESSION OF FIH‐1 309
7.1.1 FIH‐1 EXPRESSION IS NOT DOWN REGULATED IN SOME CANCERS .................................................... 310 7.1.2 SUBCELLULAR LOCALISATION OF FIH‐1 IN BREAST AND NON‐SMALL CELL LUNG CANCER MAY BE IMPORTANT
..................................................................................................................................................... 310 7.1.3 POSSIBLE MECHANISMS LEADING TO ALTERED FIH‐1 LOCALISATION ................................................ 311 7.1.4 POTENTIAL REGULATORS OF FIH‐1 LEVELS IN CANCER ................................................................... 312 7.1.5 TUMOUR PROMOTING ROLE OF FIH‐1 IN RENAL CELL CARCINOMA .................................................. 313 7.1.6 SUMMARY ............................................................................................................................. 314
7.2 REGULATION OF FIH‐1 315
7.2.1 FIH‐1 REGULATION BY HYDROXYLATION STATUS OF THE ARD POOL?............................................... 315 7.2.2 IS FIH‐1 POST‐TRANSLATIONALLY MODIFIED? .............................................................................. 316 7.2.3 2‐DE RESULTS ........................................................................................................................ 317 7.2.4 PURIFICATION AND MS ANALYSIS .............................................................................................. 317 7.2.5 METHIONINE OXIDATION ......................................................................................................... 318 7.2.6 ASPARAGINYL DEAMIDATION .................................................................................................... 319 7.2.7 ACETYLATION ......................................................................................................................... 322 7.2.8 PHOSPHORYLATION ................................................................................................................. 323 7.2.9 EVIDENCE FROM THIS THESIS THAT FIH‐1 IS PHOSPHORYLATED ....................................................... 324 7.2.10 POTENTIAL PHOSPHORYLATION OF SER36 OF FIH‐1 ................................................................... 325 7.2.11 POSSIBLE MECHANISM OF FIH‐1 REGULATION BY SER36 PHOSPHORYLATION .................................. 326 7.2.12 TANKYRASE, INSULIN SENSITIVITY AND POTENTIAL PARSYLATION OF FIH‐1 .................................... 327
7.3 FINAL CONCLUSION 330
8 REFERENCES ................................................................................................................... 335
1
CHAPTER 1
INTRODUCTION
3
1 INTRODUCTION
1.1 Regulation of the Hypoxia Inducible Factors
1.1.1 The necessity of oxygen
Oxygen is critical for aerobic metabolism and as such is necessary for the survival of all
mammals. Energy is taken from metabolites via a series of redox reactions, with
dioxygen acting as the final electron acceptor in oxidative phosphorylation, generating
approximately eighteen times as much adenosine triphosphate (ATP) per unit of glucose
as would be generated by anaerobic glycolysis. The use of oxygen, however efficient
and advantageous, is also a risk. While the majority of electrons stripped from glucose
are finally reacted with oxygen to form water, a subset of these electrons leave the
chain prior to this event and form free radicals, or highly reactive oxygen species (ROS),
such as hydrogen peroxide. ROS can be detrimental to cells as they oxidise
macromolecules, potentially leading to DNA damage and improper cell function.
Deviations from the optimal oxygen concentration of a cell, where oxidative
phosphorylation occurs efficiently and with minimal generation of ROS, are highly
undesirable in either direction, as either an increase or a decrease in oxygen levels can
increase ROS manufacture (Guzy and Schumacker, 2006). Cells must therefore maintain
oxygen concentrations to within a narrow range optimal for ATP production via
oxidative phosphorylation, but also within a range that minimises the production of
harmful ROS.
4
1.1.2 Definitions of normoxia and hypoxia
Despite the absolute requirement for oxygen in metabolic processes, mammalian
tissues have no means of storing oxygen. Rather, there must be a continual supply of
oxygen to each tissue and to each cell that matches the requirement of that cell or
tissue for oxygen. There is no oxygen level universal to all tissues in the body that can be
defined as “normoxia”. Rather, different tissues types, and different regions and cells
within each tissue, behave differently with regards to metabolic activity as they carry
out their particular functions, and therefore require a specific O2 concentration that will
match their demand for oxygen. Thus, normoxia can be defined as occurring when the
supply of oxygen to a cell or tissue is suited to that cells’ or tissues’ demand for oxygen,
or when the oxygen level is within a range optimal for each individual tissue. Hypoxia
can be defined as occurring when the demand and use of oxygen by a cell exceeds the
oxygen supply, or when the supply of oxygen falls short of the demand. Without an
adequate oxygen supply, cells cannot provide the energy required for maintaining
critical processes, and sustained hypoxia leads to cell death in the absence of adaptive
measures that enable a hypoxic cell to alter its behaviour.
1.1.3 Physiological and pathophysiological causes of hypoxia
Inadequate oxygen levels, or hypoxia, can arise either as a result of normal physiological
processes or as a consequence of disease or injury. Physiologically, hypoxia can arise
during periods of elevated respiration such as that encountered during exercise where
the consumption of oxygen by working muscle cells outpaces the supply, and also at
high altitudes when low atmospheric oxygen fails to satiate the body’s requirements
(Reviewed in (Lundby et al., 2009)). Hypoxia is also an important occurrence during
development, where it acts to signal vital processes such as angiogenesis and
maintenance of pluripotency (Lee et al., 2001; Ceradini and Gurtner, 2005).
5
There are numerous pathological situations that give rise to localised or systemic
hypoxia. Situations that restrict blood flow, such as stroke, atherosclerotic plaques and
myocardial infarction prevent adequate delivery of oxygen. Neoplastic disease, where
cell proliferation occurs at faster rate than vascular growth, results in regions of hypoxia
within solid tumours. Hypoxia thus becomes a limiting factor in tumour growth,
resulting in necrotic lesions and necessitating adaptive changes to enable progression of
growth. The role of hypoxia in the development of tumours is further discussed later in
this chapter.
1.1.4 The Hypoxia Inducible Factors
Adaptive mechanisms have evolved at both the organismal and cellular levels to enable
cell survival during periods of low oxygen. All nucleated cells in the body are able to
sense decreases in oxygen levels and maintain homeostasis by mounting an adaptive
response. Of interest to this thesis is the genomic response to hypoxia. This adaptive
response involves the regulation of numerous genes that mediate a variety of cellular
processes which collectively act to address the oxygen deficit, chiefly by both
decreasing oxygen consumption and by increasing oxygen delivery to hypoxic regions.
This genomic response is mediated predominantly by the Hypoxia‐Inducible Factors
(HIFs), transcription factors that directly activate gene expression in response to
decreased intracellular oxygen levels.
The HIFs are assembled from α and β subunits and are transcriptionally active in this
heterodimeric form (Wang et al., 1995). Both subunits are members of the basic helix–
loop–helix/Per‐Arnt‐Sim (bHLH/PAS) homology family of transcription factors which also
includes the aryl hydrocarbon receptor, also known as the dioxin receptor (AHR, DR),
Single Minded 1 and 2 (SIM1 and 2; for review, see (Kewley et al., 2004), and more
6
recently neuronal PAS domain protein 4 (NPAS4; (Ooe et al., 2009). Family members are
characterised by the presence of a bHLH domain critical for binding to DNA and
dimerisation, and a PAS domain also required for dimerisation with other family
members.
The HIF‐β‐subunit, also known as the aryl hydrocarbon receptor nuclear translocator
(Arnt1), is a constitutively expressed nuclear protein. The α‐subunit however is strictly
regulated in an oxygen‐dependent manner. There are three paralogues of the HIF‐α
subunit, HIF‐1α, HIF‐2α and HIF‐3α, and three paralogues of HIF‐1 β (Arnt1, Arnt2 and
Arnt3), with either HIF‐1α or HIF‐2α able to heterodimerise with HIF‐1β to form
functional transcription factor complexes termed HIF‐1 and HIF‐2, respectively.
Whereas HIF‐1 and HIF‐2 are the transcription factors responsible for oxygen‐regulated
change in gene expression, HIF‐3α lacks a transactivation domain and has no known role
in formation of an active transcription factor complex (Gu et al., 1998; Makino et al.,
2001). Exhibiting significant structural and regulatory differences to its HIF‐1α and HIF‐
2α paralogues, a splice variant of the HIF‐3α locus known as Inhibitory PAS domain
protein (IPAS) it is instead postulated to behave as a negative regulator of HIF‐mediated
transcription via binding HIF‐1α and preventing formation of an active transcription
factor complex (Makino et al., 2001).
When HIF‐mediated transcription is required during a period of decreased intracellular
oxygen pressure, HIF‐α receives signals enabling it to become stabilised and active, and
it subsequently translocates to the nucleus to dimerise with its cognate β‐subunit (Kallio
et al., 1998). Once formed, the HIF‐1 and HIF‐2 heterodimers bind to DNA at hypoxia
response elements (HREs), (A/G)CGTG consensus sequences of DNA, located in the
regulatory regions of target genes. Here, transcriptional coactivators must be recruited
for formation of an intact initiation complex. HIF‐1α and HIF‐2α each contain two
transactivation domains; an oxygen‐regulated C‐terminal transactivation domain (CAD,
7
spanning residues 727‐826 of hHIF‐1α) and a more centrally located transactivation
domain designated N‐terminal transactivation domain (NAD, spanning residues 531‐575
of hHIF‐1α (Jiang et al., 1997; Pugh et al., 1997) Although the obligate partner HIF‐1β
has its own well defined C‐terminal transactivation domain, this TAD appears to be
dispensable for transcriptional activation in context of the HIF heterodimer (Jiang et al.,
1996). The HIF‐α‐CAD is thought to serve as the predominant transactivation domain, as
it has been found to regulate the transcription of most HIF target genes (Dayan et al.,
2006). However, there is a small cohort of genes that exclusively require the NAD for
expression, suggesting that there are distinct roles for each transactivation domain in
HIF‐mediated transcription.
1.1.5 HIF Target Genes
HIF‐1 and HIF‐2 are required to modify gene expression during periods of low oxygen
tension to enable cells to adapt to the altered conditions. This is primarily achieved by
increasing blood flow to the hypoxic region or by decreasing the use of oxygen by the
hypoxic cells. Thus, HIF target genes have roles in a variety of cellular processes
including angiogenesis, vasodilation, erythropoiesis, metabolism, and apoptosis (Figure
1.1). Over seventy genes have been determined to be directly upregulated by the HIFs
(Reviewed in (Wenger et al., 2005), however microarray experiments indicate that more
than two hundred transcripts are hypoxically regulated, though some of these may be
indirect targets, or genes regulated in a HIF‐independent manner (Elvidge et al., 2006).
The gene encoding the erythropoietin (EPO) hormone was the first hypoxically
regulated gene identified (Goldberg et al., 1988). EPO acts to counter systemic oxygen
deficiency by enhancing erythrocyte production and thereby increasing the oxygen‐
carrying capacity of blood. Dissection of the EPO enhancer provided the first glimpse
into the structure of the HRE and precipitated the discovery of many additional genes
Angiogenesis VascVEGF
Matrix Metabolism MatrAuto
and Cell MotilityAuto
Erythropoesis Eryth
Glycolysis and Glucose Transport
AldoCarb
Glucose Transport GlucLactaPhos
Proliferation and Survival
TranInsu
Survival
Apoptosis Bcl‐2
FIGURE 1.1 Cellular Processes reguList includes HIF target genes referr
cular Endothelial Growth Factor (VEGF)F Receptor (VEGF‐R; FLT‐1)
rix Metalloprotease 2 (MMP‐2)ocrine motility factor (AMF)ocrine motility factor (AMF)
hropoetin (EPO)
olase Abonic Anhydrase 9 (CA9)cose Transporters 1 and 3 (Glut1 and Glut3)ate Dehydrogenase A (LDHA)sphoglycerate Kinase 1 (PGK1)
sforming Growth Factor‐β (TGF‐β)lin‐like Growth Factor‐2 (IGF‐2)
2/E1B 19 kDa interacting protein (BNip3)
ulated by the Hypoxia Inducible Factorsred to in this thesis
11
found to contain HREs and to be regulated by the HIFs. Such genes include pro‐
angiogenic factors, for example Vascular Endothelial Growth Factor (VEGF) (Forsythe et
al., 1996) and the VEGF receptor FLT‐1 (Gerber et al., 1997) that act together to
increase oxygen transport to hypoxic tissues via the stimulation of blood vessel
branching and the growth of new vasculature. To reduce the demand of a cell for
oxygen, the HIFs also influence cellular metabolism and have been found to upregulate
the expression of glucose transporters GLUT1 and GLUT3 to maximise the amount of
metabolic substrate present in cells (Ebert et al., 1995). Target genes also include a
number of glycolytic enzymes such as aldolase A, phosphoglycerate kinase 1 (PGK1) and
lactate dehydrogenase A (LDHA) to increase glycolysis and enhance the generation of
ATP by this anaerobic mechanism (Semenza et al., 1994).
In addition to the HIF‐mediated regulation of protein‐coding genes, a group of
hypoxically induced microRNAs (miRNAs) have recently been discovered (Kulshreshtha
et al., 2007)(For a review, see (Kulshreshtha et al., 2008). Genes that have previously
been observed to be downregulated in hypoxia may actually be negatively regulated by
hypoxically‐induced miRNAs. To date there are over thirty known hypoxically‐induced
miRNAs, most notably miR‐210, which exhibits the greatest induction at hypoxia and
has been implicated in tumour progression (Huang et al., 2009b). A small number of
these hypoxically‐regulated miRNAs, including miR‐210, have been found to be directly
regulated at by HIF binding to HREs in promoter regions (Kulshreshtha et al., 2007).
1.1.6 Oxygen‐dependent regulation of HIF‐α
The HIFs are responsible for the regulation of a large number of genes involved in a
wide range of important cellular processes. Aberrant expression of these genes can
have dramatic and detrimental effects, as shall be described later in this chapter. Due to
the undesirable effects of inappropriate HIF target gene expression, transcriptional
12
activity must be strictly limited to periods of inadequate oxygen tension. It therefore
follows that the HIFs are subject to oxygen‐dependent regulation, both at the level of
protein stability and transcriptional activity, rendering them largely inactive at normoxia
but potently inducible in hypoxia. This regulation is mediated by four enzymes which
are able to detect changes in intracellular oxygen levels and post‐translationally modify
and regulate HIF‐α accordingly.
1.1.7 Oxygen‐dependent regulation of HIF‐α stability
While HIF‐α is constitutively transcribed and translated in cells, it is rapidly degraded in
normoxia. The half life of HIF‐1α is extremely short in normoxic cells, estimated to be
five minutes in cells grown at 21% oxygen. Degradation of HIF‐1α is rapid enough that
the protein, which is constitutively translated, is rarely detected (Huang et al., 1998).
When cells encounter a deficiency of oxygen however, stabilisation of HIF‐1α protein is
detected in the nuclei of cells after two minutes incubation in hypoxia (Jewell et al.,
2001). The decrease in HIF‐1α protein levels upon reoxygenation occurs just as rapidly
as stabilisation of the protein at hypoxia (Huang et al., 1996; Yu et al., 1998).
This rapid normoxic turnover is facilitated by hydroxylation of specific prolyl residues
within a central oxygen‐dependent degradation domain (ODDD) by HIF‐α ‐specific prolyl
hydroxylases (Figure 1.2) (Ivan et al., 2001; Jaakkola et al., 2001; Yu et al., 2001). There
are three HIF prolyl hydroxylase domain‐containing enzymes, PHD1, 2 and 3, also called
the HIF prolyl hydroxylases (HPHs). These enzymes hydroxylate the conserved prolyl
residues under normoxic conditions, leading to the proteasomal degradation of HIF‐α.
The Von Hippel Lindau protein (pVHL), which is the recognition component of an
ubiquitin ligase, is able to bind the hydroxyproline residues, facilitating
polyubiquitination and thus marking HIF‐α for destruction. As the PHDs require oxygen
for catalytic activity, decreases in oxygen levels reduce activity, enabling HIF‐α to evade
FIGURE 1.2 Oxygen‐dependent regulatAt normoxia the PHD enzymes use O aAt normoxia, the PHD enzymes use O2 athe ODDD of HIF‐. This hydroxylatioubiquitin ligase complex. HIF‐ is subsproteasome. During hypoxia, when oxyhydroxylate HIF‐. HIF‐a is therefore no
tion of HIF‐α stabilityand 2 OG to hydroxylate two prolyl residues withinand 2‐OG to hydroxylate two prolyl residues withinon event recruits pVHL, which is a subunit of asequently polyubiquitinated and degraded by theygen is limiting, the PHDs are unable to efficientlyot ubiquitinated and escapes degradation.
15
modification and destruction, translocate to the nucleus, and bind the β‐subunit to
begin initiation of transcription.
In addition to the oxygen‐dependent regulation of HIF stability and activity, ROS have
also been described to have a role in the regulation of HIF stability via ROS‐mediated
alteration of PHD activity. Mitochondria are a major source of ROS, which are created
when electrons leave the electron transport chain prematurely and react with oxygen to
form the superoxide anion. Particularly, complex III is known to generate superoxide
that is released into the cytoplasm, where it is converted to hydrogen peroxide
(Turrens, 2003). Hydrogen peroxide then acts to reduce activity of the PHDs, possibly by
oxidising the Fe(II) within the enzyme’s active site to Fe(III) (Gerald et al., 2004). Due to
the importance of restricting HIF activity to periods of necessity, it is not improbable
that activity of the HIF hydroxylases be subjected to a variety of regulatory inputs and
controls. The input of ROS on the stabilisation of HIF is likely to occur in combination
with intracellular oxygen levels, supported by the finding that disruption of ROS
manufacture in anoxic conditions does not prevent HIF stabilisation (Schroedl et al.,
2002).
1.1.8 Oxygen‐dependent regulation of the HIF‐α‐CAD transcriptional activity
The HIF‐α‐CAD requires interaction with the transcriptional adaptor proteins CBP/p300
in order to activate target gene expression. CBP and p300 are paralogous transcriptional
coactivators essential for linking HIF with the basal transcriptional machinery. In
addition, CPB/p300 has histone acetyltransferase activity that is needed for
modification of chromatin prior to transcription and is thus indispensable for robust
transcriptional activation. At normoxia, the CAD is repressed by a hydroxylation event
that prevents its interaction with CBP/p300. Of the two TADs present in HIF‐α, only the
CAD is subjected to this oxygen‐dependent regulatory event.
16
The hydroxylation and repression of the HIF‐α‐CAD is mediated by the asparaginyl
hydroxylase, Factor Inhibiting HIF‐1α (FIH‐1) (Lando et al., 2002b; Lando et al., 2002a;
Mahon et al., 2001; Hewitson et al., 2002). Asparaginyl hydroxylation was first found to
occur within the CAD of HIF‐2α. Mass spectrometry (MS) of HIF‐2α CAD fragments
purified from hypoxic and normoxic cells revealed an oxygen‐dependent hydroxylation
event on Asn851 (corresponding to Asn803 in hHIF‐1α) (Lando et al., 2002b). This
modification was found to impede interaction with the CH1 domain of p300, revealing
the mode by which HIF‐mediated transcription was regulated in response to changes in
oxygen availability. FIH‐1 was first identified in a yeast two hybrid (Y2H) screen as
interacting with HIF‐1α (Mahon et al., 2001) and was later shown to be the asparaginyl
hydroxylase responsible for this so called “hypoxic switch” in HIF transcriptional activity
(Hewitson et al., 2002; Lando et al., 2002a).
Like the PHDs, FIH‐1 requires oxygen for catalysis of hydroxylation. At normoxia
therefore, FIH‐1 is able to hydroxylate and thus repress HIF‐α activity by preventing
CBP/p300 binding. As intracellular oxygen levels decrease, FIH‐1 activity also decreases,
allowing HIF‐α to escape modification, retain CBP/p300 binding and activate expression
of HIF target genes that are required to address the oxygen deficit (Figure 1.3).
Structural studies of the FIH‐1/HIF‐1α‐CAD interaction show that the target asparaginyl
side chain is deeply buried within a hydrophobic pocket of the hHIF‐1α‐CAD/p300
interface and the attachment of a hydroxyl group at this position is energetically
unfavourable, thereby preventing the interaction (Freedman et al., 2002). Fluorescence
polarisation binding assays confirm the disruption of binding by hydroxylation,
demonstrating almost complete abrogation of HIF‐α‐CAD‐p300 interaction when the
CAD is hydroxylated (Cho et al., 2007).
FIGURE 1.3 FIH‐1‐mediated regulationAt normoxia, FIH uses the co‐substratesresidue within the HIF‐‐CAD, releasprecludes association with the esstranscriptional activity of the CAD. Durto efficiently catalyse the hydroxylationnon‐hydroxylated CAD and transactivat
of HIF (Taken from Lisy and Peet, 2008)s O2 and 2‐OG to hydroxylate the target asparagineing CO2 and succinate. This hydroxylation eventsential coactivators CBP/p300, repressing theing hypoxia, when oxygen is limiting, FIH is unablen of the CAD, enabling binding of CBPp300 to theion of target genes
19
1.2 Factor‐Inhibiting HIF
1.2.1 FIH‐1 expression and subcellular localisation
FIH‐1 seems to be ubiquitously expressed, with FIH‐1 protein expression detected at
similar levels in all tissue culture cell lines investigated to date (Stolze et al., 2004;
Bracken et al., 2006). In addition, immunohistochemical interrogation of an extensive
range of human tissues revealed widespread FIH‐1 protein expression (Soilleux et al.,
2005).
During normoxia, FIH‐1 is localised predominantly in the cytoplasm. Treatment of
cultured cells with hypoxia or hypoxia mimetics does not appear to alter the
cytoplasmic localisation of FIH‐1 (Metzen et al., 2003a; Linke et al., 2004; Stolze et al.,
2004). Subcellular localisation of FIH‐1 in tissues follows a similar trend, with FIH‐1
expression detected mainly in the cytoplasm of cells with lower levels detected in nuclei
(Soilleux et al., 2005).
1.2.2 FIH‐1 Structure
X‐ray crystallography has revealed the structure of human FIH‐1 as being similar to that
of other known 2‐oxoglutarate (2‐OG)‐dependent hydroxylases (Dann et al., 2002). FIH‐
1 is made up of a mixture of β‐strands and α‐helices and contains an eight‐stranded β‐
strand “jellyroll” core typical of other 2‐OG‐dependent oxygenases. The active site of
the enzyme is positioned within this core, with the Fe(II) atom required for binding
dioxygen coordinated at the centre by His199, Asp201 and His279. At the periphery of
the jellyroll lie eight α‐helices, with two α‐helices formed by the C‐terminus, creating an
interface for homodimerisation. Formation of a FIH‐1 homodimer is a requirement for
efficient substrate recognition and catalytic activity. Deletion of these two C‐terminal
20
helices precludes dimer formation and impairs hydroxylation of the HIF‐α‐CAD (Dann et
al., 2002). Likewise, disruption of homodimer formation by replacement of a
hydrophobic leucine residue with a hydrophilic arginine residue within the hydrophobic
FIH‐1 homodimer interface also impairs dimer formation and markedly reduces FIH‐1
activity in vitro (Lancaster et al., 2004).
1.2.3 Regulation of FIH‐1 activity by oxygen levels: In vitro data
Though critical in actioning the genomic response to hypoxia, the HIF transcription
factors themselves are not capable of sensing changes in intracellular oxygen
concentration. Any deviations in oxygen levels in cells are detected by the four
hydroxylases that regulate the stability and activity of HIF‐α, collectively termed the HIF
hydroxylases. The hydroxylases are all members of the Fe(II), 2‐OG and O2‐dependent
dioxygenase family. Their catalytic activity involves the splitting of diatomic oxygen,
with one oxygen atom used for hydroxylation of the target residue and the other for
oxidation of the cofactor 2‐OG, which is further decarboxylated to succinate, releasing
CO2. The direct requirement of these enzymes for molecular oxygen to modify their
substrates supports the notion of their behaving as cellular oxygen sensors that exhibit
a decline in catalytic activity concomitant with a decrease in oxygen availability.
In vitro‐determined affinities of the HIF hydroxylases for oxygen have helped gain some
insight into the level of hypoxia required for loss of catalytic activity and subsequently
for stabilisation and activation of HIF‐α. However, these studies are confounded by the
different values obtained when using peptide substrates of different lengths and
variations in experimental conditions, making it difficult to relate these values to a real
in vivo setting.
21
In brief, the earliest studies addressing HIF hydroxylase O2‐affinity in vitro with HIF
peptides assigned FIH‐1 with an approximately 2.5‐fold greater affinity for oxygen than
the PHDs, with values of approximately 90 μM for FIH‐1 (Koivunen et al., 2004) and 230‐
250 μM for PHD1‐3 (Hirsila et al., 2003), and used substrate peptides of 35 and 19
amino acyl residues, respectively. These data suggested that activity of the PHDs may
decrease prior to a decline of FIH‐1 activity when oxygen levels drop, and consequently
that HIF stabilisation precedes CAD activation and occurs at relatively moderate
hypoxia, with full activity achievable with inactivation of FIH‐1 at more severe hypoxia.
However, more recent data generated using different length substrates and a different
method argue against this model, indicating that the length of peptide substrate used
influences the determined Km values (Ehrismann et al., 2007). In these studies, there is
little difference in the O2 affinity of FIH‐1 and the PHDs in vitro, with values of 110‐237
μM and 76‐229 μM for FIH‐1 and the PHD2, respectively, using HIF‐1α substrate
peptides of varying lengths. Similarly, it has been reported that use of longer HIF‐α
peptides enhances the oxygen‐affinity of the PHDs (Koivunen et al., 2006). As enzyme‐
protein substrate binding is thought to precede binding of oxygen, various changes in
PHD conformation that occur upon binding to substrates of different lengths that may
alter the enzyme’s affinity for oxygen. As such, the value of in vitro data generated using
peptides in regards to predicting the behaviour of the HIF hydroxylases in vivo with full
length substrates is limited. Collectively however, the data do suggest that the
hydroxylases are able to respond to changes in physiological oxygen levels and thus act
as oxygen sensors.
1.2.4 Regulation of FIH‐1 activity by oxygen levels: data from cell‐based experiments
Cell‐based experiments have both supported and contradicted the in vitro‐predicted
order for loss of hydroxylase activity, and also highlighted the difficulty in extrapolating
information regarding the behaviour of full length folded proteins in a cellular
22
environment from data derived from short purified peptides in vitro (Stolze et al., 2004),
(Bracken et al., 2006).
In cells, FIH‐1 has been observed to retain activity even at severe hypoxia (<0.2% O2)
(Stolze et al., 2004). siRNA‐mediated knockdown of FIH‐1 in U2OS cells grown in hypoxic
(1% O2) conditions resulted in the upregulation of HIF targets GLUT1 and LDHA
message, and to a lesser extent an increase in VEGF mRNA (Stolze et al., 2004).
Correspondingly, overexpression of wt FIH‐1 in U2OS cells at slightly more moderate
(2% O2) and severe hypoxia (0.5% and 0.2% O2) resulted in downregulation of GLUT1,
LDHA and VEGF mRNA, indicating that FIH‐1 is able to exert a repressive effect on HIF‐
mediated transcription even at low oxygen levels. In contrast, overexpression of PHD2
had no effect upon HIF target gene expression at the same oxygen concentrations,
indicating that FIH‐1 is indeed able to retain activity even when oxygen levels are low
enough for significant loss of PHD2 activity. This is in agreement with the in vitro‐
derived Kms of the HIF hydroxylases for oxygen reported by Koivunen et al., 2004 and
Hirsila et al., 2003.
Use of an antibody raised against Asn803‐OH HIF‐1α, however, seems to indicate that
FIH‐1 is not hydroxylating HIF‐1α substrates in cells cultured in 0.5% O2 (Shin et al.,
2009). This study investigated HIF‐α hydroxylation in HEK293 cell lysates after 16 hours
culture at different oxygen levels. At normoxia, HIF‐1α protein levels were low, thus
hydroxylated protein was difficult to detect. At 1% O2 however, total HIF‐1α protein was
observed and a band detected with the hydroxylation‐specific antibody. In contrast to
the findings in U2OS cells detailed above, at 0.5% O2, HIF‐1α protein was abundant, yet
there was no hydroxylation detected in the western blots. These results suggested that
while FIH‐1 remained active at 1% O2, it lost activity in more severe oxygen deprivation
(0.5% O2), at least in this cell type under these culture conditions.
23
Investigation into the regulation of FIH‐1 activity by oxygen levels in different cell types
has suggested that FIH‐1 activity levels may depend on the cellular context (Bracken et
al., 2006). A set of experiments examining the activity of endogenous FIH‐1 and the
PHDs involved the incubation of six different cell lines expressing GalDBD‐HIF‐α‐CAD in a
range of atmospheric oxygen concentrations. FIH‐1 activity was inferred by expression
of a Gal‐Luciferase reporter (Figure 1.4), and PHD activity was indicated by
immunoblotting for HIF‐α protein to determine stabilisation. In HepG2, Caco‐2 and PC12
cells, HIF‐α became stable prior to detection of any reporter gene expression, indicating
reduced PHD activity prior to reduced FIH‐1 activity in conditions of decreasing oxygen
levels, in keeping with the data that assigned a higher Km for molecular oxygen to the
PHDs over FIH‐1 (Figure 1.4 a, b and c). In these cell lines, reporter activity was not
observed until cells were exposed to oxygen concentrations of less than 1%. In Cos‐1
and 293T cells however, HIF‐α stabilisation and reporter activity were observed at the
same oxygen level, and in HeLa cells reporter activity was observed far before
endogenous HIF‐α became stable (Figure 1.4 d, e and f). FIH‐1 protein levels were
consistent in all six cell lines, consistent with changes in reporter activity being a
product of differences in FIH‐1 activity rather than amount.
Importantly, while stabilisation of endogenous HIF‐1α was seen at 2% oxygen for all cell
types, the oxygen threshold required for activation of the Gal‐Luciferase reporter
varied, with loss of activity of FIH‐1 observed to occur at <1% oxygen in PC12, Caco‐2
and HepG2 cells, 5% oxygen in 293T cells, and 10% oxygen in HeLa cells. In other words,
these findings suggested that FIH‐1 is less sensitive to decreasing oxygen concentrations
in some cell types (PC12, Caco‐2, HepG2) and more sensitive to decreases in oxygen in
other cell types (HeLa, 293T and Cos‐1). These data suggest the existence of undefined,
cell‐specific mechanisms that contribute to regulation of FIH‐1 activity levels. For
example, the decreased activity of FIH‐1 observed in HeLa, 293T and Cos‐1 cells at
oxygen concentrations thought to be sufficient for efficient FIH‐1 catalysis could arise
0
20%
FIGURE 1.4 Transactivational activity in mammalian cells (Figure adapted fHeLa (a), 293T (b), Cos‐1 (c), PC12 (d), with 100 ng of GalDBD‐HIF‐CTAD expreof the RLTK internal control. Post‐transnormoxia or for 6 hours in normoxia foO2 to 0.5% O2 as indicated). Cells were(DP)
O2 Concentration
of the HIF‐α‐CAD in various oxygen concentrations rom Bracken et al 2006.)Caco‐2 (e) and He pG2 (f) cells were transfected pession vectors, 150 ng 5GalRE‐luciferase and 10 ng sfection, cells were incubated for 20 hours in ollowed by 14 hours at various oxygen levels (10% e also treated with the 1 mM iron chelator dipyridyl
29
from other modes of regulation that act to alter FIH‐1 activity. These could act to either
alter the affinity of FIH‐1 for oxygen or other cofactors important for catalysis of
hydroxylation, or affinity or interaction with substrates.
In summary, the data from cell‐based studies suggest that there may be other
mechanisms exerting regulatory control over FIH‐1 in some cell types, in addition to
regulation by available oxygen.
1.3 ARD‐containing substrates
1.3.1 Does FIH‐1 post‐translationally modify other proteins?
Since the elucidation of the hydroxylase‐mediated mode of HIF‐α regulation, there has
existed the important question of whether the hydroxylases have any other substrates.
At the commencement of this project, knowledge of the role of FIH‐1 in cells was
limited to its regulation of the HIF‐α proteins. There were, however, numerous efforts
underway, in the Peet lab and others, to investigate the possibility that additional
interacting proteins or substrates for FIH‐1 could exist, and there was some suggestion
in the literature that the role of the HIF hydroxylases extended beyond modification of
the HIF‐α substrates. For example, microarray data identified a small number of
transcripts affected by hydroxylase inhibition independently of HIF (Elvidge et al., 2006).
1.3.2 Identification of novel FIH‐1 substrates
The first such report of novel FIH‐1 substrates identified IκB proteins p105 and IκBα as
FIH‐1 substrates (Cockman et al., 2006). A Y2H screen enabled isolation of these novel
FIH‐1‐interacting proteins, which were subsequently found to be potential FIH‐1
30
substrates in vitro by CO2 capture assays. Hydroxylation was confirmed by MS of
proteins immunopurified from 293T cells.
1.3.3 Notch is an FIH‐1 substrate
Another good candidate for a non‐HIF substrate of FIH‐1 was the Notch receptor. Notch
is a transmembrane receptor that has long been known to have a role in cell fate
decisions in many different tissues, usually resulting in the blockage of differentiation
and maintenance of an undifferentiated cell fate (For review, see (Bolos et al., 2007).
Canonical Notch signalling involves the activation of the Notch receptor by
transmembrane ligands Delta and Jagged that are expressed on adjacent cells. This
interaction results in proteolytic cleavage events that release the Notch intracellular
domain (ICD) from the membrane, allowing for translocation of Notch ICD to the
nucleus and subsequent regulation of gene expression in concert with partner protein
CBF‐1/suppressor of hairless/Lag‐1 (CSL). The role of Notch signalling is to switch on the
transcription of target genes such Hes and Hey in order to facilitate inhibition of
differentiation. Studies had previously linked hypoxia with regulation of differentiation,
demonstrating that hypoxia influenced proliferation and differentiation of neural crest
stem cells in culture (Morrison et al., 2000) and promoted dedifferentiation of cells in
tumours (Jogi et al., 2002).
Gustafsson et al (2005) found a direct interaction between HIF‐α and Notch (Gustafsson
et al., 2005). In Notch‐responsive reporter assays, co‐transfection of HIF‐α and Notch
into P19 cells was required activation of the reporter gene. Interestingly, overexpression
of FIH‐1 led to a reduction in luciferase readout in these reporter assays, both when HIF‐
1α and Notch were coexpressed, but also in the absence of HIF‐1α at both normoxia
and hypoxia, suggesting a HIF‐independent mode of Notch regulation. While it was
possible that FIH‐1 was reducing activity of the reporter via repressive hydroxylation of
31
endogenous HIF‐1α, the low levels of HIF‐1α protein at normoxia argued against this
idea. Furthermore, an interaction between FIH‐1 and the Notch ICD was observed,
suggesting a direct mode of regulation of Notch by FIH‐1 (Gustafsson et al., 2005).
Following on from these findings, Sarah Linke investigated the possibility that Notch was
a substrate for FIH‐1. FIH‐1‐mediated hydroxylation of Notch was first suggested by
activity of Notch 1 constructs Notch Ank1‐7 and full length Notch ICD in CO2 capture
assays, and confirmed by MS of human Notch 1 Ank1‐7 purified from FIH‐1
overexpressing 293T cells (Zheng et al., 2008). MS revealed two sites of hydroxylation;
Asn1945 and Asn2012, with Asn1945 hydroxylated to a greater extent than Asn2012. At
this time an independent report of Notch hydroxylation appeared confirming the
findings of S. Linke (Coleman et al., 2007). This research confirmed FIH‐1‐mediated
hydroxylation of the two Notch 1 asparaginyl residues in cells, with Asn1945 being the
major site of hydroxylation. Mammals have four Notch proteins, Notch 1, 2, 3 and 4,
and though all are able to interact with FIH‐1, only Notch 1, 2 and 3 are hydroxylated
(Zheng et al., 2008) (Wilkins et al., 2009).
1.3.4 Function of Notch hydroxylation
Experiments aimed at delineating the effect of the Notch hydroxylation by FIH‐1 have
thus far returned ambiguous results. Mutation of the target asparaginyl residues in
Notch1 attenuates its activity in Notch‐responsive reporter assays and also reduces the
ability of Notch1 to block neuronal and myogenic differentiation, suggesting that these
residues are indeed important for Notch activity. However, overexpression of FIH‐1 also
leads to a decrease in Notch activity and a corresponding increase in differentiation, and
this effect occurs independent of the catalytic activity of FIH‐1 (Zheng et al., 2008).
32
Overexpression of either wt FIH‐1 or a D201A catalytically inactive FIH‐1 mutant both
reduce activity of a Notch‐responsive reporter (Coleman et al., 2007), also suggesting
that the inhibitory effect of FIH‐1 overexpression on Notch‐responsive reporter gene
expression occurs independently of hydroxylation, as observed previously (Gustafsson
et al., 2005). An effect of endogenous FIH‐1 on Notch activity has also not been directly
demonstrated, with FIH‐1 siRNA failing to alter Notch‐responsive reporter activity,
endogenous Notch activity, or Notch1 interaction with known binding partners
(Coleman et al., 2007). As yet, there is no defined role for the observed hydroxylation of
Notch by FIH‐1.
1.3.5 Ankyrin Repeat Domain‐containing substrates
FIH‐1‐mediated hydroxylation of Notch, p105 and IκBα substrates occurs on asparaginyl
residues located within ankyrin repeat domains (ARDs). ARDs are composed of varying
numbers of individual ankyrin repeats, where one repeat is made up of a thirty three
residue motif folded into anti‐parallel α‐helices linked by a short loop, and individual
repeats are connected via a β‐hairpin loop. The fact that this motif was present in each
of the newly identified substrates suggested that ARD hydroxylation may be a common
event. Indeed, sequence alignment of a number of ARD‐containing proteins, including
Tankyrase, Gankyrin, Myosin Phosphatase targeting subunit 1 (MYPT1), Myotrophin,
Integrin‐linked kinase‐1 (ILK‐1) and Fetal Globin Inducing Factor (FGIF) in addition to
p105 and IκBα, indicated that a conserved asparaginyl residue located at an equivalent
position within the ankyrin fold was common to each of the proteins. CO2 capture
assays suggested that peptide fragments from each of these proteins were indeed able
to elicit FIH‐1 activity in vitro (Cockman et al., 2006).
To date, the list of confirmed FIH‐1 substrates includes, in addition to HIF‐1α and HIF‐
2α: Notch1, 2 and 3 (Coleman et al., 2007; Zheng et al., 2008), p105 and IκBα (Cockman
33
et al., 2006), Ankyrin Repeat and SOCS Box Protein 4 (ASB4) (Ferguson et al., 2007) ,
Rabankyrin‐5, RNaseL, Tankyrase‐1 and 2 (Cockman et al., 2009) and MYPT1 (Webb et
al., 2009). This is in addition to a number of other ARD‐containing proteins that have
thus far only been suggested to be substrates by in vitro assays.
Aside from the HIF‐α substrates, the recently identified substrates all contain ARDs.
These domains are common among the human proteome, present in an estimated
three hundred proteins, and included in approximately 6% of all eukaryotic proteins
(Barrick et al., 2008). Asparaginyl hydroxylation may therefore be a common post
translational modification and it is likely that more FIH‐1 substrates in addition to those
listed above will be discovered in the near future.
1.3.6 Function of ARD hydroxylation
1.3.6.1 Roles of ARD substrates
The ARD substrates of FIH‐1 are involved in a diverse range of cellular activities,
including immune and inflammatory responses (NFκB), cell fate decisions (Notch
signalling), ubiquitin‐mediated proteolysis (ASB4), actin−myosin contrac lity (MYPT1),
endocytosis (Rabankyrin) and telomere regulation and vesicle trafficking (Tankyrase‐1
and ‐2). At this stage however, the function of hydroxylation and the impact of this
modification upon these cellular processes remains unknown, with only subtle
downstream effects, if any, observed.
Provided that ARD hydroxylation is found to have a significant roles, both the sheer
number of confirmed and potential substrates, taken together with the diversity in
function of these substrates, opposes the idea that FIH‐1 regulates these substrates in a
binary on/off mode similar to its regulation of HIF‐α. It also seems unlikely that all of
these processes would be similarly regulated by oxygen levels, promoting the possibility
34
of FIH‐1 either exerting an effect independent of catalytic activity, or of FIH‐1 activity
being regulated by multiple signals or stimuli. Indeed, oxygen‐dependent regulation of
FIH‐1 activity may not play a significant role in the hydroxylation of non‐HIF‐α
substrates. It has been shown that the Km of FIH‐1 for oxygen changes depending on its
substrate (Wilkins et al., 2009). With Notch1 as a substrate, the Km of FIH‐1 for oxygen is
approximately an order of magnitude lower (12 μM +/‐ 3 μM) than of FIH‐1 with HIF‐1α
as a substrate (90 μM +/‐ 25 μM), suggesting that FIH‐1‐mediated hydroxylation of
Notch will persist at oxygen levels that preclude hydroxylation and repression of HIF‐1α
(Wilkins et al., 2009). Whether such hydroxylation events are constitutive or regulated
by other, as yet unidentified signals, is currently not known.
1.3.6.2 Role of ARD hydroxylation for protein stability and structure
One suggested role for FIH‐1‐mediated hydroxylation of ARD substrates is that
hydroxylation contributes to the stability of ARDs. Asparaginyl hydroxylation was
postulated to strengthen the stability of the ARD fold via the formation of a hydrogen
bond between the hydroxyasparagine and the residue positioned two residues N‐
terminal of the asparagine (Coleman et al., 2007). However, despite the formation of a
hydrogen bond within the ankyrin fold upon hydroxylation, structural analysis of have
revealed that there is no major change to the structure of the hydroxylated compared
to unhydroxylated Notch1 ARD. Biophysical data supporting the hydroxylation‐induced
increase in ARD stability includes comparison of unfolding temperatures of
hydroxylated and non‐hydroxylated synthetic peptides, with hydroxylated peptides
having a slightly increased thermal stability. Mutation of the ‐2 amino acid residue,
usually an aspartic acid, to alanine prevented the formation of the hydrogen bond and
also reduced this increase in stability (Hardy et al., 2009). Earlier studies of IκBα
suggested that subtle changes in stability of its ARD affected the proteins ability to
interact with other partners (Truhlar et al., 2008), however more recent data have
35
contradicted this finding, showing no alteration in IκBα structure, stability, or
interaction with NFκB (Devries et al., 2010). Thus, any role for hydroxylation of ARDs by
FIH‐1 in protein stability remains controversial.
1.3.6.3 Role of ARD hydroxylation in fine‐tuning the hypoxic response
A further possibility for the role of ARD hydroxylation in cells is that competition
between FIH‐1 substrates for binding and hydroxylation may serve to contribute to the
regulation of HIF‐1α transcriptional activity. This theory is supported by several
independent studies reporting competition between HIF‐α and ARD substrates for FIH‐
1. Firstly, Notch and IκBα substrates bind FIH‐1 with a far greater affinity than HIF‐α, in
the case of Notch a 50‐fold higher affinity than that determined for HIF‐α (Wilkins et al.,
2009). Crystal structures have shown that Notch1 and HIF‐α peptides bind to largely
overlapping sites on FIH‐1, sustaining the idea that competition occurs between the two
substrates (Coleman et al., 2007). Furthermore, co‐expression of the Notch1 ICD with
the HIF‐1α‐CAD in cells dramatically reduces hydroxylation of the CAD, and
overexpression of Notch1 and Notch3 lead to an increase in Gal4DBD‐HIF‐CAD reporter
activity (Coleman et al., 2007). In agreement with these data, wt Notch 1 ICD was shown
in a separate study to enhance activity of a CAD reporter in a dose dependent manner,
an effect that was rescued by overexpression of FIH‐1 (Zheng et al., 2008). Together,
these results suggest a subtle mode of HIF regulation where Notch sequesters FIH‐1
away from HIF‐1α, thus protecting it from repression by FIH‐1. This effect is not limited
to Notch, as transfection of MYPT1 was also found to enhance transcriptional activity of
the HIF‐α‐CAD (Webb et al., 2009), and modest increases in transcriptional activity were
also observed with IκBα and RNaseL overexpression (Webb et al., 2009). siRNA‐
knockdown of endogenous IκBα was also found to have the reciprocal effect of reducing
HIF‐α activity (Shin et al., 2009).
36
Taken together, these data suggest interaction with ARD containing proteins may
reduce availability of FIH‐1 for HIF‐α hydroxylation, and therefore contribute to the
regulation of HIF target gene expression.
1.4 What is the Role of FIH‐1?
Given the large number of substrates now known for FIH‐1 and the wide array of
cellular processes these substrates are involved in, a fundamental question remains:
what is the preeminent role of FIH‐1?
Knockout mouse models of each of the three PHDs have been generated and have
provided insight into their contributions in regulation of HIF‐α stability (Takeda et al.,
2006), (Aragones et al., 2008), (Bishop et al., 2008). PHD2 null mice exhibit the most
severe phenotype, dying between 12.5 and 14.5 days post coitum with severe heart and
placental defects, a phenotype consistent with impaired HIF‐α regulation and in keeping
with the dominant role of PHD2 over PHD1 and PHD3 in hydroxylation and therefore
regulation of HIF‐1α protein levels.
1.4.1 FIH‐1 knockout mouse phenotype
In order to discern a precise role of FIH‐1 in cells, FIH‐1 null mice were generated by
Zhang et al (Zhang et al., 2010a). Published mid 2010, this research delivered surprising
results, as discussed below. FIH‐1‐/‐ mice did not present with a phenotype consistent
with general elevated HIF activity. The mice survived into adulthood and displayed no
major alteration in processes known to be regulated by the HIF pathway such as
erythropoiesis and angiogenesis. Furthermore, FIH‐1‐/‐ mouse embryonic fibroblasts
(MEFs) exhibited only small increases in expression of HIF target genes VEGF and PGK1
37
at both normoxia and at 1% atmospheric oxygen compared to wildtype MEFs, and there
was minimal to no difference between expression GLUT1 between wildtype and FIH‐/‐
cells. Curiously, deletion of both FIH‐1 and VHL from MEFs resulted in the dramatic
(~200‐fold) increase carbonic anhydrase 9 (CA9) mRNA, when only a modest increase
was observed with FIH‐1 deletion or VHL deletion alone, suggesting that FIH‐1 exerts
significant control over some genes in the absence of VHL. Other genes also
significantly upregulated by the dual deletion, including endothelin 2 (~100 fold).
FIH null mice exhibited a hypermetabolic phenotype, observable by differences
between wt and FIH‐1‐/‐ animals across a range of metabolic parameters. Whereas
activation of HIF in animals has previously been found to reduce oxygen consumption
and affect metabolism by decreasing oxidative phosphorylation and upregulating
glycolysis (Aragones et al., 2008), loss of FIH‐1 here was found to increase oxygen
consumption but to have no effect of glycolysis at normoxia. This seems to suggest that
any changes in metabolic rate were not simply due to upregulation of the HIF pathway,
but rather due to as yet undefined processes in which FIH‐1 has a part. The increased
respiration in FIH‐1‐/‐ mice was coupled with an increased heart rate and heat
production, and FIH‐1‐/‐ MEFs showed greater ATP levels and decreased levels of active
AMP‐activated protein kinase (AMPK) compared to wt MEFs.
Chronic hyperventilation was also observed in FIH‐1‐/‐ mice. Null mice exhibited a tidal
volume at normoxia similar to the increased tidal volume of wildtype animals at
hypoxia, suggesting that FIH‐1 has a role in regulating tidal volume in response to
changing atmospheric oxygen. Interestingly, when the mice were placed at hyperoxic
conditions (30% O2), their tidal volumes decreased to levels found in wildtype mice in
normoxic conditions, suggesting an overall shift in what the null mice perceived to be
normoxic oxygen levels. How FIH‐1 contributes to ventilory control is not clear.
38
Lastly, FIH‐1 null animals showed increased consumption of food and water, in keeping
with their elevated metabolic rates. However, despite this hyperphagia and also a
decrease in physical activity, FIH‐1‐/‐ mice were smaller at birth and continued to have a
smaller body size throughout adulthood compared to wildtype mice. This lower body
weight was attributed to less adipose tissue rather than decreased size of other body
structures such as muscles or organs. Loss of FIH‐1 also conferred a dramatically
increased sensitivity to insulin and subsequent protection against weight gain when the
mice were fed a high‐fat diet. Neuron‐specific deletion of FIH‐1 phenocopied these
global FIH‐1‐/‐ mice, suggesting that FIH‐1‐mediated regulation of global metabolism
occurs through the activities of FIH‐1 in neurons.
1.4.2 A cell‐type specific role for FIH‐1?
A cell type specific role for FIH‐1 has not yet been defined but may exist. Expression of
the HIF substrates is ubiquitous, and FIH‐1 itself is also expressed in all tissue
investigated to date (Soilleux et al., 2005). In the absence of any cell‐specific regulatory
pathways that regulate FIH‐1 activity, it would be logical to expect that FIH‐1 activity
would proceed in a similar manner across all cell types. However, this data clearly
suggests a novel, neuronal role for FIH‐1 in the regulation of metabolism. As the
expression of HIF target genes Vegf, Glut‐1, Bnip3 and Ca9 in neuronal FIH‐1‐/‐ mice
were unchanged, it is probable that FIH‐1 here is not acting through alterations in HIF‐
mediated transcription.
In summary, typical phenotypes relating to the upregulation of the HIF pathway were
not observed in FIH‐/‐ mice. Rather, these data suggest that FIH‐1 additionally, and
perhaps primarily, acts as a neuronal regulator of metabolism in mammals. In addition,
FIH‐1 seems to be involved in respiratory regulation, specifically in the nomination of an
oxygen “set point” that an animal considers normoxia. Particularly interesting is the
39
concept that FIH‐1 is regulated in a cell specific manner by as yet unidentified
mechanisms, and exerts control over important cellular and global processes via
mechanisms that may be HIF‐independent.
In addition, these findings increase the therapeutic potential of manipulating FIH‐1
activity, with the implication being that inhibition of FIH‐1 may be a strategy employed
for reversal of insulin resistance, type II diabetes and possibly aid in reduction of diet‐
induced obesity. The importance of FIH‐1 in these processes and the ambiguity
surrounding the nature and mechanisms behind its involvement validate the need for
further research into FIH‐1 function, activity and regulation.
1.5 Regulation of FIH‐1
At the commencement of this project, limited data was available regarding regulation of
FIH‐1 beyond regulation of enzyme activity by oxygen levels. Recent reports have
illuminated novel mechanisms that regulate FIH‐1 expression and activity, discussed
below.
1.5.1 Regulation at the FIH‐1 promoter by Protein Kinase C
Negative regulation of FIH‐1 transcription by protein kinase C (PKC) has been
demonstrated in renal cell carcinoma (RCC) (Datta et al., 2004). Here, qPCR revealed a
dose‐dependent increase in FIH‐1 message in 786‐O cells with introduction of a
dominant negative form of PKC. This was followed up by further data suggesting a
specific mechanism for decreased FIH‐1 message (Li et al., 2007b). Li and colleagues
identified a cis regulatory element in the FIH‐1 promoter that binds phosphorylated
CCAAT‐displacement protein (CDP), leading to repression of FIH‐1 transcription. PKC is
40
the kinase responsible for CDP phosphorylation, and this phosphorylation event was
demonstrated to be important for DNA binding by CDP and the subsequent repression
of FIH‐1. As a side note, expression of the fly CDP homologue Cut has, through several
independent lines of evidence, been shown to be positively regulated by Notch in flies
and to co‐localise with Notch in mice (Nepveu, 2001). Given that FIH‐1 represses Notch
activity and CDP/Cut represses FIH‐1 transcription, it is interesting to speculate whether
upregulation of CDP/Cut by Notch could represent a positive feedback loop for Notch
activity.
As RCC is VHL deficient, the transcriptional repression of FIH‐1 via the PKC/CDP
mechanism may represent a means to activate HIF‐target gene expression in RCC, and
this mechanism may be particularly important in RCC in light of the results of the FIH‐1‐/‐
knockout studies described above that found that FIH‐1 dramatically reduces certain HIF
target genes in the absence of VHL. The existence of this mechanism in other cell types,
or in normal tissues, is unknown.
1.5.2 Regulation of FIH‐1 translation by microRNA 31
miR‐31 has been reported to attenuate FIH‐1 expression in head and neck squamous
cell carcinoma (HNSCC), with a decrease in FIH‐1 message and protein observed upon
miR‐31 overexpression in HNSCC cell line SAS (Liu et al., 2010). In this study, miR‐31 was
found to be upregulated in HNSCC, and ectopic expression of miR‐31 increased
oncogenic traits such as proliferation, migration and anchorage‐independent growth of
SAS cells. Importantly, miR‐31 expression has been reported to be elevated in other
cancer types, including colorectal carcinoma (Slaby et al., 2007) and tongue carcinoma
(Wong et al., 2008). Upregulation of miR‐31 in specific tumour types may represent a
mechanism for downregulation of FIH‐1 level and thereby a mechanism for increased
41
normoxic HIF‐target gene expression that may be important in progression of some
cancer types.
Contrarily, miR‐31 expression is decreased in metastatic breast cancer. Here,
overexpression of miR‐31 in the breast cancer cell line MDA‐MB‐231 resulted in
decreased invasion and motility in in vitro assays, and decreased metastasis in vivo
(Valastyan et al., 2009). This was found to occur via the downregulation of expression of
a number of pro‐metastatic mRNAs by miR‐31. The effect of miR‐31 on FIH‐1 expression
was not examined in this study. Taken together, the two separate findings indicate that,
firstly, miR‐31 is able to decrease FIH‐1 expression and secondly, that miR‐31 expression
is specifically decreased in aggressive breast cancers, may represent a mechanism for
increased levels of FIH‐1 in those tumours where miR‐31 expression is attenuated,
though this has yet to be suggested in the literature or investigated. As miR‐31 appears
to act pleiotropically in tumour progression, the role for miR‐31 in regulation of FIH‐1
expression in other tumour types is not clear.
1.5.3 Regulation of HIF transcriptional activity by nitric oxide
Nitric oxide (NO) has also been implicated in both positive and negative regulation of
HIF‐α stability and transcriptional regulation (Berchner‐Pfannschmidt et al., 2010). In
normoxic cell culture conditions, nitric oxide donors are known to stabilise and activate
HIF‐1α (Metzen et al., 2003a; Sandau et al., 2001). In hypoxia however, NO donors have
been found to elicit a reduction in HIF‐1α accumulation and activity (Huang et al., 1999;
Liu et al., 1998). HIF‐1α is known to be directly regulated by S‐nitrosylation of cysteine
residues within the ODDD and the CAD, enhancing protein stability and increasing
activity, respectively (Li et al., 2007a; Yasinska and Sumbayev, 2003). The involvement
of NO in modulation of HIF‐α is extends beyond the direct modification of HIF‐α, as NO
has also been reported to affect the HIF hydroxylases (Metzen et al., 2003b; Park et al.,
42
2008). Use of NO donors have been observed to enhance HIF‐1α‐CAD transcriptional
activity in normoxic conditions (Metzen et al., 2003b), in contrast to earlier data
showing attenuation of activity under hypoxic conditions (Huang et al., 1999). The cause
of these discordant effects of NO on transcriptional activity of HIF‐α between normoxia
and hypoxia is not known, however FIH‐1 has been suggested to be directly inhibited by
NO (Park et al., 2008). Addition of NO donors in vitro prevents FIH‐1 from hydroxylating
HIF‐1α substrates, as seen by MS. The mechanism of inhibition is unknown, though may
involve either competition between NO and oxygen for binding to the Fe(II) within the
enzyme’s active site, or nitrosylation of FIH‐1 itself as a regulatory modification
(Chowdhury et al., 2011).
1.5.4 FIH‐1 is not regulated by tricarboxylic acid cycle intermediates
FIH‐1 consumes the tricarboxylic acid (TCA) cycle intermediate 2‐OG during catalysis of
hydroxylation. In the reaction, 2‐OG accepts one oxygen atom from molecular oxygen
and is decarboxylated, giving off CO2 and forming succinate, another TCA cycle
intermediate. Regulation of HIF hydroxylases by TCA cycle intermediates has been
demonstrated previously (Selak et al., 2005). Accumulation of succinate by inhibition of
succinate dehydrogenase negatively regulates PHD2 activity, resulting in stabilisation of
HIF‐α protein levels at normoxia. This inhibition likely occurs through competition of
succinate and 2‐OG for binding to the enzyme (Hewitson et al., 2007). The possibility of
FIH‐1 being inhibited by the citric acid cycle intermediates fumarate, succinate and
oxaloacetate was investigated (Hewitson et al., 2007). FIH‐1 however was not found to
be inhibited by physiologically relevant levels of these molecules in an in vitro setting.
43
1.5.5 Regulation of FIH‐1 subcellular distribution by MT1‐MMP and Mint3
An interaction between FIH‐1 and Mint3 has recently been demonstrated to occur in
macrophages (Sakamoto and Seiki, 2009). Mint3 (also known as APBA3), is involved in
intracellular protein transport, most notably transport of amyloid precursor protein.
FIH‐1 binds Mint3 via the same C‐terminal α‐helices used for binding to HIF‐α and thus,
as has been previously demonstrated for a number of ARD substrates, directly
competes with HIF‐α for binding and modulates expression of HIF‐target genes.
Overexpression of Mint3 in HEK293 cells increased activity of a Gal4 reporter gene,
presumably by binding FIH‐1 and thereby reducing hydroxylation of Gal4DBD‐HIF‐CAD.
Also, shRNA‐mediated knockdown of Mint3 in macrophages, results in decreased GLUT1
and PGK1 expression, inferring greater availability of FIH‐1 in the absence of Mint3 for
HIF‐α hydroxylation.
As macrophages rely on glycolysis for ATP manufacture regardless of oxygen levels, the
interaction of FIH‐1 with Mint3 provides an example of a physiologically relevant role
for oxygen‐independent modulation of FIH‐1 regulation of HIF activity. Hydroxylation of
Mint3 was not reported, indicating that this regulatory mechanism is independent of
FIH‐1 catalytic activity.
A further study by the same authors expanded on this novel mode of FIH‐1 regulation in
macrophages (Sakamoto and Seiki, 2010). The authors proposed a model where FIH‐1
first bound to the cytoplasmic tail of membrane type 1 matrix metalloprotease (MT1‐
MMP; also known as MMP‐14), and this interaction recruited FIH‐1 to Mint3. When in a
complex with Mint3, FIH‐1 was prevented from hydroxylating HIF‐α, enabling
macrophages to upregulate genes necessary for glycolysis. The significance of this
regulatory mechanism in other cell types is unclear.
44
1.5.6 Regulation of FIH‐1 distribution by protein‐protein interactions
FIH‐1 is known to interact with a number of proteins in addition to HIF‐α, and
interaction with some proteins may influence distribution of FIH‐1 in the cell. In addition
to the interactions between FIH‐1 and MT1‐MMP and Mint3 that has been suggested to
increase peri‐nuclear localisation of FIH‐1 (Sakamoto and Seiki, 2009), co‐expression of
yellow fluorescent protein‐ (YFP)‐FIH‐1 with the full length Notch1 intracellular ICD was
found to increase the amount of YFP‐FIH‐1 in the nucleus. This Notch‐dependent
alteration in FIH‐1 subcellular localisation was likely due to the high affinity interaction
of FIH‐1 for Notch, as evidenced by the fact that a Notch1 asparagine mutant that binds
FIH‐1 poorly was not able to replicate this increased nuclear localisation (Zheng et al.,
2008). IκBα has been shown to bind to and sequester HDAC‐1 and ‐3 in the cytoplasm,
impacting on gene transcription (Viatour et al., 2003). The interaction between FIH‐1
and IκBα, or between FIH‐1 and other interacting proteins, may act to retain or
transport FIH‐1 between cellular compartments, affecting colocalisation with HIF‐α
substrates.
1.5.7 Is FIH‐1 phosphorylated?
1.5.7.1 Regulation of HIF‐α activity by kinase signalling pathways
The HIF pathway is subjected to regulation more complex than the oxygen‐dependent
hydroxylation events described. Crosstalk between HIF and numerous other signalling
pathways has been described, and post translational modifications (PTMs) in addition to
asparaginyl hydroxylation have been reported to occur on the HIF‐1α CAD and affect its
transcriptional activity, often by targeting the interaction between the CAD and
CBP/p300. For example, phosphorylation of hHIF‐1α at T796 by casein kinase II has been
45
shown to increase CAD activity (Gradin et al., 2002). The proposed mechanism involves
decreased affinity of FIH‐1 for the phosphorylated CAD via the disruption of a
hydrophobic interaction between FIH‐1 and L795 of HIF‐1α, subsequently decreasing
hydroxylation (Lancaster et al., 2004), (Cho et al., 2007).
The role of other kinases in regulation of HIF‐α CAD transcriptional activity is less clear,
and remains a somewhat contentious issue in the literature. In general, several
independent lines of data suggest a trend of increased CAD‐driven transcription
following activation of phosphatidylinositol 3‐kinase(PI3K)/Akt and mitogen‐activated
protein kinase (MAPK) signalling pathways, though the precise mechanisms responsible
for increased HIF‐α activity are, in many cases, unclear. Many studies regarding the role
of kinase signalling pre‐date the identification and characterisation of FIH‐1, resulting in
a lack of query into the role of FIH‐1 as an intermediary between kinase signalling and
HIF activity.
1.5.7.2 MAPK pathway
Evidence suggests that manipulation of MAPK pathways impacts on the activity of both
HIF‐1‐ and HIF‐2‐ mediated transcription in some cell types, however only HIF‐1α has
been found to be phosphorylated by the ERK1/2 MAPKs (Richard et al., 1999). A role for
the MAPK phosphorylation of HIF‐1α has been described (Mylonis et al., 2006). Two
phosphorylation events on S641 and S634 serve to block CRM1‐dependent nuclear
export of HIF‐1α, leading to nuclear accumulation and enabling transcriptional activity.
Inhibition of MAPKs therefore leads to enhanced nuclear export of HIF‐1α and a
reduction in transcriptional activity, explaining, in part, the observed decrease in HRE‐
reporter genes upon MAPK inhibitor treatment.
46
Investigation of the effects of the MAPK pathway on HIF‐2α mediated transcription in
PC12 cells has been done using the MAPK inhibitor PD98059, which binds to and
prevents activation of the upstream MEK1 (Conrad et al., 1999). Activity of full length
HIF‐2α in a HRE‐Luciferase reporter assay was impeded by PD98059 in both normoxia
and hypoxia, without affecting HIF‐2α protein levels. Furthermore, inhibition of MAPKs
by PD98059 had no effect on the levels on the phosphorylation status of HIF‐2α,
suggesting an indirect mechanism of regulation of transcriptional activity of HIF‐2α in
PC12 cells. This work was corroborated by similar findings for HIF‐2α activity in PC12
cells, and also HIF‐1α activity in HeLa and HepG2 cells (Alvarez‐Tejado et al., 2002). Use
of PD98059 was found to produce a small reduction in hypoxic levels of HIF‐1α in
HepG2 cells, but had minimal effects on HIF‐1α or HIF‐2α protein levels. Despite these
minimal changes in protein levels, there was a consistent and significant reduction in
hypoxic HRE‐driven luciferase expression in HepG2, HeLa and PC12 cells.
The role of MAPKs in regulating HIF‐2α transcriptional activity is not clear, as HIF‐2α is
not known to be a substrate of MAPKs. In addition, a HIF‐1α‐CAD construct that does
not include the two identified MAPK sites also responds to MAPK inhibitor treatment,
observed as a decrease in transcriptional activity (Sang et al., 2003), suggesting that the
nuclear export mechanism may not be solely responsible for the upregulation of HIF
activity by MAPKs. Interestingly, when the effect of PD98059 on the NAD (HIF‐1α 530‐
658) and the CAD (HIF‐1α 786‐826) were investigated separately, PD98059 was found to
downregulate both NAD‐ and CAD‐driven reporter gene expression, however, only the
NAD was found to be phosphorylated by MAPK (Sang et al., 2003). Phosphorylation of
FIH‐1 by MAPK has apparently been investigated and found not to occur, however this
data has not been published (Sang et al., 2003). Increased HIF activity, whether by
MAPK signalling or other pathways, is involved in cancer progression, thus delineating
the exact nature of regulatory events that enhance HIF‐mediated gene expression is
particularly important. Information into transcriptional activity of the HIFs gleaned from
47
studies that employ various kinase inhibitors needs to be carefully interpreted, as
CBP/p300 has also been found to be a MAPK substrate, and its activity is known to be
affected by phosphorylation status (Sang et al., 2003).
1.5.7.3 The PI3K/Akt signalling pathway
Another kinase signalling pathway that has been implicated in regulation of HIF‐
mediated gene expression is the PI3K/Akt pathway. Akt (also known as Protein Kinase B
(PKB)), is a serine/threonine kinase that is activated by various extracellular signals such
as growth factors, hormones, cytokines, and cellular stresses, including hypoxia.
Following activation, Akt has roles in the regulation of various cellular processes such as
proliferation, survival, apoptosis, angiogenesis and glucose metabolism (For a review,
see (Franke, 2008).
There has been a significant amount of research published into the link between
PI3K/Akt signalling, HIF‐1α protein levels and HIF‐1α transcriptional activity, yet the
contribution of Akt into HIF activation remains a contentious issue in the literature
(Alvarez‐Tejado et al., 2002; Arsham et al., 2002). As with the studies using MAPK
inhibitors above, data generated using PI3K inhibitors is complicated by the fact that
phosphorylation of p300 by Akt has been found to be essential for transcriptional
activity (Huang and Chen, 2005). Therefore, results obtained by use of these inhibitors
in reporter assays are reflective of the effects of kinase inhibition on both p300 and HIF‐
α. Furthermore, Akt signalling has been reported to stabilise HIF‐α protein in
glioblastoma cell line U373 (Zundel et al., 2000), although this finding was later directly
contradicted (Arsham et al., 2002). Akt has also been implicated in the growth‐factor
dependent increase in HIF‐1α levels in prostate cancer cell lines PC‐3 and DU145 (Jiang
et al., 2001). Though controversial, the various effects of the PI3K/Akt pathway on HIF‐α
48
stability make it difficult attribute any results showing elevated HIF‐mediated gene
transcription solely to enhanced levels of HIF activity.
Where evident, the effect of the PI3K/Akt pathway on HIF target gene expression has
been cell‐specific (Shafee et al., 2009). Activation of PI3K/Akt activity was found to
either increase or decrease levels of HIF target gene CA9 mRNA in a cell‐type specific
manner, with osteosarcoma Saos‐2 cells and breast cancer MCF‐7 cells showing
increased CA9 upon infection with an adenoviral vector expressing constitutively active
PI3K, and a fibrosarcoma‐derived cell line MCH603 showing a decrease in CA9 message.
The PI3K inhibitors wortmannin and LY294002 do not have equivalent effects on
hypoxic HIF‐mediated transcription in HeLa, PC12 and HepG2 cells (Alvarez‐Tejado et
al., 2002). Use of both LY294002 and wortmannin in HeLa cells showed a dose‐
dependent decrease in HIF‐mediated transcription at hypoxia. In PC12 cells, however,
use of the same concentrations of these inhibitors had no effect on HRE‐driven
luciferase expression, and in HepG2 cells, LY294002 but not wortmannin reduced
reporter activity, though this may have been due to off‐target effects the of LY294002
inhibitor. Furthermore, HRE‐reporter assays have found that constitutively active Akt
was able to enhance hypoxic gene transcription in human glioblastoma U373 cells
without increasing HIF‐1α protein level, but not in mouse hepatoma IcIc7 cells (Arsham
et al., 2002). Though it appears that PI3K/Akt signalling is not sufficient for HIF‐
dependent transcription, and hypoxia is primarily responsible for induction of HIF‐
activity in these cell lines, evidence suggests that HIF transcriptional activity is altered,
in a cell type‐specific manner, in response to inhibition or overexpression of the
PI3K/Akt pathway. Whether this occurs via cell‐specific regulation of FIH‐1, resulting in
an altered capacity for hydroxylation of HIF‐α substrates and changes in HIF‐mediated
gene transcription, remains unknown but presents an interesting path for future
investigation.
49
1.5.7.4 Summary
Investigation of an involvement of FIH‐1 in these pathways has not been published, yet
FIH‐1 has a known role in regulating HIF transcriptional activity. A possible pathway for
kinase signalling to increase HIF‐transcriptional activity could be through modification
and inhibition of FIH‐1 activity in specific cell types, and this would be consistent with
other observations of cell‐specific differences in FIH‐1 activity (section 1.2.4)
1.6 FIH‐1 and Disease
1.6.1 Hypoxia, the HIF pathway and solid tumour growth
It is commonly accepted that solid tumours arise from an accumulation of genetic
changes that act to confer upon cells a proliferative and survival advantage, leading to
rapid expansion of tumour cell populations. This rapid growth often occurs discordantly
with growth of a functional vascular network; as such, regions of hypoxia are common
in solid tumours. Neovascularisation has been found to be a limiting factor for tumour
growth. Solid tumours have been found to not progress beyond 1‐2 mm in diameter
without the growth of a new blood supply, as beyond this size diffusion of oxygen is not
sufficient (Folkman et al., 1966; Folkman, 1990).
Assessment of tumour tissue oxygenation with oxygen‐sensitive electrodes has shown
that overall, regions of hypoxia arise in approximately 50‐60% of solid tumours (Vaupel
and Mayer, 2007). Tumour‐associated hypoxia is significant as it is associated with more
aggressive cancer phenotypes that have increased resistance to both radiation therapy
and chemotherapy through various mechanisms, including a requirement by some
chemotherapeutic agents for oxygen to become maximally cytotoxic, difficulties in drug
50
delivery and also an increased rate of development of drug resistance via enhanced
genetic instability (For a review, see (Teicher, 1994).
HIF‐1α is also involved in treatment resistance, as HIF‐1α null MEFs are more
susceptible to ionising radiation and chemotherapeutic agents than wildtype cells via
increased apoptosis (Unruh et al., 2003). Indeed, HIF‐α expression has been found to be
an important prognostic marker for a number of cancers and has been found to
contribute to a more aggressive cancer phenotype. Increased levels of HIF‐α protein has
been reported in a variety of cancer types, including breast, lung, skin, bladder, colon,
pancreas, brain, gastric, ovarian, and renal cell carcinomas (Zhong et al., 1999), (Talks et
al., 2000). Overexpression of HIF often correlates with a poor prognosis in a many of
human cancers types, including breast (Schindl et al., 2002), ovarian (Birner et al.,
2001), cervical (Burri et al., 2003), brain (Zagzag et al., 2000) and stomach cancers
(Takahashi et al., 2003). Conversely, loss of HIF‐1α impedes growth of tumours (Ryan et
al., 1998).
The HIF pathway is central in enabling tumour cells to adapt to the challenging
intratumoural microenvironment, and also in activating transcription of genes required
for carcinogenesis. Important phenotypic hallmarks of cancer include unlimited
proliferative potential, self‐generation of growth signals, insensitivity to anti‐
proliferative signals, evasion of apoptosis, upregulation of glycolysis, angiogenesis and
metastasis (Hanahan and Weinberg, 2000). Stabilisation and activation of HIF, either by
hypoxia or other means, enables tumour cells to express genes regulating these
advantageous processes. Specific well known examples include transforming growth
factor‐β (TGF‐β) (Schaffer et al., 2003) and insulin‐like growth factor 2 (IGF2) (Feldser et
al., 1999) for cell proliferation and survival, VEGF (Levy et al., 1995; Forsythe et al.,
1996) and its receptor FLT‐1 (Gerber et al., 1997) for stimulation of angiogenesis, matrix
metalloprotease MMP‐2 for extracellular matrix metabolism and autocrine motility
51
factor (AMF) for cell motility (Krishnamachary et al., 2003), as well as glucose
transporters GLUT1 and GLUT3 (Ebert et al., 1995) and various glycolytic enzymes
(Semenza et al., 1994) to increase anaerobic ATP production. HIF‐1α has a well‐known
paradoxical role in both promoting and preventing apoptosis via the activation of
transcription of both pro‐ and anti‐apoptotic genes, such as BNip3 and TGF‐α,
respectively (Sowter et al., 2001; Krishnamachary et al., 2003), adding to the complexity
of the role of HIF‐1 in tumour progression. Importantly, FIH‐1 appears to be capable of
altering HIF‐mediated expression under hypoxic conditions (Stolze et al., 2004). This
suggests that FIH‐1 may act to modulate HIF‐mediated transcription even when oxygen
is limiting, and also highlights the importance of further investigating FIH‐1 expression,
both in cancer and in normal cells.
The relative contributions of HIF‐1α and HIF‐2α to cancer progression are not
completely understood. For example, studies in RCC, a disease typified by loss or
defects in both VHL alleles, have shown that tumour progression is favoured by a
specific upregulation of the HIF‐2α isoform, as HIF‐2α was found to promote growth of
RCC xenografts, whereas HIF‐1α expression slow impaired RCC tumour progression
(Raval et al., 2005). In specific cases, activity of HIF‐1α has been suggested to impede
tumour progression by promoting apoptosis of tumour cells (Khan et al., 2011).
In sum, cancer cells exploit the HIF pathway to adapt to the hypoxic microenvironment
by activating HIF‐target genes, thus enabling angiogenesis, increasing glucose uptake
and metabolism, and enhancing other processes advantageous to cancer cells such as
cell proliferation, survival and metastasis. This activation can occur through tumour‐
associated hypoxia, or via oncogenic activation through a variety of signalling pathways.
Manipulating the HIF pathway to enable a reduction in these processes therefore
provides an important therapeutic avenue for the treatment of neoplastic disease. At
the beginning of this project, there was a deficit in knowledge regarding the expression
52
of FIH‐1 in tumour tissues. Given the importance of HIF target gene expression in
cellular processes critical for cancer progression, it was considered vital that an
investigation into FIH‐1 expression in a range of normal and tumour tissues was carried
out.
1.7 Thesis Objectives
At the commencement of this project, FIH‐1 expression had not been investigated in
any tumour tissues, nor had investigations been made into the contribution of FIH‐1 to
cancer progression. Given the extensive and important involvement of HIF
transcriptional activity in tumour progression, a role for FIH‐1 in the regulation of HIF
activity in cancer was thought possible and warranted investigation. Specifically, to
enable HIF to escape from FIH‐1 mediated repression, it could logically be expected that
FIH‐1 expression or activity would be downregulated.
It was therefore original intent of this thesis to directly address this question and
investigate the expression of FIH‐1 protein in a range of normal tissues compared to
tissues derived from solid tumours. However, following publication of data that
comprehensively addressed this same goal (Soilleux et al., 2005), the aims of this
project were redefined to investigate instead the existence of potential regulatory
mechanisms influencing the activity of FIH‐1 independently or in addition to the effect
of oxygen concentration.
Thus, the aims of this thesis can be summarised as follows:
Aim 1: To investigate the expression of FIH‐1 protein in a range of cancer cell lines,
normal and cancer tissues
53
Aim2: To investigate the existence of post‐translational modification of FIH‐1, with a
particular focus on regulatory mechanisms that may influence the activity of FIH‐1,
independently, or in addition to, the effect of oxygen concentration.
This thesis will first discuss work carried out in the fulfilment of the Aim1, followed by a
discussion relevant data that was published following this work. This thesis will then
describe work carried out to address Aim 2.
55
CHAPTER 2
MATERIALS AND
METHODS
57
2 MATERIALS AND METHODS
2.1 Chemicals and Reagents
1kb DNA plus ladder Invitrogen
40% Bis‐Acrylamide Solution Bio‐Rad
Amylose agarose Scientifix
BigDye Invitrogen
Bio‐Lyte Buffer Bio‐Rad
Bradford Assay Dye Reagent Bio‐Rad
Bromophenol Blue Sigma
Brilliant Blue R Concentrate Sigma
CHAPS Bio‐Rad
CNBr‐activated Sepharose 4B Amersham Biosciences
DAPI Stain Calbiochem
DMEM Gibco‐BRL
Fetal calf serum JRH Biosciences
FuGENE 6 Transfection Reagent Roche
HiTrap chelating HP 5ml columns GE Healthcare
HiTrap IgM HP Purification columns GE Healthcare
Immobiline DryStrip IPG Strips, 7cm,
pH 3‐11 and pH 4‐7 GE Healthcare
IPG Buffer 4‐7 GE Healthcare
Lipofectamine 2000 Invitrogen
Nickel‐IDA agarose Scientifix
Oligofectamine Invitrogen
PD‐10 Desalting Columns GE‐Healthcare
Ponceau Red Sigma
58
Precision Plus Protein Dual Colour Standards Bio‐Rad
Protein A Sepharose GE‐Healthcare
Puromycin Sigma
ReadyStrip IPG Strips, 7cm, pH 4‐7 Bio‐Rad
SyproRuby Protein Gel Stain Invitrogen
Tween 20 Sigma
2.2 Radiochemicals
[γ‐32P] ATP; 3000 Ci/mM, 10 mCi/ml, 1 mCi Perkin Elmer
2.2 Commercial Kits
2‐D Clean Up Kit Amersham Biosciences
DNA midiprep kit Qiagen
Gel Extraction Kit Qiagen
ImmobilonTM Western Chemiluminescent
HRP Substrate Millipore
Mouse Typer Sub‐Isotyping Kit Bio‐Rad
SilverQuest Invitrogen
SupersignalTM West Pico Chemiluminescent
HRP Substrate Pierce
2.3 Enzymes
AcTEVTM Protease Invitrogen
Lambda Protein Phosphatase New England Biolabs
59
Lysozyme Sigma
PFU turbo polymerase Stratagene
Restriction Enzymes New England Biolabs
T4 DNA ligase New England Biolabs
Taq Polymerase New England Biolabs
Thrombin Sigma
2.4 Antibodies
2.4.1 Primary Antibodies
9F6: Mouse monoclonal antibody generated by immunising mice
with full length human maltose‐binding protein (MBP)‐FIH‐1.
Binds unknown antigen of ~37 kDa.
Anti‐FIH‐1: Rabbit polyclonal antisera No.8 and No.9 made in our
laboratory against MBP‐FIH‐1. Blots incubated in the indicated
dilutions overnight.
Commercial affinity‐purified rabbit polyclonal antibody NB
100‐428 (Novus Biologicals). Incubated at 1:1000 dilution in
PBT overnight.
Anti‐tubulin: Rat monoclonal (Abcam ab6160). Used 1:4000 in PBT,
incubated for 1‐2 hours.
Anti‐Myc: Hybridoma supernatant 9E10. Due to batch to batch variation,
dilutions and blotting conditions vary.
Anti‐p‐Ser5 RNA PolII: Antibody raised against phosphorylated RNA Polymerase II
CTP repeat YSPTSPS (phosphor S5) (Abcam, ab5131).
60
Anti‐paxillin: Purified mouse monoclonal antibody raised against full length
human paxillin with GST tag. (Millipore (formerly Upstate)
catalogue number 05‐417)
Anti‐nucleoporin: Mouse monoclonal antibody generated against amino acids
401‐522 of human nucleoporin p62. (Santa Cruz, catalogue
number sc‐48373).
2.4.3 Secondary Antibodies
Anti‐mouse‐HRP: Goat anti‐mouse antibody (Pierce). Used 1:20 000 in PBT.
Anti‐rabbit‐HRP: Goat anti‐rabbit antibody (Pierce). Used 1:20 000 in PBT.
Anti‐rat‐HRP: Goat anti‐rat antibody (Abcam ab6845). Used 1:5000 in PBT.
Anti‐mouse‐TRITC: anti‐mouse tetramethylrhodamine B isothiocyanate (TRITC)
(DAKO). Used at 1:1000 in 3% BSA in PBT.
2.5 Buffers and Solutions
2‐D Lysis Buffer 1: 1% SDS, 25 mM Tris pH 8.5, 20 mM DTT.
2‐D Lysis Buffer 2: 13% CHAPS, 0.65% Triton‐X‐100 and 43 mM DTT
2x SDS Sample Buffer: 100 mM Tris pH 6.8, 20% glycerol, 4% SDS, 200 mM DTT,
coomassie blue
4x SDS Sample Buffer: 240 mM Tris pH 6.8, 40% glycerol, 8% SDS, 400 mM DTT,
coomassie blue
10X Thrombin
cleavage buffer:
200 mM Tris pH 8, 1.5 mM NaCl, 25 mM CaCl2.
61
Amylose Lysis Buffer: 150 mM NaCl, 20 mM Tris, pH 8
Binding Buffer: 10 mM Tris pH 7.5, 0.4 % Triton‐X‐100, 150 mM KCL, 2 mM EDTA,
1 mM PMSF
Acetate Buffer: 0.1 M AcOH (acetic acid), 0.5 M NaCl, pH 4.0
Coupling Buffer: 0.5 M NaCl, 0.1 M NaHCO3, pH 8.3
Coomassie stain: 0.03% coomassie brilliant blue, 8.75% acetic acid, 50% methanol
Desalting Buffer: 20 mM Tris pH 8, 150 mM NaCl
Destain 1: 8.75% acetic acid, 50% methanol
Destain 2: 7% acetic acid, 5% methanol
Elution Buffer: 0.5 % SDS, 50 mM beta mercaptoethanol (BME), 10 mM Tris pH 8
Equilibration
Solution 1:
6 M urea, 0.375 M Tris pH 8.5, 2 % SDS, 20 % glycerol, 10 mM
DTT.
Equilibration
Solution 2:
6 M urea, 0.375 M Tris pH 8.5, 2 % SDS, 20 % glycerol, 2.5 %
iodoacetamide.
Guanidine Extract
Buffer:
6 M guanidine, 100 mM phosphate buffer pH 7, 150 mM NaCl,
0.1% NP‐40, 20 mM imidazole, 20 mM DTT; vacuum filtered
through a 0.45 μm filter
Guanidine Wash
Buffer:
6 M guanidine, 100 mM phosphate buffer pH 7, 150 mM NaCl, 50
mM imidazole; vacuum filtered through a 0.45 μm filter
Hypotonic Buffer: 10 mM HEPES pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.4% NP‐40,
10% Ficoll‐400, with 1 mM DTT, 1 mM PMSF and 1 x protease
inhibitor cocktail added fresh
IEF Buffer: 8 M Urea, 2% CHAPS, 40 mM DTT, 1 mM NaF (added fresh)
Native Lysis Buffer: 150 mM NaCl, 1% NP‐40, 50 mM Tris, 10 mM MgCl2, 10 mM
MnCl2, supplemented with fresh 1 mM PMSF, 1x protease
inhibitor cocktail, 1 mM NaF, 1 mM Na3VO4, 20 mM β‐
glycerophosphate
Nickel Lysis Buffer: 500 mM NaCl, 20 mM Tris, pH 8, 5 mM imidazole
62
Nuclear Extract Buffer: 20 mM HEPES pH7.9, 1.5 mM MgCl2, 0.5 mM EDTA, 20% glycerol,
0.42 M KCl, with 1 mM DTT, 1 mM PMSF and 1 x protease
inhibitor cocktail added fresh
PBT: PBS, 0.1% Tween 20
Ponceau Stain: 0.5% (wt/vol) Ponceau Red, 1% (vol/vol) acetic acid
Protease Inhibitor
Cocktail (100x)
200μg/ml Aprotinin, 500μg/ml Leupeptin, 100 μg/ml Pepstatin,
400 μ/ml Bestsatin, 150 μM EDTA
Rehydration Solution: 8 M Urea, 4% CHAPS, 0.4% Triton‐X‐100 and 10 mM DTT
SB3‐10 Wash buffer: 1% SB3‐10, 100 mM phosphate buffer pH 7, 5 mM imidazole,
vacuum filtered through a 0.45 μm filter
SDS Wash buffer: 1% SDS, 100 mM phosphate buffer pH 7, 5 mM imidazole,
vacuum filtered through a 0.45 μm filter
SDS elution buffer: 1% SDS, 100 mM phosphate buffer pH 7, 250 mM imidazole,
vacuum filtered through a 0.45 μm filter
Tris/Glycine
separating gel:
10 or 12.5% polyacrylamide, 187 mM Tris pH 8.8, 0.1% SDS
Tris/Glycine stacking
gel:
5 % polyacrylamide, 62.5 mM Tris pH 6.8, 0.1 % SDS
Wash buffer: 10 mM Tris pH 8, 150 mM NaCl, 1 mM PMSF.
Western Transfer
Buffer:
50 mM Tris, 2.85% glycine
Whole cell extract
buffer:
20 mM HEPES, 0.42 M NaCl, 0.5 % NP‐40, 25 % glycerol, 1.7 mM
EDTA, 1.5 mM MgCl2.
2.7 Bacterial strains and growth media
Escherichia coli (E.coli) DH5α: Used to propagate plasmid DNA
E.coli BL21 (DE3): Used to express protein
Luria Broth (LB): 1% tryptone, 0.5% yeast extract, 1% NaCl, sterilised by autoclaving.
63
2.8 Plasmids
2.8.1 Bacterial Expression Plasmids
pET‐32a (Novagen): Encodes an N‐terminal Trx‐6His tag and also a C‐terminal 6His tag
unless a stop codon is present in the insert.
pMBP: Encodes an N‐terminal maltose‐binding protein (MBP) tag.
pMBP‐FIH‐1: Encodes MBP‐full length human FIH‐1. Constructed and kindly provided by
Dr. Richard Bruick. Described in (Lando et al., 2002a).
pET32a‐hHIF‐1α‐ID‐CAD(737‐826): Encodes 6His‐human HIF‐1α residues 737‐826.
Constructed and kindly provided by Dr. Daniel Peet. Described in (Linke et al., 2004).
pET32a(+)‐TEV‐FIH‐1: Encodes full length human FIH‐1 with an N‐terminal Trx‐6His tag.
This tag can be liberated using TEV protease. The TEV protease recognition site was
inserted into the vector by Dr. Fiona Whelan, and full length human FIH‐1 was then
cloned into the vector by Sarah Wilkins.
pET32a‐mNotch1‐ANK1‐4.5: Encodes Trx‐6His mNotch1 Ankyrin Repeats 1‐4.5 (1847‐
2027). Generated and kindly provided by Sarah Linke.
pET32a‐Myc‐mNotch1‐RAM: Encodes Trx‐6His‐Myc‐mNotch1 RAM domain (1753‐
1847)‐6His. Generated and kindly provided by Sarah Linke.
pET32a(+)‐TEV‐FIH‐1 S36A and S36D mutants
64
S36A and S36D mutations were introduced by overlap extension PCR using the following
primers. The S‐tag primer was used for pET‐32 sequencing.
Oligonucleotides:
FIH S36A Forwards: 5’ CCCAGTTGCGCAGTTATGCCTTCCCGACTAGGCCC 3’
FIH S36A Reverse: 5’ GGGCCTAGTCGGGAAGGCATAACTGCGCAACTGGG 3’
FIH‐1 S36D Forwards: 5’ CCCAGTTGCGCAGTTATGACTTCCCGACTAGGCCC 3’
FIH‐1 S36D Reverse: 5’ GGGCCTAGTCGGGAAGTCATAACTGCGCAACTGGG 3’
S‐tag: 5’ GGTTCTGGTTCTGGCCATT 3’
T7 terminator: 5’ GCTAGTTATTGCTCAGCGG 3’
2.8.2 Mammalian Expression Plasmids
pcDNA 3.1‐FIH‐1: Kindly provided by Dr. Richard Bruick and described in (Lando et al.,
2002a).
pEGFP‐C2 (BD Biosciences): Enables expression of enhanced green fluorescence protein
(GFP) in mammalian cells.
pEF‐IRES‐Puro 5‐FIH‐1: Enables expression in mammalian cells of a bicistronic message
incorporating both the puromycin resistance and FIH‐1 mRNA, with an internal
ribosome entry site (IRES) to ensure efficient translation of both messages. Generated
and kindly provided by Dr. Daniel Peet.
pEF‐IRES‐Puro 5‐Myc‐6His‐FIH‐1
FIH‐1 was digested out of pPC86‐FIH‐1 using Nco1 and Not1, and ligated into Nco1 and
Not1‐digested pEF‐IRES‐puro 5. pEF‐IRES‐puro5‐FIH‐1 was prepped and digested with
65
Nco1. Oligonucleotides coding for the desired Myc‐6His tag and containing ends
complementary to the Nco1 digested vector were annealed by first heating for 5
minutes at 100°C in the presence of 100 mM NaCl and then incubation overnight at
room temperature. Singly annealed oligonucleotides were gel‐purified and ligated serial
dilutions of the “insert” (1/10 to 1/100 000) into the digested vector, and ligations
transformed into E.coli. DNA was made from six resulting colonies and correct ligation
of the tag was assessed by diagnostic digest, revealing three clones containing the Myc‐
6His sequence in the correct orientation. Following successful sequencing, pEF‐IRES‐
puro5‐Myc‐6His‐FIH‐1 DNA was prepared and used for generation of stable HeLa cell
lines.
Oligonucleotides:
Forwards: 5’ CATGGGTGAACAAAAGCTTATTTCTGAAGAAGATTTGAACCATCATCATCATCAT
CATGC 3’
Reverse: 5’CATGGCATGATGATGATGATGATGGTTCAAATCTTCTTCAGAAATAAGCTTTTGTT
CACC 3’
pSUPER‐FIH‐1‐85, 336, 960 and 1020: Original vector obtained from Oligoengine, and
siRNA sequences cloned into vector by Anthony Fedele. siRNA constructs targeted
towards bases 85‐107, 336‐358, 960‐982 and 1020‐1042 of the FIH‐1 coding sequence
(GenBank accession number NM_017902).
2.9 siRNA‐mediated knockdown of FIH‐1
2.9.1 Plasmid‐based system
Self‐complementary oligonucleotides complementary to FIH‐1 mRNA beginning at
nucleotides 85, 336, 960 and 1020, and 22 bp in length were cloned into pSUPER
66
(Oligoengine) for intracellular expression of dsRNAs by Sarah Linke and Anthony Fedele.
In addition, a scrambled self‐complementary dsRNA containing a random 22mer with
no complementary sequences was cloned into pSUPER for use as a transfection control.
293T cells were grown in 6 well trays until the desired confluency was reached. Various
amounts (2 μg – 4 μg) of each pSUPER construct was transfected into cells using either
Lipofectamine of FuGENE according to manufacturer’s instructions, and the media
replaced 4, 6 or 8 hours post transfection. Cotransfection of 1 μg GFP was used as a
transfection control. Cells were then grown for a further 24, 48 or 72 hours. The highest
transfection efficiency obtained was 70% transfection efficiency (as evidence by GFP
positive cells) with no discernable decrease in FIH‐1 in siRNA transfected cells.
2.9.2 siRNA oligonucleotides
Two siRNA oligonucleotides were designed based on published sequences
complementary to nucleotides 91‐111 and 160‐180 relative to the start codon (GenBank
accession number NM_017902) (Stolze et al., 2004). Lyophilised siRNA duplexes were
resuspended by addition of the provided resuspension buffer (100 mM Potassium
Acetate, 30 mM HEPES‐KOH, 2 mM Magnesium Acetate, pH 7.4) to a concentration of
20 μM. siRNA oligonucleotides were transfected into cell by addition to cell culture at a
final concentration of 20 nM at 0 and 24 hour time points, using Oligofectamine
according to manufacturer’s instructions. Cells were harvested at 48 hours.
2.10 General DNA methods
2.10.1 Transformations
Competent E.coli DH5α or BL21 (DE3) cells were thawed on ice and between 10‐100 ng
of plasmid DNA was mixed with 50 μl of the competent cells. After 20 minute incubation
67
on ice, the cells were heat shocked for 1 minute at 42°C. Cells were incubated on ice for
a further 10 minutes and plated onto LB/agar plates with the appropriate selection
antibiotic. Ligation transformants were mixed with 500 μl of SOC media and incubated
at 37°C for 1 hour. Cells were gently pelleted at RT and plated onto selection LB/agar
plates. Plates were incubated at 37°C overnight.
2.10.2 DNA preparation
Single colonies from transformation plates were used to inoculate 5 ml LB (mini‐prep) or
50 ml LB (midi‐prep) overnight cultures, with 0.1 % ampicillin or carbenicillin.
Mini‐Prep: Overnight cultures were centrifuged at 14 000 rpm at RT. 150 μl of each
resuspension buffer (100 μg RNase A, 50 mM Tris (pH 8.0), 10 mM EDTA (pH 8) lysis
buffer (200 mM NaOH, 1% SDS) and neutralisation buffer (600 mM KAc (pH 5.5) were
added sequentially and samples incubated on ice for 15 minutes. After centrifugation
for 10 minutes at 14 000 rpm at 4°C, supernatants were combined with 1 ml cold EtOH
and incubated at ‐20°C for 30 minutes. Samples were centrifuged once more the pellets
washed with 1 ml cold 70% EtOH, was dried at 42°C and resuspended in 15‐20 μl of TE.
Midi‐Prep: Midipreps were performed using the QIAGEN midiprep kit according to the
provided protocol.
2.10.3 Agarose gel electrophoresis
DNA was separated on 1‐2% agarose gels.
68
2.10.4 Gel‐purification of DNA
DNA was purified from agarose gels using the QIAquick Gel Extraction Kit (QIAGEN).
2.10.5 Restriction digests
Restriction digests were carried out using NEB enzymes and buffers according to the
manufacturer’s protocols.
2.10.6 Ligations
Ligations were performed using standard procedures; Approximately 100‐200ng of
insert DNA was incubated with cut vector DNA at a 1:4 vector:insert ratio, with ATP,
ligation buffer and 5 units T4 DNA ligase in a total volume of 20 μl. Ligations were
performed either at 37°C for 4 hours or overnight at 16°C. To prevent the re‐ligation of
vector plasmid, digests were incubated with SAP for 1 hour at 37°C prior to ligation to
remove 5’ phosphate groups. SAP was inactivated by incubation at 65°C for 10 minutes.
2.10.7 Sequencing
DNA sequencing was performed using 100 ng of primer and 400 ng of template DNA
with 2 μl BigDye Version 3 and the commercial buffer in a total reaction volume of 20 μl.
The PCR program involved a denaturing step of 3.5 minutes at 96°C, and then 25 cycles
of 30 seconds at 96°C, annealing for 15 seconds at 50°C and extension for 4 minutes at
60°C. DNA products were precipitated by addition of 80 μl 75 % isopropanol,
centrifuged for 30 minutes at 14 000 rpm using a bench top centrifuge, and washed
with 250 μl of 75% isopropanol. The isopropanol wash was removed and the pellet air‐
dried. Sequence analysis was carried out by the Institute of Medical and Veterinary
Science (IMVS) Adelaide.
69
2.11 Recombinant Protein Purification Methods
2.11.1 Ni2+‐affinity purification of recombinant His‐tagged proteins
Single colonies of transformed E.coli BL21 were used to inoculate 50 ml LB overnight
cultures. 15 mls of overnight culture was then used to inoculate 500 ml LB cultures that
were grown in a shaking platform at 37°C for approximately 1‐2 hours, or until the cell
density reached an optical density at wavelength 600 nm (OD600) of approximately 0.6.
Protein expression was induced by the addition of 1 mM Isopropyl β‐D‐
thiogalactopyranoside (IPTG) and growth on a shaking plate at 37°C for 4 hours. For FIH‐
1 expression, cultures were induced and grown at 30°C for 4 hours to reduce
degradation of the protein. Cultures were then centrifuged at 3000 rpm for 5 minutes,
the supernatant decanted and the cell pellet frozen at –20°C until required. Frozen cell
pellets were then thawed on ice, resuspended in 30 mls of lysis buffer (20 mM Tris pH 8,
500 mM NaCl, 5 mM imidazole). Cells were lysed by three passes through a pre‐chilled
cell disruptor (Microfluidics), and the lysates centrifuged for 30 minutes at 14 000 rpm
(JA 25.50 rotor, Beckman Coulter centrifuge) at 4°C. 1 ml of Nickel‐IDA resin was added
to the clarified lysate and shaken at 4°C for 1‐2 hours before loading into an empty
column. Resin was washed with 50 ml of lysis buffer with freshly added 1 mM PMSF, 0.5
mM DTT, then 50 ml of lysis buffer with 10 mM imidazole. Bound protein was eluted
with 250 mM imidazole and desalted using SephadexTM G‐25 Medium PD‐10 desalting
columns equilibrated with 25 ml desalting buffer.
70
2.11.2 Amylose‐affinity purification of recombinant MBP‐tagged proteins
Overnight cultures were grown and used to inoculate 500 ml cultures for protein
expression as described above, except 500 ml cultures were induced using 0.2 mM IPTG
and grown at 30°C for 5 hours. Frozen pellets of the induced cultures were thawed on
ice when required and resuspended in 30 ml of amylase lysis buffer with fresh 1 mM
PMSF added. Cells were lysed by passing three times through a cell disruptor
(Microfluidics) that was chilled on ice. Lysates were centrifuged for 30 minutes at 14
000 rpm (JA 25.50 rotor, Beckman Coulter centrifuge) at 4°C and 1 ml of amylose
agarose (Scientifix) resin added to the clarified supernatant. After 1‐2 hours incubation
on a rocking platform at 4°C, the resin was placed in an empty PD‐10 column and
washed with 50 ml amylose lysis buffer with 1 mM PMSF, then 50 ml of amylose lysis
buffer before eluting with 10 mM maltose.
2.12 General Protein methods
2.12.1 Preparation of cell lysates
Cells were washed in three times in PBS and harvested into microfuge tubes. Cells were
spun at 1400 rpm at RT for 5 min and the pellet resuspended in WCEB (with 1 mM DTT,
1 mM PMSF and 1x protease inhibitor cocktail added fresh) and incubated on rocking
platform at 4°C for 30 minutes. Lysates were then centrifuged at 14 000 rpm for 30
minutes at 4°C and the supernatant transferred to a new microfuge tube. The protein
extract was assayed for protein concentration by the Bradford assay. Extracts were
stored at –80°C.
71
2.12.2 Preparation of nuclear and cytosolic extracts
Cells were washed three times with ice cold PBS and harvested into a microfuge tube.
Adherent cells were treated with TEN, harvested into a microfuge tube and washed a
further time with PBS. Cells were gently pelleted at 1400 rpm for 5 minutes at 4°C, and
resuspended in approximately 2.5 pellet volumes of hypotonic buffer and incubated on
ice for 5 minutes before centrifugation at 14 000 rpm for 30 minutes at 4°C. The
supernatant containing the cytosolic fraction was removed and the pellet resuspended
in approximately 1.5 pellet volumes of nuclear extract buffer. Nuclear lysis occurred
during a 45 minutes incubation on a rocking platform at 4°C, and lysates were then
centrifuged at 14 000 rpm for 30 minutes at 4°C. The supernatant containing the
nuclear fraction was retained. Quality of subfractionation was assessed by
immunoblotting using antibodies against cytosolic‐ (paxillin) and nuclear‐ (nucleoporin)
specific proteins.
2.12.3 Protein quantification
For quantification of total protein in lysates and 2‐D samples, Bradford Assays were
employed, using bovine serum albumin (BSA) as the standard and using Bradford Assay
Dye Reagent.
2.12.4 Sodium dodecyl sulfate polyacrylamide gel electrophoresis
Protein samples were heated at 95°C for 10 minutes in the presence of the appropriate
amount of either 2x or 4x SDS sample buffer. Proteins were then loaded onto 1.5 mm
thick 10% or 12.5% Tris/Glycine gels and run in SDS‐PAGE running buffer at between
120‐190 V.
72
2.12.5 Protein staining
Coomassie: Total protein was visualised by staining with Coomassie stain, then
destaining with destain 1 for 1 hour followed by further destaining with destain 2
overnight.
2.12.6 Western blotting
Nitrocellulose blotting membrane was soaked in Milli‐Q (MQ) H2O until saturated, then
equilibrated in wet transfer buffer. Following SDS‐PAGE, proteins were transferred onto
the nitrocellulose membrane by assembly of the wet transfer apparatus and application
of 250 mA current at 4°C for 60 minutes. Transfer and loading was routinely assessed by
staining protein with ponceau stain. Blots were blocked with 10 % skim milk for 1 hour
before addition of the desired primary antibody. Following primary antibody incubation,
blots were washed with varying stringencies depending on the requirements of each
antibody, then incubated with the appropriate secondary antibody. Following washing,
ECL reagents (Pierce SuperSignalTM West Pico Chemiluminescent HRP Substrate or
Millipore Immobilon TM Western Chemiluminescent HRP Substrate) were applied
according to each manufacturer’s instructions and blots exposed to X‐ray film.
2.12.7 Stripping and re‐probing western blots
Blots were washed for 10 minutes in PBS/0.1% Tween, and then 5 minutes in PBS. Blots
were then incubated in stripping solution (2% SDS, 50 mM Tris pH 7, 50 mM DTT), at
70°C for 15‐30 minutes, and stripping solution removed by 3 washes in PBS of 5 minutes
each. Blots were then re‐blocked and re‐probed.
73
2.12.8 Immunofluorescence
Cells were grown on glass coverslips in 6 well trays for at least two days prior to fixation.
Cells were fixed with 4% paraformaldehyde for 30 minutes, washed with 1x PBS, then
permeabilised with permeabilisation solution (0.1% triton‐X‐100, 0.1% Na citrate in PBS)
and incubated for 5 minutes on ice. Coverslips were washed and blocked by incubation
for 1 hour in 5% skim milk. 9F6 antibody was added at 1:100 dilution in PBS and
coverslips incubated for 3 hours at RT. After washing twice in PBS, anti‐mouse TRITC
conjugated secondary antibody (DAKO) was added at 1:1000 dilution and incubated in
darkness for 1 hour. Coverslips were washed twice before incubation in Hoechst stain
for 4 minutes. Coverslips were washed once, dried, and mounted onto microscope
slides using 90% glycerol, 20 mM Tris pH 8.5.
2.13 IgM Column Purification
HiTrap IgM Purification HP columns were used to immobilise the 9F6 antibody to the
column for use in purification of the unknown antigen. Binding of the antibody to the
column was achieved according the manufacturer’s instructions. The
immunoprecipitation (IP) carried out by preparing whole cell extract (WCE) from 293T
cells using standard protocols, and then diluting the WCE 50‐fold into the required
thiophilic binding buffer (20 mM sodium phosphate, 0.8 M ammonium sulfate
((NH4)2SO4), pH 7.5). Lysate was filtered using a 0.45 μm filer and applied to the column
at a flow rate of 1 ml / minute. The column was then washed with 3 bed volumes of
binding buffer and elution was achieved by passing 10 ml of the recommended elution
buffer (20 mM sodium phosphate) through the column to elute antibody/antigen
complexes in 1 ml fractions. Each fraction was then separated by SDS‐PAGE and blotting
with the 9F6 antibody used to identify fractions containing the desired antigen.
74
2.14 Cyanogen Bromide‐Activated Sepharose Purification of IgM
2.14.1 Preparation of cyanogen bromide‐activated sepharose
0.3 g of cyanogen bromide‐ (CNBr)‐activated sepharose 4B (Amersham Biosciences) was
prepared and used according to manufacturer’s instructions. The lyophilised powder
was hydrated in 1 mM HCL and washed with 150 ml of 1 mM HCL using a sintered glass
filter. 0.3 g of lyophilised powder gave approximately 1 ml of prepared slurry.
2.12.2 9F6 dialysis
A Pierce Slide‐A‐Lyzer Mini Dialysis unit was used to buffer exchange the 9F6 antibody
into coupling buffer. 1 ml of antibody solution was dialysed against 500 ml of coupling
buffer at 4°C on a magnetic stirrer. Dialysis occurred over 24 hours, with the buffer
replaced with 500 ml fresh coupling buffer approximately halfway through the
procedure.
2.12.3 Coupling 9F6 to cyanogen bromide‐activated sepharose
The prepared sepharose was then added to the antibody in coupling buffer and binding
allowed to occur overnight on a rotating platform at 4°C. The antisera was quantified
before and after binding, and an 8x reduction in protein concentration was evident
following incubation with CNBr‐Sepharose, indicating that 9F6 had bound the resin.
Following coupling, unbound 9F6 was washed away with 5 ml of coupling buffer and the
resin was incubated in 1 M ethanolamine pH 8.0 for 2 hours to block any remaining
active groups. The resin was then washed 5 times with buffers of alternating pH
(acetate buffer pH 4, coupling buffer pH 8.3) Antibody‐conjugated resin was stored at
4°C in 10 mM Tris pH 8.0, 140 mM NaCl, with 0.02% sodium azide (NaN3).
75
2.12.4 Capturing the 9F6 antigen from cell lysate
100 μl of 9F6‐CNBr‐sepharose prepared above was incubated with 200 μl of 293T cell
lysate prepared using whole cell extract buffer (WCEB) for 2‐4 hours at 4°C, and
unbound protein removed by 3 washes in wash buffer.
2.14.5 Elution of captured antigen
And protein that bound to the 9F6 resin was eluted by heating for 5 minutes at 55°C in
Elution Buffer. The resin was then centrifuged at 14 000 rpm for 3 minutes and the
supernatant removed and heated at 95°C with SDS sample buffer for loading onto
protein gels.
2.15 General Mammalian Cell Culture methods
2.15.1 Mammalian cell lines and media
293T: human embryonic kidney. DMEM + 10% FCS
Caco‐2: human colon adenocarcinoma. DMEM + 20% FCS
Cos‐1: monkey kidney. DMEM + 10% FCS
HeLa: human cervical carcinoma. DMEM + 10% FCS
Hep3B: human liver carcinoma. DMEM + 10% FCS
HepG2: human liver carcinoma. DMEM + 10% FCS
PC12: rat adrenal pheochromocytoma. DMEM + 10% FCS + 5% Horse Serum
MEFs: mouse embryonic fibroblasts. ES cell media + 10 mM FCS, 2 mM Glutamine.
Cells were maintained in at 37°C and 5 % CO2.
76
2.15.2 Transfection
Transfections were performed using either Oligofectamine, Lipofectamine or FuGENE 6
transfection reagents according to manufacturer’s instructions.
2.15.3 Generation of stable cell lines
HeLa cells were grown to 30% confluency in a 6 cm dish and transfected with 2 μg of the
Myc‐6His‐pEF‐IRES‐puro5 mammalian expression plasmid. Cells were grown for a
further 48 hours and then media was replaced with media containing 1 μg/ml
puromycin. Media was then changed daily with fresh media containing 1 μg/ml of
puromycin to remove dead cells, for 7 days. Cells were then trypsinised and either
frozen as a polyclonal pool, maintained as a polyclonal pool, or diluted and plated out at
a density of 0.8 cells/well in 96 well trays for generation of monoclonal stable cell lines.
Expression of FIH‐1 in the polyclonal pool was assessed by western blot and 2‐D gel
western blot. For generation of monoclonal stable cell lines, wells were continually
assessed for colony growth and after 2 weeks incubation, eleven wells were found to
each contain a single colony. These colonies were expanded, frozen down and tested
for expression by western blot and 2‐D gel western blot.
2.16 Two‐Dimensional Electrophoresis
2.16.1 Sample preparation: method 1
Cells were washed four times with ice‐cold PBS and lysed by the addition of 2‐D Lysis
Buffer 1 (typically, cells from 75 cm2 flask lysed in 250 μl volume). After addition of 2‐D
Lysis Buffer 1, cell dishes were incubated at 4°C on a rocking platform for 30 minutes,
and the lysate harvested. The lysate was then diluted by the addition of an equal
volume of 2‐D Lysis Buffer 2, and urea added to give a final concentration of 8 M.
77
Lysates were then centrifuged for 15 minutes at 14 000 rpm and quantified by Bradford
Assay. To achieve a desired concentration of 100 μg / 125 μl, lysates were diluted with
rehydration solution, to which 0.625% Bio‐Lyte Buffer (Bio‐Rad) and a trace of
Coomassie Blue was added.
2.16.2 Sample preparation: method 2
Cells were washed four times with ice‐cold PBS. Cells were lysed in flasks by the
addition of 300‐400 μl of IEF buffer to which phosphatase inhibitor NaF was added
fresh, and a rubber policeman used to scrape cells off the flask and harvest cells into 1.5
ml microfuge tubes. Lysates were centrifuged at 14 000 rpm at 4°C for 30 minutes, and
supernatants were retained. Lysates were precipitated using the 2‐D Clean Up Kit
(Amersham Biosciences) according to manufacturer’s instructions, with the following
adjustments: all centrifugation steps were performed at half speed to facilitate easier
resuspension of precipitated protein. (Centrifugation at the recommended speeds
resulted in protein pellets that were difficult to re‐suspend even after extensive
sonication using a sonicating bath). The acetone/wash buffer wash was performed by
sonicating samples in a sonication bath instead of vortexing to ensure better dispersal
of the protein pellet. A second wash was performed in wash buffer with vortexing as
recommended. The final protein pellets were resuspended in a minimum volume (125
μl) of IEF buffer and quantified. Samples were diluted to 100 μg per 125 μl with IEF
buffer.
2.16.3 Strip rehydration and isoelectric focusing
0.625 μl of ampholytes was added to each sample, along with a trace of either
coomassie or bromophenol blue as a tracking dye. The strips used in these experiments
were 7 cm strips with either a pH 3‐11 or pH 4‐7 gradient, as indicated. Samples were
pipetted along the entire length of the 7 cm IPGphor strip holders and strips applied
over the samples, ensuring there the entire strip was saturated with the sample, that
78
there were no air bubbles and that the gel was in direct contact with both electrodes in
the strip holder. Strips were overlaid with mineral oil to minimise evaporation and the
strip holder cover put in place. Active rehydration of the strip occurred for 12 hours at
50 V and 20°C. Isoelectric focusing (IEF) parameters for focusing proteins on the 7 cm
strips used were as follows: 250 V for 1 hour, 1000 V for 1 hour, then a maximum 8000
V in rapid ramping mode for a total of 20 000 VHrs, and a holding step of 500 V. The
current was limited to 50 mA.
2.16.4 Strip equilibration
Following IEF, strips were placed in 10 ml polypropylene tubes and 5 ml of equilibration
solution 1 added and strips placed on a rocker for 15 minutes at room temperature.
Equilibration solution 1 was then decanted and Equilibration solution 2 was added and
strips incubated on a rocker for another 15 minutes. Following this second equilibration
steps, strips were briefly rinsed with water to remove the equilibration solution
(containing iodoacetamide) and strips embedded atop the pre‐prepared second
dimension gels.
2.16.5 Second dimension SDS‐PAGE
1.5 mm 12.5% Tris/Glycine gels were poured and allowed to polymerise at 4°C for a
minimum of 1 hour. The 70% ethanol at the top of the gel was decanted and 1x running
buffer was poured over the top of the gel. The plastic tab overhand on the basic end of
the strip was cut off and the strip applied to the top of the gel such that the strip was
lying flush with the gel and the basic end was aligned with the right edge of the gel. A
comb with a single lane was inserted on the left of the gel and 1% agarose made with 1x
running buffer was poured over the strip and marker lane and allowed to set. Gels were
transferred into the electrophoresis tanks and run at 90 V for approximately 15
minutes, or until the proteins could be seen to have left the strip and entered the gel,
79
and then run at 120 V, typically until the 25 kDa marker reached the bottom of the gel,
to enable better separation of FIH‐1 and Myc‐6His‐FIH‐1.
2.16.6 Visualisation
Following second dimension separation, proteins were typically transferred to
nitrocellulose and results visualised by immunoblotting using the various anti‐FIH‐1
antibodies, or the anti‐Myc antibody. Coomassie stain and SyproRuby stain were used
to visualise total protein.
2.17 Immunoprecipitation
2.17.1 Cell lysate preparation
Cells were washed three times with ice‐cold PBS. Weakly adherent cells were harvested
into 1.5 ml tubes by pipetting with PBS, and samples were centrifuged for 2 minutes at
1500 rpm to pellet the cells prior to resuspension of the pellet in WCEB, supplemented
with fresh 1 mM PMSF and 1 mM DTT. Typically, 3 ml of PBS was used for harvesting
cells into 3x 1.5 ml tubes. Each pellet was resuspended in 200‐500 μl WCEB, depending
on the desired concentration of the lysate. Cells were pipetted up and down in WCEB
and incubated on a rotating platform at 4°C for 45 minutes. Lysates were then
centrifuged for 30 minutes at 14 000 rpm, and the supernatant retained.
2.17.2 Preclearing
Preclearing of cell lysates was performed prior to IP to remove proteins that bound to
Protein A Sepharose beads. This was done by addition of 50 μl of protein A slurry per 1
ml of lysate and incubation on a rotating shaker at 4°C for 2 hours.
80
2.17.3 Antigen‐antibody complex formation
The required amount of total protein (varied between experiments, see figure legends)
was diluted in binding buffer and the antibody added at the desired concentration
(varied between experiments, final dilutions indicated in figure legends). Immune
complexes were allowed to form over 2 hours or overnight by incubation on a rotating
platform at 4°C.
2.17.4 Bead rehydration and blocking
Protein A sepharose beads were rehydrated in MQ H2O according to manufacturer’s
recommended protocol. 100 mg of beads was rehydrated by addition of 1 ml of MQ H20
for 1 hour. The beads were allowed to settle, at which point the supernatant was
discarded and fresh MQ H2O added. This was repeated three times. Protein‐A‐
Sepharose was blocked by addition of 0.5% BSA and shaking for 1 hour at 4°C. Beads
were then washed three times with MQ H2O following final resuspension in binding
buffer. If beads were to be stored, 0.02% sodium azide was added.
2.17.5 Immune complex binding to resin
50‐100 μl of prepared, blocked slurry was added to the immune complex mixture for 1
or 2 hours, or overnight, rotation at 4°C.
2.17.6 Elution
IPs were pelleted by gentle centrifugation (3000 rpm for 2 minutes at 4°C) and pellets
washed three times with binding buffer. Elution of bound protein was achieved by a
variety of methods. Commonly, pelleted resin was resuspended in SDS sample buffer
and boiled for 5 minutes, prior to centrifugation and removal of the supernatant for
SDS‐PAGE.
81
2.18 Notch‐Affinity Pulldowns
2.18.1 Notch construct expression
50 ml cultures containing 0.1% carbenicillin and 2% glucose were grown overnight. The
next day, 30 ml of overnight culture was gently pelleted to remove the glucose and
resuspended in 15 ml of LB, which was then used to inoculate 500 ml of LB with 0.1%
carbenicillin. Cultures were grown at 37°C for 1.5‐2 hours, or until the OD600 reached
0.4‐0.5, at which point protein expression was induced by addition of a final
concentration of 1 mM IPTG. Protein expression continued for 2 hours at 37°C before
cultures were pelleted and pellets frozen prior to next day purification.
2.18.2 Notch construct purification
Each pellet was resuspended in lysis buffer with added 1 mM PMSF and 0.5 mM DTT to
give 30 ml of suspension/L culture. Cells were lysed by passing three times through a
cell disruptor (Microfluidics) that was chilled on ice, and lysates centrifuged for 45
minutes at 14 000 rpm. The supernatant was retained and incubated with 1 ml of nickel‐
resin slurry for 1 hour at 4°C to allow binding. The resin was washed three times with
lysis buffer with 10 ml imidazole and stored in lysis buffer with 10 mM imidazole and
0.2% sodium azide. 50 μl of resin was boiled in SDS sample buffer and eluted protein
separated by SDS‐PAGE and stained with coomassie to ensure purified 6His‐Trx‐Notch1‐
Ank1‐4.5‐6His.
2.18.3 Notch‐affinity purification
Lysates were prepared from either HeLa cells or the number 9 HeLa cell line stably
overexpressing Myc‐6His‐FIH‐1, by washing cells three times with PBS and then lysing in
WCEB with phosphatase inhibitors and PMSF. The indicated amount of lysate (200 μl to
2 ml) was incubated with the indicated amount of Notch‐loaded resin (50 μl to 200 μl)
made up as a slurry in lysis buffer with 10 mM imidazole and 0.2% sodium azide added.
82
Binding occurred over a 2 hour incubation on a rotating platform at 4°C. Unbound
protein was removed by 4x washing in lysis buffer, and bound protein was eluted using
either 250 mM Imidazole in lysis buffer, heating in SDS sample buffer, or digestion with
Thrombin. Thrombin digestions were carried out assuming the nickel resin was
saturated with 6His‐Trx‐Notch1 Ank1‐4.5‐6His. 1 ml of prepared resin can hold 10 mg of
protein and 1 unit of thrombin can theoretically cleave 500 μg of protein. A 10x excess
of thrombin was used for the cleavage of 6His‐Trx‐Notch1 Ank1‐4.5‐6His, as 10 units
was added for every 50 μl of resin in the pulldown. Thrombin digestions were carried
out for 16 hours at room temperature in thrombin cleavage buffer.
2.19 Ni2+‐Affinity Purification of FIH‐1 from HeLa cells
Each 175 cm2 flask of cells was washed with 1x PBS and lysed by the addition of 10 ml
Guanidine Extract Buffer with freshly added DTT. Flasks were incubated for 2 hours at
4°C on a rocking platform to allow for reduction of proteins. Lysates were then
combined in a Schott bottle covered in foil and alkylated by the addition of 0.925 g
iodoacetamide to 100 ml of lysate, and stirring for 2 hours at RT. Lysates were then
clarified by ultracentrifugation for 4 hours at 60 000 rpm and filtered through a 0.45
micron filter and stored overnight at 4C. A 5 ml HiTrap Chelating HP column (GE
Healthcare) was charged with Nickel and equilibrated with 25 ml Guanidine Wash buffer
before lysate was applied to the column at a flow rate of 1ml/minute. Consecutive
washes of 50 ml guanidine wash buffer, 50 ml SB3‐10 wash buffer and 50 ml SDS wash
buffer preceded elution with 10 ml of SDS elution buffer, collected in 1 ml fractions.
83
2.20 Phosphatase Treatments
2.20.1 Cell lysis
Cell lysates were prepared by first washing cells 3x with ice‐cold PBS and then pipetting
the loosely adherent cells into microfuge tubes in PBS. Cells were gently pelleted and
PBS removed, and two separate lysates were prepared by addition of either Lysis Buffer
II (150 mM NaCl, 1 % SDS, 50 mM Tris pH 7.5, 1 % NP‐40, supplemented with fresh 1
mM PMSF and protease inhibitor cocktail) or Lysis Buffer II with phosphatase inhibitors
added. Samples were rotated at 4°C for 1‐2 hours. Lysates were then centrifuged at 14
000 rpm for 30 minutes at 4°C, and total protein quantified by Bradford Assay.
2.20.2 Phosphatase treatments
200 μg of lysate was diluted to a total volume of 77.5 μl, to which 10 μl of 20 mM MnCl2
and 10 μl of 10x Lambda phosphatase buffer was added, giving a total volume of 97.5
μl. 2.5 μl (1000 units) of lambda phosphatase was added to the “+” phosphatase
reaction tube, and 2.5 μl of buffer only was added to the negative reaction tube
containing the lysate prepared with phosphatase inhibitors. Both the reaction tube and
negative control tube were incubated at 30°C for 1 hour.
2.20.3 Positive control
Following the 16 hour incubation with or without lambda phosphatase, 20 μl of each
sample (40 μg total protein) was taken, boiled with SDS sample buffer and separated by
10% Tris/glycine SDS‐PAGE. Western blotting with the RNA polymerase II CTD repeat
YSPTSPS (phospho S5) antibody (ab70158) a 1:2500 dilution was performed to visualise
success of the phosphatase treatments (217 kDa, detects at 260 kDa). Where required
by lack of non‐specific bands, a rat β‐tubulin antibody (1: 2 000) served as a loading
control.
84
2.20.4 2‐DE
The remaining 80 μl (160 μg total protein) of each sample was precipitated using the
Amersham 2‐D Clean UP kit, and final pellets resuspended in IEF buffer and quantified
by Bradford Assay. 100 μg of each sample was separated by 2‐DE as described
previously. Western blotting with anti‐Myc antibody 9E10 was used to visualise the
results.
2.20.5 Internal reference protein
A Myc‐tagged protein of different size to Myc‐6His FIH‐1 was used as an internal
standard to enable more accurate comparison of the relative positions of phosphatase‐
treated FIH‐1 compared to untreated FIH‐1. The 31 kDa Myc‐6His‐mNotch1RAM(1753‐
1859)‐6His, encompassing the RAM domain of Notch, was expressed in bacteria and
purified by Nickel affinity chromatography for this purpose. Purified Myc‐6His‐
mNotch1RAM‐6HIs was precipitated using the 2‐D clean up kit, resuspended in IEF
buffer and quantified by Bradford assay. Immediately prior to IEF of phosphatase‐
treated samples, 0.1 μg of the internal standard protein was added to each sample.
2.21 In vitro Phosphorylation Assay
The protocol for in vitro phosphorylation of FIH‐1 by cell lysate was based on a
published example of a similar experiment (Rybina et al., 1997).
2.21.1 Protein expression and purification
Trx‐6His‐FIH‐1 and 6His‐HIF1α‐CAD were bacterially expressed and purified as described
above, with the following difference: following the final wash with 10 mM imidazole,
85
resin was washed with 35 mM imidazole and then proteins were stored attached to
nickel resin at 4°C in the presence of 0.2% sodium azide.
2.21.2 Preparation of cell lysate
Lysate was prepared by washing cells 3x with ice‐cold PBS and adding native lysis buffer
directly to cell dishes. Lysis buffer was supplemented with fresh 1 mM PMSF, 1x
protease inhibitor cocktail, 1 mM NaF, 1 mM Na3VO4, 20 mM β‐glycerophosphate,
except in samples to be treated with lambda phosphatase where phosphatase inhibitors
were omitted. Following a 30 minute incubation on ice, lysates were centrifuged at 14
000 rpm for 30 minutes at 4°C and the supernatant retained.
2.21.3 Phosphorylation Assay
30 μl of the FIH‐1 or HIF‐CAD‐bound nickel resin slurry was pipette into a microfuge
tube. A concentrated ATP solution was made by adding gamma [ϒ‐32P]‐ATP to 10 mM
ATP, and this solution was added to either cell lysate or buffer only such that the final
concentration in the lysate would be 150 μM ATP and 50 μCi. 600 μl of the ATP/lysate
mixture was added to the tubes containing the resin to initiate the reaction. Tubes were
incubated on a 37°C heating block for 3 minutes and then pelleted using centrifugation.
The lysate was removed and the resin washed 3x with Native Lysis Buffer. The bound
protein was then eluted by boiling in the presence of 40 μl of SDS sample buffer. The
resin was then pelleted and the sample buffer containing eluted protein separated by
SDS‐PAGE.
2.21.4 Phosphatase treatment
As a negative control, phosphatase treated samples were included. Lysates were
prepared as described above, except phosphatase inhibitors were excluded from the
Native Lysis Buffer. Reactions occurred as described above. After the final wash, 1000 U
(2.5 μl) of lambda phosphatase was added to each sample in the presence of 100 μl of
86
lambda phosphatase reaction buffer (2 mM MnCl2, 1x lambda phosphatase buffer).
Reactions were incubated at 30°C for 1 hour, then the resin was washed 3x in Native
Lysis Buffer, and then protein eluted by boiling the resin in 40 μl of SDS sample buffer.
2.21.5 TEV cleavage
For samples where resin‐bound FIH‐1 was to be eluted by TEV cleavage, the reactions
proceeded as described above. Following the final wash of the resin, 30 μl of TEV
reaction buffer (made using supplied 20x TEV buffer and 0.1 mM DTT), including 10
units (1 μl) of TEV protease, was added to the resin and incubated for 1 hour at 30°C.
The resin was then pelleted and the supernatant retained for SDS‐PAGE. The resin was
retained and any remaining protein was eluted by boiling in 40 μl of SDS sample buffer.
87
CHAPTER 3
CHARACTERISING
FIH‐1 EXPRESSION
89
3 CHARACTERISING FIH‐1 EXPRESSION
3. 1 Introduction
At the commencement of this PhD project, there was a scarcity of data regarding the
expression of FIH‐1 in tissues. A primary aim of this work was to characterise the spatial
and temporal expression of FIH‐1 in a variety of normal mouse tissues. Given the role
that FIH‐1 plays in regulating the HIF pathway, and the ensuing impact of this pathway
on human health and disease, a further aim of this work was to investigate the
expression of FIH‐1 in a variety of tumour tissues. As HIF activity has been
demonstrated to be important in the progression and malignancy of solid tumours, the
levels of the enzyme responsible for decreasing HIF activity could logically be
hypothesised to be decreased in instances of elevated HIF‐target gene expression.
3.1.1 Expression and importance of HIF‐1α and HIF‐2α in cancer progression
HIF‐1α and HIF‐2α expression have been widely investigated in both normal and tumour
tissues and found to be widely expressed. HIF‐1α message has been found in every
human, mouse and rat tissue analysed to date (Wiener et al., 1996), whereas HIF‐2α
exhibits a much more restricted expression pattern and is expressed predominantly in
endothelial cells, and in the kidney and lung (Ema et al., 1997).
HIF‐α is also present in many different cancer types, including bladder, breast, colon,
glial, hepatocellular, ovarian, pancreatic, prostate, and renal carcinomas, and its
overexpression or normoxic stabilisation is associated with more aggressive cancer
phenotypes (Talks et al., 2000). The presence of HIF‐α protein in neoplastic tissues has
been suggested to be a result of both the hypoxic tumour microenvironment, as
90
evidenced by the perinecrotic expression pattern observed, and also as a result of
genetic alterations, as expression is often observed in oxygenated tumour regions well
situated near blood vessels. Intratumoral regions of hypoxia can stabilise and activate
the transcription factors via the traditional oxygen‐dependent pathway. HIF can also be
upregulated by growth factors independently of oxygen concentration or by loss or
inactivation of VHL. As described in more detail in Chapter 1 of this thesis, the activation
of HIF‐mediated transcription is favourable for tumour growth and cancer progression,
thus elucidating the expression of the regulator of HIF transcriptional activity in cancer
was also of interest to determine whether FIH‐1 regulation is also a contributing factor
to cancer progression. As such, one of the original aims of this project was to compare
expression of FIH‐1 between normal and diseased (i.e. neoplastic) human tissue
samples.
3.1.2 What was known about FIH‐1 expression?
At the time that these aims were defined, knowledge of FIH‐1 expression was limited to
the following pieces of research. When FIH‐1 was first discovered as a HIF‐interacting
protein, expressed sequence tag (EST) data suggested that FIH‐1 was expressed in cells
derived from a wide range of tissues including bladder, brain breast, colon, heart,
kidney, lung, lymph, bone marrow, muscle, nerve, ovary, prostate, skin, testis, tonsil,
and uterus, mirroring the ubiquitous expression of its HIF‐α substrates (Mahon et al
2001). Some years later, FIH‐1 protein expression was investigated in a number of cell
lines and levels were found to be similar in all cell lines examined (Stolze et al., 2004).
Expression was found to be predominantly cytoplasmic, however small amounts of FIH‐
1 were detected in the nucleus, and the subcellular distribution of FIH‐1 did not alter in
hypoxia, or with treatment of cells with hypoxia mimetics (Metzen et al., 2003a; Linke et
al., 2004; Stolze et al., 2004).
91
3.2 Honours Results
3.2.1 Generation of FIH‐1 monoclonal antibody (undergraduate research year, 2004)
The first approach undertaken to address the aim of investigating FIH‐1 expression in
tissues was to conduct immunoblot and immunohistochemical investigations into FIH‐1
expression in a range of normal mouse tissues. As such, anti‐FIH‐1 antibodies suitable
for each application were required.
At the commencement of this work, there were no commercially available FIH‐1
antibodies. Two polyclonal rabbit antisera, designated No.8 and No.9, had been
generated in our laboratory against MBP‐FIH‐1 and were initially used for the western
analysis of FIH‐1 expression in a range of mouse tissues (Lisy, 2004). The No.9 antibody
had been used previously by our lab for immunofluorescence of 293T cells (Linke et al.,
2004). Though the antibodies were suitable for use in blotting human cell culture
lysates, initial experiments utilising either of these antibodies in detection of FIH‐1 from
mouse tissue lysate generated poor results (Lisy, 2004). Western blots of mouse tissue
homogenates displayed high levels of background staining, with strong, dark
background bands present in lanes containing mouse tissue lysates compared with 293T
cell lysates. Despite attempts at optimising blotting conditions to improve the specificity
of the antibody for mouse FIH‐1 detection, a high level of non‐specific staining was
persistently observed. It was thus deemed unlikely that these antibodies would to be
suitable for use for the proposed investigations into FIH‐1 expression in mouse tissues.
This early work necessitated generation of an antibody suitable for detection of mouse
FIH‐1 from whole tissue lysates and for immunohistochemistry. Generation of an anti‐
FIH‐1 monoclonal antibody formed part of my Honours project (undergraduate research
project), which was completed in 2004. The work described in sections 3.2.2 and 3.2.3
92
below was carried out during Honours (2004) and was submitted as part of the thesis
entitled “Characterisation of FIH‐1 Expression” (Lisy, 2004).
3.2.2 Generation of an anti‐FIH‐1 monoclonal antibody
Three mice were immunised with purified full‐length human maltose‐binding protein
(MBP)‐FIH‐1 and their serum assessed by western blotting of 293T cell lysates with
transiently transfected FIH‐1 (FIH‐1‐293T) to confirm an immune response to the
antigen (Figure 3.1). Antisera from all three mice indicated the presence of antibodies
against FIH‐1, evidenced by the detection of an intense band corresponding to an
approximately 40 kDa protein on blots probed using the sera (Figure 3.1). Interestingly,
immune serum from each of the three mice showed detection of a doublet at this
position. Approximately 1200 hybridoma clones were subsequently generated, and
hybridoma supernatants were screened by both enzyme‐linked immunosorbent assay
(ELISA) and western blotting for FIH‐1 detection. One hybridoma, designated 9F6, gave
the strongest response (OD450 of 0.615) in the ELISA and also returned a strong, clean
band at the correct size of FIH‐1 in the western blot screen as well as a second, more
intense band at approximately 35 kDa (data not shown). Hybridoma 9F6 was therefore
maintained in culture for further use and characterisation.
3.2.3 Preliminary use of 9F6
This antibody was then used in preliminary western blots to determine FIH‐1 levels in
both cancer cell lines and in a range of normal mouse tissues (data not shown). These
blots revealed some differences in detection between the 9F6 supernatant and the No.
8 and No. 9 antibodies such that any further use of 9F6 required more rigorous
validation regarding its ability to detect FIH‐1. Specifically, western blots confirmed that
9F6 was consistently able to detect two proteins of similar molecular weight at the
293T pc
WB fWB:
kDa
Antisera from m
1 2
50
3737
Figure 3.1. Immunoblots probed usinfrom three immunised miceFIH‐1‐overexpressing 293T lysate wascut into five strips for immunodetecantibodies (No.8 and No.9) or antisera f( )FIH. (This work was performed as partthe Honours thesis entitled “CharacterLisy).
DNA3.1 FIH‐1 lysate
banti‐FIH PoAbs
No. 8 No.9
mouse #
3
ng anti‐FIH‐1 polyclonal antibodies and antisera
separated by SDS‐PAGE and the subsequent blotsction by either one of two anti‐FIH polyclonalfrom each of the three mice immunised with MBP‐of the Honours project 2004, and was included inisation of FIH‐1 Expression” submitted by Karolina
95
expected position of FIH‐1. Similarly, The No.8 and No.9 antibodies had also been
routinely observed to detect a doublet. Comparison of the two bands detected by 9F6
with the bands detected by the other FIH‐1 antibodies revealed that the larger
(approximately 40 kDa) protein bound by 9F6 corresponded in size to the band detected
by the polyclonal anti‐FIH‐1 antibodies No.8 and No.9. This “upper” band increased in
intensity with transient FIH‐1 overexpression (Figure 3.2). Each antibody also detected a
slightly smaller protein (approximately 37 kDa) that was not enhanced with FIH‐1
overexpression. 9F6 displayed an apparent preference for binding to the smaller protein
(Figure 3.2) and in some experiments, including the representative blot shown in Figure
3.2, could only detect FIH‐1 with overexpression. Increasing the stringency of western
blotting conditions by decreasing the amount of 9F6 antibody or adding skim milk was
able to remove detection of the “upper” band, by 9F6 (data not shown). The “lower”
band that was preferentially detected by 9F6 was, on the other hand, not detected by
the No.8 or No.9 antibodies.
3.2.4 Conclusions from undergraduate research
At the conclusion of the undergraduate research project, a monoclonal antibody had
been generated and had been demonstrated to bind FIH‐1 upon transient FIH‐1
overexpression in 293T cells. However, this antibody appeared to preferentially detect a
protein in mammalian cell extracts with a slightly lower molecular weight than FIH‐1.
Given that this smaller protein may have been an alternative form of FIH‐1, this
preferential detection by the 9F6 monoclonal had significant potential in characterising
this alternative form. The fact that a “lower” band was also detected with numerous
other FIH‐1‐specific antibodies supported the notion that it was an alternative form of
FIH‐1.
sfected
80
113
FIH‐1
Untrans
kDa
37
49
64
26
20
WB: N
Figure 3.2 Immunoblots of cell lysateusing either the No.8 PoAb or 9F6.L f 293T ll i h i hLysates from 293T cells with or withoseparated by SDS‐PAGE and proteinantibody No.8 or monoclonal antibody
sfected
80113
FIH‐1
Untrans
kDa
37
4964
26
20
o.8 WB: 9F6
from 293T cells ‐/+ FIH‐1‐ overexpression probed
i l f d DNA3 1 FIH 1out transiently transfected pcDNA3.1 FIH‐1 werewas detected by blotting with either polyclonal9F6, as indicated.
99
Thus, it was necessary to verify that 9F6 bound FIH‐1 and was also of key interest to
ascertain whether or not both bands detected by 9F6 represented differentially
modified forms FIH‐1. Evidence that suggested that these two bands may indeed have
both represented FIH‐1 included the observation that immune sera from all three of the
mice immunised with MBP‐hFIH‐1 were able to detect two bands in western blots, and
detection of two bands using the No.8 rabbit polyclonal antisera was also routinely
seen. This suggested that each of the immunised animals had individually generated
antibodies either against two different antigens, or against two different forms of FIH‐1.
Given that only the higher band was increased upon overexpression of FIH‐1, it was
thought likely that the lower band was either a modified form of FIH‐1 not favoured by
overexpression, or corresponded to a different protein with a common epitope.
3.3 Further characterisation of 9F6 antigen
3.3.1 9F6 can detect purified and overexpressed FIH‐1
To further confirm that 9F6 was able to detect overexpressed FIH‐1, 293T cells were
transiently transfected with MBP‐FIH‐1 and lysates were separated by SDS‐PAGE before
transfer onto nitrocellulose. Blots were cut into two parts for separate incubation in
either 9F6 supernatant or No.8 serum. Alignment of the resulting blots showed that 9F6
was able to bind the same sized protein as the No.8 antibody, supporting the ability of
9F6 to bind to MBP‐FIH‐1 (Figure 3.3 a). MBP‐FIH‐1 has a calculated size of 83.4 kDa,
and the band detected by both antibodies corresponded to a molecular weight slightly
larger than this predicted size. 9F6 was also found to be capable of detecting
recombinant MBP‐FIH‐1 (Figure 3.3 b).
1l
5l
10l
5l
293T lysate
kDa
a.
64
81
114
182
MBFIH‐
WB: No.8WB: 9F6
64
DAPFIH‐1
c.
DApcDNA 3.1‐FIH‐1p
Figure 3.3 9F6 is able to bind recombin(a) The indicated volumes of lysate froseparated by SDS‐PAGE and the subsewas blotted using either No.8 or 9F6. (and purified MBP‐FIH‐1 and lysate frseparated by SDS‐PAGE and probed wcoverslips for 3 days. pcDNA 3.1‐FIH‐1 wto coverslips at 72 hours. 5% skim milkwith 9F6 diluted 1:100 in PBS Boundwith 9F6, diluted 1:100 in PBS. Boundconjugate and nuclei visualised using DA
293T lysate
1l 10l
5l20l
0.1l
Purified MBP‐FIH
b.
P‐‐1
64
81
114
182
MBP‐FIH‐1
kDa
64
50
379F6 band
WB 9F6WB: 9F6
PI
PI
nant and overexpressed FIH‐1om 293T cells stably overexpressing MBP‐FIH werequent blot was cut in half as indicated. Each half(b) The indicated volumes of bacterially expressedrom cells stably overexpressing MBP‐FIH‐1 werewith 9F6 antibody. (c) 293T cells were grown onwas transfected in at 48 hours and cells were fixedk was used for blocking prior to detection of FIH‐1d antibody was detected using anti mouse TRITCd antibody was detected using anti‐mouse TRITCAPI stain.
103
Immunofluorescence of 293T cells transiently transfected with pcDNA3.1‐hFIH‐1 using
9F6 also suggested that 9F6 was able to bind overexpressed FIH‐1 (Figure 3.3 c). Control
transfected cells showed mainly cytoplasmic staining, in keeping with the previously
determined subcellular localisation of FIH‐1 (Linke et al., 2004). Transfection of
pcDNA3.1‐hFIH‐1 resulted in markedly greater intensity of cytoplasmic staining of
transfected cells when compared to adjacent non‐transfected cells, where cytoplasmic
staining was moderate.
3.3.2 siRNA‐mediated knockdown of FIH‐1
To ascertain whether 9F6 detected endogenous FIH‐1 and, importantly, whether the
lower band was likely to be a modified form of FIH‐1, siRNA‐mediated knockdown of
FIH‐1 was attempted. The expected outcomes of FIH‐1 knockdown were to either see a
reduction in both bands detected by 9F6, indicating that both bands were FIH‐1, a
reduction in one band only, signifying which band was FIH‐1, or no effect at all on either
band, which would indicate that neither band was FIH‐1.
3.3.2.1 Plasmid‐based siRNA
Initially, the pSUPER plasmid‐based siRNA system was used, where the desired siRNA
sequence is transcribed in cells following transfection. Four siRNA sequences that
corresponded to 22 nucleotides stretches of FIH‐1 beginning at nucleotides 85, 336, 960
and 1020 relative to the start codon (GenBank accession number NM_017902) were
cloned into pSUPER by Anthony Fedele. A random 22 nucleotide sequence that
displayed no significant homology to any human or mouse sequences was used as a
scrambled negative control. Results achieved using this method were variable.
Parameters such as cell confluency, the amount of pSUPER constructs used, the amount
of Lipofectamine or FuGENE transfection reagent used, transfection times, and growth
104
times of cells post‐transfection were altered with the aim to improve the outcome. In
summary, 293T cells were transfected at 30% or 50% confluency with 2‐4 μg of each
pSUPER construct using either Lipofectamine or FuGENE, with cotransfection of a GFP
construct to provide an estimate of transfection efficiency. Cell media was replaced
after 4 hours and cells grown for either 24, 48 or 72 hours. Despite these various
measures to improve transfection efficiency and knockdown of FIH‐1, the best
transfection efficiency observed by assessing the number of GFP‐positive cells was 70%,
and this was not associated with a decrease in FIH‐1 in siRNA‐transfected cells, as
indicated by blotting with the No.8 and No.9 antibodies and there was no decrease was
observed in the bands detected by 9F6 (data not shown).
3.3.2.2 siRNA oligonucleotides
The second method employed was the direct transfection of anti‐FIH‐1 siRNA
oligonucleotides. Two siRNA oligonucleotides were designed and synthesised according
to previously published sequences and targeted nucleotides 91–111 (F1) and 160–180 (
F2) relative to the start codon (GenBankTM accession number NM_017902), as was a
scrambled negative control (Stolze et al., 2004). 293T cells were transfected at a final
siRNA concentration of 20 nM at 0 hours and again at 24 hours prior to harvesting cells
at 48 hours and analysing FIH‐1 protein levels by western blot using both the No.8
antibody and 9F6.
Relative to untransfected and scramble transfected control, transfection of either of the
two siRNA oligonucleotides targeting different mRNA sequences of FIH‐1 decreased
levels of the 40 kDa protein detected by the No.8 polyclonal antibody, confirming that
this indeed did correspond to endogenous FIH‐1 (Figure 3.4). The “lower” band was not
detected by the No.8 antibody here. FIH‐1 knockdown with these siRNA constructs
failed to decrease the intensity of the lower molecular weight protein detected by 9F6.
ed
a.
Untransfecte
Scrambled
F1 siRNA
F2 siRNA
81
kDa
37
50
64
81
* F
*
WB: No.8
26
Figure 3.4 Protein detected by 9F6 not293T cells were transfected with a totnegative control or the two siRNA oliseparated by SDS‐PAGE and blottingantibody No.8 (a) or 9F6 (b). * denotes
ed
b.
81
Untransfecte
Scrambled
F1 siRNA
F2 siRNA
kDa
FIH‐1 37
50
64
81
“Lower” band
26
WB: 9F6
reduced by FIH‐1 specific siRNAtal concentration of 20 nM of either a scrambledigonucleotides F1 and F2. 20 μg of lysates wereperformed using either the anti‐FIH polyclonalnon‐specific band.
107
This protein was detected at a similar level in cells transfected with scrambled siRNA or
siRNA targeting FIH‐1, indicating that this lower band was is unlikely to represent FIH‐1.
There remained the possibility that this smaller protein resulted from a splicing event
where the spliced mRNA no longer contains the sequences targeted by the siRNAs. The
sequences targeted encompass nucleotides 86–106 (F1) and 172–192 (F2) of FIH‐1
mRNA, both at the beginning of the sequence where a splicing event would make only a
small reduction in molecular weight of the resulting protein. However, there is no
evidence supporting or suggesting that splice variants of FIH‐1 exist.
3.3.3 9F6 is an IgM monoclonal antibody
A more definitive strategy for determining whether 9F6 binds FIH‐1 was IP of the
antigenic protein(s) from cell lysate followed by identification using mass spectrometry
(MS). Preliminary attempts to immunoprecipitate the 9F6 antigen began during the
Honours undergraduate year. In these experiments, 9F6 was first incubated with 293T
cell lysate to enable antibody‐antigen complex formation prior to incubation with
protein A‐linked sepharose beads to bind any formed immunocomplexes. Unbound
proteins were removed by washing and any bound complexes eluted and analysed by
western blotting. 9F6 failed to immunoprecipitate any antigen from cell lysate, even
from FIH‐1‐overexpressing cells, whereas the No.8 polyclonal antibody was able to
immunoprecipitate the upper band only (i.e. the band that increased in intensity with
FIH‐1 overexpression), but not the smaller protein commonly detected by No.8 in
western blots (data not shown).
The failure of the protocol using 9F6 and protein‐A‐sepharose may have been due to
poor affinity of the 9F6 heavy chain for protein A. Different classes of antibodies possess
varying properties depending of the composition of their heavy chain. For example, IgG
immunoglobulins exist as monomers and are able to bind protein A whereas IgM
108
antibodies interact to form multimers and do not bind to protein A or protein G. It was
therefore important to classify the 9F6 constant region to determine the appropriate
matrix to use in immunoprecipitation experiments. A Mouse Typer Sub‐Isotyping Kit
was used to ascertain the isotype of 9F6. This method determined antibody class by an
ELISA where FIH‐1 was first adsorbed onto a microplate to enable binding of 9F6,
followed by incubation with rabbit anti‐mouse antibodies specific for each
immunoglobulin class. Goat anti‐rabbit antibodies conjugated to HRP revealed that 9F6
is an IgM antibody and provided a likely explanation for the inability to IP the 9F6
antigen using protein A. To circumvent this limitation, two alternative methods for
immobilising 9F6 were attempted, as described below.
3.3.4 Immunoprecipitation of 9F6 antigen by thiophilic interaction chromatography
The first method for purifying the 9F6 antigen was to utilise HiTrapTM IgM HP
Purification columns which contain a thiophilic adsorption medium coupled to the
sepharose and are used for purification of IgM antibodies from hybridoma
supernatants. The aim was to apply 9F6 supernatant to the column to bind the antibody
and then pass cell lysate over the immobilised antibody to enable capture of the
antigenic protein(s). Elution of the antibody‐antigen complex and separation on SDS‐
PAGE would be followed by band excision and mass spectrometry for protein
identification.
Thiophilic adsorption is based on the non‐covalent affinity of proteins for thioether‐
containing ligands coupled to sepharose in the presence of an antichaotropic (water
structuring) salt. The columns are packed with 2‐mercaptopyridine attached to
sepharose as the immobilised ligand and the salt used was ammonium sulphate. The
column first had to be equilibrated with thiophilic binding buffer containing 20 mM
109
sodium phosphate and 0.8 M ammonium sulphate, and the same concentration of
ammonium sulphate had to be present in the antibody solution
PD‐10 desalting columns were used to buffer exchange the antibody into the required
binding buffer, followed by filtration through a 0.45 μm filter. Following equilibration of
the column according to manufacturer’s instructions, 90 ml of the 9F6 solution (0.08
mg/ml), or a total amount of approximately 7 mg of IgM, was passed through the
column. The maximum capacity of the column is 5 mg of IgM, and an excess amount of
antibody was applied in an attempt to saturate the column. Following washing with
binding buffer to remove unbound contaminants, the cell lysate was added to the
column. In order to maintain the high concentration of salt necessary for retention of
the IgM immunoglobulin on the column, 293T cells were lysed in WCEB, and 1 ml of
lysate was then diluted 50‐fold into the thiophilic binding buffer. The column was then
washed with 3 bed volumes of binding buffer to remove unbound proteins, and any
antibody/antigen complexes were eluted by passing 10 ml of 20 mM sodium phosphate
without any ammonium sulphate over the column in 1 ml fractions. The presence of the
9F6 antigen was revealed by western blotting using 9F6, and showed that the antigen
had indeed been retained on the column and eluted in fraction 3 (Figure 3.5). Here,
293T lysate equivalent to 1% and 0.1% of the total input lysate was loaded and flow
through lysate from the column was concentrated, and the equivalent of 1% flow
through was loaded, indicating a decrease in the amount of the 9F6 antigen. Each
elution fraction was run, and elution fraction 3 was found to contain the desired
protein. However, subsequent SDS‐PAGE and coomassie or silver staining failed to
produce a visible band for excision and further analysis (data not shown). Subsequent
purifications were performed with increased amounts of input lysate (5 ml and 10 ml),
however silver stained gels failed to resolve a clear band at the expected size for the
9F6 antigen.
ut
nput
dlysate
64
81114182
50
1% Inpu
0.1% In
Dep
leted
kDa
37
26
Figure 3.5. Thiophilic Interaction Chrom90 ml (0.08 mg/ml) of 9F6 supernatanbinding buffer prior to application to a
b d t i b f 293remove unbound proteins before 293column was then washed and antibodInput and depleted lysate and 20 μl oPAGE and protein detected by 9immunoprecipitations.
d lysate
Dep
leted
1 2 3
Elution Fractions
matographynt was filtered and buffered in the appropriateHiTrap IgM column. The column was washed to
3T l t d th h th l Th
WB: 9F6
3T lysate was passed through the column. Thedy‐antigen complexes eluted in 1 ml fractions.of each elution fraction were separated by SDS‐9F6 WB. Representative of 3 independent
113
3.3.5 Immunoprecipitation of 9F6 antigen using CNBr‐activated sepharose
The second method attempted for purification of the protein of interest was to
covalently crosslink purified 9F6 to the stationary phase to enable capture of the
antigen. CNBr‐activated sepharose was used, which allows for the covalent attachment
of proteins via the proteins’ amino groups. The aim of this was to try and prepare a
more stable resin with a higher capacity to bind the 9F6 antigen.
CNBr‐activated‐sepharose was prepared according to manufacturer’s instructions. The
9F6 antibody was purified using a HiTrap IgM column as described above and then
dialysed into the required coupling buffer of 0.1 M sodium bicarbonate and 0.5 M
sodium chloride. The antibody was then bound to prepared CNBr‐activated sepharose
and the success of binding was crudely assessed by measuring the absorbance at 280
nm of the antibody solution before and after incubation with the CNBr‐sepharose. The
A280 of the solution after binding decreased four‐fold, indicating that much of the
antibody had indeed bound to the resin. Any unbound 9F6 was washed off the resin and
the resin was then blocked with 1 M ethanolamine prior to further washing and
incubation with 293T lysate. 9F6‐CNBr‐sepharose was then used to immunoprecipitate
the antigenic protein from lysate. Elution of the antigen was achieved through
disruption of the immunocomplex by incubation with a buffer containing reducing agent
BME or by elution of both antigen and antibody by boiling the resin.
Figure 3.6 is representative of the best result achieved from six purifications. Here, the
9F6 antibody was treated in three different ways prior to attachment to the CNBr‐
activated sepharose. 9F6 antisera was either directly diluted 1:2 into coupling buffer
(Lanes marked 9F6); 9F6 antisera was dialysed against coupling buffer (Lanes marked
lysed 9F6
1% input
0.1% input
Purified
&Dia
150250kDa
37
25
15010075
50
25
20
W
Figure 3.6. CNBr‐Sepharose Immunopre9F6 antibody prepared as indicated bef9F6 antibody prepared as indicated befconjugated resin was then incubated witelution and SDS‐PAGE. Each sample waslanes, in order to enable comparison betRepresentative of > 5 independent purifsamples.
lysed 9F6
Dialysed 9F6
9F6
Dialysed 9F6
Purified
&Dia
0.1% input
9F6 antigen
WB: 9F6 Coomassie
ecipitationfore binding to CNBr linked sepharose beads 9F6fore binding to CNBr‐linked sepharose beads. 9F6‐th 293T cell lysate for four hours prior to washing,s run in duplicate in addition to 1% and 0.1% inputtween WB and coomassie stained gels.fications. See text (Section 3.3.5) for details on 9F6
117
dialysed 9F6); or the antibody was purified using the HiTrap IgM columns and then
buffer exchanged by dialysis against coupling buffer (Lanes marked purified and
dialysed 9F6). Dialysis was performed by dialysing 1 ml of 9F6 supernatant against 500
ml coupling buffer at 4°C for 24 hours, with one buffer exchange. Binding of 9F6 to the
resin was determined by measuring the A280 of the antibody solution before and after
binding. Lysate prepared from 293T cells was then incubated with each preparation of
IgM‐linked sepharose for 4 hours, prior to washing, elution and SDS‐PAGE. Eluted
protein from each of the conditions is shown in Figure 3.6, where each sample was
separated and run in duplicate for both immunoblotting and coomassie stain. The
approximately 37 kDa band detected by the 9F6 antibody is clearly evident in the 0.1%
input lane, and a weak band corresponding in size to this bad is also seen in each IP in
both the immunoblot and the coomassie stained gel.
This band was excised and identified by MS as Heterogeneous Nuclear Riobonucleo
Protein B1 (hnRNP B1). hnRNP B1 is a 37 kDa protein with no significant sequence
homology with FIH‐1, as seen in a Clustal W alignment of the two human protein
sequences (data not shown). Whether hnRNP is the legitimate antigen of 9F6 or simply
a highly abundant contaminating protein is unknown, as further verification of hnRNP as
the 9F6 antigen was not carried out. It is however unlikely to represent a true antigen of
9F6 as hnRNP has been previously reported to be a common contaminant in the
purification of lowly abundant proteins of similar molecular weight, and has been found
to migrate with the desired protein in SDS‐PAGE (Jiang et al., 2008). Despite
immunoblotting showing the presence of desired proteins, the relative high abundance
of hnRNP has been found previously to occlude identification of other proteins in the
sample. As 9F6 has been observed to bind to FIH‐1 when FIH‐1 is overexpressed, it is
possible that FIH‐1 was present in the immunoprecipitated sample yet its presence was
masked by this highly abundant contaminating protein. It is also a possibility that 9F6
118
binds more strongly to another unknown protein with a similar epitope to FIH‐1, and
that identification of this protein was also occluded by the highly abundant hnRNP.
3.3.6 9F6 antigen is not FIH‐1
A conclusive method for further demonstrating that the band detected by 9F6 is not
FIH‐1 presented itself with the generation of MEF cells lacking FIH‐1 (Zhang et al 2009).
Zhang and colleagues created FIH‐1 knockout mice, and the MEFs were originally
cultured from homozygous floxed mouse embryos infected with an adenoviral vector
expressing Cre‐recombinase to generate the FIH‐1 null line. Untreated cells that
retained FIH‐1 and served as a control MEF cell line. Western blot analysis of these cell
lines would definitively demonstrate whether the lower band was an isoform of FIH‐1 or
not.
Lysates were prepared from wildtype (wt) MEFs and FIH‐1 deficient (ko) MEFs and
separated by SDS‐PAGE. Subsequent western blots of these extracts revealed that the
antigen detected by 9F6 was present at equal levels in wt and ko MEFs (Figure 3.7).
Probing the same samples using the anti‐FIH antibody No.8 and a newly available
commercial anti‐FIH‐1 antibody NB 100‐428 confirmed the absence of the 40 kDa upper
band in the ko MEFs compared to the wt MEFs. The NB 100‐428 anti‐FIH‐1 rabbit
polyclonal antibody was not able to detect any protein corresponding in size to the
lower band detected by 9F6 and commonly observed with other anti‐FIH‐1 specific
antibodies. The No.8 blot reveals a high degree of background staining at the conditions
used, including a significant band that corresponds in size to the 9F6 band, which
remained unaltered between wt and FIH‐1 ko MEFs.
This definitively demonstrated that whilst 9F6 was capable of detecting overexpressed
EFs
EFs
EFs
NB100 428 #8WB:
150
100
75
Wt M
Wt M
KO M
E
kDa
37
75
50
37
Figure 3.7 Detection of FIH‐1 in MEFs20 μg of lysates from either wt MEFprotein detected by blotting usng eithe
EFs
EFs
EFs
9F6
Wt M
KO M
E
KO M
EFIH‐1
“Lower” band
Fs or KO MEFs were separated by SDS‐PAGE andr NB 100‐428, No.8 or 9F6, as indicated.
121
FIH‐1 under certain conditions, FIH‐1 was not the predominant antigen bound by 9F6.
Given the original intention of this work was to generate an FIH‐1 specific antibody for
use in techniques such as immunohistochemistry, further work leading to the
identification of the protein or use of this antibody for analysis of FIH‐1 expression was
subsequently abandoned.
3.3.7 Use of polyclonal anti‐FIH‐1 antibody to investigate FIH‐1 expression
Due to the failure to create a monoclonal antibody useful in the detection of FIH‐1,
polyclonal antisera was used instead in a study investigating the levels of FIH‐1 protein
in six different cell lines where differences in the level of FIH‐1 activity had been
suggested from reporter assays, as described above and published in (Bracken et al.,
2006). Despite the differences observed in CAD reporter activity, inferring differences in
the activity of FIH‐1 in its hydroxylation and subsequent silencing of the CAD, FIH‐1
protein levels remained consistent in all six cell lines investigated (Figure 3.8). Where
there were subtle differences observed in protein levels relative to the α‐tubulin loading
control, these were not consistent with the activity levels of FIH‐1. For example, both
HeLa and PC12 cells have slightly less FIH‐1, and 293T cells and HepG2 cells have a
greater signal. However, FIH‐1 activity in both 293T cells and HeLa cells was observed to
be lower than in both PC12 and HepG2 cells, so any subtle differences in FIH‐1 protein
levels cannot attribute for the differences in CAD reporter activity observed (Bracken et
al., 2006).
3.4 Summary and Discussion
Characterisation of FIH‐1 expression in mouse tissues required the generation of an
antibody for use in immunoblotting of mouse tissue lysates and immunohistochemistry
Figure 3.8 FIH‐1 expression in cancer ce20 μg of total lysate from each celtransferred to nitrocellulose. Blottinglevels of FIH‐1 protein in each cell lysaFigure published in Bracken et al 2006.
ell lines (Bracken et al 2006)ll line was separated by 12.5% SDS‐PAGE andwith the anti‐FIH‐1 No.8 antibody showed theate, and α‐tubulin was used as a loading control.
125
of mouse tissues sections. One hybridoma supernatant, 9F6, was found to secrete an
antibody able to detect a protein of similar size to endogenous FIH‐1 and able to detect
overexpressed FIH‐1 in immunoblotting and immunohistochemical applications.
However, further characterisation of the antibody showed that the chief protein
detected by 9F6 was not FIH‐1, as this protein was present in FIH‐1 ko MEFs at
equivalent levels to wt MEFs. MS analysis of a protein immunoprecipitated by 9F6
identified hnRNP B1 as the potential antigen, however, as this protein is a common
contaminant in immunoprecipitation, the definitive nature of the 9F6 antigen remains
unknown.
3.4.1 Expression of FIH‐1 in cancer: paper published
During the course of this work, a research paper entitled “Use of novel monoclonal
antibodies to determine the expression and distribution of the hypoxia regulatory
factors PHD1, PHD2, PHD3 and FIH‐1 in normal and neoplastic human tissues” was
published (Soilleux et al., 2005). This paper described the immunohistochemical
interrogation of an extensive array of healthy human tissues and cells including but not
limited to lung, oesophagus, stomach, liver, gallbladder, pancreas, kidney, cervix, ovary,
tonsil, lymph node, thymus, testis, breast, placenta and umbilical cord for expression of
not only FIH‐1 but also of PHD1, PHD2 and PHD3. Additionally, a small panel of
neoplastic tissues was investigated to provide an indication of the relative expression
levels of the HIF hydroxylases between normal and cancerous tissue. While
unfortunately negating the first aim of this thesis, this research offered some interesting
findings regarding FIH‐1 and PHD1‐3 expression in healthy tissue and in solid tumour
tissue.
In brief, the data showed that expression of PHD1 and PHD3 was markedly reduced in
renal carcinoma compared with normal kidney tissue, and that PHD2 protein levels
126
were also lower in the carcinoma section versus the corresponding normal tissue.
However, expression of FIH‐1 was generally consistent between the diseased and
normal tissue samples, although the authors concluded that FIH‐1 expression was
increased in renal carcinoma compared to normal kidney tissue. As the authors did not
assess expression of HIF target genes in this study it is impossible to know whether FIH‐
1 is exerting any repressive effects upon the HIF CAD. In the absence of reduced FIH‐1
protein levels, it is interesting to speculate that in these neoplastic tissues where it is
advantageous for the cells to activate HIF target gene expression (and indeed there is a
reported selective pressure on cells to do so), that FIH‐1 may be present but that the
enzyme may not retain full activity in hydroxylating HIF‐α substrates. It is possible, due
to the haphazard nature of intratumoral vasculature resulting in regions of hypoxia, that
FIH‐1 activity may be decreased in hypoxic regions due to the limited availability of
oxygen. However, coupled with data that described physiologically significant FIH‐1
activity even at 0.2% oxygen (Stolze et al., 2004), it is also interesting to contemplate
the existence of any mechanism(s) other than oxygen levels that may be acting to
reduce the activity of FIH‐1 and therefore favour HIF activity, such as PTM of FIH‐1 that
may modulate behaviour of the enzyme.
3.4.2 Expression of FIH‐1 in breast cancer: Is subcellular localisation of FIH‐1
important?
One possible mode of regulating FIH activity may be via modulating its subcellular
distribution and affecting colocalisation with substrates. Tan et al investigated both HIF‐
1α and FIH‐1 expression in breast cancer, as well as expression of HIF target gene CA9
(Tan et al., 2007). HIF has been previously implicated in breast cancer progression, and
cancers that exhibit elevated HIF‐1α expression tend to have a more aggressive
phenotype. This comprehensive study examined expression of FIH‐1, HIF and HIF target
gene CA9 in 295 breast carcinoma tissue cores. The data revealed that the vast majority
127
of breast cancer tissue sections (81%) were positive for FIH‐1, and that of those, 48.5%
showed immunoreactivity in both the cytoplasm and the nucleus. However, 18% and
14% of tissues showed FIH exclusively in the cytoplasm or nucleus, respectively,
suggesting that movement of FIH‐1 between the cytoplasm and the nucleus may be a
regulated process. Furthermore, these differences seemed to impact upon the
expression of HIF target gene CA9 such that any nuclear FIH‐1 correlated negatively
with HIF target gene expression, and exclusive cytoplasmic FIH‐1 expression was
associated with increased CA9 expression. In tumours displaying any nuclear FIH (either
exclusively nuclear or in conjunction with cytoplasmic expression), CA9 expression was
decreased, and tumour grade was lower. This also correlated with an increased
probability of long‐term disease‐free survival. However, tumours that exhibited nuclear
exclusion of FIH‐1 were of a higher grade, had elevated CA9 expression and a higher
incidence of disease recurrence. These data suggest that when FIH‐1 is present in the
nucleus, it is able to reduce the severity of tumours, possible via decreasing HIF‐
mediated transcription of target genes. These data also suggest the possibility that FIH
may be actively excluded from the nucleus in aggressive cancer, or conversely that
cancer cells which contain mutations leading to nuclear exclusion of FIH‐1 give rise to
more aggressive tumours due to the increased expression of HIF‐target genes, and
suggests the existence of an active process that regulates FIH‐1 distribution.
3.4.3 Other studies examining FIH‐1 expression in cancer
Another study published during the course of this work described investigations into
expression of the HIF hydroxylases in non‐small cell lung cancer (Giatromanolaki et al.,
2008). FIH‐1 staining was observed in most cases and was found to be mostly
cytoplasmic, however a small number of tumours showed predominant, strong, nuclear
staining. As above, these data suggest that localisation of FIH‐1 may be a regulated
process in tumours. A separate study examined the expression of FIH‐1 was also
128
assessed in pancreatic endocrine tumours where strong cytoplasmic FIH‐1 levels
correlated with malignant and metastatic pancreatic tumours (Couvelard et al., 2008).
Overall, four studies describing FIH‐1 expression in various cancer types failed to
observe a statistically significant decrease in FIH‐1 levels, or a correlation with reduced
FIH‐1 levels and enhanced HIF target gene transcription. The obvious question of FIH‐1
inactivation in tumour tissues remains.
129
CHAPTER 4
TWO‐DIMENSIONAL
ELECTROPHORESIS
131
4 TWO‐DIMENSIONAL ELECTROPHORESIS
4.1 Introduction
4.1.1 Evidence of cell‐specific differences in FIH‐1 activity
FIH‐1 directly utilises and consumes oxygen for catalysis of hydroxylation and its
catalytic activity is known to be regulated by oxygen concentration. Additionally, there
are several lines of evidence that suggest that FIH‐1 may be regulated by additional or
complementary mechanisms.
Firstly, experiments performed by our lab showed that FIH‐1 activity, but not PHD
activity, varied in a cell type‐specific manner (Bracken et al., 2006). In brief, PC12,
HepG2 and Caco‐2 cells exhibited essentially no HIF‐α‐CAD transcriptional activity until
oxygen concentrations dropped below 1% O2, inferring efficient hydroxylation and
repression by FIH‐1. HIF‐α‐CAD activity was, however, observed in HeLa, 293T and Cos‐1
cells at much more moderate hypoxia (10% O2), and activity increased gradually as
oxygen levels decreased to <1% O2. Variations in FIH‐1 protein levels were not
responsible for these differences in CAD‐driven reporter expression as western blotting
revealed that FIH‐1 levels did not vary significantly between the cell lines. These results
suggest that FIH‐1 exhibits different levels of activity at the same oxygen level
depending on the cellular context, and perhaps suggests additional regulatory
mechanisms at work in certain cells where more strict control over HIF activity is
required.
These cell‐specific differences may represent subtleties in regulation of FIH‐1 and
therefore HIF activity that enable tissue‐specific responses to changes in oxygen levels,
as each tissue, or cell type has a normoxic oxygen value appropriate for its
132
requirements. For example, the brain is an organ which cannot be without oxygen for
more than a few minutes without acquiring irreparable damage, whereas the kidney is
reported to exist in a constant state of low oxygen tension. As such, when animals are
exposed to varying oxygen levels for 1 hour, HIF is induced in the brain at more
moderate levels of hypoxia (18%) whereas it takes more severe oxygen deprivation (6%)
before HIF is stabilised in the kidney (Stroka et al., 2001). Cells may therefore require
additional mechanisms of HIF regulation at both the stability and activity levels to
facilitate body‐wide fine‐tuning of the hypoxic response.
4.1.2 Regulation of FIH‐1 in cancer?
As described above, FIH‐1 has been found to be upregulated in some pancreatic
(Couvelard et al., 2008), renal (Soilleux et al., 2005) and breast (Tan et al., 2007)
cancers. Regulation of FIH‐1 subcellular localisation may occur in cells and affect the co‐
localisation of FIH‐1 with various substrates. FIH‐1 is predominantly a cytoplasmic
protein, though significant levels are present in the nucleus (Stolze et al., 2004; Linke et
al., 2004). There is no change in localisation of FIH‐1 between hypoxia and normoxia;
however research by Tan et al (Tan et al., 2007) has found that subcellular distribution
of FIH‐1 is altered in breast cancer cells. Subcellular localisation of FIH‐1 was
significantly associated with the severity of the tumours, with cytoplasmic and nuclear
FIH‐1 correlating with high and low tumour grades, respectively, and exclusively
cytoplasmic FIH‐1 was found to be a poor prognostic indicator for long term disease
free survival. In terms of FIH‐1 regulation, these data are important as they suggest
active exclusion of FIH‐1 from the nucleus in tumours with a more aggressive
phenotype.
Collectively, FIH‐1 has also been found to be consistently expressed in a range of human
cancers, though its role there in regulating expression of HIF target genes in not clear. It
133
is possible that FIH‐1 activity is reduced in regions of hypoxia within the solid tumour
microenvironment, although studies have shown that FIH‐1 activity persists at low
oxygen levels with HIF‐α substrates in cells (Stolze et al., 2004). Oxygen concentration
within tumours is highly heterogeneous, therefore it is worth investigating the
possibility that FIH‐1 is regulated by an oxygen‐independent mechanism in some
tumour cells and tissues. Furthermore, there is evidence that subcellular localisation of
FIH‐1 may be regulated in breast cancer. Recent reports have described some
alterations in FIH‐1 distribution via interactions with other proteins, most notably the
interactions with MT1‐MMP and Mint3 in macrophages (Sakamoto and Seiki, 2009;
Sakamoto and Seiki, 2010). When experiments for this chapter began, there were no
reports of any mechanism of regulation of sub‐cellular FIH‐1 distribution.
4.1.3 Is FIH‐1 post‐translationally modified?
The capacity of FIH‐1 to inhibit HIF‐mediated transcription under conditions of
decreasing oxygen levels has been found to be cell type specific (Bracken et al., 2006).
Also, FIH‐1 has been observed to retain activity with HIF‐α substrates in cells at severe
hypoxia (Stolze et al., 2004). These two observations suggest that, while oxygen is
required for FIH‐1 catalytic activity, oxygen levels may not be the exclusive means of
FIH‐1 regulation in all cell types. FIH‐1 may be subjected to cell type‐specific regulatory
mechanisms that act independently of oxygen levels.
Cell‐type specific differences in HIF‐α transcriptional activity have also been observed
following treatment of cells with kinase inhibitors or constitutively active kinases,
suggesting that these pathways impart a level of regulation on HIF‐α activity. A role for
direct regulation of HIF‐1α by MAPKs has been described (Mylonis et al., 2006),
however HIF‐2α‐mediated transcription is also affected by this pathway without being a
substrate. Also, a role for the PI3K/Akt pathway in modulating HIF‐α activity in a cell‐
134
specific manner has not been elucidated. As hydroxylation of the HIF‐α‐CAD by FIH‐1
regulates transcriptional activity, phosphorylation of FIH‐1 by these kinase pathways
needs to be investigated.
Together, these lines of evidence suggest that FIH‐1 may be subjected to more complex
regulatory systems than are presently known, and that these may significantly impact
HIF activity both in physiological and pathophysiological situations. The recent
discovery of numerous ARD substrates of FIH‐1 may exert additional regulatory
mechanisms upon FIH‐1 by sequestering the enzyme away from HIF substrates. The fact
that FIH‐1 is able to modify additional substrates in addition to the canonical HIF‐α
subunits may have necessitated the evolution of more stringent control over FIH‐1
activity that operates in addition to regulation of activity by oxygen levels.
Understanding the regulatory mechanisms acting upon FIH‐1 are therefore important to
gain further insight into the role of FIH‐1 in cells and its interactions with substrates, but
also possibly to assist in the treatment of neoplastic disease.
PTM of FIH‐1 may represent a mode of regulation of the enzyme that acts in addition to
regulation of catalytic activity by oxygen levels. PTMs are widely known to regulate
many characteristics of a protein, including activity state, stability, subcellular
localisation, and interactions with other proteins and molecules. Thus, the possibility
that FIH‐1 was subjected to post translational modification was investigated.
Two‐dimensional electrophoresis (2‐DE) was employed to enable investigation of post
translational modifications occurring upon FIH‐1 in cells. The aim of this work was to
separate proteins from lysates prepared from a number of different cell lines by 2‐DE
and then specifically detect FIH‐1 using anti‐FIH‐1 antibodies in order to determine
whether multiple forms of FIH‐1 exist in cells.
135
4.2 Methods Employed for Two‐Dimensional Electrophoresis
4.2.1 Overview
2‐DE enables proteins to be separated according to two parameters; isoelectric point
(pI) and molecular weight (O'Farrell, 1975). During first dimension separation, or IEF, a
protein mixture is placed onto a gel strip containing an immobilised pH gradient and an
electric field is applied, causing each protein to migrate to the pH where their overall
net charge is zero, or their pI. The presence of many different PTMs can be readily
observed by this method as the addition of chemical groups to proteins can alter the pI
and therefore final position of the protein within the pH range. For example, addition of
a PO4‐ group introduces a negative charge and therefore decreases a proteins pI,
resulting in a phosphorylated protein focusing in a more acidic position in a pH gradient
compared with proteins in an unphosphorylated state. 2‐DE, however, is limited to the
detection of modifications that confer a change in charge of the protein. Neutral
modifications, such as hydroxylation, do not affect the isoelectric point or focusing of
proteins and are therefore not able to be inferred by this method. Following the
separation of proteins according to pI in the first dimension, proteins are separated for
a second time according to molecular weight perpendicular to separation in the first
dimension, yielding a 2‐D distribution pattern of the proteins in the sample.
The basic work flow of 2‐D PAGE used in this study is described in Figure 4.1. The main
steps included in this figure and in the following discussion are sample preparation,
protein quantification, IEF, strip equilibration, SDS‐PAGE and finally protein
visualisation. Extensive optimisation of sample preparation and IEF was required to
achieve high quality, reproducible results, with the specific methods used in this work
outlined in Figure 4.1. To follow will firstly be a description of each of these steps and a
Sample
BASIC 2‐DE PROTOCOL
Preparation
Protein Quantification
First Dimension IEF
Strip Equilibration
Second Dimension
SDS‐PAGE
Visualisation
Figure 4.1 Work flow of 2‐DE
Visualisation
Figure 4.1 Work flow of 2 DE
The six major steps of 2‐DE are sample equilibration, SDS‐PAGE and visualisatio
• Method 1: Lysis into SDS‐containing buffer (2‐D Lysis Buffer 1), dilution into CHAPS and Urea buffer (2‐D Lysis Buffer 2)
Specific methods used in this thesis
and Urea buffer (2 D Lysis Buffer 2)
• Method 2: Direct Lysis into IEF buffer• Precipitation and resuspension in IEF buffer
• Bradford Assay
• Bio‐Rad Protean IEF Cell (Bio‐Rad)• Ettan IPGphor II Isoelectric Focusing System (GE Lifesciences).
• Reduction and alkylation
• Immunoblotting using anti‐FIH‐1 antibodies
• 10 or12.5 % Tris/Glycine SDS‐PAGE
antibodies• Sypro Ruby Stain, Coomassie Stain
preparation, protein quantification, IEF, strip on.
139
discussion of the reasoning behind each of the methods used, and secondly the results
generated using each method will be presented.
4.2.2 2‐DE sample preparation
Development of a rigorous sample preparation protocol is critical for successful 2‐DE.
Proteins to be separated by 2‐DE must be fully denatured, soluble, and reduced to
ensure that all intra‐ and intermolecular interactions are disrupted. Complete
denaturation of the proteins in the sample is critical for exposure of all ionisable amino
acid side chains.
Cell lysis and protein denaturation are often achieved using high concentrations of the
chaotropic agent urea. Disruption of cell membranes and protein denaturation can also
be facilitated by using SDS, however the presence of large amounts of SDS in final 2‐D
samples is undesirable due to the effect of the anionic detergent binding to proteins
and interfering with the isoelectric focusing in the first dimension. For this reason, 2‐D
samples containing SDS must be diluted by addition of excess amounts of non‐ionic
detergent or zwitterionic detergent to compete SDS off proteins in the sample (Weiss
and Gorg, 2008). Commonly used detergents that are compatible with IEF are the
zwitterionic detergent CHAPS, and non‐ionic Triton‐X‐100 and NP‐40. In addition to urea
and detergent, standard 2‐D buffers also include a reducing agent, such as DTT, and an
ampholyte mixture consisting of small amphoteric molecules which assist in the
maintenance of protein solubility during IEF in the absence of salt. As 2‐DE here was
concerned with detection of PTMs, possible phosphorylation events were preserved by
the addition of 20 mM sodium fluoride, an inhibitor of phosphoseryl and
phosphothreonyl phosphatases, and 2 mM sodium orthovanadate, an inhibitor of
phosphotyrosyl phosphatases, to the sample buffer immediately before use.
140
4.2.3 Methods of sample preparation
Two different methods for cell lysis and protein solubilisation were employed during the
course of this work, as detailed below.
4.2.3.1 Sample preparation method 1
The method for 2‐D sample preparation initially employed was based on a method
routinely used by colleagues and involved detergent lysis of cells in a buffer containing
SDS. In application of this method, cells were lysed in buffer containing 1% SDS (2‐D
Lysis Buffer 1), following dilution of the lysate into a buffer containing 13% CHAPS (2‐D
Lysis Buffer 2) and addition of 8 M urea. Though successful 2D‐PAGE was achieved using
this method, instances of high sample conductivity and poor IEF often resulted. The
current was limited to 50 mA in these experiments, and presence of ionic compounds
prevented high voltages from being reached, leading to extensive focusing times.
4.2.3.2 Sample preparation method 2
An alternative sample preparation method was researched and devised. Subsequent
experiments were performed by lysing cells directly into a buffer containing 8 M urea,
2% CHAPS and 40 mM DTT and supplemented with fresh phosphatase inhibitors (IEF
buffer) This was done to avoid the use of SDS in the cell lysis procedure as this may have
been a potential source of ionic strength in the samples, despite the subsequent
dilution of samples in CHAPS buffer. Protein precipitation was also used to reduce any
other remaining contaminating substances, and precipitated proteins were
resuspended in IEF Buffer, quantified and diluted to produce samples of 100 μg of
protein in a total volume of 125 μl.
141
4.2.4 Protein quantification
Samples were quantified using a Bradford Protein Assay, which is based on a change in
absorbance from 465 to 595 nm of the dye Coomassie Brilliant Blue G‐250 when bound
to proteins (Bradford, 1976). As some detergents have been found to interfere with
accuracy of the Bradford assay (Bradford, 1976; Compton and Jones, 1985), and as the
effect of other buffer components urea and DTT upon the assay were unknown, the
propriety of using this method for quantifying protein concentrations in 2‐D samples
was investigated. Briefly, BSA was diluted to 1‐5 mg/ml in either water or IEF buffer, and
5 μl of the standard was combined with 195 μl of the Bio‐Rad Protein Assay Reagent
and the absorbance at 595 nm measured and plotted to a standard curve for each
condition (Figure 4.2). The results showed that there was little difference in measured
absorbance using either water or buffer up to BSA concentrations of 3 mg/ml. At
concentrations beyond this however, there was a difference in the measured
absorbance and the assay did not appear to be linear at higher protein concentrations
when proteins were dissolved in IEF buffer. However, since 2‐D samples quantified
were often between 0.8‐2 μg/μl, samples within this range were predicted to be
accurately quantified by Bradford Assay. In using the assay for quantification of 2‐D
samples, the BSA standards were made up in IEF buffer to reduce any variations
between the samples and standards due to interference by buffer components.
The conclusion that the Bradford assay was a suitable method for quantifying 2‐D
samples was consistent with published data comparing standard curves of BSA in water,
8 M urea, CHAPS, DTT or IPG buffer alone, or combinations of each of these
components in Bradford, Lowry and BCA assays (Kao et al., 2008). These data also
supported the use of Bradford assays over Lowry and BCA assays for protein
quantification in 2‐DE sample buffers, concluding that the Bradford Assay was the most
1.4
Standard Curve of
0.8
1
1.2
95 nm
0
0.2
0.4
0.6A 5
0 1 2
BSA Concen
Figure 4.2 Standard Curve of BSA in eitStandard amounts of BSA were dilutedcurves using Bradford Assay Reagent.
f BSA in water or IEF buffer
water
IEF buffer
3 4 5
ntration (mg/ml)
ther water or IEF bufferin either water or IEF buffer to make two standard
145
accurate of the three colorimetric assays. Thus, the Bradford assay was employed
throughout this work with confidence for quantification of 2‐D samples.
4.2.5 Isoelectric focusing
The strips used in these experiments were 7 cm strips with an immobilised pH gradient
(IPG) of either 3‐11 or 4‐7, as indicated, with recommended sample loads of 100 μg of
protein. Prior to application to the strips, either Coomassie Blue or Bromophenol Blue
was added to the samples as a tracking dye. Though Coomassie is a charged molecule,
its use as a dye in IEF has been reported not to interfere with IEF, and in fact to result in
better resolution, particularly for proteins with basic pIs, than when using Bromophenol
Blue (Vilain et al., 2001). Coomassie Blue and Bromophenol Blue dyes were both used
interchangeably throughout this work with no discernable difference in quality of
focusing or spot resolution.
Samples were applied to IPG strips by active rehydration at 50 V for 12 hours. The
voltage was then increased to 250 V for an hour to allow for the migration of small
charged contaminants to the electrodes prior to focusing. The voltage was then
increased to 1000 V for an hour and then set at a maximum of 8000 V for a total
focusing period of 20 000 VHrs. The current was limited at 50 mA, and as such the
maximum set voltage was not usually reached and focusing was achieved over a longer
period of time at lower voltages. IEF was performed using first the Bio‐Rad Protean IEF
Cell (Bio‐Rad) and later the Ettan IPGphor II Isoelectric Focusing System (GE
Lifesciences).
146
4.2.6 Equilibration
Following IEF, it is necessary to prepare the proteins for SDS‐PAGE, which is achieved by
incubating the strips in two sequential buffers. Equilibration is required before
separation in the second dimension to allow proteins to bind to SDS for separation by
SDS‐PAGE, and also to ensure that proteins are reduced and alkylated to prevent
reoxidation during the run. Equilibration was achieved through a 20 minute incubation
in sequential equilibration buffers containing 6 M urea, 2% SDS, 0.375 M Tris and 20%
glycerol with the first containing 10 mM DTT to ensure proteins remain soluble and do
not aggregate when leaving the strip, and the second containing 2.5% iodoacetamide to
alkylate cysteine thiol groups of proteins. Iodoacetamide also alkylates any remaining
DTT, as the presence of excess DTT in the second dimension is known to cause
horizontal streaking in 2‐D gels. This step was performed without modification
throughout all experiments.
4.2.7 Second dimension separation
Following equilibration, strips were loaded onto vertical 10 or 12.5% Tris/Glycine gels
and overlaid with a 1% molten agarose solution. A lane allowing for molecular weight
markers was included and gels were run at 90 V until all proteins were seen to enter the
gel, at which point the voltage was increased to 120 V and proteins separated according
to molecular weight. This step also remained consistent throughout the course of the
experiments.
4.2.8 Visualisation
As these experiments aimed to investigate the modification of only one, lowly abundant
protein, visualisation was achieved through western blotting using three different FIH‐1‐
specific polyclonal antibodies. Sypro Ruby staining and Coomassie Blue staining were
147
used for visualisation of total protein in a limited number of 2‐D gels to ensure good
resolution across the entire gel area.
4.3 Optimisation of Two‐Dimensional Electrophoresis
2‐DE was undertaken to investigate the possibility of FIH‐1 modification, and determine
whether any such modification may contribute to regulation of FIH‐1 in a cell‐specific
manner. The data leading to this notion was generated from experiments performed in
HeLa, 293T, Cos‐1, Caco‐2, HepG2 and PC12 cells, and while it was of interest to do
experiments using lysates from multiple cell lines, initial experiments were aimed at
simply generating a successful protocol that gave reproducible results before more
comprehensive analysis would be undertaken. Thus, all early 2‐D gels utilised lysates
from HeLa cells only, as this cell line was convenient to grow, had been shown to have
similar levels of FIH‐1 to other cell lines, and sufficient quantities of lysate could be
generated without difficulty.
4.3.1 Sample preparation method 1
Concentrated lysate was prepared from HeLa cells grown under standard conditions
and harvested at approximately 90% confluency. Cells were lysed in 2‐D Lysis Buffer 1 as
described above prior to dilution in 2‐D Lysis Buffer 2 containing 13% CHAPS (2‐D
sample preparation method 1, see section 2.16.1 of Materials and Methods Chapter),
and proteins separated by isoelectric focusing across a pH 4‐7 pH gradient using the and
SDS‐PAGE.
Proteins were visualised by blotting with the anti‐FIH‐1 No.8 antibody. Blots showed
that the proteins were effectively separated and resolved, with four spots observed to
148
migrate to the expected molecular weight for FIH‐1 (Figure 4.3). The detection of
several spots of different pI and molecular weight using the anti‐FIH‐1 antibody
supported the idea that modifications of FIH‐1 occurred. FIH‐1 has a predicted pI of 5.57
and an estimated molecular weight of 40.38 kDa (calculated by the Scripps online
protein calculator at www.scripps.edu/~cdputnam/protcalc.html). These spots,
however, did appear to be in a more acidic position than expected. The strip contains an
immobilised, linear pH range of 4‐7, thus a protein with a pI of 5.57 would be expected
to focus in the more basic side of the gel, just to the right of the middle of the blot.
Detecting FIH‐1 with this antibody however did not show any proteins in that position,
rather the spots detected were focused at a position corresponding approximately to
5.1 along the linear pH range. Resolution in a more acidic position can be indicative of
modification, such as phosphorylation or deamidation, both modifications which confer
a negative charge upon a protein and thus decrease the isoelectric point.
Despite some successful 2‐DE using this method, where clear resolution of individual
spots was achieved, IEF of samples prepared using method 1 was often problematic and
gave inconsistent results (data not shown). Difficulties during IEF, such as low maximum
voltages, long times taken to reach the desired total volt hours, and generation of heat
along the strip during the IEF run, suggested the presence of ionic contaminants. An
obvious source of ionic strength in the sample was the SDS used during cell lysis. In an
attempt to improve the resolution and consistency of the method, proteins were
precipitated to remove the SDS and any other contaminating substances that may have
been present in the samples. Precipitation of samples prepared by method 1 did not
seem to improve results (data not shown.) Thus, an alternative sample preparation
method was devised.
75
4
kDa
50
37
Figure 4.3 2‐DE of HeLa cell lysates usi
HeLa lysate was made by sample prepaon pH 4 7 immobiline strips (Bio Rad)on pH 4‐7 immobiline strips (Bio‐Rad).Cell, followed by 12.5% Tris/Glycine SDantibody was used for detection of FIHFIH‐1. Representative of > 3 independe
pH 7
FIH‐1
WB: No.8
ing Method 1
aration Method 1. 100 μg of protein was separatedIEF was performed using the Bio Rad Protean IEFIEF was performed using the Bio‐Rad Protean IEF
DS PAGE. Immunoblotting with the No.8 anti‐FIH‐1H‐1. Arrow indicates predicted molecular weight ofnt experiments.
151
4.3.2 Sample preparation method 2 and precipitation
The sample preparation protocol was altered to involve direct lysis of HeLa cells into a
buffer containing 8 M urea, 2% CHAPS buffer and 40 mM DTT (2‐D sample preparation
method 2, see section 2.16.2 of Materials and Methods Chapter). To improve the purity
of the samples, an additional step of precipitation using a 2‐D Clean Up Kit was
included.
2‐DE samples were prepared using method 2, and proteins were precipitated and
dissolved in IEF buffer by sonication. This method consistently yielded well resolved
spots on small format gels, as visualised by immunoblotting with the No.8 antibody
(Figure 4.4). The results showed a similar spot profile to that observed with the previous
method, again indicating that FIH‐1 may be modified in HeLa cells. Method 2 was both
simpler and also reduced the problems encountered during previous IEF runs, such as
low voltage and extensive times required to complete IEF. Sample preparation method
2 was thus exclusively employed in addition to precipitation for preparation of all
subsequent 2‐D samples.
It is important here to note that the manufacturer’s protocol for use of the kit had to be
altered. Recommended centrifugation speeds were 12 000 x g, however use of these
speeds was found to cause precipitated protein to form pellets so compact that proper
washing and resuspension of the pellet in IEF buffer could not be achieved, even with
extensive sonication of samples in a sonicating bath. Due to the undesirable effects
associated with heating samples in urea, care was care taken to keep samples cool
during sonication. Samples were placed on ice between short sonication steps, and ice
was also added to the sonication bath. Ultimately, to circumvent the difficulties
encountered with resuspending and dissolving precipitated proteins, subsequent use of
4 kDa
50
37
75
kDa
37
WB: N
Figure 4.4 2‐DE of HeLa lysates using M
HeLa lysate was prepared using MethodUp Kit followed by resuspension in IEF b7 immobiline strips (Bio‐Rad) and IEFfollowed by 12.5% Tris/Glycine SDS PA
l lArrow indicates predicted molecular windependent experiments.
pH 7
FIH‐1
No.8 anti‐FIH‐1 polyclonal Ab
Method 2
d 2. Samples were precipitated using the 2‐D Cleanbuffer. 100 μg of protein was separated by on pH 4‐F performed using the Bio‐Rad Protean IEF cell,AGE and immunoblotting with the No.9 antibody.
h f l fweight of FIH‐1. Results are representative of > 3
155
the 2‐D Clean Up Kit involved centrifugation steps of approximately half the
recommended speed. This was found to pellet precipitated protein effectively and also
enabled easy resuspension and dissolution of pellets without the need for sonication.
4.3.3 Sample preparation using method 2 and optimised precipitation protocol
At this time, a different IEF apparatus became available. The Ettan IPGphor II Isoelectric
Focusing System (GE Lifesciences) offered the advantages of monitoring of voltage and
current over the entire focusing protocol and also provided a graphical readout of each
of these parameters. All subsequent 2‐D gels described in this thesis were generated
using the Ettan IPGphor II Isoelectric Focusing System for IEF.
To assess protein resolution across the entire gel area using the optimised sample
preparation protocol and the new IEF system, small‐format 2‐D gels of HeLa lysate were
generated and stained with Coomassie or Sypro Ruby. A representative 2‐D gel is shown
in Figure 4.5, where 100 μg of HeLa lysate was separated using a 7 cm IPG strip of pH 3‐
11, followed by 12.5% Tris/Glycine SDS‐PAGE and staining with Sypro Ruby. This
confirmed that good resolution was achieved across the middle of the pH range, with
some loss at the extreme acidic end which was to be expected, and which should not
interfere with FIH‐1 resolution. This gel was not run far in the second dimension and
reveals excellent spot resolution between 20 kDa to approximately 90 kDa, easily
encompassing FIH‐1.
Samples prepared by method 2 and precipitated were subjected to IEF using the Ettan
IPGphor II IEF System across 7 cm pH 4‐7 IPG strips and the second dimension gels were
run further to enable better separation of the spots of interest. Blotting with the No.9
anti‐FIH‐1 antibody however revealed a distinct spot profile (Figure 4.6) to that
observed previously (Figures 4.3 and 4.4). Firstly, the spots detected on the HeLa blot
100
200
250
3 kDa
75
50
100
37
25
20
Figure 4.5 Sypro Ruby stain of HeLa lys
HeLa lysate was made by Sample Prepaof protein was separated by on pH 3‐using the Ettan IPGphor II Isoelectricusing the Ettan IPGphor II Isoelectric12.5% Tris/Glycine SDS PAGE. TotalPredicted weight of FIH‐1 is indicated.
pH 11
(FIH‐1)
Sypro Ruby Stain
sate 2‐D gel
aration method 2 followed by precipitation. 100 μg‐11 IPG strip (GE Lifesciences) and IEF performedc Focusing System (GE Lifesciences) followed byc Focusing System (GE Lifesciences), followed byprotein was detected using Sypro Ruby stain.Representative of two independent experiments.
75
50
4
kDa
37
Figure 4.6 2‐DE of HeLa cell lysate
Cell lysates were prepared using Methwere precipitated and resuspended inpH 4‐7 IPG strips and IEF performedSystem, followed by 10% Tris/Glycimmunoblotting with the No.9 antibodFIH‐1. Representative of > 3 independe
pH 7
WB: No.9 anti‐FIH‐1
FIH‐1
od 2, lysing cells directly into IEF buffer. SamplesIEF buffer. 100 μg of protein was separated by onusing the Ettan IPGphor II Isoelectric Focusing
cine SDS PAGE. Results were visualised bydy. Arrow indicates predicted molecular weight ofent experiments.
161
with the No.9 antibody are clearly positioned in a more basic part of the blot than those
observed previously. Here, a string of four spots of different pIs, and possibly two
different molecular weights, were detected in a more basic region of the 2‐D blot. What
is also evident in this blot is that the four spots previously observed in the more acidic
region of the blot using the No.8 antibody are not seen here. The position of the spots
on the more basic half (extending between approximately pH 5.3‐5.6) of the blots
however is quite different to the position of spots observed in earlier experiments,
which focused in the more acidic part of the blot, around pH 5.0. The isoelectric point of
FIH‐1 as observed in the later experiments using an optimised protocol is closer to the
predicted isoelectric point of 5.57. This more basic spot profile was repeatedly observed
with both No.8 and No.9 antibodies in all subsequent 2‐D blots.
4.3.4 Different spot profiles an artefact of sample preparation
The discrepancy between the two different pIs observed could be an artefact of sample
preparation. The early experiments were confounded by ionic contamination leading to
extended focusing times and elevated temperatures during IEF. Removal of
contaminating substances by precipitation also required optimisation, and initial
experiments including a precipitation step involved sonication of samples to facilitate
dissolution in IEF buffer. This increased both handling times and also an elevated
temperature of the sample in the presence of urea before the protocol was altered to
avoid the sonication step.
Thus one possible reason for the increased acidity of the pI of FIH‐1 in early 2‐D gels
could be carbamylation of the protein. Urea decomposes in solution to ammonium
cyanate, and cyanate is able to react with available amino groups on lysine, arginine and
cysteine residues, or the N‐terminus, of a protein. This reaction involves the removal of
the positively charge amino group and attachment of a carbamoyl group, conferring a
162
more acidic isoelectric point on the carbamylated protein. The decomposition of urea
into ammonium cyanate is known to be enhanced with increasing temperatures and
over time. Implementation of each of the early protocols caused elevated temperatures
and increased handling times when proteins were in the presence of 8 M urea (either
during the IEF run in method 1, or during sonication of samples to facilitate
resuspension following precipitation in both method 1 and method 2), and would likely
also have led to increased generation of ammonium cyanate and thus carbamylation of
the proteins. Indeed, carbamylation has been found to induce dramatic shifts of protein
spots in 2D maps (McCarthy et al., 2003). Later experiments limited sample
temperatures and handling times by avoiding long IEF times and sonication steps, thus
minimising the likelihood of protein modification by urea decomposition. Carbamylation
is therefore a possible contributor to the observed difference in spot position between
early 2‐D gels and later 2‐D gels generated using the optimised protocol.
4.3.4 Summary of preliminary HeLa experiments
Sample preparation protocols and IEF parameters required extensive optimisation in
order to achieve reproducible results. Two sample preparation methods were
employed. Though a spot profile for FIH‐1 was observed in 2‐D gels generated using
method 1, this method was associated with extensive times taken to reach the
prescribed number of VHrs. Method 2 was used in conjunction with precipitation and
overall avoided the problems associated with method 1, enabling successful IEF.
However, using the 2D Clean Up Kit according to the recommended protocol resulted in
difficulties dissolving precipitated protein pellets in IEF buffer, and required additional
steps of sonication to achieve solubility. The final optimised protocol for 2‐D sample
preparation devised for lysis and solubilisation of HeLa cells involved direct cell lysis into
IEF buffer with added phosphatase inhibitors, precipitation of samples and slow
centrifugation steps to facilitate easy resuspension into IEF buffer.
163
The results obtained during optimisation of a robust 2‐DE protocol are interesting, not
only in the indication that FIH‐1 is indeed present in various forms in HeLa cells, but also
as a comparison of the effects of sample preparation on resolution of proteins in 2‐DE.
The clearly different spot profiles obtained when using different sample preparation
methods highlighted the importance of a rigorous and carefully performed protocol in
achieving reliable results and reducing the likelihood of sample preparation artefacts.
Overall, these preliminary results indicate that there are several different forms of FIH‐1
present in HeLa cells. 2‐D immunoblots generated using the optimised sample
preparation protocol consistently showed up to four spots with different pIs and two
molecular weights, supportive of the PTM of FIH‐1 in HeLa cells. The work performed so
far had enabled optimisation of a successful, reproducible 2‐DE protocol, which could
now be applied to investigate the spot profile of FIH‐1 in other cell lines.
4.4 Two‐Dimensional Electrophoresis Results
Once a reliable 2‐DE protocol had been devised, it was important to next validate the
spot profile observed in HeLa cells and investigate the modification of FIH‐1 in other cell
lines, with particular interest in discerning any differences in spot profile between cell
lines that may relate to differences in FIH‐1 regulation.
Subsequent 2‐DE experiments performed using HeLa lysates according to the optimised
protocol involving sample preparation method 2 and precipitation with reduced
centrifugation speeds corroborated the results seen in Figure 4.6. Here, NB 100‐428 was
164
used to detect FIH‐1 in blots of HeLa lysate separated by 2‐DE (Figure 4.7). This antibody
also detected spots in the basic side of the blot, similar to the spots detected in 293T
and Cos‐1 lysates using both No.8 and No.9 antibodies. FIH‐1 was observed to resolve
into at least three spots of different pI, and two different molecular weights. Again,
these results are consistent with the modification of FIH‐1 in HeLa cells.
The other cell lines to be investigated included Cos‐1, 293T, HepG2 and Caco‐2 cell lines,
due to the previously discussed data suggesting differences in FIH‐1 activity (see Section
1.2.4). Additionally, it was of interest to investigate the spot profile of FIH‐1 in the wt
and FIH‐1‐/‐ MEFs, as comparison of the spot profiles of these two samples would more
definitively demonstrate which spots were likely to be FIH‐1. Also, while the other cell
lines are derived from various cancer tissues and thus can be presumed to contain
genetic alterations which may or may not affect FIH‐1 modification, the MEFs are a
primary cell line as such may provide a more accurate representation of “normal” FIH‐1
modification.
4.4.1 2‐DE of MEF lysates
In order to definitively show that the more basic spot pattern observed in later
experiments was indeed FIH‐1, 2‐D separation of wt and FIH‐1‐/‐ MEF cell lysates was
performed and FIH‐1 detected using anti‐FIH‐1 antibody NB 100‐428 (Figure 4.8). The
wt MEF 2‐D blot shows a trail of spots that appear above the 37 kDa marker and at a pH
of approximately 5.3‐5.6 (Figure 4.8 a). These spots are absent in the FIH‐1‐/‐ MEF blot
(Figure 4.8 b), suggesting that these spots represent FIH‐1. This position is consistent
with the predicted molecular weight and pI of FIH‐1, and is consistent with the spot
profile observed for FIH‐1 on 2‐DE blots of HeLa cell lysates generated using the
4 kDa
37
50
75100
kDa
Figure 4.7 2‐DE of HeLa lysateFigure 4.7 2 DE of HeLa lysate
HeLa cell lysates were prepared usingprotein as separated by on pH 4‐7 IPGIsoelectric Focusing System, followedvisualised by immunoblotting with theof FIH‐1 is indicated. Representative of
pH 7
FIH‐1
WB: anti‐FIH NB 100‐428
g Method 2 followed by precipitation. 100 μg ofstrips and IEF performed using the Ettan IPGphor IId by 10% Tris/Glycine SDS PAGE. Results wereNB 100‐428 antibody. Predicted molecular weightf >3 independent experiments.
kDaa.
4
75
100200
50
37
4
b.
100200
kDa
75
50
37
Figure 4.8 2‐DE of wt and FIH‐/‐MEF lysa
Wt (a) and FIH‐1‐/‐ (b) MEF lysates wereby precipitation. 125 μg of each lysaperformed using the Ettan IPGphor IITris/Glycine SDS PAGE. FIH‐1 was detectof 2 independent experiments.
pH 7
wt MEFs
FIH‐1
WB: anti‐FIH NB 100‐428
pH 7
FIH‐1 ‐/‐ MEFs
WB: anti‐FIH NB 100‐428
ates
made using Method 2 and contaminants removedte was loaded onto pH 4‐7 IPG strips and IEFIsoelectric Focusing System, followed by 12.5%
ted using the NB 100‐428 antibody. Representative
169
optimised 2‐DE sample preparation protocol (Figures 4.6 and 4.7). Significant
background staining was observed in each blot, and the background spot patterns were
consistent between both wt and ko MEF blots.
This experiment definitively demonstrated that the more basic spot train observed in
multiple experiments is FIH‐1, validating the results achieved using the optimised
sample preparation method. Importantly, it demonstrated the presence of multiple
forms of FIH‐1 in a primary cell line, supporting the existence of multiple modifications
of FIH‐1.
4.4.2 2‐DE of Cos‐1, 293T and HeLa cell lysates
Lysates from 293T and Cos‐1 cells were separated by 2‐D PAGE using the optimised
sample preparation protocol (method 2 and precipitation) and IEF protocol described
above. Immunoblotting was performed using the No.8 polyclonal antibody which
detected background protein spots in 2‐D blots and also appeared to bind non‐
specifically to membranes, confounding visualisation of FIH‐1. The best example of a
blot achieved using the No.8 antibody is a 2‐DE blot of Cos‐1 lysate shown in Figure 4.9.
Here, a string of possibly four spots was detected at the predicted size of FIH‐1 and
consistent with the pI observed for FIH‐1 from HeLa cell lysates (Figures 4.6 and 4.7) and
MEF lysates (Figure 4.8). These results suggested that FIH‐1 exists in multiple states in
Cos‐1 cells, consistent with multiple PTMs.
2‐DE blots of 293T lysates were visualised by immunoblotting with the No.9 anti‐FIH‐1
antibody (Figure 4.10). A spot train consisting of three or possibly four spots of different
pI, and also a series of spots of two different molecular weights, was detected,
supportive of several modified forms of FIH‐1 from 293T cell lysates.
75
100
4 kDa
37
50
Figure 4.9 2‐DE of Cos‐1 cell lysate
Cos‐1 cell lysates were prepared usinfollowed by precipitation. 100 μg of prperformed using the Ettan IPGphor ITris/Glycine SDS PAGE. Results were visArrow indicates predicted molecular wexperiments.
pH 7
WB: No 8 anti‐FIH‐1
FIH‐1
WB: No.8 anti FIH 1
ng Method 2, lysing cells directly into IEF buffer,rotein was separated on pH 4‐7 IPG strips and IEFI Isoelectric Focusing System, followed by 10 %ualised by immunoblotting with the No.8 antibody.weight of FIH‐1. Representative of 3 independent
100
4 kDa
50
75
37
Figure 4.10 2‐DE of 293T cell lysate
Cell lysates were prepared by lysingprecipitation. 100 μg of protein was susing the Ettan IPGphor II Isoelectric FPAGE. Results were visualised by immArrow indicates predicted molecular weArrow indicates predicted molecular weexperiments .
pH 7
FIH‐1
WB: No.9 anti‐FIH‐1
g cells directly into IEF buffer (Method 2) andeparated on pH 4‐7 IPG strips and IEF performedocusing System, followed by 10% Tris/Glycine SDSmunoblotting with the No.9 anti‐FIH‐1 antibody.eight of FIH 1 Representative of > 3 independenteight of FIH‐1. Representative of > 3 independent
175
Due to the reported differences in subcellular localisation of FIH‐1 reported in breast
cancer, it was of interest to investigate any potential differences in FIH‐1 spot profile
between cytoplasmic and nuclear fractions. HeLa lysates were subjected to
subfractionation into cytosolic and nuclear extracts. Each extract was precipitated and
resuspended in IEF buffer for IEF prior to SDS‐PAGE, and FIH‐1 detected using the No.9
antibody. Specific nuclear and cytosolic markers were used to ascertain purity of the
fractions (Figure 4.11 a). The results showed that there was no difference in spot
pattern between cytoplasmic or nuclear FIH‐1 from HeLa cells (Figure 4.11 b, c).
In summary, the results achieved in implementation of the optimised sample
preparation protocol Cos‐1, 293T and HeLa cell lysates all exhibited a similar spot profile
for FIH‐1 as detected by three different anti‐FIH‐1 antibodies. The pattern of spots
observed indicates that there are up to four forms of FIH‐1 with different pI, and two
forms of FIH‐1 with different molecular weights. These results suggest that FIH‐1
protein is modified in these cell lines.
4.4.3 2‐DE of 293T, Cos‐1, Caco‐2 and HepG2 cell lysates
The other cell lines to be investigated by 2‐DE were those that were found by reporter
assay to exhibit varying levels of oxygen‐dependent regulation of FIH‐1. As HeLa, Cos‐1
and 293T cell lines belonged to the group that demonstrated HIF‐α‐CAD activity at more
moderate hypoxia (10% and 5% oxygen), it was of interest to compare the spot profile
of FIH‐1 in these cells to FIH‐1 from Caco‐2 and HepG2 cells where HIF‐α‐CAD was only
active at <1% oxygen. To this end, 293T, Cos‐1, Caco‐2 and HepG2 cells were all grown
at normal oxygen levels and lysates prepared for 2‐DE by addition of IEF buffer to cells,
followed by precipitation and resuspension in IEF buffer. Following 2‐DE, FIH‐1 was
detected using the NB 100‐428 anti‐FIH‐1 antibody, and results of 293T (a), Cos‐1 (b),
Subcellular Fraction
a.
4
b.
Anti‐paxillin
C N 75
75 50
75
100
kDa
Anti‐nucleoporin37
c.
50
75
100
4
kDa
37
50
Figure 4.11 Subcellular fractionation of
Cytoplasmic and nuclear HeLa cell lysafollowed by nuclear lysis in nuclear e(anti‐paxillin) and nuclear‐ (anti‐nucleoto proteins in each fraction being precprotein was separated by on pH 4‐7 IPGII Isoelectric Focusing System, followcytoplasmic (a) and nuclear (b) proteinsantibody. Predicted molecular weighindependent experimentsindependent experiments.
4 pH 7
FIH‐1
4 pH 7
WB: No.9 anti‐FIH‐1
FIH‐1
f HeLa cell lysates and 2‐DE
tes were prepared by cell lysis in hypotonic bufferextract buffer. Immunoblotting using cytoplasmic‐oporin) specific antibodies was performed (a) priorcipitated and resuspended in IEF buffer. 100 μg ofG strips and IEF performed using the Ettan IPGphorwed by 10% Tris/Glycine SDS PAGE. Results fors were visualised by immunoblotting with the No.9ht of FIH‐1 is indicated. Representative of 2
179
Caco‐2 (c) and HepG2 (d) cell lysates are shown in Figure 4.12. All four blots show the
detection of a series of spots with a molecular weight of approximately 40 kDa in the
basic half of the blot corresponding to a pI of between approximately 5.5 and 5.9.
These results are in agreement with the predicted size and pI of FIH‐1 and also
consistent with the results achieved previously using this optimised protocol. Three to
four forms of FIH‐1 with different pI are evident in 293T, Cos‐1 and HepG2 blots, with
five different pI forms of FIH‐1 seen in Caco‐2 blots. Additionally, as a consequence of a
clean antibody signal or superior spot resolution, two different molecular weight spots
can clearly be seen in Cos‐1, HepG2 and Caco‐2 cells. These results indicate that FIH‐1 is
subjected to modifications affecting pI in these cell lines.
The excellent resolution achieved enabled a comparison of FIH‐1 spot patterns between
the cell lines. Spot intensity in 2‐DE can roughly be equated to the relative abundance of
protein amount and therefore provide an indication of the relative amounts of each of
the modified forms of a protein. Regarding the different activity levels of the CAD, and
by inference, FIH‐1, observed by reporter assays, 293T and Cos‐1 cells belonged to the
group where more moderate hypoxia (10% ‐ 5% oxygen) was sufficient to see a decline
in FIH‐1 activity, and Caco‐2 and HepG2 cell lines belonged to the group that only
displayed CAD activity at <1% oxygen. It is therefore interesting to note the similarities
between FIH‐1 spot patterns in 293T and Cos‐1 cells, and Caco‐2 and HepG2 cells, and
the marked differences between each pair.
Strikingly, the pattern of spot intensity is altered between each pair. 2‐D blots of 293T
and Cos‐1 cell lysates each show three or four spots of different pI, where the most
basic spot is the most intense, and trend of decreasing spot intensity is observed along
the series as the spots as the pI becomes more acidic. In Caco‐2 and HepG2 lysates, four
or more different pI forms of FIH‐1 are seen, and while in Caco‐2 and HepG2 cells the
most intense spots are represented by the most basic spots, the consistent decline in
a. 293T cells
100
kDa
4 p
75
4
37
50
b. Cos‐1 cells
75
100
kDa 4
37
50
H 7
pH 7
FIH‐1
WB: anti‐FIH NB 100‐428
pH 7
FIH‐1
WB: anti‐FIH NB 100‐428
c. Caco‐2 cells
4
75
100
4 kDa
37
50
d. HepG2 cells
75
100
kDa 4
37
50
Figure 4.12 2‐DE of 293T, Cos‐1, Caco‐2Figure 4.12 2 DE of 293T, Cos 1, Caco 2
293T (a), Cos‐1 (b), Caco2 (c) and Hepinto IEF buffer. The samples were preciμg of each lysate was loaded onto pHIPGphor II Isoelectric Focusing System,detected using the commercial NB 100‐is indicated. Representative of >3 indep
pH 7 pH 7
FIH‐1
WB: anti‐FIH NB 100‐428
pH 7
FIH‐1
2 and HepG2 cell lysates
WB: anti‐FIH NB 100‐428
2 and HepG2 cell lysates
pG2 (d) cell lysates were made direct lysis of cellsipitated and pellets resuspended in IEF buffer. 1004‐7 IPG strips and IEF performed using the Ettanfollowed by 10% Tris/Glycine SDS PAGE. FIH‐1 was‐428 antibody. Predicted molecular weight of FIH‐1endent experiments.
185
intensity of increasingly acidic spots observed on 293T and Cos‐1 cells is not seen here.
Rather, in Caco‐2 2‐D blots, the spot directly left of the most basic and most intense
spot is very faint, with the second most intense spot occupying a more acidic position.
Similarly, the intensity of the spots observed on HepG2 2‐D blots occurs “out‐of‐
sequence”, where the most basic spot is as intense as the second most acidic spot, and
the intermittent spots are more faint. These differences imply that in 293T and Cos‐1
cells, the form of FIH‐1 represented by the most basic spot in the series is the most
abundant form in these cells, and that the more acidic forms of FIH‐1 occur with less
frequency. In Caco‐2 and HepG2 cells however, the most basic form of FIH‐1 is also the
most highly abundant, however, more acidic forms of FIH‐1 are present in these cells
with greater abundance than in 293T or Cos‐1 cells. What modifications are represented
by these differentially charged forms of FIH‐1 and what consequences these
modifications may have for the regulation of both FIH‐1 and HIF‐α activity in specific cell
types are of considerable interest and worthy of further investigation.
4.5 Summary and Discussion
In summary, 2‐D experiments investigating the modification of FIH‐1 in different cell
lines are consistent with the existence of different modified forms of FIH‐1. FIH‐1 was
seen to resolve as up to five distinct forms with different pIs and two different
molecular weights. Though there was some ambiguity regarding the position of FIH‐1
using different sample preparation protocols during early experiments where the
protocol was being optimised, the comparison of wt and FIH‐1‐/‐ MEF lysates gave a
clear indication of where on a 2‐D gel FIH‐1 is expected to migrate and validated the
results seen in all five of the other cell lines examined. Furthermore, this position is in
agreement with the predicted size and pI for FIH‐1. Overall, these results showed that in
186
six different cell lines and using three different antibodies, multiple forms of FIH‐1 with
different pI and molecular weight were detected in 2‐D blots.
Significantly, comparison of the spot patterns between cells seems to suggest that these
modifications may occur differentially in a cell‐type specific manner. While this is
somewhat speculative, it is interesting that the cell lines that exhibited similarities in
FIH‐1 activity also exhibit similarities in spot patterns. Returning to the reporter assay
data from Bracken et al (2006), if each individual spot detected does indeed represent
unique modification of FIH‐1, these results suggest that the dominant form of FIH‐1 in
Cos‐1 and 293T cells is the most basic form and that this corresponds to a lower level of
hydroxylation of HIF‐α in these cells. In Caco‐2 and HepG2 cells, other modifications are
aloes abundant, evidenced by high intensity of some of the more acidic spots, and this
corresponds to a greater level of hydroxylation and repression of HIF‐α transcriptional
activity by FIH‐1.
There are many PTMs known to alter a proteins’ charge. A decrease in pI, or the
presence of an acidic spot train, can either be indicative of a PTM that increases the
negative charge, or conversely decreases the positive charge, of a protein. The most
well known example is phosphorylation, which involves addition of a negatively charged
phosphate (PO4‐) group. Additionally, deamidation, where a positively charged amide
group is removed from either asparagine or glutamine residues to give aspartic acid and
glutamic acid, respectively, also causes a shift of proteins in the acidic direction through
loss of the positive charge. Acetylation of lysine residues also removes a positive charge
by removing the terminal amide group of the lysine side chain for attachment of the
acetyl group. Lesser known modifications that confer a negative charge to a protein
include the covalent attachment of negatively charged ADP‐ribose to a protein, a
process termed PARsylation, or poly ADP‐ribosylation. Conversely, other PTMs do not
alter the overall charge of a protein. While the existence of neutral PTMs is not
187
precluded by these data, such modifications would not contribute to pI changes or
distinct spots on a 2‐D gel. Though nitrosylation was mentioned in the Introduction
Chapter (See Section 1.5.3), this modification does not alter pI at neutral pH and was
therefore not considered to be a contributor to the FIH‐1 spot profile observed.
FIH‐1 appears to be modified in the cell lines investigated here. Clearly however, the
nature of the modification(s) that give rise to the different pIs of each form of FIH‐1 are
unknown. A number of modifications exist that can alter the charge of a protein and
affect its migration in 2‐DE. In order to identify the nature of FIH‐1 modification, the
next chapter investigates the nature of the possible PTMs by purification of FIH‐1 from
cells and analysis by MS.
189
CHAPTER 5
PURIFICATION OF
FIH‐1
191
5 PURIFICATION OF FIH‐1
5.1 Overview
The results of the 2‐DE experiments described in the previous chapter suggested that
there are multiple forms of FIH‐1 in cells. It was therefore of immediate interest to
formally demonstrate FIH‐1 modification and to identify the nature of the
modification(s) that gave rise to the multiple protein spots observed in 2‐D blots. The
aim of this section of work was therefore to purify FIH‐1 from mammalian cell lysate
with sufficient purity and yield to MS for characterisation of any modifications. Such
detailed analysis of PTMs would require μg quantities of purified protein as a minimum.
This was to be done collaboratively, with the MS analysis to be performed by Professor
Jeffrey Gorman and Johana Chicher at the Queensland Institute of Medical Research
(QIMR).
Several methods for FIH‐1 purification from HeLa cells were undertaken. Briefly, the
first strategy for enrichment of FIH‐1 was the IP of endogenous FIH‐1 with three anti‐
FIH‐1 antibodies. Secondly, a novel method was created that utilised the high‐affinity
interaction of FIH‐1 with an ankyrin repeat domain by using a Notch polypeptide for
purification of endogenous FIH‐1 from lysate. Following optimisation of this method,
HeLa cells stably overexpressing a Myc‐6His‐tagged FIH‐1 were created to increase yield
and purity of purified FIH‐1, and the Notch‐affinity method was applied in the
purification of overexpressed FIH‐1. Concurrently, denaturing nickel affinity
chromatography was performed to purify Myc‐6His‐FIH‐1 from stable cell lysate. Details
of each method, including steps taken to optimise each protocol and results achieved,
are described below.
192
5.2 Immunoprecipitation of Endogenous FIH‐1
The first method employed to purify FIH‐1 from cell lysate was IP of FIH‐1 from HeLa
cells. This method was chosen to enable purification of endogenous FIH‐1. HeLa cells
were selected because they were used in much of the 2‐DE experiments that gave
clean, reproducible results, and also because they are a fast growing cell line that
enabled more rapid generation of large amounts of lysate. Initially, small scale IPs were
carried out using three different anti‐FIH‐1 antibodies to assess the ability and efficiency
of each antibody to immunoprecipitate endogenous FIH‐1. Contingent upon good
results being achieved from these preliminary experiments, the overall aim was to
substantially scale up the successful protocol to provide enough material for MS
analysis.
The antibodies used for FIH‐1 IP included two rabbit polyclonal antisera generated in
house against MBP‐FIH‐1, No.8 and No.9, and the commercially available polyclonal
antibody NB 100‐428. The No.8 and No.9 antibodies had not been previously tested for
IP. Protein A sepharose was used to bind the immune complexes and unbound proteins
were removed by washing the resin several times prior to elution by heating the resin to
95°C in SDS sample buffer. Preliminary small‐scale IPs using 200 μg of HeLa lysate and
either the No.8 or No.9 PoAbs demonstrated inefficient binding of FIH‐1, as there was
no enrichment of FIH‐1 in the eluted fraction compared to input protein, and there
were high amounts of non‐specific binding of proteins observed (data not shown).
Extensive optimisation was undertaken to improve both yields and specificity. To reduce
the non‐specific binding of proteins to protein A, the lysate was precleared with protein
A sepharose. The preclearing step was first included alone, and then in combination
with blocking the resin with 0.5 μg/μl BSA for further reduction of non‐specific binding.
193
Total input protein was increased from 200 μg to 500 μg or 1 mg per IP, and the length
of time for both antibody‐antigen complex formation and antibody‐antigen complex
binding to resin was also increased from 2 hours to 6 hours or overnight. Varying
dilutions of each of the three anti‐FIH‐1 antibodies were used to seek a balance
between obtaining a good yield and minimising non‐specific binding of proteins to the
antibodies. The amount of antibody used in the IP was decreased from a dilution of
1:200 to dilutions of 1:500, 1:750 and 1:1000 to try to reduce the high levels of
background proteins binding in the IPs.
A representative experiment using No.9 and NB 100‐428 antibodies is presented in
Figure 5.1. In this IP, 900 μl of HeLa lysate (4 μg/μl) was precleared with Protein A
sepharose. Seven IPs containing 125 μl of the precleared lysate each, or 500 μg total
protein, were then initiated by addition of the antibodies at dilutions of 1:200, 1:500, 1:
750 and 1:1000 for No.9, and 1:200 and 1:400 for NB 100‐428. Immune complexes were
allowed to form during a 2 hour incubation at 4°C. Protein A sepharose was blocked
with BSA and was then added to each IP to bind the antibody‐antigen complexes. Bound
protein was eluted by heating resin in SDS sample buffer, a method which also liberated
the antibody from the resin and resulted in detection of the light (~25 kDa) and strong
detection of the heavy (~50 kDa) chains of the antibody in western blots. The first lane
in Figure 5.1 contains 10% input lysate, and the following six lanes contain the results of
the six different IPs performed. To help discern which bands are due to elution and
reduction of the antibody, an antibody only “IP” was included in this experiment (lane 8
of Figure 5.1). It is evident from this western blot that much of the “background” seen
in the IP lanes is derived from the reduction of the antibodies into their constituent
subunits. A lysate‐only negative lane was also included to ascertain if, following pre‐
clearing and blocking, proteins in the lysate were still binding non‐specifically to the
resin (lane 9). That there are no significant bands around the size of FIH‐1 present in this
lane suggested that non‐specific binding of proteins to the resin did not contribute to
No.9 IP
75100
kDa
50
37
Figure 5.1 Immunoprecipitation of FIH
25
1 2 3
Figure 5.1 Immunoprecipitation of FIHHeLa cell lysates were incubated with pprotein‐A. 500 μg of lysate was incubavolume of 1 ml in binding buffer, andhours. Protein‐A sepharose beads wcontaining 0.5% BSA at 4°C for 1 houovernight incubation at 4°C. Resin wasPAGE d l d b i bl iPAGE and analysed by immunoblottingRepresentative of 6 independent exper
PNB 100‐428 IP
h
FIH‐1
Heavy chain
Light chain
‐1
4 5 6 7 8 9WB: Anti‐FIH‐1 No.9
1protein‐A sepharose to reduce proteins that bind toated with the indicated antibody dilution in a totalantibody‐antigen complexes were formed over 2
were blocked by incubation in blocking bufferur prior to addition of 70 μl of slurry per IP andwashed and protein eluted and separated by SDS‐P di d l l i h f FIH 1 i i di dg. Predicted molecular weight of FIH‐1 is indicated.
iments.
197
the background bands observed in the blots. Comparing the results of each IP to the
10% input lane, it is clear that despite attempts made at improving the yield of FIH‐1,
the amount of FIH‐1 following IP using either antibody were poor. Decreasing the
amount of antibody in the IP has a dramatic and deleterious effect on FIH‐1 yield, even
when increasing the dilution from 1:200 to 1:500 (compare lanes 2 and 3). It is difficult
to estimate the amount of immunoprecipitated FIH‐1 compared to input as the input
lane contains two intense bands very close in size, whereas the IP lanes contain only
one. What is clear, however, is that the amount of FIH‐1 precipitated from the lysate is
far less than that contained in 10% of the input material, indicating that the efficiency of
this method in purifying FIH‐1 from cells is inadequate for subsequent analysis. This
figure is representative of six experiments. Altering other parameters, such as
increasing the amount of input protein, did not achieve significant improvements in the
amount of FIH‐1 enriched in the samples following IP without also increasing amount of
non‐specific proteins also precipitated (data not shown). This suggested that the three
polyclonal antibodies employed for these experiments exhibited poor specificity in
precipitating FIH‐1.
Western blots revealing immunodepletion of lysates following the IP reiterated the
inefficiency of the antibody binding, showing a scant reduction in FIH‐1 in lysate
following IP (data not shown). Thus, despite numerous attempts at
immunoprecipitating FIH‐1 and altering a number of parameters in the IP protocol, it
was not possible to immunoprecipitate endogenous FIH‐1 from lysate in sufficient
enough quantities to enable a scale‐up of the method and subsequent mass
spectrometric analysis. As this was thought to be due to both the inefficiency of
antibody‐FIH‐1 interaction and the low level of FIH‐1 present in the lysate, alternative
strategies were devised.
198
5.3 Affinity pulldowns of FIH‐1 using Notch1 Ankyrin Repeats 1‐4.5
5.3.1 Strategy of Notch1 Ank1‐4.5‐affinity purification
Due to the inefficiency of the anti‐FIH‐1 antibodies at immunoprecipitating endogenous
FIH‐1, an alternative method for binding and thus purifying FIH from the lysate was
devised. During the course of this work, it was discovered by our laboratory and others
that the Notch receptor was also an interacting protein and substrate of FIH‐1 (Zheng et
al., 2008) (Coleman et al., 2007). Of significance here, mouse Notch1 polypeptides were
found to bind FIH‐1 with up to 50‐fold greater affinity than HIF‐α (Wilkins et al., 2009),
and pulldown experiments assessing the binding of FIH‐1 for the HIF‐CAD and FIH‐1 for
the ankyrin repeats of Notch also showed a high affinity FIH‐1‐Notch interaction, far
greater than the affinity of FIH‐1 for the CAD (S. Linke, unpublished data). In particular,
a truncated polypeptide containing ankyrin repeats 1‐4.5 of Notch1 (Notch1 Ank1‐4.5),
which encompasses both known hydroxylation sites, was shown to interact with FIH‐1
with an even greater affinity than the full length ankyrin repeat region ( Figure 5.2 a;
(Linke, 2009).
Given this relatively high affinity and the ability to make mg quantities of recombinant
Notch1 Ank1‐4.5 protein, a Notch1 Ank1‐4.5‐affinity pulldown was employed for
purification of endogenous FIH‐1 from cell lysate. The methodology involved bacterial
expression of a Trx‐6His‐tagged Notch1 Ank1‐4.5 construct (represented in Figure 5.2
b), nickel‐affinity purification, and incubation of Notch1 Ank1‐4.5‐bound resin with HeLa
lysate. Notch1 Ank1‐4.5 would act as bait to bind FIH‐1 from the cell lysate, exploiting
the previously demonstrated stable, high‐affinity interaction of this Notch peptide for
FIH‐1 (Wilkins et al., 2009).
a.
Thrcleav
b. Schematic representation of Notch
cleav
TrxN 6His
18 kDa
Figure 5.2 Notch Ank 1‐4.5‐pulldown ofTrx‐6His‐tagged Notch proteins containTrx 6His tagged Notch proteins containbacterially expressed and purified. Equresin and incubated with purified MBPFIH‐1 visualised by an anti‐MBP westeSchematic representation of the Notchpurification of FIH‐1 (b).
rombin vage site
construct
Notch Ank 1‐4.5
vage site
6His C
20 kDa
f FIH‐1.ning the RAM domain, Ank1‐4.5 or Ank1‐7 werening the RAM domain, Ank1 4.5 or Ank1 7 wereal amounts of each protein were bound to nickelP‐FIH‐1. Proteins were eluted with imidazole andern blot (a); from Sarah Linke, unpublished data).h Ank 1‐4.5 construct used for the Notch‐affinity
201
5.3.2 Results of Notch1 Ank1‐4.5‐affinity purification
Trx‐6His‐Notch1 Ank1‐4.5‐6His was bacterially expressed and applied to nickel resin as
per the standard laboratory protocol for purification of His‐tagged proteins. A sample of
resin was taken and protein eluted by imidazole and run on a gel to visualise the
purified product (Figure 5.3 a). Coomassie staining demonstrated successful expression
and purification of the protein. For the pulldown, 50 μl of the Trx‐6His‐Notch1 Ank1‐4.5‐
6His‐loaded resin was incubated with 200 μl of HeLa cell lysate (1 mg of total protein)
for 2 hours prior to washing and elution of both Notch and FIH‐1 off the resin with 250
mM imidazole. Depletion of FIH‐1 from the lysate was used to assess efficiency of this
method (Figure 5.3 b). Equivalent amounts of input lysate and depleted lysate were
separated by SDS‐PAGE and blotted for FIH‐1. Depletion of FIH‐1 from the lysate was
evident when comparing levels of FIH‐1 in the input to FIH‐1 present in lysate after
binding (compare lane 1 with lane 5 and lane 2 with lane 6, Figure 5.3 b) and when
compared to the non‐specific background band at approximately 55 kDa which remains
unaltered before and after the pulldown. Lane 9 contains eluted FIH‐1, which is several
fold more intense than 5% input and consistent with an overall yield of FIH‐1 of greater
than 25%. These results indicate that this method was approximately 10‐fold more
efficient at purifying endogenous FIH‐1 from the lysate than IP.
5.3.3 Optimising elution of FIH‐1
Elution of both Trx‐6His‐Notch1 Ank1‐4.5‐6His and FIH‐1 proteins simultaneously using
imidazole presented a difficulty as the Notch1 Ank1‐4.5 construct was 38 kDa and
present in great excess, making separation from FIH‐1, which is 40 kDa, and
visualisation of any eluted FIH‐1 on a coomassie stained gel problematic. Several
alternative elution methods were employed with the aim of disrupting the Notch‐FIH‐1
interaction, thus liberating FIH‐1 from the resin whilst leaving Trx‐6His‐Notch1 Ank1‐
a. Purified Trx‐6His‐Notch1 Ank 1 ‐4.5‐6His
b. FIH‐1
75100
50
5%
50
75
kDa
kDa
37
Coomassie Stain
1
37
Figure 5.3 Depletion of FIH‐1 from lysaTrx‐6His‐Notch1 Ank 1 ‐4.5‐6His was expchromatography. 20 μl of eluted proteinvisualised by Coomassie stain (a). (b) shNotch‐affinity purification. 50 μl of Notcμl (1 mg) of HeLa lysate and bound for 9binding to enable assessment of FIH‐1 dwith 250 mM imidazole for 15 minutes, Predicted molecular weight of FIH‐1 is iexperimentsexperiments.
Depletion following Notch‐pulldown
I L D l t d L t H‐1
*
% 2.5% 1% 0.5% 5% 2.5% 1% 0.5%
Input Lysate Depleted Lysate
Eluted FIH
n.s.
WB: anti‐FIH‐1 NB 100‐428
1 2 3 4 5 6 7 8 9
FIH‐1
te by Notch Ank1‐4.5 pulldownpressed in bacteria and purified by Nickel affinity n was separated by 10 % SDS‐PAGE and proteins ows depletion of FIH‐1 from HeLa lysate following ch Ank 1‐4.5‐ bound nickel resin was added to 200 90 minutes at 4°C. The lysate was retained after depletion. Bound FIH‐1 was eluted by incubation and results visualised by anti‐FIH‐1 immunoblot. indicated. Representative of >3 independent
205
4.5‐6His attached. First, 8 M urea was used to denature the proteins and thus destroy
FIH‐1 and Notch binding whilst retaining the electrostatic interaction between the
histidine tags and the nickel. Secondly, 1% SDS was used as an alternative denaturant to
attempt to disturb the protein‐protein interaction without disturbing Notch from the
nickel resin. Thirdly, the Trx‐6His‐Notch1 Ank1‐4.5‐6His construct contained a thrombin
cleavage site between the N‐terminal Trx‐6His tag and the beginning of the Notch
ankyrin repeat sequence (see Figure 5.2 b) for liberation of the Notch fragment from
the affinity tag. Cleavage with thrombin would thus separate the tagged protein into
two smaller fragments of 18 and 20 kDa each, sufficiently different in size from 40 kDa
FIH‐1. This method, however, would presumably only liberate a proportion of bound
FIH‐1, as the Notch construct contained both N‐ and C‐terminal 6His tags. Removal of
the remaining Notch (and thus any remaining FIH‐1) would require a second elution
step of imidazole that would remove the C‐terminally‐bound Notch from the resin.
Results achieved when employing these various elution methods are presented in
Figure 5.4. Included in the figure is a ponceau stain of the membrane to crudely
demonstrate the abundance of Notch1 Ank1‐4.5 in each sample following the various
elution methods. The sensitivity of ponceau stain enabled visualisation of the Notch
polypeptide, which would be the most abundant protein in each sample, but not FIH‐1
which would be present at much lower levels (Figure 5.4 a). In Figure 5.4, lane 1 shows
5% input and the following lanes contain the results of Notch pulldowns eluted with
imidazole (lane 2), 1% SDS (lane 3), 8 M urea (lane 4) and thrombin digestion (lane 5).
The ponceau stain revealed that a significant amount of Notch1 Ank1‐4.5 was eluted
with imidazole, as expected, but also to a lesser extent with the SDS and Urea elution
methods. The thrombin‐digestion elution method was the only strategy that liberated
FIH‐1 from the resin without coelution of a 38 kDa Notch protein. Though it is not clear
how complete the cleavage of Notch1 Ank1‐4.5 by thrombin was, no full length Trx‐
6His‐Notch1 Ank1‐4.5‐6His was detected by ponceau staining, which suggested that this
Elution method
a. 5% input
75
50
kDa
25
50
37
20
NA
20
Ponceau Stain
1 2 3 4 5
Figure 5.4 Optimisation of elution of N50 μl of Notch1 Ank1‐4.5‐ bound nickeand bound for 90 minutes at 4°C. Foleither: 250 mM imidazole, 1% SDS, 8 Mwas separated by SDS‐PAGE, transferPonceau (a). FIH‐1 was detected using N( ) g
Elution method
b.
5% input
75
50FIH 1
kDa
25
50
37
20
FIH‐1
Notch Ank 1‐4.5
20
WB: anti‐FIH‐1 NB 100‐428
1 2 3 4 5
Notch1 Ank1‐4.5 pulldownel resin was added to 200 μl (1 mg) of HeLa lysatelowing washing, bound protein was eluted usingM Urea, or thrombin, as indicated. Eluted proteinrred to nitrocellulose and the blot stained withNB 100‐428 (b).( )
209
method liberated only cleaved Notch fragments. Imidazole elution achieved the
greatest FIH‐1 yield (Figure 5.4 b), however it also resulted in the greatest level of Trx‐
6His‐Notch1 Ank1‐4.5‐6His elution and was therefore deemed to be an unsuitable
method for FIH‐1 elution.
The amount of FIH‐1 eluted by the various methods seems to be a result of how much
Trx‐6His‐Notch1 Ank1‐4.5‐6His was also eluted off the Nickel resin, as where 1% SDS
and 8 M urea yielded less FIH‐1, they also co‐eluted with less Trx‐6His‐Notch1 Ank1‐4.5‐
6His. This suggests that the methods employed did not disrupt only the FIH‐1‐Notch
interaction, but that the elution is a result of a disruption in Notch binding to the nickel
resin. In Lane 5 (Figure 5.4 b), the amount of FIH‐1 eluted by thrombin digestion is
significantly less than the amount of FIH‐1 recovered using the other methods.
However, as this was the only method able to achieve elution of FIH‐1 without large
amounts of the Notch fragment, this method was used for subsequent pulldowns.
Figure 5.5 shows a separate pulldown where, following elution of FIH‐1 with thrombin
(lane 2), a second elution step performed using 250 mM imidazole (lane 3) liberated a
similar amount of FIH‐1 to the thrombin method, showing that only approximately half
of the total bound FIH‐1 was freed by thrombin cleavage of the N‐terminal tag of
Notch1 Ank1‐4.5 containing both N‐ and C‐terminal 6His tags. The thrombin cleavage
appeared to be complete, as there was no 38 kDa Notch protein evident on a ponceau
stain of the blot in the lane of protein eluted by imidazole following incubation of the
resin with thrombin (lane 3 and lane 5). This complete cleavage of the interfering 38
kDa protein engendered a two‐step elution protocol that first used thrombin to cleave
Trx‐6His‐Notch1 Ank1‐4.5‐6His into 20 kDa and 18 kDa fragments, followed by an
imidazole elution that enabled liberation of FIH‐1 bound to Notch that was attached to
nickel resin via its C‐terminal 6His tag. In this manner, all the FIH‐1 captured by Notch
ut after
after
+ Lysate ‐ Lysatea.
75100
50
10% inpu
Thrombin
Imidazole
Thrombin
Thrombin
Imidazole
Thrombin
1 2 3 4 5
50
37
25
20
Ponceau
Figure 5.5 Elution of FIH‐1 by thrombin50 μl of Notch1 Ank1‐4.5‐ bound nickel50 μl of Notch1 Ank1 4.5 bound nickel200 μl of buffer only, and bound for 90was eluted by incubation of the resintemperature. Following thrombin cleavincubated with 250 mM imidazole to eseparated by SDS‐PAGE, transferred to(a). FIH‐1 was detected using NB 100‐42
t after
after
+ Lysate ‐ Lysateb.
10% inpu
kDa
75100
50
Thrombin
Imidazole
Thrombin
Thrombin
Imidazole
Thrombin
Thrombin
1 2 3 4 5
50
37
25
20
WB: anti‐FIH‐1 NB 100‐428u
n cleavage of Notch1 Ank1‐4.5resin was added to 200 μl (1 mg) of HeLa lysate orresin was added to 200 μl (1 mg) of HeLa lysate or
0 minutes at 4°C. Following washing, bound proteinn with 10 units thrombin for 16 hours at roomvage, the supernatant was removed and the resinelute the remaining protein. Eluted protein wasnitrocellulose and the blot stained with Ponceau
28 (b).
213
affinity could be eluted without interference from an excessive amount of similarly
sized purified Notch1 Ank1‐4.5 protein.
5.3.4 Summary of Notch Ank1‐4.5‐affinity purification of FIH‐1
In summary, a method for the efficient purification of FIH‐1 from HeLa cells using Notch
Ank1‐4.5‐affinity was created and optimised. This method was scaled up to purification
of endogenous FIH‐1 from up to 2 ml of HeLa lysate (~5 mg/ml) prepared from two 175
cm2 flasks. Yields from these purifications, however, did not deliver enough FIH‐1 to
enable detection by SDS‐PAGE using coomassie stain and Sypro Ruby stain, and not
surprisingly the yields did not return enough FIH‐1 for a detailed characterisation of
PTMs by MS (data not shown). The next step therefore was to create a cell line stably
overexpressing FIH‐1 to increase the amount of FIH‐1 in the lysate to facilitate greater
yields.
5.4 Generation of Stable Myc‐6His‐hFIH‐1 HeLa Cell Line
5.4.1 Strategy
Purifying a sufficient amount endogenous FIH‐1 by either IP or Notch‐affinity pulldown
for MS analysis was made difficult by the low abundance of FIH‐1 in cells. Increasing
inputs of lysate into each method (up to 1.5 mg/Notch‐affinity pulldown and 1 mg per
IP) did increase yields of FIH‐1, but at the sacrifice of specificity, resulting in both
increased FIH‐1 and increased background binding, despite measures undertaken to
increase stringency. Thus, stable cell lines overexpressing Myc‐6xHistidine (6His)‐
tagged FIH‐1 were generated to enhance the proportion of FIH‐1 in the lysate. The Myc‐
tag was included to enable reliable, clean detection of the tagged protein by anti‐Myc
214
antibodies and the 6His tag would enable purification of Myc‐6His‐FIH‐1 by Ni2+‐affinity
chromatography.
The vector chosen for creation of the stable cell lines was pEF‐IRES‐puro5. This vector
expresses a bicistronic message incorporating both the puromycin resistance mRNA
required for selection and the mRNA of the protein to be overexpressed. Expression is
driven by a strong human elongation factor‐1α (EF‐1α) promoter and an internal
ribosome entry site (IRES) ensures that both messages are translated efficiently.
5.4.2 Generation of stable Myc‐6His‐hFIH‐1 HeLa polyclonal cell line
Full length human FIH‐1 was excised from pPC86‐FIH‐1 by Nco1 and Not1 digestion and
ligated into pEF‐IRES‐puro5 upstream of the IRES. Oligonucleotides containing a
sequence coding for the Myc‐6His tag and for ligation at the N‐terminus of FIH‐1
sequence were designed. The oligonucleotides were annealed and gel purified prior to
ligation into the vector. Test digests and sequencing both confirmed the correct
integration of the tag N‐terminal to FIH‐1, and the construct was subsequently
transfected into HeLa cells. Cells expressing FIH‐1 and puromycin N‐acetyltransferase
(PAC) were selected for by addition of 1 μg/ml puromycin in the cell culture media, and
expression of FIH‐1 by the polyclonal stable cell pool confirmed by anti‐Myc and anti‐
FIH‐1 western blots. Figure 5.6 shows that compared to endogenous FIH‐1 in HeLa cells,
the polyclonal cells expressed the transgene at high levels, at least 10‐fold higher when
compared to endogenous FIH‐1. It was of interest to observe that endogenous FIH‐1
migrated as two bands, as seen previously, and that overexpressed Myc‐6His‐FIH‐1,
visible above endogenous FIH‐1 and detected using both anti‐FIH‐1 and anti‐Myc
antibodies, also migrated as a doublet.
50
kDa
37
WB: anti‐Myc
Figure 5.6 Stable overexpression of MyThe polyclonal pool of cells stablyexpression by anti‐Myc (9E10) and antcompared to lysates of untransfected H
Myc‐6His‐FIH‐1y
WB: NB 100‐428
Endogenous FIH‐1
yc‐6His‐FIH‐1 in polyclonal HeLa cell lineexpressing Myc‐6His‐FIH‐1 was examined for
ti‐FIH‐1 (NB 100‐428) western bots of lysates, andeLa cells.
217
2‐DE of stable cell lysate was performed to confirm that the spot profile for Myc‐6His‐
FIH‐1 was consistent to that observed for endogenous FIH‐1 (Figure 5.7). 50 μg of lysate
from the Myc‐6His‐FIH‐1 stably‐overexpressing polyclonal pool was separated by 2D‐E.
Figure 5.7 a shows an anti‐Myc immunoblot of one such 2‐D gel, where a spot “train” of
9 individual spots at approximately the correct size for Myc‐6His‐FIH‐1, which is 42.2
kDa, was observed. The anti‐Myc antibody again revealed detection of two different
molecular weight forms of Myc‐tagged FIH‐1, suggesting that the modification of FIH‐1
responsible for the discrete migration of two bands observed previously is also
occurring on overexpressed protein. A separate experiment where FIH‐1 was visualised
by anti‐FIH‐1 immunoblot enabled comparison of the overexpressed FIH‐1 with
endogenous FIH‐1 (Figure 5.7 b). Myc‐6His‐FIH‐1 is positioned in a more basic region of
the 2‐D gel compared with endogenous FIH‐1, as expected given the tag increases the
predicted pI to ~5.98 (compared with 5.57 for FIH‐1; Scripps Protein Calculator). Only
two spots are visible for endogenous FIH‐1 due to the short exposure time of the
membrane to film to enable a clearer image of overexpressed FIH‐1, and also due to the
relatively low level of endogenous FIH‐1 compared to the tagged FIH‐1. Together, these
data indicate that the Myc‐6His‐tagged overexpressed FIH‐1 displays a similar spot
profile to endogenous FIH‐1 in 2‐D gels, taking into account differences in size and pI
due to the introduction of the tag.
5.4.3 Generation of Myc‐6His‐FIH‐1 monoclonal HeLa cell lines
From the above 2‐DE westerns, it was clear that Myc‐6His‐FIH‐1 was being expressed at
much greater levels than endogenous FIH‐1. Too great an overexpression was
undesirable as it was important not to saturate any modifying mechanisms acting upon
FIH‐1 in the cell. Monoclonal cell lines were thus created, with the objective of assessing
each individual line for expression of Myc‐6His‐FIH‐1 relative to endogenous FIH‐1 and
the identification of a lower expressing line than the polyclonal pool. Use of a
a
75
a.
kDa
4 p
50
37
b. 4 pH
75
kDa
50
37
Figure 5.7 2‐DE of lysate from stable MLysates were prepared from the Mycsample preparation Method 2 and precquantification, 50 μg of each lysate wastrips were applied to 12.5% Tris/Gldetected using the anti‐Myc antibodydetected using the NB100‐428 antibody
H 7
Myc‐6His‐FIH‐1
WB: anti‐Myc 9E10
H 7
WB: anti‐FIH‐1 No. 8
Myc‐6His‐FIH‐1
Endogenous FIH‐1
Myc‐6His‐FIH‐1 polyclonal poolc6His‐FIH‐1 polyclonal HeLa cell line using 2‐DEcipitation. Following resuspension in IEF buffer andas loaded onto pH 4‐7 IPG strips for IEF and thenycine gels for SDS PAGE. Transgenic FIH‐1 was(a) and both endogenous and tagged FIH‐1 werey (b). Representative of 3 individual experiments.
221
monoclonal cell line was advantageous as it allowed for consistent expression and
therefore modification of FIH‐1 in the entire population of cells, and therefore
consistent modification of FIH‐1 in a sample.
Stable monoclonal cells lines were generated by performing limiting dilutions of cells
from the polyclonal pool and plating out a calculated 0.8 cells/well in a 96 well tray.
Wells that grew colonies derived from single cells were expanded and assessed for
expression of Myc‐6His‐FIH‐1 by anti‐Myc western blot (Figure 5.8). The expression of
tagged FIH‐1 was seen to vary, as seen when comparing expression between
monoclonal lines 3, 10 and 11, which express a low level of Myc‐6His‐FIH‐1, versus the
higher expression levels of monoclonal cell line 6. Cell line 8 was discarded due to the
presence of an additional band at approximately 40 kDa, and lines 5 and 7 discarded
due to morphological differences compared to parent HeLa cells.
Lysates from monoclonal lines 2, 6 and 9 were interrogated by 2‐DE to determine if
their spot profile was similar to endogenous FIH‐1 and to that observed for FIH‐1 in the
polyclonal lysate, and also to assess relative levels of expression compared to
endogenous FIH‐1. The results of 2‐D analysis of monoclonal stable cell lines 2 and 9
and the polyclonal stable line are presented in Figure 5.9 a, b and c, respectively. Here,
100 μg of lysate from each cell line was separated by 2‐D PAGE, and the results
visualised using the NB 100‐428 anti‐FIH‐1 antibody to enable detection of both
overexpressed Myc‐6His‐FIH‐1 and endogenous FIH‐1. What is evident is that
comparatively, expression of transgenic FIH‐1 is greater than endogenous FIH‐1 in both
the stable cell lines, but lower than the dramatic overexpression of Myc‐6His‐FIH‐1 in
the polyclonal line. To achieve less intense detection of overexpressed FIH‐1, blots were
washed and exposed for shorter times. While this enabled clear visualisation of
individual protein spots of Myc‐6His‐FIH‐1, detection of endogenous FIH‐1 was lost
(Figure 5.9, right hand panels). The spot profile of stable cell line number 9 was most
Monoc
1 2 3 4
75
100
kDa
37
50
Figure 5.8 Comparison of Myc‐6His‐FIHLysates were prepared from untransfectexpressing Myc‐6His‐FIH‐1. 20 μg total using anti‐Myc antibody 9E10. The non‐non‐specific
clonal lines:
4 5 6 7 8 9 10 11
* n.s.
Myc‐6His‐FIH‐1
H‐1 expression in stable monoclonal cell lines
WB: anti‐Myc
ted HeLa cells and 12 monoclonal cell lines stably protein was loaded per lane and blots probed ‐specific band served as a loading control. *n.s. =
a 4 pH
50
75
100
kDaa. 4 pH
37
25
b. 4 pH
50
75
100
kDa
37
25
4 pH c.kDa
50
75
100
37
25
WB: anti‐F
Figure 5.9 2‐DE of lysates from monocloLysates were prepared from Myc‐6His‐Fpool (c) using 2‐DE sample preparation Mbuffer to 100 μg in 125 μl for separatwestern blot using the commercialextensively and ECL reagents re‐applied
th i ht)on the right).
7
STABLE CELL LINE NUMBER 2
7
7
STABLE CELL LINE NUMBER 9
7
POLYCLONAL STABLEPOLYCLONAL STABLE CELL LINE
FIH‐1 NB 100‐428
onal and polyclonal Myc‐6His‐FIH‐1 stable cell linesIH‐1 stable cell lines 2 (a), 9 (b) and the polyclonalMethod 2, then precipitated and resuspended in IEFion by 2D‐E. Results were visualised by anti‐FIH‐1NB 100‐428 antibody. Blots were then washedto generate a less intense signals (blown up panels
227
similar to the spot patterns observed in other cell lines previously. In addition, this cell
line expressed Myc‐6His‐FIH‐1 significantly above endogenous FIH‐1, but modestly
enough that saturating any cellular mechanism or enzymes that act in a cell to modify
FIH‐1 was less likely to be a problem compared with the high overexpression observed
in the polyclonal pool. Thus, stable cell line number 9 was selected for purification of
Myc‐6His‐FIH‐1 using Ni2+‐affinity chromatography. Although not discussed in this
chapter, the Notch‐affinity method was also used for purification of Myc‐6His‐FIH‐1,
and though this method did appear to be more efficient at binding FIH‐1 from the lysate
than Ni2+‐affinity, it was ultimately not feasible due to the co‐elution of the Notch
protein.
5.5 Ni2+‐Affinity Chromatography
5.5.1 Small scale Ni2+‐affinity purifications
Denaturing Ni2+‐affinity purifications were carried out on lysate from the number 9
stable cell line. Briefly, this procedure involved the lysis of cells in a denaturing
guanidine buffer containing DTT, followed by alkylation using iodoacetamide. Reduced
and alkylated lysates were clarified by ultracentrifugation and filtered prior to
application over a Ni2+‐charged 5 ml HiTrap column. Bound protein was eluted off the
column by 250 mM imidazole and elution fractions were assessed for purified Myc‐6His‐
FIH‐1 by anti‐Myc or anti‐FIH‐1 western blot and coomassie gels. Fractions found to
include Myc‐6His‐FIH‐1 were pooled, frozen and sent to collaborators at QIMR for MS.
Initially, purifications were performed from lysate prepared from four 175 cm2 flasks of
confluent Myc‐6His‐FIH‐1 stable cell line number 9 (#9) cells (Figure 5.10 a). 15 ml of
guanidine lysis buffer was added to each flask to give a total of 60 ml of lysate from
which to purify Myc‐6His‐FIH‐1. Western blots showed that elution fractions 2, 3 and 4
Elution Fractions (10 l)a.
1 2 3 4 5 6 7 8
75
100
150
kDa
37
50
Ant
Elution Fractions (10 l)
b.
75
100
150
kDa1 2 3 4 5
( )
37
50
Anti‐Myc WB
Figure 5.10 Ni2+‐Affinity purification of MFour flasks (a) or twelve flasks (b) of stabconfluency and then lysed in guanidine lyconfluency and then lysed in guanidine lyby ultracentrifugation and filtered. Myc‐6affinity chromatography, and eluted in 1 protein was visualised by anti‐Myc immu
Elution Fractions (45 l)
9 10 2 3 4
75
100150
250
kDa
37
50Myc‐6His‐FIH
ti‐Myc WB Coomassie stain
Elution Fractions (45 l)
2 3 4
75
100
150250
kDa
37
50
75
Myc‐6His‐FIH
Coomassie stain
Myc‐6His‐FIH‐1ble cell line number 9 were grown to ~90% ysis buffer Lysates were reduced alkylated clarifiedysis buffer. Lysates were reduced, alkylated, clarified 6His‐FIH‐1 was purified from the lysate by nickel ml fractions using 250 mM imidazole. Eluted unoblot and coomassie stained gels.
231
all contained significant amounts of Myc‐6His‐FIH‐1, and that no FIH‐1 was detected in
later elution fractions. Coomassie‐stained SDS‐PAGE gels of these three fractions
showed the presence of a distinct yet faint band at the predicted molecular weight for
Myc‐6His‐FIH‐1. Non‐specific binding of proteins to the column also occurred, as
evidenced by numerous background bands present in each elution fractions.
5.5.2 Scale‐up of Ni2+‐affinity purifications
Purifications were scaled up, and the results of one such purification are presented in
Figure 5.10 b. Here, 180 ml of lysate was made from twelve 175 cm2 flasks of #9 stable
HeLa cells and Myc‐6His‐FIH‐1 purification was carried out as described above. Bound
protein was eluted with 10 ml of elution buffer collected in 1 ml fractions. Blotting with
the anti‐Myc antibody showed purified FIH‐1 eluting in fractions 2, 3, and 4, with most
protein eluting off the column in fraction 3, and no FIH‐1 present in fractions 5 and
beyond (Figure 5.10 b). Coomassie stained gels of each of the fractions showed that a
band was present at the expected size for Myc‐6His‐FIH‐1; however this band is faint
relative to other bands present on the gel, and it is not the most intense band in the
lane. Also, increasing the amount of input lysate three‐fold here, from lysate made from
four flasks (Figure 5.10 a) to lysate prepared from twelve flasks (Figure 5.10 b), did not
appear to significantly increase the amount of FIH‐1 obtained following the purification.
Further purifications were performed and the purified fractions assessed for Myc‐6His‐
FIH‐1 by immunoblot and coomassie‐stained gel as above. Results from purifications
consistently showed elution of Myc‐6His‐FIH‐1 in fractions 2‐4 by immunoblot, and the
presence of a clear, yet not intense, band at the expected size in coomassie‐stained
gels. Purification of Myc‐6His‐FIH‐1 from lysate prepared from twelve 175 cm2 flasks of
again showed the presence of faint bands on coomassie‐stained gels at the expected
size (Figure 5.11 b). Despite the presence of more intense background bands, the band
a.
1 32 54 6
Elution Fractions (10l)
50
kDa
50
37
anti‐Myc WB
Figure 5.11 Ni2+‐Affinity purification ofTwelve flasks of stable cell line numberin guanidine lysis buffer Lysates werein guanidine lysis buffer. Lysates wereand filtered. Myc‐6His‐FIH‐1 waschromatography, and eluted in 1 mimmunoblot was used to assess whichindicated volumes of fraction 3 were seeither stained using coomassie or tranElution fractions 3 and 4 were pooled, fp
b.
2 l
10 l
45 l
150250
Volume of elutionfraction 3
kDa kDa
150250
Myc‐50
75
100
150
50
75
100
150
Myc6His‐FIH
37
25i M C i
37
25
f Myc‐6His‐FIH‐1r 9 were grown to ~90% confluency and then lysedreduced alkylated clarified by ultracentrifugation
anti‐Myc WB
Coomassie
reduced, alkylated, clarified by ultracentrifugationpurified from the lysate by nickel affinity
ml fractions using 250 mM imidazole. Anti‐Mych fractions contained Myc‐6His‐FIH‐1 (a), and theeparated by SDS‐PAGE and the gels cut in half andnsferred to nitrocellulose for anti‐Myc blotting (b).frozen, and sent to QIMR for MS.
235
thought to be Myc‐6His‐FIH‐1 was separated enough from other visible protein bands to
enable excision of a single band and MS analysis. Elution fractions 3 and 4 from this
purification were pooled and sent for MS analysis. The yields of FIH‐1 in these bands
were low and difficult to detect and significantly lower than required for detailed
proteomic analysis.
5.5.3 Optimisation of Ni2+‐affinity chromatography
Several strategies were employed with the aim of improving both the yield and purity of
FIH‐1 purified by this method. Further work performed to improve results from Ni2+‐
affinity chromatography included further increases in input lysate. Lysate was prepared
from thirty 175 cm2 flasks of #9 Myc‐6His‐FIH‐1 stable cells lysed in 10 ml of lysis buffer
per flask, however this method failed to deliver any improvements in yield of FIH‐1
(data not shown). To increase purity of eluted protein, imidazole gradient washes were
employed rather than a single concentration of 250 mM imidazole for elution. A
technique used to increase efficiency of purification of FIH‐1 from the lysate was to
switch to a batch purification format rather than the previously used column
purification method. This can be advantageous as it allows for a longer time for binding
of the target protein to the resin. Additionally, while the concentration of lysate is
limited in column purification, lysates used in batch purification can be much more
concentrated and thus more lysate can be used for the purifications. Where 10 ml of
guanidine extract buffer was used for lysis of one 175 cm2 flask, here, 1 ml of whole cell
extract buffer was used for lysis of the same amount of cells. Lysis in WCEB also enables
gels of lysate before and after the purification to be run for assessment of depletion of
FIH‐1.
A number of batch purifications were performed. The most successful batch purification
result is presented in Figure 5.12. Here, three separate purifications were carried out,
% Input
Dep
leted Lysate
μl (10%) eluate
10 μl (20%) eluate
1%
D 5
75
100
150
kDa
Lane: 1 2 3 4
50
37
Figure 5.12 Batch Purification of Myc‐6Three purifications were performedincubated with 100 μl of prepared Ni‐4°C. The resin was washed three times
Lane: 1 2 3 4
Anti‐Myc WB
4 C. The resin was washed three timesIn sample 1, protein was eluted using 5eluted by heating resin for 10 minutes awas washed with increasing imidazolemM prior to elution by heating resin fbuffer. All samples were alkylated prindicated amounts of each sample we
i i f i ll lstaining or transfer to nitrocellulose an6His‐FIH‐1 were excised and sent to Qbands that were excised.* Due to a batch of Glycine Tris SDS pthis period exhibited convoluted fronts,
ample 1
ample 2
ample 3
Sa Sa Sa
Myc‐6His‐FIH‐1
5 6 7 8 9*
6His‐FIH‐1where 1.5 ml (~5.5 mg of total protein) was‐IDA resin for two hours on a rotating platform ats, with the final wash containing 10 mM imidazole.
5 6 7 8 9Coomassie
s, with the final wash containing 10 mM imidazole.0 μl of 250 μM imidazole. In sample 2, protein wasat 70°C in 50 μl SDS‐PAGE sample buffer. Sample 3concentration of 20 mM, 30 mM, 40 mM and 50for 10 minutes at 70°C in 50 μl SDS‐PAGE samplerior to SDS‐PAGE. SDS‐PAGE gels containing theere run and gels cut in half for either Coomassied i bl i d dind anti‐Myc blotting. Bands corresponding to Myc‐QIMR for MS analysis. Dotted red boxes indicate
prepared with the incorrect pH, all gels run during, as seen here.
239
each consisting of 1.5 ml of lysate, containing 5.5 mg of total protein, incubated with
100 μl of prepared Ni‐IDA resin. Protein was bound to the resin during a two hour
incubation on a rotating platform at 4°C, and then the resin was subjected to two
washes in wash buffer 1 (150 mM NaCl, 20 mM Tris, 10% glycerol, 1% NP‐40) followed
by one wash in wash buffer 2, which has 10 mM imidazole added. Following these
common washing steps, each sample was treated differently. Protein in sample 1 was
eluted using 50 μl of 250 μM imidazole (Lane 5), and protein in sample 2 was eluted by
heating resin for 10 minutes at 70°C in 50 μl SDS‐PAGE sample buffer (Lane 7). The third
sample was washed with wash buffers of increasing imidazole concentration (20 mM,
30 mM, 40 mM and 50 mM imidazole; Lane 9). Previous work had established that
washes of up to 50 mM imidazole could be applied to significantly improve purity of
eluted protein without major losses of the desired protein (data not shown). Following
the imidazole gradient washes, bound protein was eluted by heating resin for 10
minutes at 70°C in 50 μl SDS‐PAGE sample buffer, seen in the lane marked Sample 3
(Lane 9). Eluted material was compared on western blots and coomassie‐stained gels,
and results indicated that there was a clear band present at the expected position for
FIH‐1 in all three samples, and while there was still a lot of background evident, the
amount of Myc‐6His‐FIH‐1 purified relative to background proteins was improved.
Looking at the depletion of FIH‐1 from the input lysate in the anti‐Myc immunoblot, this
method also appeared to be capturing most of the FIH‐1 from the lysate (Figure 5.12,
compare lane 1 and 2).
The bands corresponding to the position of FIH‐1 in Figure 5.12 were excised and sent
for MS. The gel slices were subjected to in‐gel tryptic digestion, eluted and were
analysed by matrix‐assisted laser desorption ionisation‐(MALDI)‐MS to determine the
identities of the major proteins present. The MS analysis however showed that FIH‐1
was not the main protein present in any of these bands. The bands in Lanes 5 and 7
were found to contain FIH‐1, however FIH‐1 was not the most abundant protein in
240
either sample, and Mascot searches based on peptide mass fingerprints assigned FIH‐1
with relatively low scores of 62 and 94 for FIH‐1 in samples 1 and 2 (lanes 5 and 7),
respectively. No FIH‐1 peptides were detected in the band in lane 9. A score of 67 or
greater is deemed to be significant by the Mascot search algorithm, and there were
numerous other proteins with much higher scores and hence likely greater abundance
in each of the samples. FIH‐1 was therefore being purified by this method and was
present in the bands, but it was not the major component of the bands.
As such, there was not sufficient amounts of FIH‐1 present in the purified samples to
enable detailed characterisation and subsequent identification of any modifications.
Despite excellent depletion of Myc‐6His‐FIH‐1 observed in using the batch purification
method, yields of FIH‐1 were still too low and FIH‐1 not pure enough to facilitate
detailed characterisation by MS analysis. The batch purification method also
compromised the purity of the eluted protein. This could perhaps be improved in future
work by applying lysate to the nickel resin in batch format, and then transferring the
resin to a column for washing with low imidazole buffers prior to elution.
5.5.4 Further optimisation of Ni2+‐affinity purification
A further purification was performed with research assistant Colleen Bindloss, who had
extensive experience with purifying the dioxin receptor using large‐scale denaturing
Ni2+‐affinity chromatography. Lysate was prepared from ten 175 cm2 flasks as described
above, with the modification of no imidazole in the guanidine lysis buffer. Purifications
proceeded as per the usual protocol, with the exception that the concentration of
imidazole in the guanidine wash buffer was reduced 10‐fold from 50 mM to 5 mM in an
attempt at improving Myc‐6His‐FIH‐1 yields. Bound protein was eluted in 2 ml fractions,
with samples from each fraction run on SDS‐PAGE gels for coomassie stain and anti‐Myc
immunoblot assessment of Myc‐6His‐FIH‐1 content (Figure 5.13). The results of this
F ti 2 3 4
10μ L Elution
75
100
150
Fraction: 2 3 4 kDa
37
50
Anti‐Myc WB
Figure 5.13 Ni2+‐Affinity purification ofTen 175cm2 flasks of stable cell line nulysed in guanidine lysis buffer with no imby ultracentrifugation and filtered. Mycaffinity chromatography, and washes py g p y, pof 5 mM. Bound protein was eluted inimmunoblot was used to assess whiindicated volumes of fraction 3 were seeither stained using coomassie or traElution fractions 2 and 3 were pooled, f
2 3 4
40μ L
75
100
150
2 3 4 kDa
37
50
Coomassie
Myc‐6His‐FIH‐1
f Myc‐6His‐FIH‐1umber 9 were grown to ~90% confluency and thenmidazole. Lysates were reduced, alkylated, clarifiedc‐6His‐FIH‐1 was purified from the lysate by nickelperformed with a reduced imidazole concentrationp2 ml fractions using 250 mM imidazole. Anti‐Mycich fractions contained Myc‐6His‐FIH‐1, and theeparated by SDS‐PAGE and the gels cut in half andansferred to nitrocellulose for anti‐Myc blotting.frozen, and sent to QIMR for MS.
243
purification revealed that, despite eliminating imidazole from the lysis buffer and
reducing the amount of imidazole in the wash buffer, the purity of each fraction was
improved, with less non‐specific background bands observed in the coomassie‐stained
gel. Western blots showed that only fractions 2 and 3 contained FIH‐1 protein, with
fraction 3 containing the most FIH‐1. Fraction 3 also contains a distinct band at the
expected size for Myc‐6His‐FIH‐1.
5.5.5 Mass spectrometry Results
Fractions 2 and 3 were sent to the QIMR and fraction 3 was analysed by both liquid
chromatography (LC)‐MALDI‐ MS and also LC‐electrospray ionisation (ESI)‐LTQ‐Orbitrap
MS. Using each technique, 50% and 71% of FIH‐1 was covered, respectively, with a
combined total of 73% sequence coverage. Tryptic peptides covered are represented in
Table 5.1 and Figure 5.14. There was evidence of PTMs, with four asparaginyl
deamidation events identified on Asn58, Asn87, Asn110 and Asn151 and four
occurrences of methionine oxidation at Met108, Met160, Met325 and Met 343 of
human FIH‐1 (Figure 5.14). These modifications can occur in vivo but also in samples
during handling.
No phosphorylation was detected, however this does not preclude the existence of
phosphorylation events on FIH‐1. Complete sequence coverage was not achieved, so
modified residues may have occurred in regions of the protein not covered.
Modifications may have additionally been difficult to detect due to a low abundance of
modified versus unmodified FIH‐1 in the sample. The low amounts of FIH‐1 in the
samples may have interfered with the detection of modifications that occurred with a
low stoichiometry, as weak signals would have been difficult to discern in samples
where there was a low signal‐to‐noise ratio.
Tryptic SequenceTryptic Peptide
Sequence
P1 mgeqk
P2 liseedlnhhhhhhamaat
P3 eeagalgpawdesqlr
P4 sysfptrpipr
P5 l dP5 lsqsdpr
P6 aeelieneepvvltdtnlv
P7 wdleylqenigngdfsvys
P8 flyydek
P9 k
P10 manfqnfkpr
P11 snr
P12 eemk
P13 fhefvek
P14 lqdiqqr
P15 ggeer
P16 l l tl dtP16 lylqqtlndtvgr
P17 k
P18 ivmdflgfnwnwink
P19 qqgk
P20 r
P21 gwgqltsnllligmegnvt
P22 gyk
P23 r
P24 cilfppdqfeclypypvhh
P25 qsqvdfdnpdyer
P26 fpnfqnvvgyetvvgpgdv
P27 t kP27 gaptpk
P28 r
P29 ieyplk
P30 ahqk
P31 vaimr
P32 niek
P33 mlgealgnpqevgpllntm
P34 gr
P35 yn
Table 5.1 Tryptic peptides of Myc‐6His‐Shaded boxes indicate sequences cover
CoveredCovered by MS
-
taaeavasgsgepr +
+
+
-
vypalk +
asthk -
+
-
+
-
-
+
+
-
++
-
-
-
-
tpahydeqqnffaqik +
-
-
hpcdr +
+
vlyipmywwhhiesllnggititvnfwyk -
-
-
+
-
-
-
mik -
-
-
‐FIH‐1 covered by MSred by MS
Figure 5.14 Schematic representation oof MS coverage of FIH‐1 and PTMs identified
249
5.5.6 Continuation of FIH‐1 purification
Due to the significant amount of time dedicated to the purification of FIH‐1 and the
limited time available towards the end of the body of work described in this thesis,
purification of FIH‐1 was continued by Michael Nastasie as part of an Honours (4th year
undergraduate) project. This work involved continuation of the Ni2+‐affinity
chromatography method of purifying tagged FIH‐1 from the #9 stable Myc‐6His‐FIH‐1
overexpressing cell line previously generated, and optimisation included altering the
both the amount and concentration of input lysate in an attempt to increase the
efficiency of the purification. Interestingly, in keeping with the results obtained here,
Michael Nastasie found that increasing total input lysate did not deliver greater yields of
purified Myc‐6His‐FIH‐1 (Nastasie, 2009). Comparison of eluted protein from
purifications performed from twelve 175 cm2 flasks and thirty six 175 cm2 flasks did not
reveal any increase in recovered FIH‐1. The concentration of the lysate however was
shown to be significant in determining final yields of FIH‐1, as lysing cells in smaller
volumes of lysis buffer (7.5 ml compared with 15 ml per 175 cm2 flask) was shown to
increase the amount of purified Myc‐6His‐FIH‐1, even when using lysate prepared from
only twelve 175 cm2 flasks. Despite these adjustments to the purification protocol, the
yields of FIH‐1 were not able to be significantly improved, and no further MS data was
able to be obtained.
5.5.7 Summary of Ni2+‐affinity chromatography
The work performed thus far included two methods used to purify endogenous FIH‐1
from HeLa Lysate, generation of stable cell lines overexpressing tagged FIH‐1 and
optimisation of large scale Ni2+‐purifications of the tagged FIH‐1. Using the latter
methods, bands of the expected size for Myc‐6His‐FIH‐1 were able to be detected on
coomassie‐stained gels. Optimisation of column purification led to MS data from tryptic
250
peptides covering a total of 73% of the protein, and showing events of deamidation and
oxidation. In order to try to increase the amount of FIH‐1 available for analysis, batch
purification was performed. Despite the observation of clear bands at the correct size
on a gel, subsequent MS analysis revealed that this technique reduced the purity of
purified FIH‐1, as FIH‐1 protein was present in these bands, but at a low level compared
to other more highly abundant proteins.
5.7 Summary and Discussion
The aim of the work presented in this chapter was to purify a sufficient amount of FIH‐1
from cells for identification of the any FIH‐1 modifications by MS. To this end, a variety
of methods were attempted at purifying endogenous FIH‐1 from HeLa cells.
Immunoprecipitation of endogenous FIH‐1 did not yield sufficient amounts of protein
due to poor efficiency of FIH‐1 binding by the three polyclonal antibodies used.
Difficulties were encountered when the IPs were modestly scaled up, as increasing the
amount of input lysate correspondingly increased the amount of background staining.
Further attempts at immunoprecipitating FIH‐1 from cells could include use of an
alternative anti‐FIH‐1 monoclonal antibody that has since become commercially
available (Santa Cruz sc‐271780). Alternatively, the anti‐Myc monoclonal antibody 9E10
could be used for IP of Myc‐6His‐FIH‐1 from stably overexpressing cells. Use of
monoclonal antibodies rather than the polyclonal antibodies used here may improve
the specificity of binding and may increase the yield of FIH‐1 obtained. In addition to IPs,
a novel method exploiting the high affinity FIH‐1 Notch1 Ank1‐4.5 interaction was
devised for the purification of FIH‐1 from cells. The efficiency of the Notch Ank1‐4.5‐
affinity method was excellent, however the difficulties encountered with elution of FIH‐
1 precluded successful use and continuation of this method.
251
Monoclonal cell lines overexpressing a tagged fusion protein were generated for larger
scale Ni2+‐affinity purifications. Overall, almost twenty large‐scale FIH‐1 purifications
were performed using Myc‐6His‐FIH‐1‐overexpressing cell lines by large scale
denaturing Ni2+‐affinity chromatography. Significant amounts of Myc‐6His‐FIH‐1 were
difficult to purify using this method. Optimisation steps included increasing the amount
of input lysate, lysing in a smaller volume of lysis buffer to reduce the volume of the
lysate, altering the stringency of washes, altering the elution conditions, and switching
to batch purification instead of column purification. Purified Myc‐6His‐FIH‐1 was
analysed by MS/MS, with the best results achieving 73% FIH‐1 sequence coverage FIH‐1.
PTMs identified included methionine oxidation and asparaginyl deamidation.
5.7.1 Methionine oxidation
Four events of methionine oxidation on FIH‐1 were identified, occurring on Met108,
Met160, Met325 and Met343. Methionine oxidation involves the ROS‐mediated
oxidation of the sulphur atom in the methionine side chain to methionine sulfoxide. As
oxidation of methionine residues does not result in a change in a proteins overall
charge, this modification would not have contributed to the pI variation of FIH‐1
observed in 2‐D gels. Furthermore, methionine oxidation is a widely known artefact of
sample handling and is often encountered during the processing of MALDI‐TOF MS
samples (Potgieter et al., 1997). Based on the work described in this thesis, it is
unknown if this modification occurred in cells or during Myc‐6His‐FIH‐1 purification or
subsequent MS analysis.
5.7.2 Asparaginyl deamidation
Four deamidated asparaginyl residues, Asn58, Asn87, Asn110 and Asn151, were also
identified by MS analyses of purified Myc‐6His‐hFIH‐1. Deamidation of proteins results
252
in a series of spots with increasingly acidic pIs concurrent with an increasing number of
deamidated asparaginyl residues. Each spot may represent a mixture of proteins with
deamidated residues at different positions, but with the same overall number of
deamidated asparagines, conferring the same overall charge to the protein. Thus,
deamidation could be responsible for some of the isoelectric heterogeneity observed in
2‐DE experiments, as the addition of negative charges is in keeping with the series of
increasingly acidic spots seen for FIH‐1 in 2‐DE blots. However, like methionine
oxidation, asparagine deamidation can occur both in cells and during sample
preparation. Because deamidation is a known, common occurrence in both 2‐DE and
MS (Krokhin et al., 2006), it is unclear whether these events are physiologically relevant
for FIH‐1 or are merely artefacts of sample processing.
Deamidation can occur both enzymatically and non‐enzymatically, though the
commonly held view is that most deamidation events occur nonenzymatically, both in
vivo and in vitro. Only a small number of mammalian deamidases have been discovered
and these have a limited substrate range. Examples include nicotinamide deamidase
(Petrack et al., 1965) and Protein NH2‐terminal Asparagine Deamidase (PNAD) (Stewart
et al., 1994). PNAD only catalyses deamidation of N‐terminal asparagines and has no
observed activity with interior asparaginyl residues, while nicotinamide deamidase has
no demonstrated protein substrates at all.
Nonenzymatic deamidation is thought to occur when the nitrogen atom that forms the
peptide bond between an asparaginyl residue and the following residue, attacks the
carbonyl carbon atom of the asparagine, forming an intermediate succinimide ring with
the loss of the amide group. This intermediate ring structure is then cleaved at one of
two positions, giving either aspartic acid or isoaspartic acid (Capasso et al., 1989). As
such, for deamidation to proceed, the asparaginyl residue must be located within a
flexible region of a protein to enable initial attack and the subsequent formation of the
253
intermediate succinimide ring. Protein structure is a significant contributor to
deamidation rate, with flexible regions in proteins being more amenable to
deamidation. Asparaginyl residues that reside within highly structured regions of
proteins are not as readily deamidated (Rivers et al., 2008).
The positions of the identified deamidated asparaginyl residues in FIH‐1 were examined
to assist in speculation about the likelihood of each deamidation event occurring in cells
versus in denatured protein samples. As Asn87, Asn110 and Asn151 occur in
unstructured loops on the surface of FIH‐1 (Figure 5.15), it is conceivable that these
residues may be subjected to deamidation in the context of intact, native FIH‐1. Asn58,
on the other hand, is located within an α‐helix. α‐helical structure has been reported to
inhibit deamidation of asparaginyl residues that reside within a helix, therefore Asn58 is
less likely to undergo deamidation within native FIH‐1. As the Myc‐6His‐FIH‐1
purifications were carried out in denaturing conditions, it may be that deamidation of
Asn58 occurred in the sample following protein denaturation. In support of this, it has
been reported that while the half‐life for deamidation is relatively long for asparaginyl
residues within structured parts of proteins, both denaturation and tryptic digestion of
proteins prior to MS dramatically accelerates the rate of deamidation through loss of
structure that may have provided steric hindrance to deamidation, enabling the
reaction to proceed (Rivers et al., 2008). The loss of FIH‐1 structure by denaturation
during cell lysis and then tryptic digestion prior to MS may have contributed to the
deamidation events identified by MS, particularly in the case of Asn58 that lies within a
structured part of the protein.
The amino acid sequence flanking the asparaginyl residue also impacts on deamidation
rates. It has been reported that side of the amino acid residue immediately C‐terminal
to the asparaginyl residue has a greater effect on deamidation rate that the N‐terminal
residue side chain (Robinson and Robinson, 2001a). The presence of small residues
Figure 5.15Structure of human FIH‐1 retrieved from(MMDB ID 33168) and displayed using Chttp://www.ncbi.nlm.nih.gov/structure
m the Molecular Modelling Database (MMDB) Cn3D 4.1 program accessed on . Deamidated Asn residues are indicated.
257
immediately C‐terminal to an asparagine increase the rate of in vitro deamidation of
that asparagine in peptides, with asparaginyl residues within Asn‐Gly sequences the
most readily deamidated, particularly when they reside in an unstructured region of a
protein. Asparaginyl residues that lie adjacent to bulky, hydrophobic residues are least
likely to be deamidated, probably due to steric interference of the adjacent side chain in
the access of the backbone nitrogen to the carbonyl carbon.
The effect of neighbouring amino acid residues on rates of asparaginyl residues in
peptides has experimentally determined (Robinson and Robinson, 2001a). Based on
these studies, predictions can be made regarding the rates of deamidation of
asparagine residues in FIH‐1. Asn87 is flanked by two glycine residues, and is thus likely
to be readily deamidated due to the lack of steric interference of the small flanking
residues on the deamidation reaction. Aspartic acid, which follows Asn151, has been
determined to have moderate effects on asparaginyl deamidation, and the presence of
a glutamic acid residue following Asn58 predicted to have greater inhibitory effects. The
phenylalanine residue C‐terminal to Asn110 is likely to cause the greatest interference
to asparaginyl deamidation. Taken together, both the location of Asn87 in a flexible
region on the surface of FIH‐1 and the presence of two glycine residues immediately
adjacent to Asn87, suggest that this residue may be deamidated in cells.
Estimations of non‐enzymatic deamidation half‐lives of all asparaginyl residues in FIH‐1
were calculated using an algorithm developed by N. Robinson (Robinson, 2002), freely
available at http://deamidation.entrewave.com, and a published crystal structure of
FIH‐1 (Dann et al., 2002)(Protein Data Bank ID code 1MZE; Figure 5.15). This algorithm
is based on the extensive studies investigating the deamidation rates of asparaginyl
residues in pentapeptides with different combinations of neighbouring amino acids
(Robinson and Robinson, 2001a), and also of the rates of deamidation as they occur in
proteins with known structure (Robinson and Robinson, 2001b). The estimated rates
258
apply for proteins in a pH 7.4, 37°C, 0.15 M Tris buffer and deamidation half‐life is given
by (100) x deamidation coefficient (CD), in days. There is precedent in the literature for
the accurate prediction of this algorithm for biologically relevant deamidated
asparaginyl residues (Deverman et al., 2002). Using this algorithm, Asn87 of FIH‐1 has
the shortest predicted half‐life of all asparaginyl residues (Table 5.2) and is therefore
predicted to be most readily deamidated.
Taken together, the identification of deamidation events on Asn87 by MS, the position
of this residue in a flexible, unstructured region on the surface of FIH‐1, and the short
estimated half‐life given by the deamidation prediction algorithm, suggest that it is
possible that this deamidation event occurs in cells and is not simply an artefact of
sample processing procedures. From the data presented in this thesis however, it
cannot be determined whether these deamidation events occurred in vivo prior to 2‐DE,
or prior to MS, or in vitro as a result of sample processing.
5.7.3 Phosphorylation
In addition to deamidation, the different pIs of proteins observed in a spot “train” could
be due to an increasing number of phosphorylation events, which also introduce a
negative charge to a protein. The addition of phosphate groups onto a protein causes
an acidic shift in pI, with each subsequent phosphate group adding further negative
charges, resulting in a “train” of spots. The fact that phosphorylation was not detected
in by MS does not preclude the possibility of this modification contributing to the
differences in pI of FIH‐1, as the complete FIH‐1 sequence was not covered and there
was not a sufficient amount of the protein to enable detection of modifications that
may have occurred at low frequency. Modifications occurring at low levels can also be
difficult to detect by MS, where detection of modified protein peaks may be
confounded by highly abundant peaks attributed to unmodified protein. To address this
Position Sequence
58 Glu‐Asn‐Glu 8
68 Thr‐Asn‐Leu 2
84 Glu‐Asn‐Ile 4
87 Gly‐Asn‐Gly 0
110 Ala‐Asn‐Phe 1
113 Gln‐Asn‐Phe 2
119 S A A 2119 Ser‐Asn‐Arg 2
151 Leu‐Asn‐Asp 3
166 Phe‐Asn‐Trp 1
168 Trp‐Asn‐Trp 5
171 Ile‐Asn‐Lys 2y
185 Ser‐Asn‐Leu 1
194 Gly‐Asn‐Val 1
205 Gln‐Asn‐Phe 4
246 Asp‐Asn‐Pro 1
254 Pro‐Asn‐Phe 1
257 Gln‐Asn‐Val 1
286 Leu‐Asn‐Gly 1
294 Val‐Asn‐Phe 2
321 A A Il 9321 Arg‐Asn‐Ile 9
332 Gly‐Asn‐Pro 4
341 Leu‐Asn‐Thr 1
349 Tyr‐Asn 5
Table 5.2 Prediction of Asn half life for Prediction of deamidation half life baseavailable at http://deamidation.entrew.
CD MS Coverage? Deamidation?
2.778 YES YES
98.303 YES
01.032 YES
.337 YES YES
01.849 YES YES
8.261 YES
0 618 NO0.618 NO
.086 YES YES
19.750 YES
6.042 YES
02.339 YES
65.556 YES
77.058 YES
3.397 YES
94.618 YES
89.426 NO
55.994 NO
1.118 NO
7.647 NO
77 133 NO77.133 NO
06.278 YES
14.842 YES
.635 NO
FIH‐1ed on an algorithm developed by Robinson (2002) and ave.com. See Section 5.7.2 for further discussion.
261
possibility in terms of phosphorylation, enrichment of phospho‐peptides could be
carried out prior to MS to increase the ratio of any phosphorylation FIH‐1 in the samples
to improve the likelihood of detecting phosphorylation.
5.7.4 Final summary of FIH‐1 purification and MS results
MS data indicated deamidation of asparaginyl residues and methionine oxidation as two
modifications that occurred on FIH‐1. Both deamidation and oxidation of proteins can
occur in cells or in samples during sample preparation procedures and it is not known if
these modifications occurred on FIH‐1 in cells or occurred during processing of the
samples.
Ultimately, the numerous strategies employed for the purification of both endogenous
FIH‐1 and Myc‐6His‐FIH‐1 did not deliver adequate amounts of protein for in‐depth MS
analysis that would have enabled comprehensive mapping of PTMs. More detailed
characterisation of PTMs would require greater amounts of purified FIH‐1 and may
require further optimisation of the Ni2+‐affinity purification method, such as scaling up
the input whilst simultaneously lysing in a smaller volume of lysis buffer. Continuation
of this work could also include using an alternative, more highly overexpressing Myc‐
6His‐FIH‐1 stable cell line from which to purify FIH‐1. In assessing the different
expression levels of FIH‐1 in the various stable monoclonal cell lines created, there was
a significant difference in expression levels between some of the cell lines. The #9 stable
cell line was chosen as it overexpressed FIH‐1 at levels moderately higher than
endogenous FIH‐1. There were, however, other monoclonal cell lines, for example
clonal line number 6, which expressed higher levels of Myc‐6His‐FIH‐1 (see Figure 5.8).
Purifying FIH‐1 from a cell line that was shown to overexpress FIH‐1 at higher levels
would increase the amount of FIH‐1 available for purification and could perhaps
increase the final yield of FIH‐1 purified by Ni2+‐affinity chromatography. While this
262
strategy may enable higher yields of FIH‐1 to be obtained, it may not facilitate
identification of PTMs, as the proportion of modified FIH‐1 to unmodified FIH‐1 may be
decreased by saturating out cellular modification machinery.
Methionine oxidation does not affect any ionisable groups and is therefore not
expected to alter the pI of a protein, thus this modification would not have contributed
to the pI heterogeneity observed. Deamidation, on the other hand, introduces a
negative charge due to loss of an amide group and gain of a carboxyl group, and
deamidated proteins thus have a more acidic isoelectric point and can appear on 2‐D
maps as a spot train where each protein spot in the acidic direction is due to increasing
numbers of deamidated residues. As such, deamidation of FIH‐1 may contribute to the
spot profile observed in 2‐DE blots. Phosphorylation of proteins also involves the
addition of negative charges and a concurrent decrease in pI. Due to the incomplete
results obtained by purification and MS of FIH‐1, and due to the possibility that the
spots observed in 2‐D gels could be due to either deamidation or phosphorylation, a
more directed approach was undertaken to determine whether FIH‐1 could be
phosphorylated. Given the importance of phosphorylation in regulating protein activity
and its involvement in numerous signalling pathways, it was of interest to investigate
FIH‐1 was subjected to this important modification.
263
CHAPTER 6
INVESTIGATING
POTENTIAL
PHOSPHORYLATION
OF FIH‐1
265
6 INVESTIGATING POTENTIAL
PHOSPHORYLATION OF FIH‐1
6.1 Overview
The 2‐DE experiments presented in Chapter 4 demonstrated that there are multiple
forms of FIH‐1 in all samples analysed. Purification and MS analyses identified
asparaginyl deamidation and methionyl oxidation of FIH‐1. As complete coverage of
FIH‐1 was not achieved by these analyses, the possibility of PTMs occurring elsewhere
on FIH‐1 remained.
The spot profile for FIH‐1 was suggestive of step‐wise decreases in pI. Such pI effects
can be mediated by PTMs that affect the charge of a protein by either introducing
negative charges or removing positive charges. Accordingly, deamidation of FIH‐1 is in
keeping with the spot pattern observed. However, as deamidation is a widely known
artefact of both 2‐DE and MS, it is not known if this PTM is physiologically relevant for
FIH‐1. Phosphorylation of proteins also results in decreased pI and could therefore also
be contributing to the differences in FIH‐1 pI.
Covalent attachment and removal of phosphate groups on proteins represents the best‐
characterised and most widely recognised PTM that regulates protein behaviour.
Phosphorylation of proteins is a highly regulated process that has known effects on the
activities of many cellular proteins. There are over 500 known mammalian kinases, and
over 150 known phosphatases (Cohen, 2002). These enzymes together act to regulate
what is termed the “phosphoproteome”, the entire cellular complement of
266
phosphoproteins, including the phosphorylation status of each protein at each site in
response to various stimuli and signalling events.
Anecdotal evidence links FIH‐1 to proteins involved in kinase signalling. Firstly, Integrin‐
linked kinase‐1 (ILK‐1) has been suggested to be a substrate of FIH‐1. Sequence
alignment of the ARs of genuine FIH‐1 substrates and putative substrates suggested that
ILK‐1 could be hydroxylated by FIH‐1. Assessment of hydroxylation in vitro showed that
ILK‐1 was a weak substrate (Cockman et al., 2006). This does not preclude a direct
interaction between FIH‐1 and ILK‐1 however, as Notch4 has been found to interact
strongly with FIH‐1 and not be hydroxylated (Wilkins et al., 2009). No further data has
been published showing whether FIH‐1 and ILK‐1 directly interact, but given that ILK‐1
contains an ARD it would be interesting to see if they co‐IP. Secondly, unpublished data
suggests that FIH‐1 interacts with Protein Phosphatase 1 Regulatory subunit 12C
(PP1R12C). Rachel Hampton‐Smith conducted a Y2H assay using FIH‐1 as bait to isolate
FIH‐1 interacting proteins. A number of proteins were identified as FIH‐1 interactors,
many containing ARDs. Some, such as IκBα, have since been corroborated in the
literature (Cockman et al., 2006). PP1R12C contains four ankyrin repeats and was one of
the proteins pulled out by the Y2H assay, and subsequent work showed that
overexpression of PP1R12C in 293T cells increased expression of a Gal‐HRE reporter
gene, indirectly inferring an interaction with FIH‐1 reduces FIH‐1‐mediated HIF‐CAD
repression (R. Hampton‐Smith, unpublished data).
Also, in addition to the regulation of HIF transcriptional activity by oxygen levels via FIH‐
1, there is evidence that MAPK and PI3K/Akt kinase pathways can influence the
transcriptional activity of HIF, as discussed in Section 1.5.7 of this thesis. In many cases,
the direct mode of regulation of these signalling pathways has not been elucidated.
267
Two different experimental approaches were undertaken to investigate the possible
phosphorylation of FIH‐1. First, phosphatase treatment of cell lysate was performed to
enable a comparison of spot profile between phosphatase treated and untreated cell
lysates by 2‐DE. Second, in vitro phosphorylation assays were performed, involving the
incubation of bacterially expressed and purified FIH‐1 with lysate in the presence of [ϒ‐
32P]‐ATP.
6.2 Phosphatase Treatments
6.2.1 Phosphatase treatments
To ascertain whether phosphorylation might be responsible for one or more of the
different spots observed for FIH‐1, lysates were made from the Myc‐6His‐FIH‐1
overexpressing stable cell line #9 for treatment with lambda phosphatase.
Overexpressed FIH‐1 was investigated simply due to the superior detection of Myc‐
tagged proteins with the 9E10 over detection of endogenous FIH‐1 with any of the FIH‐
1‐specific antibodies. The broad spectrum protein phosphatase lambda used to
dephosphorylate proteins in the lysate, and phosphatase inhibitors included in the
control samples to preserve any phosphorylation on FIH‐1.
To control for successful phosphatase treatment, small aliquots of both phosphatase
treated and untreated lysates were taken and run on a gel for western blotting with an
antibody targeted to phosphorylated Ser5 of RNA polymerase II. Phosphatase treatment
resulted in loss of the approximately 250 kDa band detected by this antibody, and this
positive control was performed for all experiments
Treated and untreated lysates were separated by 2‐DE and results visualised by blotting
with anti‐Myc antibodies (Figure 6.1 b). The results of a representative experiment are
b.a.
‐ +S
PPase
kDa
100
75
150
250 pSer‐RNA Pol II
50
100
75
50
WB: pSer‐RNA Pol II
50
37
100
75
50
37
75
37
Figure 6 1 Lambda Phosphatase TreatmFigure 6. 1 Lambda Phosphatase TreatmCell lysates were prepared from stable Min the presence or absence of phosphator 2.5 μl of buffer was added to + PPa200 μg total protein in a volume of 60 μμg removed for the Phospho‐Ser RNA Pand resuspended in IEF buffer, and 100 p ,Results were visualised by anti‐Myc WBexperiments.
4 7pI
PPa
0
0
5
‐ PPase
0
7
0
5
+ PPase
0
7
5
7
Anti‐Myc WB
mentment Myc‐6His‐FIH‐1 overexpressing cell line #9 either tase inhibitors. 1000 units of lambda phosphatase ase and PPase samples, respectively, containing μl. Reactions occurred over an hour at 30°C, and 20 Pol II WB (a). Remaining samples were precipitated μg separated by 2‐DE using 7 cm pH 4‐7 IPG strips. μg p y g p p
B (b). Representative of > 3 independent
271
presented in Figure 6.1. Figure 6.1 a shows successful loss by phosphatase treatment of
the band detected by the phospho‐specific antibody, and Figure 6.1 b shows the spot
pattern of phosphatase treated and untreated lysates as detected by the anti‐Myc
antibody. Some obvious differences were observed in the spot profile between the
untreated and treated lysates. In the blot of untreated lysates (top panel), a train of four
spots of the same molecular weight but different pIs was apparent, however with
phosphatase treatment (lower panel), only two spots with different pIs were observed,
and there was also a change in the molecular weight of the most basic spot, with this
most basic spot possibly being two spots close in size. The reduction in the number of
spots resolved on the 2‐D gels from four spots to two with phosphatase treatment
suggests that phosphatase treatment caused a loss of two modifications that led to
changes in the pI of Myc‐6His‐FIH‐1.
Loss of phosphate groups and the corresponding negative charges from a protein
results in a more basic pI. Such a change would be evident on a 2‐D gel as a shift of
protein spots to a more basic position. Indeed, from these data it does appear as
though the two most acidic spots in the untreated 2‐D blot are the ones lost in the
phosphatase‐treated blot (Figure 6.1 b). However, despite great care being taken during
application of IPG strips to the second dimension gels to ensure that the pH gradients of
the two gels would align, it is difficult to conclude with certainty from this experiment
whether the two spots observed following phosphatase treatment have migrated to a
more basic position than the spots present on the untreated blot, as there are no other
spots present on the gel to serve as a reference.
6.2.2 Phosphatase treatment with internal reference protein
To determine the nature of the shift in pI of Myc‐6His‐FIH following phosphatase
treatment, it was necessary to include an internal standard to provide a point of
272
comparison. A Myc‐tagged Notch construct encompassing the RAM domain was
selected as it would be able to be detected by the same anti‐Myc antibody. Myc‐6His‐
Notch 1753‐1859 is approximately 30.8 kDa and has a predicted isoelectric point of 5.21
(Protein Calculator v3.3, http://www.scripps.edu/~cdputnam/protcalc.html). As such it
was sufficiently different in size and pI from tagged FIH‐1 to enable distinct detection,
yet had a predicted pI within the pH 4‐7 range of the IPG strips that were used for these
experiments. The protein was expressed in and purified from bacteria. Purified protein
was assessed on Coomassie‐stained SDS‐PAGE gels (Figure 6.2 a) prior to precipitation
of the protein using the 2‐D Clean Up Kit and resuspension of protein pellets in IEF
buffer. Different amounts of the internal standard protein were added to 20 μg of #9
stable cell lysate and run on a gel and blotted with the anti‐Myc 9E10 antibody in order
to gauge the relative strength of bands detected for purified Myc‐Notch‐Ram and Myc‐
6His‐FIH‐1 from lysate (Figure 6.2 b). From this experiment, it was decided that less than
0.5 μg of purified RAM domain would be sufficient to serve as an internal standard.
Phosphatase‐treated and untreated lysates were prepared for 2‐DE as described above,
and 0.25 μg of the purified Notch‐RAM construct was added to the samples
immediately prior to IEF. The anti‐Myc blots of the resulting 2‐D gels presented in Figure
6.3 b show an excess of internal standard protein. Despite this, excellent resolution of
spots was achieved for the standard protein and also for Myc‐tagged FIH‐1, enabling a
direct comparison of the position of FIH‐1 spots between phosphatase‐treated and
untreated blots relative to the well‐resolved spots of the standard.
The spots marked on Figure 6.3 b as ‘1’, ‘2’ and ‘3’ resolved well in both 2‐D gels and
were thus used as the internal reference for this experiment. Aligning the two blots
according to these spots clearly shows a shift in the position of the Myc tagged‐FIH‐1
spots to a more basic position. The dashed line superimposed over the blots shows that,
in the upper, untreated panel, there are three clearly resolved spots on the left of the
MWkDa
Purified Myc‐6His‐Notch
1l 2l 5l 10
a.
50
37
25
Purified Myc
Coomassie St
25
b.
50
0.1 0.5 1 2 MWkDa
Purified Myc
37
Figure 6.2 Purification and anti‐Myc imreference protein
Myc‐6His‐Notch‐RAM was bacteriallchromatography. Purity of the proteing p y y pVarying amounts (between 0.1 μg anadded to 20 μg of lysate from the Mywere separated by SDS‐PAGE and result
h‐RAM
0l
Myc‐6His‐Notch‐RAM
‐6His‐Notch‐RAM
tain
Myc‐6His‐FIH‐1
2.5 0.5 1 2 2.5 g purified RAM
+ 20 g Stable #9 lysate
‐6His‐Notch‐RAM
WB: anti‐Myc
Myc‐6His‐Notch‐RAM
mmunoblot of the Myc‐6His‐Notch‐RAM
ly expressed and purified by nickel‐affinityn was assessed by SDS‐PAGE and coomassie (a).y ( )nd 2.5 μg) of purified Myc‐6His‐Notch‐RAM wasyc‐6His‐FIH‐1 overexpressing cell line #9. Proteinsts visualised by anti‐Myc WB (b).
+
PPase
a.
100
150
250
‐ +
50
100
75
37
WB pSer RNA PWB: pSer‐RNA P
pSer‐RNA Pol II
ol IIol II
4
50
b.
kDa
37
50
37
Figure 6. 3 Phosphatase treatment andCell lysates were prepared from stable Mthe presence or absence of phosphatasthe presence or absence of phosphatas2.5 μl of buffer was added to + PPaseμg total protein in a volume of 100 μl. Rremoved for the Phospho‐Ser RNA Poand resuspended in IEF buffer, and 0.5μg separated by 2‐DE using 7cm pH4‐7(b) . Representative of 3 independent ex
7pI
‐ PPase
Myc‐6His‐FIH‐1
Myc‐6His‐FIH‐1
12 3
d 2‐DE of HeLa cell lysatesMyc‐6His‐FIH‐1 overexpressing cell line #9 either inse inhibitors 1000 units of lambda phosphatase or
+ PPase12 3
se inhibitors. 1000 units of lambda phosphatase ore and PPase samples, respectively, containing 200Reactions occurred over an hour at 30°C, and 40 μgl II WB (a). Remaining samples were precipitatedμg of Myc‐6His Notch‐RAM added prior to IEF. 100IPG strips. Results were visualised by anti‐Myc WBxperiments.
279
line, with the most basic spot being the weakest and the central spot showing the
greatest intensity. In the lower panel, showing the phosphatase‐treated samples, the
position of the spots has shifted such that there is now one more basic spot to the right
of the dashed line, indicating loss of a phosphate group. Additionally, the intensity of
the spots has changed such that the spot immediately to the left of the dashed line is
the most intense in the phosphatase‐treated blot, indicating an increase in the
abundance of the form of FIH‐1 at this pI. Overall, phosphatase treatment here results
in the appearance of a more basic spot, and a shift in intensity of spots in a basic
direction.
0.1 μg of the purified Notch‐RAM internal standard protein was included in a further
experiment where lambda phosphatase treated lysates were compared with control
lysates by 2‐DE (Figure 6.4). Comparison of the position of Myc‐His‐FIH‐1 in
phosphatase treated and control 2‐D blots relative to the internal standard again shows
a decrease in the number of spots with different pIs from four spots to two. This
indicates that phosphatase treatment of lysates reduced the pI heterogeneity of FIH‐1
(Figure 6.4). To better visualise a shift in pI of phosphatase treated protein, a dashed
line has been superimposed over an expanded image of each blot. With phosphatase
treatment (lower panel), there is again the appearance of a more basic spot to the right
of the dashed line, and a loss of the more acidic spots seen in the untreated blot,
suggesting that the pI of FIH‐1 becomes more basic upon phosphatase treatment. This
basic shift in pI upon phosphatase treatment is consistent with the phosphorylation of
FIH‐1 in HeLa cells. Additionally, phosphatase treatment was observed to alter the
migration of FIH‐1 in the second dimension (Figure 6.4 and also Figure 6.1), consistent
with an alteration in charge of the protein upon phosphatase treatment.
‐ P
4 pH
50
37
kDa
50
37
+ PWB: anti‐Myc
Figure 6. 4 Phosphatase treatment andCell lysates were prepared from stable(1000 units) of lambda phosphatase orPPase samples, respectively, containinReactions occurred over an hour aReactions occurred over an hour aresuspended in IEF buffer and 0.1 μg oseparated by 2‐DE using 7 cm pH 4‐7 IPpH 4‐7 IPG strips Results were visindependent experiments.
PPase
7 ‐ PPase
Myc‐6His FIH‐1
+ PPase
PPase
Myc‐6His FIH‐1
d 2‐DE of HeLa cell lysatesMyc‐6His‐FIH‐1 overexpressing cell line #9. 2.5 μlr 2.5 μl of buffer was added to + PPase and ‐ ng 200 μg total protein in a volume of 100 μl.at 30°C Samples were then precipitated andat 30 C. Samples were then precipitated andof Myc‐6His Notch‐RAM added prior to IEF. 100 μgPG strips. and 100 μg separated by 2‐DE using 7cmsualised by anti‐Myc WB. Representative of 3
283
6.2.3 Summary of phosphatase treatments
In summary, phosphatase treatments of HeLa cell lysate have indicated that lambda
phosphatase causes Myc‐6His‐FIH‐1 protein to shift to a more basic region of the pH
gradient compared with untreated FIH‐1. This is suggestive of loss of one or more
phosphate groups from the protein and is consistent with phosphorylation of FIH‐1 in
HeLa cells.
6.3 Phosphorylation Assay
6.3.1 Strategy for in vitro phosphorylation assay
Metabolic labelling of proteins in cells with a 32P is commonly carried out to provide
evidence that a protein in phosphorylated. This typically involves the incubation of cells
in phosphate‐free media supplemented with [ϒ‐32P]‐ATP to enable incorporation of the
32P isotope into any phosphate moieties that are attached to proteins. Following
labelling, cells are lysed and the protein of interest is immunoprecipitated from the
lysate for separation on SDS‐PAGE, immunoblotting and autoradiography. Due to the
inherent difficulties in IP of FIH‐1 described in Chapter 5 of this thesis, this approach
was not attempted for FIH‐1.
Instead, in order to provide evidence of phosphorylation, a phosphorylation assay was
devised where bacterially expressed and purified FIH‐1 could be phosphorylated in vitro
by application of cell lysate to FIH‐1 bound to nickel resin. Similar to a kinase assay, the
assay would be performed using [ϒ‐32P]‐ATP to radiolabel phosphorylated protein, and
results visualised by coomassie‐stained gel and autoradiography. Purified HIF‐1α‐CAD,
known to be phosphorylated on Thr796, was employed as a positive control. This
experiment was based on similar procedures used in the literature (Rybina et al., 1997).
284
6.3.2 In vitro phosphorylation of Trx‐6His‐FIH‐1
Trx‐6His‐hFIH‐1 and 6His‐hHIF‐1α‐CAD were separately expressed in BL21 E.coli cells.
Following cell lysis and centrifugation of lysates, FIH‐1 or the HIF‐1α‐CAD were bound to
nickel resin and the resin washed to remove unbound proteins. A 30 μl sample of slurry
was heated in the presence of SDS sample buffer to assess levels and purity of the
protein, and the remaining FIH‐1 was left attached to the resin and stored at 4°C in the
presence of sodium azide. In order to allow for a similar amount of each substrate
protein to be included in the assay, bound substrate proteins were quantified by elution
from a 10 μl sample of each resin and measuring the absorbance of eluted protein at
280 nm.
Lysate from HeLa cells was prepared using Native Lysis buffer (see Section 2.5 of
Materials and Methods Chapter), supplemented with phosphatase inhibitors. An ATP
solution was made up with a mixture of “cold” and “hot” [ϒ‐32P]‐ATP and added to 500
μl lysate (or to buffer only) such that the final concentration of ATP per sample was 150
μM and 50 μCi. To initiate reactions, the ATP solution was added to the lysate (or to the
buffer only for negative control samples), and the lysate/buffer added to Trx‐6His‐hFIH‐
1‐ or 6His‐hHIF‐1α‐CAD‐bound resin. Reactions were incubated at 37°C for 3 minutes.
The resin was then washed three times with 1 ml of native lysis buffer and FIH‐1 was
eluted by heating the resin in the presence of SDS sample buffer. Proteins were
separated by SDS‐PAGE and visualised by Coomassie stain, and the gels were then
sealed in plastic wrap to which hot ‘marker’ strips were affixed adjacent to gels. These
strips were simply small strips of filter paper to which 1 μl of a 1:100 000 dilution of the
[ϒ‐32P]‐ATP had been applied. Following exposure to phosphor storage screens, each gel
was carefully removed from the cassette to ensure that the position of the gel relative
to the markers remained constant, and the gels were scanned still wrapped in plastic
285
film with the markers attached. This image was used to accurately align the Coomassie
gel and autoradiograph images. The “hot” markers and plastic film were then removed
from the gel and the gel rescanned to generate a better image without the plastic film,
and this image was the one ultimately used in generation of a finalised figure.
Representative results are presented in Figure 6.5. This figure is presented in two parts,
with the upper panel (Figure 6.5 a) included to illustrate alignment of the coomassie gel
and the autoradiograph, and the lower panel (Figure 6.5 b) showing the final figure. This
process was carried out for each of the experiments, however the preparatory stages of
putting each figure together are shown only for this first experiment. This method
permitted confidence that the bands on the coomassie gels lined up precisely with
bands observed on the autoradiograph.
Comparison of Coomassie‐stained gel and the autoradiograph suggested that FIH‐1 was
phosphorylated by a kinase or kinases supplied by HeLa lysate (Figure 6.5 b). Trx‐6His‐
FIH‐1 has a molecular weight of approximately 57 kDa, and can be seen, as indicated, on
the Coomassie‐stained gel. A strong signal in the corresponding position of the
autoradiograph indicates the incorporation of 32P into a protein with molecular weight
similar to FIH‐1, and given the abundance of the FIH‐1 protein, is most likely
phosphorylation of FIH‐1. There is also significant phosphorylation of a background
band of approximately 30 kDa. The 6His‐HIF‐1α‐CAD positive control was also
phosphorylated in this assay. Addition of lambda phosphatase in a separate experiment
resulted in loss of the radioactive signal (Figure 6.6).
6.3.3 TEV cleavage of Trx‐6His‐FIH‐1
Due to the presence of a significant background band in the 6His‐HIF‐1α‐CAD lanes of a
similar size to Trx‐6His‐FIH‐1, and also due to the high levels of background proteins
1. Autoradiography 2. Sc
a.
“hot” marker strips*
HIF CADTrx‐6His FIH‐1
b.
37 kDa
50 kDa75 kDa
‐ + ‐ +
Coomassie
25 kDa
Figure 6.5 In vitro Phosphorylation AssRecombinant 6His‐HIF‐1α‐CAD and Tchromatography and left bound to resieach assay was estimated to be 20 μgreaction tubes, and native lysis bufferwere initiated by addition of the 150 μfor 3 minutes at 37°C. The resin was ththe resin in SDS sample buffer. SDS‐Pexposed by autoradiography. (a) descrthe results. Representative of > 3 indep
can Coomassie Gel 3. Overlay images
HIF CADTrx‐6His FIH‐1
Lysate
FIH‐1
‐ + ‐ +
e Autorad
HIF‐1‐CAD
sayTrx‐6His‐FIH‐1 were purified by nickel affinityin. The amount of substrate‐bound resin added toof protein. 600 μl of HeLa cell lysate was added tor only added to negative control tubes. ReactionsμM, 50 μCi ATP cocktail. Reactions were incubatedhen washed 3 times and proteins eluted by heatingPAGE gels were stained with coomassie and thenibes how each figure was made, and (b) presentspendent experiments.
+ + + +
HIF CAD FIH‐1
‐ + + ‐ + + + +
37 kDa
50 kDa
75 kDa
Coomas
25 kDa
Figure 6.6 In vitro Phosphorylation AssRecombinant 6His‐HIF‐1α‐CAD and Tchromatography and left bound to ressubstrate protein was put in a tube anonly added to negative control tubes. Ry g50 μCi ATP cocktail. Reactions were incwashed 3 times with native lysis buffeReactions were incubated for 1 hour atSDS sample buffer. Results were visualisby autoradiography. Representative of t
HIF CAD FIH‐1
Lysate‐ + + ‐ + + +
LysatePPase+
‐ + + ‐ + +
Trx‐6His‐FIH‐1
ssie Autorad
say with phosphatase treatmentTrx‐6His‐FIH‐1 were purified by nickel affinityin. Resin estimated to be bound to 20 μg of eachnd 500 μl of HeLa lysate added. Native lysis bufferReactions were initiated by addition of the 150 μM,y μ ,cubated for 3 minutes at 37°C. The resin was thener prior to addition of λ PPase buffer +/‐ λ PPase.30°C. Proteins were eluted by heating the resin insed by SDS‐PAGE and coomassie staining, followedtwo independent experiments.
291
observed in purified Trx‐6His‐FIH‐1, the proteins were again bacterially expressed and
purified, and this time the resin was washed with 35 mM imidazole to try to reduce
non‐specific proteins. Also, proteins in subsequent experiments were eluted by addition
of 500 mM imidazole rather than by heating in SDS sample buffer in an attempt to
increase purity of the eluted protein samples.
Subsequent phosphorylation assays indicated that the background band observed with
6His‐HIF‐1α‐CAD had indeed been successfully removed using this method, increasing
the confidence that the signal observed in FIH‐1 lanes was attributed to FIH‐1 and not a
background band (data not shown). However, background bands in FIH‐1 lanes
persisted. TEV cleavage of the Trx‐6His tag from FIH‐1 was performed to both ascertain
if the phosphorylation occurred on FIH‐1 or on the tag, and also to alter the position of
FIH‐1 on the gel to enable visualisation of radioactivity away from the background
bands.
TEV cleavage was achieved by first eluting Trx‐6His‐FIH‐1 from the resin using 30 μl of
500 mM imidazole, and then adding TEV protease in TEV reaction buffer to 30 μl of
eluted protein and incubating for 1 hour at 30°C. Coomassie gels demonstrated that TEV
cleavage achieved good separation of the Trx‐6His tag from FIH‐1 (Figure 6.7). The
approximately 57 kDa Trx‐6His‐FIH‐1 band is significantly reduced with TEV digestion,
indicating that digestion is essentially complete, though there may be low levels of
undigested FIH‐1 remaining. An approximately 40 kDa FIH‐1 band is observed in the
lane containing TEV‐digested protein, and there is a faint band seen in the
corresponding position in the autoradiograph. This provides evidence for a low level of
phosphorylation of FIH‐1.
One difficulty with this assay was the comparatively low yields of Trx‐6His‐FIH‐1
obtained by bacterial expression and purification compared with yields of 6His‐HIF‐1α‐
HIF CAD FIH‐1
‐ + ‐ + + +
50 kDa
75 kDa
25 kDa
37 kDa
Coomassie
Figure 6. 7 TEV Digestion of in vitro ph6His‐HIF‐1α‐CAD and Trx‐6His‐FIH‐1 weremained bound to nickel resin. Resinprotein was put in a tube and 500 μl ofinitiated by the addition of the [ϒ‐32P]‐A37°C for 3 minutes. Resins were then wμl of 500 mM imidazole or by incubatwere visualised by SDS‐PAGE and coomRepresentative of two independent exp
HIF CAD FIH‐1
‐ + ‐ + + +
Lysate
TEV protease
Trx‐6His‐FIH‐1
Autorad
FIH‐1
osphorylated proteinere bacterially expressed and purified, and proteinestimated to be bound to 20 μg of each substrateHeLa lysate or 500 μl of buffer added. Assays wereATP cocktail (150 μM, ~ 50 μCi), and incubated atwashed and protein eluted by either addition of 30tion with TEV protease at 30°C for 1 hour. Resultsassie staining, followed by autoradiography.periments.
295
CAD. For example, prior to the TEV cleavage experiments described above, proteins
were eluted from 10 μl of resin by addition of 30 μl of 250 mM imidazole, and the
proteins quantified to give an indication of the amount of resin that would need to be
used to give equal amounts of protein. For example, for the protein preparations used
in Figure 6.7, there was significantly more HIF‐1α‐CAD (approximately 2.18 μg/μl) than
FIH‐1 (approximately 0.45 μg/μl) when the proteins were eluted from the same amount
of resin. Thus, nearly five times as much FIH‐1‐resin was needed compared with HIF‐α‐
CAD‐resin to give similar amounts of the substrate proteins in the assay. This may have
contributed to the significantly higher amounts of background proteins and radioactivity
observed in FIH‐1 samples, as there was more unbound resin available to bind to non‐
specifically to proteins in the lysate. Perhaps further work could include larger scale
expression of the protein and purification and/or less Ni‐IDA resin used for purification
to increase the load of FIH‐1 on the resin.
6.3.4 In silico prediction of phosphorylation sites
Following on from the phosphatase treatments and phosphorylation assays that both
suggested that FIH‐1 is phosphorylated, bioinformatic tools were used to predict
possible sites of phosphorylation of the protein. Two online tools were used for this
purpose; Scansite (scansite.mit.edu/) and PhosphoMotif Finder
(www.hprd.org/PhosphoMotif_finder). Scansite uses an algorithm to find known motifs
in a protein sequence (Obenauer et al., 2003). Each potential motif is then ranked by a
position‐specific scoring matrix using data provided by peptide library screens. These
screens involved the modification of degenerate peptides with a central, “target” amino
acid, by a known domain, and provided information regarding the contribution of
flanking amino acids to a recognition or substrate motif. PhosphoMotif, on the other
hand, does not utilise any algorithm in the prediction of kinase consensus sequences
(Amanchy et al., 2007). Rather, this tool simply searches for and returns any sequences
296
present in the query protein that match with any published kinase or phosphatase
binding motifs.
Using a combination of these tools, Ser36 was chosen as a likely site of phosphorylation
by the serine/threonine kinase Akt. Ser36 is surface exposed and within an unstructured
region of FIH‐1, making it a good candidate for phosphorylation. Figure 6.8 shows the
structure of human FIH‐1 (MMDB ID 33168) retrieved from the Molecular Modelling
Database (MMDB) and displayed using Cn3D 4.1 program produced by NCBI (National
Centre for Biotechnology Information). This view looks down the beta barrel of FIH‐1
into its catalytic core, displaying the coordinated Fe(II) atom within the catalytic pocket.
Ser36 is indicated in yellow at the top of the structure, in an extended loop on the
surface of the protein immediately prior to a β‐strand.
6.3.5 In vitro phosphorylation assay with Ser36 FIH‐1 mutants
A serine to alanine mutant at position 36 of human FIH‐1 was made and used in an in
vitro phosphorylation assay to assess the ability of mutant FIH‐1 to be phosphorylated
by cell lysate. In addition to this mutation, Michael Nastasie (an Honours student in the
laboratory) mutated two additional residues of human FIH‐1, Ser13 and Thr302, based
on independent bioinformatic analyses, and these mutants were also included in the
assay.
Mutant and wt Trx‐6His‐FIH‐1 proteins were bacterially expressed and purified. Once
again including the HIF‐α‐CAD as a positive control, HeLa cell lysate and the [ϒ‐32P]‐ATP
cocktail were added to wt FIH‐1 and S13A, S36A and T302A FIH‐1 mutants and HIF‐α‐
CAD to allow for phosphorylation. Resin‐bound proteins were then washed and eluted
by the addition of 30 μl of 500 mM imidazole. Eluted proteins were separated by SDS‐
Figure 6. 8Structure of human FIH‐1 retrieved from(MMDB ID 33168) and displayed using Chttp://www.ncbi.nlm.nih.gov/structurestructure .
m the Molecular Modelling Database (MMDB) Cn3D 4.1 program accessed on . Ser36 is indicated in yellow at the top of the
299
PAGE, stained with Coomassie, and phosphorylation visualised by autoradiography
(Figure 6.9).
Whereas the Coomassie‐stained gel showed the presence of a single band for FIH‐1, the
autoradiograph showed that there were two bands close in molecular weight at a
similar position. Alignment of the gel and autoradiograph indicated that the upper,
larger band in the autoradiograph aligned precisely with the FIH‐1 band in the
Coomassie‐stained gel, once again consistent with phosphorylation of recombinant FIH‐
1 in this assay. In assessing phosphorylation of the three FIH‐1 mutants, S13A and
T302A mutants both displayed two bands by autoradiography similar to those seen for
wt FIH‐1. S13A and T302A mutant lanes both contained the larger molecular weight
band on the autoradiograph at the same position as the FIH‐1 band in the coomassie‐
stained gel. The S36A FIH‐1 mutant lane, however, only contained the smaller
molecular weight band in the autoradiograph, which does not correspond in size to FIH‐
1 in the coomassie‐stained gel. The S36A FIH‐1 mutant did not produce the larger
molecular weight band in the autoradiograph, suggesting that an alanine substitution at
this position significantly reduced or prevented phosphorylation of FIH‐1. These
preliminary results suggest that this residue is important and required for FIH‐1
phosphorylation, either as a direct phospho‐acceptor residue itself or as part of critical
sequence or structural recognition or binding motifs required for phosphorylation at
another distal site.
6.4 Summary and Discussion
6.4.1 Evidence in support of FIH‐1 phosphorylation
Two approaches were undertaken to investigate the possibility that FIH‐1 is
phosphorylated. Firstly, lysates treated with lambda phosphatase were separated by 2‐
wt
A A
A
FIH‐1 muta
‐ + ‐ + + + +
HIF CAD FIH‐1
S13A
S36A
T302
75
kDa
25
37
50
CCooma
Figure 6. 9 In vitro phosphorylation assFIH‐1.6His‐HIF‐1α‐CAD and wt and mutant Tand purified and protein left bound toand purified, and protein left bound toprotein was combined with 500 μl ofinitiated by the addition of the [ϒ‐32P]‐A37°C for 3 minutes. Resins were then wof 500 mM imidazole Results werefollowed by autoradiography.
A
ants
wt
A A
A
FIH‐1 mutants
+ Lysate
T302
FIH‐1
‐ + ‐ + + + +
HIF CAD FIH‐1
S13A
S36A
T302A
FIH 1
i Autoradssie Autorad
say comparing phosphorylation of wt and mutant
Trx‐6His‐FIH‐1 proteins were bacterially expressedo nickel resin 20 μg of each resin bound substrateo nickel resin. 20 μg of each resin‐bound substrateHeLa lysate or 500 μl of buffer only. Assays wereATP cocktail (150 μM, ~ 50 μCi), and incubated atwashed and proteins eluted with addition of 30 μlvisualised by SDS‐PAGE and coomassie staining,
303
DE to observe any changes in pI of FIH‐1 spots that might indicate loss of one or more
phosphate groups. The addition of an internal reference protein enabled accurate
comparison of phosphatase‐treated and untreated samples, and revealed a shift of the
spots to a more basic position, consistent with the loss of a negative charge(s) and
therefore the loss of one or more phosphate groups. These data suggest that FIH‐1 is
phosphorylated in HeLa cells. The fact that there were still multiple spots of different pI
observed following successful phosphatase treatment may suggest that multiple
modifications in addition to phosphorylation give rise to the isoelectric point
heterogeneity seen for FIH‐1, such as deamidation.
Secondly, a phosphorylation assay was devised to provide evidence for phosphorylation
of FIH‐1. Recombinant Trx‐6His‐hFIH‐1 was incubated with HeLa cell lysate in the
presence of [ϒ32‐P]‐ATP, and results indicated that FIH‐1 was indeed phosphorylated by
a kinase or kinases supplied by the lysate in this assay. This experiment is not
necessarily an accurate representation of what occurs in a cell, as lysates were prepared
from whole cells and incubated with an excess of purified FIH‐1. This experiment does,
however, indicate that FIH‐1 is able to accept a phosphate group under these conditions
and is therefore a candidate for in vivo phosphorylation. Taken together, these results
provide compelling evidence that FIH‐1 may be subject to phosphorylation events in
cells.
6.4.2 Possible site of phosphorylation and involvement of Akt
In addition to these broad experiments looking for FIH‐1 phosphorylation, a number of
bioinformatics tools were used for the prediction of PTMs. Ser36 of FIH‐1 was found to
be a good candidate for Akt‐mediated phosphorylation. In preliminary experiments
investigating the possible phosphorylation of these specific residues, a S36A FIH‐1
mutant exhibited an attenuated signal in the phosphorylation assay compared with wt
304
FIH‐1 and S13A and T302A FIH‐1 mutants. Whether this is because Ser36 is
phosphorylated itself, or if the mutation disrupts the binding, interaction or
modification of FIH‐1 at another site, is not clear.
The predicted site of FIH‐1 phosphorylation at Ser36 deviates from the Akt consensus
site of RXRXXS, with a glutamine instead of an arginine at position ‐5 relative to the
target serine (31‐ QLRSYS ‐36). Serine 36 is conserved in mouse and rat FIH‐1, and cow
and zebrafish FIH‐1 have a threonine in the equivalent position, a residue which is also
readily able to be phosphorylated by Akt (Figure 6.10). Interestingly, while the serine at
position 36 is not conserved in Triboleum castaneum (red flour beetle), which has a well
conserved FIH‐1, that position is occupied by an aspartic acid. Substitution of
phosphorylated residues with acid residues is known to mimic the effects of
phosphorylation via the introduction of a negative charge (For examples, see
(Maciejewski et al., 1995; Trautwein et al., 1993). This residue is not conserved in
Xenopus laevis, with a glutamine in this position.
The involvement of Akt in regulation of HIF‐mediated transcription remains
controversial. Both positive and negative regulation of HIF target genes have been
observed following PI3K/Akt activation depending on the cell type (Shafee et al., 2009),
and inhibiting the pathway was found to have cell‐specific effects on HRE‐reporter gene
transcription (Alvarez‐Tejado et al., 2002). Importantly, HIF‐1α itself is not a substrate of
Akt, suggesting that regulation of HIF‐mediated transcription mediated by Akt occurs
indirectly (Zundel et al., 2000). Based on the preliminary findings presented in this
chapter, investigation into the Akt‐mediated phosphorylation of FIH‐1 should be
continued.
HsFIH SMmFIH SRnFIH SBtFIH SBtFIH SDrFIH SXlFIH STcFIH S
*
Figure 6.10. Multiple species alignmen
Hs: Homo sapiens (human), Mm: Mus mBos taurus (cow), Dr: Danio rerio (zebraTribolium castaneum (red flour beetle)
SQLRSYSSQLRSYSSQLRSYSSQLRSYTSQLRSYTSQLRQYTSQLRSYQSQLRKYD**** *
t of potential Akt1 site at Serine 36 of FIH‐1
musculus (mouse), Rn: Rattus novergicus (rat), Bt:
fish), Xl: Xenopus laevis (African clawed frog), Tc:
307
CHAPTER 7
DISCUSSION AND
CONCLUDING
REMARKS
309
7 DISCUSSION AND CONCLUDING REMARKS
7. 1 Expression of FIH‐1
FIH‐1 is a significant contributor to the regulation of HIF transcriptional activity.
Hydroxylation of HIF‐α by FIH‐1 prevents coactivator recruitment and thereby precludes
activation of transcription of HIF‐CAD‐regulated genes. As expression of numerous HIF
target genes are generally considered to be important in solid tumour progression, it
was of interest to investigate the levels of FIH‐1 in solid tumours. It was hypothesised
that, as FIH‐1 was important in regulating transcriptional activity of the HIF‐α‐CAD,
decreased FIH‐1 levels in neoplastic tissues may boost HIF‐driven transcription and may
be advantageous for cancer progression.
One of the original aims of this thesis was to characterise the expression of FIH‐1 in a
range of normal tissues, and to then investigate the expression of FIH‐1 in cancer
tissues. In relation to this aim, generation of an anti‐FIH‐1 monoclonal antibody was
attempted. Characterisation of the resulting antibody however delivered unsatisfactory
results, demonstrating that, while the antibody was capable of detecting overexpressed
FIH‐1, the primary antigen detected by the antibody was not FIH‐1. Instead, use of an
existing polyclonal antibody was optimised and used to assess FIH‐1 expression in a
number of cancer cell lines (Figure 3.9) and published in Bracken et al (2006). FIH‐1 was
found to be expressed at a similar level in all cell lines investigated, despite data
inferring cell‐specific differences in repression of the HIF‐α‐CAD by FIH‐1, and suggesting
regulation of FIH‐1 activity rather than protein levels in different cell types.
During the course of that work, data detailing the expression of FIH‐1 in a range of
normal human tissues and some tumour tissues was published (Soilleux et al., 2005).
310
Several subsequent publications also examined the expression of FIH‐1 in breast cancer
(Tan et al., 2007; Tan et al., 2009), non‐small cell lung cancer (Giatromanolaki et al.,
2008) and pancreatic cancer (Couvelard et al., 2008). These data hint that the role of
FIH‐1 in cancer progression is likely to be a complicated story to unravel.
7.1.1 FIH‐1 expression is not down regulated in some cancers
Cancer tissues investigated thus far have not shown a decrease in FIH‐1 protein levels;
in fact, FIH‐1 levels have been reported to be slightly elevated in renal cancer compared
to normal kidney tissue (Soilleux et al., 2005), despite a reported mechanism of PKC‐
mediated repression of FIH‐1 transcription in RCC cell lines. Strong FIH‐1 staining has
also been observed in follicular lymphoma and pancreatic endocrine tumours
(Couvelard et al., 2008). These findings suggest inactivation of FIH‐1 activity rather than
downregulation of transcription or protein levels may be important for elevated HIF
activity in tumour tissues. Regulation of FIH‐1 activity may be a consequence of
decreased levels of oxygen in hypoxic regions of solid tumours, or may be due to other,
as yet undefined regulatory processes that affect FIH‐1 activity or protein‐protein
interactions rather than protein levels.
7.1.2 Subcellular localisation of FIH‐1 in breast and non‐small cell lung cancer may be
important
The subcellular distribution of FIH‐1 has been found to be important in breast cancer
(Tan et al., 2007). FIH‐1 also shows strong, consistent staining in non‐small cell lung
cancer, with some differences in subcellular localisation (Giatromanolaki et al., 2008). In
lung cancer, expression of the PHDs was found to be strongly nuclear only in
combination with strong cytoplasmic staining, whereas high levels of nuclear FIH‐1 were
observed without strong cytoplasmic staining in a small number of cases. Localisation of
311
FIH‐1 in cells from normal tissues is generally higher in the cytoplasm, with lower levels,
if any, present in the nuclei of cells from a wide range of tissues. The only exception to
this has been the reported low level FIH‐1 staining in the nuclei of pancreatic acini cells
in the absence of any cytoplasmic staining (Soilleux et al., 2005). Together, these data
suggest that localisation of FIH‐1 may be a regulated process. However, whether any
mechanism(s) exist that regulate FIH‐1 trafficking between the cytoplasm and the
nucleus is not yet known.
7.1.3 Possible mechanisms leading to altered FIH‐1 localisation
Macrophages are a unique example of a cell type where HIF target genes are
constitutively expressed regardless of oxygen concentration as part of the normal
metabolic strategy. In macrophages, it has been found that FIH‐1 is sequestered away
from HIF‐α substrates via binding to first to MT1‐MMP at the Golgi, and then to binding
to Mint3 (Sakamoto and Seiki, 2010). This frees HIF‐α from hydroxylation and enables it
to remain active so it can upregulate glycolysis regardless oxygen levels in the cell.
While this is important in macrophages that are required at sites of inflammation and
injury to allow normoxic gene expression and therefore persistent upregulation of
glycolysis, it may also a normal physiological process that can become hijacked in
tumours, enabling HIF to escape FIH‐1 mediated repression without altering levels of
FIH‐1 protein.
Interestingly, the expression of MT1‐MMP has been investigated in a number of tumour
types and has been found to correlate with malignant cancer phenotypes (Sounni et al.,
2002a; Deryugina et al., 2002; Jiang et al., 2006). MT1‐MMP has a well‐defined role in
local extracellular matrix digestion and invasion, and is therefore known to assist in
tumour metastasis and progression through well defined mechanisms (Reviewed in
(Itoh, 2006). MT1‐MMP also has roles in tumour angiogenesis which are reported to be
312
independent to a role in endothelial cell migration and instead rely on enhanced VEGF
expression, though the precise pathway of this upregulation is unknown (Sounni et al.,
2002b; Sounni et al., 2004). In light of the recent data from Sakamoto and Seiki (2010),
it is an intriguing possibility that MT1‐MMP may exert an inhibitory effect on FIH‐1
through sequestering the enzyme from HIF‐α substrates in tumours to facilitate the
expression of pro‐angiogenic genes such as VEGF, similar to the pro‐glycolytic role of
this mechanism in macrophages.
7.1.4 Potential regulators of FIH‐1 levels in cancer
Strong expression of FIH‐1 has been demonstrated for certain cancer types, however no
process has been directly described to be responsible for increasing FIH‐1 protein levels.
One mechanism for decreased FIH‐1 expression in tumours involves the repression of
FIH‐1 message by miR‐31 in HNSCC (Liu et al., 2010). In HNSCC, increased miR‐31 is
correlated with increased HIF‐1 activity and the promotion of important oncogenic
traits such as proliferation and migration. Importantly, FIH‐1 expression in HNSCC SAS
cells was found to decrease cell proliferation and migration, suggesting that, in this cell
type, FIH‐1 has a tumour‐suppressing role.
The expression of miR‐31 in cancer is variable, with some cancer types, including
HNSSC, reported to upregulate miR‐31 (Liu et al., 2010), and others cancer types
reported to specifically downregulate miR‐31 expression, as observed in aggressive
breast cancer (Valastyan et al., 2009). In breast cancer, the downregulation of miR‐31
alleviates the repression of pro‐metastatic genes, thereby enhancing metastasis. Ectopic
overexpression of miR‐31 in metastatic breast cancer cells results in marked impairment
of the ability to colonise metastases in mouse xenograft models (Valastyan et al., 2009).
As both up‐ and downregulation have been reported to enhance oncogenicity, miR‐31
313
has dual functions in tumour promotion and tumour suppression that appear to be
context‐dependent.
miR‐31 is reported to target the 3’UTR of FIH‐1 message and repress translation.
Downregulation of miR‐31 in breast cancer, or other types of cancer where miR‐31 is
downregulated, may represent a mechanism of enhanced FIH‐1 protein levels. Whether
expression of FIH‐1 is increased in metastatic breast cancer or other cancers due to
attenuated miR‐31 expression has not been investigated. miR‐31 is reported to be
downregulated in gastric cancer (Zhang et al., 2010b) and to be involved in the
resistance of prostate cancer cell lines to chemotherapy‐induced apoptosis (Bhatnagar
et al., 2010), however the expression of FIH‐1 in these cancer types has not been
investigated.
7.1.5 Tumour promoting role of FIH‐1 in renal cell carcinoma
A role for elevated FIH‐1 in cancer is not yet clear, though it is possible that FIH‐1 may
exert tumour type or cell type‐specific effects on cancer progression depending on the
gene expression profile of specific tumour types. Evidence for a dual role for FIH‐1 in
cancer comes from research into clear cell renal cell carcinoma (CCRCC), where FIH‐1
has been suggested to play a tumour‐promoting role via the preferential repression of
HIF‐1α over HIF‐2α (Khan et al., 2011). Most cases of CCRCC arise from biallelic loss or
inactivation of VHL, leading to constitutively stable HIF. HIF‐1α activity in RCC cell lines
has been found to increase apoptosis, and therefore to be a negative factor in CCRCC
progression. In keeping with this, HIF‐1α expression is highest in early stages of renal
carcinoma, whereas HIF‐2α is the dominant isoform expressed in more advanced solid
tumours (Raval et al., 2005). In 786‐O cells, a RCC cell line that only expresses the HIF‐2α
isoform, FIH‐1 is not observed to have any impact on HIF‐2α ‐mediated gene expression,
314
indicating that HIF‐2α has an attenuated response to hydroxylation and silencing by FIH‐
1.
This may be a product of the amino acid differences between HIF‐1α and HIF‐2α that
renders HIF‐2α a much poorer FIH‐1 substrate. HIF‐1α peptide substrates are more
efficiently hydroxylated by FIH than HIF‐2α substrates in vitro (Koivunen et al., 2004).
Although there is quite high conservation at the amino acid level between the HIF‐1α
and HIF‐2α CADs, HIF‐2α has a valine in the equivalent position to the important Ala804
of HIF‐1α. Mutational analyses that swapped the residues in this position (Ala804Val in
HIF‐1α and Val848Ala in HIF‐2α) were sufficient to completely reverse the in vitro
hydroxylation rates mediated by FIH‐1 such that HIF‐1α Ala804Val exhibited a Vmax
similar to that of wild‐type HIF‐2α, and Val848Ala HIF‐2α an elevated Vmax much like
that of wild‐type HIF‐1α (Bracken et al., 2006).
Given that in some RCC cell lines, FIH‐1 is not able to regulate HIF‐2 activity, any role
that FIH‐1 may play in tumour progression may depend on which is the dominant HIF‐α
isoform expressed in a tumour. Expression of FIH‐1 in HNSCC was found to impede
tumour progression, whereas FIH‐1 was found to be a positive factor in renal
carcinoma. Interaction with HIF‐α and/or other substrates in cancer cells may render
FIH‐1 as a tumour suppressor or tumour promoter depending on the context. Any
regulatory mechanism(s) that impart alterations into FIH‐1 protein levels, such as miR‐
31 and PKC, subcellular distribution, such as MT1‐MMP and Mint3, or activity, may
also be acting in a context‐dependent manner.
7.1.6 Summary
The emerging story of FIH‐1 expression and involvement in cancer progression is a
complex one, with extensive research required to gain a better understanding of its
315
role. In cancer tissues examined to date, there is no evidence of a significant decrease in
FIH‐1 protein levels in solid tumours. Rather, FIH‐1 appears to be expressed at either
similar or higher levels in tumour tissues compared to matched normal tissue controls.
Therefore, it can be hypothesised that, if HIF‐α is to escape the repressive hydroxylation
of its CAD, the activity rather than protein level of FIH‐1 may be downregulated. The
subcellular localisation of FIH‐1 may be important in tumour progression, with lack of
nuclear FIH‐1 correlating with more aggressive cancer phenotypes. This not only
suggests a possible role for FIH‐1 in cancer progression, but also hints at the existence
of cellular machinery that acts to regulate the location of FIH‐1 in cells. The remainder
of this work aimed to elucidate whether FIH‐1 was subjected to regulation by post
translational modification in the cell lines that exhibited variations of FIH‐1 activity with
the HIF‐α substrates.
7.2 Regulation of FIH‐1
7.2.1 FIH‐1 regulation by hydroxylation status of the ARD pool?
The initial data that led to the hypothesis that FIH‐1 was regulated by mechanisms other
than oxygen levels was that it was seen that the CAD displayed different levels of
activity in different cell types at the same oxygen threshold. More recently, a number of
publications have described protein‐protein interactions between FIH‐1 and other, ARD‐
containing proteins. These interactions can result in hydroxylation of a substrate by FIH‐
1, such as with Notch1‐3 (Coleman et al., 2007) and the IκB proteins p105 and IκBα
(Cockman et al., 2006), or no hydroxylation, such as occurs with Notch 4 (Wilkins et al.,
2009). While no concrete role has been attributed to these interactions to date, a
suggestion has been that, through binding to FIH‐1, these proteins regulate HIF‐α by
reducing the availability of FIH‐1 for hydroxylating and therefore repressing the CAD.
Consistent with this, increases in HIF‐α‐CAD‐driven transcription have been observed
316
with overexpression in Notch1 and Notch3 (Coleman et al., 2007; Zheng et al., 2008),
MYPT1 (Webb et al., 2009), and IκBα knockdown reduces HIF‐mediated gene expression
(Shin et al., 2009). Such a mechanism may, in part, contribute to the observed
differences in CAD activity seen across the different cell lines, as FIH‐1 may be
dissimilarly sequestered by ARD proteins depending on the expression and
hydroxylation status of ARD proteins in each cell type, giving rise to subtle changes in
gene expression.
7.2.2 Is FIH‐1 post‐translationally modified?
Early observations from FIH‐1 immunoblots contributed to the notion that FIH‐1 may be
post‐translationally modified. In immunoblots, FIH‐1 is often observed as a doublet in
blots from various cell line lysates and using a number of different antibodies. The
migration of FIH‐1 as a doublet was also observed in Myc‐6His‐FIH‐1 stable
overexpressing cell lines made, further confirming the existence of two differently
migrating forms of FIH‐1 (For examples, see Figures 5.7 and 5.8). Separation of two
bands, a slower migrating band (upper) and a faster migrating band (lower) can be
indicative of PTM that causes a change in charge of a protein. For example, deamidation
of asparagine or glutamine residues and phosphorylation alter the charge of a protein
and therefore can alter the migration of proteins through gels during electrophoresis.
Self‐hydroxylation of FIH‐1 has been reported, though any impacts on enzyme activity
have not been described as a consequence of this modification (Chen et al., 2008).
There have been no reports of PTM of FIH‐1 that may act to modulate enzymatic
activity or otherwise regulate behaviour of FIH‐1. In light of work that has suggested
regulation of FIH‐1 beyond regulation of enzyme activity by oxygen levels, this thesis
sought to investigate whether FIH‐1 was post‐translationally modified.
317
7.2.3 2‐DE results
The 2‐DE experiments described in Chapter 4 of this thesis provide evidence for PTM of
FIH‐1. A number of PTMs are known to give rise to differences in the charge of a
protein, for example phosphorylation, deamidation, glycosylation and acetylation. The
resolving power of 2‐DE enables the separation of proteins with PTMs that affect their
charge state, however it does not enable differentiation between PTMs that confer
equivalent changes to isoelectric point. For example, deamidation and phosphorylation
both introduce increased negative charge to a protein, causing a more acidic pI, as was
observed for the spot profile for FIH‐1. As both modifications confer a single negative
charge to a protein, it is not possible to distinguish between these two modifications on
a 2‐D gel.
2‐DE immunoblots from six cell lines indicate that there are multiple forms of
endogenous FIH‐1 with different pIs in each cell line investigated (Figures 4.7, 4.8 and
4.12). Tagged and overexpressed FIH‐1 also showed similar pI heterogeneity (Figures 5.7
and 5.9). These results were consistent with PTM events on FIH‐1 that alter the
protein’s pI.
7.2.4 Purification and MS analysis
Further work aimed to characterise the nature of the FIH‐1 modifications that gave rise
to the multiple spots seen in 2‐DE blots. Large‐scale denaturing Ni2+‐affinity
chromatography was carried out to purify overexpressed Myc‐6His‐FIH‐1 from a stably
overexpressing HeLa cell line, with purified protein analysed by LC‐MALDI‐MS and ESI‐
LTQ‐Orbitrap MS. MS analysis successfully covered 73% of the protein and identified
several PTMs. Four methionine oxidation events and four asparagine deamidation
events were detected. The relevance of these modifications is discussed below.
318
7.2.5 Methionine oxidation
Oxidation of Met108, Met 160, Met325 and Met343 of human FIH‐1 was determined by
MS analysis of Myc‐6His‐FIH‐1 purified from HeLa cells. From the work presented in this
thesis however, it is unknown whether FIH‐1 methionyl oxidation occurs in cells or was
simply an artefact of sample handling.
Oxidation of methionine residues that occurs physiologically is thought to accumulate
as proteins age. Cellular methionine sulfoxide reductases act to reverse this
modification, and this cycle of oxidation and reduction has been postulated to provide a
“sink” for cellular ROS and enable cells and whole organisms to withstand oxidative
stresses (Moskovitz et al., 2001; Moskovitz et al., 1998). This role of methionine
residues as antioxidants is thought to be a major physiological role for this modification.
As this modification is thought to occur non‐specifically on all proteins with surface
exposed methionine residues, oxidation of FIH‐1 could occur physiologically where
methionine residues are freely accessible by ROS.
Methionine oxidation can alter the conformation of a protein by altering surface
hydrophobicity (Chao et al., 1997). This can lead to changes in enzyme activity, and as
methionine oxidation typically accumulates with age of a protein, here, methionine
oxidation may mediate an age‐related loss of enzyme activity. Age‐related loss of FIH‐1
enzyme activity in vitro is certainly observed in hydroxylation assays following storage of
purified proteins in the lab (Sarah Wilkins, personal communication). There are reports
of roles for methionine oxidation in alteration of enzyme activity in vitro (Levine et al.,
2000), but limited data regarding a role for methionine oxidation beyond protein aging
exists. Enzymatic reversal of methionine oxidation has been found to efficiently restore
activity of proteins in vitro, however data suggests that methionine sulfoxide reductases
319
are unlikely to be able to restore native conformation and full activity to enzymes in an
in vivo setting (Sun et al., 1999).
Methionine oxidation of FIH‐1 may well occur as a consequence of protein aging in cells,
and data from structure and function studies of other enzymes suggests that, if
methionine oxidation does indeed occur in cells, it could have an impact on FIH‐1
conformation and behaviour. However, as this modification is common to all proteins
with surface‐available methionines, and is known to accumulate with age, it is probable
that any in vivo methionine oxidation of FIH‐1 occurs as a natural consequence of
protein aging rather than as a controlled mechanism of FIH‐1 regulation. As such, this
modification is unlikely to contribute to any cell‐specific alterations in FIH‐1 activity.
7.2.6 Asparaginyl deamidation
Four asparaginyl residues within FIH‐1 were found to be deamidated; Asn58, Asn87,
Asn110 and Asn151. Deamidation removes the positively charged amide group from
asparaginyl residue side chains. This alteration in charge is in keeping with the
increasing acidity of pI seen for FIH‐1 in 2‐D blots and may therefore be a contributor to
the differences in pI of FIH‐1 seen in these experiments. However, as deamidation can
occur during 2‐DE and MS, it is not clear whether this modification occurs in cells or in
denatured protein samples.
Physiologically, deamidation of asparagine and glutamine residues has been associated
with protein aging and degradation (Robinson and Robinson, 2004). Asparaginyl
deamidation is an irreversible modification and has been postulated to serve as a
“molecular clock” that assists in the time‐dependent regulation of various cellular
processes such as protein turnover. Indeed, a negative correlation exists between the
number of asparagine and glutamine residues in a protein and a proteins’ half‐life
(Robinson et al., 1970). Deamidation has consequences for the local conformation of a
320
protein due to the introduction of a negative charge. In general, deamidation is thought
to ‘loosen’ the structure of a protein, thereby increasing susceptibility to proteolytic
attack and facilitating degradation. Enzyme activity can also be affected by
deamidation, with studies into human phenylalanine hydroxylase demonstrating that
deamidation affected the catalytic activity of the enzyme and also increased its
susceptibility to tryptic digestion (Solstad and Flatmark, 2000). Deamidation has also
been found to affect protein‐protein interactions. One significant reported role for
deamidation in vivo includes the suppression of the antiapoptotic activity of Bcl‐2 family
member Bcl‐xL in response to DNA damage (Deverman et al., 2002). This is thought to
occur via conformational changes induced by two deamidation events, leading to a
reduced ability of Bcl‐xL to bind and therefore block the proapoptotic activities of other
Bcl‐2 proteins. Clearly, physiological deamidation of proteins can have significant
consequences for the behaviour and activities of proteins in vivo.
Based on a combination of structural observations of the position of each deamidated
asparaginyl residue within FIH‐1, the identity of the adjacent amino acid residues, and
software prediction of deamidation half life of each asparaginyl residue, the likelihood
of each deamidation event occurring in vivo was estimated. Asn87 was deemed as the
most likely residue to undergo physiological deamidation as it was present on the
surface of FIH‐1 in an unstructured region, was flanked by small glycine residue and
were predicted to have the shortest deamidation half‐life of all asparagine residues in
the FIH‐1 protein sequence. Therefore, it is possible that deamidation of this residue
occurs in cells. This does not preclude deamidation of the other asparagine residues
determined by MS to be deamidated.
From structural data, Asn87 is not known to bind cofactors or participate in
configuration of the enzyme’s active site, or make contacts with HIF‐α substrates.
Multiple species alignment of FIH‐1 does indicate conservation of this residue between
321
human, mouse, rat, cow and xenopus. Future work could be performed to ascertain
whether this deamidation event, or deamidation of the other asparagine residues found
to be deamidated by MS, are of physiological significance. Knowledge of proportion of
each deamidation event on FIH‐1 could provide some information regarding the
likelihood of any physiological role. The asparaginyl residues thought to undergo
deamidation could be mutated to aspartic acid, thus mimicking the modification, and
assays could be performed in cells and in vitro to assess the activity and behaviour of
“deamidated” FIH‐1. For example, estimation of the half‐life of mutated FIH‐1 could
establish if deamidation contributed to accelerated degradation of FIH‐1. In vitro
hydroxylation assays could determine whether the rate of catalytic activity is affected
with a variety of FIH‐1 substrates. Co‐IPs with known interacting partners could be
performed to assess the impact of deamidation on FIH‐1 protein interactions in vivo,
and HRE‐luciferase reporter assays could indicate the effect of FIH‐1 deamidation on
gene regulation via binding and hydroxylation of HIF‐α substrates.
As it is not known whether these deamidation events occurred in cells or during
processing, future work could also attempt to minimise any occurrence of deamidation
in 2‐DE samples and during IEF, with the intent to then repeat 2‐DE experiments and
compare results obtained with earlier data. 2‐DE inherently requires proteins to be fully
denatured to enable accurate separation of proteins based on charge during IEF. This
requirement for denatured proteins increases the likelihood of deamidation occurring
within samples as denaturation reduces the steric hindrance to deamidation provided
by protein structure. As rates of deamidation in samples increases with temperature
and over time (Scotchler and Robinson, 1974), the need to keep 2‐D samples as cold as
possible and to limit handling and IEF times is critical in reducing deamidation.
Artefactual deamidation that occurs during IEF has been suggested to be significantly
reduced by performing the first dimension separation at 10°C (Mateos et al., 2009), and
future work could include preparation of 2‐DE samples at 4°C and repeating 2‐DE at this
322
lower temperature to assess any impacts on the FIH‐1 spot profile. Additionally, acidic
or alkaline pH has been found to accelerate the rate of deamidation, with a pH of 7
found to reduce deamidation of proteins in cell lysate, and a pH of 6 found to decrease
rates of deamidation of peptides in vitro, (Deverman et al., 2002; Scotchler and
Robinson, 1974). Therefore future experiments could include measurement and
adjustment of pH of 2‐DE samples prior to IEF. Repetition of 2‐DE on samples with an
adjusted pH at 4°C should minimise any deamidation of proteins occurring in 2‐D
samples, and comparison of results obtained with the adjusted protocol could help
ascertain whether the multiple forms of FIH‐1 observed in 2‐DE experiments described
in this thesis are the result of in‐sample deamidation or deamidation in cells.
7.2.7 Acetylation
Acetylation of internal lysine residues may also be considered as a PTM that alters the
pI of a protein, as acetylation neutralises the positive charge of the lysine. Lysine
acetylation is a dynamic, regulated PTM, with catalysis of lysine acetylation performed
by a group of enzymes collectively called lysine acetyltransferases (KATs). Numerous
lysine deacetylases act to reverse the modification. Histone acetylation is commonly
known to regulate chromatin structure and function, with the neutralisation of lysines
by acetylation loosening the interaction of histones with DNA, providing a structure
permissive to transcription. Many non‐histone protein are also acetylated, with non‐
histone acetylation found to be important in the regulation of various cellular processes
(Reviewed in (Sadoul et al., 2011).
Potential acetylation of FIH‐1 requires investigation, particularly as deacetylase enzymes
HDAC4, HDAC5 and HDAC7 have been found to modulate the transcriptional activity of
HIF‐1α (Kato et al., 2004; Seo et al., 2009). Transfection of HDAC4 and HDAC5 into HeLa
cells upregulated VEGF expression, increased expression of a GRE‐luciferase reporter
323
gene and enhanced the interaction between HIF‐1α and p300 without increasing levels
of HIF‐1α protein (Seo et al., 2009). Conversely, HDAC inhibitors increase transcriptional
activity of endogenous HIF at concentrations that do not increase levels of HIF‐1α
protein. Rather, inhibition of deacetylation decreases HIF‐α/p300 complex formation,
an effect independent of decreased acetylation of HIF‐α itself (Fath et al., 2006). A
N803A mutant, however, was still able to be transcriptionally repressed by HDAC
inhibition, suggesting that hydroxylation by FIH‐1 is not involved.
Use of online acetylation prediction tools indicated that FIH‐1 may be a candidate for
lysyl acetylation. The tools “Prediction of Acetylation on Internal Lysines” (PAIL;
bdmpail.biocuckoo.org/index.php; (Li et al., 2006), and PredMod
(www.cs.cornell.edu/w8/~amrita/predmod.html; (Basu et al., 2009) both predicted
acetylation of Lys176 and Lys304 of FIH‐1. As neither of these lysine residues was
covered by MS analyses, the acetylation‐status of FIH‐1 is not yet known. Currently, a
role for FIH‐1 in the alterations in HIF‐α activity by HDACs has not been described.
7.2.8 Phosphorylation
Kinase signalling pathways have been implicated in regulation of HIF‐α transcriptional
activity, as discussed in the Introduction Chapter of this thesis. Specifically, Akt was
found to exhibit cell‐specific effects on HIF‐mediated transcription without directly
phosphorylating the HIF‐α subunit. MAPKs were also found to influence HIF‐2α activity
without HIF‐2α being a MAPK substrate. As Akt and MAPK have also been found to
phosphorylate p300 (Huang and Chen, 2005; Sang et al., 2003), which is required for
transcriptional activation by HIF, it is difficult to confer a role for phosphorylation in
transcriptional regulation of HIF‐α solely to p300‐independent mechanisms. The fact
kinase pathway manipulation exerts cell‐specific effects on HIF activity, however, argues
324
for transcriptional regulation by phosphorylation in addition to regulation of the
ubiquitous transcriptional coactivators.
7.2.9 Evidence from this thesis that FIH‐1 is phosphorylated
The migration of FIH‐1 as a horizontal series of spots of increasingly acidic pIs on a 2‐D
field can be indicative of successive phosphorylation events that add negative charges
to proteins. To investigate the possibility that phosphorylation was responsible for the
of different pI forms observed in 2‐D blots, phosphatase treatment of cell lysate was
performed to assess the effect of loss of potential phosphate groups on the pattern of
FIH‐1 spots observed. Phosphatase treatment of cell lysate showed a basic shift in spot
position, supporting the notion that a phosphorylation contributes to different forms of
FIH‐1 with different pIs. Phosphatase treatment was not observed to consolidate the
spots into a single, more basic spot. Rather, phosphatase treatment was seen to shift
the position of all four spots one position to the right (Figure 6.3) or to reduce the
number of spots to two spots of more basic pI (Figure 6.4). This may be suggestive of a
combination of different modifications on FIH‐1 such that dephosphorylation only
reduces the pI heterogeneity whilst other modifications persist and contribute to a train
of spots. For example, it is possible that a single phosphorylation event occurs and is
lost upon phosphatase treatment, yet FIH‐1 still gives rise to several spots of different pI
as a result of, for example, a number of deamidated residues that remain. Overall,
results from phosphatase treatments and 2‐DE supported phosphorylation of FIH‐1.
In vitro phosphorylation assays corroborated the data from phosphatase treatments,
providing further evidence for phosphorylation of FIH‐1. These experiments indicated
that recombinant FIH‐1 was phosphorylated in vitro, albeit at a low level, by an
unknown kinase or kinases present in HeLa cell lysate. Collectively, these results suggest
that FIH‐1 may be phosphorylated in HeLa cells. Perhaps a future avenue for the
325
characterisation of FIH‐1 phosphorylation could include the in vitro phosphorylation of
purified FIH‐1 followed by MS analysis.
7.2.10 Potential phosphorylation of Ser36 of FIH‐1
Currently, the identity of the kinase or kinases responsible for any potential
modifications, the site of modification, or the signalling pathway and stimulus for the
modification are unknown. Possibilities include the MAPK and Akt pathways, both of
which have been implicated in HIF‐mediated gene expression, as discussed below.
Use of phosphorylation‐prediction programs suggested that Ser36 of FIH‐1 may be a
candidate for phosphorylation. Interestingly, the S36A FIH‐1 mutant exhibited markedly
reduced incorporation of 32P, indicating that this mutant was not phosphorylated to the
same extent as wildtype FIH‐1 or the two other mutants assayed. Though preliminary,
these data warrant further investigation to ascertain, first of all, whether this residue is
indeed a phosphoacceptor, and secondly, the kinase responsible for any modification.
As Ser36 lies within a conserved Akt1 recognition site, future experiments, such as
kinase assays and MS of in vitro phosphorylated FIH‐1, may seek to investigate the
phosphorylation of FIH‐1 by Akt1 as a starting point. If the results of such experiments
indicated FIH‐1 phosphorylation by Akt, is would interesting to speculate further on the
effects of such a modification. Studies have suggested that, in some cells, inhibition of
Akt reduces HIF‐mediated gene expression and transfection of a constitutively active
form of Akt increases HIF‐mediated transcription. Therefore, a logical hypothesis
follows that, if FIH‐1 is indeed an Akt substrate, phosphorylation of FIH‐1 by Akt may
somehow act to reduce FIH‐1‐mediated hydroxylation of HIF‐α, facilitating enhanced
HIF transcriptional activity in some cell types. The effects of Akt activation in a cell are
diverse, with Akt regulating a number of cellular processes including angiogenesis, cell
326
survival and proliferation and also glucose metabolism, processes also regulated by HIF.
It remains a possibility that, in certain cell types, FIH‐1 may be able to respond both to
oxygen levels and also to growth factors, hormones and cytokines via Akt to dually
modulate these overlapping cellular processes.
7.2.11 Possible mechanism of FIH‐1 regulation by Ser36 phosphorylation
In addition to prediction of Ser36 as a site for Akt‐mediated phosphorylation, phospho‐
Ser36 is predicted to be form a 14‐3‐3 mode 1 binding motif. The 14‐3‐3 protein family
consists of seven isoforms: 14‐3‐3β, ϒ, ϵ, σ, ζ, τ and η. These proteins form homo‐and
heterodimers with other 14‐3‐3 family members, and dimers then interact with other
proteins to mediate a range of effects (Reviewed in (Dougherty and Morrison, 2004). As
each monomer contains its own protein‐binding domain, each dimer is able to interact
with two separate proteins simultaneously. As such, 14‐3‐3 proteins can act to facilitate
a diverse range of protein‐protein interactions.
The 14‐3‐3 family of proteins regulate other proteins through phospho‐Ser‐ or –Thr‐
dependent interactions (For a review, see (Mackintosh, 2004). Known effects of 14‐3‐3
binding include modulation of protein‐protein interactions, regulation of subcellular
localisation, and alteration of enzyme activity. As an example of 14‐3‐3‐mediated
protein regulation, the consequences for phosphorylation by Akt and subsequent
binding of a 14‐3‐3 protein have been described for a members of the Forkhead Box
domain (Fox) family of transcription factors, FoxO1, FoxO4 and FoxO3a (Brunet et al.,
1999; Obsilova et al., 2005; Zhao et al., 2004). Following phosphorylation of FoxO
proteins by Akt, 14‐3‐3 binding obscures a nuclear localisation sequence (NLS)
(confirmed for Foxo4; (Obsilova et al., 2005), thus leading to cytoplasmic retention and
prevention of DNA binding.
327
14‐3‐3 proteins interact with other proteins via RSXpSXP binding motifs. A perfect 14‐3‐
3 binding site exists flanking the Ser36 residue (33‐RSYSFP‐38), and this site is
completely conserved between human, mouse and rat FIH‐1. Should Ser36 of FIH‐1 be
found to be phosphorylated, investigating a possible interaction with 14‐3‐3 proteins
would be an interesting next step. Mechanisms that regulate FIH‐1 location in a cell are
of particular interest and importance given the recent reports linking FIH‐1 subcellular
distribution with survival times in breast cancer (Tan et al., 2007). Interestingly, 14‐3‐3
proteins are associated with cancer progression, most notably 14‐3‐3σ. Cytoplasmic
expression of 14‐3‐3σ is strongly associated with increased malignancy of breast cancer
and poor survival outcomes (Simpson et al., 2004). Due to the pathophysiological
significance of FIH‐1 subcellular localisation, investigating potential phosphorylation of
FIH‐1 by Akt and any subsequent effects of such a modification, including the possibility
of interactions with 14‐3‐3 proteins and consequences for FIH‐1 localisation, is a
worthwhile avenue for future continuation with this project. A potential means of
nuclear exclusion and/or cytoplasmic retention of FIH‐1 could enable stabilised, nuclear
HIF‐α to evade hydroxylation of its CAD, enabling the enhanced levels of HIF targe gene
expression seen in breast cancer.
7.2.12 Tankyrase, insulin sensitivity and potential PARsylation of FIH‐1
In addition to regulation of HIF transcriptional activity, FIH‐1 also has roles in metabolic
control (Zhang et al., 2010a). This role is currently poorly understood. In literature
searches for evidence of FIH‐1 regulation by phosphorylation, effects of kinase
pathways on HIF‐transcriptional activity were the most logical choice, as the role of FIH‐
1 upon HIF‐activity has been well characterised. However, it is likely that, in some cells,
FIH‐1 has HIF‐independent effects. The surprising findings delivered from the mouse
knockout study certainly support a role for FIH‐1 in neuronal‐specific regulation of
metabolism, ventilation and insulin sensitivity (Zhang et al., 2010a). Both global and
328
neuronal FIH‐1‐/‐ mice displayed an enhanced sensitivity to insulin, evidenced by a
greater and more rapid decrease in blood glucose levels following insulin injection. At
this stage the mechanism of this is unknown.
One of the ARD‐containing proteins recently found to be an FIH‐1 substrate is Tankyrase
(Cockman et al., 2006; Cockman et al., 2009; Yang et al., 2011). Tankyrase consists of an
ARD that mediates protein‐protein interactions and also contains a PARP domain that is
responsible for the covalent attachment of poly‐ADP‐ribose groups onto proteins,
termed PARsylation. Tankyrase is known to undergo self‐PARsylation in addition to
modification of interacting proteins (Chi and Lodish, 2000).
Interestingly, Tankyrase‐/‐ mice have a hypermetabolic phenotype with similar
characteristics to FIH‐1‐/‐ mice. Tankyrase‐deficient mice exhibit reduced adiposity
despite hyperphagia, and also show greater O2 consumption and elevated body
temperature compared to wt mice (Yeh et al., 2009). These concordant phenotypes
suggest that both FIH‐1 and Tankyrase are required for normal regulation of metabolism
and furthermore, given that they are known to interact, that they may be acting to
regulate metabolism via a common mechanism or pathway.
Insulin‐regulated glucose transport in muscle, adipose and liver cells is facilitated by
Glut4 transporters in the PM. Glut4 resides primarily within intracellular vesicles termed
Glut4 storage vesicles (GSVs), and translocates to the PM upon insulin stimulation
(Reviewed in (Bryant et al., 2002)). Insufficient recruitment of Glut4 to the PM is
associated with insulin resistance, indicative of the importance of this GSV‐PM Glut4
cycling mechanism in insulin‐regulated glucose uptake. The exact molecular machinery
involved in transport and cycling of Glut4 between the GSVs and the PM are not clear,
however Tankyrase is known to be involved. Tankyrase has been found to bind, via its
ARD, to another GSV‐associated protein, insulin‐regulated aminopeptidase (IRAP). Both
329
Tankyrase and IRAP appear to be required for Glut4 translocation as knockdown of
either Tankyrase or IRAP in adipocytes was found to attenuate the transport of Glut4 to
the PM (Yeh et al., 2007). Whether FIH‐1 has any role in the PM translocation of Glut4 in
response to insulin is not known.
Clearly, the exact contributions of Tankyrase and FIH‐1 to regulation of metabolism are
yet to be elucidated. The fact that Tankyrase is a substrate of FIH‐1, and that both FIH‐1
and Tankyrase knockout mice exhibit significant phenotypic overlap, suggest that
perhaps the interaction between these two proteins has a role in metabolic regulation
in insulin‐sensitive tissues. As Tankyrase is known to PARsylate interacting proteins,
future experiments could include investigation into the possible PARsylation of FIH‐1 by
Tankyrase. PARsylation involves the attachment of negatively charged ADP‐ribose
moieties, leading to an acidic shift in pI (for example, see (Visochek et al., 2005) and
consistent with the pattern of FIH‐1 spots in 2‐DE blots. Inhibition of PARsylation can be
achieved with the use of small molecule inhibitors of poly‐ADP‐ribosylation, such as
PJ34 (Suarez‐Pinzon et al., 2003) or XAV939 (Huang et al., 2009a). Further experiments
could involve treatment of cells with PARP inhibitors followed by 2‐DE and blotting for
FIH‐1. A shift of spots and a reduction in spot number to a more basic pI would indicate
loss of a poly‐ADP‐ribose group. Inhibition of the PARP‐domain of Tankyrase using was
found to reduce the PM translocation of Glut4 and also decrease glucose uptake in
response to insulin (Yeh et al., 2007). As such, if FIH‐1 is found to be PARsylated by
Tankyrase, this PTM could have a significant impact on the activities of FIH‐1 in a cell.
7.2.13 Linking FIH‐1 with insulin‐regulated glucose uptake
Though beyond the scope of this thesis, there is further anecdotal evidence that links
FIH‐1 with known mediators of insulin‐responsive glucose homeostasis. As already
described in the introduction chapter, PKC has been found to downregulate FIH‐1
transcription in RCC cell lines (Datta et al., 2004). PKC has also been implicated in
330
insulin‐stimulated glucose transport. Overexpression of PKC in 3T3‐L1 adipocytes leads
to increase in insulin‐stimulated glucose transport (Bandyopadhyay et al., 1997).
Though the connection between PKC and glucose transport are unclear, a correlation
between overexpressed PKC, decreased FIH‐1 and enhanced insulin sensitivity is in
keeping with the enhanced insulin sensitivity of FIH‐1‐/‐ mice.
Finally, Akt is known to regulate glucose uptake in response to insulin signalling. Akt2
knockout mice have an insulin‐resistant phenotype, showing decreased glucose
clearance in response to insulin (Cho et al., 2001). Ectopic expression of Akt was found
to increase the Glut4 in the PM fraction in 3T3‐L1 adipocytes (Tanti et al., 1997; Kohn et
al., 1996), and expression of a constitutively active form of Akt1 in L6 skeletal muscle
cells was also found to enhance glucose uptake via increased Glut4 in the PM (Hajduch
et al., 1998). Should FIH‐1 be found to be a substrate of Akt, it would be interesting to
examine if Akt‐mediated phosphorylation of FIH‐1 would have any impact in insulin‐
stimulated glucose uptake or Glut4 PM translocation.
7.3 Final Conclusion
The work described in this thesis provides novel evidence in support of the PTM of FIH‐1
in a range of cell types. Furthermore, there is evidence that these modifications may
differ between different cell types, in keeping with the cell‐specific levels of FIH‐1
activity displayed by FIH‐1. Any regulatory mechanisms that are found to alter FIH‐1
behaviour could certainly be restricted to specific cells or tissues. Different tissues, and
different cells within tissues, could naturally be thought to require different levels of
FIH‐1 activity in keeping with various metabolic requirements and tissue‐specific
“normoxic” oxygen levels. Furthermore, it is possible that additional modes of FIH‐1
regulation may occur in cells where there is a physiological requirement for further,
331
oxygen‐independent FIH‐1 regulatory mechanisms, such as in cells where the role of
FIH‐1 extends beyond its role in regulating HIF transcriptional activity. For example,
neuronal cells may require discrete regulatory mechanisms of FIH‐1 pertaining to
control of metabolic processes.
As yet, the nature of PTMs of FIH‐1 remains unknown. Of immediate importance will be
the definitive identification of the PTMs that occur on FIH‐1. Deamidation of specific
asparaginyl residues of FIH‐1 should be further investigated to determine if these
events are physiologically relevant. In light of the data presented in this thesis,
phosphorylation of FIH‐1 by Akt or other kinases should also be further explored. The
idea that the FIH‐1‐Tankyrase interaction both PARsylates FIH‐1 and also assists in the
regulation of insulin‐dependent glucose uptake presents an interesting connection
between FIH‐1 and metabolic regulation, and also warrants further investigation.
At the beginning of this research, FIH‐1 was known as the asparaginyl hydroxylase
responsible for regulating HIF‐transcriptional activity. Now, an extensive list of novel
FIH‐1 substrates have been identified, FIH‐1 knockout mice have revealed a surprising
role for FIH‐1 in the regulation of global metabolism, and examination of FIH‐1
expression indicates a role in cancer progression. Investigating the regulation of this
enzyme remains essential to enhance the understanding of both physiological and
pathophysiological processes. It is suggested that FIH‐1 is subjected to PTMs, however a
complete picture of FIH‐1 modifications remains to be established. Further experiments
to address this deficit in knowledge are eagerly anticipated.
333
REFERENCES
335
8 REFERENCES
Alvarez‐Tejado, M., Alfranca, A., Aragones, J., Vara, A., Landazuri, M. O. &del Peso, L. (2002). Lack of evidence for the involvement of the phosphoinositide 3‐kinase/Akt pathway in the activation of hypoxia‐inducible factors by low oxygen tension. J Biol Chem 277(16): 13508‐13517.
Amanchy, R., Periaswamy, B., Mathivanan, S., Reddy, R., Tattikota, S. G. &Pandey, A. (2007). A curated compendium of phosphorylation motifs. Nat Biotechnol 25(3): 285‐286.
Aragones, J., Schneider, M., Van Geyte, K., Fraisl, P., Dresselaers, T., Mazzone, M., Dirkx, R., Zacchigna, S., Lemieux, H., Jeoung, N. H., Lambrechts, D., Bishop, T., Lafuste, P., Diez‐Juan, A., Harten, S. K., Van Noten, P., De Bock, K., Willam, C., Tjwa, M., Grosfeld, A., Navet, R., Moons, L., Vandendriessche, T., Deroose, C., Wijeyekoon, B., Nuyts, J., Jordan, B., Silasi‐Mansat, R., Lupu, F., Dewerchin, M., Pugh, C., Salmon, P., Mortelmans, L., Gallez, B., Gorus, F., Buyse, J., Sluse, F., Harris, R. A., Gnaiger, E., Hespel, P., Van Hecke, P., Schuit, F., Van Veldhoven, P., Ratcliffe, P., Baes, M., Maxwell, P. &Carmeliet, P. (2008). Deficiency or inhibition of oxygen sensor Phd1 induces hypoxia tolerance by reprogramming basal metabolism. Nat Genet 40(2): 170‐180.
Arsham, A. M., Plas, D. R., Thompson, C. B. &Simon, M. C. (2002). Phosphatidylinositol 3‐kinase/Akt signaling is neither required for hypoxic stabilization of HIF‐1 alpha nor sufficient for HIF‐1‐dependent target gene transcription. J Biol Chem 277(17): 15162‐15170.
Bandyopadhyay, G., Standaert, M. L., Zhao, L., Yu, B., Avignon, A., Galloway, L., Karnam, P., Moscat, J. &Farese, R. V. (1997). Activation of protein kinase C (alpha, beta, and zeta) by insulin in 3T3/L1 cells. Transfection studies suggest a role for PKC‐zeta in glucose transport. J Biol Chem 272(4): 2551‐2558.
Barrick, D., Ferreiro, D. U. &Komives, E. A. (2008). Folding landscapes of ankyrin repeat proteins: experiments meet theory. Curr Opin Struct Biol 18(1): 27‐34.
Basu, A., Rose, K. L., Zhang, J., Beavis, R. C., Ueberheide, B., Garcia, B. A., Chait, B., Zhao, Y., Hunt, D. F., Segal, E., Allis, C. D. &Hake, S. B. (2009). Proteome‐wide prediction of acetylation substrates. Proc Natl Acad Sci U S A 106(33): 13785‐13790.
Berchner‐Pfannschmidt, U., Tug, S., Kirsch, M. &Fandrey, J. (2010). Oxygen‐sensing under the influence of nitric oxide. Cell Signal 22(3): 349‐356.
Bhatnagar, N., Li, X., Padi, S. K., Zhang, Q., Tang, M. S. &Guo, B. (2010). Downregulation of miR‐205 and miR‐31 confers resistance to chemotherapy‐induced apoptosis in prostate cancer cells. Cell Death Dis 1(12): e105.
336
Birner, P., Schindl, M., Obermair, A., Breitenecker, G. &Oberhuber, G. (2001). Expression of hypoxia‐inducible factor 1alpha in epithelial ovarian tumors: its impact on prognosis and on response to chemotherapy. Clin Cancer Res 7(6): 1661‐1668.
Bishop, T., Gallagher, D., Pascual, A., Lygate, C. A., de Bono, J. P., Nicholls, L. G., Ortega‐Saenz, P., Oster, H., Wijeyekoon, B., Sutherland, A. I., Grosfeld, A., Aragones, J., Schneider, M., van Geyte, K., Teixeira, D., Diez‐Juan, A., Lopez‐Barneo, J., Channon, K. M., Maxwell, P. H., Pugh, C. W., Davies, A. M., Carmeliet, P. &Ratcliffe, P. J. (2008). Abnormal sympathoadrenal development and systemic hypotension in PHD3‐/‐ mice. Mol Cell Biol 28(10): 3386‐3400.
Bolos, V., Grego‐Bessa, J. &de la Pompa, J. L. (2007). Notch signaling in development and cancer. Endocr Rev 28(3): 339‐363.
Bracken, C. P., Fedele, A. O., Linke, S., Balrak, W., Lisy, K., Whitelaw, M. L. &Peet, D. J. (2006). Cell‐specific regulation of hypoxia‐inducible factor (HIF)‐1alpha and HIF‐2alpha stabilization and transactivation in a graded oxygen environment. J Biol Chem 281(32): 22575‐22585.
Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein‐dye binding. Anal Biochem 72: 248‐254.
Brunet, A., Bonni, A., Zigmond, M. J., Lin, M. Z., Juo, P., Hu, L. S., Anderson, M. J., Arden, K. C., Blenis, J. &Greenberg, M. E. (1999). Akt promotes cell survival by phosphorylating and inhibiting a Forkhead transcription factor. Cell 96(6): 857‐868.
Bryant, N. J., Govers, R. &James, D. E. (2002). Regulated transport of the glucose transporter GLUT4. Nat Rev Mol Cell Biol 3(4): 267‐277.
Burri, P., Djonov, V., Aebersold, D. M., Lindel, K., Studer, U., Altermatt, H. J., Mazzucchelli, L., Greiner, R. H. &Gruber, G. (2003). Significant correlation of hypoxia‐inducible factor‐1alpha with treatment outcome in cervical cancer treated with radical radiotherapy. Int J Radiat Oncol Biol Phys 56(2): 494‐501.
Capasso, S., Mazzarella, L., Sica, F. &Zagari, A. (1989). Deamidation via cyclic imide in asparaginyl peptides. Pept Res 2(2): 195‐200.
Ceradini, D. J. &Gurtner, G. C. (2005). Homing to hypoxia: HIF‐1 as a mediator of progenitor cell recruitment to injured tissue. Trends Cardiovasc Med 15(2): 57‐63.
Chao, C. C., Ma, Y. S. &Stadtman, E. R. (1997). Modification of protein surface hydrophobicity and methionine oxidation by oxidative systems. Proc Natl Acad Sci U S A 94(7): 2969‐2974.
Chen, Y. H., Comeaux, L. M., Eyles, S. J. &Knapp, M. J. (2008). Auto‐hydroxylation of FIH‐1: an Fe(ii), alpha‐ketoglutarate‐dependent human hypoxia sensor. Chem Commun (Camb) (39): 4768‐4770.
337
Chi, N. W. &Lodish, H. F. (2000). Tankyrase is a golgi‐associated mitogen‐activated protein kinase substrate that interacts with IRAP in GLUT4 vesicles. J Biol Chem 275(49): 38437‐38444.
Cho, H., Ahn, D. R., Park, H. &Yang, E. G. (2007). Modulation of p300 binding by posttranslational modifications of the C‐terminal activation domain of hypoxia‐inducible factor‐1alpha. FEBS Lett 581(8): 1542‐1548.
Cho, H., Mu, J., Kim, J. K., Thorvaldsen, J. L., Chu, Q., Crenshaw, E. B., 3rd, Kaestner, K. H., Bartolomei, M. S., Shulman, G. I. &Birnbaum, M. J. (2001). Insulin resistance and a diabetes mellitus‐like syndrome in mice lacking the protein kinase Akt2 (PKB beta). Science 292(5522): 1728‐1731.
Chowdhury, R., Flashman, E., Mecinovic, J., Kramer, H. B., Kessler, B. M., Frapart, Y. M., Boucher, J. L., Clifton, I. J., McDonough, M. A. &Schofield, C. J. (2011). Studies on the Reaction of Nitric Oxide with the Hypoxia‐Inducible Factor Prolyl Hydroxylase Domain 2 (EGLN1). J Mol Biol 410(2): 268‐279.
Cockman, M. E., Lancaster, D. E., Stolze, I. P., Hewitson, K. S., McDonough, M. A., Coleman, M. L., Coles, C. H., Yu, X., Hay, R. T., Ley, S. C., Pugh, C. W., Oldham, N. J., Masson, N., Schofield, C. J. &Ratcliffe, P. J. (2006). Posttranslational hydroxylation of ankyrin repeats in IkappaB proteins by the hypoxia‐inducible factor (HIF) asparaginyl hydroxylase, factor inhibiting HIF (FIH). Proc Natl Acad Sci U S A 103(40): 14767‐14772.
Cockman, M. E., Webb, J. D., Kramer, H. B., Kessler, B. M. &Ratcliffe, P. J. (2009). Proteomics‐based identification of novel factor inhibiting hypoxia‐inducible factor (FIH) substrates indicates widespread asparaginyl hydroxylation of ankyrin repeat domain‐containing proteins. Mol Cell Proteomics 8(3): 535‐546.
Cohen, P. T. (2002). Protein phosphatase 1‐‐targeted in many directions. J Cell Sci 115(Pt 2): 241‐256.
Coleman, M. L., McDonough, M. A., Hewitson, K. S., Coles, C., Mecinovic, J., Edelmann, M., Cook, K. M., Cockman, M. E., Lancaster, D. E., Kessler, B. M., Oldham, N. J., Ratcliffe, P. J. &Schofield, C. J. (2007). Asparaginyl hydroxylation of the Notch ankyrin repeat domain by factor inhibiting hypoxia‐inducible factor. J Biol Chem 282(33): 24027‐24038.
Compton, S. J. &Jones, C. G. (1985). Mechanism of dye response and interference in the Bradford protein assay. Anal Biochem 151(2): 369‐374.
Conrad, P. W., Freeman, T. L., Beitner‐Johnson, D. &Millhorn, D. E. (1999). EPAS1 trans‐activation during hypoxia requires p42/p44 MAPK. J Biol Chem 274(47): 33709‐33713.
Couvelard, A., Deschamps, L., Rebours, V., Sauvanet, A., Gatter, K., Pezzella, F., Ruszniewski, P. &Bedossa, P. (2008). Overexpression of the oxygen sensors PHD‐1, PHD‐2, PHD‐3, and FIH Is associated with tumor aggressiveness in pancreatic endocrine tumors. Clin Cancer Res 14(20): 6634‐6639.
338
Dann, C. E., 3rd, Bruick, R. K. &Deisenhofer, J. (2002). Structure of factor‐inhibiting hypoxia‐inducible factor 1: An asparaginyl hydroxylase involved in the hypoxic response pathway. Proc Natl Acad Sci U S A 99(24): 15351‐15356. Epub 12002 Nov 15313.
Datta, K., Li, J., Bhattacharya, R., Gasparian, L., Wang, E. &Mukhopadhyay, D. (2004). Protein kinase C zeta transactivates hypoxia‐inducible factor alpha by promoting its association with p300 in renal cancer. Cancer Res 64(2): 456‐462.
Dayan, F., Roux, D., Brahimi‐Horn, M. C., Pouyssegur, J. &Mazure, N. M. (2006). The oxygen sensor factor‐inhibiting hypoxia‐inducible factor‐1 controls expression of distinct genes through the bifunctional transcriptional character of hypoxia‐inducible factor‐1alpha. Cancer Res 66(7): 3688‐3698.
Deryugina, E. I., Soroceanu, L. &Strongin, A. Y. (2002). Up‐regulation of vascular endothelial growth factor by membrane‐type 1 matrix metalloproteinase stimulates human glioma xenograft growth and angiogenesis. Cancer Res 62(2): 580‐588.
Deverman, B. E., Cook, B. L., Manson, S. R., Niederhoff, R. A., Langer, E. M., Rosova, I., Kulans, L. A., Fu, X., Weinberg, J. S., Heinecke, J. W., Roth, K. A. &Weintraub, S. J. (2002). Bcl‐xL deamidation is a critical switch in the regulation of the response to DNA damage. Cell 111(1): 51‐62.
Devries, I. L., Hampton‐Smith, R. J., Mulvihill, M. M., Alverdi, V., Peet, D. J. &Komives, E. A. (2010). Consequences of IkappaB alpha hydroxylation by the factor inhibiting HIF (FIH). FEBS Lett 584(23): 4725‐4730.
Dougherty, M. K. &Morrison, D. K. (2004). Unlocking the code of 14‐3‐3. J Cell Sci 117(Pt 10): 1875‐1884.
Ebert, B. L., Firth, J. D. &Ratcliffe, P. J. (1995). Hypoxia and mitochondrial inhibitors regulate expression of glucose transporter‐1 via distinct Cis‐acting sequences. J Biol Chem 270(49): 29083‐29089.
Ehrismann, D., Flashman, E., Genn, D. N., Mathioudakis, N., Hewitson, K. S., Ratcliffe, P. J. &Schofield, C. J. (2007). Studies on the activity of the hypoxia‐inducible‐factor hydroxylases using an oxygen consumption assay. Biochem J 401(1): 227‐234.
Elvidge, G. P., Glenny, L., Appelhoff, R. J., Ratcliffe, P. J., Ragoussis, J. &Gleadle, J. M. (2006). Concordant regulation of gene expression by hypoxia and 2‐oxoglutarate‐dependent dioxygenase inhibition: the role of HIF‐1alpha, HIF‐2alpha, and other pathways. J Biol Chem 281(22): 15215‐15226.
Ema, M., Taya, S., Yokotani, N., Sogawa, K., Matsuda, Y. &Fujii‐Kuriyama, Y. (1997). A novel bHLH‐PAS factor with close sequence similarity to hypoxia‐inducible factor 1alpha regulates the VEGF expression and is potentially involved in lung and vascular development. Proc Natl Acad Sci U S A 94(9): 4273‐4278.
Fath, D. M., Kong, X., Liang, D., Lin, Z., Chou, A., Jiang, Y., Fang, J., Caro, J. &Sang, N. (2006). Histone deacetylase inhibitors repress the transactivation potential of hypoxia‐
339
inducible factors independently of direct acetylation of HIF‐alpha. J Biol Chem 281(19): 13612‐13619.
Feldser, D., Agani, F., Iyer, N. V., Pak, B., Ferreira, G. &Semenza, G. L. (1999). Reciprocal positive regulation of hypoxia‐inducible factor 1alpha and insulin‐like growth factor 2. Cancer Res 59(16): 3915‐3918.
Ferguson, J. E., 3rd, Wu, Y., Smith, K., Charles, P., Powers, K., Wang, H. &Patterson, C. (2007). ASB4 is a hydroxylation substrate of FIH and promotes vascular differentiation via an oxygen‐dependent mechanism. Mol Cell Biol 27(18): 6407‐6419.
Folkman, J. (1990). What is the evidence that tumors are angiogenesis dependent? J Natl Cancer Inst 82(1): 4‐6.
Folkman, J., Cole, P. &Zimmerman, S. (1966). Tumor behavior in isolated perfused organs: in vitro growth and metastases of biopsy material in rabbit thyroid and canine intestinal segment. Ann Surg 164(3): 491‐502.
Forsythe, J. A., Jiang, B. H., Iyer, N. V., Agani, F., Leung, S. W., Koos, R. D. &Semenza, G. L. (1996). Activation of vascular endothelial growth factor gene transcription by hypoxia‐inducible factor 1. Mol Cell Biol 16(9): 4604‐4613.
Franke, T. F. (2008). PI3K/Akt: getting it right matters. Oncogene 27(50): 6473‐6488.
Freedman, S. J., Sun, Z. Y., Poy, F., Kung, A. L., Livingston, D. M., Wagner, G. &Eck, M. J. (2002). Structural basis for recruitment of CBP/p300 by hypoxia‐inducible factor‐1 alpha. Proc Natl Acad Sci U S A 99(8): 5367‐5372.
Gerald, D., Berra, E., Frapart, Y. M., Chan, D. A., Giaccia, A. J., Mansuy, D., Pouyssegur, J., Yaniv, M. &Mechta‐Grigoriou, F. (2004). JunD reduces tumor angiogenesis by protecting cells from oxidative stress. Cell 118(6): 781‐794.
Gerber, H. P., Condorelli, F., Park, J. &Ferrara, N. (1997). Differential transcriptional regulation of the two vascular endothelial growth factor receptor genes. Flt‐1, but not Flk‐1/KDR, is up‐regulated by hypoxia. J Biol Chem 272(38): 23659‐23667.
Giatromanolaki, A., Koukourakis, M. I., Pezzella, F., Turley, H., Sivridis, E., Bouros, D., Bougioukas, G., Harris, A. L. &Gatter, K. C. (2008). Expression of prolyl‐hydroxylases PHD‐1, 2 and 3 and of the asparagine hydroxylase FIH in non‐small cell lung cancer relates to an activated HIF pathway. Cancer Lett.
Goldberg, M. A., Dunning, S. P. &Bunn, H. F. (1988). Regulation of the erythropoietin gene: evidence that the oxygen sensor is a heme protein. Science 242(4884): 1412‐1415.
Gradin, K., Takasaki, C., Fujii‐Kuriyama, Y. &Sogawa, K. (2002). The transcriptional activation function of the HIF‐like factor requires phosphorylation at a conserved threonine. J Biol Chem 277(26): 23508‐23514. Epub 22002 Apr 23530.
340
Gu, Y. Z., Moran, S. M., Hogenesch, J. B., Wartman, L. &Bradfield, C. A. (1998). Molecular characterization and chromosomal localization of a third alpha‐class hypoxia inducible factor subunit, HIF3alpha. Gene Expr 7(3): 205‐213.
Gustafsson, M. V., Zheng, X., Pereira, T., Gradin, K., Jin, S., Lundkvist, J., Ruas, J. L., Poellinger, L., Lendahl, U. &Bondesson, M. (2005). Hypoxia requires notch signaling to maintain the undifferentiated cell state. Dev Cell 9(5): 617‐628.
Guzy, R. D. &Schumacker, P. T. (2006). Oxygen sensing by mitochondria at complex III: the paradox of increased reactive oxygen species during hypoxia. Exp Physiol 91(5): 807‐819.
Hajduch, E., Alessi, D. R., Hemmings, B. A. &Hundal, H. S. (1998). Constitutive activation of protein kinase B alpha by membrane targeting promotes glucose and system A amino acid transport, protein synthesis, and inactivation of glycogen synthase kinase 3 in L6 muscle cells. Diabetes 47(7): 1006‐1013.
Hanahan, D. &Weinberg, R. A. (2000). The hallmarks of cancer. Cell 100(1): 57‐70.
Hardy, A. P., Prokes, I., Kelly, L., Campbell, I. D. &Schofield, C. J. (2009). Asparaginyl beta‐hydroxylation of proteins containing ankyrin repeat domains influences their stability and function. J Mol Biol 392(4): 994‐1006.
Hewitson, K. S., Lienard, B. M., McDonough, M. A., Clifton, I. J., Butler, D., Soares, A. S., Oldham, N. J., McNeill, L. A. &Schofield, C. J. (2007). Structural and mechanistic studies on the inhibition of the hypoxia‐inducible transcription factor hydroxylases by tricarboxylic acid cycle intermediates. J Biol Chem 282(5): 3293‐3301.
Hewitson, K. S., McNeill, L. A., Riordan, M. V., Tian, Y. M., Bullock, A. N., Welford, R. W., Elkins, J. M., Oldham, N. J., Bhattacharya, S., Gleadle, J. M., Ratcliffe, P. J., Pugh, C. W. &Schofield, C. J. (2002). Hypoxia‐inducible factor (HIF) asparagine hydroxylase is identical to factor inhibiting HIF (FIH) and is related to the cupin structural family. J Biol Chem 277(29): 26351‐26355. Epub 22002 May 26331.
Hirsila, M., Koivunen, P., Gunzler, V., Kivirikko, K. I. &Myllyharju, J. (2003). Characterization of the human prolyl 4‐hydroxylases that modify the hypoxia‐inducible factor. J Biol Chem 278(33): 30772‐30780. Epub 32003 Jun 30773.
Huang, L. E., Arany, Z., Livingston, D. M. &Bunn, H. F. (1996). Activation of hypoxia‐inducible transcription factor depends primarily upon redox‐sensitive stabilization of its alpha subunit. J Biol Chem 271(50): 32253‐32259.
Huang, L. E., Gu, J., Schau, M. &Bunn, H. F. (1998). Regulation of hypoxia‐inducible factor 1alpha is mediated by an O2‐dependent degradation domain via the ubiquitin‐proteasome pathway. Proc Natl Acad Sci U S A 95(14): 7987‐7992.
Huang, L. E., Willmore, W. G., Gu, J., Goldberg, M. A. &Bunn, H. F. (1999). Inhibition of hypoxia‐inducible factor 1 activation by carbon monoxide and nitric oxide. Implications for oxygen sensing and signaling. J Biol Chem 274(13): 9038‐9044.
341
Huang, S. M., Mishina, Y. M., Liu, S., Cheung, A., Stegmeier, F., Michaud, G. A., Charlat, O., Wiellette, E., Zhang, Y., Wiessner, S., Hild, M., Shi, X., Wilson, C. J., Mickanin, C., Myer, V., Fazal, A., Tomlinson, R., Serluca, F., Shao, W., Cheng, H., Shultz, M., Rau, C., Schirle, M., Schlegl, J., Ghidelli, S., Fawell, S., Lu, C., Curtis, D., Kirschner, M. W., Lengauer, C., Finan, P. M., Tallarico, J. A., Bouwmeester, T., Porter, J. A., Bauer, A. &Cong, F. (2009a). Tankyrase inhibition stabilizes axin and antagonizes Wnt signalling. Nature 461(7264): 614‐620.
Huang, W. C. &Chen, C. C. (2005). Akt phosphorylation of p300 at Ser‐1834 is essential for its histone acetyltransferase and transcriptional activity. Mol Cell Biol 25(15): 6592‐6602.
Huang, X., Ding, L., Bennewith, K. L., Tong, R. T., Welford, S. M., Ang, K. K., Story, M., Le, Q. T. &Giaccia, A. J. (2009b). Hypoxia‐inducible mir‐210 regulates normoxic gene expression involved in tumor initiation. Mol Cell 35(6): 856‐867.
Itoh, Y. (2006). MT1‐MMP: a key regulator of cell migration in tissue. IUBMB Life 58(10): 589‐596.
Ivan, M., Kondo, K., Yang, H., Kim, W., Valiando, J., Ohh, M., Salic, A., Asara, J. M., Lane, W. S. &Kaelin, W. G., Jr. (2001). HIFalpha targeted for VHL‐mediated destruction by proline hydroxylation: implications for O2 sensing. Science 292(5516): 464‐468. Epub 2001 Apr 2005.
Jaakkola, P., Mole, D. R., Tian, Y. M., Wilson, M. I., Gielbert, J., Gaskell, S. J., Kriegsheim, A., Hebestreit, H. F., Mukherji, M., Schofield, C. J., Maxwell, P. H., Pugh, C. W. &Ratcliffe, P. J. (2001). Targeting of HIF‐alpha to the von Hippel‐Lindau ubiquitylation complex by O2‐regulated prolyl hydroxylation. Science 292(5516): 468‐472. Epub 2001 Apr 2005.
Jewell, U. R., Kvietikova, I., Scheid, A., Bauer, C., Wenger, R. H. &Gassmann, M. (2001). Induction of HIF‐1alpha in response to hypoxia is instantaneous. Faseb J 15(7): 1312‐1314.
Jiang, B. H., Jiang, G., Zheng, J. Z., Lu, Z., Hunter, T. &Vogt, P. K. (2001). Phosphatidylinositol 3‐kinase signaling controls levels of hypoxia‐inducible factor 1. Cell Growth Differ 12(7): 363‐369.
Jiang, B. H., Semenza, G. L., Bauer, C. &Marti, H. H. (1996). Hypoxia‐inducible factor 1 levels vary exponentially over a physiologically relevant range of O2 tension. Am J Physiol 271(4 Pt 1): C1172‐1180.
Jiang, B. H., Zheng, J. Z., Leung, S. W., Roe, R. &Semenza, G. L. (1997). Transactivation and inhibitory domains of hypoxia‐inducible factor 1alpha. Modulation of transcriptional activity by oxygen tension. J Biol Chem 272(31): 19253‐19260.
Jiang, D., Zhou, Y., Moxley, R. A. &Jarrett, H. W. (2008). Purification and identification of positive regulators binding to a novel element in the c‐Jun promoter. Biochemistry 47(35): 9318‐9334.
342
Jiang, W. G., Davies, G., Martin, T. A., Parr, C., Watkins, G., Mason, M. D. &Mansel, R. E. (2006). Expression of membrane type‐1 matrix metalloproteinase, MT1‐MMP in human breast cancer and its impact on invasiveness of breast cancer cells. Int J Mol Med 17(4): 583‐590.
Jogi, A., Ora, I., Nilsson, H., Lindeheim, A., Makino, Y., Poellinger, L., Axelson, H. &Pahlman, S. (2002). Hypoxia alters gene expression in human neuroblastoma cells toward an immature and neural crest‐like phenotype. Proc Natl Acad Sci U S A 99(10): 7021‐7026.
Kallio, P. J., Okamoto, K., O'Brien, S., Carrero, P., Makino, Y., Tanaka, H. &Poellinger, L. (1998). Signal transduction in hypoxic cells: inducible nuclear translocation and recruitment of the CBP/p300 coactivator by the hypoxia‐inducible factor‐1alpha. Embo J 17(22): 6573‐6586.
Kao, S. H., Wong, H. K., Chiang, C. Y. &Chen, H. M. (2008). Evaluating the compatibility of three colorimetric protein assays for two‐dimensional electrophoresis experiments. Proteomics 8(11): 2178‐2184.
Kato, H., Tamamizu‐Kato, S. &Shibasaki, F. (2004). Histone deacetylase 7 associates with hypoxia‐inducible factor 1alpha and increases transcriptional activity. J Biol Chem 279(40): 41966‐41974.
Kewley, R. J., Whitelaw, M. L. &Chapman‐Smith, A. (2004). The mammalian basic helix‐loop‐helix/PAS family of transcriptional regulators. Int J Biochem Cell Biol 36(2): 189‐204.
Khan, M. N., Bhattacharyya, T., Andrikopoulos, P., Esteban, M. A., Barod, R., Connor, T., Ashcroft, M., Maxwell, P. H. &Kiriakidis, S. (2011). Factor inhibiting HIF (FIH‐1) promotes renal cancer cell survival by protecting cells from HIF‐1alpha‐mediated apoptosis. Br J Cancer 104(7): 1151‐1159.
Kohn, A. D., Summers, S. A., Birnbaum, M. J. &Roth, R. A. (1996). Expression of a constitutively active Akt Ser/Thr kinase in 3T3‐L1 adipocytes stimulates glucose uptake and glucose transporter 4 translocation. J Biol Chem 271(49): 31372‐31378.
Koivunen, P., Hirsila, M., Gunzler, V., Kivirikko, K. I. &Myllyharju, J. (2004). Catalytic properties of the asparaginyl hydroxylase (FIH) in the oxygen sensing pathway are distinct from those of its prolyl 4‐hydroxylases. J Biol Chem 279(11): 9899‐9904. Epub 2003 Dec 9829.
Koivunen, P., Hirsila, M., Kivirikko, K. I. &Myllyharju, J. (2006). The length of peptide substrates has a marked effect on hydroxylation by the hypoxia‐inducible factor prolyl 4‐hydroxylases. J Biol Chem 281(39): 28712‐28720.
Krishnamachary, B., Berg‐Dixon, S., Kelly, B., Agani, F., Feldser, D., Ferreira, G., Iyer, N., LaRusch, J., Pak, B., Taghavi, P. &Semenza, G. L. (2003). Regulation of colon carcinoma cell invasion by hypoxia‐inducible factor 1. Cancer Res 63(5): 1138‐1143.
343
Krokhin, O. V., Antonovici, M., Ens, W., Wilkins, J. A. &Standing, K. G. (2006). Deamidation of ‐Asn‐Gly‐ sequences during sample preparation for proteomics: Consequences for MALDI and HPLC‐MALDI analysis. Anal Chem 78(18): 6645‐6650.
Kulshreshtha, R., Davuluri, R. V., Calin, G. A. &Ivan, M. (2008). A microRNA component of the hypoxic response. Cell Death Differ 15(4): 667‐671.
Kulshreshtha, R., Ferracin, M., Wojcik, S. E., Garzon, R., Alder, H., Agosto‐Perez, F. J., Davuluri, R., Liu, C. G., Croce, C. M., Negrini, M., Calin, G. A. &Ivan, M. (2007). A microRNA signature of hypoxia. Mol Cell Biol 27(5): 1859‐1867.
Lancaster, D. E., McNeill, L. A., McDonough, M. A., Aplin, R. T., Hewitson, K. S., Pugh, C. W., Ratcliffe, P. J. &Schofield, C. J. (2004). Disruption of dimerization and substrate phosphorylation inhibit factor inhibiting hypoxia‐inducible factor (FIH) activity. Biochem J 383(Pt. 3): 429‐437.
Lando, D., Peet, D. J., Gorman, J. J., Whelan, D. A., Whitelaw, M. L. &Bruick, R. K. (2002a). FIH‐1 is an asparaginyl hydroxylase enzyme that regulates the transcriptional activity of hypoxia‐inducible factor. Genes Dev 16(12): 1466‐1471.
Lando, D., Peet, D. J., Whelan, D. A., Gorman, J. J. &Whitelaw, M. L. (2002b). Asparagine hydroxylation of the HIF transactivation domain a hypoxic switch. Science 295(5556): 858‐861.
Lee, Y. M., Jeong, C. H., Koo, S. Y., Son, M. J., Song, H. S., Bae, S. K., Raleigh, J. A., Chung, H. Y., Yoo, M. A. &Kim, K. W. (2001). Determination of hypoxic region by hypoxia marker in developing mouse embryos in vivo: a possible signal for vessel development. Dev Dyn 220(2): 175‐186.
Levine, R. L., Moskovitz, J. &Stadtman, E. R. (2000). Oxidation of methionine in proteins: roles in antioxidant defense and cellular regulation. IUBMB Life 50(4‐5): 301‐307.
Levy, A. P., Levy, N. S., Wegner, S. &Goldberg, M. A. (1995). Transcriptional regulation of the rat vascular endothelial growth factor gene by hypoxia. J Biol Chem 270(22): 13333‐13340.
Li, A., Xue, Y., Jin, C., Wang, M. &Yao, X. (2006). Prediction of Nepsilon‐acetylation on internal lysines implemented in Bayesian Discriminant Method. Biochem Biophys Res Commun 350(4): 818‐824.
Li, F., Sonveaux, P., Rabbani, Z. N., Liu, S., Yan, B., Huang, Q., Vujaskovic, Z., Dewhirst, M. W. &Li, C. Y. (2007a). Regulation of HIF‐1alpha stability through S‐nitrosylation. Mol Cell 26(1): 63‐74.
Li, J., Wang, E., Dutta, S., Lau, J. S., Jiang, S. W., Datta, K. &Mukhopadhyay, D. (2007b). Protein kinase C‐mediated modulation of FIH‐1 expression by the homeodomain protein CDP/Cut/Cux. Mol Cell Biol 27(20): 7345‐7353.
Linke, S. (2009). Substrate Specificity of Factor Inhibiting HIF‐1 (FIH‐1).
344
Linke, S., Stojkoski, C., Kewley, R. J., Booker, G. W., Whitelaw, M. L. &Peet, D. J. (2004). Substrate requirements of the oxygen‐sensing asparaginyl hydroxylase factor‐inhibiting hypoxia‐inducible factor. J Biol Chem 279(14): 14391‐14397.
Lisy, K. (2004). Characterisation of FIH‐1 Expression.
Liu, C. J., Tsai, M. M., Hung, P. S., Kao, S. Y., Liu, T. Y., Wu, K. J., Chiou, S. H., Lin, S. C. &Chang, K. W. (2010). miR‐31 ablates expression of the HIF regulatory factor FIH to activate the HIF pathway in head and neck carcinoma. Cancer Res 70(4): 1635‐1644.
Liu, Y., Christou, H., Morita, T., Laughner, E., Semenza, G. L. &Kourembanas, S. (1998). Carbon monoxide and nitric oxide suppress the hypoxic induction of vascular endothelial growth factor gene via the 5' enhancer. J Biol Chem 273(24): 15257‐15262.
Lundby, C., Calbet, J. A. &Robach, P. (2009). The response of human skeletal muscle tissue to hypoxia. Cell Mol Life Sci 66(22): 3615‐3623.
Maciejewski, P. M., Peterson, F. C., Anderson, P. J. &Brooks, C. L. (1995). Mutation of serine 90 to glutamic acid mimics phosphorylation of bovine prolactin. J Biol Chem 270(46): 27661‐27665.
Mackintosh, C. (2004). Dynamic interactions between 14‐3‐3 proteins and phosphoproteins regulate diverse cellular processes. Biochem J 381(Pt 2): 329‐342.
Mahon, P. C., Hirota, K. &Semenza, G. L. (2001). FIH‐1: a novel protein that interacts with HIF‐1alpha and VHL to mediate repression of HIF‐1 transcriptional activity. Genes Dev 15(20): 2675‐2686.
Makino, Y., Cao, R., Svensson, K., Bertilsson, G., Asman, M., Tanaka, H., Cao, Y., Berkenstam, A. &Poellinger, L. (2001). Inhibitory PAS domain protein is a negative regulator of hypoxia‐inducible gene expression. Nature 414(6863): 550‐554.
Mateos, A., Girardet, J. M., Molle, D., Dary, A., Miclo, L. &Gaillard, J. L. (2009). Two‐dimensional cartography of equine beta‐casein variants achieved by isolation of phosphorylation isoforms and control of the deamidation phenomenon. J Dairy Sci 92(6): 2389‐2399.
McCarthy, J., Hopwood, F., Oxley, D., Laver, M., Castagna, A., Righetti, P. G., Williams, K. &Herbert, B. (2003). Carbamylation of proteins in 2‐D electrophoresis‐‐myth or reality? J Proteome Res 2(3): 239‐242.
Metzen, E., Berchner‐Pfannschmidt, U., Stengel, P., Marxsen, J. H., Stolze, I., Klinger, M., Huang, W. Q., Wotzlaw, C., Hellwig‐Burgel, T., Jelkmann, W., Acker, H. &Fandrey, J. (2003a). Intracellular localisation of human HIF‐1 alpha hydroxylases: implications for oxygen sensing. J Cell Sci 116(Pt 7): 1319‐1326.
Metzen, E., Zhou, J., Jelkmann, W., Fandrey, J. &Brune, B. (2003b). Nitric oxide impairs normoxic degradation of HIF‐1alpha by inhibition of prolyl hydroxylases. Mol Biol Cell 14(8): 3470‐3481.
345
Morrison, S. J., Csete, M., Groves, A. K., Melega, W., Wold, B. &Anderson, D. J. (2000). Culture in reduced levels of oxygen promotes clonogenic sympathoadrenal differentiation by isolated neural crest stem cells. J Neurosci 20(19): 7370‐7376.
Moskovitz, J., Bar‐Noy, S., Williams, W. M., Requena, J., Berlett, B. S. &Stadtman, E. R. (2001). Methionine sulfoxide reductase (MsrA) is a regulator of antioxidant defense and lifespan in mammals. Proc Natl Acad Sci U S A 98(23): 12920‐12925.
Moskovitz, J., Flescher, E., Berlett, B. S., Azare, J., Poston, J. M. &Stadtman, E. R. (1998). Overexpression of peptide‐methionine sulfoxide reductase in Saccharomyces cerevisiae and human T cells provides them with high resistance to oxidative stress. Proc Natl Acad Sci U S A 95(24): 14071‐14075.
Mylonis, I., Chachami, G., Samiotaki, M., Panayotou, G., Paraskeva, E., Kalousi, A., Georgatsou, E., Bonanou, S. &Simos, G. (2006). Identification of MAPK phosphorylation sites and their role in the localization and activity of hypoxia‐inducible factor‐1alpha. J Biol Chem 281(44): 33095‐33106.
Nastasie, M. (2009). Exploring the Post Translational Modification of the Asparaginyl Hydroxylase FIH.
Nepveu, A. (2001). Role of the multifunctional CDP/Cut/Cux homeodomain transcription factor in regulating differentiation, cell growth and development. Gene 270(1‐2): 1‐15.
O'Farrell, P. H. (1975). High resolution two‐dimensional electrophoresis of proteins. J Biol Chem 250(10): 4007‐4021.
Obenauer, J. C., Cantley, L. C. &Yaffe, M. B. (2003). Scansite 2.0: Proteome‐wide prediction of cell signaling interactions using short sequence motifs. Nucleic Acids Res 31(13): 3635‐3641.
Obsilova, V., Vecer, J., Herman, P., Pabianova, A., Sulc, M., Teisinger, J., Boura, E. &Obsil, T. (2005). 14‐3‐3 Protein interacts with nuclear localization sequence of forkhead transcription factor FoxO4. Biochemistry 44(34): 11608‐11617.
Ooe, N., Motonaga, K., Kobayashi, K., Saito, K. &Kaneko, H. (2009). Functional characterization of basic helix‐loop‐helix‐PAS type transcription factor NXF in vivo: putative involvement in an "on demand" neuroprotection system. J Biol Chem 284(2): 1057‐1063.
Park, Y. K., Ahn, D. R., Oh, M., Lee, T., Yang, E. G., Son, M. &Park, H. (2008). Nitric oxide donor, (+/‐)‐S‐nitroso‐N‐acetylpenicillamine, stabilizes transactive hypoxia‐inducible factor‐1alpha by inhibiting von Hippel‐Lindau recruitment and asparagine hydroxylation. Mol Pharmacol 74(1): 236‐245.
Petrack, B., Greengard, P., Craston, A. &Sheppy, F. (1965). Nicotinamide Deamidase from Mammalian Liver. J Biol Chem 240: 1725‐1730.
346
Potgieter, H. C., Ubbink, J. B., Bissbort, S., Bester, M. J., Spies, J. H. &Vermaak, W. J. (1997). Spontaneous oxidation of methionine: effect on the quantification of plasma methionine levels. Anal Biochem 248(1): 86‐93.
Pugh, C. W., O'Rourke, J. F., Nagao, M., Gleadle, J. M. &Ratcliffe, P. J. (1997). Activation of hypoxia‐inducible factor‐1; definition of regulatory domains within the alpha subunit. J Biol Chem 272(17): 11205‐11214.
Raval, R. R., Lau, K. W., Tran, M. G., Sowter, H. M., Mandriota, S. J., Li, J. L., Pugh, C. W., Maxwell, P. H., Harris, A. L. &Ratcliffe, P. J. (2005). Contrasting properties of hypoxia‐inducible factor 1 (HIF‐1) and HIF‐2 in von Hippel‐Lindau‐associated renal cell carcinoma. Mol Cell Biol 25(13): 5675‐5686.
Richard, D. E., Berra, E., Gothie, E., Roux, D. &Pouyssegur, J. (1999). p42/p44 mitogen‐activated protein kinases phosphorylate hypoxia‐inducible factor 1alpha (HIF‐1alpha) and enhance the transcriptional activity of HIF‐1. J Biol Chem 274(46): 32631‐32637.
Rivers, J., McDonald, L., Edwards, I. J. &Beynon, R. J. (2008). Asparagine deamidation and the role of higher order protein structure. J Proteome Res 7(3): 921‐927.
Robinson, A. B., McKerrow, J. H. &Cary, P. (1970). Controlled deamidation of peptides and proteins: an experimental hazard and a possible biological timer. Proc Natl Acad Sci U S A 66(3): 753‐757.
Robinson, N. E. (2002). Protein deamidation. Proc Natl Acad Sci U S A 99(8): 5283‐5288.
Robinson, N. E. &Robinson, A. B. (2001a). Molecular clocks. Proc Natl Acad Sci U S A 98(3): 944‐949.
Robinson, N. E. &Robinson, A. B. (2001b). Prediction of protein deamidation rates from primary and three‐dimensional structure. Proc Natl Acad Sci U S A 98(8): 4367‐4372.
Robinson, N. E. &Robinson, A. B. (2004). Amide molecular clocks in drosophila proteins: potential regulators of aging and other processes. Mech Ageing Dev 125(4): 259‐267.
Ryan, H. E., Lo, J. &Johnson, R. S. (1998). HIF‐1 alpha is required for solid tumor formation and embryonic vascularization. Embo J 17(11): 3005‐3015.
Rybina, I. V., Liu, H., Gor, Y. &Feinmark, S. J. (1997). Regulation of leukotriene A4 hydrolase activity in endothelial cells by phosphorylation. J Biol Chem 272(50): 31865‐31871.
Sadoul, K., Wang, J., Diagouraga, B. &Khochbin, S. (2011). The tale of protein lysine acetylation in the cytoplasm. J Biomed Biotechnol 2011: 970382.
Sakamoto, T. &Seiki, M. (2009). Mint3 enhances the activity of hypoxia‐inducible factor‐1 (HIF‐1) in macrophages by suppressing the activity of factor inhibiting HIF‐1. J Biol Chem 284(44): 30350‐30359.
347
Sakamoto, T. &Seiki, M. (2010). A membrane protease regulates energy production in macrophages by activating hypoxia‐inducible factor‐1 via a non‐proteolytic mechanism. J Biol Chem 285(39): 29951‐29964.
Sandau, K. B., Fandrey, J. &Brune, B. (2001). Accumulation of HIF‐1alpha under the influence of nitric oxide. Blood 97(4): 1009‐1015.
Sang, N., Stiehl, D. P., Bohensky, J., Leshchinsky, I., Srinivas, V. &Caro, J. (2003). MAPK signaling up‐regulates the activity of hypoxia‐inducible factors by its effects on p300. J Biol Chem 278(16): 14013‐14019. Epub 12003 Feb 14013.
Schaffer, L., Scheid, A., Spielmann, P., Breymann, C., Zimmermann, R., Meuli, M., Gassmann, M., Marti, H. H. &Wenger, R. H. (2003). Oxygen‐regulated expression of TGF‐beta 3, a growth factor involved in trophoblast differentiation. Placenta 24(10): 941‐950.
Schindl, M., Schoppmann, S. F., Samonigg, H., Hausmaninger, H., Kwasny, W., Gnant, M., Jakesz, R., Kubista, E., Birner, P. &Oberhuber, G. (2002). Overexpression of hypoxia‐inducible factor 1alpha is associated with an unfavorable prognosis in lymph node‐positive breast cancer. Clin Cancer Res 8(6): 1831‐1837.
Schroedl, C., McClintock, D. S., Budinger, G. R. &Chandel, N. S. (2002). Hypoxic but not anoxic stabilization of HIF‐1alpha requires mitochondrial reactive oxygen species. Am J Physiol Lung Cell Mol Physiol 283(5): L922‐931.
Scotchler, J. W. &Robinson, A. B. (1974). Deamidation of glutaminyl residues: dependence on pH, temperature, and ionic strength. Anal Biochem 59(1): 319‐322.
Selak, M. A., Armour, S. M., MacKenzie, E. D., Boulahbel, H., Watson, D. G., Mansfield, K. D., Pan, Y., Simon, M. C., Thompson, C. B. &Gottlieb, E. (2005). Succinate links TCA cycle dysfunction to oncogenesis by inhibiting HIF‐alpha prolyl hydroxylase. Cancer Cell 7(1): 77‐85.
Semenza, G. L., Roth, P. H., Fang, H. M. &Wang, G. L. (1994). Transcriptional regulation of genes encoding glycolytic enzymes by hypoxia‐inducible factor 1. J Biol Chem 269(38): 23757‐23763.
Seo, H. W., Kim, E. J., Na, H. &Lee, M. O. (2009). Transcriptional activation of hypoxia‐inducible factor‐1alpha by HDAC4 and HDAC5 involves differential recruitment of p300 and FIH‐1. FEBS Lett 583(1): 55‐60.
Shafee, N., Kaluz, S., Ru, N. &Stanbridge, E. J. (2009). PI3K/Akt activity has variable cell‐specific effects on expression of HIF target genes, CA9 and VEGF, in human cancer cell lines. Cancer Lett 282(1): 109‐115.
Shin, D. H., Li, S. H., Yang, S. W., Lee, B. L., Lee, M. K. &Park, J. W. (2009). Inhibitor of nuclear factor‐kappaB alpha derepresses hypoxia‐inducible factor‐1 during moderate hypoxia by sequestering factor inhibiting hypoxia‐inducible factor from hypoxia‐inducible factor 1alpha. FEBS J 276(13): 3470‐3480.
348
Simpson, P. T., Gale, T., Reis‐Filho, J. S., Jones, C., Parry, S., Steele, D., Cossu, A., Budroni, M., Palmieri, G. &Lakhani, S. R. (2004). Distribution and significance of 14‐3‐3sigma, a novel myoepithelial marker, in normal, benign, and malignant breast tissue. J Pathol 202(3): 274‐285.
Slaby, O., Svoboda, M., Fabian, P., Smerdova, T., Knoflickova, D., Bednarikova, M., Nenutil, R. &Vyzula, R. (2007). Altered expression of miR‐21, miR‐31, miR‐143 and miR‐145 is related to clinicopathologic features of colorectal cancer. Oncology 72(5‐6): 397‐402.
Soilleux, E. J., Turley, H., Tian, Y. M., Pugh, C. W., Gatter, K. C. &Harris, A. L. (2005). Use of novel monoclonal antibodies to determine the expression and distribution of the hypoxia regulatory factors PHD‐1, PHD‐2, PHD‐3 and FIH in normal and neoplastic human tissues. Histopathology 47(6): 602‐610.
Solstad, T. &Flatmark, T. (2000). Microheterogeneity of recombinant human phenylalanine hydroxylase as a result of nonenzymatic deamidations of labile amide containing amino acids. Effects on catalytic and stability properties. Eur J Biochem 267(20): 6302‐6310.
Sounni, N. E., Baramova, E. N., Munaut, C., Maquoi, E., Frankenne, F., Foidart, J. M. &Noel, A. (2002a). Expression of membrane type 1 matrix metalloproteinase (MT1‐MMP) in A2058 melanoma cells is associated with MMP‐2 activation and increased tumor growth and vascularization. Int J Cancer 98(1): 23‐28.
Sounni, N. E., Devy, L., Hajitou, A., Frankenne, F., Munaut, C., Gilles, C., Deroanne, C., Thompson, E. W., Foidart, J. M. &Noel, A. (2002b). MT1‐MMP expression promotes tumor growth and angiogenesis through an up‐regulation of vascular endothelial growth factor expression. Faseb J 16(6): 555‐564.
Sounni, N. E., Roghi, C., Chabottaux, V., Janssen, M., Munaut, C., Maquoi, E., Galvez, B. G., Gilles, C., Frankenne, F., Murphy, G., Foidart, J. M. &Noel, A. (2004). Up‐regulation of vascular endothelial growth factor‐A by active membrane‐type 1 matrix metalloproteinase through activation of Src‐tyrosine kinases. J Biol Chem 279(14): 13564‐13574.
Sowter, H. M., Ratcliffe, P. J., Watson, P., Greenberg, A. H. &Harris, A. L. (2001). HIF‐1‐dependent regulation of hypoxic induction of the cell death factors BNIP3 and NIX in human tumors. Cancer Res 61(18): 6669‐6673.
Stewart, A. E., Arfin, S. M. &Bradshaw, R. A. (1994). Protein NH2‐terminal asparagine deamidase. Isolation and characterization of a new enzyme. J Biol Chem 269(38): 23509‐23517.
Stolze, I. P., Tian, Y. M., Appelhoff, R. J., Turley, H., Wykoff, C. C., Gleadle, J. M. &Ratcliffe, P. J. (2004). Genetic analysis of the role of the asparaginyl hydroxylase factor inhibiting hypoxia‐inducible factor (HIF) in regulating HIF transcriptional target genes. J Biol Chem 279(41): 42719‐42725.
349
Stroka, D. M., Burkhardt, T., Desbaillets, I., Wenger, R. H., Neil, D. A., Bauer, C., Gassmann, M. &Candinas, D. (2001). HIF‐1 is expressed in normoxic tissue and displays an organ‐specific regulation under systemic hypoxia. Faseb J 15(13): 2445‐2453.
Suarez‐Pinzon, W. L., Mabley, J. G., Power, R., Szabo, C. &Rabinovitch, A. (2003). Poly (ADP‐ribose) polymerase inhibition prevents spontaneous and recurrent autoimmune diabetes in NOD mice by inducing apoptosis of islet‐infiltrating leukocytes. Diabetes 52(7): 1683‐1688.
Sun, H., Gao, J., Ferrington, D. A., Biesiada, H., Williams, T. D. &Squier, T. C. (1999). Repair of oxidized calmodulin by methionine sulfoxide reductase restores ability to activate the plasma membrane Ca‐ATPase. Biochemistry 38(1): 105‐112.
Takahashi, R., Tanaka, S., Hiyama, T., Ito, M., Kitadai, Y., Sumii, M., Haruma, K. &Chayama, K. (2003). Hypoxia‐inducible factor‐1alpha expression and angiogenesis in gastrointestinal stromal tumor of the stomach. Oncol Rep 10(4): 797‐802.
Takeda, K., Ho, V. C., Takeda, H., Duan, L. J., Nagy, A. &Fong, G. H. (2006). Placental but not heart defects are associated with elevated hypoxia‐inducible factor alpha levels in mice lacking prolyl hydroxylase domain protein 2. Mol Cell Biol 26(22): 8336‐8346.
Talks, K. L., Turley, H., Gatter, K. C., Maxwell, P. H., Pugh, C. W., Ratcliffe, P. J. &Harris, A. L. (2000). The expression and distribution of the hypoxia‐inducible factors HIF‐1alpha and HIF‐2alpha in normal human tissues, cancers, and tumor‐associated macrophages. Am J Pathol 157(2): 411‐421.
Tan, E. Y., Campo, L., Han, C., Turley, H., Pezzella, F., Gatter, K. C., Harris, A. L. &Fox, S. B. (2007). Cytoplasmic location of factor‐inhibiting hypoxia‐inducible factor is associated with an enhanced hypoxic response and a shorter survival in invasive breast cancer. Breast Cancer Res 9(6): R89.
Tan, E. Y., Yan, M., Campo, L., Han, C., Takano, E., Turley, H., Candiloro, I., Pezzella, F., Gatter, K. C., Millar, E. K., O'Toole, S. A., McNeil, C. M., Crea, P., Segara, D., Sutherland, R. L., Harris, A. L. &Fox, S. B. (2009). The key hypoxia regulated gene CAIX is upregulated in basal‐like breast tumours and is associated with resistance to chemotherapy. Br J Cancer 100(2): 405‐411.
Tanti, J. F., Grillo, S., Gremeaux, T., Coffer, P. J., Van Obberghen, E. &Le Marchand‐Brustel, Y. (1997). Potential role of protein kinase B in glucose transporter 4 translocation in adipocytes. Endocrinology 138(5): 2005‐2010.
Teicher, B. A. (1994). Hypoxia and drug resistance. Cancer Metastasis Rev 13(2): 139‐168.
Trautwein, C., Caelles, C., van der Geer, P., Hunter, T., Karin, M. &Chojkier, M. (1993). Transactivation by NF‐IL6/LAP is enhanced by phosphorylation of its activation domain. Nature 364(6437): 544‐547.
350
Truhlar, S. M., Mathes, E., Cervantes, C. F., Ghosh, G. &Komives, E. A. (2008). Pre‐folding IkappaBalpha alters control of NF‐kappaB signaling. J Mol Biol 380(1): 67‐82.
Turrens, J. F. (2003). Mitochondrial formation of reactive oxygen species. J Physiol 552(Pt 2): 335‐344.
Unruh, A., Ressel, A., Mohamed, H. G., Johnson, R. S., Nadrowitz, R., Richter, E., Katschinski, D. M. &Wenger, R. H. (2003). The hypoxia‐inducible factor‐1 alpha is a negative factor for tumor therapy. Oncogene 22(21): 3213‐3220.
Valastyan, S., Reinhardt, F., Benaich, N., Calogrias, D., Szasz, A. M., Wang, Z. C., Brock, J. E., Richardson, A. L. &Weinberg, R. A. (2009). A pleiotropically acting microRNA, miR‐31, inhibits breast cancer metastasis. Cell 137(6): 1032‐1046.
Vaupel, P. &Mayer, A. (2007). Hypoxia in cancer: significance and impact on clinical outcome. Cancer Metastasis Rev 26(2): 225‐239.
Viatour, P., Legrand‐Poels, S., van Lint, C., Warnier, M., Merville, M. P., Gielen, J., Piette, J., Bours, V. &Chariot, A. (2003). Cytoplasmic IkappaBalpha increases NF‐kappaB‐independent transcription through binding to histone deacetylase (HDAC) 1 and HDAC3. J Biol Chem 278(47): 46541‐46548.
Vilain, S., Cosette, P., Charlionet, R., Hubert, M., Lange, C., Junter, G. A. &Jouenne, T. (2001). Substituting Coomassie Brilliant Blue for bromophenol blue in two‐dimensional electrophoresis buffers improves the resolution of focusing patterns. Electrophoresis 22(20): 4368‐4374.
Visochek, L., Steingart, R. A., Vulih‐Shultzman, I., Klein, R., Priel, E., Gozes, I. &Cohen‐Armon, M. (2005). PolyADP‐ribosylation is involved in neurotrophic activity. J Neurosci 25(32): 7420‐7428.
Wang, G. L., Jiang, B. H., Rue, E. A. &Semenza, G. L. (1995). Hypoxia‐inducible factor 1 is a basic‐helix‐loop‐helix‐PAS heterodimer regulated by cellular O2 tension. Proc Natl Acad Sci U S A 92(12): 5510‐5514.
Webb, J. D., Muranyi, A., Pugh, C. W., Ratcliffe, P. J. &Coleman, M. L. (2009). MYPT1, the targeting subunit of smooth‐muscle myosin phosphatase, is a substrate for the asparaginyl hydroxylase factor inhibiting hypoxia‐inducible factor (FIH). Biochem J 420(2): 327‐333.
Weiss, W. &Gorg, A. (2008). Sample solublization buffers for two‐dimensional electrophoresis. Methods Mol Biol 424: 35‐42.
Wenger, R. H., Stiehl, D. P. &Camenisch, G. (2005). Integration of oxygen signaling at the consensus HRE. Sci STKE 2005(306): re12.
Wiener, C. M., Booth, G. &Semenza, G. L. (1996). In vivo expression of mRNAs encoding hypoxia‐inducible factor 1. Biochem Biophys Res Commun 225(2): 485‐488.
351
Wilkins, S. E., Hyvarinen, J., Chicher, J., Gorman, J. J., Peet, D. J., Bilton, R. L. &Koivunen, P. (2009). Differences in hydroxylation and binding of Notch and HIF‐1alpha demonstrate substrate selectivity for factor inhibiting HIF‐1 (FIH‐1). Int J Biochem Cell Biol 41(7): 1563‐1571.
Wong, T. S., Liu, X. B., Wong, B. Y., Ng, R. W., Yuen, A. P. &Wei, W. I. (2008). Mature miR‐184 as Potential Oncogenic microRNA of Squamous Cell Carcinoma of Tongue. Clin Cancer Res 14(9): 2588‐2592.
Yang, M., Chowdhury, R., Ge, W., Hamed, R. B., McDonough, M. A., Claridge, T. D., Kessler, B. M., Cockman, M. E., Ratcliffe, P. J. &Schofield, C. J. (2011). Factor‐inhibiting hypoxia‐inducible factor (FIH) catalyses the post‐translational hydroxylation of histidinyl residues within ankyrin repeat domains. FEBS J 278(7): 1086‐1097.
Yasinska, I. M. &Sumbayev, V. V. (2003). S‐nitrosation of Cys‐800 of HIF‐1alpha protein activates its interaction with p300 and stimulates its transcriptional activity. FEBS Lett 549(1‐3): 105‐109.
Yeh, T. Y., Beiswenger, K. K., Li, P., Bolin, K. E., Lee, R. M., Tsao, T. S., Murphy, A. N., Hevener, A. L. &Chi, N. W. (2009). Hypermetabolism, hyperphagia, and reduced adiposity in tankyrase‐deficient mice. Diabetes 58(11): 2476‐2485.
Yeh, T. Y., Sbodio, J. I., Tsun, Z. Y., Luo, B. &Chi, N. W. (2007). Insulin‐stimulated exocytosis of GLUT4 is enhanced by IRAP and its partner tankyrase. Biochem J 402(2): 279‐290.
Yu, A. Y., Frid, M. G., Shimoda, L. A., Wiener, C. M., Stenmark, K. &Semenza, G. L. (1998). Temporal, spatial, and oxygen‐regulated expression of hypoxia‐inducible factor‐1 in the lung. Am J Physiol 275(4 Pt 1): L818‐826.
Yu, F., White, S. B., Zhao, Q. &Lee, F. S. (2001). HIF‐1alpha binding to VHL is regulated by stimulus‐sensitive proline hydroxylation. Proc Natl Acad Sci U S A 98(17): 9630‐9635.
Zagzag, D., Zhong, H., Scalzitti, J. M., Laughner, E., Simons, J. W. &Semenza, G. L. (2000). Expression of hypoxia‐inducible factor 1alpha in brain tumors: association with angiogenesis, invasion, and progression. Cancer 88(11): 2606‐2618.
Zhang, N., Fu, Z., Linke, S., Chicher, J., Gorman, J. J., Visk, D., Haddad, G. G., Poellinger, L., Peet, D. J., Powell, F. &Johnson, R. S. (2010a). The asparaginyl hydroxylase factor inhibiting HIF‐1alpha is an essential regulator of metabolism. Cell Metab 11(5): 364‐378.
Zhang, Y., Guo, J., Li, D., Xiao, B., Miao, Y., Jiang, Z. &Zhuo, H. (2010b). Down‐regulation of miR‐31 expression in gastric cancer tissues and its clinical significance. Med Oncol 27(3): 685‐689.
Zhao, X., Gan, L., Pan, H., Kan, D., Majeski, M., Adam, S. A. &Unterman, T. G. (2004). Multiple elements regulate nuclear/cytoplasmic shuttling of FOXO1: characterization of phosphorylation‐ and 14‐3‐3‐dependent and ‐independent mechanisms. Biochem J 378(Pt 3): 839‐849.
352
Zheng, X., Linke, S., Dias, J. M., Gradin, K., Wallis, T. P., Hamilton, B. R., Gustafsson, M., Ruas, J. L., Wilkins, S., Bilton, R. L., Brismar, K., Whitelaw, M. L., Pereira, T., Gorman, J. J., Ericson, J., Peet, D. J., Lendahl, U. &Poellinger, L. (2008). Interaction with factor inhibiting HIF‐1 defines an additional mode of cross‐coupling between the Notch and hypoxia signaling pathways. Proc Natl Acad Sci U S A 105(9): 3368‐3373.
Zhong, H., De Marzo, A. M., Laughner, E., Lim, M., Hilton, D. A., Zagzag, D., Buechler, P., Isaacs, W. B., Semenza, G. L. &Simons, J. W. (1999). Overexpression of hypoxia‐inducible factor 1alpha in common human cancers and their metastases. Cancer Res 59(22): 5830‐5835.
Zundel, W., Schindler, C., Haas‐Kogan, D., Koong, A., Kaper, F., Chen, E., Gottschalk, A. R., Ryan, H. E., Johnson, R. S., Jefferson, A. B., Stokoe, D. &Giaccia, A. J. (2000). Loss of PTEN facilitates HIF‐1‐mediated gene expression. Genes Dev 14(4): 391‐396.
ADDENDA
1. On page 9 references should be listed for Figure 1.1 as follows. VEGF: Forsythe et
al., 1996; VEGF‐R: Gerber et al., 1997; MMP‐2 and AMF: Krishnamachary et al.,
2003; EPO: Goldberg et al., 1988; CA9: Grabmaier et al., 2004; Glut1 and Glut3:
Ebert et al., 1995; Aldolase A, LDHA and PGK1: Semenza et al., 1994; TGF‐β:
Schaffer et al., 2003; IGF‐2: Feldser et al., 1999; BNip3: Sowter et al., 2001
2. On page 6, references should be added for HIF‐β, HIF‐1 and HIF‐2 as follows. HIF‐
β: Hoffman et al., 1991 ; HIF‐1: Wang and Semenza, 1995; HIF‐2: Ema et al., 1997
and Tian et al., 1997.
3. On page 11, the word “at” should be deleted from the last line of the second
paragraph to read “regulated by HIF binding to HREs”.
4. On page 30, the word “the” should be included in the third line of the second
paragraph of section 1.3.3 to read “was required for the activation of the
reporter gene”.
5. On page 52, the word “the” should be included in the first line of the second
paragraph of section 1.7 to read “It was therefore the original intent of this
thesis...”.
6. On page 53, the word “of” should be included in the second to last sentence to
read “...followed by a discussion of relevant data...” .
7. On page 67, it should be specified that agarose gels were run in TBE buffer.
8. On page 91, the word “to” should be removed from the last sentence of the
second paragraph to read “It was thus deemed unlikely that these antibodies
would be suitable for use...”.
9. On page 132, the word “in” should be replaced by the word “is” in the last line to
read “...expression of HIF target genes is not clear”.
10. On page 140, a full stop should be included at the end of the first sentence of
section 4.2.3.2, following the word buffer.
11. On page 161, the word “charge” should be amended to “charged” in the last line
to read “...positively charged amino group...”.
12. On page 195, lane 9 of Figure 5.1 should be labelled “Lysate only”.
13. On page 213, the word “of” should be included in the first line of section 5.4.1 to
read “Purifying a sufficient amount of endogenous FIH‐1...”.
14. On pages 217 and 227, Figures 4.12 b, c and d should be referenced when
comparing present results with previous results.
15. On page 249, the word “the” should be deleted from the sixth line to read
“...optimisation included altering both the amount and concentration...”.
16. On page 258, the word “in” should be removed from the sixth last line to read “...
was not detected by MS...”.
17. On page 262, the word “whether” should be included in the last sentence to read
“...it was of interest to investigate whether FIH‐1 was subjected to this important
modification”.
18. On page 266, the words “in vitro” should be italicised on the fourth line of the
second paragraph.
19. On page 267, the word “was” should be inserted in the last sentence of the
second paragraph to read “...protein phosphatase lambda was used...”.
20. On page 283, the word “in” should be replaced with the word “is” in the first
sentence of section 6.3.1 to read “... a protein is phosphorylated”.
21. Page 307 was unintentionally left blank due to a formatting error.
22. In the second paragraph on page 313, the word “to” should be removed from the
fourth line to read “...then binding to Mint3”; the word “of” should be inserted
into the sixth line to read “...regardless of oxygen levels...”; the word “be” should
be included in the ninth line to read “...it may also be a normal physiological
process...”.
23. The following reference should be included in the Reference (page 335):
Ema, M., Taya, S., Yokotani, N., Sogawa, K., Matsuda, Y. & Fujii‐Kuriyama, Y. (1997). A
novel bHLH‐PAS factor with close sequence similarity to hypoxia‐inducible factor
1alpha regulates the VEGF expression and is potentially involved in lung and vascular
development. Proc Natl Acad Sci USA 94(9): 4273‐4278.
Grabmaier, K., MC, A de Weijert, M.C., Verhaegh, G.W., Schalken, J.A. & Oosterwijk,
E. (2004). Strict regulation of CAIX(G250/MN) by HIF‐1alpha in clear cell renal cell
carcinoma. Oncogene 23(33):5624‐5631.
Hoffman, E. C., Reyes, H., Chu, F. F., Sander, F., Conley, L. H., Brooks, B. A. &
Hankinson, O. (1991). Cloning of a factor required for activity of the Ah (dioxin)
receptor. Science 252(5008): 954‐958.
Lisy, K. & Peet, D.J. (2008). Turn me on: regulating HIF transcriptional activity. Cell
Death Differ 15(4):642‐649.
Tian, H., McKnight, S. L. & Russell, D. W. (1997). Endothelial PAS domain protein 1
(EPAS1), a transcription factor selectively expressed in endothelial cells. Genes Dev
11(1): 72‐82.
Wang, G. L. & Semenza, G. L. (1995). Purification and characterization of hypoxia‐
inducible factor 1. J Biol Chem 270(3): 1230‐1237.