1. No category

CHARACTERIZATION OF THE NOVEL
HUMAN PROLYL 4-HYDROXYLASES
AND ASPARAGINYL HYDROXYLASE
THAT MODIFY THE HYPOXIAINDUCIBLE FACTOR
MAI JA
H I R SI L Ä
Faculty of Medicine,
Department of Medical Biochemistry
and Molecular Biology;
Biocenter Oulu,
University of Oulu
OULU 2004
MAIJA HIRSILÄ
CHARACTERIZATION OF THE NOVEL
HUMAN PROLYL 4-HYDROXYLASES
AND ASPARAGINYL HYDROXYLASE
THAT MODIFY THE HYPOXIAINDUCIBLE FACTOR
Academic Dissertation to be presented with the assent of
the Faculty of Medicine, University of Oulu, for public
discussion in the Auditorium of the Department of
Pharmacology and Toxicology, on December 3rd, 2004,
at 10 a.m.
O U L U N Y L I O P I S TO, O U L U 2 0 0 4
Copyright © 2004
University of Oulu, 2004
Supervised by
Docent Johanna Myllyharju
Professor Kari Kivirikko
Reviewed by
Doctor Eric Huang
Doctor Eric Metzen
ISBN 951-42-7574-8 (nid.)
ISBN 951-42-7575-6 (PDF) http://herkules.oulu.fi/isbn9514275756/
ISSN 0355-3221
OULU UNIVERSITY PRESS
OULU 2004
http://herkules.oulu.fi/issn03553221/
Hirsilä, Maija, Characterization of the novel human prolyl 4-hydroxylases and
asparaginyl hydroxylase that modify the hypoxia-inducible factor
Faculty of Medicine, Department of Medical Biochemistry and Molecular Biology, Biocenter Oulu,
University of Oulu, P.O.Box 5000, FIN-90014 University of Oulu, Finland
2004
Oulu, Finland
Abstract
HIF prolyl 4-hydroxylases (HIF-P4Hs) and HIF asparaginyl hydroxylase (FIH) are novel members
of the 2-oxoglutarate dioxygenase family that play key roles in the regulation of the hypoxiainducible transcription factor (HIF). They hydroxylate specific proline and asparagine residues in
HIF-α, leading to its proteasomal degradation and inhibition of its transcriptional activity,
respectively. These enzymes are inhibited in hypoxia, and as a consequence HIF-α becomes
stabilized, forms a dimer with HIF-β, attains its maximal transcriptional activity and induces
expression of many genes that are important for cell survival under hypoxic conditions.
The three HIF-P4Hs and FIH were expressed here as recombinant proteins in insect cells and
purified to near homogeneity. All these enzymes were found to require long peptide substrates. The
three HIF-P4Hs and FIH acted differently on the various potential hydroxylation sites in the HIF-α
isoforms. The HIF-P4Hs acted well on sequences with cores distinctly different from the core motif
-Leu-X-X-Leu-Ala-Pro-, suggested based on sequence analysis studies, the alanine being the only
relatively strict requirement in addition to the proline itself. Acidic residues around the hydroxylation
site also played a distinct role. These results together with those of others provide evidence that there
is no conserved core motif for the hydroxylation by HIF-P4Hs.
The Km values of the HIF-P4Hs for O2 were slightly above its atmospheric concentration, while
the Km of FIH was about one-third of these values but still 2.5 times that of the type I collagen P4H.
The HIF-P4Hs are thus effective oxygen sensors, as even small decreases in the amount of O2 affect
their activities, while a more severe decrease is required to inhibit FIH activity. Small molecule
inhibitors of the collagen P4Hs also inhibited the HIF-P4Hs and FIH but with distinctly different Ki
values, indicating that it should be possible to develop specific inhibitors for the HIF-P4Hs and FIH.
The HIF-P4Hs were found to bind the iron cosubstrate more tightly than FIH and the collagen
P4Hs, and the chelator desferrioxamine was an ineffective inhibitor of the HIF-P4Hs in vitro. Several
metals were effective competitive inhibitors of FIH but they were ineffective inhibitors of the HIFP4Hs. The well-known stabilization of HIF-1α by cobalt and nickel is thus not due to a simple
competitive inhibition of the HIF-P4Hs, and is probably at least in part due to HIF-P4H-independent
mechanisms. The effective inhibition of FIH by these metals nevertheless indicates that the stabilized
HIF-1α is transcriptionally fully active.
Keywords: 2-oxoglutarate dioxygenase, asparaginyl hydroxylase, enzyme inhibitors,
oxygen homeostasis, prolyl 4-hydroxylase
Acknowledgements
This work was carried out at the Collagen Research Unit of the Department of Medical
Biochemistry and Molecular Biology, University of Oulu, and Biocenter Oulu during the
years 2001-2004.
I wish to express my deepest gratitude to my supervisor, Docent Johanna Myllyharju
for her excellent supervision and advice throughout this study. It has been a pleasure and
priviledge to work in her group. I am most grateful to my other supervisor, Professor Kari
Kivirikko for his broad knowledge and optimism even during the difficult moments of
these projects. I want to thank Professor Taina Pihlajaniemi for providing excellent
research facilities and inspiring scientific atmosphere in the department. I owe my sincere
thanks to Professor Peppi Karppinen for her knowledge, positive attitude and friendship.
It was a pleasure to share all the projects with her. Professor Ilmo Hassinen deserves my
warmest thanks for his advice, especially in the beginning of these projects. I wish to
express my respect to Professor Leena Ala-Kokko for her distinguished scientific work.
I would like to thank Dr. Eric Huang and Dr. Eric Metzen for their valuable comments
on this thesis. I also acknowledge Malcolm Hicks for his revision of language of the
thesis and the original articles.
I wish to thank Dongxia Li, Mitchell Brenner, Charles Yang, Leon Xu and Todd
Seeley for their efforts on these projects and I thank Stephen Klaus for his advice with the
cell culture. Michael Arend is acknowledged for his help in synthesizing some of the
compounds. I warmly thank Docent Volkmar Günzler for his efforts and valuable advice.
The staff of the Department is acknowledged, especially Auli Kinnunen and MarjaLeena Kivelä for their secretarial assistance, and Pertti Vuokila, Marja-Leena Karjalainen
and Seppo Lähdesmäki for their help in many practical matters. I warmly thank the
people helping us with the computers, especially Risto Helminen for his fast and skillfull
help. I am grateful to Riitta Jokela and Anne Kokko for their excellent technical
assistance in the daily labwork. Eeva Lehtimäki, Raija Juntunen and Tanja Aatsinki are
also acknowledged.
My warmest thanks go to Katriina Keskiaho, Antje Neubauer and Päivi Tiainen who
shared the office with me. Because of you it was always nice to come to work. We had
discussions concerning scientific problems and especially besides science and we had
nice moments also outside the lab. I thank all our group members for greating a good
spirit in our group. Especially I thank Outi Pakkanen for giving me the hint that our
group might need new researchers when I was still wondering my future in the previous
job.
I warmly thank Anne Latvanlehto and Harri Elamaa for sharing the organizing process
of the dissertation and the party.
I owe my deepest gratitude to my parents, Marja-Liisa and Taisto, for their love and
support throughout my life. My brother Pekka deserves a big hug for just being my
brother. Finally, I owe my loving thanks to my husband Jari Mäki-Runsas for his love and
encouragement throughout these years together, on my way from high school to PhD. He
has been my “leisure activities instructor” taking me to have holidays both in South and
North. Without him I would have been married only to my work.
This work was supported by the Health Science Council and the Finnish Centre of
Excellence Programme 2002-2005 of the Academy of Finland, Sigrid Juselius
Foundation, Emil Aaltonen Foundation, Oulu University Scholarship Foundation,
Biocenter Oulu and FibroGen Inc.
Oulu, August 2004
Maija Hirsilä
Abbreviations
AAH
ARD1
ARNT
bHLH
BSA
CAD
CoA
cDNA
C-P4H
DFO
EGF
EPO
ER
FIH
Flp
HIF
HIF-P4H
HRE
ID
IGF
IPAS
IUGR
Ki
Km
mRNA
NAD
ODDD
PAS
PBCV-1
PDI
PER
aspartyl (asparaginyl) β hydroxylase
HIF-α acetyltransferase
aryl hydrocarbon nuclear translocator
basic helix-loop-helix
bovine serum albumin
C-terminal transactivation domain of HIF-α
coenzyme A
complementary DNA
collagen prolyl 4-hydroxylase
desferrioxamin
epidermal growth factor
erythropoietin
endoplasmic reticulum
factor inhibiting HIF
fluoroproline
hypoxia-inducible factor
HIF prolyl 4-hydroxylase
hypoxia response element
inhibitory domain
insulin-like growth factor
inhibitory PAS
intrauterine growth retardation
inhibitory constant
Michaelis-Menten constant
messenger RNA
N-terminal transactivation domain of HIF-α
oxygen-dependent degradation domain
Per-Arnt-Sim
Paramecium bursaria Chlorella virus-1
protein disulphide isomerase
Drosophila periodic clock protein
PHY
pVHL
Rbx1
Rpb
ROS
SIM
Tm
VEGF
Vmax
X (in -Gly-X-Y-)
Y (in -Gly-X-Y-)
C. elegans collagen prolyl 4-hydroxylase α subunit
von Hippel-Lindau tumour-suppressor protein
RING-box protein
subunit of RNA polymerase
reactive oxygen species
Drosophila single-minded protein
midpoint of thermal transition
vascular endothelial growth factor
maximum initial velocity
any amino acid
any amino acid
List of original articles
This thesis is based on the following articles, which are referred to in the text by their
Roman numerals:
I
Hirsilä M, Koivunen P, Günzler V, Kivirikko KI & Myllyharju J (2003)
Characterization of the human prolyl 4-hydroxylases that modify the hypoxiainducible factor. J Biol Chem 278: 30772-30780.
II
Koivunen P, Hirsilä M, Günzler V, Kivirikko KI & 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: 9899-9904.
III Li D, Hirsilä M, Koivunen P, Brenner MC, Xu L, Yang C, Kivirikko KI &
Myllyharju J (2004) Many amino acid substitutions in a HIF-1α-like peptide cause
only minor changes in its hydroxylation by the HIF prolyl 4-hydroxylases.
Substitution of 3,4-dehydroproline or azetidine-2-carboxylic acid for the proline
leads to a high rate of uncoupled 2-oxoglutarate decarboxylation. J Biol Chem, in
press.
IV Hirsilä M, Koivunen P, Xu L, Seeley T, Kivirikko KI & Myllyharju J (2004) Effect
of metals on the hydroxylases in the oxygen sensing pathway. Cobalt is only a weak
inhibitor of HIF prolyl 4-hydroxylases, while cobalt and zinc are effective inhibitors
of the asparaginyl hydroxylase. Submitted
The articles published in the Journal of Biological Chemistry (2003; 278: 30772-30780,
2004; 279: 9899-9904 and 2004; Li et al., in press) are reproduced by permission of the
copyright owner, the American Society for Biochemistry and Molecular Biology.
Contents
Abstract
Acknowledgements
Abbreviations
List of original articles
Contents
1 Introduction ...................................................................................................................15
2 Review of the literature .................................................................................................17
2.1 Hypoxia-inducible factor ........................................................................................17
2.1.1 Function and tissue distribution.......................................................................17
2.1.2 Knock-out mouse models ................................................................................19
2.1.3 Structure ..........................................................................................................21
2.1.4 Regulation .......................................................................................................23
2.1.5 Effect of metal ions on HIF-1..........................................................................26
2.1.6 Role of HIF in disease .....................................................................................28
2.2 Prolyl 4-hydroxylases .............................................................................................30
2.2.1 Occurrence and functions of 4-hydroxyproline in proteins other than HIF.....30
2.2.2 Vertebrate collagen prolyl 4-hydroxylases.......................................................32
2.2.3 Nematode collagen prolyl 4-hydroxylases ......................................................34
2.2.4 Prolyl 4-hydroxylases from other organisms...................................................36
2.2.5 Catalytic properties..........................................................................................37
2.2.6 Reaction mechanism........................................................................................38
2.2.7 Human HIF prolyl 4-hydroxylases ..................................................................40
2.2.7.1 Biological functions..................................................................................40
2.2.7.2 Gene expression and regulation................................................................41
2.2.7.3 Molecular properties and cellular location ...............................................42
2.2.7.4 Catalytic properties...................................................................................43
2.2.8 Nematode and Drosophila HIF-P4Hs..............................................................44
2.3 HIF asparaginyl hydroxylase..................................................................................45
2.3.1 Biological function ..........................................................................................45
2.3.2 Structural properties ........................................................................................45
2.3.3 Catalytic properties and substrate specificity ..................................................46
2.4 Aspartyl (asparaginyl) β hydroxylase .....................................................................47
3 Outlines of the present research.....................................................................................49
4 Materials and methods...................................................................................................50
4.1 Cloning and expression of recombinant HIF-P4Hs and FIH (I-IV) .......................50
4.1.1 Cloning and generation of expression vectors.................................................50
4.1.2 Expression of the recombinant enzymes .........................................................51
4.1.3 Purification of the recombinant enzymes ........................................................51
4.2 Characterization of the catalytic and inhibitory properties of HIF-P4Hs
and FIH (I-IV) .......................................................................................................52
4.3 Liquid chromatography/mass spectrometry analysis of peptide
hydroxylation (III) .................................................................................................53
4.4 Stabilization of HIF-1α in human cell lines (IV)....................................................53
4.5 Analysis of VEGF production (IV).........................................................................53
4.6 PCR analysis of HIF-P4H mRNA expression (I) ...................................................54
5 Results ...........................................................................................................................55
5.1 Cloning of the HIF-P4H 1- 3 and FIH cDNAs (I, II) .............................................55
5.2 Recombinant expression of the HIF-P4Hs and FIH (I, II)......................................55
5.3 Purification of recombinant HIF-P4Hs and FIH (I, II, IV).....................................56
5.4 Optimization of activity assays for the HIF-P4Hs 1-3 and FIH (I, II, IV)..............56
5.5 Catalytic centre activities of HIF-P4Hs and FIH (I, II, IV) ....................................57
5.6 The 2-oxoglutarate-binding arginine of HIF-P4Hs cannot be replaced
by a lysine (I).........................................................................................................58
5.7 Hydroxylation of peptides of varying lengths and representing different
HIF-α isoforms by the HIF-P4Hs and FIH (I, II)..................................................58
5.8 Effects of amino acid substitutions in a HIF-1α peptide on the hydroxylation
by the HIF-P4Hs (III) ............................................................................................60
5.8.1 Substitution of the Leu-Glu-Met-Leu-Ala residues preceding the
hydroxylatable Pro564........................................................................................60
5.8.2 Substitutions of alanine for acidic HIF-1α residues ........................................62
5.8.3 Substitution of 3,4-dehydroproline or L-azetidine-2-carboxylic acid
for Pro564 ...........................................................................................................63
5.9 Km values of HIF-P4Hs and FIH for the cosubstrates (I, II, IV).............................63
5.10 Inhibition of the HIF-P4Hs and FIH by 2-oxoglutarate analogues (I, II) .............65
5.11 Effect of desferrioxamine and metals on the HIF-P4Hs and FIH (IV) .................66
5.11.1 Effect of desferrioxamine on the activities of the HIF-P4Hs.........................66
5.11.2 Inhibition of the HIF-P4Hs and FIH by metals..............................................67
5.11.3 Effect of cobalt and DFO on the synthesis of HIF-P4H-2 and
in vitro activity of the synthesized enzyme.......................................................68
5.11.4 Effect of cobalt and iron on the stability of HIF-P4Hs 2 and 3......................68
5.11.5 HIF-1α stabilization in human HEK293 and Hep3B cells ............................68
5.11.6 Effect of cobalt, nickel and zinc on the production of VEGF by human
HEK293 and Hep3B cells.................................................................................69
5.12 Analysis of the expression of mRNAs for the three human
HIF-P4H isoenzymes (I).......................................................................................69
6 Discussion .....................................................................................................................71
6.1 Catalytic centre activities of HIF-P4Hs and FIH....................................................71
6.2 Substrate requirements of the HIF-P4Hs and FIH..................................................72
6.3 Cosubstrate requirements of the HIF-P4Hs and FIH ..............................................75
6.4 Inhibition of the HIF-P4Hs and FIH by 2-oxoglutarate analogues.........................77
6.5 Effect of desferrioxamine and metals on the HIF-P4Hs and FIH ...........................77
6.6 Expression and alternative splicing of HIF-P4H mRNAs ......................................79
7 Future prospects.............................................................................................................80
References
Original articles
1 Introduction
The hypoxia-inducible factor HIF is a transcriptional complex that plays a central role in
the regulation of oxygen homeostasis. It is composed of two subunits: a hypoxiaregulated α subunit with three isoforms in humans and an oxygen-insensitive β subunit.
Under normal oxygen conditions the HIF-α subunit is rapidly degraded in proteasomes
after its expression due to post-translational hydroxylation of at least one of two
conserved proline residues within a polypeptide segment termed the oxygen-dependent
degradation domain (ODDD), the two proline residues in HIF-1α being present in -LeuX-X-Leu-Ala-Pro-Ala- and -Leu-X-X-Leu-Ala-Pro-Tyr- sequences. Under hypoxic
conditions the hydroxylation and degradation of the α subunit are blocked, allowing HIFα to accumulate in the nucleus and form a complex with HIF-β, leading to the activation
of hypoxia-inducible genes. A novel enzyme family, the HIF prolyl 4-hydroxylases (HIFP4Hs), with three isoforms in humans, has recently been identified and found to function
in the hydroxylation of HIF-α. The HIF-P4Hs belong to the family of 2-oxoglutarate
dioxygenases and require Fe2+, 2-oxoglutarate, O2 and ascorbate in their reaction.
Transcriptional activation in an oxygen-dependent manner is another key event that
regulates the activity of HIF. The enzyme responsible for this regulation is a HIF-specific
asparaginyl hydroxylase that is identical to a previously identified factor inhibiting HIF
(FIH). Under normoxia this enzyme inhibits the full transactivation function of HIF by
hydroxylating a specific asparagine residue in HIF-α. FIH is likewise a 2-oxoglutarate
dioxygenase and its reaction thus also requires Fe2+, 2-oxoglutarate, O2 and ascorbate. In
hypoxia the asparagine residue remains unhydroxylated, which allows binding of the
transcriptional coactivator p300.
Some divalent metal ions such as Co2+ and Ni2+ are known to stabilize HIF-α and lead
to increased expression of HIF target genes, e.g. those for erythropoietin (EPO) and
vascular endothelial growth factor (VEGF). Although the mechanisms involved in the
HIF-α stabilization caused by metal ions are not fully understood, one possible
mechanism is inhibition of the hydroxylases of the oxygen sensing pathway, as several
divalent metal ions are known to inhibit the collagen prolyl 4-hydroxylases (C-P4Hs) and
lysyl hydroxylases that belong to the same family of 2-oxoglutarate dioxygenases as the
HIF-P4Hs and FIH. On the other hand, evidence has been presented indicating that cobalt
and nickel may prevent degradation of HIF-α by becoming directly bound to this
polypeptide.
The aims of the present study were to amplify the cDNAs of the three human HIFP4H isoenzymes and FIH and express them as recombinant proteins in insect cells.
Simple assays were developed to measure their activities, with the aim of studying the
catalytic properties of these novel enzymes and their inhibition with several 2oxoglutarate analogues known to inhibit the C-P4Hs. The sequence requirements of the
three human HIF-P4Hs were studied by systematically varying each residue preceding
the hydroxylatable proline. An additional goal was to determine the effects of several
divalent metal ions on the activities of the recombinant HIF-P4Hs and FIH and their
possible effects on the stability of HIF-α and the expression of VEGF, one of the HIF
target genes, in cultured cells.
2 Review of the literature
2.1 Hypoxia-inducible factor
Oxygen homeostasis represents an important challenge for the development and
physiology of both vertebrate and invertebrate organisms. Direct diffusion of O2 is
sufficient for oxygenation in simple invertebrates, but the large body size of humans and
other vertebrates has led to the development of a complex infrastructure for O2 supply.
The regulation of these different systems provides the basis for O2 homeostasis. Cellular
O2 concentrations are precisely controlled by the coordinated regulation of a variety of
genes to maintain adequate substrates for oxidative phosphorylation and other essential
metabolic reactions. The main regulator of O2 homeostasis in the animal kingdom is the
transcriptional activator HIF.
2.1.1 Function and tissue distribution
HIF-1 is expressed in response to lowered oxygen concentration (hypoxia) in most cell
types (Wenger et al. 1996, Wiener et al. 1996) and it activates the transcription of genes
that are crucial for cell survival under hypoxic conditions (Wang et al. 1995). HIF-1 was
first shown to be the key hypoxic regulator of EPO (Semenza & Wang 1992, Wang &
Semenza 1995), but it has subsequently been found to induce expression of many
proteins involved in the control of glucose metabolism, cell proliferation and survival and
O2 transport and delivery (Figure 1). The number of known HIF-1 target genes is growing
continuously.
HIF-1 recognizes its target genes by binding to the hypoxia response element (HRE),
which is a sequence of less than 100 residues including one or more HIF-1 binding sites
containing the core sequence 5’-RCGTG-3’. Functional HRE consists of a pair of
contiguous transcription factor binding sites, one of which is bound by HIF-1, and it is
possible that these two sites may function cooperatively at the level of either DNA
binding or transactivation. (Semenza et al. 1996.)
18
Fig. 1. Target genes of HIF-1. The numbers in the figure are references to: 1. Ebert et al.
1995, 2. Chen et al. 2001, 3. Zelzer et al. 1998, 4. O’Rourke et al. 1996, 5. Semenza et al. 1996,
6. Semenza et al. 1994, 7. Mathupala et al. 2001, 8. Riddle et al. 2000, 9. Firth et al. 1994, 10.
Minchenko et al. 2002, 11. Graven et al. 1999, 12. Wykoff et al. 2000, 13. Feldser et al. 1999,
14. Tazuke et al. 1998, 15. Zaman et al. 1999, 16. Bruick 2000, 17. Caniggia et al. 2000, 18.
Semenza & Wang 1992, 19. Jelkmann 1992, 20. Wang & Semenza 1993, 21. Rolfs et al. 1997,
22. Lok & Ponka 1999, 23. Bianchi et al. 1999, 24. Tacchini et al. 1999, 25. Mukhopadhyay et
al. 2000, 26. Norris & Millhorn 1995, 27. Levy et al. 1995, 28. Liu et al. 1995, 29. Forsythe et
al. 1996, 30. Gerber et al. 1997, 31. Eckhart et al. 1997, 32. Lee et al. 1997, 33. Melillo et al.
1995, 34. Palmer et al. 1998, 35. Hu et al. 1998, 36. Minchenko & Caro 2000, 37. Kietzmann et
al. 1999, 38. Cormier-Regard et al. 1998, 39. Nguyen & Claycomb 1999, 40. Furuta et al. 2001,
and 41. Ambrosini et al. 2002.
Two other HIF-α subunits in addition to HIF-1α have been characterized. HIF-2α, which
was identified and cloned shortly after the cloning of HIF-1α (Tian et al. 1997, Ema et al.
1997, Flamme et al. 1997, Hogenesch et al. 1997), shares several target genes with the
latter but does not have a role in the up-regulation of glycolytic genes, and no genes
uniquely regulated by HIF-2α have been identified yet (Hu et al. 2003). Studying the
homozygous hif2a knock-out mice has revealed a syndrome of multiple organ pathology,
metabolic abnormalities, increase in the amount reactive oxygen species (ROS) and
reduced expression of genes encoding the antioxidant enzymes (AOE), however,
suggesting that HIF-2α regulates the AOE gene expression (Scortegagna et al. 2003). The
third α subunit isoform, HIF-3α, was first identified in the mouse (Gu et al. 1998), the
human HIF-3α being identified in 2001 (Hara et al. 2001). All the HIF-α subunit
isoforms are members of the same basic-helix-loop-helix (bHLH)-Per-Arnt-Sim (PAS)
protein superfamily, and share a number of structural and biochemical similarities,
19
although little is known about the expression and functions of HIF-3α. Transfected HIF3α has been shown to suppress the hypoxia-driven expression of target genes when the
expression level of HIF-β is limited, which may be attributable to competition between
the three HIF-α isoforms (Hara et al. 2001). It has also been suggested that HIF-3α may
function as a negative regulator of hypoxia-induced gene expression (Hara et al. 2001,
Makino et al. 2002). An alternatively spliced form of HIF-3α, IPAS (inhibitory PAS
protein), plays a role as negative regulator of gene expression in the Purkinje cells of the
cerebellum and in the corneal epithelium of the eye under hypoxic conditions, forming
heterodimers with HIF-αs and resulting in the failure of HIFs to bind to HREs (Makino et
al. 2001, 2002). This IPAS function is important for the avascular phenotype of the
cornea in the eye.
The expression patterns of the three HIF-α isoforms are different. Northern blot
analyses have revealed that each HIF-α mRNA has a unique tissue-specific distribution,
with HIF-1α expression being highest in the kidney and heart, HIF-2α expression in the
placenta, lung, heart and liver, and HIF-3α expression in the skeletal muscle, adult
thymus, lung, brain, heart and kidney (Ema et al. 1997, Hogenesch et al. 1997, Gu et al.
1998). HIF-2α seems to have a significant role in lung development, expression of its
mRNA being markedly enhanced in this tissue, whereas mRNA expression of HIF-1α
remains at a low level (Ema et al. 1997). Interestingly, HIF-2α expression is parallelled
by the expression of VEGF mRNA, suggesting that the latter may also, or even
predominantly, be regulated by HIF-2α (Ema et al. 1997). HIF-1α is expressed at
constant level during placental development, whereas expression of the HIF-2α mRNA
increases during gestational age (Rajakumar & Conrad 2000). There is an increased level
of vascularization with advancing gestation, which is in agreement with the high level of
HIF-2α mRNA expression in endothelial cells (Ema et al. 1997, Tian et al. 1997). In
tissues where both HIF-1α and HIF-2α are expressed, this occurs mostly in different cell
types, so that in the brain, for example, HIF-1α is expressed in neuronal cells and HIF-2α
in non-parenchymal cells (Wiesener et al. 2003). HIF-1α and HIF-2α also have distinct
functions in different types of cancer and differ according to the cell type, suggesting that
these differences are important for tumour development (Sowter et al. 2003). In a breast
carcinoma cell line, for example, HIF-1α was required for the hypoxic induction of many
target genes and HIF-2α could not be substituted for it in the regulation of any of the
studied genes when HIF-1α was inactivated. In contrast, HIF-2α was the major HIF-α
form in a renal carcinoma cell line.
2.1.2 Knock-out mouse models
Knock-out mouse models have been used to study the roles of HIF-1α and HIF-2α both
in embryonic development and in postnatal physiology.
Homozygous hif1a knock-out mice have severe defects in their embryonic
development leading to embryonic death at mid-gestation on account of major
malformations of the heart and vasculature (Iyer et al. 1998, Ryan et al. 1998, Kotch et
al. 1999). Complete loss of HIF-1α results in developmental arrest, myocardial
hyperplasia, regression and dilation of vessels, especially in the head region, and marked
20
cell death within the cephalic mesenchyme related to defects in neural tube closure (Iyer
et al. 1998, Ryan et al. 1998, Kotch et al. 1999). HIF-1 is thus a highly important
determinant of embryonic morphogenesis, especially vascularization. The vascular
defects observed in hif1a-/- mice could be assumed to be due to a deficiency of VEGF
caused by the absence of HIF-1α expression, but surprisingly, elevated VEGF mRNA
expression was demonstrated in hif1a-/- embryos, and instead the vascular defects were
found to correlate with mesenchymal cell death (Kotch et al. 1999). It is not clear,
however, whether the loss of endothelial cells is a primary or secondary event in the
pathogenesis of vascular regression. Conditional HIF-1α knock-out studies with mouse
myeloid cells showed that HIF-1α is essential for myeloid cell function in vitro and
inflammation responses in vivo (Cramer et al. 2003). Inflammatory responses mediated
by myeloid cells were decreased in HIF-1α deficient cells suggesting that HIF-1α
regulates glycolysis and energy metabolism in myeloid cells. HIF-1α is required for the
infiltration, oedema formation and tissue destruction caused by granulocytes and
macrophages in inflammatory responses (Cramer et al. 2003). Another study using
conditional inactivation of the hif1a gene in avascular tissues showed that mice lacking
HIF-1α are smaller than their wild-type littermates and die soon after birth (Schipani et
al. 2001). The knock-out mice had tracheal and skeletal abnormalities with massive death
of chondrocytes, thus suggesting that HIF-1α is required for chondrocyte growth and
survival (Schipani et al. 2001).
Homozygous hif2a knock-out mice have variable defects in the development of their
cardiovascular and respiratory systems and in several other organs (Tian et al. 1998, Peng
et al. 2000, Compernolle et al. 2002, Scortegagna et al. 2003). Vasculogenesis is normal
in these mice, but severe defects were detected in embryonic vascular remodelling, with
vessels fusing improperly or failing to assemble, leading to embryonic death (Peng et al.
2000). Normal cardiovascular function, in both embryonic and neonatal development,
requires catecholamines, and hif2a knock-out mice have been shown to have lowered
catecholamine levels (Tian et al. 1998). This leads to a slower heart beat and foetal death
caused by circulatory failure, indicating an important role for HIF-2α in the regulation of
catecholamine homeostasis and protection from lethal heart failure and defects in
postvasculogenetic remodelling (Tian et al. 1998, Peng et al. 2000). In another knock-out
model, neonate homozygous mice were shown to have severe respiratory failure and die
soon after birth due to lung collapse (Compernolle et al. 2002). The expression of VEGF
in alveolar cells was shown to be reduced, with impaired lung maturation and defective
surfactant production, suggesting that VEGF, and also HIF-2α, by regulating VEGF
expression, must play an important role in the regulation of lung maturation by increasing
the conversion of glygocen stores to surfactants. This agrees with high levels of HIF-2α
expression in the lungs and the suggestion that VEGF expression may actually be
regulated by HIF-2α (Ema et al. 1997). Pathological abnormalities have also been
reported in the eye, skeletal muscle, heart, liver, testis and bone marrow, of homozygous
hif2a knock-out mice, these tissues being sites of high energy demand and sensitive to
high oxidative stress (Scortegagna et al. 2003). Abnormalities were also found in the
mitochondria, the observations suggesting a systemic disorder due to a mitochondrial
disease. In agreement with this suggestion, the mice have high lactate and low glucose
levels and abnormalities in the functions of the citric acid cycle and fatty acid oxidation.
The mitochondrial dysfunction is likely to be caused by increased oxidative stress (for a
21
review, see Wallace 1999), and tissues of HIF-2α null mice were found to have elevated
levels of ROS. Furthermore, the expression levels of HIF-2α-regulated genes encoding
antioxidant enzymes were reduced, leading to an impaired response to oxidative stress.
Decreased expression of antioxidant genes is thus likely to be an important factor in the
hif2a-/- phenotype (Scortegagna et al. 2003).
Heterozygous hif1a and hif2a knock-out mice were found to have impaired
development or a total loss of development of pulmonary hypertension and right
ventricular hypertrophy caused by chronic hypoxia (Yu et al. 1999, Brusselmans et al.
2003). Heterozygous hif1a knock-out mice showed a delayed increase in erythrocyte
levels in response to hypoxia relative to wild-type mice (Yu et al. 1999), while
erythrocyte levels increased normally in heterozygous hif2a knock-out mice
(Brusselmans et al. 2003), supporting data according to which the EPO gene is a target of
HIF-1 (Wang & Semenza 1993). Under conditions of prolonged hypoxia the
heterozygous hif1a mice lost more weight than the wild-type ones, while the
heterozygous hif2a mice lost weight only slightly and survived even better than the wildtype ones (Yu et al. 1999, Brusselmans et al. 2003). HIF-2α has been shown to have a
less significant role than HIF-1α in the upregulation of glycolytic genes (Brusselmans et
al. 2001), indicating that the latter is more important for regulation of the mechanisms
influencing body weight. In another model of heterozygous hif1a knock-out mice it was
found that HIF-1α is an important regulator of the carotid body-mediated increased
ventilation response in hypoxia (Kline et al. 2002). In these heterozygous mice the
increase in ventilation is not impaired but occurs via chemoreceptors other than the
carotid body, as the carotid bodies of hif1a+/- mice are unresponsive to hypoxic
stimulation. These findings indicate that HIF-1α has unique functions in the carotid body
that cannot be replaced by HIF-2α, even though it is also expressed in the carotid body
(Tian et al. 1998). These observations from knock-out studies suggest that the two HIF-α
subunits have distinct roles, perhaps related to their differences in cell-type specific
expression.
2.1.3 Structure
HIFs are heterodimers composed of α and β subunits that belong to the basic-helix-loophelix-Per-Arnt-Sim (bHLH-PAS) superfamily, the HIF-β subunit also being known as the
mammalian aryl hydrocarbon receptor nuclear translocator (ARNT) (Wang et al. 1995).
The bHLH-PAS proteins are transcriptional regulators that control a variety of
developmental and physiological processes, the most common functional unit of these
proteins being a heterodimer (reviewed in Crews 1998). These proteins share a common
conserved sequence structure having one bHLH domain and a PAS domain in the N
terminus (Figure 2). The basic region of the bHLH domain binds DNA, while the HLH
region is involved in dimerization (Ferré-D’Amare et al. 1993). The PAS domain is
divided into two subdomains, PAS-A and PAS-B, separated by a spacer (Crews et al.
1988), and this domain also mediates dimerization between the two subunits and
increases the specificity of the function of the bHLH-PAS transcription factor (Huang et
al. 1993, Pongratz et al. 1998). The C-terminal half of the 826-amino-acid HIF-1α
22
contains two transactivation domains, separated by an inhibitory domain (Jiang et al.
1997). The N-terminal transactivation domain (NAD) and the C-terminal transactivation
domain (CAD) consist of amino acids 531-575 and 786-826, respectively, in HIF-1α. The
activity of these transactivation domains is increased by hypoxia, leading to increased
transcriptional activity mediated by HIF-1. Deletion analysis of HIF-1α has revealed that
residues 576-785 between the two transactivation domains are involved in the inhibition
of HIF-1α transactivation in normoxia, and it was therefore suggested that these residues
constitute an inhibitory domain (ID), which inhibits the transcriptional activity of HIF-1
in normoxia. (Jiang et al. 1997.) This inhibitory function is most likely, however, due to
an asparaginyl hydroxylase FIH, which inhibits the transactivation of HIF-1α by
hydroxylating an asparagine residue within the CAD in normoxia (Hewitson et al. 2002,
Lando et al. 2002a, b). FIH interacts with HIF-1α residues 757-826 thus encompassing
part of the ID (Mahon et al. 2001), which could explain why deletion of ID causes
increase in transcriptional activity of HIF-1 in normoxia (Jiang et al. 1997). The oxygendependent degradation domain (ODDD, amino acids 401-603) is involved in the
regulation of HIF-α stability, controlling its degradation by the ubiquitin-proteasome
pathway (Huang et al. 1998).
The human HIF-2α and HIF-3α are polypeptides of 870 and 667 residues,
respectively. They show a high similarity to HIF-1α in the N-terminal bHLH and PAS
domains, while the sequence in the C terminus is more variable between these proteins
(Ema et al. 1997, Flamme et al. 1997, Hara et al. 2001). HIF-3α lacks the structures for
transactivation found in the C terminus of HIF-1α and HIF-2α (Gu et al. 1998). but
posseses a leucine zipper domain not found in the other isoforms that mediates DNAbinding and protein-protein interactions (Maynard et al. 2003). Another unique structure
in HIF-3α among the three isoforms is the -Leu-X-X-Leu-Leu-, which is common in
nuclear receptor cofactors (Maynard et al. 2003). These two structures suggest that HIF3α may bind DNA sequences and interact with proteins that are not among those
recognized by the other two HIF-α isoforms. HIF-2α and HIF-3α also dimerize with the
ARNT subunit, and the resulting heterodimer recognizes the HREs in a manner similar to
HIF-1α (Ema et al. 1997, Tian et al. 1997, Gu et al. 1998).
HIF-1α
826 aa
bHLH
PAS A
PAS B
Dimerization & DNA binding
HIF-2α
870 aa
HIF-3α
667 aa
HIF-β
774 or
789 aa
bHLH
PAS A
PAS B
bHLH
PAS A
PAS B
bHLH
PAS A
NAD
ID
826
786
603
575
531
ODDD
299
249
156
106
70
17
401
23
CAD
Transactivation & Regulation
NAD
NAD
ID
CAD
ID
PAS B
Fig. 2. Structures of the bHLH-PAS proteins HIF-1α, HIF-2α, HIF-3α, and HIF-β. They are
all composed of a basic-helix-loop-helix (bHLH) domain, two subdomains of PAS (Per-ArntSim) and N-terminal (NAD) and C-terminal (CAD) transactivation domains, separated by an
inhibitory (ID) domain. The oxygen-dependent degradation domain (ODDD) is involved in
the regulation of HIF-α stability.
2.1.4 Regulation
The activity of HIF-1 is determined by the expression, stability and activity of the α
subunit. HIF-1β is expressed ubiquitously in excess over HIF-1α and is maintained at
constant levels with no increase in protein expression in hypoxia, whereas the protein
levels of HIF-1α are highly induced under hypoxic conditions, the HIF-1 DNA-binding
activity correlating with the HIF-1α levels (Huang et al. 1996). Regulation of HIF-1α
expression and activity in vivo occurs at multiple levels, including protein expression and
stabilization (Wang et al. 1995, Jiang et al. 1996b, Huang et al. 1996, 1998, Kallio et al.
1999), nuclear localization (Kallio et al. 1998) and transactivation (Jiang et al. 1996a,
1997, Pugh et al. 1997 Kallio et al. 1998) and also at the level of mRNA expression.
mRNA levels in several organs in vivo are markedly higher in hypoxia than under
normoxic conditions (Wiener et al. 1996, Lee et al. 2000).
There is a peak in nuclear levels of HIF-1α in hypoxia, but the post-hypoxic decay of
this protein is very rapid, with a half-life of only a few minutes, indicating instability in
normoxic cells (Wang et al. 1995). In normoxia HIF-1α is polyubiquitinated and
subjected to degradation in the 26S proteasome (Kallio et al. 1999), this degradation
being controlled by its ODDD (Pugh et al. 1997, Huang et al. 1998). The stability of
24
HIF-1α is regulated by the interaction of the ODDD domain with the von Hippel-Lindau
tumour suppressor gene product pVHL (Maxwell et al. 1999), which requires the
hydroxylation of specific proline residues of the ODDD domain by HIF-P4Hs which are
also known as PHDs, EGLNs and HPHs (Bruick & McKnight 2001, Epstein et al. 2001,
Ivan et al. 2001, 2002, Jaakkola et al. 2001, Yu et al. 2001). HIF-1α has two pVHLbinding domains, the N-terminal region covering residues 380-417 (Masson et al. 2001)
and the C-terminal residues 556-572 (Tanimoto et al. 2000), with hydroxylatable proline
residues 402 and 564, respectively. pVHL then constitutes a functional E3 ubiquitin
ligase together with elongins B and C, cullin 2 and the RING-box protein Rbx1
(Lonegran et al. 1998, Iwai et al. 1999, Kamura et al. 1999, Lisztwan et al. 1999), which
leads to ubiquitination and degradation of HIF-1α (for a review, see Huang & Bunn
2003). In addition to proline hydroxylation, a specific asparagine, residue 803 within the
CAD, is hydroxylated by FIH (Mahon et al. 2001, Lando et al. 2002b). This inhibits the
transactivation function of HIF-1α in normoxia by inhibiting the binding of the
transcriptional coactivator p300. In hypoxia, the levels of O2 are not sufficient for the
proline and asparagine hydroxylations, and as a result HIF-1α escapes ubiquitination and
proteasomal destruction, which leads to increased protein levels, induced nuclear
localization, dimerization with the β subunit and interaction with transcriptional
coactivators to activate transcription of the target genes (for a review, see Fedele et al.
2002). However, it has been shown that the enzymatic activity of HIF-P4Hs is inhibited
in graded manner with respect to decreasing oxygen concentration, so that it is likely that
there will be enough oxygen for hydroxylase activity in many physiological hypoxia
conditions (Epstein et al. 2001). Under even mildly hypoxic conditions an E3 ubiquitin
ligase Siah2 controls the stability of HIF-P4H isoenzymes 1 and 3, but not HIF-P4H-2,
and targets the enzymes to ubiquitin-dependent degradation in proteasomes, leading to
increased HIF-1α levels (Nakayama et al. 2004). It was therefore concluded that HIFP4H-2 regulates HIF-1α expression in normoxia and the other two HIF-P4Hs in
moderate hypoxia when there is still a sufficient amount of oxygen for enzyme activity.
HIF-2α is likely to be regulated in a similar manner to HIF-1α, as it has the proline
and asparagine hydroxylation sites, but the regulation mechanism in the case of the third
isoform is less well understood. HIF-3α also seems to be targeted for degradation by
proline hydroxylation (Maynard et al. 2003), and in addition its regulation involves
alternative splicing. One alternatively spliced form of HIF-3α, IPAS, is composed of the
bHLH and PAS domains but lacks the transactivation domains, and hypoxia has been
shown to induce the process of alternative splicing (Makino et al. 2001, 2002).
Hypoxia plays a role in the regulation of all three HIF-α isoforms at the level of their
mRNAs. The mRNA levels of HIF-1α in several organs in vivo are markedly higher
under hypoxic than normoxic conditions, and HIF-1α mRNA can be detected in
myocardial specimens from patients with acute ischaemia or myocardial infarction
(Wiener et al. 1996, Lee et al. 2000). It has been shown that prolonged hypoxia
differentially regulates the expression of the HIF-1α and HIF-2α mRNAs at least in the
epithelial cells of the lung (Uchida et al. 2004). Acute hypoxia similarly induces the
expression of both proteins, but the stability of HIF-1α mRNA is reduced in prolonged
hypoxia, leading to diminished protein levels. In contrast, the mRNA levels of HIF-2α
significantly increase during prolonged hypoxia, most probably due to increased
transcriptional activity, because no changes in mRNA stability were observed (Wiesener
25
et al. 2003, Uchida et al. 2004). It has been shown in another study, however, that there is
no change in the expression of HIF-1α mRNA during hypoxic exposure of mice,
although the HIF-1α protein levels a re increased markedly (Stroka et al. 2001). mRNA
expression of HIF-3α is highly induced by moderate hypoxia in a number of tissues
(Heidbreder et al. 2003), while more severe hypoxia is required for induced HIF-1α
mRNA expression (Wiener et al. 1996). This suggests that HIF-3α may have a protective
role during early or moderate hypoxia.
The two oxygen-dependent regulatory pathways suggest either that multiple levels of
regulation allow graded responses to subtle changes in O2 concentration or that two
independent regulatory events may ensure that the hypoxic response pathway is tightly
controlled (Bruick & McKnight 2002). The HIF-P4Hs and FIH are not the only oxygen
sensors and mediators of HIF-1α stability and transcriptional activity, however, for
another post-translational modification involved in the regulation of HIF-α stability is
acetylation by the acetyltransferase ARD1 (Jeong et al. 2002). This acetylates a specific
lysine in the ODDD (Lys532 in HIF-1α) by transferring an acetyl group from acetyl-CoA
through direct interaction under normoxic conditions. This acetylation increases the
interaction of HIF-α with pVHL and subsequent degradation. The acetylated lysine
residue is conserved in HIF-1α and HIF-2α, which are oxygen-sensitive, but not in HIF3α, which is also stable in normoxia (Hara et al. 2001). This novel regulatory event is
required for maximal pVHL-mediated degradation of HIF-α. In addition to the posttranslational modifications, several growth factors, including insulin-like growth factor
(IGF) and epidermal growth factor (EGF), activate the signal transduction pathway that
leads to increased translation of HIF-1α (for a review, see Semenza 2002).
Phosphorylation of a conserved threonine in HIF-α CAD (Thr796 in HIF-1α) increases the
affinity between HIF-1α and the transcriptional coactivator p300 (Gradin et al. 2002) by
preventing the hydroxylation of Asn803 by FIH (Lancaster et al. 2004). The tumour
suppressor protein p53 limits the hypoxia-induced expression of HIF-1α by E3 ubiquitin
ligase-mediated ubiquitination and degradation, while loss of p53 enhances HIF-1α
levels leading to overexpression of VEGF, as observed in several human cancers (Ravi et
al. 2000). In addition to hypoxia, the endogenous metabolites of glycolysis and the citric
acid cycle, pyruvate and oxaloacetate, stabilize HIF-1α and stimulate HIF-1 nuclear
translocation and HIF-mediated gene expression (Dalgard et al. 2004). It has also been
suggested that ROS may play a role in the control of HIF-1α, but the exact mechanism
and their importance in response to hypoxia are unclear. It has been shown that hypoxia
stimulates ROS generation at the mitochondrial complex III, leading to stabilization of
HIF-1α (Chandel et al. 1998, 2000, Schroedl et al. 2002). Superoxide is generated at
complex III in hypoxia and is converted to H2O2 in the cytosol, leading to activation of
phosphatidylinositol-3-kinase and phosphatases and consequent HIF-1α stabilization,
suggesting that ROS may be the signal for hypoxia response. Cells lacking mitochondrial
DNA have a normal response to hypoxia, measured at the level of HIF-1α protein
stabilization, nuclear translocation and its transcriptional activation activity, however,
suggesting that oxygen sensing and the stabilization of HIF-α does not require an active
mitochondrial respiratory chain electron-transfer pathway (Srinivas et al. 2001, Vaux et
al. 2001). Results of another study suggest totally the opposite possibility, namely that
hypoxia leads to decreased ROS levels. Hydroxyl radicals were shown to be generated
within the ER in an iron-dependent Fenton reaction under normoxic conditions (H2O2 +
26
2+
3+
-
Fe → Fe + •OH + OH ), •OH generation being minimal in hypoxia. HIF-1α was also
found to be located within the ER in normoxia. Inhibition of the Fenton reaction by an
•OH scavenger led to nuclear localization and enhanced activation of HIF-1α and
subsequent expression of HIF-1α target genes upon the inhibition of HIF-P4H activity
and interaction with pVHL. (Liu et al. 2004.)
2.1.5 Effect of metal ions on HIF-1
Certain divalent metals such as cobalt, nickel and manganese have been shown to mimic
hypoxia and to cause induced expression of hypoxia-regulated genes and stabilization of
HIF-α. Various hypotheses have been developed to describe the mechanisms involved.
EPO production is induced by the transition metals cobalt and nickel. Manganese, but
not cadmium, tin or zinc, has likewise been found to induce EPO production, but less
effectively than cobalt and nickel. The stimulation of EPO expression by hypoxia and
metals is thought to occur through a common pathway, and it was hypothesized in early
studies that a heme protein may be involved in this mechanism, because EPO production
in hypoxia was shown to be inhibited by carbon monoxide and because the hypoxia and
metal-induced EPO synthesis was markedly decreased when heme synthesis was
inhibited. (Goldberg et al. 1988.) In addition to EPO, the expression of other hypoxiaregulated genes, e.g. those for VEGF (Goldberg & Schneider 1994, Ladoux & Frelin
1994) and the glycolytic enzymes (Firth et al. 1994, Semenza et al. 1994, Ebert et al.
1996), are similarly regulated by transition metals, supporting the notion of a common
mechanism. In this model the metal ions would substitute for iron in the heme and lock
the heme protein in its deoxygenated form, which would activate a signalling cascade
leading to HIF-1 activation and subsequent gene expression. Incorporation of metals into
the heme should occur competitively with iron, and it has been shown that there is indeed
competition between iron and cobalt (Ho & Bunn 1996). According to this heme
hypothesis, zinc may be another transition metal capable of stimulating HIF-1-induced
gene expression, as it effectively replaces the iron that is bound to heme. Zinc failed to
induce EPO expression under normoxia, however (Goldberg et al. 1988). Later it was
demonstrated that zinc induces the accumulation of HIF-1α but inhibits the nuclear
translocation of HIF-β resulting in inactivation of HIF-1 (Chun et al. 2000). The heme
model cannot explain, however, why the iron chelator desferrioxamin (DFO) also induces
the expression of hypoxia-inducible genes, as the heme-bound iron is not available for
chelation (Wang & Semenza 1993).
In another model the HIF-stimulating effect of transition metals, especially cobalt, has
been shown to be caused by ROS (Chandel et al. 1998). Generation of ROS increases
during hypoxia or CoCl2 treatment, causing oxidative stress in the cells, and the resulting
ROS would then activate HIF-1. It has been shown, however, that even though ROS are
produced in cells during exposure to hypoxia-mimicking metals, their formation is not
involved in HIF-1 activation (Salnikow et al. 2000).
The mechanism behind cellular oxygen sensing was resolved in 2001 when the three
P4Hs that hydroxylate HIF-1α were identified. The activity of these enzymes was shown
to be inhibited by cobaltous ions and DFO in vitro by demonstrating that CoCl2 and DFO
27
treatments reduced the binding of HIF-α to pVHL. As the HIF-P4Hs belong to a nonheme Fe2+-containing enzyme family, it was suggested that the inhibition may be caused
by cobalt through substitution at the iron centre. (Epstein et al. 2001.) Several divalent
metal ions are competitive inhibitors of the C-P4Hs with respect to Fe2+ (Rapaka et al.
1976, Ryhänen et al. 1976, Tuderman et al. 1977. Kivirikko & Pihlajaniemi 1998), which
supports the model of cobalt binding to the iron centre of the HIF-P4Hs. This model is in
agreement with the observations that both cobalt and iron chelators increase HIF activity.
Additional models for the mechanisms involved in the cobalt-induced HIF activation
have been presented, however. It has been suggested that cobalt may reduce intracellular
levels of ascorbate by converting it to dehydroascorbate, which would lead to inactivation
of HIF-P4Hs as the iron required in the enzymatic reaction would remain in the oxidized
ferryl ion form (Salnikow et al. 2004). Other studies have shown that in addition to
inhibiting the activity of the HIF-P4Hs, cobalt can become directly bound to HIF-α,
thereby inhibiting HIF-α-pVHL interaction and the subsequent degradation of HIF-α
(Yuan et al. 2001, 2003). Matrix-assisted laser desorption/ionization-time of flight mass
spectrometry analyses have shown that cobalt becomes bound to a specific amino acid
sequence within the ODDD of HIF-α, the carboxyl side groups of the acidic amino acids
probably providing the binding sites. Cobalt can become bound even to hydroxylated
HIF-α providing a model in which HIF-α-pVHL interaction is also inhibited in cases
where a subset of HIF-α is hydroxylated. Therefore HIF-α stabilization by CoCl2 differs
from that caused by the iron chelators, which act only by chelating the iron required for
the HIF-P4H activity. (Yuan et al. 2003.) Cobalt can stabilize HIF-1α even in cells
having non-functional pVHL, indicating that HIF-1α stabilization by cobalt may also
occur through a pVHL-independent mechanism (Kanaya & Kamitani 2003). The Nterminal region of HIF-1α (residues 1-330) has a second ubiquitination site in addition to
that in the C-terminus (residues 344-698), which contains the hydroxylated Pro564. As the
N-terminal region lacks the pVHL-binding site and is located far away from the Cterminal region that is ubiquitinated in a pVHL-dependent manner, it is likely that the Nterminal region is ubiquitinated independently of pVHL. Further studies revealed that
cobalt may also prevent pVHL-independent degradation, and cobalt was shown to
become bound to the two PAS domains in the N-terminal region. (Kanaya & Kamitani
2003.) An additional mechanism for the cobalt effect involves up-regulation of HIF-1α
synthesis through the phosphatidylinositol-3- kinase (PI3K) pathway (Chachami et al.
2004), which has been previously suggested to be involved in the regulation of the HIF1α transcription (Zhong et al. 2000). The mechanisms by which other metals mimic
hypoxia have not been studied in detail, and it is therefore unknown whether their
mechanisms of function are similar to those of cobalt or whether they act by different
mechanisms. It has been suggested, however, that nickel, like cobalt, may inhibit the
HIF-P4Hs by depleting intracellular ascorbate levels (Salnikow et al. 1999, 2004).
Another possibility is activation of the PI3K signalling pathway by nickel (Li et al.
2004), as also suggested for cobalt.
28
2.1.6 Role of HIF in disease
HIF-1 has an important role in many pathophysiological conditions in which oxygen
homeostasis is disrupted, e.g. heart diseases, cancer, cerebrovascular diseases and chronic
obstructive pulmonary diseases, all of which are important causes of death in western
countries.
Ischaemia and infarction occur when tissue perfusion is insufficient to meet metabolic
demands, and VEGF provides a mechanism to optimize tissue perfusion according to the
demands. VEGF has an important role in angiogenesis and its expression is increased in
most cell types in response to hypoxia. In myocardial ischaemia the viability of the
myocardium is irreversibly affected by oxygen and glucose deprivation, with infarction
as the end result (for a review, see Semenza 2000a). The development of collateral blood
vessels downstream of the coronary stenosis is induced, allowing myocardial perfusion
and thereby influencing the incidence and severity of myocardial infarction. The degree
of this coronary collateralization is correlated with the extent of VEGF expression, which
in turn is preceded by the HIF-1α mRNA and protein expression induced during acute
ischaemia and early infarction (Semenza 2000a). Many human patients with coronary
artery stenosis, especially aged subjects, nevertheless fail to develop collateral vessels,
because of an impairment of VEGF production due to decreased HIF-1 activity in
response to hypoxia (Rivard et al. 2000). This individual variation in adaptive responses
to ischaemia may determine the risk of myocardial infarction and its severity. It may be
possible to identify individuals with decreased VEGF expression, and such patients could
be good candidates for therapeutic angiogenesis by stimulating HIF-1α expression (for a
review, see Semenza 2000b). In cerebral ischaemia HIF-1α expression is induced in the
viable tissue surrounding the infarction, to an extent that is correlated with the expression
of glycolytic enzymes, suggesting that the induction of glycolytic metabolism may
promote the survival of neurons in the region surrounding the damaged tissue (Semenza
2000a).
Alveolar hypoxia leads to pulmonary hypertension in patients with chronic obstructive
lung disease (Semenza 2000a). Hypoxia induces remodelling of the pulmonary arterioles,
which results in reduced lumen diameter and increased resistance to blood flow leading
to progressive heart failure. It has been suggested that inhibition of HIF-1 activity locally
in the lung could be used as a therapeutic means of treating or preventing pulmonary
hypertension in individuals at risk (Semenza 2000a).
In diabetes, occlusion of the retinal vessels leads to neovascularization caused by
ischaemia, a process in which VEGF expression plays an important role (Pierce et al.
1995). It is also known that there are increased levels of HIF-1α in the ischaemic retina
that show temporal and spatial correlations with the increased expression of VEGF
(Ozaki et al. 1999). These findings are consistent with the hypothesis that HIF-1 plays a
role in the up-regulation of VEGF in the ischaemic retina.
HIF-1 plays a role in pre-eclampsia and intrauterine growth retardation (IUGR), which
are pregnancy disorders (Semenza 2000a). The central defect in pre-eclampsia, which is a
leading cause of foetal and maternal morbidity and mortality, is the failure of trophoblasts
to invade the myometrium and induce remodelling of the uterine arteries during early
placentation. After the first 10-12 weeks of pregnancy, the placenta and foetus are
29
exposed to maternal blood and at this stage the trophoblasts actively invade the maternal
decidua, the developmental switch from proliferative to invasive being controlled by
cellular O2 concentrations (Genbacev et al. 1997, Caniggia et al. 2000). Improper downregulation of HIF-1 activity is thought to play a role in pre-eclampsia, while the noninvasive trophoblast phenotype is maintained by HIF-1-mediated TGFβ3 expression
(Caniggia et al. 2000). Intrauterine growth retardation is another leading cause of foetal
morbidity and mortality, being caused by placental and foetal hypoxia due to decreased
placental perfusion (Semenza 2000a). IGFs have an important role in foetal growth, and
expression of their inhibitor, IGF-binding protein 1, is induced in hypoxia and IUGR
(Semenza 2000a).
Tumours need vascularization in order to grow beyond a few cubic millimetres, and
increased expression of VEGF is essential for this angiogenesis. Tumour propagation and
malignant progression are also associated with intratumoral hypoxia, however, which is
caused by structurally and functionally abnormal vascularization of the tumour, resulting
in irregular perfusion (for a review, see Hockel & Vaupel 2001). Intratumoral hypoxia is a
challenge for cancer treatment, because it leads to tumour resistance to therapy (Shannon
et al. 2003). Marked overexpression of HIF-1α has been observed in many human
tumours, including brain gliomas and colon, breast, lung and prostate carcinomas (Zhong
et al. 1999, Zagzag et al. 2000). This overexpression is correlated with vascular density,
indicating that HIF-1 activity leads to angiogenic switch in hypoxic tumours (Zagzag et
al. 2000, Huss et al. 2001). HIF-1α overexpression is due to both intratumoral hypoxia
and genetic alterations, such as the loss of pVHL and p53 functions, which leads to
decreased ubiquitination of HIF-α (for a review, see Semenza 2002). Loss of pVHL
function also leads to von Hippel-Lindau disease, which is a heritable cancer syndrome
caused by mutations in the gene for VHL, leading to continuous expression of HIF-1.
Patients have a high risk of developing various benign and malignant tumours of the
central nervous system, kidneys, adrenal glands, pancreas and reproductive adnexal
organs. (Lonser et al. 2003.) Based on studies with renal clear cell carcinomas, it has
been suggested, however, that rather HIF-2α than HIF-1α is the oncogene in renal cancer
(Kondo et al. 2002, Maranchie et al. 2002). It was also speculated that due to the
expression pattern of HIF-2α, it could be responsible for the observed organ specificity of
the VHL disease (Maranchie et al. 2002). Tumours with HIF-1α overexpression are
highly aggressive, as the expression of several genes controlling tumour cell survival, are
regulated by HIF-1 (Shannon et al. 2003). Patients with glioblastoma multiforme, for
example, which shows the highest HIF-1α expression levels among brain tumours, have
survival times of less than one year (Semenza 2002). HIF-1α immunohistochemistry
could be used to analyse tumour biopsies, provide prognostic information and identify
patients requiring aggressive therapy. Inhibition of the activity of HIF-1 in tumours
provides possibilities for developing new therapeutic strategies, the most appropriate
target population for HIF-1 inhibitors being those cancer types which show high levels of
HIF-1α overexpression correlated with unsuccessful treatment and mortality (Semenza
2002). It has been suggested that overexpression of HIF-P4Hs could be used to inhibit the
stabilization of HIF-1α in hypoxia and thereby to inhibit the vascularization and tumour
growth (Erez et al. 2003). In addition to chemotherapeutic agents, gene therapy could be
one approach for targeting hypoxic regions of tumours by means of hypoxia-targeted
transgene expression, for which purpose a hypoxia-dependent replicative adenovirus has
30
been developed, an attenuated oncolytic adenovirus that selectively lyses cells under
hypoxic conditions (Post & Van Meir 2003). Hypoxia-targeted delivery and pro-drug
activation could also be used. For example, an adenovirus expressing human cytochrome
P450 is regulated by a promoter that is responsive to optimized hypoxia and delays
tumour growth in response to the pro-drug cyclophosphamide (Binley et al. 2003). These
hypoxia-targeted suicide gene therapies are aimed at activating cytotoxic agents at the
tumour site while sparing normal tissue from damage.
2.2 Prolyl 4-hydroxylases
Oxygenases catalyze reactions in which an oxygen atom is incorporated from molecular
oxygen into an organic substrate. These enzymes commonly use iron as a cofactor, either
as part of a heme group or in a non-heme form. Oxygenases can be divided into
dioxygenases, in which both oxygen atoms of the O2 molecule are incorporated into the
substrate, and mono-oxygenases, in which only one oxygen atom is incorporated while
the other is reduced to water. The 2-oxoglutare-dependent dioxygenases make up the
largest subgroup of mononuclear non-heme Fe2+-dependent oxygenases (Prescott & John
1996). Most of them have an absolute requirement for 2-oxoglutarate and require Fe2+
and ascorbate for optimal function. These enzymes occur widely, being found in many
organisms from prokaryotes to eukaryotes, one form of P4H also being expressed by the
Paramecium bursaria Chlorella virus-1 (PBCV-1) (Eriksson et al. 1999). 2-Oxoglutarate
dioxygenases catalyze a range of oxidative reactions, including hydroxylations,
saturations, ring closures and rearrangements (Prescott & John 1996). The bestcharacterized 2-oxoglutarate dioxygenase in mammals is the P4H that catalyzes the posttranslational hydroxylation of proline residues in collagens. It was believed that C-P4Hs
are the only P4Hs in animals until a novel P4H with three isoforms catalyzing the
hydroxylation of proline residues within the α subunit of the transcription factor HIF was
recently identified (Bruick & McKnight 2001, Epstein et al. 2001, Ivan et al. 2002).
2.2.1 Occurrence and functions of 4-hydroxyproline in proteins
other than HIF
4-Hydroxyproline is found in vertebrates mainly in collagens and other proteins with
collagen-like sequences (for reviews, see Kivirikko & Pihlajaniemi 1998, Myllyharju &
Kivirikko 2001, 2004). The collagens are a large family of extracellular matrix proteins
that provide structural integrity for various organs and tissues. In addition, they are
involved in other important biological functions, such as the healing of wounds and
fractures, inhibition of angiogenesis and tumour growth, and organ morphogenesis
(Prockop & Kivirikko 1995, O’Reilly et al. 1997, Petitclerc et al. 2000, Lin et al. 2001,
Myllyharju & Kivirikko 2001, 2004). Some of these functions, however, are performed
by the non-collagenous domains found in many collagens, like the inhibition of
angiogenesis and tumour growth that are performed by restin and endostatin, which are
31
fragments of collagen types XV and XVIII, respectively, and by the C-terminal
noncollagenous domain of type IV collagen (O’Reilly et al. 1997, Ramchandran et al.
1999, Petitclerc et al. 2000). At least 27 vertebrate proteins containing altogether 42
different polypeptide chains have so far been defined as collagens (for a review, see
Myllyharju & Kivirikko 2004). All collagens consist of three polypeptides, called α
chains, that contain repeating -Gly-X-Y- triplets, but the lengths of the chains and the
residues in the X and Y positions vary according to the collagen type and domain (for a
review, see Myllyharju & Kivirikko 2004). These α chains in a collagen trimer can all be
identical or the trimer can consist of two or three different chains. The α chains are coiled
into a left-handed helix, and the three helical chains are then coiled around each other
into a right-handed triple helix. The existence of glycine as every third residue in the
collagen sequence is essential for the triple-helical structure, because as the smallest
amino acid, it fits in the middle of the helix where the three chains come together (see
Prockop & Kivirikko 1995). Proline is often found in the X position of the -Gly-X-Yrepeat and 4-hydroxyproline in the Y position. The hydroxylation of proline residues is
required for stabilization of the collagen triple helix at normal body temperature. The
midpoint of the thermal transition from helix to coil (Tm) for non-hydroxylated type I
procollagen is 24°C, while the Tm for the fully hydroxylated type I procollagen molecule
is 39°C (Berg & Prockop 1973, Jimenez et al. 1973, Rosenbloom et al. 1973). Thus
without proline hydroxylation, the collagen triple helix denatures at normal body
temperature. There are some small but distinct differences in 4-hydroxyproline content
between collagen types, content in the most abundant collagen, type I, being about 100
residues per 1000 (Kivirikko et al. 1992).
The mechanisms by which the 4-hydroxyproline residues stabilize the collagen triple
helix are not totally understood, but several proposals have been made. X-ray
crystallographic studies have shown that each triple helix of a collagen-like peptide is
surrounded by an extensive network of hydrogen bonds between water molecules and
peptide acceptor groups. 4-Hydroxyproline residues have a critical role in this water
network by binding two water molecules and thus forming an interchain link to the amide
oxygen of another 4-hydroxyproline (Bella et al. 1994). However, when the 4hydroxyprolines in the Y positions of the -Gly-X-Y- triplets were replaced by 4(R)
fluoroproline (Flp), which is incapable of forming hydrogen bonds, the stability of the
(Pro-Flp-Gly)10 triple helix was enhanced, exceeding the thermal stability of any known
collagen of a similar size (Holmgren et al. 1998). The result indicated that the thermal
stability was due to inductive effects caused by the most electronegative fluorine that
favours the trans conformation of the hydroxyprolyl peptide bond. It can therefore be
predicted that 4-hydroxyproline residues act in the same way and stabilize the collagen
triple helix by a stereoelectronic effect. In addition, several structural and host-guest
studies of collagen have been made in order to resolve the basis for the stability of the
triple helix, and crystallization studies, for example, have shown that there are only 15
water molecules in the (Pro-Hyp-Gly)10 structure, arguing against stability being provided
by water molecules (for a review, see Jenkins & Raines 2002).
In addition to the collagens, the collagen superfamily includes a heterogeneous group
of more than 20 other proteins having collagen-like sequences. This group includes the
C1q subcomponent of complement, a C1q-like factor, adiponectin, at least eight
collectins and three ficolins, the collagen-like tail structure of acetylcholinesterase, three
32
macrophage receptors, two EMILINS (elastic fibre-associated glycoproteins), a srchomologous-and-collagen protein and ectodysplasin (for a review, see Myllyharju &
Kivirikko 2004). Additional proteins reported to contain 4-hydroxyproline are elastin
(Rosenbloom 1987), hydroxyproline-luteinizing hormone releasing hormone (Gautron et
al. 1992) and hydroxyproline-lysyl-bradykinin (Sasaguri 1988). As in the collagens, the
4-hydroxyproline is also found in these proteins almost exclusively in the Y positions of
the -Gly-X-Y- sequences (Kivirikko et al. 1992).
2.2.2 Vertebrate collagen prolyl 4-hydroxylases
C-P4Hs (EC 1.14.11.2, procollagen-proline, 2-oxoglutarate 4-dioxygenase) are located
within the endoplasmic reticulum (ER) and catalyze the formation of 4-hydroxyproline in
collagens and other proteins with collagen-like sequences (for reviews, see Kivirikko &
Myllyharju 1998, Kivirikko & Pihlajaniemi 1998, Myllyharju 2003).
The active C-P4H in vertebrates is an α2β2 tetramer in which the two catalytic sites are
located within the α subunits and the β subunits are identical to the enzyme and
chaperone protein disulphide isomerase (PDI, EC 5.3.4.1) (Pihlajaniemi et al. 1987,
Kivirikko & Myllyharju 1998, Kivirikko & Pihlajaniemi 1998, Myllyharju 2003). The
molecular weight of the tetramer is about 240 000, the molecular weights of the two types
of monomer being about 63 000 (α subunit) and 58 000 (β subunit). C-P4H was for a
long time assumed to be of only one type, but two additional α subunit isoforms have
now been cloned and characterized from human, mouse and rat tissues (Helaakoski et al.
1995, Annunen et al. 1997, Kukkola et al. 2003, Van Den Diepstraten et al. 2003), while
the previously known α subunit has been cloned from man (Helaakoski et al. 1989), and
the rat (Hopkinson et al. 1994), mouse (Helaakoski et al. 1995) and chicken (Bassuk et
al. 1989). All three α subunit isoforms have been shown to combine with PDI and to
form the [α(I)]2β2, [α(II)]2β2 and [α(III)]2β2 tetramers, known as the type I, II and III
enzymes, respectively. Coexpression studies of the type I and type II α subunits with the
PDI polypeptide in insect cells showed that they do not form a α(I)α(II)β2 tetramer
(Annunen et al. 1997).
The human α(I), α(II) and α(III) subunits consist of 517, 514 and 525 amino acids,
respectively, and are synthesized in forms containing a signal peptide of 17-21 additional
residues (Helaakoski et al. 1989, Annunen et al. 1997, Kukkola et al. 2003, Van Den
Diepstraten et al. 2003). The overall sequence identity between the processed human α(I)
and α(II) subunits is 64%, that between the human α(I) and α(III) subunits 35% and that
between the human α(II) and α(III) subunits 37%, the highest degree of identity being
found in the catalytically important C-terminal region, where that between the α(I) and
α(II) subunits is as high as 80%, while that between the α(III) and the α(I) or α(II)
subunit is 56-57%. All three α subunits have four conserved critical residues at the
catalytic site: two histidines and one aspartate that bind the Fe2+ atom and a lysine that
binds the C-5 carboxyl group of 2-oxoglutarate (Lamberg et al. 1995, Myllyharju &
Kivirikko 1997). In addition, a third histidine residue is also conserved in all three human
α subunits (Annunen et al. 1997, Kukkola et al. 2003; for reviews see Kivirikko &
Myllyharju 1998, Myllyharju 2003), this third histidine probably being involved in both
33
binding of the C-1 carboxyl group of the 2-oxoglutarate to the iron atom and
decarboxylation of this cosubstrate (Myllyharju & Kivirikko 1997). Five cysteine
residues are conserved in all the human α subunits, but α(II) contains an additional
cysteine between the fourth and fifth conserved ones and α(III) has an additional cysteine
located between the first and second conserved ones (Kukkola et al. 2003). The α(I)
subunit has two intrachain disulphide bonds formed between the second and third and the
fourth and fifth cysteine residues (John & Bulleid 1994, Lamberg et al. 1995), these
being needed to maintain the structure of the α subunit. No interchain bonds are formed
between the α and PDI subunits, however. The human α subunits contain two asparagine
residues that act as attachment sites for oligosaccharides (Helaakoski et al. 1989,
Annunen et al. 1997, Kukkola et al. 2003), but the extent of their glycosylation varies
(Kedersha et al. 1985, Vuori et al. 1992a). Glycosylation of the α subunit has no role in
the assembly of the type I enzyme tetramer or the catalytic activity of the type I or type
III enzyme tetramer (Lamberg et al. 1995, Kukkola et al. 2003).
The three C-P4H isoenzymes have distinct differences in their expression patterns in
tissues. Type I C-P4H is the main form in most cells and in tissues of mesenchymal
origin, while type II is the main form in chondrocytes, endothelial cells and the cells of
epithelial structures (Annunen et al. 1998, Nissi et al. 2001). The type I isoenzyme is the
main form in fibroblasts and fibroblastic cells in many tissues, for example, and is the
only form in skeletal myocytes and smooth muscle cells and represents about 90% of the
total C-P4H activity in the kidneys, heart, liver, skeletal muscle and skin, while the type
II enzyme represents over 70% of the C-P4H activity in cartilage (Annunen et al. 1998).
The expression of type III isoenzyme has not been studied in detail so far. The highest
levels of the human α(III) mRNA expression were found in the human placenta, adult
liver and foetal skin in Northern hybridization studies, and in the placenta and foetal
kidney, liver and lung in PCR analysis of multitissue cDNA panels, but the levels in all
the tissues studied were much lower than those of the α(I) and α(II) mRNAs (Kukkola et
al. 2003). The α(III) mRNA is also expressed in human vascular smooth muscle cells,
being found especially in the fibrous cap of atherosclerotic carotid arteries (Van Den
Diepstraten et al. 2003). It is currently unknown, however, whether the type III is
expressed at higher levels than the other two isoenzymes in any cells or tissues.
The α subunits have a distinct peptide-substrate-binding domain that is separate from
the catalytic domain and is located between residues 138 and 244 in the human α(I)
subunit (Myllyharju & Kivirikko 1999), and NMR studies have revealed that this domain
consists of five α helices and one short β strand, being thus very distinct from the
previously characterized proline-rich peptide-binding modules that consist mainly of β
strands (Hieta et al. 2003). The recombinant peptide-binding domain has recently been
crystallized for structural studies (Pekkala et al. 2003). The catalytic domain is located in
the C-terminal region of the α subunits and contains the Fe2+ and 2-oxoglutarate binding
residues described above.
The PDI that serves as the β subunit in the vertebrate C-P4H tetramers (Pihlajaniemi
et al. 1987) is a catalyst of disulphide bond formation, breakage and rearrangement
during protein folding in the ER (for a recent review, see Freedman et al. 2002). The
secreted proteins enter the ER in an unfolded state and exit when they are correctly
folded and assembled. Other cellular and extracellular locations as well as ER have also
been reported for PDI (for a review, see Turano et al. 2002). Proteins belonging to the
34
PDI family have been found on the cell surface, in the extracellular space, cytosol and
nucleus, and in some cases this non-ER location is related to the redox properties of the
proteins.
The human PDI polypeptide consists of 491 amino acids and is synthesized in a form
containing a signal peptide of 17 additional residues (Pihlajaniemi et al. 1987). PDI is a
modular protein consisting of four structural domains, a, b, b’ and a’, and an additional
linker region between b’ and a’ and a C-terminal acidic extension c (Edman et al. 1985,
Pihlajaniemi et al. 1987, Kemmink et al. 1997, Freedman et al. 1998, Pirneskoski et al.
2004). The homologous a and a’ domains show significant sequence similarity to
thioredoxin, a polypeptide that is involved in redox functions in the cytoplasm, and they
contain the thioredoxin fold (Kemmink et al. 1996). The homologous b and b’ domains
have no sequence similarity to thioredoxin or to the domains a and a’, but NMR studies
have shown that they also have the thioredoxin fold (Kemmink et al. 1997, Freedman et
al. 2002). The a and a’ domains each contain the active site motif -Cys-Gly-His-Cys-,
and the isolated domains can act as reductases or oxidases. The other domains are also
needed for full activity of the PDI polypeptide, however, the effect being most significant
in complex disulphide isomerizations (Darby et al. 1998).
The main function of PDI in the C-P4H α2β2 tetramer is to keep the highly insoluble α
subunits in a non-aggregated, catalytically active form (Vuori et al. 1992a, John et al.
1993, Veijola et al. 1996b), this function probably being related to the chaperone function
of PDI, as mutations that inactivate the two catalytic sites of the PDI polypeptide have no
effect on tetramer formation (Vuori et al. 1992b). A minimum requirement for the
assembly of an active C-P4H tetramer is fulfilled by the PDI domain pair b’a’
(Pirneskoski et al. 2001). The α subunit has no ER retention signal, and thus another
function of PDI in the tetramer is to keep the enzyme within the ER, as PDI has the Cterminal ER-retention signal -Lys-Asp-Glu-Leu and as deletion of this sequence from the
PDI polypeptide in recombinant expression studies in insect cells led to significant
secretion of both the free polypeptide and the enzyme tetramer into the culture medium
(Vuori et al. 1992b). PDI is expressed in excess over the α subunit, and thus regulation of
the amounts of the active enzyme tetramer occurs mainly through regulation of the α
subunit (for a review, see Kivirikko & Myllyharju 1998).
2.2.3 Nematode collagen prolyl 4-hydroxylases
The cuticle or exoskeleton of the nematode Caenorhabditis elegans is composed of crosslinked collagens (Cox et al. 1981), and the genome encodes about 175 cuticle collagen
chains (Johnstone 2000, Myllyharju & Kivirikko 2004). C. elegans has four genes
encoding C-P4H α subunit-like polypeptides, three of which, termed PHY-1, PHY-2 and
PHY-3, have now been characterized (Veijola et al. 1994, Friedman et al. 2000, Hill et al.
2000, Winter & Page 2000, Riihimaa et al. 2002, Myllyharju et al. 2002). The processed
PHY-1 and PHY-2 polypeptides consist of 543 and 523 amino acids, respectively (Veijola
et al. 1994, Friedman et al. 2000, Hill et al. 2000, Winter & Page 2000), and are likely to
have a similar domain structure to the human α subunit isoforms, the highest
conservation being found within the C-terminal catalytic domains, which contain all the
35
residues essential for catalytic activity. PHY-1 and PHY-2 are 57% identical to each other,
while the identity between PHY-1 and PHY-2 and the human α(I) and α(II) ranges from
43% to 46% (Veijola et al. 1994, Winter & Page 2000). The third C. elegans α subunit
isoform, PHY-3, is much shorter than PHY-1 and PHY-2, consisting of only 295 amino
acids, and its sequence is only 17% identical to residues 256-542 in PHY-1 and 18-20%
identical to the corresponding residues in PHY-2 and the human α(I) and α(II) subunits,
conservation being highest within the C-terminal region that covers PHY-3 amino acids
150-295 (Riihimaa et al. 2002.). C. elegans has two genes, pdi-1 and pdi-2, encoding PDI
isoforms (Winter & Page 2000), and coexpression studies of PHY-1 and PDI-2 or human
PDI in insect cells led to the formation of an active αβ dimer instead of the tetramer
found in vertebrates (Veijola et al. 1994, 1996a). PHY-1 and PHY-2 also form a PHY1/PHY-2/(PDI-2)2 mixed tetramer with PDI-2, but neither PHY-1 nor PHY-2 forms any
tetramer in the absence of the other (Myllyharju et al. 2002). This mixed tetramer is also
the main C-P4H form in wild-type C. elegans in vivo, with small amounts of the PHY1/PDI-2 dimer but no PHY-2/PDI-2 dimer, while very small amounts of the latter dimer
are found in phy-1-/- nematodes (Myllyharju et al. 2002). Recombinant PHY-3 forms an
active enzyme only when coexpressed with C. elegans PDI-1, an isoform that does not
form any active enzyme with the PHY-1 and PHY-2 polypeptides (Myllyharju et al. 2002,
Riihimaa et al. 2002).
The phy-1 and phy-2 genes are expressed in the hypodermal cells at times when
collagen is being synthesized, indicating an essential role in the formation of the cuticle
at all developmental stages (Winter & Page 2000). PHY-3 is involved in the
hydroxylation of proline residues in the early embryo, most likely in the collagens of the
egg shell (Riihimaa et al. 2002).
The cuticle of the filarial nematode Brugia malayi, like those of other nematodes, is
composed of collagen-like molecules. A B. malayi C-P4H has been cloned and
characterized and found to have some highly unusual properties. The α subunit, PHY-1,
is soluble and active without any PDI polypeptide and forms an α4 tetramer (Winter et al.
2003). Lack of the PDI polypeptide is not a unique property, however, as short
monomeric P4Hs of 210 and 283 residues have been characterized from PBCV-1
(Eriksson et al. 1999) and Arabidopsis thaliana (Hieta & Myllyharju 2002). The B.
malayi PHY-1 is much longer than these P4Hs, however, the processed polypeptide
consisting of 524 amino acids and having its highest homology with the C. elegans PHY1 polypeptide, with an identity of 76%. The cysteine residues and all the catalytically
critical residues are conserved between these two polypeptides, the catalytic and
inhibition properties of the B. malayi homotetramer being very similar to those of the C.
elegans PHY-1/PHY-2/(PDI-2)2 mixed tetramer. The B. malayi PHY-1 is likely to
function in the modification of cuticular collagens in the same way as the C. elegans
mixed tetramer. (Winter et al. 2003.)
A C-P4H has also been characterized from another filarial nematode Onchocerca
volvulus, the PHY-1 polypeptide of which has many characteristics in common with the
C-P4H α subunits from C. elegans and vertebrates (Merriweather et al. 2001). It is
believed that detailed characterization of the enzymes from B. malayi and O. volvulus
might lead to the development of specific inhibitors for these parasites.
36
2.2.4 Prolyl 4-hydroxylases from other organisms
About 20 C-P4H α subunit-related genes have been identified in Drosophila
melanogaster (Adams et al. 2000, Abrams & Andrew 2002), an unexpectedly high
number because D. melanogaster is known to have only three types of collagen (see
Myllyharju 2003, Myllyharju & Kivirikko 2004). It is therefore possible that some of
these P4Hs act on proline residues in proteins other than the collagens. Ten of the α
subunit-like genes have been cloned and their expression patterns analysed (Annunen et
al. 1999, Abrams & Andrew 2002). Six of these genes have tissue-specific embryonic
expression, one is expressed in the mouth part precursors and two in the proventriculus
and epidermis. One of the genes has a broader expression, including the anterior and
posterior midgut primordial, the fat body, the haemocytes and the epidermis. (Abrams &
Andrew 2002.) One of the α subunits has been characterized in more detail and the
processed polypeptide was found to consist of 516 amino acids and to have 34% and 35%
sequence identity to the human α(I) and α(II) subunits, respectively, and 31% identity to
the C. elegans PHY-1 (Annunen et al. 1999). Four of the five critical residues at the
cosubstrate-binding sites are conserved, but the residue assisting in the binding of the C-1
carboxyl group of the 2-oxoglutarate and the decarboxylation of this cosubstrate is an
arginine instead of a histidine, this difference probably being the reason for the 2 to 3fold higher Km for 2-oxoglutarate. Differences were also found in the amino acid
sequence of the peptide-binding domain, which probably explains the relatively poor
hydroxylation of (Pro-Pro-Gly)10, a good substrate for the vertebrate enzymes.
Coexpression studies indicate that the D. melanogaster enzyme is an α2β2 tetramer, thus
resembling the vertebrate enzymes. (Annunen et al. 1999.)
4-Hydroxyproline is found in many plant glycoproteins, and P4Hs have been partially
characterized from several plant species (Kivirikko et al. 1992), including Phaseolus
vulgaris (Wojtaszek et al. 1999) and green algaes, such as Chlamydomonas reinhardii
(Kaska et al. 1987), Volvox carteri (Kaska et al. 1988) and Enteromorpha prolifera
(Simpson et al. 1986). The P4Hs from the unicellular and multicellular green algae have
been shown to be monomeric (Kaska et al. 1987, 1988).
A P4H has recently been cloned and characterized from Arabidopsis thaliana (Hieta &
Myllyharju 2002). The genome of this species encodes six P4H α subunit-like
polypeptides, one of which was shown to be a 283-residue soluble monomer with 21%
sequence identity to the catalytically important C-terminal region of the human α(I)
subunit. All the catalytically critical residues identified in animal enzymes are conserved
in this plant P4H. The recombinant A. thaliana enzyme hydroxylated poly(L-proline)
effectively, which is the most distinct difference in catalytic properties between plant and
animal P4Hs. In addition to hydroxylating poly(L-proline), the A. thaliana enzyme also
hydroxylated many other proline-rich peptides and surprisingly even the collagen-like
peptides (Pro-Pro-Gly)10 and (Ala-Pro-Gly)5. (Hieta & Myllyharju 2002.)
It has been believed that viral and bacterial proteins have no 4-hydroxyproline, but a
C-P4H α subunit-like polypeptide has also been identified from a eukaryotic algal virus,
Paramecium bursaria Chlorella virus-1 (PBCV-1) (Eriksson et al. 1999). The enzyme is
a monomer and its 242-residue polypeptide shows distinct sequence similarity to the Cterminal region of the animal α subunits. The recombinant enzyme hydroxylated (Pro-
37
Pro-Gly)10, poly(L-proline) and several synthetic peptides corresponding to proline-rich
repeats coded by the viral genome, suggesting that PBCV-1 expresses proteins in which
proline residues become hydroxylated (Eriksson et al. 1999).
2.2.5 Catalytic properties
All P4Hs require Fe2+, 2-oxoglutarate, O2 and ascorbate (reviewed in Kivirikko &
Myllyharju 1998, Kivirikko & Pihlajaniemi 1998, Myllyharju 2003), the Km values for
these cosubstrates varying within relatively narrow limits between the C-P4Hs and
related enzyme forms from animals, viruses and plants (Veijola et al. 1994, Vuori et al.
1992a, Helaakoski et al. 1995, Annunen et al. 1997, 1999, Eriksson et al. 1999, Hieta &
Myllyharju 2002, Myllyharju et al. 2002, Kukkola et al. 2003, Winter et al. 2003).
The requirement for non-heme iron is very specific, and this cation is loosely bound,
the activity of the purified enzyme being completely dependent on added Fe2+ (for
reviews, see Kivirikko & Myllylä 1980, Kivirikko et al. 1992). Several divalent cations,
e.g. palladium, calcium, copper and zinc, inhibit C-P4Hs competitively with respect to
Fe2+, Zn2+ being the most potent inhibitor (Rapaka et al. 1976, Ryhänen et al. 1976,
Tuderman et al. 1977). 2-Oxoglutarate is a highly specific and absolute requirement and
is decarboxylated during the hydroxylation (Kivirikko & Myllylä 1980, Kivirikko et al.
1992). Many small molecules, such as pyridine 2,4-dicarboxylate, pyridine 2,5dicarboxylate, N-oxaloglycine and oxalylalanine (Majamaa et al. 1984, Cunliffe et al.
1992, Baader et al. 1994) are competitive inhibitors with respect to 2-oxoglutarate.
Ascorbate is also an absolute requirement, being needed in the uncoupled
decarboxylation of 2-oxoglutarate as an alternative oxygen acceptor in order to prevent
oxidation of the enzyme-bound Fe2+ or residues in the enzyme protein between catalytic
cycles (see, Myllylä et al. 1984, Kivirikko et al. 1992, Kivirikko & Pihlajaniemi 1998).
Molecular oxygen is required to supply the oxygen for the hydroxyl group.
The catalytic site of the C-P4Hs comprises a set of separate locations for the binding
of the various cosubstrates and the peptide substrate (Hanauske-Abel & Günzler 1982,
Hanauske-Abel & Popowicz 2003). The Fe2+ is bound to the α subunit by three residues,
His412, Asp414 and His483 in the human α(I) subunit (Lamberg et al. 1995, Myllyharju &
Kivirikko 1997). while the 2-oxoglutarate binding site can be divided into three distinct
subsites (Majamaa et al. 1984, Kivirikko et al. 1992). Subsite I is a positively charged
side chain of the α subunit that binds the C-5 carboxyl group of the 2-oxoglutarate, this
residue being Lys493 in the human α(I) (Myllyharju & Kivirikko 1997). Subsite II
comprises two cis-positioned coordination sites of the enzyme-bound Fe2+ and is chelated
by the C-1-C-2 moiety, while subsite III involves a hydrophobic binding site in the C-3C-4 region of the cosubstrate. Molecular oxygen is thought to be bound to the ferrous ion
end-on in an axial position (Hanauske-Abel & Günzler 1982, Hanauske-Abel &
Popowicz 2003). The binding site for ascorbate is partially identical to the subsite II for
2-oxoglutarate, as it also contains the two cis-positioned coordination sites of the
enzyme-bound iron (Majamaa et al. 1986).
The Km values of the human type I C-P4H for its cosubstrates are 2 µM for Fe2+, 20
µM for 2-oxoglutarate and 300 µM for ascorbate (Myllyharju & Kivirikko 1997). There
38
are some differences in Km values between the three human isoenzymes and the C-P4Hs
and related enzymes from other organisms. The Km for Fe2+ differs four-fold between the
human type I and type III isoenzymes, that of the latter being only 0.5 µM (Kukkola et al.
2003). Among enzymes from other organisms, the Km values for Fe2+ vary between 0.4
µM in the case of PBCV-1 and 16 µM in that of A. thaliana, respectively (Eriksson et al.
1999, Hieta & Myllyharju 2002) The Km values of the three human isoenzymes for 2oxoglutarate are identical, but those of the Brugia malayi α4 tetramer, the Caenorhabditis
elegans mixed tetramer, one Drosophila melanogaster enzyme and the characterized A.
thaliana P4H are three to six times higher. The Km values of the human isoenzymes and
enzymes from other organisms for ascorbate are essentially the same. (Vuori et al. 1992a,
Annunen et al. 1997, 1999, Eriksson et al. 1999, Hieta & Myllyharju 2002, Myllyharju et
al. 2002, Kukkola et al. 2003, Winter et al. 2003.)
The minimum sequence requirement for the hydroxylation of a proline residue by a
vertebrate C-P4H is fulfilled by an X-Pro-Gly triplet, these enzymes with only a few
exceptions hydroxylating only proline residues preceding a glycine. The chain length of
the peptide substrate has a marked effect on the Km, which decreases with increasing
chain length (for reviews, see Kivirikko & Myllyharju 1998, Kivirikko & Pihlajaniemi
1998, Myllyharju 2003). The Km of the human type I enzyme for the tripeptide Pro-ProGly is 20 mM, whereas the Km values for (Pro-Pro-Gly)5 and (Pro-Pro-Gly)20 are 2 mM
and 50 µM, respectively, when expressed as concentrations of the -X-Pro-Gly- units
(Kivirikko & Myllyharju 1998). A similar effect of substrate chain length has also been
noticed with viral and plant P4Hs (Eriksson et al. 1999, Hieta & Myllyharju 2002).
Poly(L-proline) is a competitive inhibitor of the vertebrate and D. melanogaster C-P4Hs,
the Ki values decreasing with increasing chain length (Helaakoski et al. 1995, Annunen et
al. 1997, 1999, Kukkola et al. 2003), although this polypeptide is most effective in the
case of vertebrate type I enzyme. The C. elegans and B. malayi C-P4Hs are particularly
ineffectively inhibited by poly(L-proline) (Veijola et al. 1994, Myllyharju et al. 2002,
Winter et al. 2003), and interestingly, poly(L-proline) is a substrate for viral and plant
P4Hs (Eriksson et al. 1999, Hieta & Myllyharju 2002). The differences between the
enzymes in their Km and Ki values for various peptide substrates and inhibitors suggest
that there must be some distinct differences in the structures of the peptide-binding
domains between the animal C-P4Hs and related enzymes.
2.2.6 Reaction mechanism
In the reaction catalyzed by the P4Hs 2-oxoglutarate is stoichiometrically decarboxylated
during hydroxylation, with one atom of the O2 molecule being incorporated into the
succinate and the other into the hydroxyl group formed on the proline residue (Figure 3)
(see Kivirikko & Myllylä 1980, Kivirikko et al. 1992, Kivirikko & Myllyharju 1998,
Kivirikko & Pihlajaniemi 1998, Myllyharju 2003). Kinetic studies have shown that the
reaction occurs in an ordered way (Myllylä et al. 1977, Tuderman et al. 1977). First Fe2+
is bound to the enzyme, followed by binding of the 2-oxoglutarate, O2 and the peptide
substrate, release of the products occurring in the order of the hydroxylated peptide, CO2,
succinate and Fe2+, in which the Fe2+ is not released between most of the catalytic cycles.
39
When the proline is being hydroxylated, ascorbate is not used stoichiometrically and the
enzyme can catalyze a number of reaction cycles in its absence, but the reaction rate then
ceases rapidly, probably due to the formation of Fe3+·O-, which needs to be reduced by
ascorbate to make the enzyme available for further catalytic cycles, ascorbate thus being
needed to reactivate the enzyme. P4Hs also catalyze uncoupled decarboxylation of 2oxoglutarate, i.e. decarboxylation without hydroxylation of the substrate. Ascorbate acts
as an alternative oxygen acceptor in these uncoupled cycles and is then consumed
stoichiometrically (Myllylä et al. 1984), the rate of this uncoupled reaction observed in
the absence of any peptide substrate being only 1-4% of the rate of the complete reaction
seen in the presence of a saturating substrate concentration (Tuderman et al. 1977).
Uncoupled reaction cycles may be caused by incorrect conformation of the substrate at
the catalytic site, and some peptides, such as poly(L-proline) increase the rate of the
uncoupled reaction (Myllylä et al. 1984). In addition, the biological substrates of the C4Hs contain non-hydroxylatable sequences, and when the catalytic site encounters such a
sequence uncoupled decarboxylation of 2-oxoglutarate takes place (Tuderman et al.
1977, Myllylä et al. 1984).
Fig. 3. The reactions catalyzed by prolyl 4-hydroxylase. The 2-oxoglutarate is
stoichiometrically decarboxylated during proline hydroxylation (A). In an uncoupled reaction
ascorbate is stoichiometrically consumed either in the presence (B) or absence (C) of the
substrate.
40
2.2.7 Human HIF prolyl 4-hydroxylases
As indicated above, under normoxic conditions the HIF-α subunit is targeted for
degradation by a ubiquitin-ligase complex that recognizes a hydroxylated proline residue
(Ivan et al. 2001, Jaakkola et al. 2001). This post-translational modification is performed
by a family of novel HIF-P4Hs forming the egl-nine (egln) gene family (Bruick &
McKnight 2001, Epstein et al. 2001, Ivan et al. 2002).
The well-characterized C-P4Hs do not hydroxylate proline residues in the HIF-α
subunit, because the substrate contexts of the modified proline residues in collagens are
different from those surrounding the proline residues in HIF-α (Jaakkola et al. 2001). The
hydroxylated prolines in collagens are located in -X-Pro-Gly- sequences (for a review,
see Myllyharju & Kivirikko 2004), while the prolines in HIF-1α are present in -Leu-XX-Leu-Ala-Pro402-Ala- and -Leu-X-X-Leu-Ala-Pro564-Tyr- sequences (Ivan et al. 2001,
Jaakkola et al. 2001, Masson et al. 2001). In addition, the C-P4Hs are located within the
ER while the enzymes hydroxylating HIF-α are cytoplasmic or nuclear (Huang et al.
2002, Metzen et al. 2003).
The HIF-P4Hs were identified by searches of the C. elegans and mammalian
databases for candidates having the predicted β barrel jelly roll motif common to the 2oxoglutarate dioxygenases and the iron and 2-oxoglutarate binding sites in the same
relationship as found in other members of the enzyme family (Bruick & McKnight 2001,
Epstein et al. 2001). The three human protein products were initially named prolyl
hydroxylase domain (PHD) enzymes 1, 2 and 3, Egl-nine (EGLN) 2, 1 and 3 or HIF
prolyl hydroxylases (HPHs) 3, 2 and 1, but are called here HIF-P4Hs 1, 2 and 3,
respectively. These enzymes produced in a rabbit reticulocyte lysate in vitro transcription
translation system were shown to be active and to promote the capture of pVHL by HIF1α (Epstein et al. 2001, Ivan et al. 2002). In another study the human HIF-P4Hs were
expressed in bacterial cells and purified enzymes were shown to hydroxylate a HIF-1α
peptide. The enzyme activity was greatly increased by addition of the known cosubstrates
of C-P4Hs, ascorbate, iron and 2-oxoglutarate, and inhibited by CoCl2. These data
suggested that the HIF-P4Hs found by database query are the prolyl 4-hydroxylases
modifying HIF-1α. (Bruick & McKnight 2001.)
2.2.7.1 Biological functions
When a normal oxygen concentration is available for cells, at least one of the two
prolines, Pro402 (Masson et al. 2001) and Pro564 (Jaakkola et al. 2001), in the ODDD of
HIF-1α is hydroxylated by the HIF-P4H isoenzymes, leading to binding of HIF-1α to
pVHL and subsequent proteasomal degradation.
Although all three HIF-P4Hs can hydroxylate HIF-1α in vitro, their in vivo specificity
for HIF-1α differs. It has been reported that either HIF-P4H-2 (Huang et al. 2002, Berra
et al. 2003, Metzen et al. 2003, Appelhoff et al. 2004) or HIF-P4H-3 (Bruick &
McKnight 2001) has the highest specific activity. HIF-P4H-2 is the most abundant in
41
most of the cell types studied in normoxia, and its silencing alone is sufficient to induce
HIF-1α in normoxia and increase its target gene expression (Berra et al. 2003, Appelhoff
et al. 2004). Phylogenetic analysis also shows that HIF-P4H-2 is the ancestral form of the
egln gene family (Taylor 2001). It is therefore likely that HIF-P4H-2 is the key enzyme
hydroxylating HIF-1α in vivo, but that isoenzymes 1 and 3 may also regulate HIF-1α in
some tissues or under certain physiological conditions. It is likely that HIF-P4H-3
regulates mainly HIF-2α, especially under hypoxic conditions, in which this enzyme has
been shown to retain significant activity (Appelhoff et al. 2004).
It is unknown to date whether HIF-P4Hs can hydroxylate proline residues in other
proteins in addition to the HIF-αs, but it is significant that HIF-P4H-1 is induced by
oestrogens and stimulates cell proliferation (Seth et al. 2002). Also, the rat homologue of
human isoenzyme 3, SM-20, was first identified as a transcript induced by plateletderived growth factor in smooth muscle cells (Wax et al. 1994) and as a mitochondrial
protein involved in apoptosis (Lipscomb et al. 2001) and growth regulation (Madden et
al. 1996). Furthermore, the C. elegans HIF-P4H EGL-9 was first identified as being
responsible for an egg-laying defect (Trent et al. 1983).
The large subunit of RNA polymerase II (Rpb1) is involved in the transition from
transcription to elongation. Regions of Rpb1 and its adjacent subunit Rpb6 share
sequence and structural similarity with the pVHL-binding domain of the ODDD of HIF1α (Kuznetsova et al. 2003). pVHL becomes bound to hyperphosphorylated Rpb1 in a
proline hydroxylation-dependent manner, the formation of the pVHL-Rpb1 complex
being increased in the presence of Fe2+, 2-oxoglutarate and ascorbate and being inhibited
in the presence of iron chelators and known inhibitors of the C-P4Hs. Binding of pVHL
to phosphorylated Rpb1 leads to ubiquitination of Rpb1 and its degradation, and it has
been suggested that the ubiquitination of Rpb1 might have a role in the regulation of
efficient elongation through elongation-pause and arrest of specific genes (Kuznetsova et
al. 2003). Even though the mechanism of targeting the Rpb1 for ubiquitination and
subsequent degradation is similar to that of the HIF-1α, it is not known whether the
proline hydroxylation of Rpb1 is due to HIF-P4Hs or some so far unidentified P4H.
2.2.7.2 Gene expression and regulation
The three HIF-P4Hs have unique but overlapping tissue expression patterns. In normoxia
the mRNA for HIF-P4H-1 is highly expressed in the placenta and with slightly lower
levels in the brain, liver, heart and adipose tissue (Oehme et al. 2002, Cioffi et al. 2003).
The HIF-P4H-2 mRNA is abundantly expressed in many tissues, with the highest
expression observed in the heart and adipose tissue (Lieb et al. 2002, Oehme et al. 2002,
Cioffi et al. 2003), while the level of expression of HIF-P4H-3 mRNA in general is rather
low in most tissues (del Peso et al. 2003, Appelhoff et al. 2004), with highest expression
in the heart and placenta (Oehme et al. 2002, Cioffi et al. 2003). A striking difference in
the expression patterns of the isoenzymes is seen in the testis, where the mRNA for HIFP4H-1 is highly expressed whereas the expression of the other two isoenzymes is barely
detectable (Lieb et al. 2002, Oehme et al. 2002). The protein expression levels of HIFP4Hs 2 and 3 parallel those of the mRNA, but the protein level of isoenzyme 1 is lower
42
than would be predicted from the mRNA levels (Appelhoff et al. 2004).
The mRNA and protein expression levels of HIF-P4Hs 2 and 3 are induced by
hypoxia, suggesting a possible role for these isoenzymes in a negative feedback pathway
responsible for enhanced degradation of HIF-α upon reoxygenation after prolonged
hypoxia (Epstein et al. 2001, Berra et al. 2003, Cioffi et al. 2003, D’Angelo et al. 2003,
del Peso et al. 2003, Metzen et al. 2003, Appelhoff et al. 2004, Marxsen et al. 2004),
whereas the mRNA and corresponding protein levels of HIF-P4H-1 remain unchanged or
are even reduced by hypoxia (Appelhoff et al. 2004). This hypoxic induction might be
mediated by the HIF transcriptional activity, suggesting that HIF may control the stability
of its own α subunit by controlling the expression of HIF-P4Hs 2 and 3 (Berra et al.
2003, Cioffi et al. 2003). In support of this hypothesis HIF-P4Hs 2 and 3 were found to
be target genes of HIF-α (del Peso et al. 2003, Marxsen et al. 2004), HIF-1α inducing
specifically HIF-P4H-2 and HIF-2α regulating HIF-P4H-3 (Aprelikova et al. 2004).
In addition to oxygen, some other physiological stimuli are also involved in the
regulation of the expression of HIF-P4Hs. Oestrogen causes induction of HIF-P4H-1
(Seth et al. 2002), and the glycolytic and citric acid cycle metabolites pyryvate and
oxaloacetate up-regulate the mRNA levels of isoenzymes 2 and 3 (Dalgard et al. 2004).
These metabolites also stabilize HIF-1α and increase HIF-mediated gene expression, and
may therefore regulate another feedback loop limiting the induction of gene expression in
hypoxia and in adaptation to re-oxygenation (Dalgard et al. 2004). HIF-P4H-3 is likely to
be a substrate of the cytosolic TRiC chaperonin, which is known to form a complex with
pVHL, raising a possibility that the induction of HIF-P4H-3 may have an effect on this
pVHL-TRiC association (Masson et al. 2004).
As mentioned above, Siah2, which is an E3 ubiquitin ligase, regulates the stability of
HIF-P4H isoenzymes 1 and 3 during hypoxia (Nakayama et al. 2004). Even in mild
hypoxia these enzymes become degraded in proteasomes after ubiquitination, as a result
of which the level of HIF-1α expression is elevated. The association between Siah2 and
HIF-P4H-2 is very weak, and it can be concluded from the results of other studies that
isoenzyme 2 probably regulates HIF-1α in normoxia while the other two isoenzymes,
mainly HIF-P4H-3, regulate its availability under hypoxic conditions (Berra et al. 2003,
Appelhoff et al. 2004, Nakayama et al. 2004). The expression of Siah2 is induced upon
exposure to even moderate hypoxia, and therefore Siah2 inhibits HIF-P4H activity under
conditions where there could still be sufficient amounts of oxygen for enzyme activity
(Nakayama et al. 2004).
2.2.7.3 Molecular properties and cellular location
The polypeptides of the HIF-P4H isoenzymes 1, 2 and 3 consist of 407, 426 and 239
amino acids, respectively (Taylor 2001). These three isoenzymes show a 42-59%
sequence identity to each other but no significant sequence similarity to the C-P4Hs.
Secondary structure predictions indicate that the HIF-P4Hs probably form the common
jelly roll structure (eight β strands) with the critical Fe2+-binding residues, the two
histidines and one aspartate identified in the C-P4Hs as a conserved -His-X-Asp-…-Hismotif (His297, Asp299 and His358 in HIF-P4H-1). In the HIF-P4Hs the residue that binds the
43
C-5 carboxyl group of the 2-oxoglutarate is an arginine in position +9 with respect to the
second iron-binding histidine whereas in the C-P4Hs it is a lysine in position +10 (Bruick
& McKnight 2001, Epstein et al. 2001, Ivan et al. 2002.)
The HIF-P4Hs are cytoplasmic and nuclear proteins, HIF-P4H-1 being located
exclusively in the nucleus, HIF-P4H-2 being mainly present in the cytoplasm and HIFP4H-3 being found in both the cytoplasm and nucleus (Huang et al. 2002, Metzen et al.
2003). The existence of a fourth isoenzyme possessing a transmembrane domain has been
postulated, and it has been predicted that this protein may be located in the ER
membranes in an orientation in which its catalytic site is in the cytoplasm (Oehme et al.
2002). It has not been directly shown to hydroxylate HIF-αs, however.
2.2.7.4 Catalytic properties
The HIF-P4Hs belong to the 2-oxoglutarate dioxygenases and thus require Fe2+, 2oxoglutarate, O2 and ascorbate. One oxygen atom is incorporated into the substrate
forming the hydroxyl group, while the other oxygen atom is incorporated into the
succinate formed in the decarboxylation of the 2-oxoglutarate (Lando et al. 2003), the
reaction mechanism thus being similar to that of the C-P4Hs (see section 2.2.6).
After hydroxylation of at least one of the two proline residues, HIF-1α becomes bound
to the pVHL in an extended conformation. In the case of the C-terminal hydroxylation
site, residues 560-567 (-Glu-Met-Leu-Ala-4Hyp-Tyr-Ile-Pro-) and 571-577 (-Asp-PheGln-Leu-Arg-Ser-Phe-) form two distinct binding sites, the first being the primary
binding site, and it is likely that pVHL may interact similarly with the N-terminal site
(Hon et al. 2002, Min et al. 2002). It seems also very likely that the substrate region of
HIF-α becomes bound to the HIF-P4Hs in a similar extended conformation. Two distinct
proline residues, Pro402 and Pro564 in HIF-1α, both located in -Leu-X-X-Leu-Ala-Prosequence motifs, are hydroxylated (Ivan et al. 2001, Jaakkola et al. 2001, Masson et al.
2001), HIF-2α and HIF-3α being subject to similar hydroxylation and regulation to HIF1α (Chan et al. 2002, Maynard et al. 2003). The C-terminal hydroxylation site in both
HIF-1α and HIF-2α is hydroxylated most effectively by HIF-P4H-3 (Appelhoff et al.
2004). In contrast, HIF-P4H-3 hydroxylates the N-terminal proline in both of the HIF-αs
ineffectively or not at all (Epstein et al. 2001, Appelhoff et al. 2004). No marked
difference can be seen in the efficiency of the other two isoenzymes with respect to these
C and N-terminal prolines in HIF-1α and HIF-2α (Appelhoff et al. 2004).
The HIF-P4Hs require a stretch of about 20 HIF amino acids for selective recognition
(Ivan et al. 2001, Jaakkola et al. 2001, Yu et al. 2001) as HIF-1α peptides of this length
can support site-specific proline hydroxylation and subsequent binding to pVHL. It seems
likely that not all these 20 residues are essential for substrate recognition by the HIFP4Hs, as mutation studies have unexpectedly revealed that probably the only fully
obligatory residue for proline hydroxylation is the proline itself, while mutations were
tolerated at -5, -2 and -1 positions relative to the proline in the -Leu-X-X-Leu-Ala-Promotif (Huang et al. 2002). However, the proline itself is not sufficient for hydroxylation,
as Pro567 in HIF-1α is not hydroxylated even though it is only three residues away from
the proline targeted by the HIF-P4Hs (Ivan et al. 2001, Jaakkola et al. 2001).
44
Furthermore, the acidic residues C-terminal to proline appear to be important for
recognition (Huang et al. 2002). Leucine 574 is required for the hydroxylation of Pro564,
but it does not affect the hydroxylation of the N-terminal Pro402 (Kageyama et al. 2004).
Mutation of this residue inhibits, but does not completely eliminate, prolyl hydroxylation,
most likely due to involvement of additional residues in hydroxylation. Interestingly,
Leu574 is required for the interaction with HIF-P4H-2 (Kageyama et al. 2004), which is
thought to be the major prolyl 4-hydroxylase controlling the normoxic degradation of
HIF-1α (Berra et al. 2003).
Several small molecule inhibitors of the C-P4Hs, such as the 2-oxoglutarate analogues
N-oxalylglycine and dimethyloxalylglycine, inhibit the activity of HIF-P4Hs in vitro and
in the nematode C. elegans in vivo at low molecular concentrations when tested in HIF1α and pVHL capture assays (Epstein et al. 2001, Jaakkola et al. 2001, Ivan et al. 2002,
Mole et al. 2003). Iron chelators and transition metals such as Co2+ have also been
reported to inhibit hydroxylase activity (Epstein et al. 2001).
2.2.8 Nematode and Drosophila HIF-P4Hs
The HIF-pVHL-P4H system also exists in the nematode C. elegans, suggesting that it
evolved before the development of the systemic oxygen supply systems to regulate
oxygen homeostasis at the cellular level. C. elegans HIF-P4H is a product of the egl-9
gene (Epstein et al. 2001), which has previously been identified as a gene involved in the
egg-laying defect phenotype (Trent et al. 1983). Analysis of mutant worms with defective
egl-9 alleles demonstrated that EGL-9 is involved in the regulation of C. elegans HIF
(CeHIF), and pVHL capture assays showed that EGL-9 functions as the HIF-P4H
targeting HIF-α to pVHL, this activity being dependent on 2-oxoglutarate, iron and
oxygen and inhibited by Co2+. The critical hydroxylated residue in CeHIF-α, Pro621, is
present in the same -Leu-X-X-Leu-Ala-Pro- sequence as the hydroxylatable proline
residues in the human HIF-α subunits, but the C. elegans and human enzymes are not
functionally interchangeable, implying differences in substrate specificity. C. elegans
probably has only one type of HIF-P4H, because mutant egl-9 led to total loss of CeHIF
regulation. (Epstein et al. 2001.)
The fruit fly D. melanogaster has a hypoxic response pathway similar to that in
mammalian cells, as suggested by the identification of the homologues of HIF-1α (Sima)
(Bacon et al. 1998) and the β subunit (dHIF-1beta) (Ma & Haddad 1999) and pVHL
(dVHL) (Adryan et al. 2000). Supresssion of the D. melanogaster HIF-P4H under
normoxic conditions resulted in up-regulation of Sima and subsequent elevation of the
expression of a hypoxia-inducible gene, indicating that HIF-P4H also regulates the HIFdependent hypoxia response pathway in Drosophila (Bruick & McKnight 2001, LavistaLlanos et al. 2002). The HIF-P4H of D. melanogaster is required for the cyclindependent protein kinase complex Cyclin D/Cdk4-induced cell growth, suggesting that
HIF-P4H is also a regulator of cell growth and a mediator for CycD/Cdk4. HIF-P4H
functions downstream of CycD/Cdk4 and is sufficient to increase growth when
overexpressed, indicating that CycD/Cdk4 and HIF-P4H function in a common pathway.
45
It is possible that Cdk4 phosphorylates HIF-P4H, and growth induction probably requires
the hydroxylation activity of HIF-P4H. (Frei 2004, Frei & Edgar 2004.)
2.3 HIF asparaginyl hydroxylase
2.3.1 Biological function
The activity of HIF-α in mammals is also regulated by transcriptional activation in an
oxygen-dependent manner. Under normoxic conditions hydroxylation of a specific
asparagine residue in the CAD, Asn803 in human HIF-1α or Asn851 in human HIF-2α,
blocks the interaction of HIF-α with the transcriptional coactivator p300, thus preventing
transcriptional activation (Lando et al. 2002b). The enzyme responsible for
hydroxylation, asparaginyl hydroxylase is identical to a previously characterized protein
known as factor inhibiting HIF (FIH) (Mahon et al. 2001, Hewitson et al. 2002, Lando et
al. 2002a).
FIH is widely expressed at similar levels in different human tissues, its expression,
unlike that of HIF-P4Hs 2 and 3, being unaffected by hypoxia (Stolze et al. 2004). Full
hypoxic induction of HIF activity requires inhibition of FIH activity, leading to enhanced
transcriptional activity of HIF, in parallel with inhibition of the HIF-P4Hs, leading to
stabilization of HIF-1α. It has been suggested that the active site of FIH may not act
directly as the oxygen sensor but that the oxygen recognition mechanism might require a
complex formed by HIF-1α and VHL (Mahon et al. 2001, Lando et al. 2003, Lee et al.
2003). FIH is active over a wide range of oxygen concentrations, from normoxia even to
severe hypoxia (Mahon et al. 2001, Lando et al. 2002a, Stolze et al. 2004). This contrasts
with the functioning of the HIF-P4Hs, which do not reduce the expression of HIF target
genes at oxygen concentrations below 2%. The result indicates that FIH and HIF-P4Hs
have different oxygen-dependent characteristics in vivo. (Stolze et al. 2004.) The cellular
localization of FIH is in the cytoplasm, and the substrates of FIH, HIF-1α and HIF-2α,
are cytoplasmic under normoxic conditions, being translocated into the nucleus in
hypoxia. FIH does not co-localize in the nucleus along with the HIF-α subunits, however
(Linke et al. 2004, Stolze et al. 2004), suggesting that asparagine hydroxylation of HIF-α
occurs in the cytoplasm prior to its nuclear entry (Stolze et al. 2004).
It has been suggested that FIH may act in fine-tuning the HIF response, since its
effects are modest by comparison with those of HIF-P4Hs in normoxia, and in hypoxia
FIH regulates the function of the hypoxia-inducible HIF-P4Hs 2 and 3 but not the HIFP4H-1, which is not affected by hypoxia, and therefore FIH may indirectly affect the
stability of HIF-α (Stolze et al. 2004).
2.3.2 Structural properties
FIH is a 349-amino-acid protein expressed in multiple cell types derived from brain,
46
embryo, heart, kidney, lung, muscle and skin, for example (Mahon et al. 2001). FIH
possesses the jellyroll core and the conserved iron-binding –His-X-(Asp/Glu)-…-Hismotif common to the other 2-oxoglutarate dioxygenases, but its overall structure differs
significantly from those of the HIF-P4Hs and the other 2-oxoglutarate dioxygenases
(Dann et al. 2002, Hewitson et al. 2002, Lando et al. 2002b, Elkins et al. 2003, Lee et al.
2003). The jellyroll motif is formed from β strands 8-11 and 14-17, and four additional β
strands form an antiparallel β sheet that flanks the jellyroll core, while β1, β2 and α1
form a sheet-helix-sheet motif that is conserved in almost all the 2-oxoglutarate
dioxygenases (Elkins et al. 2003). FIH has two separate domains: domain I, comprising
residues 1-299, and domain II, residues 300-349 (Lee et al. 2003). Domain I is composed
of the central β barrel, which is surrounded by six α helices, while domain II consists of
two C-terminal α helices. FIH is a dimer, with the dimer interface involving the two Cterminal helices of each molecule in an interlocking arrangement stabilized by
hydrophobic interactions (Dann et al. 2002, Elkins et al. 2003, Lee et al. 2003). This
dimeric structure is required for its catalytic activity (Dann et al. 2002, Lancaster et al.
2004).
At the centre of the β barrel structure is a large cavity lined with the residues involved
in the coordination of the ferrous ion, His199, Asp201 and His279, one side of this facial
triad having an opening for the 2-oxoglutarate that is bound by Lys214 (Dann et al. 2002,
Elkins et al. 2003, Lee et al. 2003). Thus the binding site for 2-oxoglutarate in FIH is
different from that in the other 2-oxoglutarate dioxygenases in which the 2-oxoglutarate
binding residue is an arginine, with the exception of the C-P4Hs, where it is also a lysine,
in position +10 with respect to the second iron-binding histidine (Myllyharju & Kivirikko
1997, Myllyharju 2003). A long groove extends from the active site towards the
interconnecting loop between domains I and II lined by residues, many of which are
hydrophobic, this groove acting as the binding site for HIF-α (Elkins et al. 2003, Lee et
al. 2003). The asparagine to be hydroxylated in HIF-α is sandwiched between the FIH
residue Tyr102 and the bound ferrous ion and forms hydrogen bonds with the side chains
of the FIH residues Arg238 and Gln239 and six additional hydrogen bonds stabilizing the
binding of FIH to HIF-1α (Elkins et al. 2003). FIH contains a distinct binding site for
pVHL, which is located N-terminal to the HIF-α binding site. This pVHL binding site is
within the first of the two segments of FIH (segments 1 and 2 comprise residues 1-89 and
90-126, respectively) and FIH may provide the link for interaction between pVHL and
the HIF-1α CAD (Lee et al. 2003). This finding is in agreement with other data showing
that pVHL interacts with the HIF-1α CAD only in the presence of FIH (Mahon et al.
2001).
2.3.3 Catalytic properties and substrate specificity
FIH belongs to the same family of 2-oxoglutarate dioxygenases as the P4Hs and thus
requires iron, 2-oxoglutarate, molecular oxygen and ascorbate (Hewitson et al. 2002).
The Km values of FIH for a 52-amino-acid peptide corresponding to HIF-1α CAD
residues 775-826 and for 2-oxoglutarate, when measured using a 2-oxoglutarate
decarboxylation assay, are both 10 µM (Hewitson et al. 2002). FIH is inhibited by N-
47
oxalylglycine, which is a known inhibitor of the HIF-P4Hs and other 2-oxoglutarate
dioxygenases, and also by Co2+ and Zn2+ (Hewitson et al. 2002). The requirement for
dioxygen as a cosubstrate indicates that FIH may act as a cellular oxygen sensor, and it
has been shown that when the atmospheric oxygen concentration is below 5% there is a
clear reduction in HIF-1α hydroxylation (Hewitson et al. 2002). This suggests that the
activity of FIH, like those of the HIF-P4Hs, is limited by the oxygen concentration. Other
studies have revealed, however, that FIH is active even under severely hypoxic conditions
(0.2% O2) (Mahon et al. 2001, Lando et al. 2002a, Stolze et al. 2004).
The only known substrates of FIH are HIF-1α and HIF-2α. Interaction assays have
shown that FIH becomes bound to the CAD of HIF-1α in a region that does not contain
the hydroxylatable asparagine residue (Mahon et al. 2001). Instead, the asparagine is
located 20-30 residues C-terminal to the putative FIH binding region, suggesting that in
order to hydroxylate the asparagine efficiently, FIH has to become bound to a region
adjacent to the asparagine motif. A region surrounding an -Arg-Leu-Leu- motif found in
both HIF-1α and HIF-2α might be important for FIH targeting (Lando et al. 2003), the
motif being critical for silencing the HIF-1α CAD in normoxia (O’Rourke et al. 1999).
Structural studies have shown the existence of two distinct FIH-CAD interaction sites,
one contains the hydroxylation site (CAD795-806) and the other is more C-terminal,
containing residues 813-822, kinetic analyses suggesting a weaker binding at CAD813-822.
The crystal structure of FIH also reveals that the region exactly around the hydroxylation
site of CAD does not form direct interactions with FIH, and in agreement with this
finding, CAD peptides longer than 20 residues are required for FIH to become bound to
CAD (Elkins et al. 2003). Mutation studies have shown that Val802 is essential for
effective catalysis of hydroxylation but does not significantly affect substrate binding
(Linke et al. 2004). Mutation of Val802 to Ala caused local structural rearrangements
around Asn803, confirming the importance of Val802 in the reaction catalyzed by FIH,
while other residues around the Asn803 are not individually essential for hydroxylation
(Linke et al. 2004). FIH is highly selective for asparagine over aspartic acid, with only a
very low activity obtained in 2-oxoglutarate turnover assays upon mutation of Asn803 to
Asp803 (Hewitson et al. 2002). The hydroxylation occurs at the β carbon of the asparagine
residue, producing a threo-isomer, in contrast to the product of the other known aspartic
acid/asparagine hydroxylases, which is an erythro-isomer (McNeill et al. 2002a).
2.4 Aspartyl (asparaginyl) β hydroxylase
The asparaginyl hydroxylase FIH involved in the regulation of HIF activity is distinct
from the previously characterized aspartyl (asparaginyl) β hydroxylase (AAH). AAH
hydroxylates specific aspartic acid and asparagine residues at the β position within
epidermal growth factor-like domains in several proteins, including receptors and
receptor ligands, and in extracellular matrix molecules (Wang et al. 1991, Lavaissiere et
al. 1996). As AAH hydroxylates receptors involved in cell growth and differentiation, it
is likely that its overexpression will be associated with or contribute to malignant
transformation by promoting proliferation and cell cycle progression, inhibiting apoptosis
and enhancing tumour cell migration (Lavaissiere et al. 1996, Ince et al. 2000, Sepe et al.
48
2002).
AAH likewise belongs to the enzyme family of 2-oxoglutarate dioxygenases, its
reaction thus requiring Fe2+, 2-oxoglutarate, molecular oxygen and ascorbate (Wang et al.
1991). The human AAH consists of 757 amino acids and contains four distinct regions
(Korioth et al. 1994, Dinchuk et al. 2000). AAH is a type 2 integral membrane protein
with a short N-terminal domain projecting into the cytoplasm. The N-terminal domain is
followed by a transmembrane domain and a highly charged lumenal region. The Cterminal region contains the aspartyl hydroxylase catalytic domain, which possesses
dibasic glycine and histidine motifs critical for catalytic activity (Jia et al. 1994,
McGinnis et al. 1996).
3 Outlines of the present research
The HIF-P4Hs play a major role in the regulation of oxygen homeostasis, as they
hydroxylate proline residues in HIF-α, this hydroxylation leading to capture of HIF-α by
the pVHL and subsequent proteasomal degradation. Hypoxia inhibits these
hydroxylations so that HIF-α escapes degradation and is translocated into the nucleus,
where it forms an active dimer with HIF-β and activates the transcription of a number of
hypoxia-inducible genes. FIH is another enzyme involved in the regulation of HIF, as
hydroxylation of a specific asparagine residue in HIF-α prevents binding of the
transcriptional coactivator p300, thus inhibiting the full transcriptional activity of HIF.
The HIF-P4Hs and FIH have been identified only recently and have been found to belong
to the enzyme family of 2-oxoglutarate dioxygenases. When this work was initiated, only
limited data were available on the kinetic and inhibitory properties of these enzymes. One
objective of the present work was therefore to express HIF-P4H isoenzymes and FIH as
recombinant proteins for their kinetic characterization, the specific aims being:
1. to express the HIF-P4Hs as recombinant enzymes, determine their catalytic and
inhibitory characteristics and compare them with those of the well-characterized CP4Hs,
2. to express FIH as a recombinant enzyme, determine its catalytic and inhibitory
characteristics and compare them with those of the HIF-P4Hs and C-P4H,
3. to purify the recombinant HIF-P4Hs and FIH to determine their catalytic centre
activities,
4. to study the peptide substrate requirements of the HIF-P4Hs and FIH, and
5. to determine the effects of different divalent metal ions on the activities of the HIFP4Hs and FIH.
4 Materials and methods
Detailed descriptions of the materials and methods are presented in the original papers IIV.
4.1 Cloning and expression of recombinant HIF-P4Hs and FIH (I-IV)
4.1.1 Cloning and generation of expression vectors
The cDNAs encoding the human HIF-P4Hs 1, 2 and 3 were obtained by PCR
amplification from colon, aorta and lung Marathon-Ready cDNAs (Clontech). Full-length
HIF-P4H-1 and 3 cDNAs were obtained, but the PCR product of HIF-P4H-2 had an
insertion of 51 nucleotides replacing the sequence coding for amino acids Cys58-Pro175.
This product was named the HIF-P4H-2 PCR variant. In order to generate a full-length
HIF-P4H-2 cDNA, exon 1 of HIF-P4H-2 containing the missing region was amplified
from genomic DNA. The full-length HIF-P4H 1-3 cDNAs and the HIF-P4H-2 PCR
variant were cloned into the pVL and pAcG3X (N-terminal GST-tag) baculovirus
expression vectors (Pharmingen), and the HIF-P4H-3 cDNA was also cloned into a
pBlueBacHis (N-terminal His-tag) baculovirus vector (Pharmingen). For expression in E.
coli, the HIF-P4H-1 and 3 cDNAs were cloned into a pMALc expression vector (fusion
with N-terminal maltose binding protein) (New England Biolabs). Site-directed
mutagenesis of the HIF-P4H-1 cDNA was performed using QuickChange™ mutagenesis
(Stratagene).
The full-length FIH cDNA was similarly obtained by PCR from Marathon-Ready
cDNAs and cloned into pVL1392, pBlueBacHis (N-terminal His tag) and pAcG3X (Nterminal GST tag) baculovirus expression vectors (Pharmingen).
In order to generate FLAGHis-tagged HIF-P4Hs and FIH and V5His-tagged FIH,
FLAGHis and V5His tags amplified from plasmids d28e6 and d28e5 (FibroGen Inc.),
respectively, were cloned into the 3’ ends before the stop codons of the HIF-P4H-1-3 and
FIH cDNAs in the pVL vectors.
51
4.1.2 Expression of the recombinant enzymes
The recombinant baculovirus expression vectors were co-transfected into Spodoptera
frugiperda Sf9 cells with BaculoGold DNA (Pharmingen) by calcium phosphate
transfection, and the recombinant viruses were amplified. Sf9 or High Five insect cells
(Invitrogen) were cultured as monolayers in TNM-FH medium (Sigma) supplemented
with 10% foetal bovine serum (BioClear), or alternatively the High Five cells were
cultured in suspension in Sf900IISFM serum-free medium (Gibco). To produce
recombinant proteins, cells seeded at a density of 5 x 106/100-mm plate or 0.5-1 x 106/ml
were infected with the recombinant viruses at a multiplicity of 5. The cells were
harvested 72 h after infection, homogenized in a buffer containing Triton X-100 and
centrifuged. The pellets were further solubilized in 1% SDS, and aliquots of all the
soluble fractions were analysed by SDS-PAGE under reducing conditions followed by
Coomassie Blue staining or Western blotting.
For expression of the HIF-P4H isoenzymes 1 and 3 in E. coli, the recombinant
expression plasmids were transformed into the E. coli BL21(DE3) strain (Novagen). The
maltose-binding protein fusion polypeptides were expressed according to the pMAL
protein fusion and purification system manual (New England Biolabs).
4.1.3 Purification of the recombinant enzymes
Recombinant HIF-P4Hs and FIH with a C-terminal FLAGHis tag were purified by antiFLAG M2 affinity chromatography (Sigma). Triton X-100 extracts of insect cells
expressing the recombinant enzymes were passed through the Anti-FLAG M2 column
several times and allowed to react with the affinity material for 1 h before elution with a
buffer containing 100 µg/ml of the FLAG peptide (Sigma). The elution buffer was kept in
the column for 30 min before the fractions were collected. In the case of the HIF-P4Hs, 1
or 5 µM FeSO4 was added to all buffers during the purification to prevent inactivation.
His-tagged HIF-P4H-3 and FIH-FLAGHis were also purified by chelating Sepharose
charged with Ni2+ (ProBond, Invitrogen). The Triton X-100 extract was allowed to react
with the Sepharose for 1 h before elution with a gradient of 0.02-0.5 M imidazole. The
fractions collected were analysed by SDS-PAGE under reducing conditions followed by
Coomassie Blue staining. Fractions containing the purified enzymes were pooled, and in
the case of the Ni2+-charged chelating Sepharose-purified enzymes, the imidazole was
removed by changing the buffer using an Ultrafree-15 centrifugal filter device with a
Biomax-10 filter (Millipore). The concentrations of the purified enzymes were
determined (Roti-Quant, Carl Roth). GST-tagged HIF-P4Hs were purified with GSTrap
FF chromatography according to the manufacturer’s instructions (Amersham
Biosciences). The maltose-binding protein fusion polypeptides expressed in E. coli were
purified according to the pMAL protein purification manual (Amersham Biosciences).
SP sepharose chromatography was used to partially purify the non-tagged HIF-P4H-2.
The Triton X-100-soluble cell fraction was diluted 1:4 in 20 mM HEPES buffer, pH 6.5,
and the sample was loaded into an SP Fast Flow column (Amersham Biosciences). HIFP4H-2 was eluted at a NaCl concentration of 200 mM.
52
4.2 Characterization of the catalytic and inhibitory properties of
HIF-P4Hs and FIH (I-IV)
Enzyme activity was assayed by a modified method based on the hydroxylation-coupled
decarboxylation of 2-oxo[1-14C]glutarate (Kivirikko & Myllylä 1982). In the HIF-P4H
activity assays, 1 ml of the reaction mixture contained 5 µM FeSO4, 160 µM 2oxoglutarate, 2 mM ascorbate, 60 µg catalase, 100 µM dithiothreitol, 2 mg bovine serum
albumin and 50 mM Tris-HCl buffer, adjusted to pH 7.8 at 25°C, and 25-50 µM of a
synthetic 19-amino-acid peptide corresponding to the C-terminal hydroxylation site of
HIF-1α as a substrate. The same reaction mixture was used in the FIH activity assays
with the exceptions that 5-10 µM FeSO4, 100 µM 2-oxoglutarate and 200 µM of a
synthetic 35-amino-acid peptide corresponding to the asparaginylation site of HIF-1α as
a substrate were used. The peptide substrates were obtained from Innovagen, SynPep or
Mimotopes.
The Km values of HIF-P4Hs and FIH for the various HIF peptide substrates and the
cosubstrates were determined by varying the concentration of a substrate or cosubstrate
while the concentrations of the others were kept constant at saturating levels, with the
exception that the concentration of O2 was that of air and thus non-saturating. For
determination of the Km values for O2, 1-45% O2 gas mixtures (AGA) were used to
equilibrate the reaction mixtures.
Ki values for 2-oxoglutarate analogues were determined by adding the inhibitors at 4-5
constant concentrations, whereas the concentration of 2-oxoglutarate was varied. The
IC50 values of the HIF-P4Hs for different metal ions were determined by varying the
metal concentrations while the concentrations of the other reaction components were kept
constant. The effect of increasing concentrations of the iron chelator desferrioxamine
(DFO) on the HIF-P4Hs was studied at a Fe2+ concentration of 5 µM. The Ki values of
FIH for metals were determined by adding the metal ions at 4-5 constant concentrations
and varying the concentration of Fe2+ or the peptide substrate. The IC50 values of the
recombinant type I C-P4H for metals were determined using the standard C-P4H activity
assay with (Pro-Pro-Gly)10 (Peptide Institute) as a substrate (Kivirikko & Myllylä 1982).
The inactivation of HIF-P4H-2 and 3 by cobalt was studied by incubating Triton X100 extracts of insect cells expressing these enzymes in the presence of 100 µM cobalt
for 4 and 24 h at 4°C. In control experiments the cell extracts were incubated in a buffer
containing no metal or in the presence of 100 µM iron. The final cobalt concentration in
the enzyme activity assay was 2 µM, and the Fe2+ concentration was 5 µM, with the
exception of the control sample preincubated in the presence of Fe2+, in which case the
final iron concentration was 7 µM.
The activity of a recombinant HIF-P4H-2 synthesized in the presence of cobalt or
DFO was studied by expressing the enzyme in insect cells for 48 h in the presence of
100-1000 µM CoCl2 or 100-300 µM DFO. After cell harvesting, the soluble cell extracts
were used directly for enzyme activity assays in a 1:15 dilution using reaction mixtures
containing no added Fe2+, or 50 or 600 µM Fe2+. The amount of enzyme present in the
soluble fractions was analysed by reducing SDS-PAGE followed by Western blotting
with a HIF-P4H-2 antibody (anti-PHD2, Bethyl Laboratories) and ECL™
immunodetection reagents (Amersham Biosciences). The levels of HIF-P4H-2 expression
53
were measured by densitometry of the ECL films using a GS-710 calibrated imaging
densitometer (Bio-Rad).
4.3 Liquid chromatography/mass spectrometry analysis
of peptide hydroxylation (III)
Hydroxylation of a HIF-1α peptide 554DTDLDLEALAPYIPADDDFQ573 or peptides in
which 3,4-dehydroproline or L-azetidine-2-carboxylic acid was substituted for the
hydroxylatable proline was performed by the standard protocol with minor changes in the
concentrations of 2-oxoglutarate and the substrates and with a partially purified HIFP4H-2 as the enzyme. The reaction was carried out for 4 h at 37°C, after which
hydroxylation of the peptides was analysed by LC-MS using a Finnigan LCQ™ DUO
LC/MS instrument (Thermo Electron) with electrospray ionization in negative mode.
Chromatography was carried out using a YMC Pro Pack C18 (150 x 2.0 mm, 3µ, 120 Å)
column with a gradient of 1-75% acetonitrile in 0.1% formic acid at a flow rate of 0.2
ml/min.
4.4 Stabilization of HIF-1α in human cell lines (IV)
Human embryonic kidney cells, HEK293, and human hepatocellular carcinoma cells,
Hep3B (American Type Culture Collection), were grown in Dulbecco’s modified Eagle’s
medium (Biochrom AG) and Eagle’s minimal essential medium (Sigma), respectively,
both supplemented with 10% foetal bovine serum (Bioclear). The cells were subcultured
at a density of 1 x 105 cells/cm2 and allowed to grow overnight. They were then washed
twice with 0.15 M NaCl and 0.02 M phosphate, pH 7.4, and the medium was changed to
Optimem 1 (Invitrogen). The cells were exposed to increasing concentrations of various
metals for 20 h. In control experiments the cells were treated with 25 µM 3-carboxy-4oxo-3,4-dehydro-1,10-phenantroline (Fibrogen Inc.) or subjected to hypoxia by culturing
them in an airtight incubator (Billups-Rothenberg). Medium samples were collected and
total cell extracts were obtained by lysing the cells in 150 mM NaCl, 0.1% SDS and 20
mM Tris-HCl, pH 6.8, supplemented with complete protease inhibitor cocktail (Roche).
The cell extracts were analysed by SDS-PAGE under reducing conditions, followed by
Western blotting with a monoclonal antibody against human HIF-1α (BD Biosciences)
and ECL™ Detection Reagents (Amersham Biosciences).
4.5 Analysis of VEGF production (IV)
The Quantikine human VEGF immunoassay (R&D Systems) was used to measure the
concentration of VEGF in medium samples collected after incubation of HEK293 and
Hep3B cells in the presence of various concentrations of CoCl2, NiCl2 and ZnCl2.
54
4.6 PCR analysis of HIF-P4H mRNA expression (I)
PCR analysis of human multiple tissue cDNA (MTC™) panel I and human foetal MTC
(Clontech) were used to study the expression of HIF-P4H isoenzymes in various tissues
according to the manufacturer’s protocol. Specific 5’ and 3’ primers were used to amplify
full-length HIF-P4H-1 and HIF-P4H-3 cDNAs, while in the case of HIF-P4H-2 the
forward primer was from the 3’ end of exon 1, starting at nucleotide 532. The PCR
products were sequenced to verify their identities.
5 Results
5.1 Cloning of the HIF-P4H 1- 3 and FIH cDNAs (I, II)
In order to express the three human HIF-P4H isoenzymes and FIH as recombinant
proteins, PCR amplification was used to obtain the corresponding cDNAs from human
colon, aorta and lung cDNA pools with specific primers. In the case of HIF-P4H-2 the
PCR product obtained corresponded to a variant in which the nucleotides encoding
residues Cys58-Pro175 were missing and were replaced by a 51-nucleotide stretch
encoding -Leu-Leu-Gly-Gly-Tyr-Arg-Phe-Ala-Phe-Ser-Trp-Asn-Ser-Asp-Glu-Arg-Ala-.
This stretch was found to represent the non-coding strand of the HIF-P4H-2 cDNA
between bp 450-500 in a 3’ to 5’ direction. The same PCR variant was also obtained
when amplification of the missing sequence was attempted from genomic DNA with
Advantage-2 polymerase, and the correct, full-length PCR product could only be
obtained with the Advantage-GC2 polymerase. This region of the HIF-P4H-2 cDNA has
a high GC content and is therefore likely to have a strong tendency to form secondary
structures, which make amplification of the correct sequence difficult.
5.2 Recombinant expression of the HIF-P4Hs and FIH (I, II)
The untagged full-length HIF-P4Hs and the HIF-P4H-2 PCR variant, GST fusion protein
forms of all the isoenzymes, His-tagged HIF-P4H-3 and untagged FIH, His-FIH, FIHFLAGHis, FIH-V5His and GST-FIH were expressed as recombinant proteins in insect
cells. The cells were harvested 72 h after infection and homogenized, and the soluble and
insoluble fractions were analysed by SDS-PAGE followed by Coomassie Blue staining.
Most of the HIF-P4H-2 polypeptide was found in the soluble fraction, while the vast
majority of the other HIF-P4H polypeptides remained in the insoluble fraction, although
distinct bands corresponding to HIF-P4H-1 and GST-fused forms of HIF-P4H-1 and HIFP4H-3 were also detected in the soluble fraction (Figure 3 in I). The HIF-P4H-2 PCR
variant, GST-HIF-P4H-2 PCR variant, HIF-P4H-3 and His-HIF-P4H-3 polypeptides were
56
not detectable in the soluble fraction in the SDS-PAGE gels, but a definitive level of HIFP4H activity was found in the soluble fraction of insect cells expressing these
polypeptides (see below and Table I in I).
The level of His-FIH expression was lower than for the FIH, FIH-FLAGHis, FIHV5His and GST-FIH polypeptides (Figure 1A in II). About half of the recombinant HisFIH and FIH-FLAGHis polypeptides were found in the soluble fraction, whereas the
majority of the FIH, FIH-V5His and GST-FIH polypeptides were in the insoluble
fraction. Western blot analysis showed that FIH-V5His was highly degraded and FIHFLAGHis and GST-FIH also showed some degradation (Figure 1B in II).
5.3 Purification of recombinant HIF-P4Hs and FIH (I, II, IV)
An initial attempt was made to purify the recombinant HIF-P4H isoenzymes 1 and 3 by
means of GST or His tags (I), but in both cases purification was unsuccessful, as the
majority of the recombinant enzymes did not become bound to the affinity column or the
enzymes became very loosely bound and were eluted during the washing steps.
Isoenzymes 1 and 3 were also expressed as maltose-binding fusion proteins in E. coli, but
no enzyme activity was detected with the purified polypeptides.
FIH-FLAGHis was purified by either anti-FLAG affinity or chelating Sepharose
column chromatography (II). Essentially pure protein was obtained by both methods
(Figure 2 in II). The HIF-P4Hs were also expressed as FLAGHis-tagged recombinant
polypeptides (IV) and successfully purified by anti-FLAG affinity chromatography
(Figure 1 in IV), but HIF-P4Hs 1 and 3 in particular easily became inactivated during
purification and could not be stored for long. This inactivation could be partly prevented
by adding 1 or 5 µM Fe2+ to all buffers during the purification.
5.4 Optimization of activity assays for the HIF-P4Hs 1-3
and FIH (I, II, IV)
The most commonly used assay for the C-P4Hs, based on the decarboxylation of
radioactively labelled 2-oxo-[1-14C]glutarate, was adopted and optimized in order to
measure the activities of the HIF-P4Hs and FIH (I, II). In this assay a synthetic peptide is
hydroxylated and the radioactivity of the 14CO2 formed during the hydroxylation-coupled
decarboxylation of 2-oxo-[1-14C]glutarate is measured.
The standard substrate used in the HIF-P4Hs assays was a 19-amino-acid synthetic
peptide corresponding to the C-terminal hydroxylation site around Pro564 of HIF-1α and
comprising its residues 556-574. The assay was found to be linear with increasing
concentrations of cell extracts containing the recombinant enzymes for up to 6000-9000
dpm of 14CO2 generated (Figure 4A in I), and with time for at least 25 min (Figure 4B in
I). Variation of the concentration of the peptide substrate or any of the cosubstrates gave
typical Michaelis-Menten kinetics (Figure 4C, D in I).
57
The HIF-P4H activities obtained varied in the range 1560-16,360 dpm/100 µl of
extractable cell protein depending on the isoenzyme and type of expression construct
(Table I in I). The differences in the activity values do not represent differences in the
specific activities of the recombinant enzymes, as the polypeptides differed in their
solubility in the extraction buffer. The highest activity was obtained with the full-length
HIF-P4H-2, the majority of which was found in the soluble fraction. Polypeptides
expressed as GST fusion proteins gave higher activities than the non-fusion proteins, the
increase being marked in the case of HIF-P4H-3 (Table I in I). The increases in the
activities are probably due to the better solubility of the GST fusion proteins in the
extraction buffer (Figure 3 in I).
The standard synthetic substrate used in the FIH assays corresponded to the residues
788-822 in the CAD of HIF-1α containing the Asn803 hydroxylation site. The optimized
assay was linear with increasing concentrations of enzyme in the cell extracts expressing
the recombinant enzyme or up to 800 nM pure enzyme (Figure 3A in II), and with time
for at least 20 min (Figure 3B in II). The reaction gave typical Michaelis-Menten kinetics
when the concentration of the substrate or any of the cosubstrates was varied (Figure 4 in
II). High levels of enzyme activity were found in the extracts of insect cells expressing
any of the recombinant FIH polypeptides, except GST-FIH (Table I in II).
Since an addition of bovine serum albumin to the reaction mixture is required to
achieve the maximal reaction velocity of the purified C-P4Hs (Kivirikko & Myllylä
1982), the effect of BSA on the activities of HIF-P4H-2 in a crude cell extract and a
purified form was studied at different temperatures (IV). The BSA addition was not
required for maximal activity when the crude insect cell extract was used as the enzyme
source (Table 2 in IV), but it significantly increased the amount of activity obtained when
purified HIF-P4H-2 was used (Table 2 in IV).
5.5 Catalytic centre activities of HIF-P4Hs and FIH (I, II, IV)
The catalytic centre activities of the recombinant HIF-P4Hs could not be exactly
determined when crude cell extracts were used as the enzyme sources, but they were
estimated to be more than 40-55 mol/mol/min (I). These estimates were later verified
with purified recombinant enzymes, which were found to have catalytic centre activities
of at least 40-50 mol/mol/min (Table 1 in IV). Recombinant HIF-P4H-1 produced in E.
coli has been reported to have a catalytic centre activity of only 12 x 10-6 mol/mol/min
(McNeill et al. 2002b), and our data are in agreement with this value, indicating that HIFP4Hs expressed in E. coli have a very low activity or are totally inactive. The catalytic
centre activities of FIH purified by either anti-FLAG affinity or chelating Sepharose
column chromatography were 85-135 and 70-120 mol/mol/min, respectively (II), much
higher than those obtained for recombinant FIH expressed in E. coli (Hewitson et al.
2002). The catalytic centre activities of the HIF-P4Hs determined here are about onethird and one-tenth of the highest values for FIH and type I C-P4H (Vuori et al. 1992a),
but are close to the values reported for the lysyl hydroxylases, which belong to the same
family of 2-oxoglutarate dependent dioxygenases (Rautavuoma et al. 2002) (Table 1 in
IV).
58
The HIF-P4Hs and FIH, like the C-P4Hs (Tuderman et al. 1977) and lysyl
hydroxylases (Puistola et al. 1980), were found to catalyze an uncoupled decarboxylation
of 2-oxoglutarate in the presence of all cosubstrates but in the absence of the peptide
substrate (IV). In the case of the C-P4Hs the rate of this reaction is 0.5-1% of the rate of
the hydroxylation-coupled decarcarboxylation (Tuderman et al. 1977), and the
corresponding uncoupled decarboxylation rates obtained for the purified HIF-P4Hs were
3-4% (Table 1 in IV), being 4-6-fold higher than those of the C-P4Hs but close to those
of the lysyl hydroxylases (Puistola et al. 1980). The uncoupled decarboxylation rate of
FIH was 0.4 % (Table 1 in IV), being thus slightly lower than that of the C-P4Hs.
5.6 The 2-oxoglutarate-binding arginine of HIF-P4Hs cannot be
replaced by a lysine (I)
The residue binding the C-5 carboxyl group of 2-oxoglutarate is a lysine in the C-P4Hs
(Myllyharju & Kivirikko 1997) but an arginine in the HIF-P4Hs. The lysine residue in
type I C-P4H can be replaced by an arginine, but the mutation leads to 85% inactivation
(Myllyharju & Kivirikko 1997). In our studies, substitution of lysine for the Arg367 in
HIF-P4H-1 inactivated the enzyme completely indicating that the arginine is an absolute
requirement for 2-oxoglutarate binding in the HIF-P4Hs (Table I in I).
5.7 Hydroxylation of peptides of varying lengths and representing
different HIF-α isoforms by the HIF-P4Hs and FIH (I, II)
All the HIF-P4Hs hydroxylated effectively the 19-amino-acid synthetic peptide
556
DLDLEMLAPYIPMDDDFQL574, corresponding to the C-terminal hydroxylation site
of HIF-1α (Table II in I). Omission of two residues from the N terminus had only a small
effect on this hydroxylation efficiency, but omission of two additional residues, resulting
in a 15-residue peptide, caused a marked increase in the Km values with isoenzymes 1 and
2. Shortening by two further residues led to increased Km values with all three
isoenzymes, although the increase was modest in the case of isoenzyme 3. Shortening of
the peptide by two or four residues at its C terminus increased the Km values of HIF-P4Hs
1 and 2 markedly but that of HIF-P4H-3 only slightly. Omission of one additional residue
to give a 14-amino-acid peptide led to a distinct increase in the Km in the case of HIFP4H-3 as well, while 13 and 11-residue peptides shortened at their C termini were not
hydroxylated by any of the isoenzymes. An 11-residue peptide shortened by four residues
at each end was hydroxylated by all of the isoenzymes with markedly increased Km
values, whereas an 11-residue peptide shortened by three residues at the N terminus and
five at the C terminus was hydroxylated only by isoenzymes 1 and 3. An 8-residue
peptide lacking six residues at the N terminus and five at the C terminus was not
hydroxylated by any of the isoenzymes. An addition of five residues to the C terminus of
the 19-amino-acid peptide did not improve the Km or Vmax values.
59
A 19-residue synthetic peptide corresponding to the N-terminal hydroxylation site
around Pro402 of HIF-1α was hydroxylated by HIF-P4Hs 1 and 2 with Km values that
were 20-50-fold higher than those for the C-terminal peptide, while HIF-P4H-3 did not
hydroxylate this peptide at all (Table III in I). One obvious difference between the two
hydroxylation sites is that Pro402 is located in a -Leu-Ala-Pro-Ala- sequence while Pro564
is present in a -Leu-Ala-Pro-Tyr- sequence. The differences in hydroxylation between the
two peptides are not due to the difference in the residue after the proline, however, as
Tyr565→ Ala and Tyr565→ Gly substitutions caused only minor increases in the Km values
(Table III in I).
The HIF-P4Hs were also found to hydroxylate peptides related to human HIF-2α and
HIF-3α and C. elegans HIF-α (Table IV in I). A 19-residue peptide corresponding to the
C-terminal hydroxylation site of HIF-2α was a slightly less effective substrate than the
corresponding HIF-1α peptide, while a 20-residue peptide corresponding to the Nterminal hydroxylation site of HIF-2α was hydroxylated ineffectively by all three
isoenzymes, but distinctly more effectively than the corresponding HIF-1α peptide, and
was even hydroxylated by HIF-P4H-3, which did not hydroxylate the N-terminal site of
HIF-1α at all. The difference between the two HIF-2α hydroxylation sites was thus much
smaller than that between the two HIF-1α sites. An additional sequence in HIF-2α with a
-Leu-Ala-Pro-Val- motif was not hydroxylated by any of the isoenzymes. All the
isoenzymes hydroxylated effectively a peptide corresponding to the C-terminal site of
HIF-3α, and all three human HIF-P4Hs also hydroxylated a 20-residue peptide
corresponding to the C. elegans HIF-α hydroxylation site, but with high Km and low Vmax
values.
FIH was found to require particularly long substrates, so that its Km for a 35-residue
HIF-1α peptide 788DESGLPQLTSYDCEVNAPIQGSRNLLQGEELLRAL822 was 100
µM and omission of only four residues from the C terminus reduced the enzyme activity
to only 26% of this (Table III in II). Shortening of the peptide by four and eight additional
residues reduced the relative activity by 37% and 15%, respectively. It is currently
unknown why a peptide lacking eight residues was hydroxylated more effectively than
one lacking four residues. The enzyme activity obtained with a peptide lacking 16
residues at the C terminus was only 9% of the original level, while shortening by seven
residues at the N terminus and nine at the C terminus reduced the activity to 4%. A 52amino-acid peptide comprising HIF-1α residues 775-826 was not significantly better as a
substrate than the 35-residue peptide.
Peptides related to HIF-2α were much less efficient substrates for FIH than ones
related to HIF-1α, so that the activities obtained with 35 and 19-residue HIF-2α peptides
were only 7% and 1%, respectively, of that obtained with the 35-residue HIF-1α peptide
(Table III in II).
60
5.8 Effects of amino acid substitutions in a HIF-1α peptide on the
hydroxylation by the HIF-P4Hs (III)
5.8.1 Substitution of the Leu-Glu-Met-Leu-Ala residues preceding the
hydroxylatable Pro564
The substrate sequence requirements of the HIF-P4Hs were studied further by
systematically varying each residue in the Leu-Glu-Met-Leu-Ala- sequence preceding the
Pro564 in the 20-residue peptide 554DTDLDLEMLAPYIPMDDDFQ573. It differs from the
19-residue peptide used in the study of the effect of substrate length on hydroxylation by
HIF-P4Hs, in that it contains two additional residues in its N terminus and lacks one in its
C terminus and has the hydroxylatable proline in the middle of the peptide. The Km
values of a purified version of this peptide for HIF-P4Hs 1, 2 and 3 are 15, 25 and 5 µM
while those for the 19-residue peptide are 8, 7 and 7 µM, respectively. The Vmax values
for all three isoenzymes were unaffected. Due to the high cost of synthetic peptides, the
experiments were first performed using crude peptide preparations of less than 50%
purity, cell lysates as sources of HIF-P4Hs 1 and 3, and partially purified HIF-P4H-2.
These initial experiments showed that substitution of leucines 559 and 562 had in
many cases only relatively minor effects (Table 1). Substitution of Leu559 gave identical
data for isoenzymes 1 and 2, the replacement of this residue with tryptophan giving a
markedly higher reaction rate than was obtained with leucine (120-130%), while the rates
obtained with tyrosine and methionine were similar to that with leucine. The lowest rates
(10-30%) were obtained with asparagine, threonine, histidine, arginine and lysine, all the
other substitutions giving rates of 30-80% (Table 1). In the case of HIF-P4H-3, two
residues, isoleucine and phenylalanine, gave significantly higher rates (130%) than were
obtained with leucine. The lowest rate (10-30%) was obtained with lysine, all the other
residues giving 30-80% activity (Table 1).
Substitions of Leu562 again gave similar data for HIF-P4H-1 and 2, and very similar
data for HIF-P4H-3 (Table 1). No residue was better than leucine in this position for any
of the isoenzymes. Reaction rates similar to that with leucine were obtained with valine,
isoleucine, phenylalanine and arginine in the case of HIF-P4Hs 1 and 2, and with valine,
isoleucine, phenylalanine, tyrosine and methione in the case of HIF-P4H-3. The lowest
reaction rates were obtained with glutamate, glycine, proline and aspartate for all
isoenzymes and also with glutamine for isoenzymes 1 and 2. All the other residues gave
reaction rates of 30-80%.
Substitutions of Glu560, Met561 and Ala563 gave very similar, although not identical,
data for HIF-P4Hs 1 and 2 (Table 1). No residue was better than the glutamate in position
560, but most residues were as good as it, while tryptophan, valine, phenylalanine,
tyrosine, cysteine, arginine and histidine gave reaction rates of 50-80%, the lowest rates,
30-50%, being obtained with lysine and proline. In the case of HIF-P4H-3, several amino
acids gave significantly higher reaction rates than glutamate, namely glycine, serine,
valine, alanine, glutamine, asparagine and lysine, while all the other amino acids gave
similar reaction rates to the control peptide. Significant differences were thus found
61
between the isoenzymes in their substrate preferences, as lysine and arginine were among
the most efficient substitutes for Glu560 in the case of HIF-P4H-3 but the least effective
ones in the case of isoenzymes 1 and 2.
Met561 could be replaced with most amino acids with no significant changes in the
reaction rates for HIF-P4Hs 1 and 2, only proline giving a lower rate (about 50%) (Table
1). Most amino acids gave similar or higher rates than methionine for HIF-P4H-3, only
aspartate, glutamate and proline giving rates of 50-80% (Table 1).
Ala563 was by far the strictest requirement (Table 1). No other residue gave a rate of
80-100% in the case of HIF-P4H-1, while only serine and three residues, isoleucine,
valine and serine, gave reaction rates similar to that with alanine in the case of HIF-P4Hs
2 and 3, respectively.
The initial results were confirmed by analysing some of the peptides at a purity higher
than 85% and using highly purified preparations of HIF-P4Hs 1 and 2, whereas those of
HIF-P4H-3 were still mostly crude cell extracts because of the instability of the purified
enzyme. Determination of Km and Vmax values for the purified peptides confirmed the
data obtained with the unpurified one, with some minor exceptions (compare tables 1-3 in
III). In addition, five peptides with more than one substitution were designed on the basis
of data on some of the substitutions that alone had given no effect or only a very slight
effect in the initial studies. All five peptides had Km values for the three isoenzymes that
were similar to or less than twice that of the control peptide (Table 3 in III). In the case of
HIF-P4H-3 the Vmax values for all five peptides were likewise very similar to that for the
control peptide (90-120%), while in the case of HIF-P4H-2 one double-mutant peptide
(Leu559 → Ile, Leu562 → Ile) gave a Vmax of 70%, the other four giving a Vmax of 100%.
The Vmax values of these peptides for HIF-P4H-1 were somewhat lower than that of the
control peptide, two double-substituted peptides (Leu559 → Trp, Leu562 → Val and Leu559
→ Tyr, Leu562 → Ile) giving Vmax values of 90% and 80%, respectively, while the other
three peptides gave a Vmax of 60-70%.
These data indicate that the three HIF-P4Hs, especially HIF-P4H-3, can act well on
sequences distinctly different from the -Leu-X-X-Leu-Ala-Pro- consensus motif that has
been suggested in previous studies as being required for hydroxylation (Epstein et al.
2001, Masson et al. 2001).
62
Table 1. Relative reaction rates of the HIF-1α-like peptide with 19a substitutions of the
five residues preceding the hydroxylatable Pro564
Activity %
Eb
L
Mb
Ab
L
HIF-P4Hs 1 and 2
>120
W
80-120
Y, L, M
Most amino acids
50-80
I, F, A, E, D
W, V, F, Y, C, R, H P
Y, T, K, M
30-50
C, P, G, V, Q, S
K, P
N, C, W, H, A, S V, Q
10-30
N, T, H, R, K
Most amino acids L, V, I, F, R
<10
A, S
T, K
Q, E, G
M, G, R, I
P, D
Other amino acids
HIF-P4H-3
>120
I, F
G, S, V, A, Q, N, R, Y, W, L, V
80-120
W, Y, L, M, A, V Other amino acids
Most amino acids L, V, I, F, Y, M
I, A, V, S
50-80
E, D, C, P, Q, S, T
D, E, P
R, K, H, C, Q, T
30-50
G, N, H, R
K, N, W, H, A, S, E, M, F, G, L
10-30
K
E, G
W, D
P, D
N, P
K
R, T, C
Q
<10
a
The 19 amino acid residues are A, C, D, E, F, G, H, I, K, M, N, P, Q, R, S, T, V, W, Y. bSimilar, but not
identical data were obtained for HIF-P4Hs 1 and 2.
5.8.2 Substitutions of alanine for acidic HIF-1α residues
Six out of the seven acidic residues in the 20-residue peptide
554
DTDLDLEMLAPYIPMDDDFQ573 were subjected to alanine substitution, the Nterminal aspartate being excluded. These experiments were performed with crude
peptides. The results obtained were similar for each of the HIF-P4Hs, and detailed results
are therefore given only for isoenzyme 2.
Any single acidic residue could be replaced by alanine with no decrease in the
substrate properties (Table 4 in III). Also, any two acidic residues could be replaced with
no decrease in the reaction rate unless one of them was the first of the three consecutive
aspartates in the PYIPMDDD sequence, namely Asp569. This substitution reduced the
reaction rate to 50-80%. Even three acidic residues could be replaced with no significant
decrease in the reaction rate if two of them were on the N-terminal side of Pro564 and one
on the C-terminal side, but not Asp569. If two of the residues were on the C-terminal side
and one on the N-terminal side, the reaction rates were 50-80%, unless the pair on the Cterminal side was Asp569 + Asp570, in which case the rate was lower. Substitution of all
three acidic residues on the N-terminal side, and some combinations of four substituted
residues, gave rates that were only 10-30%. Substitution of all three C-terminal residues,
or of four residues with three of them on the C-terminal side, gave the lowest rates (less
than 10%).
63
5.8.3 Substitution of 3,4-dehydroproline or L-azetidine-2-carboxylic
acid for Pro564
Replacement of the hydroxylatable proline in a peptide substrate of the C-P4Hs with 3,4dehydroproline is known to lead to a rate of uncoupled decarboxylation that is similar to
the hydroxylation rate with the non-modified peptide (Wu et al. 2000). Corresponding
substitution studies were therefore also performed for the HIF-P4Hs.
Crude peptides in which 3,4-dehydroproline or L-azetidine-2-carboxylic acid was
substituted for Pro564 (Figure 1 in III) gave 2-oxoglutarate decarboxylation rates with all
three HIF-P4Hs that were similar to or even higher than the hydroxylation-coupled
decarboxylation rate obtained with the control peptide. Substitution of other proline
analogues or some other amino acids for Pro564 caused no significant decarboxylation.
Experiments were also performed with purified peptides to determine the Km and Vmax
values and to find out whether the high 2-oxoglutarate decarboxylation rates obtained
with the 3,4-dehydroproline and L-azetidine-2-carboxylic acid peptides indeed
represented uncoupled decarboxylation. The Vmax values of all the HIF-P4Hs for the 3,4dehydroproline peptide were higher (120%) than for the control peptide, while the Km
values were similar to that for the control peptide or slightly lower (Table 5 in III). The
Vmax values of HIF-P4Hs 1 and 2 for the L-azetidine-2-carboxylic acid peptide were 70%
of that of the control peptide and that of HIF-P4H-3 110%, the Km values of all three
isoenzymes being 3 to 6-fold relative to that of the control peptide.
Liquid chromatography and mass spectrometry were used to study hydroxylation of
the peptides. The elution position of the hydroxylated control peptide differed distinctly
from that of the non-hydroxylated peptide (Figure 2A, B in III), whereas the mobilities of
the 3,4-dehydroproline (Figure 2 C, D in III) and L-azetidine-2-carboxylic acid peptides
were identical with and without enzyme incubation. Mass spectral data showed that the
control peptide was hydroxylated whereas the 3,4-dehydroproline (Figure 2 in III) and Lazetidine-2-carboxylic acid peptides were not. It is therefore evident that the high rates of
2-oxoglutarate decarboxylation observed with the 3,4-dehydroproline and L-azetidine-2carboxylic acid peptides were not due to hydroxylation-coupled decarboxylation.
5.9 Km values of HIF-P4Hs and FIH for the cosubstrates (I, II, IV)
Km values for the cosubstrates were determined using extracts from insect cells
expressing the HIF-P4Hs and purified preparations of FIH as enzyme sources. The Km
values of the HIF-P4Hs for 2-oxoglutarate were about 3-fold and those for ascorbate
about half relative to those of type I C-P4H, with no significant differences between the
HIF-P4H isoenzymes, while the corresponding values for FIH were very close to those of
type I C-P4H (Table 2). The Km of FIH for Fe2+ was 0.5 µM, one-fourth of that of type I
C-P4H (Table 2). Determination of this Km for the HIF-P4Hs was not possible with the
cell extracts as enzyme sources, as they retained partial activity even in the absence of
added iron (I). This suggests that the iron atom may be more tightly bound to the HIFP4Hs than to FIH and the C-P4Hs. Purified HIF-P4Hs also retained partial activity even
64
in the absence of added iron (IV), so that it was not possible to determine their true Km
values for Fe2+. The apparent Km values determined for preparations that already had
considerable activity levels without any iron addition were very low, only 0.03 µM in the
cases of HIF-P4Hs 1 and 2, and about 0.1 µM in the case of HIF-P4H-3 (Table 2). These
values were 5-15-fold and 20-65-fold lower than the Km values for FIH and type I CP4H, respectively. Maximal activities of both crude and purified HIF-P4Hs were obtained
with 1-5 µM Fe2+, the activities decreasing at higher Fe2+ concentrations.
The clearest difference between the HIF enzymes and the C-P4Hs was found in their
Km values for O2, those of the HIF-P4Hs being 230-250 µM, i.e. slightly above the
concentration of dissolved oxygen in air (as shown for HIF-P4H-1 in Figure 4), while
that of type I C-P4H was only one-sixth of these values (Table 2). The Km of FIH for O2
was 90 µM, markedly lower than those of the HIF-P4Hs, but over 2-fold relative to that
of type I C-P4H.
Table 2. Km values of the HIF-P4Hs, FIH and type I C-P4H (C-P4H-I) for cosubstrates
2-Oxoglutarate
Km, µM
Ascorbate
Km, µM
O2
Km, µM
Fe2+
Km, µM
HIF-P4H-1
HIF-P4H-2
60
60
170
180
230
250
0.03
0.03
HIF-P4H-3
55
140
230
0.1
FIH
25
260
90
0.5
C-P4H-I
20a
300a
40
2a
Enzyme
a
Myllyharju & Kivirikko 1997.
65
Fig. 4. Km of HIF-P4H-1 for oxygen is slightly above the concentration of dissolved O2 in air.
5.10 Inhibition of the HIF-P4Hs and FIH by 2-oxoglutarate
analogues (I, II)
Distinct differences were found between the HIF-P4Hs, FIH and type I C-P4H in their
inhibition by several 2-oxoglutarate analogues. Pyridine-2,4-dicarboxylate and pyridine2,5-dicarboxylate are well characterized and efficient inhibitors of the C-P4Hs, and
pyridine-2,4-dicarboxylate also inhibited the HIF-P4Hs and FIH, but less effectively,
whereas pyridine-2,5-dicarboxylate was only a very weak inhibitor of the HIF-P4Hs,
with Ki values exceeding 300 µM, although its Ki for FIH was 50 µM (Table 3). Distinct
differences were also seen between the HIF-P4Hs and FIH in their inhibition by other
compounds. The two most potent inhibitors of the HIF-P4Hs, 3-hydroxypyridine-2carbonylglycine and N-((3-hydroxy-6-choloroquinolin-2-yl)carbonyl)glycine, were
distinctly less efficient inhibitors of FIH, while two potent inhibitors of FIH,
oxalylglycine and 3,4-dihydroxybenzoate, were distinctly less efficient inhibitors of the
HIF-P4Hs. Furthermore, differences were found between the three HIF-P4Hs, HIF-P4H1 having higher Ki values than the other two isoenzymes for all the compounds tested.
66
Table 3. Ki values of the HIF-P4Hs, FIH and C-P4H-I for certain 2-oxoglutarate
analogues
Inhibitor
Pyridine-2,4-dicarboxylate
Pyridine-2,5-dicarboxylate
3-Hydroxypyridine-2-carbonyl-
HIF-P4H-1 HIF-P4H-2
Ki, µM
Ki, µM
HIF-P4H-3
Ki, µM
FIH
Ki, µM
C-P4H-I
Ki, µM
40
>300
7
>300
8
>300
30
50
2a
0.8a
15
2
1
>300
0.4b
50
8
10
2
1.9c
>300
>300
>300
10
5c
30
10
10
ND
2b
0.8
0.2
0.2
>300
0.06d
glycine
Oxalylglycine
3,4-Dihydroxybenzoic acid
3-Carboxy-4-oxo-3,4-dihydro1,10-phenanthroline
N-((3-Hydroxy-6-chloroquinolin2-yl)carbonyl)glycine
ND= not determined. a Kivirikko & Myllyharju 1998, Kivirikko & Pihlajaniemi 1998. b IC50, Ivan et al. 2002. c
Baader et al. 1994. d IC50.
5.11 Effect of desferrioxamine and metals on the HIF-P4Hs
and FIH (IV)
5.11.1 Effect of desferrioxamine on the activities of the HIF-P4Hs
The iron chelator desferrioxamine (DFO) is known to stabilize the α subunit of HIF,
leading to activation of its target genes (Ivan et al. 2001, Jaakkola et al. 2001), and it has
also been shown to inhibit the ability of reticulocyte lysates programmed with HIF-P4Hs
to promote interaction between HIF-1α and pVHL (Epstein et al. 2001). When the effect
of increasing concentrations of DFO on the activity of the recombinant HIF-P4Hs was
studied in the presence of 5 µM Fe2+, it was found that up to 1 mM DFO did not inhibit
crude HIF-P4H-2 at all and inhibited crude HIF-P4H-1 by only 10-20%, the inhibition
level remaining constant between 10 µM and 1 mM DFO. Crude HIF-P4H-3 was already
inhibited by 50% with 10 µM DFO (Figure 2A in IV), but was not completely inhibited at
higher DFO concentrations, 25% of its activity remaining at 1 mM DFO. Purified type I
C-P4H was also 50% inhibited by 10 µM DFO, but total inhibition was already achieved
at 100 µM DFO. Purified HIF-P4Hs 1 and 2 were 40-50% inhibited by 10 µM DFO, the
level of inhibition of HIF-P4H-1 increasing only to about 60% and that of HIF-P4H-2
remaining steady around 40-45% at up to 1 mM DFO (Figures 2B and C in IV). These
results indicate that HIF-P4Hs 1 and 2 bind Fe2+ very strongly, which is in agreement
with the very low apparent Km values of these enzymes for Fe2+.
67
5.11.2 Inhibition of the HIF-P4Hs and FIH by metals
Several divalent metal ions inhibit the C-P4Hs competitively with respect to iron, zinc
being the most potent inhibitor (Rapaka et al. 1976, Ryhänen et al. 1977, Tuderman et al.
1977, Kivirikko & Pihlajaniemi 1998). Cobalt and nickel are well-known stabilizers of
HIF-α (Semenza 2001). It was therefore assumed upon the discovery of the HIF-P4Hs
that this stabilization was caused by competitive inhibition of the HIF-P4Hs by the metals
concerned. We studied here the inhibition of the HIF-P4Hs and FIH by various metals.
In the case of HIF-P4Hs 1 and 2, IC50 values for the metals were determined with
both crude and purified enzymes, whereas the IC50 values of HIF-P4H-3 were determined
only with crude enzyme samples due to the marked inactivation of the purified HIF-P4H3 in freezing and during storage at 4°C. Cobalt is known to stabilize HIF-1α efficiently at
a 100 µM concentration (Wang & Semenza 1993). Surprisingly, cobalt was found to
inhibit all three HIF-P4Hs rather ineffectively, especially HIF-P4H-2, as 100 µM cobalt
inhibited crude HIF-P4H-2 by only 20% and even 2 mM cobalt did so by only 44%
(Figure 3A, Table 4 in IV). The IC50 values of crude HIF-P4Hs 1 and 3 for cobalt were 60
and 70 µM (Table 4 in IV), respectively, but the inhibition remained at a level of 50-65%
even with 500 µM cobalt. Purified HIF-P4H isoenzymes 2 and 1 had IC50 values of 100
µM and 40 µM, respectively, but inhibition increased only slightly at higher cobalt
concentrations (Figure 3B, Table 4 in IV). The IC50 of type I C-P4H for cobalt was 15
µM, with 500 µM cobalt causing 80% inhibition.
Nickel was found to be an even less effective inhibitor of the HIF-P4Hs than cobalt
(Table 4 in IV). Of the other metals studied, zinc was an efficient inhibitor of HIF-P4H-3,
but an inefficient inhibitor of HIF-P4Hs 1 and 2 (Figure 3 and Table 4 in IV). It
nevertheless inhibited the HIF-P4Hs more efficiently than cobalt at high concentrations
(Figure 3 in IV). Cadmium also inhibited all three HIF-P4Hs, while magnesium and
manganese were the least potent inhibitors (Table 4 in IV).
Inhibition of FIH was studied with the same metal ions. Purified FIH preparations
showed only relatively low levels of activity without added iron, the Km of FIH for iron
being about 0.5 µM, i.e. 5 to 15-fold higher than those of the HIF-P4Hs. It was thus
possible to determine the mode of inhibition. The metals were found to inhibit purified
FIH competitively with respect to iron (Figure 4 in IV) and non-competitively with
respect to the peptide substrate. FIH was inhibited by most of the metals tested more
effectively than were the HIF-P4Hs (Table 4 in IV). Zinc was the most potent inhibitor of
FIH, with a Ki of 0.5 µM, which is essentially the same as that of type I C-P4H, and 6fold lower than the IC50 of crude HIF-P4H-3 and 60 and 260-fold lower than the
corresponding values for purified HIF-P4Hs 1 and 2, respectively (Table 4 in IV). Cobalt
and nickel were also potent inhibitors of FIH, with Ki values of 1 and 4 µM, respectively,
these values being 40 and 80-fold lower than the lowest corresponding IC50 values of any
of the HIF-P4Hs, and about 10-fold lower than those of type I C-P4H (Table 4 in IV).
The Ki of FIH for cadmium was 10 µM, and therefore identical to that of type I C-P4H
and almost the same as the IC50 of the purified HIF-P4H-1, while the other two HIF-P4Hs
had about 10-fold higher IC50 values (Table 4 in IV). Unlike the HIF-P4Hs and type I CP4H, FIH was also inhibited efficiently by manganese, with a Ki of 10 µM (Table 4 in
IV). Magnesium was the only metal ion tested that was a poor inhibitor of FIH.
68
5.11.3 Effect of cobalt and DFO on the synthesis of HIF-P4H-2 and in
vitro activity of the synthesized enzyme
The activity of a recombinant HIF-P4H synthesized in the presence of cobalt and DFO
was studied by expressing FLAGHis-tagged HIF-P4H-2 in insect cells for 48 h in the
presence of increasing amounts of cobalt or DFO. Analysis of cell extracts by SDS-PAGE
followed by Western blotting showed that while the addition of 1 mM cobalt or 100 µM
DFO had no effect on the amount of HIF-P4H-2 synthesized (Figure 4 in IV, lanes 2 and
3), the presence of 300 µM DFO caused about a 50% reduction in the amount of the
enzyme protein as determined by densitometry (Figure 4 in IV, lane 4).
The recombinant HIF-P4H-2 produced in the presence of cobalt or DFO was found to
become significantly inactivated during synthesis, 100, 500 and 1000 µM cobalt leading
to inactivation by about 50%, 60% and 75%, respectively, and 100 and 300 µM DFO by
75-80% (Table 5 in IV). The activity of the enzyme synthesized in the presence of cobalt
could not be reduced further or increased by adding increasing concentrations of cobalt or
iron, respectively, to the reaction mixture, whereas the enzyme synthesized in the
presence of DFO could be partially reactivated by adding iron (Table 5 in IV). In other
words, the enzyme synthesized in the presence of relatively high cobalt concentrations
became permanently inactivated, whereas that synthesized in the presence of DFO, i.e. in
the absence of free iron, could be reactivated by iron additions, although very
ineffectively.
5.11.4 Effect of cobalt and iron on the stability of HIF-P4Hs 2 and 3
Crude preparations of HIF-P4Hs 2 and 3 were pre-incubated in the presence of 100 µM
cobalt or iron for 4 and 24 h at 4°C in order to study the effect of this on enzyme activity.
Activity measurements were then carried out with samples diluted 1:50 in the reaction
mixture, thus resulting in a 2 µM final concentration of cobalt, which is much lower than
the IC50 values for these HIF-P4Hs. HIF-P4Hs 2 and 3 lost about 50% and 60% of their
activity, respectively, after incubation with 100 µM cobalt for 4 h, and were almost or
totally inactivated (95% and 100%) after 24 h incubation (Figure 5 in IV). Control
samples of HIF-P4H-2 and HIF-P4H-3 incubated in a buffer containing no metal ions lost
about 45% and 60% of their activities in 24 h (Figure 5 in IV). Iron protected the
enzymes from inactivation, the activities of HIF-P4Hs 2 and 3 being about 130% and
110% of the controls after a 4 h preincubation in the presence of iron and 170% and
190% after 24 h (Figure 5 in IV).
5.11.5 HIF-1α stabilization in human HEK293 and Hep3B cells
Stabilization of HIF-1α by cobalt, nickel and zinc in HEK293 and Hep3B cells was
studied by culturing the cells for 20 h in the presence of increasing concentrations of the
metals. In control experiments, cells were treated with a P4H inhibitor, 3-carboxy-4-oxo-
69
3,4-dihydro-1,10-phenanthroline (Ivan et al. 2002), or cultured under hypoxic
conditions.Total cell extracts were analyzed by SDS-PAGE followed by Western blotting
with a HIF-1α antibody. As expected, in addition to treatment with the P4H inhibitor or
culturing in hypoxia, stabilization of HIF-1 was seen upon treatment with 100 µM cobalt
and 200 µM nickel in both cell types, although lower concentrations produced no
detectable stabilization. Zinc caused no stabilization of HIF-1α when tested up to 1 mM.
5.11.6 Effect of cobalt, nickel and zinc on the production of VEGF
by human HEK293 and Hep3B cells
Induction of VEGF expression in HEK293 and Hep3B cells by cobalt, nickel and zinc
was studied by measuring the amount of VEGF in medium samples collected from
comparable numbers of cells cultured in the presence of increasing concentrations of the
metals. All three metals induced VEGF expression in both cell types, but the effect was
more marked in the HEK293 than in the Hep3B cells in the case of cobalt and nickel,
whereas no distinct differences were found between the cell types in the case of zinc
(Table 6 in IV). Culturing of HEK293 and Hep3B cells in the presence of 10 µM cobalt
increased their VEGF production to 220% and 120%, respectively, while 50 µM cobalt
increased the values to 350% and 180%. The level of VEGF expression rose further at
higher cobalt concentrations, the maximal increases, of 2100% in HEK293 cells and
350% in Hep3B cells, being obtained with 300 µM cobalt, while still higher cobalt
concentrations were less effective. Nickel caused only a slight increase in VEGF
expression at a 50 µM concentration, the highest increases (2190% and 350%) being
obtained with 500 µM nickel. The VEGF response to low concentrations of zinc was
moderate in both cell types, and 300 µM zinc was already toxic to the cells.
5.12 Analysis of the expression of mRNAs for the three human
HIF-P4H isoenzymes (I)
The only PCR product obtained for HIF-P4H-1 corresponded to the full-length cDNA,
whereas in the case of HIF-P4Hs 2 and 3 smaller fragments originating from the
corresponding mRNAs were amplified from the colon, aorta and lung cDNA pools in
addition to the full-length products. In order to study whether these smaller fragments
represented alternatively spliced forms, the expression of HIF-P4H mRNAs in various
tissues was analysed further by PCR amplification of foetal and adult human multiple
tissue cDNA panels.
Full-length HIF-P4H-1 cDNA was obtained from all the tissues studied, with no
smaller PCR products detected (Figure 2A in I). The level of HIF-P4H-1 mRNA
expression was highest in the adult brain, placenta, lung and kidney. A full-length HIFP4H-2 PCR product starting from the end of exon 1 was likewise amplified from all the
tissues studied (Figure 2B in I), the highest level of mRNA expression being found in the
adult heart, brain, lung and liver and in the foetal brain, heart, spleen and skeletal muscle,
70
but in addition to the full-length product, two smaller fragments were obtained from
several tissues, and sequencing indicated that these represented splicing variants lacking
either exon 3 or 4 sequences. Nevertheless, the levels of expression of these splicing
variants were clearly lower than those of the full-length mRNA. Full-length HIF-P4H-3
was amplified from all the tissues studied, with the highest expression levels in the adult
heart, brain, placenta, lung and skeletal muscle and in the foetal heart, spleen and skeletal
muscle (Figure 2C in I), and one additional smaller PCR product was obtained in about
equal amounts to the full-length HIF-P4H-3 product in all the tissues studied. Sequencing
demonstrated that this represented a splicing variant lacking nucleotides 79-357. All the
alternatively spliced forms of the isoenzyme 2 and 3 mRNAs followed the GT/AG rule.
The alternatively spliced forms were found to encode inactive enzymes (Table I in I).
6 Discussion
6.1 Catalytic centre activities of HIF-P4Hs and FIH
The three HIF-P4Hs were purified here to near homogeneity as active enzymes for the
first time. Others have expressed HIF-P4H-1 in E. coli, but the catalytic centre activity of
the purified enzyme was very low, 12 x 10-6 mol/mol/min (Epstein et al. 2001, McNeill et
al. 2002b). The catalytic centre activities of the HIF-P4Hs purified here were markedly
higher than this, about 40-50 mol/mol/min, although they were still lower than the
activities of 135 and 400 mol/mol/min recorded for the purified FIH obtained here and
for type I C-P4H (Vuori et al. 1992a), respectively. The catalytic centre activities
measured for the purified HIF-P4Hs were in good agreement with estimates based on the
activities of these enzymes in crude cell extracts. Although the results for the HIF-P4Hs
and FIH were corrected for saturating concentrations of the peptide substrate and
cosubstrates, these values are also dependent on the substrates used in the assay, and
therefore it is not known whether higher catalytic centre activities could be obtained with
the full-length HIF-1α than with the 19 and 35-residue peptides used in the standard
assays of HIF-P4H and FIH activity, respectively. Furthermore, some amino acid
substitutions in the HIF-1α-like peptide increased the Vmax values of all three HIF-P4Hs
to 120% while the Vmax values of the three HIF-P4Hs with the 20-residue HIF-3α peptide
were 120-150% of those with the 19-residue HIF-1α peptide. It is thus obvious that the
catalytic centre activities of the HIF-P4Hs would be higher, at least about 70
mol/mol/min, if a more effective substrate were used. The present data do not support
previous results obtained by mobility shift and pull-down assays with a 122-residue HIF1α fragment as a substrate suggesting that HIF-P4H-2 has clearly higher activity than the
other two isoenzymes (Huang et al. 2002). Instead, the catalytic centre activities of the
three isoenzymes were almost identical, and the highest activity was measured for HIFP4H-3. These results do not mean, however, that all three isoenzymes would be equally
effective in vivo, where their activities are also affected by other factors, such as the
concentrations of the various enzyme proteins and the cellular locations of the three
isoenzymes.
72
6.2 Substrate requirements of the HIF-P4Hs and FIH
The C-P4Hs act more effectively on long than short peptides, the chain length influencing
the Km values, whereas no significant differences are observed in the Vmax values for
short and long peptides (Kivirikko & Pihlajaniemi 1998). The minimum requirement for
hydroxylation by the C-P4Hs is fulfilled by tripeptides with the sequence X-Pro-Gly. The
present data indicate that the HIF-P4Hs and FIH also require long peptides, their
minimum requirements being much longer than that of the C-P4Hs. The HIF-P4Hs and
FIH were also found to hydroxylate other human HIF-α isoforms besides HIF-1α, but
with distinct differences.
The shortest HIF-1α-like peptide hydroxylated by all three HIF-P4H isoenzymes in
vitro was an 11-residue peptide, i.e. a version of the 19-residue control peptide shortened
by four residues at each end, while another 11-residue peptide shortened by three residues
at the N terminus and five at the C terminus was hydroxylated only by HIF-P4Hs 1 and 3.
Systematic shortening of the 19-amino-acid peptide at the N and/or C terminus indicated
that all three isoenzymes required two residues (LA) preceding the proline and at least
five (YIPMD) after the proline, although none of them would actually hydroxylate the
hypothetically shortest 8-residue peptide LAPYIPMD. It is thus evident that the HIFP4Hs require more than eight residues for hydroxylation, and a peptide of about 20residues appears to be an optimal substrate, as the addition of five residues to the C
terminus of the 19-residue peptide did not improve the hydroxylation efficiency.
Epstein et al. (2001) have suggested that the second hydroxylation site in HIF-1α,
Pro402, may be hydroxylated less effectively than Pro564, and that only HIF-P4Hs 1 and 2
can catalyze this hydroxylation. They used HIF-P4Hs produced in a reticulocyte lysate to
hydroxylate HIF-1α mutants, and the hydroxylated polypeptides were subsequently
subjected to capture by pVHL. Substitution of alanine for Pro402 had only a minor effect
on capture when the C-terminal Pro564 was intact, whereas substitution of glycine for
Pro564 with Pro402 intact reduced the capture markedly and was promoted only by
isoenzymes 1 and 2. No capture occurred when both prolines were subjected to mutation.
Our data explain these results by demonstrating that the Km values of HIF-P4Hs 1 and 2
for the peptide corresponding to Pro402 were 20-50 times higher than those for the Cterminal peptide with Pro564, and that HIF-P4H-3 did not hydroxylate the N-terminal
peptide at all. Similar results were also obtained in recent in vivo transfection assays,
which showed that HIF-P4H-3 acts very ineffectively on the N-terminal hydroxylation
site in both HIF-1α and HIF-2α (Appelhoff et al. 2004). It is therefore possible that
hydroxylation of Pro402 is less efficient than that of Pro564 also in vivo and that the former
may not become hydroxylated in many situations at all. It has been shown, however, that
in vivo Pro402 is critical for HIF-1α stability. Although the interaction of pVHL with the
N-terminal hydroxylation site is weaker than with the C-terminal site, it was capable of
targeting HIF-1α for degradation in normoxia when Pro564 was deleted (Huang et al.
1998), therefore suggesting an independent interaction of pVHL with the Pro402 (Huang
et al. 2002). Simultaneous mutation of both hydroxylatable prolines resulted in a
markedly more stable HIF-1α than Pro564 mutation alone, supporting the important role
of hydroxylation of the N-terminal site in HIF-1α stability in vivo. (Masson et al. 2001).
An obvious difference between the two hydroxylation sites of HIF-1α is that the residue
73
564
following Pro is a tyrosine whereas that following Pro402 is an alanine. This difference
does not explain the difference in hydroxylation efficiency, however, as substitution of
alanine or glycine for Tyr565 caused only a slight increase in the Km values. Our result
indicating that the Tyr565 → Ala substitution has no effect on hydroxylation is in
agreement with that of a study in which the same substitution did not affect pVHL
binding (Ivan et al. 2001), although opposite results regarding capture by pVHL have
also been reported (Bruick & McKnight 2001, Jaakkola et al. 2001).
HIF-P4Hs were also found to hydroxylate peptides related to HIF-2α, HIF-3α and C.
elegans HIF-α. The HIF-2α sequence contains three possible hydroxylation sites. The
peptide corresponding to the C-terminal was a good substrate for HIF-P4H-3 and also a
relatively good substrate for the other two isoenzymes, whereas the N-terminal peptide
was an inefficient substrate for all three isoenzymes. The third potential hydroxylation
site in HIF-2α with a LAPV motif was not hydroxylated by any of the HIF-P4Hs. The
hydroxylation site in HIF-3α is highly similar to the C-terminal site in HIF-1α and a
corresponding HIF-3α-like peptide was hydroxylated very effectively by all three HIFP4H isoenzymes. The function of HIF-3α is currently poorly understood, but the present
results and those reported by others (Maynard et al. 2003) suggest that it is also regulated
by proline hydroxylation.
The two critical prolines of HIF-1α, Pro564 and Pro402, are located in -Leu-Glu-MetLeu-Ala-Pro-Tyr- and -Leu-Thr-Leu-Leu-Ala-Pro-Ala- sequences, respectively, the C and
N-terminal hydroxylation sites in HIF-2α in -Leu-Glu-Thr-Leu-Ala-Pro-Tyr- and -LeuAla-Gln-Leu-Ala-Pro-Thr- sequences, and the hydroxylated prolines in HIF-3α and the
C. elegans HIF-α in -Leu-Glu-Met-Leu-Ala-Pro-Tyr- and -Leu-Ser-Cys-Leu-Ala-ProPhe- sequences, respectively. Therefore all these sequences around the hydroxylatable
proline have a common sequence -Leu-X-X-Leu-Ala-Pro-, which has been suggested to
be required for hydroxylation by the HIF-P4Hs (Epstein et al. 2001, Masson et al. 2001).
Some mutation studies have indicated that substitution of alanine for Leu562 or glycine for
Ala563 in a HIF-1α peptide may prevent hydroxylation (Bruick & McKnight 2001, Ivan et
al. 2001), whereas it has been shown elsewhere that substitutions of Leu559, Leu562 and
Ala563 can be tolerated, suggesting that the hydroxylatable proline may be the only
obligatory residue (Huang et al. 2002). The present results are in agreement with this, as
they show that HIF-P4Hs do not strictly require the suggested -Leu-X-X-Leu-Ala-Procore motif, but instead, the two leucines can be replaced by many other amino acids with
no marked differences in hydroxylation by any of the HIF-P4Hs. The residues in the two
X positions, glutamate and methionine in HIF-1α, could be replaced by almost any amino
acid, as expected. Surprisingly though, there were eight amino acids that were better for
HIF-P4H-3 than the authentic glutamate residue and four that were better than Met561
itself. Ala563 was found to be the strictest requirement, as it could not be fully replaced by
any amino acid in the case of HIF-P4H-1, only by serine in the case of HIF-P4H-2, and
only by three residues, isoleucine, valine and serine, in the case of HIF-P4H-3. Leu574 has
been previously found to be required for hydroxylation of Pro564, especially in the case of
HIF-P4H-2 (Kageyama et al. 2004) and it was shown in the present study that shortening
the 19-residue peptide by two residues from the C terminus, including Leu574, causes a
marked increase in the Km values of HIF-P4Hs 1 and 2. As the peptide used in the amino
acid substitution study done here, however, was missing exactly Leu574 from its C
terminus, it was not possible to study its substitution in detail in the scope of this study.
74
The acidic residues present in the 20-residue peptide used as a substrate were also found
to play a distinct role, but the substitution of alanine for any six of them or in some
combinations even for three of them, had no negative effects. These data indicate that it
will be impossible to find additional substrates for the HIF-P4Hs by means of
computerized protein sequence searches, because of the lack of a specific core sequence.
The data obtained here indicate that the HIF-P4Hs require even longer peptide
substrates than the C-P4Hs, the minimum requirement being a peptide of more than eight
residues. The data from the amino acid substitution experiments, however, indicate that
the minimum sequence requirement for hydroxylation by HIF-P4Hs is -Ala-Pro-.
Together these results and those of Huang and co-workers (2002) suggest that the -AlaPro- sequence is a necessary but not sufficient requirement for hydroxylation. HIF-1α
may become bound to the HIF-P4Hs in a similar extended conformation to that required
for its binding to pVHL (Hon et al. 2002, Min et al. 2002), so that the HIF-P4Hs may
have a long substrate-binding site and interaction between them and HIF-1α may occur at
multiple sites, with preferences for certain positions, such as those occupied by Leu559
and Leu562, those occupied by acidic residues and that occupied by Leu574 (Kageyama et
al. 2004).
The third potential hydroxylation site of HIF-2α that was studied here but was not
hydroxylated by any of the three isoenzymes contains the minimum sequence
requirement -Ala-Pro- but the suggested core motif has the sequence -Phe-Gln-Pro-LeuAla-Pro- instead of -Leu-X-X-Leu-Ala-Pro-. Even though single substitutions of
phenylalanine for the first leucine in the core sequence of the HIF-1α-like peptide and
glutamine for the following glutamate and proline for the methionine caused only slight
decreases in the hydroxylation efficiency of the HIF-P4Hs, or even an increase, the
possibility cannot be excluded that these three substitutions together may reduce the
hydroxylation efficiency markedly, but no triple-mutant peptide of this conformation was
studied in the present work. An additional major difference in the third HIF-2α peptide is
that it has only one acidic residue. Our data showed that up to three of the seven acidic
residues in the HIF-1α peptide could be replaced by alanine with no negative effects, but
replacement of all but one of them would have inactivated the peptide completely.
Furthermore, the only acidic residue of this HIF-2α peptide is in the position occupied by
Leu574 in HIF-1α, which has been reported to be important for interaction (Kageyama et
al. 2004). The fact that this third HIF-2α peptide was not hydroxylated supports the
suggestion that several other residues in addition to -Ala-Pro- are required for HIF-P4Hsubstrate interaction.
FIH requires even longer substrates than the HIF-P4Hs, as the omission of only four
residues from the C terminus of the 35-residue HIF-1α peptide was enough to reduce the
enzyme activity to only 26%. The present data are in agreement with results of
crystallization studies on FIH in which a HIF-1α CAD peptide at least 20 residues in
length was required to obtain FIH-CAD crystals (Elkins et al. 2003). The same
crystallization data revealed the presence of two interaction sites in HIF-1α, site 1
involving residues 795-806, around the hydroxylatable asparagine, and site 2 located
more towards the C terminus, comprising residues 813-822. It has been suggested that the
latter may represent a weaker binding site (Elkins et al. 2003), but the present data
indicate that this second site is also required for effective interaction between FIH and
HIF-1α, as the omission of only four residues (819-822) from the C terminus of the 35-
75
residue peptide caused a significant decrease in enzyme activity. Residues 788-794,
preceding interaction site 1, were also found to be important, as their omission had a
marked effect.
FIH was also found to hydroxylate peptides corresponding to the hydroxylation site of
HIF-2α, but far less efficiently than those related to HIF-1α. This suggests that FIH may
act less effectively on HIF-2α than on HIF-1α in vivo and that the transcriptional
coactivator p300 may thus form a complex more easily with the HIF-2α, promoting the
expression of its target genes. Specific FIH inhibitors may therefore be more effective on
genes regulated by HIF-1α than on those regulated by HIF-2α. No conserved substrate
core sequence has been defined for FIH, and the only residue that has been reported to be
required in addition to the hydroxylatable Asn803, is Val802 in the HIF-1α sequence (Linke
et al. 2004). The sequence around the hydroxylatable asparagine is highly conserved
between the two human HIF-α isoforms, as also between isoforms of different
mammalian species, the residue preceding asparagine being valine in all the HIF-αs
studied (Linke et al. 2004). However, as shown here, the hydroxylatable asparagine and
the valine preceding it are not sufficient for hydroxylation, supporting the suggestion that
substrate recognition by FIH requires other sequence determinants at separate locations
(Mahon et al. 2001, Elkins et al. 2003), leading to a requirement for long peptide
substrates.
6.3 Cosubstrate requirements of the HIF-P4Hs and FIH
Our results show distinct differences between the HIF-P4Hs, FIH and the C-P4Hs in their
Km values for reaction cosubstrates, the most significant difference being found with
respect to oxygen. The Km of the HIF-P4Hs for oxygen is slightly above the atmospheric
concentration of this molecule and 6 times higher than that of type I C-P4H, while the Km
of FIH for oxygen is distinctly lower than those of the HIF-P4Hs but still clearly higher
than that of type I C-P4H. These data indicate that the HIF-P4Hs and FIH act as true
oxygen sensors. Even minor decreases in O2 concentration are likely to influence the
activities of the HIF-P4Hs, leading to increased HIF-α stabilization and activation of
HIF-regulated genes. A larger decrease in O2 concentration is required to reduce FIH
activity, which then allows the binding of p300 to HIF-α and leads to maximal
transcriptional activity.
It has been reported previously that there is a marked decrease in asparagine
hydroxylation of HIF-1α in hypoxia, supporting the role of FIH as an oxygen sensor
(Hewitson et al. 2002), but there are also data that argue against such role for FIH. It has
been suggested that its activity may not be regulated solely by its affinity to oxygen, as
some FIH activity has remained even at low oxygen concentrations (Mahon et al. 2001,
Lando et al. 2002a, Stolze et al. 2004). Instead, the oxygen dependence of hydroxylation
by FIH could be due to its substrate recognition mechanism, which requires formation of
a complex with pVHL and HIF-1α (Lee et al. 2003). In normoxia FIH would bind to
HIF-1α only after pVHL had formed a complex with the latter through hydroxyproline,
and in hypoxia FIH could not bind to HIF-1α in the absence of an HIF-1α-pVHL
complex. It is important to understand the meaning of a Km value, the concentration at
76
which half of the maximal reaction velocity is attained. The high Km values of HIF-P4Hs
and FIH for oxygen do not mean that these enzymes would be inactive at lower O2
concentrations, for they in fact showed some activity even with the lowest oxygen
concentrations used in the in vitro activity assays, 10 and 20 µM, and recent in vivo
analyses have shown that HIF-P4H-2 can act on HIF-α at 2% O2, and FIH is active even
at 0.2% O2 (Stolze et al. 2004), results that are not contradicted by the in vitro ones
obtained here. The present results and those of the above in vivo study indicate that the
two types of hydroxylase involved in the HIF response have different demands in terms
of oxygen concentration, FIH being capable of functioning at lower concentrations. The
differences in Km values between the HIF-P4Hs and type I C-P4H are consistent with the
functions of the two classes of P4Hs, i.e the HIF-P4Hs act as effective oxygen sensors, so
that even small decreases in O2 influence their activities, whereas the C-P4Hs must be
able to act even in situations entailing low O2 concentrations, e.g. in wounds and in
tissues with low vascularity.
The Km values of the HIF-P4Hs for 2-oxoglutarate were 2 to 3-fold higher than those
of FIH and type I C-P4H. In FIH (Dann et al. 2002, Elkins et al. 2003, Lee et al. 2003)
and in the C-P4Hs (Kukkola et al. 2003, Van Den Diepstraten et al. 2003 and for review,
see Myllyharju 2003) the residue binding the C-5 carboxyl group of 2-oxoglutarate is a
lysine, whereas in almost all other 2-oxoglutarate dioxygenases (Myllyharju 2003),
including the HIF-P4Hs (Bruick & McKnight 2001, Epstein et al. 2001, Ivan et al. 2002),
this residue is an arginine which cannot be replaced in the HIF-P4Hs by a lysine, as
shown here. There are nevertheless no data to indicate whether the clear difference in Km
values for 2-oxoglutarate is due to the differences in the 2-oxoglutarate-binding residue
or some other properties of the catalytic site.
Ascorbate is known to be required for full catalytic activity of the C-P4Hs both in
vitro and in vivo (Kivirikko & Myllyharju 1998, Kivirikko & Pihlajaniemi 1998,
Myllyharju 2003). Nevertheless, it is not needed in the actual hydroxylation reaction but
is required in the uncoupled decarboxylation of 2-oxoglutarate, in which it is consumed
stoichiometrically (Myllylä et al. 1984). The highly reactive ferryl ion formed at the
catalytic site during the first half-reaction of the catalytic cycle leads to enzyme
inactivation, but when ascorbate is present, it serves as an alternative oxidizable substrate
and the enzyme is converted back to its active form (Myllylä et al. 1984, Kivirikko &
Pihlajaniemi 1998). As the HIF-P4Hs were also found to catalyze uncoupled 2oxoglutarate decarboxylation, it seems highly likely that the main function of ascorbate in
the HIF-P4H reaction is the same as in the C-P4H reaction.
The Km values of the HIF-P4Hs for iron could not be determined with crude enzyme
preparations due to their high activity levels even in the absence of added iron. Even the
purified HIF-P4Hs had considerable activity without any iron additions, and apparent Km
values could therefore only be determined using such preparations. The Km values of
HIF-P4Hs 1 and 2 for iron were more than 15-fold and 65-fold lower than those of FIH
and type I C-P4H, respectively, and that of HIF-P4H-3 5-fold and 20-fold lower. The
optimal iron concentration in the HIF-P4H reaction mixture, 1-5 µM, was also much
lower than the 50 µM required for the C-P4Hs (Kivirikko & Myllylä 1982). The present
data therefore indicate that the HIF-P4Hs differ distinctly from FIH and the C-P4Hs in
binding iron more tightly.
77
6.4 Inhibition of the HIF-P4Hs and FIH by 2-oxoglutarate analogues
Small molecules that would increase the levels of HIF or enhance its transcriptional
activity are believed to be beneficial in diseases characterized by acute or chronic
ischaemia, such as myocardial infarction, stroke, diabetes and peripheral vascular disease
(Bruick & McKnight 2001, Epstein et al. 2001, Ivan et al. 2002). Previous work has
already shown that some small molecule inhibitors of C-P4Hs also inhibit HIF-P4Hs
(Epstein et al. 2001, Ivan et al. 2002) and FIH (Hewitson et al. 2002, Elkins et al. 2003),
and in the present study distinct differences were found between the HIF-P4Hs, FIH and
type I C-P4H in their inhibition by several 2-oxoglutarate analogues. All the compounds
tested were less effective inhibitors of the HIF-P4Hs and FIH than of the C-P4Hs, which
is not surprising, because they were selected from those known to inhibit C-P4Hs
effectively. Differences in inhibition properties were also found between the HIF-P4Hs
and FIH, in that the two most potent inhibitors of the HIF-P4Hs, for example, were the
least efficient inhibitors of FIH. These data suggest that there must be differences in the
structures of the 2-oxoglutarate-binding sites between the three classes of hydroxylase,
and that it should therefore be possible to develop specific inhibitors for the HIF-P4Hs
and FIH that would be therapeutically beneficial. The 2-oxoglutarate-binding residue of
the HIF-P4Hs and C-P4Hs and many other 2-oxoglutarate dioxygenases is located in the
eighth β strand of the jellyroll motif (Bruick & McKnight 2001, Epstein et al. 2001, Ivan
et al. 2002, Kukkola et al. 2003, Myllyharju 2003), whereas that of FIH is in the fourth
strand (Dann et al. 2002, Elkins et al. 2003, Lee et al. 2003). It is likely, however, that
several residues at the 2-oxoglutarate-binding site influence the binding properties of
various inhibitors and not just the actual 2-oxoglutarate-binding residue.
6.5 Effect of desferrioxamine and metals on the HIF-P4Hs and FIH
The fact that iron chelator DFO inhibited type I C-P4H much more effectively than it did
the HIF-P4Hs is in agreement with the strong iron binding of the HIF-P4Hs. DFO
inhibited HIF-P4H-3 most effectively, which agrees with the finding that the apparent Km
of this isoenzyme for iron is higher than those of the other two. It is possible that DFO
may not be able to enter the cosubstrate binding sites of any of the HIF and collagen
P4Hs, the effective inhibition of the C-P4Hs being probably due to chelation of the free
iron from the reaction mixture. The effect of DFO on the HIF-P4Hs will be much weaker,
however, because these enzymes bind the iron more tightly and retain a significant level
of their activity even when all the free iron has been chelated from the reaction mixture.
When DFO was added to cultured insect cells synthesizing HIF-P4H-2, the inhibition
was more effective, even though complete inhibition was still not obtained, unlike in the
case of the synthesis of C-P4Hs in the presence of DFO (Kivirikko & Myllylä 1982,
Kivirikko & Pihlajaniemi 1998). The activity of the iron-free C-P4Hs can be restored by
adding iron to the reaction mixture (Kivirikko & Myllylä 1982, Kivirikko & Pihlajaniemi
1998), whereas such an addition of iron did not reactivate HIF-P4H-2. This agrees with
the findings that addition of iron prevents inactivation even of crude HIF-P4Hs and that it
stabilizes the HIF-P4Hs during their purification. The present data differ distinctly from a
78
previous report that pVHL capture of a HIF-1α peptide hydroxylated with reticulocyte
lysates programmed with the HIF-P4Hs was already strongly inhibited with 100 µM DFO
in the lysate (Epstein et al. 2001).
Several models have been presented to explain the mode of HIF-α stabilization by
metal ions. It has been suggested that cobalt becomes bound to the catalytic site of HIFP4Hs instead of iron and therefore inhibits their activity (Epstein et al. 2001), that cobalt
may stabilize HIF-α by binding directly to it and thereby preventing HIF-α - pVHL
interaction (Yuan et al. 2001, 2003), and that stabilization may be independent of pVHL
(Kanaya & Kamitani 2003). In addition, cobalt and nickel have been reported to cause
inhibition of HIF-P4Hs by depleting the intracellular ascorbate required for the enzyme
reaction (Salnikow et al. 2004) and that they activate the phosphatidylinositol-3-kinase
pathway involved in the regulation of HIF-1α transcription (Chachami et al. 2004, Li et
al. 2004).
As many metals inhibit C-P4Hs competitively with respect to iron (Kivirikko &
Pihlajaniemi 1998), it was expected that the metals would also effectively inhibit the
HIF-P4Hs and FIH. Several metals were found to be effective competitive inhibitors of
FIH, with Ki values similar to or even lower than their IC50 for type I C-P4H. Data
obtained from structural studies have also revealed that metals inhibit FIH competitively
at the iron binding site (Elkins et al. 2003). Surprisingly, however, inhibition of the HIFP4Hs by metals was ineffective, particularly that of HIF-P4H-2. Crude HIF-P4H-2
preparations were only 20-25% inhibited by 100 µM and 200 µM concentrations of
cobalt and nickel, respectively, while these concentrations caused stabilization of HIF-1α
in cultured HEK and Hep3B cells to a level detectable by Western blotting. Due to the
high activity of the purified HIF-P4Hs even in the absence of added iron, their mode of
inhibition by metals could not be determined. The recombinant HIF-P4H-2 synthesized in
the presence of cobalt was nevertheless to a large extent inactive and could not be
reactivated by the addition of iron to the in vitro enzyme reaction, suggesting that cobalt
does not inhibit HIF-P4Hs competitively with respect to iron, a result that agrees with
that of Salnikow et al. (2004). Cobalt also rapidly inactivated crude HIF-P4Hs during
storage at 4°C, this effect perhaps being due in part to the replacement of iron at the
catalytic site and in part to other toxic effects of cobalt on the HIF-P4Hs.
The function of the HIF-P4Hs as oxygen sensors in vivo requires that they should be
present in cells at close to the limiting concentrations, so that even a slight decrease in
oxygen concentration will depress the activities of these enzymes to an extent that leads
to HIF-1α stabilization. It is therefore possible that even small decreases in the activities
of the HIF-P4Hs brought about by cobalt and nickel may be sufficient to stabilize HIF1α. It is also likely, however, that the other mechanisms suggested for the metal-induced
stabilization of HIF-1α may have contributed to or been even more important than
inhibition of the HIF-P4Hs. Whatever the responsible mechanisms are, once HIF-1α is
stabilized in the presence of cobalt or nickel, it is likely to be transcriptionally fully active
due to the highly effective inhibition of FIH by these metals. This may explain the fact
that even low cobalt concentrations are enough to increase EPO production in Hep3B
cells (Goldberg et al. 1988).
Zinc was a more effective inhibitor of all three HIF-P4Hs than nickel and cobalt, but it
did not stabilize HIF-1α in cultured HEK and Hep3B cells to detectable level, even
though 500 µM zinc has been reported to stabilize HIF-1α in Hep3B cells, albeit failing
79
to induce EPO production (Chun et al. 2000). In our study zinc was found to be an
effective inhibitor of FIH, and 5-100 µM concentrations increased VEGF production
slightly in both HEK and Hep3B cells. Higher concentrations were already toxic to the
cells, and no VEGF could be detected. It has been found previously that although zinc
induces HIF-1α stabilization, 500 µM zinc inactivates HIF-1 by inhibiting nuclear
translocation of the β subunit (Chun et al. 2000). Zinc has also been shown to induce
alternative splicing of HIF-1α, forming a variant which inhibits HIF-1 activity and
reduces the expression of HIF target genes (Chun et al. 2001). It seems possible that the
inactivation of HIF-1 may not be complete with low zinc concentrations, and therefore
HIF could induce the expression of its target genes such as VEGF, as observed in the
present study. This would be consistent with the finding that zinc is an effective inhibitor
of FIH, leading to an increase in the transcriptional activity of HIF.
6.6 Expression and alternative splicing of HIF-P4H mRNAs
The regulation of the HIF system is complex, comprising regulation of HIF stability and
activity. It is known so far that under normoxic conditions the stability and activity of
HIF is regulated by the oxygen-dependent enzymatic reactions catalyzed by the HIFP4Hs and FIH, which lead to proteasomal degradation and inhibition of the
transcriptional activity of HIF (Bruick & McKnight 2001, Epstein et al. 2001, Ivan et al.
2001, 2002, Jaakkola et al. 2001, Mahon et al. 2001, Masson et al. 2001, Yu et al. 2001,
Fedele et al. 2002, Hewitson et al. 2002, Lando et al. 2002a, Elkins et al. 2003). In
hypoxia, the HIF hydroxylating enzymes are inactive and HIF can maximally induce the
expression of its target genes. These two pathways are nevertheless unlikely to be the
only determinants affecting HIF-α stability and activity, and several other possible
regulatory mechanisms have been presented. Our results show that the HIF-P4H-2 and 3
mRNAs undergo alternative splicing that probably leads to the synthesis of catalytically
inactive HIF-P4H polypeptides in addition to the active enzymes, thus suggesting a
further complexity in the regulation of HIF. Interestingly, another alternative splicing
occurring in the HIF system has been described, namely the regulation of HIF-3α
involves alternative splicing which is increased under hypoxic conditions (Makino et al.
2001, 2002). It has been also shown that zinc causes alternative splicing of HIF-1α,
leading to reduced expression of HIF target genes (Chun et al. 2001). Further studies will
be required to investigate whether alternative splicing of the HIF-P4H mRNAs is affected
by the amount of oxygen or the presence of HIF-P4H inhibitors, for instance. It is also
interesting that no evidence for alternative splicing was found in the case of the mRNA of
HIF-P4H-1, which is thought to play a less important role in the regulation of HIF-α
stability.
The present results concerning the expression of HIF-P4H mRNA in various tissues
support the existence of tissue-specific but overlapping expression patterns, as suggested
elsewhere (Lieb et al. 2002, Oehme et al. 2002, Cioffi et al. 2003). As shown by
Appelhoff and co-workers (2004) the cell-type specific expression patterns of the HIFP4Hs together with their differential induction patterns indicate differential roles for the
isoenzymes in regulating the levels of HIF-α under varying conditions.
7 Future prospects
Even though the mRNA expression patterns of HIF-P4Hs have been studied previously,
only relatively limited data are available on the protein expression of these enzymes. In
one study where antibodies to HIF-P4Hs were used to detect HIF-P4H proteins in
extracts of different cell types, the results suggested that the protein expression patterns
of isoenzymes 2 and 3 parallel their mRNA expression patterns whereas that of
isoenzyme 1 cannot be predicted from its mRNA expression (Appelhoff et al. 2004).
These results also showed that both the mRNA and protein expression of HIF-P4Hs 2 and
3 are induced in hypoxia. It has also been established that HIF-P4H-2 is the main
regulator of HIF-1α, and that HIF-2α is regulated mainly by HIF-P4H-3 together with
HIF-P4H-1. More data would be required, however, to determine the possible individual
tasks of the isoenzymes and to verify whether their function is tissue-specific, i.e.
whether one of the isoenzymes is the dominant or only functional enzyme in a certain
tissue. Some preliminary data have been presented in a scientific seminar on knock-out
studies of HIF-P4Hs, suggesting that HIF-P4H-2 may play an important role in
embryogenesis, whereas knock-out mice in which the other two isoenzymes were the
objects of intervention were healthy and fertile. More work, perhaps exploiting
conditional gene inactivation, would be needed to study the role of isoenzymes 1 and 3 in
particular in tissues after birth. By employing conditional gene targeting, it would be
possible to restrict the inactivation of the target gene to a certain organ and
developmental stage.
It would also be interesting to know whether HIF is the only substrate of HIF-P4Hs
and FIH. The task of finding other possible substrates will be a challenging one, however,
as shown by the present along with some previous data, due to the lack of specific
conserved core sequences. Instead of computerized protein searches, however, proteinprotein interaction screening studies could be exploited in order to identify additional
substrates.
Examination of the inhibitory properties of HIF-P4Hs and FIH by means of various
small-molecule inhibitors could provide good data for developing specific inhibitors for
these enzymes that would be beneficial for treatment or preconditioning purposes in
connection with diseases involving ischaemia. The present results show that there are
differences between the classes of hydroxylases studied, and even between the HIF-P4H
81
isoenzymes. It has been predicted that by specific inhibition of a certain HIF-P4H
enzyme it may be possible to selectively modify the HIF response. For example,
inhibiting the expression of HIF-P4H-2 would provide extensive specific inhibition of the
HIF response in normoxic tissues, whereas inhibition of HIF-P4H-3 might cause a more
restricted response in certain tissues (Appelhoff et al. 2004).
References
Abrams EN & Andrew DJ (2002) Prolyl 4-hydroxylase α-related proteins in Drosophila
melanogaster:
tissue-specific
embryonic
expression
of
the
99F8-9
cluster.
Mech Dev 112: 165-171.
Adams MD, Celniker SE, Holt RA, Evans CA, Gocayne JD, Amanatides PG, Scherer SE, Li PW,
Hoskins RA, Galle RF, George RA, Lewis SE, Richards S, Ashburner M, Henderson SN,
Sutton GG, Wortman JR, Yandell MD, Zhang Q, Chen LX, Brandon RC, Rogers YH, Blazej
RG, Champe M, Pfeiffer BD, Wan KH, Doyle C, Baxter EG, Helt G, Nelson CR, Gabor GL,
Abril JF, Agbayani A, An HJ, Andrews-Pfannkoch C, Baldwin D, Ballew RM, Basu A,
Baxendale J, Bayraktaroglu L, Beasley EM, Beeson KY, Benos PV, Berman BP, Bhandari D,
Bolshakov S, Borkova D, Botchan MR, Bouck J, Brokstein P, Brottier P, Burtis KC, Busam
DA, Butler H, Cadieu E, Center A, Chandra I, Cherry JM, Cawley S, Dahlke C, Davenport LB,
Davies P, de Pablos B, Delcher A, Deng Z, Mays AD, Dew I, Dietz SM, Dodson K, Doup LE,
Downes M, Dugan-Rocha S, Dunkov BC, Dunn P, Durbin KJ, Evangelista CC, Ferraz C,
Ferriera S, Fleischmann W, Fosler C, Gabrielian AE, Garg NS, Gelbart WM, Glasser K, Glodek
A, Gong F, Gorrell JH, Gu Z, Guan P, Harris M, Harris NL, Harvey D, Heiman TJ, Hernandez
JR, Houck J, Hostin D, Houston KA, Howland TJ, Wei MH, Ibegwam C, Jalali M, Kalush F,
Karpen GH, Ke Z, Kennison JA, Ketchum KA, Kimmel BE, Kodira CD, Kraft C, Kravitz S,
Kulp D, Lai Z, Lasko P, Lei Y, Levitsky AA, Li J, Li Z, Liang Y, Lin X, Liu X, Mattei B,
McIntosh TC, McLeod MP, McPherson D, Merkulov G, Milshina NV, Mobarry C, Morris J,
Moshrefi A, Mount SM, Moy M, Murphy B, Murphy L, Muzny DM, Nelson DL, Nelson DR,
Nelson KA, Nixon K, Nusskern DR, Pacleb JM, Palazzolo M, Pittman GS, Pan S, Pollard J,
Puri V, Reese MG, Reinert K, Remington K, Saunders RD, Scheeler F, Shen H, Shue BC,
Siden-Kiamos I, Simpson M, Skupski MP, Smith T, Spier E, Spradling AC, Stapleton M,
Strong R, Sun E, Svirskas R, Tector C, Turner R, Venter E, Wang AH, Wang X, Wang ZY,
Wassarman DA, Weinstock GM, Weissenbach J, Williams SM, WoodageT, Worley KC, Wu D,
Yang S, Yao QA, Ye J, Yeh RF, Zaveri JS, Zhan M, Zhang G, Zhao Q, Zheng L, Zheng XH,
Zhong FN, Zhong W, Zhou X, Zhu S, Zhu X, Smith HO, Gibbs RA, Myers EW, Rubin GM &
Venter JC (2000) The genome sequence of Drosophila melanogaster. Science 287: 2185-2195.
Adryan B, Decker HJ, Papas TS & Hsu T (2000) Tracheal development and the von Hippel-Lindau
tumor suppressor homolog in Drosophila. Oncogene 19: 2803-2811.
Ambrosini G, Nath AK, Sierra-Honigmann MR & Flores-Riveros J (2002) Transcriptional
activation of the human leptin gene in response to hypoxia. Involvement of hypoxia-inducible
factor 1. J Biol Chem 277: 34601-34609.
83
Annunen P, Autio-Harmainen H & Kivirikko KI (1998) The novel type II prolyl 4-hydroxylase is
the main enzyme form in chondrocytes and capillary endothelial cells, whereas the type I
enzyme predominates in most cells. J Biol Chem 273: 5989-5992.
Annunen P, Helaakoski T, Myllyharju J, Veijola J, Pihlajaniemi T & Kivirikko KI (1997) Cloning
of the human prolyl 4-hydroxylase α subunit isoform α(II) and characterization of the type II
enzyme tetramer. The α(I) and α(II) subunits do not form a mixed α(I)α(II)β2 tetramer. J Biol
Chem 272: 17342-17348.
Annunen P, Koivunen P & Kivirikko KI (1999) Cloning of the α subunit of prolyl 4-hydroxylase
from Drosophila and expression and characterization of the corresponding enzyme tetramer
with some unique properties. J Biol Chem 274: 6790-6796.
Appelhoff RJ, Tian Y-M, Raval RJ, Turley H, Harris AL, Pugh CW, Ratcliffe PJ & Gleadle JM
(2004) Differential function of the prolyl hydroxylases, PHD1, 2 and 3 in the regulation of
hypoxia-inducible factor (HIF), J Biol Chem 279: 38458-38465.
Aprelikova O, Chandramouli GV, Wood M, Vasselli JR, Riss J, Maranchie JK, Linehan WM,
Barret JC (2004) J Cell Biochem 92: 491-501.
Baader E, Tschank G, Baringhaus KH, Burghard H & Günzler V (1994) Inhibition of prolyl 4hydroxylase by oxalyl amino acid derivatives in vitro, in isolated microsomes and in embryonic
chicken tissues. Biochem J 300: 525-530.
Bacon NC, Wappner P, O'Rourke JF, Bartlett SM, Shilo B, Pugh CW & Ratcliffe PJ (1998)
Regulation of the Drosophila bHLH-PAS protein Sima by hypoxia: functional evidence for
homology with mammalian HIF-1α. Biochem Biophys Res Commun 249:811-816.
Bassuk JA, Kao WW, Herzer P, Kedersha NL, Seyer J, DeMartino JA, Daugherty BL, Mark GE 3rd
& Berg RA (1989) Prolyl 4-hydroxylase: molecular cloning and the primary structure of the α
subunit from chicken embryo. Proc Natl Acad Sci USA 86: 7382-7386.
Bella J, Eaton M, Brodsky B & Berman HM (1994) Crystal and molecular structure of a collagenlike peptide at 1.9 Å resolution. Science 266: 75-81.
Berg RA & Prockop DJ (1973) The thermal transition of a non-hydroxylated form of collagen.
Evidence for a role for hydroxyproline in stabilizing the triple-helix of collagen. Biochem
Biophys Res Commun 52: 115-20.
Berra E, Benizri E, Ginouvès A, Volmat V, Roux D & Pouysségur J (2003) HIF prolyl-hydroxylase
2 is the key oxygen sensor setting low steady-state levels of HIF-1α in normoxia. EMBO J 22:
4082-4090.
Bianchi L, Tacchini L & Cairo G (1999) HIF-1-mediated activation of transferrin receptor gene
transactivation by iron chelation. Nucleic Acids Res 27: 4223-4227.
Binley K, Askham Z, Martin L, Spearman H, Day D, Kingsman S & Naylor S (2003) Hypoxiamediated tumour targeting. Gene Ther 10: 540-549.
Bruick RK (2000) Expression of the gene encoding the proapoptotic Nip3 protein is induced by
hypoxia. Proc Natl Acad Sci USA 97: 9082-9087.
Bruick RK & McKnight SL (2001) A conserved family of prolyl-4-hydroxylases that modify HIF.
Science 294: 1337-1340.
Bruick RK & McKnight SL (2002) Oxygen sensing gets a second wind. Science 295: 807-808.
Brusselmans K, Bono F, Maxwell P, Dor Y, Dewerchin M, Collen D, Herbert JM & Carmeliet P
(2001) Hypoxia-inducible factor-2α (HIF-2α) is involved in the apoptotic response to
hypoglycaemia but not to hypoxia. J Biol Chem 276: 39192-39196.
Brusselmans K, Compernolle V, Tjwa M, Wiesener MS, Maxwell PH, Collen D & Carmeliet P
(2003) Heterozygous deficiency of hypoxia-inducible factor-2α protects mice against
pulmonary hypertension and right ventricular dysfunction during prolonged hypoxia. J Clin
Invest 111: 1519-1527.
84
Caniggia I, Mostachfi H, Winter J, Gassmann M, Lye SJ, Kuliszewski M & Post M (2000)
Hypoxia-inducible factor-1 mediates the biological effects of oxygen on human trophoblast
differentiation through TGFβ3. J Clin Invest 105: 577-587.
Chachami G, Simos G, Hatziefthimiou A, Bonanou S, Molyvdas P-A & Paraskeva E (2004) Cobalt
induces HIF-1α expression in ASM cells by a ROS and PI3K dependent manner. Am J Respir
Cell Mol Biol, in press.
Chan DA, Sutphin PD, Denko NC & Giaccia AJ (2002) Role of prolyl hydroxylation in
oncogenically stabilized hypoxia-inducible factor-1α. J Biol Chem 277: 40112-40117.
Chandel NS, Maltepe E, Goldwasser E, Mathieu CE, Simon MC & Schumacker PT (1998)
Mitochondrial reactive oxygen species trigger hypoxia-induced transcription. Proc Natl Acad
Sci USA 95: 11715-11720.
Chandel NS, McClintock DS, Feliciano CE, Wood TM, Melendez JA, Rodiguez AM &
Schumacker PT (2000) Reactive oxygen species generated at mitochondrial complex III
stabilize hypoxia-inducible factor-1α during hypoxia. J Biol Chem 275: 25130-25138.
Chen C, Pore N, Behrooz A, Ismail-Beigi F & Maity A (2001) Regulation of glut1 mRNA by
hypoxia-inducible factor-1. Interaction between H-ras and hypoxia. J Biol Chem 276: 95199525.
Chun Y-S, Choi E, Kim GT, Lee MJ, Lee MJ, Lee SE, Kim MS & Park JW (2000) Zinc induces
the accumulation of hypoxia-inducible factor (HIF)-1α, but inhibits the nuclear translocation of
HIF-1β, causing HIF-1 inactivation. Biochem Biophys Res Commun 268: 652-656.
Chun Y-S, Choi E, Yeo E-J, Lee JH, Kim M-S & Park J-W (2001) A new HIF-1 alpha variant
induced by zinc ion suppresses HIF-1-mediated hypoxic responses. J Cell Sci 114: 4051-4061.
Cioffi CL, Liu XQ, Kosinski PA, Garay M & Bowen BR (2003) Differential regulation of HIF-1α
prolyl-4-hydroxylase genes by hypoxia in human cardiocvascular cells. Biochem Biophys Res
Commun 303: 947-953.
Compernolle V, Brusselmans K, Acker T, Hoet P, Tjwa M, Beck H, Plaisance S, Dor Y, Keshet E,
Lupu F, Nemery B, Dewerchin M, Van Veldhoven P, Plate K, Moons L, Collen D & Carmeliet
P (2002) Loss of HIF-2α and inhibition of VEGF impair fetal lung maturation, whereas
treatment with VEGF prevents fatal respiratory distress in premature mice. Nat Med 8: 702-710.
Cormier-Regard S, Nguyen SV & Claycomb WC (1998) Adrenomedullin gene expression is
developmentally regulated and induced by hypoxia in rat ventricular cardiac myocytes. J Biol
Chem 273: 17787-17792.
Cox GN, Kusch M & Edgar RS (1981) Cuticle of Caenorhabditis elegans: its isolation and partial
characterization. J Cell Biol 90: 7-17.
Cramer T, Yamanishi Y, Clausen BE, Forster I, Pawlinski R, Mackman N, Haase VH, Jaenisch R,
Corr M, Nizet V, Firestein GS, Gerber HP, Ferrara N & Johnson RS (2003) HIF-1α is essential
for myeloid cell-mediated inflammation.Cell 112: 645-657.
Crews ST (1998) Control of cell lineage-specific development and transcription by bHLH-PAS
proteins. Genes Dev 12: 607-620.
Crews ST, Thomas JB & Goodman CS (1988) The Drosophila single-minded gene encodes a
nuclear
protein
with
sequence
similarity
to
the
per
gene
product.
Cell 52: 143-151.
Cunliffe CJ, Franklin TJ, Hales NJ & Hill GB (1992) Novel inhibitors of prolyl 4-hydroxylase. 3.
Inhibition by the substrate analogue N-oxaloglycine and its derivatives. J Med Chem 35: 26522658.
Dalgard CL, Lu H, Mohyeldin A & Verma A (2004) Endogenous 2-oxoacids differentially regulate
expression of oxygen sensors. Biochem J 380: 419-424.
D’Angelo G, Duplan E, Boyer N, Vigne P & Frelin C (2003) Hypoxia up-regulates prolyl
hydroxylase activity. J Biol Chem 278: 38183-38187.
85
Dann CE, Bruick RK & Deisenhofer J (2002) Structure of factor-inhibiting hypoxia-inducible
factor 1: an asparaginyl hydroxylase involved in the hypoxic response pathway. Proc Natl Acad
Sci USA 99: 15351-15356.
Darby NJ, Penka E & Vincentelli R (1998) The multi-domain structure of protein disulfide
isomerase is essential for high catalytic efficiency. J Mol Biol 276: 239-247.
del Peso L, Castellanos MC, Temes E, Martín-Puig S, Cuevast Y, Olmos G & Landázuri MO
(2003) The von Hippel-Lindau/hypoxia-inducible factor (HIF) pathway regulates the
transcription of the HIF-proline hydroxylase genes in response to low oxygen. J Biol Chem 278:
48690-48695.
Dinchuk JE, Henderson NL, Burn TC, Huber R, Ho SP, Link J, O’Neil KT, Focht RJ, Scully MS,
Hollis JM, Hollis GF & Friedman PA (2000) Aspartyl β-hydroxylase (Asph) and an
evolutionarily conserved isoforms of Asph missing the catalytic domain share exons with
junction. J Biol Chem 275: 39543-39554.
Ebert BL, Firth JD & Ratcliffe PJ (1995) Hypoxia and mitochondrial inhibitors regulate expression
of glucose transporter-1 via distinct cis-acting sequences. J Biol Chem 276: 29083-29089.
Ebert BL, Gleadle JM, O’Rourke JF, Bartlett SM, Poulton J & Ratcliffe PJ (1996) Isoenzymespecific regulation of genes involved in energy metabolism by hypoxia: similarities with the
regulation of erythropoietin. Biochem J 313: 809-814.
Eckhart AD, Yang N, Xin X & Faber JE (1997) Characterization of the α1B-adrenergic receptor
gene promoter region and hypoxia regulatory elements in vascular smooth muscle. Proc Natl
Acad Sci USA 94: 9487-9492.
Edman JC, Ellis L, Blacher RW, Roth RA & Rutter WJ (1985) Sequence of protein disulphide
isomerase and implications of its relationship to thioredoxin.Nature 317: 267-270.
Elkins JM, Hewitson KS, McNeill LA, Seibel JF, Schlemminger I, Pugh CW, Ratcliffe PJ &
Schofield CJ (2003) Structure of factor-inhibiting hypoxia-inducible factor (HIF) reveals
mechanism of oxidative modification of HIF-1α. J Biol Chem 278: 1802-1806.
Ema M, Taya S, Yokotani N, Sogawa K, Matsuda Y & Fujuu-Kuriyama Y ( 1997) A novel bHLHPAS factor with close similarity to hypoxia-inducible factor 1α regulates the VEGF expression
and is potentially involved in lung and vascular development. Proc Natl Acad Sci USA 94:
4273-4278.
Epstein ACR, Gleadle JM, McNeill JM, Hewitson KS, O’Rourke J, Mole DR, Mukherji M, Metzen
E, Wilson MI, Dhanda A, Tian Y-M, Masson N, Hamilton DL, Jaakkola P, Barstead R,
Hodgkin J, Maxwell PH, Pugh CW, Schofield CJ & Ratcliffe PJ (2001) C. elegans EGL-9 and
mammalian homologs define a family of dioxygenases that regulate HIF by prolyl
hydroxylation. Cell 107: 43-54.
Erez N, Milyavsky M, Eilam R, Shats I, Goldfinger N & Rotter V (2003) Epression of prolylhydroxylase-1 (PHD1/EGLN2) suppresses hypoxia inducible factor-1α activation and inhibits
tumor growth. Cancer Res 63: 8777-8783.
Eriksson M, Myllyharju J, Tu H, Hellman M & Kivirikko KI (1999) Evidence for 4-hydroxyproline
in viral proteins. Characterization of a viral prolyl 4-hydroxylase and its peptide substrates. J
Biol Chem 274: 22131-22134.
Fedele AO, Whitelaw ML & Peet DJ (2002) Regulation of gene expression by the hypoxiainducible factors. Mol Interv 2: 229-243.
Feldser D, Agani F, Iyer NV, Pak B, Ferreira G & Semenza GL (1999) Reciprocal positive
regulation of hypoxia-inducible factor 1α and insulin-like growth factor 2. Cancer Res 59:
3915-3918.
Ferré-D'Amare AR, Prendergast GC, Ziff EB & Burley SK (1993) Recognition by Max of its
cognate DNA through a dimeric b/HLH/Z domain. Nature 363: 38-45.
86
Firth JD, Ebert BL, Pugh CW & Ratcliffe PJ (1994) Oxygen-regulated control elements in the
phosphoglycerate kinase 1 and lactate dehydrogenase A genes: similarities with the
erythropoietin 3’ enhancer. Proc Natl Acad Sci USA 91: 6496-6500.
Flamme I, Fröhlich T, von Reutern M, Kappel A, Damert A & Risau W (1997) HRF, a putative
basic helix-loop-helix-PAS-domain transcription factor is closely related to hypoxia-inducible
factor 1α and developmentally expressed in blood vessels. Mech Dev 63: 51-60.
Forsythe JA, Jiang BH, Iyer NV, Agani F, Leung SW, Koos RD & Semenza GL (1996) Activation
of vascular endothelial growth factor gene transcription by hypoxia-inducible factor. Mol Cell
Biol 16: 4604-4613.
Freedman RB, Gane PJ, Hawkins HC, Hlodan R, McLaughlin SH & Parry JW (1998) Experimental
and theoretical analyses of the domain architecture of mammalian protein disulphide-isomerase.
Biol Chem 379: 321-328.
Freedman RB, Klappa P & Ruddock LW (2002) Protein disulfide isomerases exploit synergy
between catalytic and specific binding domains. EMBO Rep 3: 136-140.
Frei C (2004) Cyclin D/Cdk4: New insights from Drosophila. Cell Cycle 3: 558-560.
Frei C & Edgar BA (2004) Drosophila Cyclin/cdk4 requires Hif-1 prolyl hydroxylase to drive cell
growth. Dev Cell 6: 241-251.
Friedman L, Higgin JJ, Moulder G, Barstead R, Raines RT & Kimble J (2000) Prolyl 4hydroxylase is required for viability and morphogenesis in Caernohabdits elegans. Proc Natl
Acad Sci USA 97: 4736-4741.
Furuta GT, Turner JR, Taylor CT, Hershberg RM, Comeford K, Narravula S, Podolsky DK &
Colgan SP (2001) Hypoxia-inducible factor 1-dependent induction of intestinal trefoil factor
protects barrier function during hypoxia. J Exp Med 193:1027-1034.
Gautron JP, Leblanc P, Bluet-Pajot MT, Pattou E, L'Heritier A, Mounier F, Ponce G, Audinot V,
Rasolonjanahary R & Kordon C (1992) A second endogenous molecular form of mammalian
hypothalamic luteinizing hormone-releasing hormone (LHRH), (hydroxyproline9)LHRH,
releases luteinizing hormone and follicle-stimulating hormone in vitro and in vivo. Mol Cell
Endocrinol 85: 99-107.
Genbacev O, Zhou Y, Ludlow JW & Fisher SJ (1997) Regulation of human placental development
by oxygen tension. Science 277: 1669-1672.
Gerber HP, 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: 23659-23667.
Goldberg MA, Dunning SP & Bunn HF (1988) Regulation of the erythropoietin gene: evidence that
the oxygen sensor is a heme protein. Science 242: 1412-1415.
Goldberg MA & Schneider TJ (1994) Similarities between the oxygen-sensing mechanisms
regulating the expression of vascular endothelial growth factor and erythropoietin. J Biol Chem
269: 4355-4359.
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: 23508-23514.
Graven KK, Yu Q, Pan D, Roncarati JS & Farber HW (1999) Identification of an oxygen
responsive enhancer element in the glyceraldehyde-3-phosphate dehydrogenase gene. Biochim
Biophys Acta 1447: 208-218.
Gu YZ, Moran SM, Hogenesch JB, Wartman L & Bradfield CA (1998) Molecular characterization
and chromosomal localization of a third α-class hypoxia inducible factor subunit, HIF3α.Gene
Expr 7: 205-213.
Hanauske-Abel HM & Günzler V (1982) A stereochemical concept for the catalytic mechanism of
prolylhydroxylase: applicability to classification and design of inhibitors. J Theor Biol 94: 421455.
87
Hanauske-Abel HM & Popowicz AM (2003) The HAG mechanism: a molecular rationale for the
therapeutic application of iron chelators in human diseases involving the 2-oxoacid utilizing
dioxygenases Curr Med Chem 10: 1005-1019.
Hara S, Hamada J, Kobayashi C, Kondo Y & Imura N (2001) Expression and characterization of
hypoxia-inducible factor (HIF)-3α in human kidney: suppression of HIF-mediated gene
expression by HIF-3α. Biochem Biophys Res Commun 287: 808-813.
Heidbreder M, Fröhlich F, Jöhren O, Dendorfer A, Qadri F & Dominiak P (2003) Hypoxia rapidly
activates HIF-3α mRNA expression. FASEB J 17: 1541-1543.
Helaakoski T, Vuori K, Myllylä R, Kivirikko KI & Pihlajaniemi T (1989) Molecular cloning of the
α-subunit of human prolyl 4-hydroxylase: the complete cDNA-derived amino acid sequence
and evidence for alternative splicing of RNA transcripts. Proc Natl Acad Sci USA 86: 43924396.
Helaakoski T, Annunen P, Vuori K, MacNeil IA, Pihlajaniemi T & Kivirikko KI (1995) Cloning,
baculovirus expression, and characterization of a second mouse prolyl 4-hydroxylase α-subunit
isoform: Formation of an α2β2 tetramer with the protein disulfide-isomerase/β subunit. Proc
Natl Acad Sci USA 92: 4427-4431.
Hewitson KS, McNeill LA, Riordan MV, Tian Y-M, Bullock AN, Welford RW, Elkins JM,
Oldham NJ, Bhattacharya S, Gleadle JM, Ratcliffe PJ, Pugh CW & Schofield CJ (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: 26351-26355.
Hieta R, Kukkola L, Permi P, Pirilä P, Kivirikko KI, Kilpeläinen I & Myllyharju J (2003) The
peptide-substrate-binding domain of human collagen prolyl 4-hydroxylases. Backbone
assignments, secondary structure, and binding of proline-rich peptides. J Biol Chem 278:
34966-34974.
Hieta R & Myllyharju J (2002) Cloning and characterization of a low molecular weight prolyl 4hydroxylase from Arabidopsis thaliana. Effective hydroxylation of proline-rich, collagen-like,
and hypoxia-inducible transcription factor α-like peptides. J Biol Chem 277: 23965-23971.
Hill KL, Harfe BD, Dobbins CA & L’Hernault SW (2000) dpy-18 encodes an α-subunit of prolyl
4-hydroxylase in Caenorhabditis elegans. Genetics 155: 1139-1148.
Ho VT & Bunn HF (1996) Effects of transition metals on the expression of the erythropoietin gene:
further evidence that the oxygen sensor is a heme protein. Biochem Biophys Res Commun 223:
175-180.
Hockel M & Vaupel P (2001) Tumor hypoxia: definitions and current clinical, biologic, and
molecular aspects. J Natl Cancer Inst 93: 266-276.
Hogenesch JB, Chan WK, Jackiw VH, Brown RC, Gu Y, Pray-Grant M, Perdew GH & Bradfields
CA (1997) Characterization of a subset of the basic-helix-loop-helix-PAS superfamily that
interacts with components of the dioxin signalling pathway. J Biol Chem 272: 8581-8593.
Holmgren SK, Taylor KM, Bretscher LE & Raines RT (1998) Code for collagens’ stability
deciphered. Nature 392: 666-667.
Hon W-C, Wilson MI, Harlos K, Claridge TDW, Schofield CJ, Pugh CW, Maxwell PH, Ratcliffe
PJ, Stuart DI & Jones EY (2002) Structural basis for the recognition of hydroxyproline in HIF1α by pVHLNature 417: 975-978.
Hopkinson I, Smith SA, Donne A, Gregory H, Franklin TJ, Grant ME & Rosamond J (1994) The
complete cDNA derived sequence of the rat prolyl 4-hydroxylase α subunit. Gene 149: 391392.
Hu C-J, Wang L-Y, Chodosh LA, Keith B & Simon C (2003) Differential roles of hypoxiainducible factor 1α (HIF-1α) and HIF-2α in hypoxic gene regulation. Mol Cell Biol 23: 93619374.
88
Hu J, Discher DJ, Bishopric NH & Webster KA (1998) Hypoxia regulates expression of the
endothelin-1 gene through a proximal hypoxia-inducible factor-1 binding site on the antisense
strand. Biochem Biophys Res Commun 245: 894-899.
Huang J, Zhao Q, Mooney SM & Lee FS (2002) Sequence determinants in hypoxia-inducible
factor-1α for hydroxylation by the prolyl hydroxylases PHD1, PHD2, and PHD3. J Biol Chem
277: 39792-39800.
Huang LE, Arany Z, Livingston DM & Bunn HF (1996) Activation of hypoxia-inducible
transcription factor depends primarily upon redox-sensitive stabilization of its α subunit. J Biol
Chem 271: 32253-32259.
Huang LE & Bunn HF (2003) Hypoxia-inducible factor and its biomedical relevance. J Biol Chem
278: 19575-19578.
Huang LE, Gu J, Schau M & Bunn HF (1998) Regulation of hypoxia-inducible factor 1α is
mediated by an O2-dependent degradation domain via the ubiquitin-proteasome pathway. Proc
Natl Acad Sci USA 95: 7987-7992.
Huang LE, Pete EA, Schau m, Milligan J & Gu J (2002) Leu-574 of HIF-1α is essential for the von
Hippel-Lindau (VHL)-mediated degradation pathway. J Biol Chem 277: 41750-41755.
Huang ZJ, Park RE & Rosbash M (1993) PAS is a dimerization domain common to Drosophila
period and several transcription factors. Nature 364: 259-262.
Huss WJ, Hanharan CF, Barrios RJ, Simons JW & Greenberg NM (2001) Angiogenesis and
prostate cancer: identification of a molecular progression switch. Cancer Res 61: 2736-2743.
Ince N, de la Monte SM & Wands JR (2000) Overexpression of human aspartyl(asparaginyl) βhydroxylase is associated with malignant transformation. Cancer Res 60: 1261-1266.
Ivan M, Haberberger T, Gervasi DC, Michelson KS, Günzler V, Kondo K, Yang H, Sorokina I,
Conaway RC, Conaway JW & Kaelin WG Jr (2002) Biochemical purification and
pharmacological inhibition of a mammalian prolyl hydroxylase acting on hypoxia-inducible
factor. Proc Natl Acad Sci USA 99: 13459-13464.
Ivan M, Kondo K, Yang H, Kim W, Valiando J, Ohh M, Salic A, Asara JM, Lane WS & Kaelin
WG Jr (2001) HIFα targeted for VHL-mediated destruction by proline hydroxylation:
implication of O2 sensing. Science 292: 464-468.
Iwai K, Yamanaka K, Kamura T, Minato N, Conaway RC, Conaway JW, Klausner RD & Pause A
(1999) Identification of the von Hippel-lindau tumor-suppressor protein as part of an active E3
ubiquitin ligase complex. Proc Natl Acad Sci USA 96: 12436-12441.
Iyer NV, Kotch LE, Agani F, Leung SW, Laughner E, Wenger RH, Gassmann M, Gearhart JD,
Lawler AM, Yu AY & Semenza GL (1998) Cellular and developmental control of O2
homeostasis by hypoxia-inducible factor 1α. Genes Dev 12: 149-162.
Jaakkola P, Mole DR, Tian Y-M, Wilson MI, Gielbert J, Gaskell SJ, von Kriegsheim A, Hebestreit
HF, Mukherji M, Schofield CJ, Maxwell PH, Pugh CW & Ratcliffe PJ (2001) Targeting of HIFα to the von Hippel-Lindau ubiquitylation complex by O2-regulated prolyl hydroxylation.
Science 292: 468-472.
Jelkmann W (1992) Erythropoietin: structure, control of production, and function. Physiol Rev 72:
449-489.
Jenkins CL & Raines RT (2002) Insights on the conformational stability of collagen. Nat Prod Rep
19: 49-59.
Jeong JW, Bae MK, Ahn MY, Kim SH, Sohn TK, Bae MH, Yoo MA, Song EJ, Lee KJ & Kim KW
(2002) Regulation and destabilization of HIF-1α by ARD1-mediated acetylation. Cell 111: 709720.
89
Jia S, McGinnis K, VanDusen WJ, Burke CJ, Kuo A, Griffin PR, Sardana MK, Elliston KO, Stern
AM & Friedman PA (1994) A fully active catalytic domain of bovine aspartyl (asparaginyl) βhydroxylase expressed in Escherichia coli: characterization and evidence for the identification
of an active-site region in vertebrate α-ketoglutarate-dependent dioxygenases. Proc Natl Acad
Sci USA 91: 7227-7231.
Jiang B-H, Rue E, Wang GL, Roe R & Semenza GL (1996a) Dimerization, DNA binding, and
transactivation properties of hypoxia inducible factor 1. J Biol Chem 271: 17771-17778.
Jiang B-H, Semenza GL, Bauer C & Marti H (1996b) Hypoxia-inducible factor 1 levels vary
exponentially over a physiologically relevant range of O2 tension. Am J Cell Physiol 271: 11721180.
Jiang B-H, Zheng JZ, Leung SW, Roe R & Semenza GL (1997) Transactivation and inhibitory
domains of hypoxia-inducible factor 1α. Modulation of transcriptional activity by oxygen
tension. J Biol Chem 272: 19253-19260.
Jimenez S, Harsch M & Rosenbloom J (1973) Hydroxyproline stabilizes the triple helix of chick
tendon collagen. Biochem Biophys Res Commun 52:106-114.
John DCA & Bulleid NJ (1994) Prolyl 4-hydroxylase: defective assembly of α-subunit mutants
indicates that assembled α-subunits are intramolecularly disulfide bonded. Biochemistry 33:
14018-14025.
John DCA, Grant ME & Bulleid NJ (1993) Cell-free synthesis and assembly of prolyl 4hydroxylase: the role of the β-subunit (PDI) in preventing misfolding and aggregation of the αsubunit. EMBO J 12: 1587-1595.
Johnstone IL (2000) Cuticle collagen genes. Expression in Caenorhabditis elegans.
Trends Genet 16: 21-27.
Kageyama Y, Koshiji M, TO KKW, Tian Y-M, Ratcliffe PJ & Huang LE (2004) Leu-574 of
human HIF-1α is a molecular determinant of prolyl hydroxylation. FASEB J 18: 1028-1030.
Kallio PJ, 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-1α. EMBO J 17: 6573-6586.
Kallio PJ, Wilson WJ, O’Brien S, Makino Y & Poellinger L (1999) Regulation of the hypoxiainducible transcription factor 1α by the ubiquitin-proteasome pathway. J Biol Chem 274: 65196525.
Kamura T, Koepp DM, Conrad MN, Skowyra D, Moreland RJ, Iliopoulos O, Lane WS, Kaelin WG
Jr, Elledge SJ, Conaway RC, Harper JW & Conaway JW (1999) Rbx1, a component of the VHL
tumor suppressor complex and SCF ubiquitin ligase. Science 284: 657-661.
Kanaya K & Kamitani T (2003) pVHL-independent ubiquitination of HIF1α and its stabilization
by cobalt ion. Biochem Biophys Res Commun 306: 750-755.
Kaska DD, Günzler V, Kivirikko KI & Myllylä R (1987) Characterization of a low-relativemolecular-mass prolyl 4-hydroxylase from the green alga Chlamydomonas reinhardii. Biochem
J 241: 483-490.
Kaska DD, Myllylä R, Günzler V, Gibor A & Kivirikko KI (1988) Prolyl 4-hydroxylase from
Volvox carteri. A low-Mr enzyme antigenically related to the α subunit of the vertebrate
enzyme. Biochem J 256: 257-263.
Kedersha NL, Tkacz JS & Berg RA (1985) Biosynthesis of prolyl hydroxylase: evidence for two
separate dolichol-media pathways of glycosylation. Biochemistry 24: 5960-5967.
Kemmink J, Darby NJ, Dijkstra K, Nilges M & Creighton TE (1996) Structure determination of the
N-terminal thioredoxin-like domain of protein disulfide-isomerase using multidimensional
heteronuclear 13C/15N NMR-spectroscopy. Biochemistry 35: 7684-7691.
Kemmink J, Darby NJ, Dijkstra K, Nilges M & Creighton TE (1997) The folding catalyst protein
disulfide isomerase is constructed of active and inactive thioredoxin modules. Curr Biol 7: 239245.
90
Kietzmann T, Roth U & Jungermann K (1999) Induction of the plasminogen activator inhibitor-1
gene expression by mild hypoxia via a hypoxia response element binding the hypoxia-inducible
factor-1 in rat hepatocytes. Blood 94: 4177-4185.
Kivirikko KI & Myllyharju J (1998) Prolyl 4-hydroxylases and their protein disulfide isomerase
subunit. Matrix Biol 16: 357-368.
Kivirikko KI & Myllylä R (1980) Hydroxylation of prolyl and lysyl residues. In: Freedman RB &
Hawkins HC (eds) Enzymology of Post-translational Modification of Proteins, Academic Press,
London, pp 53-104.
Kivirikko KI & Myllylä R (1982) Posttranslational enzymes in the biosynthesis of collagen:
intracellular enzymes. Methods Enzymol 82: 245-304.
Kivirikko KI, Myllylä R & Pihlajaniemi T (1992) Hydroxylation of proline and lysine residues in
collagens and other animal and plant proteins. In: Harding JJ & Grabbe MJC (eds) PostTranslational Modifications of Proteins, CRC Press, Boca Raton, FL, pp 1-51.
Kivirikko KI & Pihlajaniemi T (1998) Collagen hydroxylases and the protein disulfide isomerase
subunit of prolyl 4-hydroxylases. Adv Enzymol Relat Areas Mol Biol 72: 325-398.
Kline DD, Peng Y, Manalo DJ, Semenza GL & Phabhakar NR (2002) Defective carotid body
function and impaired ventilator responses to chronic hypoxia in mice partially deficient for
hypoxia-inducible factor 1α. Proc Natl Acad Sci USA 99: 821-826.
Kondo K, Klco J, Nakamura E, Lechpammer M & Kaelin WG Jr (2002) Inhibition of HIF is
necessary for tumor suppression by the von Hippel-Lindau protein. Cancer Cell 1: 237-246.
Korioth F, Gieffers C & Frey J (1994) Cloning and characterization of the human gene encoding
aspartyl β-hydroxylase. Gene 150: 395-399.
Kotch LE, Iyer NV, Laughner E & Semenza GL (1999) Defective vascularization of HIF-1α-null
embryos is not associated with VEGF deficiency but with mesenchymal cell death. Dev Biol
209: 254-267.
Kukkola L, Hieta R, Kivirikko KI & Myllyharju J (2003) Identification and characterization of a
third human, rat and mouse collagen prolyl 4-hydroxylase isoenzyme. J Biol Chem 278: 4768547693.
Kuznetsova AV, Meller J, Schnell PO, Nash JA, Ignacak ML, Sanchez Y, Conaway JW, Conaway
RC, Czyzyk-Krzeska MF (2003) von Hippel-Lindau protein binds hyperphosphorylated large
subunit of RNA polymerase II through a proline hydroxylation motif and targets it for
ubiquitination. Proc Natl Acad Sci USA 100: 2706-2711.
Ladoux A & Frelin C (1994) Cobalt stimulates the expression of vascular endothelial growth factor
mRNA in rat cardiac cells. Biochem Biphys Res Commun 204: 794-798.
Lamberg A, Pihlajaniemi T & Kivirikko KI (1995) Site-directed mutagenesis of the α subunit of
human prolyl 4-hydroxylase. Identification of three histidine residues critical for catalytic
activity. J Biol Chem 270: 9926-9931.
Lancaster DE, McNeill LA, McDonough MA, Aplin RT, Hewitson KS, Pugh CW, Ratcliffe PJ &
Schofield CJ (2004). Disruption of dimerisation and substrate phosphorylation inhibit factor
inhibiting hypoxia-inducible factor (FIH) activity. Biochem J, in press.
Lando D, Gorman JJ, Whitelaw ML & Peet DJ (2003) Oxygen-dependent regulation of hypoxiainducible factors by prolyl and asparaginyl hydroxylation. Eur J Biochem 270: 781-790.
Lando D, Peet DJ, Gorman JJ, Whelan DA, Whitelaw ML & Bruick RK (2002a) FIH-1 is an
asparaginyl hydroxylase enzyme that regulates the transcriptional activity of hypoxia-inducible
factor. Genes Dev 16: 1466-1471.
Lando D, Peet DJ, Whelan DA, Gorman JJ & Whitelaw ML (2002b) Asparagine hydroxylation of
the HIF transactivation domain: a hypoxic switch. Science 295: 858-861.
Lavaissiere L, Jia S, Nishiyama M, de la Monte S, Stern AM, Wands JR & Friedman PA (1996)
Overexpression of human aspartyl(asparaginyl)β-hydroxylase in hepatocellular carcinoma and
cholangiocarcinoma. J Clin Invest 98: 1313-1323.
91
Lavista-Llanos S, Centanin L, Irisarri M, Russo DM, Gleadle JM, Bocca SN, Muzzopappa M,
Ratcliffe PJ & Wappner P (2002) Control of the hypoxic response in Drosophila melanogaster
by the basic helix-loop-helix PAS protein similar.Mol Cell Biol 22: 6842-6853.
Lee C, Kim SJ, Jeong DG, Lee SM & Ryu SR (2003) Structure of human FIH-1 reveals a unique
active site pocket and interaction sites for HIF-1 and von Hippel-Lindau. J Biol Chem 278:
7558-7563.
Lee PJ, Jiang BH, Chin BY, Iyer NV, Alam J, Semenza GL & Choi AM (1997) Hypoxia-inducible
factor-1 mediates transcriptional activation of the heme oxygenase-1 gene in response to
hypoxia. J Biol Chem 272: 5375-5381.
Lee SH, Wolf PL, Escudero R, Deutsch R, Jamieson SW & Thistlethwaite PA (2000) Early
expression of angiogenesis factors in acute myocardial ischemia and infarction. N Engl J Med
342: 626-633.
Levy AP, Levy NS, Wegner S & Goldberg MA (1995) Transcriptional regulation of the rat
vascular endothelial growth factor gene by hypoxia. J Biol Chem 270: 13333-13340.
Li J, Davidson G, Huang Y, Jiang BH, Shi X, Costa M & Huang C (2004) Nickel compounds act
through phosphatidylinositol-3-kinase/Akt-dependent, p70(S6k)-independent pathway to induce
hypoxia inducible factor transactivation and Cap43 expression in mouse epidermal Cl41 cells.
Cancer Res 64: 94-101.
Lieb ME, Menzies K, Moschella MC, Ni R & Taubman MB (2002) Mammalian EGLN genes have
distinct patterns of mRNA expression and regulation. Biochem Cell Biol 80: 421-426.
Lin Y, Zhang S, Rehn M, Itäranta P, Tuukkanen J, Heljasvaara R, Peltoketo H, Pihlajaniemi T &
Vainio S (2001) Induced repatterning of type XVIII collagen expression in ureter bud from
kidney to lung type: association with sonic hedgehog and ectopic surfactant protein C.
Development 128: 1573-1585.
Linke S, Stojkoski C, Kewley RJ, Booker GW, Whitelaw ML & Peet DJ (2004) Substrate
requirements of the oxygen-sensing asparaginyl hydroxylase factor-inhibitin hypoxia-inducible
factor. J Biol Chem 279: 14391-14397.
Lipscomb EA, Sarmiere PD & Freeman RS (2001) SM-20 is a novel mitochondrial protein that
causes caspase-dependent cell death in nerve growth factor-dependent neurons. J Biol Chem
276: 5085-5092.
Lisztwan J, Imbert G, Wirbeleauer C, Gstaiger M & Krek W (1999) The von Hippel-Lindau tumor
suppressor protein is a component of an E3 ubiquitin protein ligase activity. Genes Dev 13:
1822-1833.
Liu Q, Berchner-Pfannschmidt U, Möller U, Brecht M, Wotzlaw C, Acker H, Jungermann K &
Kietzmann T (2004) A Fenton reaction at the endoplasmic reticulum is involved in the redox
control of hypoxia-inducible gene expression.Proc Natl Acad Sci USA 101: 4302-4307.
Liu Y, Cox SR, Morita T & Kourembanas S (1995) Hypoxia regulates vascular endothelial growth
factor gene expression in endothelial cells. Identification of a 5' enhancer. Circ Res 77: 638-643.
Lok CN & Ponka P (1999) Identification of a hypoxia response element in the transferrin receptor
gene. J Biol Chem 274: 24147-24152.
Lonegran KM, Iliopoulos O, Ohh M, Kamura T, Conaway RC, Conaway JW & Kaelin WG Jr
(1998) Regulation of hypoxia-inducible mRNAs by the von Hippel-Lindau tumor suppressor
protein requires binding to complexes containing elongins B/C and Cul2. Mol Cell Biol 18:
732-741.
Lonser RR, Glenn GM, Walther M, Chew EY, Libutti SK, Linehan WM & Oldfield EH (2003) von
Hippel-Lindau disease. Lancet 361: 2059-2067.
Ma E & Haddad GG (1999). Isolation and characterization of the hypoxia-inducible factor 1beta in
Drosophila melanogaster. Brain Res Mol Brain Res 73: 11-16.
Madden SL, Galella EA, Riley D, Bertelsen AH & Beaudry GA (1996) Induction of cell growth
regulatory genes by p53.Cancer Res 56: 5384-5390.
92
Mahon PC, Hirota K & Semenza GL (2001) FIH-1: a novel protein that interacts with HIF-1α and
VHL to mediate repression of HIF-1 transcriptional activity. Genes Dev 15: 2675-2686.
Majamaa K, Günzler V, Hanauske-Abel HM, Myllylä R & Kivirikko KI (1986) Partial identity of
the 2-oxoglutarate and ascorbate binding sites of prolyl 4-hydroxylase. J Biol Chem 261: 78197823.
Majamaa K, Hanauske-Abel HM, Günzler V & Kivirikko KI (1984) The 2-oxoglutarate binding
site of prolyl 4-hydroxylase. Identification of distinct subsites and evidence for 2-oxoglutarate
decarboxylation in a ligand reaction at the enzyme-bound ferrous ion. Eur J Biochem 138: 239245.
Makino Y, Cao R, Svensson K, Bertilsson G, Åsman 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: 550-554.
Makino Y, Kanopka A, Wilson WJ, Tanaka H & Poellinger L (2002) Inhibitory PAS domain
protein (IPAS) is a hypoxia-inducible splicing variant of the hypoxia-inducible factor-3α locus.
J Biol Chem 277: 32405-32408.
Maranchie JK, Vasselli JR, Riss J, Bonifacino JS, Linehan WM & Klausner RD (2002) The
contribution of VHL substrate binding and HIF-1α to the phenotype of VHL loss in renal cell
carcinoma. Cancer Cell 1: 247-255.
Marxsen JH, Stengel P, Doege K, Heikkinen P, Jokilehto T, Wagner T, Jelkmann W, Jaakkola P &
Metzen E (2004) Hypoxia-inducible factor-1 promotes its degradation by induction of HIF-α prolyl-4-hydroxylases. Biochem J 381: 761-767.
Masson N, Appelhoff RJ, Tuckerman JR, Tian Y-M, Demol H, Puype M, Vandekerckhove J,
Ratcliffe PJ & Pugh CW (2004). FEBS Lett 570: 166-170.
Masson N, Willam C, Maxwll PH, Pugh CW & Ratcliffe PJ (2001) Independent function of two
destruction domains in hypoxia-inducible factor-α chains activated by prolyl hydroxylation.
EMBO J 20: 5197-5206.
Mathupala SP, Rempel A & Pedersen PL (2001) Glucose catabolism in cancer cells: identification
and characterization of a marked activation response of the type II hexokinase gene to hypoxic
conditions. J Biol Chem 276: 43407-43412.
Maxwell PH, Wiesener MS, Chang G-W, Clifford SC, Vaux EC, Cockman ME, Wykoff CC, Pugh
CW, Maher ER & Ratcliffe PJ (1999) The tumour suppressor protein VHL targets hypoxiainducible factors for oxygen-dependent proteolysis. Nature 399: 271-275.
Maynard MA, Qi H, Chung J, Lee EL, Kondo Y, Hara S, Conaway RC, Conaway JW & Ohh M
(2003) Multiple splice variants of the human HIF-3α locus are targets of the von Hippel-Lindau
E3 ubiquitin ligase complex. J Biol Chem 278: 11032-11040.
McGinnis K, Ku GM, VanDusen WJ, Garsky FJ, Stern AM & Friedman PA (1996) Site-directed
mutagenesis of residues in a conserved region of bovine aspartyl (asparaginyl) β-hydroxylase:
evidence that histidine 675 has a role in binding Fe2+. Biochemistry 35: 3957-3967.
McNeill LA, Hewitson KS, Claridge TD, Seibel JF, Horsfall LE & Schofield CJ (2002a) Hypoxiainducible factor asparaginyl hydroxylase (FIH-1) catalyses hydroxylation at the β-carbon of
asparagine-803. Biochem J 367: 571-575.
McNeill LA, Hewitson KS, Gleadle JM, Horsfall LE, Oldham NJ, Maxwell PH, Pugh CW,
Ratcliffe PJ & Schofield CJ (2002b) The use of dioxygen by HIF prolyl hydroxylase (PHD1).
Bioorg Med Chem Lett 12: 1547-1550.
Melillo G, Musso T, Sica A, Taylor LS, Cox GW & Varesio L (1995) A hypoxia-responsive
element mediates a novel pathway of activation of the inducible nitric oxide synthase promoter.
J Exp Med 182: 1683-1693.
Merriweather A, Günzler V, Brenner M & Unnasch TR (2001) Characterization and expression of
enzymatically active recombinant filarial prolyl 4-hydroxylase.Mol Biochem Parasitol 116: 185197.
93
Metzen E, Brechner-Pfannschmidt U, Stengel P, Marxsen JH, Stolze Klinger M, Huang WQ,
Wotzlaw C, Hellwig-Burgel T, Jelkmann W, Acker H & Fandrey J (2003) Intracellular
localisation of human HIF-1α hydroxylases: implications for oxygen sensing. J Cell Sci 116:
1319-1326.
Min J-Y, Yang H, Ivan M, Gertler F, Kaelin WG Jr & Pavletich NP (2002) Structure of an HIF-1αpVHL complex: hydroxyproline recognition in signalling. Science 296: 1886-1889.
Minchenko A & Caro J (2000) Regulation of endothelin-1 gene expression in human microvascular
endothelial cells by hypoxia and cobalt: role of hypoxia responsive element. Mol Cell Biochem
208: 53-62.
Minchenko A, Leshchinsky I, Opentanova I, Sang N, Srinivas V, Armstead V & Caro J (2002)
Hypoxia-inducible factor-1-mediated expression of the 6-phosphofructo-2-kinase/fructose-2,6bisphosphatase-3 (PFKFB3) gene. Its possible role in the Warburg effect. J Biol Chem 277:
6183-6187.
Mole DR, Schlemminger I, McNeill LA, Hewitson KS, Pugh CW, Ratcliffe PJ & Schofield CJ
(2003) 2-oxoglutarate analogue inhibitors of HIF prolyl hydroxylase. Bioorg Med Chem Lett
13: 2677-2680.
Mukhopadhyay CK, Mazumder B & Fox PL (2000) Role of hypoxia-inducible factor-1 in
transcriptional activation of ceruloplasmin by iron deficiency. J Biol Chem. 275: 21048-21054.
Myllyharju J (2003) Prolyl 4-hydroxylases, the key enzymes of collagen biosynthesis. Matrix Biol
22: 15-24.
Myllyharju J & Kivirikko KI (1997) Characterization of the iron- and 2-oxoglutarate binding sites
of the human prolyl 4-hydroxylase. EMBO J 16: 1173-1180.
Myllyharju J & Kivirikko KI (1999) Identification of a novel proline-rich peptide-binding domain
in prolyl 4-hydroxylases. EMBO J 18: 306-312.
Myllyharju J & Kivirikko KI (2001) Collagens and collagen-related diseases. Ann Med 33: 7-21.
Myllyharju J & Kivirikko KI (2004) Collagens, modifying enzymes and their mutations in humans,
flies and worms. Trends Genet 20: 33-43.
Myllyharju J, Kukkola L, Winter AD & Page AP (2002) The exoskeleton collagens in
Caernohabditis elegans are modified by prolyl 4-hydroxylases with unique combinations of
subunits. J Biol Chem 277: 29187-29196.
Myllylä R, Majamaa K, Günzler V, Hanauske-Abel HM & Kivirikko KI (1984) Ascorbate is
consumed stoichiometrically in the uncoupled reactions catalyzed by prolyl 4-hydroxylase and
lysyl hydroxylase. J Biol Chem 259: 5403-5405.
Myllylä R, Tuderman L & Kivirikko KI (1977) Mechanism of the prolyl hydroxylase reaction. 2.
Kinetic analysis of the reaction sequence. Eur J Biochem 80: 349-357.
Nakayama K, Frew IJ, Hagensen M, Skals M, Habelbah H, Bhoumik A, Kadoya T, ErdjumentBromage H, Tempst P, Frappel PB, Bowtell DD & Ronai Z (2004) Siah2 regulates stability of
prolyl-hydroxylases, controls HIF-1α abundance, and modulates physiological responses to
hypoxia. Cell 117: 941-952.
Nguyen SV & Claycomb WC (1999) Hypoxia regulates the expression of the adrenomedullin and
HIF-1 genes in cultured HL-1 cardiomyocytes. Biochem Biophys Res Commun 265: 382-386.
Nissi R, Autio-Harmainen H, Marttila P, Sormunen R & Kivirikko KI (2001) Prolyl 4-hydroxylase
isoenzymes I and II have different expression patterns in several human tissues. J Histochem
Cytochem 49: 1143-1153.
Norris ML & Millhorn DE (1995) Hypoxia-induced protein binding to O2-responsive sequences on
the tyrosine hydroxylase gene. J Biol Chem 270:23774-23779.
Oehme F, Ellinghaus P, Kolkhof P, Smith TJ, Ramakrishnan S, Hütter J, Schramm M & Flamme I
(2002) Overexpression of PH-4, a novel putative proline 4-hydroxylase, modulates activity of
hypoxia-inducible transcription factors. Biochem Biophys Res Commun 296: 343-349.
94
O’Reilly MS, Boehm T, Shing Y, Fukai N, Vasios G, Lane WS, Flynn E, Birkhead JR, Olsen BR
& Folkman J (1997) Endostatin: an endogenous inhibitor of angiogenesis and tumor growth.
Cell 88: 277-285.
O’Rourke JF, Pugh CW, Bartlett SM & Ratcliffe PJ (1996) Identification of hypoxically inducible
mRNAs in HeLa cells using differential-display PCR. Role of hypoxia-inducible factor-1. Eur J
Biochem 241: 403-410.
O’Rourke JF, Tian Y-M, Ratcliffe PJ & Pugh CW (1999) Oxygen-regulated and transactivating
domains in endothelial PAS protein 1: comparison with hypoxia-inducible factor-1α. J Biol
Chem 274: 2060-2071.
Ozaki H, Yu AY, Della N, Ozaki K, Luna JD, Yamada H, Hackett SF, Okamoto N, Zack DJ,
Semenza GL & Campochiaro PA (1999) Hypoxia inducible factor-1α is increased in ischemic
retina: temporal and spatial correlation with VEGF expression. Invest Ophthalmol Vis Sci 40:
182-189.
Palmer LA, Semenza GL, Stoler MH & Johns RA (1998) Hypoxia induces type II NOS gene
expression in pulmonary artery endothelial cells via HIF-1. Am J Physiol 274: L212-L219.
Pekkala M, Hieta R, Kursula P, Kivirikko KI, Wierenga RK & Myllyharju J (2003) Crystallization
of the proline-rich-peptide binding domain of human type I collagen prolyl 4-hydroxylase. Acta
Crystallogr D 59: 940-942.
Peng J, Zhang L, Drysdale L & Fong G (2000) The transcription factor EPAS-1/hypoxia-inducible
factor 2α plays an important role in vascular remodelling. Proc Natl Acad Sci USA 97: 83868391.
Petitclerc E, Boutaud A, Prestayko A, Xu J, Sado Y, Ninomiya Y, Sarras MP Jr, Hudson BG &
Brooks PC (2000) New functions for non-collagenous domains of human collagen type IV.
Novel integrin ligands inhibiting angiogenesis and tumor growth in vivo. J Biol Chem 275:
8051-8061.
Pierce EA, Avery RL, Foley ED, Aiello LP & Smith LEH (1995) Vascular endothelial growth
factor/vascular permeability factor expression in a mouse model of retinal neovascularization.
Proc Natl Acad Sci USA 92: 905-909.
Pihlajaniemi T, Helaakoski T, Tasanen K, Myllylä R, Huhtala M-L, Koivu J & Kivirikko KI (1987)
Molecular cloning of the β-subunit of human prolyl 4-hydroxylase. This subunit and protein
disulphide isomerase are products of the same gene. EMBO J 3: 643-649.
Pirneskoski A, Klappa P, Lobell M, Williamson RA, Byrne L, Alanen HI, Salo KE, Kivirikko KI,
Freedman RB & Ruddock LW (2004) Molecular characterisation of the principal substrate
binding site of the ubiquitous folding catalyst protein disulphide isomerase. J Biol Chem 279:
10374-10381.
Pirneskoski A, Ruddock LW, Klappa P, Freedman RB, Kivirikko KI & Koivunen P (2001)
Domains b' and a' of protein disulfide isomerase fulfill the minimum requirement for function
as a subunit of prolyl 4-hydroxylase. The N-terminal domains a and b enhance this function and
can be substituted in part by those of ERp57. J Biol Chem 276: 11287-11293.
Pongratz I, Antonsson C, Whitelaw ML & Poellinger L (1998) Role of the PAS domain in
regulation of dimerization and DNA binding specificity of the dioxin receptor. Mol Cell Biol
18: 4079-4088.
Post DE & Van Meir EG (2003) A novel hypoxia-inducible factor (HIF) activated oncolytic
adenovirus for cancer therapy. Oncogene 22: 2065-2072.
Prescott AG & John P (1996) Dioxygenases: molecular structure and role in plant metabolism.
Annu Rev Plant Physiol Plant Mol Biol 47: 245-271.
Prockop DJ & Kivirikko KI (1995) Collagens: molecular biology, diseases, and potentials for
therapy. Annu Rev Biochem 64: 403-434.
95
Pugh CW, O’Rourke JF, Nagao M, Gleadle JM & Ratcliffe PJ (1997) Activation of hypoxiainducible factor-1; definition of regulatory domains within the α subunit. J Biol Chem 272:
11205-11214.
Puistola U, Turpeenniemi-Hujanen TM, Myllyla R, Kivirikko KI (1980) Studies on the lysyl
hydroxylase reaction. I. Initial velocity kinetics and related aspects.
Rajakumar A & Conrad KP (2000) Expression, ontogeny, and regulation of hypoxia-inducible
transcription factors in the human placenta. Biol Reprod 63: 559-569.
Ramchandran R, Dhanabal M, Volk R, Waterman MJ, Segal M, Lu H, Knebelmann B & Sukhatme
VP (1999) Antiangiogenic activity of restin, NC10 domain of human collagen XV: comparison
to endostatin. Biochem Biophys Res Commun 255: 735-739.
Rapaka RS, Sorensen KR, Lee SD & Bhatnagar RS (1976) Inhibition of hydroxyl proline synthesis
by palladium ions. Biochim Biophys Acta 429: 63-71.
Rautavuoma K, Takaluoma K, Passoja K, Pirskanen A, Kvist A-P, Kivirikko KI & Myllyharju J
(2002) Characterization of three fragments that constitute the monomers of the human lysyl
hydroxylase isoenzyme 1-3. J Biol Chem 277: 23084-23091.
Ravi R, Mookerjee B, Bhujwalla ZM, Sutter CH, Artemov D, Zeng Q, Dillehay LE, Madan A,
Semenza GL & Bedi A (2000) Regulation of tumor angiogenesis by p53-induced degradation of
hypoxia-inducible factor 1α. Genes Dev 14: 34-44.
Riddle SR, Ahmad A, Ahmad S, Deep SS, Malkki M, Schneider BK, Allen CB & White CW
(2000) Hypoxia induces hexokinase II gene expression in human lung cell line A549. Am J
Physiol Lung Cell Mol Physiol 278: L407-L416.
Riihimaa P, Nissi R, Page AP, Winter AD, Keskiaho K, Kivirikko KI & Myllyharju J (2002) Egg
shell collagen formation in Caenorhabditis elegans involves a novel prolyl 4-hydroxylase
expressed in spermatheca and embryos and possessing many unique properties. J Biol Chem
277: 18238-18243.
Rivard A, Berthou-Soulie L, Principe N, Kearney M, Curry C, Branellec D, Semenza GL & Isner
JM (2000) Age-dependent defect in vascular endothelial growth factor expression is associated
with reduced hypoxia-inducible factor 1 activity. J Biol Chem 275: 29643-29647.
Rolfs A, Kvietikova I, Gassmann M & Wenger RH (1997) Oxygen-regulated transferrin expression
is mediated by hypoxia-inducible factor-1. J Biol Chem 272: 20055-20062.
Rosenbloom J (1987) Molecular cloning and gene structure of elastins. Methods Enzymol 144:
259-288.
Rosenbloom J, Harsch M & Jimenez S (1973) Hydroxyproline content determines the denaturation
temperature of chick tendon collagen. Arch Biochem Biophys 158: 478-484.
Ryan HE, Lo J & Johnson RS (1998) HIF-1α is required for solid tumor formation and embryonic
vascularization. EMBO J 17: 3005-3015.
Ryhänen L (1976) Lysyl hydroxylase. Further purification and characterization of the enzyme from
chick embryos and chick embryo cartilage. Biochim Biophys Acta 438: 71-89.
Salnikow K, Donald SP, Bruick RK, Zhitkovich A, Phang JM & Kasprzak KS (2004) Deplteion of
intracellular ascorbate by the carcinogenic metals nickel and cobalt results in the induction of
hypoxic stress. J Biol Chem, in press.
Salnikow K, Su W, Blagosklonny MV & Costa M (2000) Carcinogenic metals induce hypoxiainducible factor-stimulated transcription by reactive oxygen species-independent mechanism.
Cancer Res 60: 3375-3378.
Salnikow K, Won GA, Melillo G, Blagosklonny MV & Costa M (1999) Nickel-induced
transformation shifts the balance between HIF-1 and p53 transcription factors. Carcinogenesis
20: 1819-1823.
Sasaguri M, Ikeda M, Ideishi M & Arakawa K (1988) Identification of [hydroxyproline3]bradykinin released from human plasma and plasma protein Cohn's fraction IV-4 by trypsin.
Biochem Biophys Res Commun 157: 210-217.
96
Schipani E, Ryan HE, Didrickson S, Kobayashi T, Knight M & Johnson RS (2001) Hypoxia in
cartilage: HIF-1α is essential for chondrocyte growth arrest and survival. Genes Dev 15: 28652876.
Schroedl C, McClintock DS, Budinger GR, Chandel NS (2002) Hypoxic but not anoxic
stabilization of HIF-1α requires mitochondrial reactive oxygen species. Am J Physiol Lung Cell
Mol Physiol 283: 922-931.
Scortegagna M, Ding K, Oktay Y, Gaur A, Thurmond F, Yan LJ, Marck BT, Matsumoto AM,
Shelton JM, Richardson JA, Bennett MJ & Garcia JA (2003) Multiple organ pathology,
metabolic abnormalities and impaired homeostasis of reactive oxygen species in Epas1-/- mice.
Nat Genet 35:331-340.
Semenza GL (2000a) HIF-1 and human disease: one highly involved factor. Genes Dev 14: 19831991.
Semenza GL (2000b) Surviving ischemia: adaptive responses mediated by hypoxia-inducible factor
1. J Clin Invest 106: 809-812.
Semenza GL (2001) HIF-1, O2, and the 3 PHDs: how animal cells signal hypoxia to the nucleus.
Cell 107: 1-3.
Semenza GL (2002) HIF-1 and tumor progression: pathophysiology and therapeutics. Trends Mol
Med 8: S62-67.
Semenza GL, Jiang BH, Leung SW, Passantino R, Concordet JP, Maire P & Giallongo A (1996)
Hypoxia response elements in the aldolase A, enolase 1, and lactate dehydrogenase A gene
promoters contain essential binding sites for hypoxia-inducible factor 1. J Biol Chem 271:
32529-32537.
Semenza GL, Roth PH, Fang HM & Wang GL (1994) Transcriptional regulation of genes encoding
glycolytic enzymes by hypoxia-inducible factor 1. J Biol Chem 269: 23757-23763.
Semenza GL & Wang GL (1992) A nuclear factor induced by hypoxia via de novo protein
synthesis binds to the human erythropoietin gene enhancer at a site required for transcriptional
activity. Mol Cell Biol 12: 5447-5454.
Sepe PS, Lahousse SA, Gemelle B, Chang H, Maeda T, Wands JR, de la Monte SM (2002) Role of
the aspartyl-asparaginyl-β-hydroxylase gene in neuroblastoma cell motility. Lab Invest 82: 881891.
Seth P, Krop I, Porter D & Polyak K (2002) Novel estrogen and tamoxifen induced genes identified
by SAGE (serial analysis of gene expression). Oncogene 21: 836-843.
Shannon AM, Bouchier-Hayes DJ, Condron CM & Toomey D (2003) Tumour hypoxia,
chemotherapeutic resistance and hypoxia-related therapies. Cancer Treat Rev 29: 297-307.
Simpson RJ, Moritz RL, Nice EC, Grego B, Yoshizaki F, Sugimura Y, Freeman HC & Murata M
(1986) Complete amino acid sequence of plastocyanin from a green alga, Enteromorpha
prolifera. Eur J Biochem 157: 497-506.
Sowter HM, Raval R, Moore P, Ratcliffe PJ & Harris AL (2003) Predominant role of hypoxiainducible transcription factor (Hif)-1α versus Hif-2α in regulation of the transcriptional
response to hypoxia. Cancer Res 63: 6130-6134.
Srinivas V, Leshchinsky I, Sang N, King MP, Minchenko A & Caro J (2001) Oxygen sensing and
HIF-1 ativation does not require an active mitochondrial respiratory chain electron-transfer
pathway. J Biol Chem 276: 21995-21998.
Stolze IP, Tian Y-M, Appelhoff RJ, Turley H, Wykoff CC, Gleadle JM & Ratcliffe PJ (2004)
Genetic analysis of the role of the asparaginyl hydroxylase FIH in regulating HIF transcriptional
target genes. J Biol Chem 279: 42719-42725.
Stroka DM, Burkhardt T, Desbaillets I, Wenger RH, Neil DAH, Bauer C, Gassmann M & Candinas
D (2001) HIF-1 is expressed in normoxic tissues and displays an organ-specific regulation
under systemic hypoxia. FASEB J 15: 2445-2453.
97
Tacchini L, Bianchi L, Bernelli-Zazzera A & Cairo G (1999) Transferrin receptor induction by
hypoxia. HIF-1-mediated transcriptional activation and cell-specific post-transcriptional
regulation. J Biol Chem 274: 24142-24146.
Tanimoto K, Makino Y, Pereira T & Poellinger L (2000) Mechanism of regulation of the hypoxiainducible factor-1α by the von Hippel-Lindau tumor suppressor protein. EMBO J 19: 42984309.
Taylor MS (2001) Characterization and comparative analysis of the EGLN gene family. Gene 275:
125-132.
Tazuke SI, Mazure NM, Sugawara J, Carland G, Faessen GH, Suen LF, Irwin JC, Powell DR,
Giaccia AJ & Giudice LC (1998) Hypoxia stimulates insulin-like growth factor binding protein
1 (IGFBP-1) gene expression in HepG2 cells: a possible model for IGFBP-1 expression in fetal
hypoxia. Proc Natl Acad Sci USA 95: 10188-10193.
Tian H, Hammer RE, Matsumoto AM, Russell DW & McKnight SL (1998) The hypoxiaresponsive transcription factor EPAS1 is essential for catecholamine homeostasis and protection
against heart failure during embryonic development. Genes Dev 12: 3320-3324.
Tian H, McKnight SL & Russell DW (1997) Endothelial PAS domain protein 1 (EPAS1), a
transcription factor selectively expressed in endothelial cells. Genes Dev 11: 72-82.
Trent C, Tsuing N & Horvitz HR (1983) Egg-laying defective mutants of the nematode
Caenorhabditis elegans. Genetics 104: 619-647.
Tuderman L, Myllylä R & Kivirikko KI (1977) Mechanism of the prolyl hydroxylase reaction.
Role of co-substrates. Eur J Biochem 80: 341-348.
Turano C, Coppari S, Altieri F & Ferraro A (2002) Proteins of the PDI family: unpredicted non-ER
locations and functions. J Cell Physiol 193: 154-163.
Uchida T, Rossignol F, Matthay MA, Mounier R, Couette S, Clottes E & Clerici C (2004)
Prolonged hypoxia differentially regulates hypoxia-inducible factor (HIF)-1α and HIF-2α
expression in lung epithelial cells. J Biol Chem 279: 14871-14878.
Van Den Diepstraten C, Papay K, Bolender Z, Brown A & Pickering JG (2003) Cloning of a novel
prolyl 4-hydroxylase subunit expressed in the fibrous cap of human atherosclerotic plaque.
Circulation 108: 508-511.
Vaux EC, Metzen E, Yeates KM & Ratcliffe PJ (2001) Regulation of hypoxia-inducible factor is
preserved in the absence of a functioning mitochondrial respiratory chain. Blood 98: 296-302.
Veijola J, Annunen P, Koivunen P, Page AP, Pihlajaniemi T & Kivirikko KI (1996a) Baculovirus
expression of two protein disulphide isomerase isoforms from Caenorhabditis elegans and
characterization of prolyl 4-hydroxylases containing one of these polypeptides as their β
subunit.Biochem J 317: 721-729.
Veijola J, Koivunen P, Annunen P, Pihlajaniemi T & Kivirikko KI (1994) Cloning, baculovirus
expression, and characterization of the α subunit of prolyl 4-hydroxylase from the nematode
Caernohabditis elegans. J Biol Chem 269: 26746-26753.
Veijola J, Pihlajaniemi T & Kivirikko KI (1996b) Co-expression of the α subunit of human prolyl
4-hydroxylase with BiP polypeptide in insect cells leads to the formation of soluble and
insoluble complexes. Soluble α subunit-BiP complexes have no prolyl 4-hydroxylase activity.
Biochem J 315: 613-618.
Vuori K, Pihlajaniemi T, Marttila M & Kivirikko KI (1992a) Characterization of the human prolyl
4-hydroxylase tetramer and its multifunctional protein disulfide-isomerase subunit synthesized
in a baculovirus expression system. Proc Natl Acad Sci USA 89: 7467-7470.
Vuori K, Pihlajaniemi T, Myllylä R & Kivirikko KI (1992b) Site-directed mutagenesis of human
protein disulphide isomerase: effect on the assembly, activity and endoplasmic reticulum
retention of human prolyl 4-hydroxylase in Spodoptera frugiperda insect cells. EMBO J 11:
4213-4217.
Wallace DC (1999) Mitochondrial diseases in man and mouse. Science 283: 1482-1488.
98
Wang GL, Jiang B-H, Rue EA & Semenza GL (1995) Hypoxia-inducible factor 1 is a basic-helixloop-helix-PAS heterodimer regulated by cellular O2 tension. Proc Natl Acad Sci USA 92:
5510-5514.
Wang GL & Semenza GL (1993) Desferrioxamine induces erythropoietin gene expression and
hypoxia-inducible factor 1 DNA-binding activity: implications for models of hypoxia signal
transduction. Blood 82: 3610-3615.
Wang GL & Semenza GL (1995) Purification and characterization of hypoxia-inducible factor 1. J
Biol Chem 270: 1230-1237.
Wang Q, VanDusen WJ, Petroski CJ, Garsky VM, Stern AM & Friedman PA (1991) Bovine liver
aspartyl β-hydroxylase. J Biol Chem 266: 14004-14010.
Wax SD, Rosenfield CL & Taubman MB (1994) Identification of a novel growth factor-responsive
gene in vascular smooth muscle cells. J Biol Chem 269: 13041-13047.
Wenger RH, Rolfs A, Marti HH, Guénet J-L & Gassmann M (1996) Nucleotide sequence,
chromosomal assignment and mRNA expression of mouse hypoxia-inducible factor 1-α.
Biochem Biophys Res Commun 223: 54-59.
Wiener CM, Booth G & Semenza GL (1996) In vivo expression of mRNAs encoding hypoxiainducible factor 1. Biochem Biophys Res Commun 225: 485-488.
Wiesener MS, Jürgensen JS, Rosenberger C, Scholze CK, Horstrup JH, Warnecke C, Mandriota S,
Bechmann I, Frei UA, Pugh CW, Ratcliffe PJ, Bachmann S, Maxwell PH & Eckardt KU (2003)
Widespread hypoxia-inducible expression of HIF-2α in distinct cell populations of different
organs. FASEB J 17: 271-273.
Winter AD, Myllyharju J & Page AP (2003) A hypodermally expressed prolyl 4-hydroxylase from
the filarial nematode Brugia malayi is soluble and active in the absence of protein disulfide
isomerase. J Biol Chem 278: 2554-2562.
Winter AD & Page AP (2000) Prolyl 4-hydroxylase is an essential procollagen-modifying enzyme
required for exoskeleton formation and the maintenance of body shape in the nematode
Caenorhabditis elegans. Mol Cell Biol 20: 4084-4093.
Wojtaszek P, Smith CG & Bolwell GP (1999) Ultrastructural localisation and further biochemical
characterisation of prolyl 4-hydroxylase from Phaseolus vulgaris: comparative analysis. Int J
Biochem Cell Biol 31: 463-477.
Wu M, Moon H-S, Pirskanen A, Myllyharju M, Kivirikko KI & Begley TP (2000) Mechanistic
studies on prolyl-4-hydroxylase: the vitamin C requiring uncoupled oxidation. Bioorg Med
Chem Lett 10: 1511-1514.
Wykoff CC, Beasley NJ, Watson PH, Turner KJ, Pastorek J, Sibtain A, Wilson GD, Turley H,
Talks KL, Maxwell PH, Pugh CW, Ratcliffe PJ & Harris AL (2000) Hypoxia-inducible
expression of tumor-associated carbonic anhydrases. Cancer Res 60: 7075-7083.
Yu AY, Shimoda LA, Iyer NV, Huso DL, Sun X, McWilliams R, Beaty T, Sham JS, Wiener CM,
Sylvester JT & Semenza GL (1999) Impaired physiological responses to chronic hypoxia in
mice partially deficient for hypoxia-inducible factor 1α. J Clin Invest 103: 691-696.
Yu F, White SB, Zhao Q & Lee FS (2001) HIF-1α is regulated by stimulus sensitive proline
hydroxylation. Proc Natl Acad Sci USA 98: 9630-9635.
Yuan Y, Beitner-Johnson D & Millhorn DE (2001) Hypoxia-inducible factor 2α binds to cobalt in
vitro. Biochem Biophys Res Commun 288: 849-854.
Yuan Y, Hilliard G, Ferguson T & Millhorn DE (2003) Cobalt inhibits the interaction between
hypoxia-inducible factor-α and von Hippel-Lindau protein by direct binding to hypoxiainducible factor-α. J Biol Chem 278: 15911-15916.
Zagzag D, Zhong H, Scalzitti JM, Laughner E, Simons JW & Semenza GL (2000) Expression of
hypoxia-inducible factor 1α in brain tumors: association with angiogenesis, invasion, and
progression. Cancer 88: 2606-2618.
99
Zaman K, Ryu H, Hall D, O’Donovan K, Lin KI, Miller MP, Marquis JC, Baraban JM, Semenza
GL & Ratan RR (1999) Protection from oxidative stress-induced apoptosis in cortical neuronal
cultures by iron chelators is associated with enhanced DNA binding of hypoxia-inducible
factor-1 and ATF-1/CREB and increased expression of glycolytic enzymes, p21 (waf1/cip1),
and erythropoietin. J Neurosci 19: 9821-9830.
Zelzer E, Levy Y, Kahana C, Shilo BZ, Rubinstein M & Cohen B (1998) Insulin induces
transcription of target genes through the hypoxia-inducible factor HIF-1α/ARNT. EMBO J 17:
5085-5094.
Zhong H, Chiles K, Feldser D, Laughner E, Hanrahan C, Georgescu MM, Simons JW & Semenza
GL (2000) Modulation of hypoxia-inducible factor 1α expression by the epidermal growth
factor/phosphatidylinositol 3-kinase/PTEN/AKT/FRAP pathway in human prostate cancer cells:
implications for tumor angiogenesis and therapeutics. Cancer Res 60: 1541-1545.
Zhong H, De Marzo AM, Laughner E, Lim M, Hilton DA, Zagzag D, Buechler P, Isaacs WB,
Semenza GL & Simons JW (1999) Overexpression of hypoxia-inducible factor 1α in common
human cancers and their metastases. Cancer Res 59: 5830-5835.