Lysyl hydroxylases 1 and 2 : Characterization of their in vivo ... - Oulun

ACTA
Marjo Hyry
OULU 2012
D 1159
UNIVERSITATIS OULUENSIS
D
LYSYL HYDROXYLASES 1 AND 2
CHARACTERIZATION OF THEIR IN VIVO ROLES
IN MOUSE AND THE MOLECULAR LEVEL
CONSEQUENCES OF THE LYSYL HYDROXYLASE 2
MUTATIONS FOUND IN BRUCK SYNDROME
UNIVERSITY OF OULU GRADUATE SCHOOL;
UNIVERSITY OF OULU, FACULTY OF MEDICINE,
INSTITUTE OF BIOMEDICINE,
DEPARTMENT OF MEDICAL BIOCHEMISTRY AND MOLECULAR BIOLOGY;
BIOCENTER OULU;
CENTER FOR CELL-MATRIX RESEARCH
MEDICA

ACTA UNIVERSITATIS OULUENSIS
D Medica 1159
MARJO HYRY
LYSYL HYDROXYLASES 1 AND 2
Characterization of their in vivo roles in mouse and the
molecular level consequences of the lysyl hydroxylase 2
mutations found in Bruck syndrome
Academic dissertation to be presented with the assent
of the Doctoral Training Committee of Health and
Biosciences of the University of Oulu for public defence
in Auditorium F101 of the Department of Physiology
(Aapistie 7), on 8 June 2012, at 10 a.m.
UNIVERSITY OF OULU, OULU 2012

Copyright © 2012
Acta Univ. Oul. D 1159, 2012
Supervised by
Professor Johanna Myllyharju
Reviewed by
Professor Heather Yeowell
Professor Helen Skaer
ISBN 978-951-42-9841-7 (Paperback)
ISBN 978-951-42-9842-4 (PDF)
ISSN 0355-3221 (Printed)
ISSN 1796-2234 (Online)
Cover Design
Raimo Ahonen
JUVENES PRINT
TAMPERE 2012

Hyry, Marjo, Lysyl hydroxylases 1 and 2. Characterization of their in vivo roles in
mouse and the molecular level consequences of the lysyl hydroxylase 2 mutations
found in Bruck syndrome
University of Oulu Graduate School; University of Oulu, Faculty of Medicine, Institute of
Biomedicine, Department of Medical Biochemistry and Molecular Biology; Biocenter Oulu;
Center for Cell-Matrix Research, P.O. Box 5000, FI-90014 Oulu, Finland
Acta Univ. Oul. D 1159, 2012
Oulu, Finland
Abstract
The extracellular matrix is not just a scaffold for cells and tissues, but rather a dynamic part of the
human body. Characteristics of collagens, the major protein components of the extracellular
matrix, are determined already during synthesis and mutations in genes encoding collagens,
unbalance of regulators or dysfunction of collagen modifying enzymes, for instance, can lead to
severe clinical complications. Certain hydroxylysine residues formed by lysyl hydroxylases (LHs)
function in collagens as precursors of collagen cross-links that stabilize collagenous structures and
thereby tissues. In humans, a deficiency of LH1, which is known to hydroxylate lysines in the
helical regions of collagen polypeptides, causes Ehlers-Danlos syndrome VIA (EDS VIA). It is
characterized e.g. by progressive kyphoscoliosis and hypermobile joints. Mutations in LH2, which
is known to hydroxylate lysines in the telopeptides of collagen polypeptides, cause Bruck
syndrome type 2 (BS2). BS2 patients suffer from fragile bones and congenital joint contractures,
for instance, but the syndrome is usually not lethal.
In this work we have generated and analyzed genetically modified LH1 and LH2 null mouse
lines to study the in vivo functions and roles of these enzymes. Analyses concentrated also on
collagen cross-links that were determined from several null or heterozygous mouse tissues. In the
present work we also studied the effects of known BS2 mutations on recombinant human LH2
polypeptides to understand the molecular pathology of the syndrome.
As an animal model for human EDS VIA, LH1 null mice had certain characteristics typical for
EDS VIA, such as muscular hypotonia, but generally the symptoms were milder. Like EDS VIA
patients, the mice have an increased risk of arterial ruptures and ultrastructural changes can be seen
in the wall of the aorta, explained by inadequate helical lysine hydroxylation accompanied by a
changed cross-linking state of tissues. Similarly, analysis of the LH2 null mouse line demonstrated
the importance of the enzyme in cross-link formation. We showed that even a reduced amount of
LH2 in adult mice changes the cross-linking pattern in tissues and a total lack of the enzyme leads
to embryonic lethality. Furthermore, we demonstrated that LH2 is particularly important in tissue
structures supporting blood vessels in the developing mouse embryo or in extraembryonic tissues.
Finally, our in vitro studies with recombinant human LH2 polypeptides revealed that the known
BS2 mutations severely affect the activity of the enzyme thus explaining the clinical symptoms of
the patients, but the mutations do not lead to a total inactivation of the enzyme, which may be
critical for the survival of patients.
Keywords: Bruck syndrome, collagen, collagen cross-link, connective tissue disorder,
Ehlers-Danlos syndrome type VIA, lysyl hydroxylase

Hyry, Marjo, Lysyylihydroksylaasit 1 ja 2.
Oulun yliopiston tutkijakoulu; Oulun yliopisto, Lääketieteellinen tiedekunta, Biolääketieteen
laitos, Lääketieteellinen biokemia ja molekyylibiologia; Biocenter Oulu; Center for Cell-Matrix
Research, PL 5000, 90014 Oulu
Acta Univ. Oul. D 1159, 2012
Oulu
Tiivistelmä
Solunulkoinen matriksi ei ole ainoastaan soluja ja kudoksia tukeva rakenne, vaan se on dynaaminen
osa ihmiskehoa. Kollageenien, solunulkoisen matriksin yleisimpien proteiinien ominaisuudet
määräytyvät jo kollageenien synteesivaiheessa ja mutaatiot kollageeneja koodittavissa
geeneissä, säätelytekijöiden epätasapaino tai esimerkiksi kollageeneja muokkaavien entsyymien
toimintahäiriöt voivat johtaa vaikeisiin kliinisiin komplikaatioihin. Tietyt lysyylihydroksylaasien
(LH) muodostamat hydroksilysiinitähteet toimivat kollageeneissa kollageeniristisidosten esiasteina.
Ristisidokset vakauttavat kollageenirakenteita ja siten myös kudoksia. LH1 hydroksyloi
lysiinejä kollageenipolypeptidien kolmoiskierteisellä alueella ja ihmisellä entsyymin puutos
aiheuttaa tyypin VIA Ehlers-Danlosin syndrooman (EDS VIA), jossa potilailla on esimerkiksi
etenevää kyfoskolioosia ja yliliikkuvat nivelet. Mutaatiot LH2-entsyymissä, joka hydroksyloi
lysiinejä kollageenipolypeptidien telopeptidialueilla, aiheuttavat tyypin 2 Bruckin syndrooman
(BS2). BS2-potilaat kärsivät mm. luiden hauraudesta ja niveljäykkyydestä, mutta syndrooma ei
yleensä ole letaali.
Tässä työssä loimme ja analysoimme geneettisesti muunnellut LH1 ja LH2 hiirilinjat, joiden
kyseinen LH-geeniaktiivisuus on hiljennetty. Linjojen avulla halusimme tutkia näiden entsyymien
toimintaa ja merkitystä in vivo. Analyysit keskittyivät myös kollageeniristisidoksiin, joita tutkittiin
useista poistogeenisten tai heterotsygoottisten hiirten kudoksista. Ymmärtääksemme
BS2:n molekyylipatologiaa, tutkimme tässä työssä myös tunnettujen BS2-mutaatioiden vaikutuksia
ihmisen LH2-rekombinanttiproteiinissa.
EDS VIA:n eläinmallina LH1 poistogeenisillä hiirillä on joitakin ominaisuuksia, kuten lihashypotonia,
jotka ovat tyypillisiä EDS VIA:lle, mutta yleisesti oireet ovat lievempiä. Kuten EDS
VIA-potilailla, hiirillä on kohonnut valtimoiden repeytymisriski ja aortan seinämän ultrarakenteessa
voidaankin havaita muutoksia. Oireita voidaan selittää riittämättömällä kollageenien kolmoiskierteisen
alueen lysiinien hydroksylaatiolla, joka muuttaa kollageenien ristisidostilaa
kudoksissa. Myös LH2-hiirilinjan analysointi osoitti kyseisen entsyymin tärkeyden ristisidosten
muodostamisessa. Jo alentunut LH2:n määrä aikuisissa hiirissä muuttaa kudosten kollageeniristisidoksia
ja täydellinen entsyymin puuttuminen johtaa sikiön kuolemaan. Lisäksi osoitimme,
että LH2 on erityisen tärkeä kudosrakenteissa, jotka tukevat kehittyvän hiiren sikiön tai sikiön
ulkopuolisten kudosten verisuonia. In vitro-tutkimukset ihmisen LH2-rekombinanttiproteiinilla
paljastivat, että tunnetut BS2-mutaatiot vaikuttavat erittäin haitallisesti entsyymin toimintaan,
mikä selittää potilaiden kliiniset oireet, mutta mutaatiot eivät kuitenkaan aiheuta entsyymin täydellistä
inaktivaatiota, mikä voi olla kriittistä potilaiden selviytymisen kannalta.
Asiasanat: Bruckin syndrooma, kollageeni, kollageeniristisidos, lysyylihydroksylaasi,
sidekudossairaus, tyypin VIA Ehlers-Danlosin syndrooma

Acknowledgements
This research was carried out at the Institute of Biomedicine, Department of
Medical Biochemistry and Molecular Biology, University of Oulu, during years
2005–2012.
I wish to express my deepest gratitude to my excellent supervisor, Professor
Johanna Myllyharju, who gave me the opportunity to start my scientific career
and prepare my thesis in her group. I have been privileged to have her as a
supervisor as she has provided all the guidance, support, trust and encouragement
that the project required. Academy Professor Emeritus Kari Kivirikko deserves
my most sincere thanks for his amazing attitude towards science and for his
endless optimism, which truly gives inspiration and basis for the scientific career
as well as life itself. I want to thank warmly Docent Raija Soininen for her
enthusiasm, trust, encouragement, and all the effort in the mouse projects.
I am grateful to Professor Emeritus Heather Yeowell and Professor Helen
Skaer for their careful revision of this thesis and valuable comments on it. M.Sc.
Sandra Hänninen is acknowledged for her expert revision of the language.
We are not able to do science on our own. I want to express my respect and
thanks to other key persons, Professor Emeritus Ilmo Hassinen, Professor Taina
Pihlajaniemi, Professor Seppo Vainio, Professor Leena Ala-Kokko, Docent Minna
Männikkö, Docent Peppi Karppinen, Dr. Aki Manninen and Dr. Lauri Eklund, of
the Department of Medical Biochemistry and Molecular Biology for providing
excellent research facilities and atmosphere to make science.
I want to thank all my co-authors for the pleasant collaboration and their
essential contribution to this work. I want to express my deepest thanks to Dr.
Kati Takaluoma for giving me hands-on introduction to the world of collagens
and science itself. I also want to thank her for supervising my M.Sc. project and
for enjoyable moments we have had together. M.D. Juha Lantto and M.D. Hanna-
Leena Talsta are thanked for their contribution to these projects and nice moments
in the office. I wish to express my gratitude to Professor Ruud Bank for his help
with cross-links and Docent Raija Sormunen and Dr. Ilkka Miinalainen for their
expertise and help with the electron microscope.
This work would have not been possible without the outstanding technical
assistance of Liisa Äijälä during all these years starting from lab basics to a
novice. I also want to thank the whole staff at the department for creating the
great atmosphere to work and for their given assistance along the way.
Particularly Auli Kinnunen, Pertti Vuokila, Risto Helminen and Seppo
7

Lähdesmäki deserve my warmest acknowledgements for their professional and
kind help. I am also really grateful for the people in the Biocenter Oulu
Transgenic core facility and the Laboratory Animal Center for the excellent
collaboration and good care of our mice.
All former and present members of our research group are heartily thanked
for being such wonderful colleagues. It has been a privilege to work with you and
share all those memorable sunny and cloudy days during these years. You are the
ones who have made our group so inspiring, supportive and warm. I couldn’t ask
for more. I also wish to thank Johanna Korvala for sharing the unforgettable time
preparing the theses and organizing the dissertation party.
All of this would not be possible without the sincere and endless support of
family and friends. I sincerely want you to know that you have always brought
out the best in me. It’s not easy to find words to thank all of my friends who have
been with me during all the good days but also during those days when I have
needed you the most. Every day I wonder with smile on my face how lucky I am
to have all of you in my life. And finally, I owe my deepest gratitude to my family,
my parents and sisters and their families. Without your unconditional love,
support and encouragement throughout all my life none of this would be possible.
I love you.
This work was supported by the Biocenter Oulu and the Emil Aaltonen
Foundation.
Oulu, April 2012 Marjo Hyry
8

Abbreviations
AGE advanced glycation end-product
bp base pair
BS Bruck syndrome
C carboxy
CD circular dichroism
cDNA complementary DNA
C-P4H collagen prolyl 4-hydroxylase
CRTAP cartilage-associated protein
Da Dalton(s)
deH-HLNL dehydro-hydroxylysinonorleucine
deH-LNL dehydro-lysinonorleucine
E embryonic day
ECM extracellular matrix
EDS Ehlers-Danlos syndrome
ER endoplasmic reticulum
FACIT fibril-associated collagens with interrupted triple helices
FKBP10 FK506 binding protein 10
GGT hydroxylysyl glucosyltransferase
GT hydroxylysyl galactosyltransferase
HIF hypoxia-inducible factor
HLKNL hydroxylysino-5-ketonorleucine
HP hydroxylysyl pyridinoline
HPLC reverse-phase high-pressure liquid chromatography
Hyl hydroxylysine
JMJD6 Jumonji domain-6 protein
Km
Michaelis-Menten constant
LH lysyl hydroxylase
LKNL lysino-5-ketonorleucine
LO lysyl oxidase
LP lysyl pyridinoline
MMP matrix metalloproteinase
mRNA messenger RNA
N amino
NC noncollagenous
OI osteogenesis imperfecta
9

P3H prolyl 3-hydroxylase
PCR polymerase chain reaction
PDI protein disulfide isomerase
PLOD procollagen-lysine 2-oxoglutarate 5-dioxygenase; human gene
Plod procollagen-lysine 2-oxoglutarate 5-dioxygenase; mouse gene
Q-PCR real-time quantitative PCR
RM Reichert’s membrane
TEM transmission electron microscopy
TGF-β transforming growth factor-β
TIA T-cell-restricted intracellular antigen
X, in -Gly-X-Y- any amino acid
Y, in -Gly-X-Y- any amino acid
10

List of original articles
This thesis is based on the following articles, which are referred to in the text by
their Roman numerals:
I Takaluoma K, Hyry M, Lantto J, Sormunen R, Bank RA, Kivirikko KI, Myllyharju J
& Soininen R (2007) Tissue-specific changes in the hydroxylysine content and crosslinks
of collagens and alterations in fibril morphology in lysyl hydroxylase 1 knockout
mice. J Biol Chem 282(9): 6588–6596.
II Hyry M, Lantto J & Myllyharju J (2009) Missense mutations that cause Bruck
syndrome affect enzymatic activity, folding, and oligomerization of lysyl hydroxylase
2. J Biol Chem 284(45): 30917–30924.
III Hyry M, Soininen R, Bank RA, Sormunen R, Miinalainen I, Talsta HL, Lantto J &
Myllyharju J (2012) Lack of lysyl hydroxylase 2 in mouse leads to early lethality
during embryonic development. Manuscript.
11

Contents
Abstract
Tiivistelmä
Acknowledgements 7
Acknowledgements 7
Abbreviations 9
List of original articles 11
Contents 13
1 Introduction 17
2 Review of the literature 19
2.1 Extracellular matrix ................................................................................. 19
2.1.1 The molecular multiplicity of the extracellular matrix and
its connection to disorders ............................................................ 19
2.2 Collagens ................................................................................................. 20
2.2.1 The collagen family ...................................................................... 21
2.2.2 Collagen biosynthesis ................................................................... 24
2.3 Collagen hydroxylysines and their function ............................................ 28
2.3.1 Glycosylation of hydroxylysines .................................................. 29
2.3.2 Collagen cross-links ..................................................................... 31
2.4 Lysyl hydroxylases .................................................................................. 36
2.4.1 Lysyl hydroxylase isoenzymes ..................................................... 37
2.4.2 Molecular and catalytic properties ............................................... 42
2.5 Human diseases related to lysyl hydroxylases and animal models ......... 46
2.5.1 Ehlers-Danlos type VI .................................................................. 46
2.5.2 Bruck syndrome 1 and 2 ............................................................... 49
2.5.3 Fibrotic diseases and cancer ......................................................... 53
2.5.4 A disease caused by mutations in LH3 and LH3 null mice .......... 55
2.5.5 Inactivation of LH in other animal models ................................... 57
3 Outlines of the present study 59
4 Materials and methods 61
5 Results 63
5.1 LH1 knock-out mice have increased risk of lethal arterial
ruptures (I) .............................................................................................. 63
5.1.1 Generation of LH1 null mice and LH1 expression in adult
mouse tissues ................................................................................ 63
13

5.1.2 LH1 null mice have no kyphoscoliosis but are hypotonic
and have increased mortality due to arterial ruptures ................... 64
5.1.3 Abnormal collagen fibrils in LH1 null mice ................................. 64
5.1.4 Decreased LH activity in LH1 null mice affects the
hydroxylysine content in various tissues differently .................... 65
5.1.5 Abnormal hydroxylation of triple helical lysine residues
affects the cross-linking pattern in the LH1 null mouse
tissues ........................................................................................... 66
5.2 Analysis of the molecular level consequences of Bruck
syndrome mutations by the use of recombinant LH2 proteins (II) .......... 67
5.2.1 Expression of wild-type and recombinant LH2(long)
polypeptides with Bruck syndrome mutations in insect
cells and their purification ............................................................ 67
5.2.2 The Bruck syndrome mutant R594H and G597V
LH2(long) polypeptides have problems in dimerization
and are prone to degradation ........................................................ 68
5.2.3 Folding of the R594H and G597V mutant LH2(long)
polypeptides is impaired and their proteolytic sensitivity is
increased ....................................................................................... 69
5.2.4 All Bruck syndrome mutations affect the activity of LH2 ........... 70
5.3 Lysyl hydroxylase 2 is essential for the embryonic development
of mouse (III) .......................................................................................... 71
5.3.1 LH2 null mice die after embryonic day 10.5 ................................ 71
5.3.2 LH2 is expressed ubiquitously in the E10.5 embryos and
in extraembryonic tissues and its lack seems to lead to
certain changes in collagen expression ......................................... 72
5.3.3 Expression of the Plod1 and Plod3 genes is not
upregulated in the Plod2 gt/gt embryos ........................................... 73
5.3.4 LH2 is widely expressed in adult Plod2 wt/gt tissues ...................... 74
5.3.5 Changes in lysyl hydroxylation and cross-linking of
collagen in adult Plod2 wt/gt mice ................................................... 74
6 Discussion 77
6.1 LH1 null mice provide a tool to study tissue-specific
consequences of the lack of LH1 activity ............................................... 77
6.2 Mutations causing Bruck syndrome lead to decreased LH2
activity, explaining the symptoms of diagnosed Bruck syndrome
type 2 patients ......................................................................................... 81
14

6.3 Inactivation of mouse Plod2 leads to lethality during early
embryonic development .......................................................................... 83
7 Conclusions and future prospects 89
References 91
Original articles 111
15

1 Introduction
Collagens are the most abundant proteins in the human body and comprise the
major structural backbone in the extracellular matrix (ECM) of tissues. Several
collagen-specific enzymes are required in collagen biosynthesis, including lysyl
hydroxylases (LHs). Hence, it is natural that changes in collagens or enzymes
involved in collagen synthesis can lead to severe consequences. Collagen-related
changes in the ECM cause many common conditions such as fibrosis, but they
also cause rare disorders because of inherited mutations in ECM genes, such as
osteogenesis imperfecta (OI) or the rarer Bruck syndrome (BS).
Humans and mice have three different LH isoforms i.e. LH1, LH2 and LH3.
Mutations in all three LH isoforms are known to cause severe connective tissue
disorders in humans demonstrating their importance. LH1 mutations have long
been known to cause the Ehlers-Danlos syndrome VIA (EDS VIA), which is
characterized e.g. by progressive kyphoscoliosis and hypermobile joints. The
discovery of LH2 was soon accompanied by identification of mutations in LH2
that cause the BS type 2. BS2 patients suffer from symptoms like fragile bones
and congenital joint contractures. Recently, LH3 mutations have also been
connected to two human disorders, and some years earlier a mouse LH3 knockout
was observed to be embryonically lethal, suggesting the importance of LHs
also in mice.
Hydroxylysine residues formed by LHs function in collagens as attachment
sites for carbohydrate residues and certain residues participate in collagen crosslinks.
Moreover, the hydroxylation state of the lysines at the cross-link sites
determines the cross-linking pattern and thus partly the mechanical properties of
tissues. Although in vitro studies have revealed much information about LHs, the
knowledge of in vivo functions of LH1 and LH2 has been restricted mostly to
analyses of human patients. Our aim in this work was to generate LH1 and LH2
null mouse lines to obtain more detailed information of the in vivo functions of
these enzymes. The mouse lines were also expected to provide more information
about the EDS VIA and BS2 syndroms at the molecular level and about the
intricate cross-link formation. In addition, we studied the effects of the known
BS2 mutations on recombinant human LH2 polypeptides to understand the
molecular pathology of the syndrome. Our results show that the LH1 null mouse
line is valuable in the analysis of the in vivo function of LH1, as well as allowing
better understanding of the pathology of EDS VIA. Surprisingly, our work
demonstrates that LH2 is essential for the embryonic development of the mouse.
17

Our data also show that the mutations found in human BS2 patients decrease the
enzyme activity in vitro by two different mechanisms, but do not lead to complete
LH2 inactivation, which is likely to explain the survival of the BS2 patients. Both
mouse lines also provide novel information on the roles of LH1 and LH2 in
determining the collagen cross-linking pattern in tissues.
18

2 Review of the literature
2.1 Extracellular matrix
The extracellular matrix (ECM) is a complex environment that surrounds cells
and enables the multicellularity of organisms. ECM consists of molecules that are
secreted by the adjacent cells and thus the environment is tailored for the
requirements of the cells and the tissue. ECM functions as a scaffold providing
strength and support against physical stress, but it also has an essential role in cell
behavior. It interacts with signaling molecules and growth factors, which regulate
cellular events such as determination, differentiation, proliferation, migration and
survival. The ECM is highly dynamic and it undergoes constant remodeling
during life and particularly in many disease states. Failures in the renewal process
underlie many pathological states like fibrosis, atherosclerosis and tumorigenesis.
(For reviews, see Kielty & Grant 2002, Daley et al. 2008, Butcher et al. 2009,
Tsang et al. 2010).
2.1.1 The molecular multiplicity of the extracellular matrix and its
connection to disorders
The main components of the ECM are collagens and noncollagenous proteins
such as proteoglycans, glycoproteins and elastins, but the different mechanical
requirements of tissues determine the ratios of these proteins. Thus, the molecular
diversity in the ECM is overwhelming and these structurally very different
molecules are themselves responsible for the dynamic nature of the ECM. Many
ECM proteins are large and complex consisting of several individually folded
domains that are highly conserved between species. The synthesis, organization
and interplay of the ECM components are strict. The synthesis is controlled by
growth factors such as transforming growth factor-β (TGF-β) and the degradation
by proteases like cysteine and serine proteases and matrix metalloproteinases
(MMPs). It is clear that the activity of these molecules must be under precise
control and misregulation can easily lead to severe consequences. (For reviews,
see Daley et al. 2008, Badylak et al. 2009, Butcher et al. 2009, Hynes 2009,
Järveläinen et al. 2009, Tsang et al. 2010). Recently it was suggested that in
addition to signaling molecules and growth factors, basic features such as the
19

stiffness of the ECM also affect differentiation and cell behavior (Engler et al.
2006, Butcher et al. 2009).
One minimal inherited or acquired change in one ECM component can lead
to physicochemical and thus to mechanical and behavioral changes in the tissue or
developing individual followed by a possible disorder or even lethality
(Järveläinen et al. 2009, Levental et al. 2009). The rare Ehlers-Danlos syndrome
(EDS) type VIA is a severe connective tissue disorder and an example of an
inherited monogenic disorder. It is caused by mutations in one of the collagenmodifying
enzymes, lysyl hydroxylase 1 (LH1, see 2.2.2 and 2.4). It is
characterized by progressive kyphoscoliosis, hypermobile joints and muscle
hypotonia, and patients have an increased risk of arterial ruptures. (For reviews,
see Yeowell & Walker 2000, Steinmann et al. 2002). ECM changes are distinct
also in common diseases that can be acquired or caused by a combination of
acquired and genetic factors, not forgetting the interactions between the ECM and
growth factors and other signaling molecules. In coronary heart disease, for
instance, the ECM components participate in the formation of the atherosclerotic
plaque. Changes in the ECM composition have been connected also to
carcinogenesis and metastasis of the tumor cells. On the other hand, probably all
diseases affect the quality and/or quantity of the ECM. Thus, exact knowledge
about individual ECM molecules, their function and interplay with growth factors
and signaling molecules is beneficial not only for the drug development and
treatment of ECM disorders, but it can be utilized also in the treatment of a wide
variety of other diseases. A deficiency in structural ECM proteins is challenging
to treat, but the regulators provide a potential tool when their function is known.
(Hynes 2009, Järveläinen et al. 2009, Levental et al. 2009, Van Agtmael &
Bruckner-Tuderman 2010).
2.2 Collagens
Collagens are the major insoluble fibrous proteins in the ECM and the most
abundant proteins in the human body. Thus, collagens have a remarkable role in
the maintenance of various structures as well as in other essential ECM activities.
Collagens are involved in many disorders with numerous mutations in collagen
genes behind them. Collagen biosynthesis requires specific enzymes and
mutations in these enzymes increase the number of collagen-related disorders.
(For reviews, see Myllyharju & Kivirikko 2001, Myllyharju & Kivirikko 2004,
20

Gelse et al. 2003, Van Agtmael & Bruckner-Tuderman 2010, Marini et al. 2010,
Parkin et al. 2011).
2.2.1 The collagen family
Humans have at least 28 different collagen types, which consist of over 40
different α chains (Myllyharju & Kivirikko 2004, Gordon & Hahn 2010). The
heterogeneity of the collagen family members is also increased by alternative
splicing of the primary transcripts coding for collagen polypeptides (Myllyharju
& Kivirikko 2001, McAlinden et al. 2005, Seppinen & Pihlajaniemi 2011). All
collagen molecules consist of three identical or different polypeptides called α
chains, which are wrapped around each other to form a triple helix. The triple
helical structure forms as a result of the repeating Gly-X-Y triplet, which
represents a collagenous domain. As the smallest amino acid, the glycine residue
is positioned in the central axis of the triple helix and it is the only amino acid that
fits into that restricted space. In fact, the most harmful mutations in the collagen
genes are often those that replace the glycine residue in the triple helical domain.
In the collagenous (Gly-X-Y)n sequence, the X and Y can be any amino acid
except glycine, but often proline is in the X position and 4-hydroxyproline in the
Y position. (Brodsky & Persikov 2005, Shoulders & Raines 2009, Gordon &
Hahn 2010).
All collagen molecules have at least one collagenous domain in their structure,
but the amount and lengths of these parts vary between collagen types. Over 90%
of type I collagen for instance is collagenous, while collagen XII contains only
10% of triple helix. Besides collagenous domains, collagens contain also
noncollagenous (NC) parts which often have important functions. These domains
may be involved in the binding of molecules or in the assembly of supramolecular
structures, or they may have their own functions when released from the collagen,
an example being endostatin. Endostatin is cleaved from the C-terminal NC1
domain of type XVIII collagen and is able to inhibit angiogenesis and tumor
growth. Similar domains called restin and arresten are also present in collagens
XV and IV, respectively. (Myllyharju & Kivirikko 2001, Gelse et al. 2003,
Ricard-Blum & Ruggiero 2005, Ricard-Blum 2011). Collagenous domains are not
unique to collagens. There are several proteins such as adiponectin, collectin,
ficolin and macrophage receptor that contain a collagen-like domain in their
structure, but they are not called collagens. (Myllyharju & Kivirikko 2004,
Ricard-Blum 2011).
21

Collagens can be divided into fibrillar and non-fibrillar collagens and further
into several subfamilies (see Fig. 1). The collagen superfamily consists of the
fibril-forming collagens such as type I, II, III and V collagens, fibril-associated
collagens with interrupted triple helices (FACITs) such as type IX and XII
collagens, collagens that form hexagonal networks (type VIII and X collagens),
network-forming collagens (type IV collagens), collagens that form beaded
filaments (type VI collagen), collagens that form anchoring fibrils (type VII
collagen), endostatin-producing collagens (type XV and XVIII collagens) and
transmembrane collagens such as collagens XIII and XXV. (For reviews, see
Myllyharju & Kivirikko 2004, Kadler et al. 2007, Gordon & Hahn 2010). Type
XXVIII collagen is the most recently identified collagen in the superfamily. It
cannot be placed clearly into any collagen subfamily mentioned above as it has
properties typical of more than one of those groups. (Veit et al. 2006). Type XXIX
collagen, reported in 2007, is likely a member of the collagen VI family rather
than a new collagen type (Söderhäll et al. 2007, Fitzgerald et al. 2008).
22

a. Fibril-forming collagens, types I, II, III, V, XI, XXIV and XXVII
Genes: COL1A1, COL1A2; COL2A1; COL3A1; COL5A1, COL5A2, COL5A3,
COL5A4; COL11A1, COL11A2; COL24A1; COL27A1
Triple helical region
N pro-
C propeptidepeptide
300 nm
100 nm
c. Collagens forming hexagonal networks, types VIII and X
Genes: COL8A1, COL8A2; COL10A1
VIII
100 nm
Fig. 1. Members of the collagen superfamily and their known supramolecular
assemblies. Reprinted from Myllyharju & Kivirikko 2004 with permission of the
publisher.
X
Fibrillar collagens
100 nm
d. The family of type IV collagens
e. Type VI collagen forming beaded filaments
Genes: COL4A1, COL4A2, COL4A3, COL4A4, COL4A5, COL4A6 Genes: COL6A1, COL6A2, COL6A3
7S
Dimer
VI
Dimer Tetramer Beaded filament
100 nm
100 nm
Dimer
200 nm
Anchoring fibril
Basement
membrane
7S
Tetramer
200 nm
Anchoring
plaque
b. FACIT and related collagens, types IX, XII,
XIV, XVI, XIX, XX, XXI, XXII and XXVI
Genes: COL9A1, COL9A2, COL9A3; COL12A1; COL14A1;
COL16A1; COL19A1; COL20A1; COL21A1; COL22A1;
COL26A1
IX
Type II fibril
GAG
GAG
XII and XIV
100 nm 100 nm
ficolin ectodysplasin
Fibril-forming (fibrillar) collagens form the most abundant subfamily of collagens
and this group includes seven different collagen types. Quantitatively the most
common types are type I, II and III collagens, whereas types V, XI, XXIV and
XXVII are minor forms. (Gelse et al. 2003, Myllyharju & Kivirikko 2004,
Ricard-Blum & Ruggiero 2005). The fibril diameter of the collagen molecules
100 nm
f. Type VII collagen forming anchoring fibrils
Gene: COL7A1
Genes: COL13A1; COL17A1; COL23A1;
VII
COL25A1
100 nm
Type I fibril
Genes: COL15A1; COL18A1
GAG
23

varies from 12 nm to over 500 nm and depends on the tissue and the stage of
development (Kadler et al. 2007). Tissues that must withstand high stress levels
contain a higher proportion of collagen fibers with a large diameter. Thus,
diameter increases at the expense of flexibility. (Bailey et al. 1998). Fibrillar
collagens, excluding types XXIV and XXVII, share an about 1000 amino acid
residues long domain which has a perfect collagenous (Gly-X-Y)n sequence
without any interruptions (Gordon & Hahn 2010). In addition, they all have N-
and C-terminal propeptides, which are cleaved before fibril formation, and may
have also some smaller NC domains. However, the cleavage of the N propeptide
is not always essential. It can also be partially cleaved or entirely intact such as in
the early embryonic form of type II collagen (IIA). The role of N propeptide in
the mature collagen is not known, but the globular N-terminal end of the N
propeptides may have binding functions. One common structure in the globular
N-terminal end is a cysteine-rich domain, which is shown to be able to bind
growth factors such as certain bone morphogenetic proteins and TGF-β1.
(Fleischmajer et al. 1990, Ng et al. 1993, Zhu et al. 1999, Larraín et al. 2000
Ricard-Blum & Ruggiero 2005).
Fibrillar collagens provide stability and integrity to the tissues and they are
ubiquitously found in tissues such as bone, cartilage and skin. Type I collagen is
usually a heterotrimer consisting of two α1(I) chains and one α2(I) chain. It forms
about 90% of the organic mass of bone and together with type V collagen it is
responsible for the structural backbone of bone. Type I collagen is also the major
collagen type in many other tissues such as tendon, skin and ligaments. Although
type I collagen is the most common collagen type in many tissues, the tissues can
be mechanically very different. In skin, for example, type I collagen forms
compounds with type III collagen providing tensile strength for the tissue, and in
bone, load bearing ability as well as tensional and torsional properties are
accompanied by calcification. (Gelse et al. 2003). The stability of the collagenous
structures is also affected by the type of cross-links predominating between
collagen molecules (Knott & Bailey 1998, Bank et al. 1999).
2.2.2 Collagen biosynthesis
Collagen biosynthesis from messenger RNA (mRNA) to an active protein is a
complex process characterized by many co- and post-translational modifications
that are unique only for collagens and proteins containing collagen-like domains
(see Fig. 2). The biosynthesis of fibril-forming collagens is best known and thus
24

used often as an example of collagen synthesis in general. Collagen biosynthesis
can be divided to intracellular and extracellular stages. Intracellular steps include
the synthesis and modification of a procollagen molecule from a gene transcript,
whereas extracellular steps convert procollagen molecules to collagen molecules
and furthermore to more extensive final structures. (Kielty & Grant 2002, Gelse et
al. 2003).
Cytoplasm
Nucleus
Extracellular
space
Endoplasmic
reticulum
Fig. 2. Main steps in the biosynthesis of fibril-forming collagens. Modified from
Myllyharju & Kivirikko 2004.
-OH
-OH
Intracellular co- and post-translational modifications
-OH
OH
-OH
-OH
-O-Gal-Glc
-
OH OH OH
OH OH
-
-
-
-
-
-
O-Gal-Glc
-OH
-OH
-O-Gal
-OH
S
S
S
S
Synthesis & co- and
post-translational modifications
Collagen prolyl 4-hydroxylase
-OH Prolyl 3-hydroxylase
Lysyl hydroxylase
-O-Gal-Glc
-O-Gal
Collagen glycosyltransferases
Assembly of 3 procollagen chains &
disulfide bonding
Protein disulfide isomerase
Assembly of triple helix
Cleavage of propeptides
N and C proteinases
Assembly into collagen fibrils &
cross-link formation
Lysyl oxidase
Collagen polypeptide chains are translated in the rough endoplasmic reticulum
(ER) as pre-pro-α-chains. The growing polypeptide chain is directed into the
lumen of the ER according to an N-terminal signal sequence. The signal sequence
is removed after translocation into the lumen by a signal peptidase, and the
produced pro-α-chain is immediately available for further co- and posttranslational
modifications. These modifications are performed within the
cisternae of ER and they consist mainly of hydroxylations and glycosylations.
Collagen prolyl 4-hydroxylases (C-P4Hs), prolyl 3-hydroxylases (P3Hs) and lysyl
25

hydroxylases (LHs) are members of the 2-oxoglutarate dioxygenase family, and
they catalyze hydroxylations of certain proline and lysine residues in procollagen
chains to 4-hydroxyproline, 3-hydroxyproline and hydroxylysine, respectively
(see Fig. 2). Glycosylations of some of the hydroxylysyl residues to
galactosylhydroxylysine and glucosylgalactosyl-hydroxylysine are catalyzed by
collagen galactosyltransferase and glugosyltransferase, respectively. (Kielty &
Grant 2002, Canty & Kadler 2005, Myllyharju 2005).
4-Hydroxyproline residues are essential for the thermal stability of the
collagen triple helix. Collagen α-chains that are not properly hydroxylated by C-
P4H do not hold together in a triple helix conformation in physiological
conditions. (Kivirikko & Pihlajaniemi 1998, Myllyharju & Kivirikko 2001). The
role of 3-hydroxyproline was unknown for a long time until a decreased prolyl 3hydroxylation
in CRTAP (cartilage-associated protein) deficiency was shown to
be connected to osteogenesis imperfecta (OI) (Morello et al. 2006, Marini et al.
2010). 3-Hydroxyproline residues are now suggested to have a role in the
assembly of collagen fibrils (see below, Weis et al. 2010). Hydroxylysine residues
are needed for the formation of covalent collagen cross-links and they serve as
sites for glycosylation (Kielty & Grant 2002). In addition to the collagenous
domain, hydroxylysine residues are found in the short noncollagenous telopeptide
regions at the ends of the collagen molecule (Kivirikko et al. 1992). LHs and
hydroxylysines will be discussed in more detail later (see 2.3 and 2.4).
C-P4Hs, P3Hs and LHs require the cofactors Fe 2+ , 2-oxoglutarate, molecular
oxygen, and ascorbate (vitamin C) for the enzymatic reaction, and the activity of
these enzymes is required for the formation of functionally normal collagens
(Myllyharju & Kivirikko 2004, Marini et al. 2010). In humans the lack of vitamin
C, for instance, leads to inadequate hydroxylations of collagens, the retaining of
collagen chains within the ER, and thus to poor connective tissue renewal which
is characteristic of scurvy, for instance (Canty & Kadler 2005). Intracellular
modifications of collagen also consist of reactions that are common for many
noncollagenous proteins, such as glycosylation of certain asparagine residues and
formation of disulfide bonds (Myllyharju & Kivirikko 2004, Kielty & Grant 2002,
Gelse et al. 2003).
Trimerization and triple helix formation of collagens
During post-translational modifications three newly synthesized pro-α-chains
connect via their C propeptide areas to form an attachment point, the nucleus. The
26

association is guided according to specific recognition sequences. (Lees et al.
1997, Kielty & Grant 2002). The trimeric structure is stabilized by formation of
interchain disulfide bonds between the C propeptides. This reaction is catalyzed
by the multifunctional protein disulfide isomerase (PDI). (Kielty & Grant 2002,
Myllyharju & Kivirikko 2004). PDI also has another stabilizing feature when it
acts as a chaperone and binds to newly synthesized collagen polypeptide chains
preventing their aggregation (Wilson et al. 1998, Myllyharju & Kivirikko 2001).
Collagen biosynthesis also requires other molecular chaperones such as the
essential collagen-specific chaperone Hsp47. Hsp47 inhibits procollagen
aggregation by binding procollagen in the ER and facilitating triple helix
formation. (Canty & Kadler 2005, Ishida & Nagata 2011).
Triple helix formation begins when C propeptides have associated and the
nucleus is formed, the cis peptide bonds involving proline are converted to the
trans form by the peptidyl-prolyl cis-trans isomerase, and when each pro-α-chain
contains about 100 4-hydroxyproline residues. Triple helix formation propagates
from the C terminus to the N terminus like a zipper. (Myllyharju & Kivirikko
2001, Canty & Kadler 2002, Kielty & Grant 2002). Generally the triple helix
structure prevents further hydroxylations, but one of the three LH isoenzymes,
LH3, has been recently reported to modify also native triple helical collagen
molecules in the extracellular space (Kielty & Grant 2002, Salo et al. 2006a).
Although the enzymatic modifications of collagen polypeptides described above
are necessary, overmodification is harmful. A delay in triple helix formation can
lead to intracellular overmodification of the collagen polypeptides followed by a
perturbation of the final triple helix structure (Torre-Blanco et al. 1992,
Raghunath et al. 1994). The triple helical procollagen molecules are transported
from the ER to the extracellular space across the Golgi stacks probably via
cisternal maturation (Myllyharju & Kivirikko 2001, Kielty & Grant 2002, Canty
& Kadler 2005).
Extracellular stage
The assembly of collagen molecules into fibrils is thought to be an extracellular
step in collagen biosynthesis. However, the assembly may begin already during
the transportation from the ER. (Canty & Kadler 2005). A key point in the fibril
formation is the cleavage of the globular N and C propeptides of the procollagen
molecules. This is carried out by two specific metalloproteinases, a procollagen N
proteinase and a procollagen C proteinase, respectively. The produced collagen
27

molecules then assemble spontaneously into a fibrillar structure in the case of
fibril-forming collagens. Finally, collagen cross-links are formed within and
between collagen molecules to stabilize the formed structures. (Myllyharju &
Kivirikko 2001, Kielty & Grant 2002, Canty & Kadler 2005). This final crosslinking
step begins with an oxidative deamination of the telopeptide lysine or
hydroxylysine residue by lysyl oxidase (LO), which is followed by certain
spontaneous reactions (Knott & Bailey 1998, Mäki 2009). In tissues the fibrils are
usually organized further into higher-order three-dimensional structures (Canty &
Kadler 2005).
Specific features in the biosynthesis of non-fibrillar collagens
Although the basic features of the biosynthesis of non-fibrillar collagens are
identical to those described above with fibrillar collagens, distinct differences also
exist. Final macromolecular assemblies determine the direction of the synthesis.
Some collagen types are subjected to additional modifications like additions of
glycosaminoglycan side chains. It is also assumed that the triple helix of
transmembrane collagens may be propagated from the N to C terminus. Many
collagens retain an N- or C-terminal noncollagenous domain or both in their
structures, and thus these domains are not cleaved during extracellular
modifications. (Myllyharju & Kivirikko 2001, Kielty & Grant 2002).
2.3 Collagen hydroxylysines and their function
Hydroxylysine (Hyl) residues are usually found in the Y position of the repeating
X-Y-Gly triplet sequence in collagens and the collagenous domains of several
noncollagenous proteins. The repeating Gly-X-Y sequences are written here as
X-Y-Gly due to the hydroxylation properties of collagen hydroxylases, the
minimum sequence requirement for hydroxylation by e.g. LH being X-Lys-Gly.
Hydroxylysine residues are also found in X-Hyl-Ser and X-Hyl-Ala sequences in
the telopeptide regions, i.e. in the ends of the α chains of some fibril-forming
collagens. (Kivirikko et al. 1992, Myllyharju 2005). A hydroxylysine residue has
been discovered in an X-Hyl-Ala sequence also in the NC1 domain of type IV
collagen, where it was suggested to participate in the formation of an Shydroxylysyl-methionine
cross-link, a sulfilimine bond, with a particular
methionine residue (Vanacore et al. 2005, Vanacore et al. 2009). In addition to
collagenous polypeptides, a single hydroxylysine residue is known to be present
28

e.g. in anglerfish somatostatin-28 (Kivirikko et al. 1992). Recently hydroxylysine
residues were found in histone peptide fragments and in proteins contributing to
RNA splicing, where the hydroxylation reaction is suggested to be catalyzed by
the Jumonji domain-6 protein (JMJD6) and not by LHs (Chang et al. 2007,
Webby et al. 2009).
The lysyl hydroxylation is usually not complete and some lysine residues
located in the Y position of X-Y-Gly triplets or in telopeptides remain
unhydroxylated. The hydroxylysine content of collagens depends on the collagen
type and tissue as well as the physiological and pathological state of the tissue.
(Kivirikko et al. 1992, Myllyharju 2005). Generally the highest hydroxylation
content is observed in type IV collagen, where almost 90% of lysine residues in
the helical area are hydroxylated, whereas in collagen III, for instance, the value
is less than 20% (Kivirikko et al. 1992). The significance of hydroxylysine
residues can be seen in various diseases. The extent of hydroxylation is decreased
in Bruck Syndrome (Bank et al. 1999, van der Slot et al. 2003) and EDS VIA
(Steinmann et al. 2002) and increased in fibrotic diseases (Brinckmann et al. 1996,
Brinckmann et al. 1999, van der Slot et al. 2003).
Hydroxylysine residues have at least two major functions in collagen
molecules. They act as attachment sites for either galactose or glucosylgalactose
carbohydrate residues. However, the purpose of these glycosylations is not fully
understood. Hydroxylysine residues are also essential in the formation of
intermolecular stabilizing covalent collagen cross-links and thus in supporting the
structure of the collagen molecule and its supramolecular assemblies. (Kivirikko
& Pihlajaniemi 1998, Kielty & Grant 2002).
2.3.1 Glycosylation of hydroxylysines
Carbohydrate units linked to hydroxylysine residues of collagen molecules are
either monosaccharide galactoses or disaccharide glucosylgalactoses (Myllyharju
2005). As the hydroxylation of lysine residues varies between collagen types,
tissue distribution and physiological state, similarly the extent of the
hydroxylysine glycosylation as well as the ratio of galactose and
glucosylgalactose vary (Kivirikko & Myllylä 1979, Kivirikko et al. 1992). Type
IV and VI collagens contain the highest amount of glycosylated hydroxylysines
with almost complete glycosylation (Kivirikko & Myllylä 1979, Chu et al. 1988),
and recently the inadequate glycosylation of these collagen types was reported to
prevent their normal assembly and secretion (Sipilä et al. 2007).
29

Two different transferase activities are required during the two glycosylation
steps. In the first step the formation of a bond between the hydroxylysine residue
and galactose moiety is catalyzed by hydroxylysyl galactosyltransferase (GT).
Some of the formed galactosylhydroxylysines are further modified by galactosyl
hydroxylysyl glucosyltransferase (GGT) to form a glucosylgalactosyl
hydroxylysine residue. (Myllyharju 2005). LH3, but not LH1 or LH2, is now
known to have both collagen glycosyltransferase activities in addition to its LH
activity (Heikkinen et al. 2000, Rautavuoma et al. 2002, Wang et al. 2002a). It
was speculated earlier that in addition to LH3 there are likely to be additional
collagen glycosyltransferases in tissues, especially as the GT activity of LH3 is
low (Rautavuoma et al. 2002, Rautavuoma et al. 2004) and two
glycosyltransferase GLT25 family members, GLT25 domain 1 and 2 (GLT25D1
and GLT25D2), were recently observed to have strong collagen GT activity
(Schegg et al. 2009). GLT25D1 is localized in the ER like LHs (Liefhebber et al.
2010). It has been suggested that the GLT25D1 could be responsible for the GT
reactions for example in bone type I collagen (Sricholpech et al. 2011).
The function of glycosylated hydroxylysines
Although the hydroxylysine glycosylation of collagens was observed decades ago,
its exact biological role is still not fully understood (Kivirikko & Myllylä 1979).
However, studies of type I and II collagens have revealed that the carbohydrate
residues may affect the collagen fibril diameter and collagen fibrillogenesis
(Torre-Blanco et al. 1992, Notbohm et al. 1999, Brinckmann et al. 1999,
Sricholpech et al. 2011), and an older study of type I collagen suggests that the
carbohydrate moieties could prevent excess cross-link formation (Yamauchi et al.
1982). Nevertheless, glycosylated hydroxylysines have been shown to participate
in the formation of cross-links (Henkel et al. 1976, Hanson & Eyre 1996,
Brinckmann et al. 1999). The importance of glycosylated hydroxylysine residues
is also demonstrated by studies with LH3 knock-out mice. The mice die during
embryonic development due to lack of hydroxylysine glycosylation activity and
subsequent impaired production of type IV collagen (Rautavuoma et al. 2004,
Ruotsalainen et al. 2006). As indicated earlier, the under-glycosylation of highly
glycosylated collagen types IV and VI affected their assembly and secretion
(Sipilä et al. 2007). Cellular studies have also shown that glycosylation catalyzed
by LH3 is important for the organization of the ECM as well as cell growth and
viability (Risteli et al. 2009, Wang et al. 2009). Mutation studies of mannose-
30

inding lectins that contain collagenous domains have also revealed that the
glycosylation of hydroxylysines is important in the folding and secretion of the
protein (Heise et al. 2000). However, just as collagens have numerous different
functions, the role of hydroxylysine-linked carbohydrates is presumably more
diverse than discussed above. For example, at least in type II collagen
glycosylated hydroxylysine residues can act as antigens and induce autoimmune
responses such as autoimmune arthritis (Myers et al. 2004), and they are also
suggested to affect the binding of type IV collagen to the melanoma receptor
(Lauer-Fields et al. 2003).
2.3.2 Collagen cross-links
Collagen cross-links are essential in the determination of the physical and
chemical properties of collagens and their significance is demonstrated in many
diseases (Bank et al. 1999, Brinckmann et al. 1999, Steinmann et al. 2002, Eyre
& Wu 2005, Levental et al. 2009). Cross-links can be formed between similar
collagen molecules but also between different collagen types (Eyre & Wu 2005).
Two opinions exist whether one cross-link is formed between two or three
collagen molecules (Bailey et al. 1998). Cross-links can be separated into
enzyme-mediated or non-enzymatic according to the initiation of the cross-link
formation (Reiser et al. 1992). Undesirable, occasional sugar-related cross-links
derived from the non-enzymatic pathway are connected to aging and diabetes, for
instance, while the enzymatic cross-linking initiated by LO is essential for proper
tissue function (Reiser et al. 1992, Eyre & Wu 2005, Robins 2007, Mäki 2009).
At least fibril-forming collagen types I, II, III, V and XI as well as certain
other collagen types are cross-linked by a mechanism requiring LO enzyme. The
formation of these cross-links can follow two basic pathways, the lysine aldehyde
(allysine) or hydroxylysine aldehyde (hydroxyallysine) pathway, where certain
lysine or hydroxylysine residues, respectively, function as precursors. (Knott &
Bailey 1998, Eyre & Wu 2005). At least types I, II and III of the fibrillar
collagens have four lysine residues that participate in the cross-link formation.
Two of the residues are located in the triple helical area and two in the telopeptide
regions, one residue in each telopeptide. (Knott & Bailey 1998). The
alternatively-spliced form of LH2, LH2(long), is the LH that hydroxylates the
telopeptide lysines in collagens, while LH1 is mainly responsible for helical
lysine hydroxylation (Uzawa et al. 1999, Eyre et al. 2002, Mercer et al. 2003, van
der Slot et al. 2003, Pornprasertsuk et al. 2004, Takaluoma et al. 2007). The
31

hydroxylation of lysine residues in the telopeptide area determines whether the
allysine or hydroxyallysine pathway is utilized for cross-linking, the
hydroxylation state of the helical lysine determines the final form for cross-link,
and LO completes the procedure (Knott & Bailey 1998, Eyre & Wu 2005). Thus,
LHs and LO are the regulatory enzymes in the cross-linking pathway. Many
general and connective tissue-specific factors such as TGF-β, connective tissue
growth factor, tumor necrosis factor-α, vitamin B6, and the active form of vitamin
D regulate the activity of these enzymes and thus also cross-linking. (Saito &
Marumo 2010). The cross-linking pathways initiated by LO are illustrated in Fig.
3.
Fig. 3. Lysyl oxidase-mediated collagen cross-linking. Modified from Eyre & Wu 2005.
32
Divalent
cross-links
Mature
cross-links
Allysine pathway
+ hLys
LO
deH-LNL
tLys
Allysine
Aldimines
+ hHyl
deH-HLNL
+ hHis
HHL
?
Hydroxyallysine pathway
Lysyl
pyrrole
+ hLys
+ hHyl
LKNL HLKNL
Ketoimines
LP
tHyl
LO
Hydroxyallysine
+ Hydroxyallysine /
HLKNL
HP
Hydroxylysyl
pyrrole

Lysyl oxidase-mediated cross-linking
Enzymatic cross-linking begins with an oxidative deamination of a telopeptide
lysine or hydroxylysine residue by LO (Knott & Bailey 1998). The formed
aldehydes allysine or hydroxyallysine, respectively, are further modified
spontaneously during subsequent steps of the two separate pathways (Knott &
Bailey 1998, Eyre & Wu 2005) (see Fig. 3).
In allysine pathway the telopeptide lysine residue reacts with a triple helical
lysine or hydroxylysine residue yielding a divalent cross-link dehydrolysinonorleucine
(deH-LNL) or dehydro-hydroxylysinonorleucine (deH-HLNL),
respectively (Knott & Bailey 1998). The deH-HLNL of these immature aldimines
can react further with a helical histidine residue to form a mature, trifunctional
cross-link histidinohydroxylysinonorleucine (HHL), which is common in the skin
for instance (Eyre & Wu 2005, Robins 2007).
In hydroxyallysine pathway the telopeptide hydroxylysine residue can react
with a triple helical lysine or hydroxylysine residue to form a
lysino-5-ketonorleucine (LKNL) or a hydroxylysino-5-ketonorleucine (HLKNL)
cross-link, respectively (Knott & Bailey 1998). Depending on the theory, these
ketoimines are further incorporated with a lysine aldehyde or a hydroxylysine
aldehyde or two ketoimines are condensed to form mature pyridinoline and
pyrrole cross-links (Knott & Bailey 1998, Eyre & Wu 2005, Robins 2007). Thus,
the pyridinolines lysylpyridinoline (LP) and hydroxylysylpyridinoline (HP) are
formed from the reaction between LKNL or HLKNL, respectively, and
hydroxyallysine or another HLKNL (Knott & Bailey 1998, Eyre & Wu 2005).
Both cross-links are common in stiff tissues, but LP is predominant in calcified
tissues and HP in cartilage, for instance. Lysyl pyrrole and hydroxylysyl pyrrole
are formed in turn from LKNL or HLKNL, respectively, and allysine or
deH-HLNL. (Bailey et al. 1998, Knott & Bailey 1998, Eyre & Wu 2005).
Recently, a new maturation product, named arginoline, was detected from the
cartilage collagens. The arginoline cross-link is formed from the ketoimine by an
oxidation reaction followed by an addition of free arginine. It is a functionally
divalent but nonreducible cross-link. About half of the mature cross-links in the
type II collagen are arginoline cross-links instead of HP. (Eyre et al. 2010). A
hydroxylysine residue is also involved in the covalent sulfilimine bond found in
collagen IV. This cross-link is formed between a hydroxylysine and a methionine.
(Vanacore et al. 2005, Vanacore et al. 2009).
33

Cross-links in tissues
The cross-linking pattern as well as the degree of it seem to be more tissuespecific
than collagen type-specific (Knott & Bailey 1998, Eyre & Wu 2005).
Cross-links in skin and other soft tissues are mainly products from the allysine
pathway, whereas cross-links derived from the hydroxyallysine pathway
predominate in tissues that require strength and stiffness such as bone, dentin,
cartilage and tendon (Knott & Bailey 1998, Kielty & Grant 2002, Eyre & Wu
2005). In addition, physiological processes affect the cross-linking state as seen in
the relatively fast cervix remodelling during pregnancy, for instance (Akins et al.
2011). In skin the level of lysine hydroxylation is low and the main form of the
divalent cross-link is deH-HLNL, which can mature later to HHL (Knott & Bailey
1998, Avery & Bailey 2005, Robins 2007). Despite the high physical
requirements of bone, only a small amount of mature pyridinoline cross-links
exist in bone collagens, but the existing cross-links are formed from lysines that
are hydroxylated, i.e. ketoimines and pyridinolines (Knott & Bailey 1998). This is
suggested to be caused by the increased proportion of pyrrole cross-links (Bailey
et al. 1998). In addition, bone tissue is under constant remodelling with the halflife
of collagen being low when compared to the slowly renewing cartilage, where
lysine residues are highly hydroxylated and collagen cross-links well matured
(Knott & Bailey 1998, Avery & Bailey 2005, Eyre et al. 2010). The half-life of
type II collagen predominating in cartilage is estimated to be about 100 years
(Kielty & Grant 2002, Avery & Bailey 2005).
The cross-linking pattern with a large amount of divalent cross-links in bone
is suggested to have an important role in the correct mineralization (Yamauchi &
Katz 1993, Knott et al. 1997, Bank et al. 1999, Pornprasertsuk et al. 2005), but
mineralization is also suggested to affect the maturation of collagen cross-links
(Knott & Bailey 1998). The nucleation of calcium apatite crystals begins in areas
adjacent to cross-linking sites, for instance (Yamauchi & Katz 1993). In addition,
the level of telopeptide hydroxylation and cross-linking may also affect the
collagen fibril diameter (Brinckmann et al. 1999, Bailey 2002, Pornprasertsuk et
al. 2005). Studies of osteoporotic bones have revealed over-hydroxylated
collagens leading to increased immature cross-linking, finer collagen fibrils,
increased collagen synthesis as well as degradation, and reduced calcification,
which in turn lead to the increased fragility of bone (Bailey 2002, Pornprasertsuk
et al. 2005). Recently, decreased levels of LH2, mature HP and LP cross-links and
two matrix proteins in mouse cervical tissue were accompanied by increased
34

collagen fibril diameter (Akins et al. 2011). In addition to enzymatic collagen
cross-links, the role of non-enzymatic, sugar related cross-links is also
emphasized in the determination of bone quality (see below) (Saito & Marumo
2010).
Cross-links can be determined from tissue samples, but the analysis of
collagen turnover products from urine and blood is also a useful method to detect
some connective tissue disorders (Knott & Bailey 1998, Eyre & Wu 2005, Saito
& Marumo 2010). Pyridinolines and ketoimines originate mainly from
mineralized tissues and are thus valuable markers of the quality of these tissues
and ongoing collagen turnover in them (Knott & Bailey 1998). However, the
analysis is not always that straightforward and pyridinoline cross-links in samples
are not always from bone collagen. For example, growing children get higher
values due to the breakdown of mineralized cartilage in growth plates rather than
bone. (Eyre & Wu 2005).
Age-related, non-enzymatic cross-links
Lysine residues are also involved in the non-enzymatic cross-links between
collagen molecules in advanced glycation end-products (AGEs), such as
glucosepane and pentosidine (Saito & Marumo 2010). AGEs are formed
spontaneously by a reaction between reducing sugars and proteins of the body
(Susic 2007, Saito & Marumo 2010). The formation requires time so that it affects
long-lived proteins of the body, such as collagen (Susic 2007). As collagens
mature after enzymatic cross-linking, the metabolic turnover decreases and allows
for non-enzymatic cross-linking (Bailey et al. 1998). In contrast to the enzymemediated
cross-links that are a prerequisite for normal tissue function, AGEmediated
cross-links are connected to aging and they are harmful for tissues
(Robins 2007). These cross-links are suggested to have a role in diabetes mellitus,
cardiovascular diseases and osteoporosis, for instance (Susic 2007, Saito &
Marumo 2010). They increase the stiffness of the tissue and affect the resistance
of the tissue to normal enzymatic degradation (Susic 2007). In bone AGEs make
collagen fibers brittle and they may also affect the activity of osteoblasts and
osteoclasts. AGE-related cross-links can be determined from tissue samples and,
just as turnover products of enzymatic cross-links, also e.g. pentosidine can be
analyzed from serum or urine samples. Serum and urinary levels of pentosidine
do not directly reflect the content in bone but may correlate to fracture risk. (Saito
& Marumo 2010).
35

2.4 Lysyl hydroxylases
Hydroxylysine residues in polypeptides are derived from lysine residues by
enzymatic reaction and not from free hydroxylysines (Kivirikko & Pihlajaniemi
1998). The reaction involves an oxidative decarboxylation of 2-oxoglutarate that
is catalyzed by LH, which was first purified to homogeneity from chick embryos
and human placental tissue and cloned from chick (Turpeenniemi-Hujanen et al.
1980, Turpeenniemi-Hujanen et al. 1981, Myllylä et al. 1991, Kivirikko &
Pihlajaniemi 1998). At present, three LH isoforms, LH1, LH2 and LH3, have
been characterized from vertebrates including human (Hautala et al. 1992a,
Valtavaara et al. 1997, Valtavaara et al. 1998, Passoja et al. 1998a), mouse
(Ruotsalainen et al. 1999), rat (Armstrong & Last 1995, Mercer et al. 2003) and
zebrafish (Schneider & Granato 2006, Schneider & Granato 2007). However,
only a single LH is found in the nematode Caenorhabditis elegans (Norman &
Moerman 2000) and the fly Drosophila melanogaster (Bunt et al. 2010). LHs
together with C-P4Hs and P3Hs belong to the 2-oxoglutarate-dependent
dioxygenase family and the genes encoding LHs are called PLODs
(protocollagen-lysine, 2-oxoglutarate, 5-dioxygenase) (Kivirikko & Pihlajaniemi
1998, Kielty & Grant 2002).
Recently a Jumonji domain-6 protein (JMJD6) was also reported to have
lysyl hydroxylation activity (Webby et al. 2009). Like collagen-modifying
hydroxylases, JMJD6 is suggested to be an iron- and 2-oxoglutarate-dependent
dioxygenase that belongs to the subfamily of Jumonji C domain-containing
hydroxylases (Chang et al. 2007). These enzymes in turn are connected to
diseases such as cancer (for a review, see Kampranis & Tsichlis 2009). The latest
data suggest that JMJD6 hydroxylates lysine residues in histones and in proteins
contributing to RNA splicing (Webby et al. 2009) and that the enzyme may also
bind and modify single-stranded RNA (Hong et al. 2010).
LHs are localized to the lumen of the ER (Myllylä et al. 1991, Kellokumpu et
al. 1994) even though they do not contain any retention sequence typical for
proteins remaining in the ER (Myllylä et al. 1991, Hautala et al. 1992a). In
addition, any cycling or retrieval from post-ER compartments is not involved
(Suokas et al. 2003). LH1 and LH3 have been shown to be associated with the ER
membrane (Kellokumpu et al. 1994, Salo et al. 2006a). LH1 contains a distinct Cterminal
segment that is suggested to be important both for its localization into
the lumen of the ER as well as its association with ER membranes (Kellokumpu
et al. 1994, Suokas et al. 2000, Suokas et al. 2003). Thus, the retention within the
36

ER and the association with the ER membrane are found to be coupled
phenomena as both characteristics are due to the C-terminal segment (Suokas et
al. 2000). This C-terminal region shows high sequence homology in all three LH
isoforms (Valtavaara et al. 1998, Passoja et al. 1998a), and it is assumed that all
three LH isoenzymes utilize the same mechanism in their localization to the ER
(Suokas et al. 2000, Suokas et al. 2003), which is also supported by the latest data
concerning LH3 (Wang et al. 2012). In addition to the ER, the active form of LH3
is found in the extracellular space and the localization seems to have tissuespecificity.
The intracellular form is preferred at least in the kidney, whereas the
extracellular form is prevalent in the liver. In addition, LH3 is detected in the
serum. (Salo et al. 2006a).
2.4.1 Lysyl hydroxylase isoenzymes
The existence of LH isoenzymes and especially separate forms for the
hydroxylation of the helical and telopeptide regions was debated for a long time
(Barnes et al. 1974, Ihme et al. 1984, Royce & Barnes 1985, Gerriets et al. 1993).
The presence of hydroxylysine residues in distinct amino acid sequences in
helical and telopeptide regions, for instance, led to the suggestion that different
forms of LHs may exist (Barnes et al. 1974, Royce & Barnes 1985). Moreover,
EDS VIA patients were reported to have immense variability in the lysine
hydroxylation levels between various tissues and collagen types, values ranging
from total lack of hydroxylation to normal levels (Ihme et al. 1984), and highly
purified LH was reported to be able to hydroxylate essentially completely the
triple helical lysines in type I protocollagen, while telopeptide lysines were
unhydroxylated (Royce & Barnes 1985). Finally, two new isoforms, LH2
(Valtavaara et al. 1997) and LH3 (Valtavaara et al. 1998, Passoja et al. 1998a),
were identified and later LH2 was recognized to have two different splicing forms
(Yeowell & Walker 1999a). PLOD1, PLOD2 and PLOD3, the genes encoding
human LH1, LH2 and LH3 (Hautala et al. 1992a, Szpirer et al. 1997, Valtavaara
et al. 1998), respectively, as well as the corresponding mouse genes (Plod1, Plod2
and Plod3) are located in different chromosomes (Sipilä et al. 2000).
According to a phylogenetic study, it has been suggested that the mouse
Plod1 and Plod2 isoforms are derived from one gene by two gene duplications,
Plod3 being the ancestral gene (Ruotsalainen et al. 1999). At the amino acid level
all three human LH isoforms share about 50% identity (Valtavaara et al. 1998). If
individual isoforms are compared (LH1-LH2, LH1-LH3, LH2-LH3) the identities
37

are even higher, ranging from about 60% to over 75% in both human and mouse
(Valtavaara et al. 1997, Valtavaara et al. 1998, Passoja et al. 1998a, Ruotsalainen
et al. 1999). Between species, individual mouse LHs are about 90% identical to
the corresponding human LHs at the amino acid level (Ruotsalainen et al. 1999).
The most concerved areas are located in the C-terminal region, where the amino
acids essential for LH activity are located, and in the central region of the
molecule (Pirskanen et al. 1996, Valtavaara et al. 1997, Passoja et al. 1998a,
Passoja et al. 1998b, Valtavaara et al. 1998, Rautavuoma et al. 2002).
Lysyl hydroxylase 1
LH1 was the first LH isoenzyme cloned from chick and human and it was named
LH before another isoform, LH2, was discovered (Hautala et al. 1992a,
Valtavaara et al. 1997). The gene coding for human LH1, PLOD1, is located on
chromosome 1 and it contains 19 exons (Hautala et al. 1992a, Heikkinen et al.
1994). PLOD1 encodes a polypeptide of 727 amino acids including a signal
peptide of 18 residues (Hautala et al. 1992a). In mouse the gene is localized to
chromosome 4 (Sipilä et al. 2000). Interestingly, introns 9 and 16 of PLOD1 show
extensive homology because of many Alu sequences (Heikkinen et al. 1994).
These sequences are suggested to be involved in many human gene
rearrangements and therefore in some hereditary diseases (Deininger & Batzer
1999). In the case of PLOD1, Alu sequences have been connected to one of the
most common mutations causing EDS VIA (Heikkinen et al. 1994, Pousi et al.
1994, Heikkinen et al. 1997, Deininger & Batzer 1999).
Lysyl hydroxylase 2
The human PLOD2 gene coding for LH2 is localized to chromosome 3 (Szpirer et
al. 1997) and the corresponding mouse gene Plod2 to chromosome 9 (Sipilä et al.
2000). Mutations in the PLOD2 gene cause Bruck syndrome (BS) type 2 (see
2.5.2., van der Slot et al. 2003, Ha-Vinh et al. 2004). LH2 is the only LH isoform
that is known to have different splicing variants (Yeowell & Walker 1999a). The
two different forms are named LH2(short) and LH2(long) or LH2a and LH2b,
respectively (Yeowell & Walker 1999a, Walker et al. 2005). LH2(long) was
characterized two years later than the short form (Valtavaara et al. 1997, Yeowell
& Walker 1999a). In the short form exon 13A between exons 13 and 14 is spliced
out, yielding a product of 737 amino acids that includes an N-terminal signal
38

peptide (Valtavaara et al. 1997, Yeowell & Walker 1999a). Exon 13A is 63 bp
long corresponding to 21 amino acids, and thus the long form is 758 amino acids
long (Yeowell & Walker 1999a). The amino acids encoded by the additional exon
13A in LH2(long) are suggested to change the peptide binding properties of LH2
(Risteli et al. 2004).
The expression of LH2 splice variants shows tissue specificity (Yeowell &
Walker 1999a, Salo et al. 2006b). In humans LH2(long) is expressed in all tissues
studied, while the short form is expressed in certain tissues and only together with
the long form (Yeowell & Walker 1999a, Walker et al. 2005). The LH2 splicing
pattern is suggested to be connected to the specificity for collagen telopeptide
lysine hydroxylation (Mercer et al. 2003, van der Slot et al. 2003). LH2, and
particularly LH2(long), is reported to be the collagen telopeptide LH responsible
for the initiation of the formation of pyridinoline cross-links (Uzawa et al. 1999,
Mercer et al. 2003, van der Slot et al. 2003, Pornprasertsuk et al. 2004, van der
Slot et al. 2004, Takaluoma et al. 2007). The role of the short form is unknown. In
fibrotic diseases such as systemic sclerosis, the splicing of LH2 is changed so that
LH2(long) expression is increased (van der Slot et al. 2003, van der Slot et al.
2004). Accordingly, overaccumulation of collagen in fibrotic conditions is
accompanied by overhydroxylation of the collagen telopeptides and an increased
amount of pyridinoline cross-links (Trojanowska et al. 1998, Brinckmann et al.
1999, Brinckmann et al. 2001, van der Slot et al. 2004).
The prevalence and clinical significance of fibrosis have led to increasing
investigation of the factors regulating splicing of the LH2 mRNA. Early studies
of LH2 splicing with a protein synthesis inhibitor cycloheximide demonstrated
that the splicing pattern can be changed by affecting the synthesis of a protein
factor necessary for the splicing (Walker et al. 2005). Phylogenetic and sequence
analyses have revealed that introns flanking exon 13A of the LH2 transcript and
exon 13A itself contain certain regulatory splicing sequences (Yeowell & Walker
1999a, Yeowell et al. 2009, Seth & Yeowell 2010, Seth et al. 2011). A branch
point sequence, suggested to participate in the exon inclusion, is followed by a
polypyrimidine tract that could contribute to the exclusion of exon 13A (Yeowell
& Walker 1999a, Yeowell et al. 2009). In addition, a sequence similar to a U-rich
element that is known to function as a binding site for TIA (T-cell-restricted
intracellular antigen) proteins is detected downstream of exon 13A (Förch et al.
2000, Del Gatto-Konczak et al. 2000, Le Guiner et al. 2001, Yeowell et al. 2009).
The binding of TIA proteins to U-rich elements has been connected to the
inclusion of alternatively spliced exons in general (Förch et al. 2000, Del Gatto-
39

Konczak et al. 2000, Le Guiner et al. 2001). Actually, two TIA proteins, TIA-1
and TIAL1, are shown to be essential for efficient inclusion of the alternatively
spliced exon 13A in LH2(long) mRNA and thus for the regulation of LH2(long)
expression (Yeowell et al. 2009). The intron upstream of exon 13A also contains
highly conserved motifs that bind Fox proteins (Baraniak et al. 2006, Seth &
Yeowell 2010). These proteins are known to regulate alternative splicing
(Baraniak et al. 2006), and in the splicing of LH2 Fox-2 has also been shown to
be required for the inclusion of exon 13A and thus production of the LH2(long)
mRNA (Seth & Yeowell 2010). Recently exon 13A itself was also reported to
contain elements affecting the alternative splicing. Exon 13A is suggested to
contain at least one enhancer and two silencer elements of exonic splicing that
could bind regulatory proteins. (Seth et al. 2011).
Lysyl hydroxylase 3
The PLOD3 gene encoding human LH3 is located on chromosome 7 (Valtavaara
et al. 1998). The gene contains 19 exons and the size of the LH3 polypeptide is
738 amino acids including a 24-residue signal peptide (Valtavaara et al. 1998,
Rautavuoma et al. 2000). The gene contains Alu sequences in the introns like that
of LH1, but the Alu sequences are shorter (Rautavuoma et al. 2000). In mouse the
gene is localized to chromosome 5 (Sipilä et al. 2000). LH3 is a multifunctional
enzyme possessing GT and GGT activities in addition to its LH activity
(Heikkinen et al. 2000, Rautavuoma et al. 2002, Wang et al. 2002a, Rautavuoma
et al. 2004, Ruotsalainen et al. 2006, Salo et al. 2006a).
The glycosyltransferase activities of LH3 are localized in the N-terminus in
contrast to the LH activity which exists in the C-terminus of the enzyme
(Pirskanen et al. 1996, Passoja et al. 1998b, Heikkinen et al. 2000, Rautavuoma
et al. 2002, Wang et al. 2002a, Wang et al. 2002b). LH3 differs from other LH
isoforms also in its cellular localization. In addition to its location in the lumen of
the ER, it is found as an active enzyme in the extracellular space and in
association with collagenous proteins (Kellokumpu et al. 1994, Salo et al. 2006a).
Thus, in addition to the general conception that LHs modify collagen
polypeptides only in the ER lumen before the formation of the triple helical
structure, it has been suggested that LH3 could modify also native collagen
molecules in the extracellular space, which furthermore indicates a possible role
for LH3 in ECM remodeling (Kielty & Grant 2002, Salo et al. 2006a). Actually,
cellular studies have suggested that glycosylation activity in the extracellular
40

space could dramatically affect cell morphology and cell behavior, although it
depends on the cell type (Wang et al. 2009). However, LH3 is suggested to prefer
collagen polypeptide chains rather than native triple helical collagen molecules as
substrates (Sricholpech et al. 2011). The enzyme is reported to be secreted from
cells via two pathways by using a Golgi complex or by bypassing it (Wang et al.
2012). LH3 is also found from serum samples indicating that the GGT activity
present in serum could be derived from LH3 (Salo et al. 2006a).
Expression of different lysyl hydroxylase isoforms
Expression levels of LH isoenzymes in human and mouse have been studied at
the mRNA and protein levels. LHs are widely expressed in different tissues and
developmental stages. Straightforward conclusions about any specific LH
isoenzyme expression patterns are essentially impossible as the results vary
markedly between different studies. The expression patterns of LH isoforms in
certain adult human and mouse tissues are shown in Table 1.
Table 1. Expression of lysyl hydroxylase isoenzymes in selected adult human and
mouse tissues.
Tissue LH1 LH2 LH3
human mouse human mouse human mouse
Blood vessels yes 2 n.d. yes 7 n.d. n.d. yes 9
Brain yes 1 yes 6,8 yes 3,7 yes 6,8,10 yes 4,5 yes 6,10
Cartilage yes 2 n.d. yes 7 n.d. n.d. yes 9
Heart yes 1 yes 6,8 yes 3 yes 6,8,10 yes 4,5 yes 6
Kidney n.d. no 8 /yes 6 yes 3,7 yes 6,8,10 yes 4,5 yes 6,9,10
Liver yes 1 yes 6,8 yes 3,7 no 6,8,10 yes 4,5 yes 6,9,10
Lung yes 1,2 yes 6,8 yes 3,7 no 8 /yes 6,10 yes 4,5 yes 6,9,10
Pancreas n.d. n.d. yes 3 yes 10 yes 4,5 yes 10
Placenta yes 1,2 n.d. yes 3,7 n.d. yes 4,5 yes 9
Skeletal muscle yes 1,8 yes 6,8 yes 3,8 no 8 /yes 6,10 yes 4,5 yes 6,10
1 (Heikkinen et al. 1994); 2 (Yeowell et al. 1994); 3 (Valtavaara et al. 1997); 4 (Passoja et al. 1998a);
5 (Valtavaara et al. 1998); 6 (Ruotsalainen et al. 1999); 7 (Yeowell & Walker 1999a); 8 (Hjalt et al. 2001);
9 (Rautavuoma et al. 2004); 10 (Salo et al. 2006b); n.d. = not determined.
In rat tissues mRNA expression of all LH isoforms including both LH2 splicing
variants was detected in almost all tissue types examined, indicating that LHs do
not have tissue-specific expression in the rat (Mercer et al. 2003). In humans,
both LH2 variants are expressed in the cartilage, frontal lobe, kidney, liver, spleen
41

and placenta (Yeowell & Walker 1999a). LH2(long) is expressed in all tissues
studied so far and it is the main transcript in all tissues studied except the kidney
and spleen (Yeowell & Walker 1999a). In mouse the short form of LH2 is
predominant in the testis, brain and kidney (Salo et al. 2006b). Thus, the
expression of the two human and mouse LH2 splicing variants is suggested to be
tissue-specific (Yeowell & Walker 1999a, Salo et al. 2006b).
All LH isoforms have been shown to be expressed with similar tissue
distributions in the mouse embryos from embryonic day (E) 7 onwards, the first
time point analyzed. LH1 expression was high throughout the analyzed period,
E7-E17, but LH2 was expressed strongly only at E7 and remained low after that.
LH3 was highly expressed at E7 and E15, the expression level being lower at the
other time points. (Salo et al. 2006b). Earlier data from LH3 null mice with the
Plod3 gene disrupted with a β-galactosidase reporter demonstrated intense X-gal
staining in heterozygous and null embryos at E8.5-9.5, the staining being
localized also to placental blood vessels (Rautavuoma et al. 2004). At E9.5 and
E11.5 all LH isoforms are expressed in the dorsal aorta, somites, tail bud, brain
and at sites of the developing maxilla and mandible of branchial arch as well as in
the limb bud at E11.5. Expression of the two LH2 splicing forms is suggested to
be developmentally regulated during mouse embryogenesis as the short form
predominates until E11.5 and the long form after that. (Salo et al. 2006b).
However, in humans, only the long form is present in fetal skin cell lines, but both
forms are seen in embryonic kidney cells (Yeowell & Walker 1999a, Walker et al.
2005).
2.4.2 Molecular and catalytic properties
An active form of LHs is a homodimer, the size of one monomer being about
85000 Da (Turpeenniemi-Hujanen et al. 1980, Turpeenniemi-Hujanen et al. 1981,
Myllylä et al. 1988, Rautavuoma et al. 2002). The dimerization of human LH3
has been reported to be dependent on amino acids 541-547, but amino acids from
other sites may also contribute (Heikkinen et al. 2011). However, dimerization
seems not to be a prerequisite for the gycosylation activity of LH3 (Heikkinen et
al. 2011). LHs consist of three domains (Rautavuoma et al. 2002) and the Cterminal
domain contains essential amino acids required for the LH activity
(Pirskanen et al. 1996, Passoja et al. 1998b). The glycosylation activity of LH3 is
located in its N-terminal domain (Rautavuoma et al. 2002, Wang et al. 2002a,
Wang et al. 2002b).
42

Reaction mechanism and catalytically important amino acids
As a member of the 2-oxoglutarate dioxygenase family, LHs, as well as C-P4Hs
and P3Hs, require Fe 2+ , 2-oxoglutarate, molecular oxygen and ascorbate for the
enzymatic reaction, and the reaction mechanism of all these collagen
hydroxylases resembles each other. Fe 2+ , 2-oxoglutarate, O2 and the peptide
substrate are bound in this order during the enzyme reaction. Reaction products
are released in sequence as well, although Fe 2+ remains usually unreleased
between catalytic cycles. The substrate is released first, followed by CO2 and
succinate. (Myllyharju 2005). A simplified representation of the hydroxylation of
a lysine by LH is shown in Fig. 4.
NH CH C
CH2 CH2 CH2 O
CH NH
2 2
HOOC
NH CH
O
C
COOH
+
HOOC
O
O2 Fe
Ascorbate
CH2 CH2 HC OH
+
HOOC
CH2NH2 2+
+ +
Lysine 2-Oxoglutarate
Hydroxylysine Succinate
Fig. 4. A schematic representation of the hydroxylation of lysine by lysyl hydroxylase.
In the lysyl hydroxylation reaction, molecular oxygen binds to the Fe 2+ in the
catalytic site of LH. One oxygen molecule of the O2 is transferred to the hydroxyl
group formed in the lysine, while another oxygen atom is incorporated into the
succinate produced in the reaction. (Kivirikko et al. 1992, Kivirikko &
Pihlajaniemi 1998). Ascorbate is essential for lysyl hydroxylation, although it is
not consumed stoichiometrically in the complete hydroxylation reaction. Collagen
hydroxylases can perform many catalytic cycles in the absence of ascorbate, but it
is needed as an obligatory element to reactivate the enzymes. This is because the
enzymes occasionally catalyze uncoupled decarboxylation of 2-oxoglutarate,
where ascorbate is consumed stoichiometrically. (Myllylä et al. 1984).
The substrate binding site of LH is suggested to be an open, hydrophilic
pocket where acidic amino acids have a significant role in binding (Risteli et al.
2004). At the amino acid level individual human and mouse LH isoforms are
about 60–75% identical, and when all isoforms are compared together the identity
is about 50% (Valtavaara et al. 1997, Valtavaara et al. 1998, Passoja et al. 1998a
Ruotsalainen et al. 1999). The C-terminal region is especially highly conserved in
LHs, containing several critical amino acids that are found in other 2-oxoglutarate
CO 2
43

dioxygenase family members as well (Pirskanen et al. 1996, Myllyharju &
Kivirikko 1997, Valtavaara et al. 1997, Passoja et al. 1998a, Passoja et al. 1998b,
Valtavaara et al. 1998, Vranka et al. 2004). It is assumed that Fe 2+ is bound to LH
with three linkages (Kivirikko & Pihlajaniemi 1998) via the amino acids histidine
638, aspartate 640 and histidine 690 in processed human LH1 (Pirskanen et al.
1996). Many divalent cations, such as Zn 2+ , function as competitive inhibitors
with respect to Fe 2+ (Kivirikko & Pihlajaniemi 1998). In addition, arginine 700
(in processed human LH1) is a positively charged residue that ionically binds the
C5 carboxyl group of 2-oxoglutarate (Hanauske-Abel & Günzer 1982, Passoja et
al. 1998b). LHs also contain several conserved cysteine residues and seven of
them are shown to be important for the enzyme activity of LH1 (Yeowell et al.
2000a).
There is only a limited amount of data available about catalytically important
amino acids for the glycosyltransferase activity of LH3. However, indications
about critical amino acids have been obtained by comparing the cDNA sequences
of the human and mouse LH3s to the only LH of Caenorhabditis elegans, which
is reported to possess both LH and glycosylation activity (Norman & Moerman
2000, Wang et al. 2002a, Wang et al. 2002b). 29 amino acids are conserved in
these sequences of which cysteine 144 and leucine 208 in the human sequence
were shown to be important for the LH3 GGT activity in all three species (Wang
et al. 2002b). The cysteine but not the leucine residue is also important for the GT
activity of the enzyme (Wang et al. 2002a). In addition, a conserved motif at
amino acids 187–191 in the human sequence containing important aspartate
residues is suggested to be a potential binding site for Mn 2+ in the GGT active site
(Wang et al. 2002a, Wang et al. 2002b). The aspartate-containing short site may
represent a DxD motif that is commonly found in many glycosyltransferases
(Unligil et al. 2000, Gastinel et al. 2001, Wang et al. 2002b).
Substrate specificity
The hydroxylysine content of collagens varies between collagen types, tissues and
normal physiological states of the tissue but also in certain disorders. These
observations have indicated some kind of substrate specificity as well as tissue
specificity of the LH isoenzymes. (Ihme et al. 1984, Kivirikko et al. 1992, Bank
et al. 1999, Eyre et al. 2002). The minimal sequence requirement for the function
of LHs is the X-Lys-Gly triplet (Kivirikko et al. 1992). All isoforms have been
shown to hydroxylate synthetic peptides representing collagenous domains of
44

type I and IV collagens (Takaluoma et al. 2007) and they seem not to have strict
sequence requirements (Risteli et al. 2004, Takaluoma et al. 2007). In addition, all
isoenzymes are able to hydroxylate lysines in the triple helical domain of
recombinant type I procollagen produced in insect cells, but LH2 is the only
isoform capable of hydroxylating the telopeptide lysines in this system
(Takaluoma et al. 2007). However, overexpression studies with LH2 in fibroblasts
demonstrated that LH2 may not be able to hydroxylate lysines in the triple helical
area (Wu et al. 2006). Hydroxylysine residues in collagen telopeptides are found
in X-Hyl-Ser and X-Hyl-Ala sequences, which differ from the collagenous
X-Hyl-Gly sequence (Myllyharju 2005). However, these X-Hyl-Ser and
X-Hyl-Ala sequences seem not to be the only requirement for the hydroxylation
by LH2 as in in vitro studies the enzyme was not able to hydroxylate a short
synthetic peptide representing the type I collagen telopeptide, but rather required
a longer substrate as demonstrated in the insect cell coexpression experiments
with a procollagen I chain (Takaluoma et al. 2007).
Despite the lack of a strict substrate specificity, several observations suggest
preferential roles for each LH isoenzyme. LH2 hydroxylates the telopeptide
lysines, but may also hydroxylate helical lysines (Uzawa et al. 1999, Mercer et al.
2003, van der Slot et al. 2003, Pornprasertsuk et al. 2004, Takaluoma et al. 2007),
LH1 hydroxylates helical lysines (Eyre et al. 2002, Steinmann et al. 2002) and
LH3 glycosylates hydroxylysines (Heikkinen et al. 2000, Rautavuoma et al. 2002,
Wang et al. 2002a, Rautavuoma et al. 2004, Ruotsalainen et al. 2006, Salo et al.
2006a), but its exact role in lysyl hydroxylation is not known. LH3 is capable of
hydroxylating helical lysines of type I procollagen at least in a recombinant
overexpression system (Takaluoma et al. 2007), and it is suggested to be
responsible for the hydroxylation of lysine residues that are further glycosylated
by the same enzyme (Sipilä et al. 2007). LH1 has been suggested to have some
telopeptide hydroxylation activity (Mercer et al. 2003, Walker et al. 2004a), but
opposite results have also been obtained (Royce & Barnes 1985, Takaluoma et al.
2007). In addition, analysis of EDS VIA patients with null LH1 mutations have
revealed some hydroxylation of collagenous domains indicating that LH2 or LH3
or both could function as helical LHs as well (Eyre et al. 2002).
LH3 is the only isoenzyme that is reported as likely to be able to modify also
native collagen molecules in the extracellular space (Salo et al. 2006a). However,
in vitro studies on skin and tendon collagens have suggested that the native type I
collagen is not the most appropriate substrate for LH3, which rather prefers
collagen polypeptide chains (Sricholpech et al. 2011).
45

2.5 Human diseases related to lysyl hydroxylases and animal
models
The kyphoscoliotic form of EDS (EDS VI) was for a long time the only disorder
connected to LHs. Studies of the syndrome have provided important information
about the function of LHs and the importance of lysyl hydroxylation in collagens
(Yeowell & Walker 2000, Steinmann et al. 2002). Currently mutations are found
in all LH isoforms leading to severe connective tissue disorders (Yeowell &
Walker 2000, van der Slot et al. 2003, Salo et al. 2008). In addition,
overexpression of the long form of LH2 is connected to fibrosis (van der Slot et al.
2003, van der Slot et al. 2004).
2.5.1 Ehlers-Danlos type VI
The kyphoscoliotic type of EDS or EDS VI is an autosomal recessive, inheritable
connective tissue disorder. EDS VI patients are clinically characterized by
articular hypermobility, kyphoscoliosis, hyperextensible skin, bad scarring, easy
bruising, and muscular hypotonia. (Yeowell & Walker 2000, Steinmann et al.
2002). Patients suffer also from ophthalmological complications and have an
increased risk of spontaneous arterial ruptures (Wenstrup et al. 1989, Steinmann
et al. 2002). Two different forms of EDS VI have been identified. In EDS VIA the
biochemical basis for this disease is a deficiency of LH1 activity, caused by
mutations in the PLOD1 gene. Therefore, the amount of triple helical
hydroxylysine is decreased and thus collagen cross-linking is changed (see Figs. 3
and 5). (Yeowell & Walker 2000, Steinmann et al. 2002). EDS VIB patients have
the same phenotype as EDS VIA patients, but the LH activity is normal
(Steinmann et al. 2002, Walker et al. 2004b) and currently the molecular
pathology of this form is unknown. Moreover, a mutation in the zinc transporter
gene was recently reported to cause EDS VI-like symptoms with abnormal
collagen hydroxylation and urinary pyridinoline levels but with normal LH and
C-P4H activities in cell extracts (Giunta et al. 2008).
Genetic background of Ehlers-Danlos syndrome type VIA
Various different mutations in the PLOD1 gene have been identified from EDS
VIA patients. Mutations consist of a single base pair (bp) mutation to large
duplications and deletions. (Yeowell & Walker 2000, Steinmann et al. 2002). The
46

most common mutation connected to Alu-Alu recombination is a large duplication
of seven exons (Pousi et al. 1994, Heikkinen et al. 1997, Brinckmann et al. 1998,
Yeowell et al. 2000b, Giunta et al. 2005a). Another predominant form is a
nonsense mutation in exon 14 (Yeowell & Walker 1997, Walker et al. 1999,
Yeowell & Walker 1999b, Pousi et al. 2000, Yeowell et al. 2000b). Several
compound heterozygote mutations have been reported as well (Yeowell & Walker
2000). The first found compound heterozygote patient also confirmed the
autosomal recessive inheritance of the syndrome (Ha et al. 1994). LH1 has
several alternative splicing pathways that can either bypass or compensate for
mutations and thus cells may sustain partial LH1 activity (Yeowell & Walker
2000). According to one study small amounts of differentially spliced LH1
transcripts are also detected from normal cells (Heikkinen et al. 1999), although
LH2 is considered to be the only LH isoform having splicing variants (Yeowell &
Walker 1999a).
Molecular biology
The LH activity in patients with EDS VIA is markedly reduced (Yeowell &
Walker 2000, Steinmann et al. 2002). However, activity levels vary significantly
between tissues, individuals and even between assays performed (Steinmann et al.
2002). There are differences in hydroxylation levels also between collagen types
as type I and III collagens are severely affected, while collagen types II and V are
basically normally hydroxylated in EDS VIA patients, generally (Ihme et al. 1984,
Eyre et al. 2002). Although the collagen is normally secreted from cells, the
decreased hydroxylation and thus glycosylation of collagens can be seen as
increased electrophoretic mobility in SDS-PAGE analysis (Jarish et al. 1998,
Walker et al. 1999, Giunta et al. 2005a, Giunta et al. 2005b). In addition, the
solubility rate of tissue collagens in denaturing solvents is suggested to be related
to the level of hydroxylation (Steinmann et al. 2002).
The degree of collagen hydroxylation is suggested to correlate with the
clinical status of the EDS VIA patient (Ihme et al. 1984). The skin for instance is
severely affected in the patients, while in cartilage the consequences are milder
(Eyre & Glimcher 1972, Pinnell et al. 1972, Eyre et al. 2002). The LH activity
determined from skin fibroblasts of patients is usually less than 25% of controls
(Yeowell & Walker 2000, Steinmann et al. 2002). In EDS VIA patients the lysine
residues in collagen telopeptides are normally hydroxylated, while the lysines in
the triple helical area are underhydroxylated, affecting the collagen cross-linking
47

and thus the characteristics of collagenous structures (see Fig. 5) (Steinmann et al.
2002, Eyre & Wu 2005). Although the pyridinoline levels in skin are generally
low, in normal skin the mature HP cross-links are suggested to be four times more
prevalent than LP cross-links, whereas in EDS VIA patients HP cross-links are
absent (see Figs. 3 and 5) (Pasquali et al. 1995, Pasquali et al. 1997, Knott &
Bailey 1998). The changed excretion pattern of LP and HP cross-links derived
from degraded collagens especially from tissues such as bone increases the LP/HP
ratio in the urine, which can be determined and used as a diagnostic marker
(Pasquali et al. 1994, Pasquali et al. 1995, Steinmann et al. 1995).
In EDS VIB patients the hydroxylation level of helical lysines in skin
fibroblasts is decreased almost as much as in EDS VIA (Steinmann et al. 2002,
Uzawa et al. 2003). Interestingly, some of the EDS VIB patients are reported to
have decreased LH2 or LH3 mRNA levels (Walker et al. 2004b).
Diagnosis and treatment of Ehlers-Danlos syndrome type VIA
During childhood EDS VIA patients are usually suspected to have a
neuromuscular disease because of the hypotonia and delayed gross motor
development, and the correct diagnosis is thus often obtained late (Steinmann et
al. 2002, Giunta et al. 2005a). In addition to clinical symptoms, the disease can be
diagnosed by measuring LH activity in cultured fibroblasts or LP/HP ratio in the
urine, as indicated earlier (Pasquali et al. 1994, Pasquali et al. 1995, Steinmann et
al. 1995). One study has introduced a method to confirm the diagnosis from RNA
and DNA samples (Giunta et al. 2005a) and successful prenatal exclusions of the
syndrome are also reported (Yeowell & Walker 1999b, Yeowell et al. 2000b).
Although the syndrome is not treatable, some patients have had some benefit
from pharmacological doses of ascorbate (vitamin C) (Dembure et al. 1987). It
has been shown to increase the collagen synthesis and LH activity both in normal
and EDS VIA fibroblasts (Dembure et al. 1987, Yeowell et al. 1995, Pasquali et
al. 1997, Yeowell et al. 1997). In addition, ascorbate is shown to affect also the
cross-link content in fibroblasts (Pasquali et al. 1997).
48

Normal
EDS VIA
hLys
hHyl
tHyl
tHyl
Fig. 5. Simplified illustration of collagen cross-linking in a stiff tissue, such as bone, in
a normal situation and in Ehlers-Danlos syndrome type VIA and Bruck syndrome.
2.5.2 Bruck syndrome 1 and 2
tHyl
tHyl
tLys
tLys
hHyl
Bruck syndrome (BS) is an extremely rare, autosomal recessive connective tissue
disorder with features typical for osteogenesis imperfecta (OI). About 30 patients
have been reported to suffer from BS. BS patients typically have osteoporosis,
fragile bones, congenital contractures of joints, progressive kyphoscoliosis and
short stature, but they usually have normal dentinogenesis and no hearing loss.
BS can be best distinguished from OI by the presence of congenital joint
contractures and the absence of hearing loss. (Viljoen et al. 1989, Brenner et al.
1993, McPherson & Clemens 1997, Brady & Patton 1997, Breslau-Siderius et al.
1998, Blacksin et al. 1998, Leroy et al. 1998, Bank et al. 1999, van der Slot et al.
2003, Ha-Vinh et al. 2004, Datta et al. 2005, Berg et al. 2005, Mokete et al. 2005,
Yapicioğlu et al. 2009, Shaheen et al. 2010, Kelley et al. 2011). Typical features
Bruck syndrome
49

of OI include increased bone fragility and short stature, impaired hearing,
dentinogenesis imperfecta and hypermobile joints (for a review, see Basel &
Steiner 2009). Most OI cases are caused by mutations in the genes coding for the
pro-α1 or pro-α2 chains of type I collagen, but mutations are detected also in
members of the protein complex responsible for collagen prolyl 3-hydroxylation
and recently also from a SERPINH1 gene that encodes the collagen chaperone
Hsp47 (Basel & Steiner 2009, Christiansen et al. 2010).
Genetic background
BS is a genetically heterogeneous disease. Mutation analysis of BS families has
led to two different linkages. BS type 1 (BS1) is linked to chromosome 17 and BS
type 2 (BS2) to the PLOD2 gene in chromosome 3. (Bank et al. 1999, van der
Slot et al. 2003, Ha-Vinh et al. 2004). Three different point mutations have been
found so far from exon 17 of the PLOD2 gene that encodes LH2 (van der Slot et
al. 2003, Ha-Vinh et al. 2004), while the genetic background and the molecular
mechanism of BS1 have been elusive for a long time. Recently, mutations in
FKBP10, a gene encoding FK506 binding protein 10 (FKBP10, also known as
FKBP65) were suggested to cause BS1 or a subtype of OI (Alanay et al. 2010,
Alanay & Krakow, 2010, Shaheen et al. 2010, Kelley et al. 2011). The FKBP10
gene is located in chromosome 17, but not at the location suggested earlier to
cause BS1 (Bank et al. 1999, Kelley et al. 2011).
Molecular biology
BS subtypes can be distinguished according to the genetic defect (Bank et al.
1999, van der Slot et al. 2003). BS2 is caused by a defective telopeptide LH,
LH2(long) (Bank et al. 1999, van der Slot et al. 2003, Ha-Vinh et al. 2004,
Takaluoma et al. 2007), while the molecular basis of BS1 is still unknown. In BS
patients the lysine residues in the triple helical area of type I collagen are
normally hydroxylated, while the telopeptide lysines are underhydroxylated (see
Fig. 5) (Bank et al. 1999). The hydroxylation state of the telopeptide lysines
affects the cross-linking profile of collagen molecules (Eyre & Wu 2005). In the
bone of the patients, for instance, the amount of telopeptide hydroxylysinederived
mature pyridinoline cross-links, HP and LP, in type I collagen is reduced
(see Figs. 3 and 5). In contrast, the amount of lysinonorleucine, a derivative of an
immature, difunctional cross-link originated from unhydroxylated telopeptide
50

lysines via the allysine route, is reported to be increased 5-fold. Interestingly,
cross-linking of type I collagen has tissue-specific features as in contrast to the
bone, the type I collagen in the ligament of a BS1 patient is normally cross-linked
and thus hydroxylated. (Bank et al. 1999). The decreased number of stable
pyridinoline cross-links and increased collagen turnover in bone can be analyzed
from urine samples (Ha-Vinh et al. 2004).
Morphological analyses of severely affected bone tissue of BS patients have
revealed wormian bones in the skull, cystic changes at old fracture sites and a
mixed callus as well as normal or immature mineralization in bone and osteoids
that vary in thickness (Viljoen et al. 1989, Brenner et al. 1993, McPherson &
Clemens 1997, Brady & Patton 1997, Blacksin et al. 1998, Leroy et al. 1998,
Ha-Vinh et al. 2004, Datta et al. 2005, Mokete et al. 2005). At the cell level
osteoblasts have swollen mitochondria and dilated ER and the diameter of
collagen fibrils in osteoids vary (Brenner et al. 1993). The cross-linking pattern is
suggested to have an important role in the correct mineralization of calcified
tissues (Yamauchi & Katz 1993, Knott et al. 1997, Bank et al. 1999,
Pornprasertsuk et al. 2005). Unsurprisingly, the mineral content in BS bone is
shown to be reduced and the size of hydroxyapatite crystals is increased (Brenner
et al. 1993). Thus, it is suggested that the fragility of BS bones is likely to be due
to a changed cross-linking pattern together with inadequate mineralization (Bank
et al. 1999). Biochemical studies show also increased extractability of type I
collagen in BS bone tissue (Brenner et al. 1993). However, the amounts of type I
and III collagens are reported to be normal and also their structure and synthesis
are seemingly normal (Brenner et al. 1993, McPherson & Clemens 1997,
Breslau-Siderius et al. 1998, Leroy et al. 1998, Bank et al. 1999, Kelley et al.
2011), which differs from OI where collagens are usually overmodified due to a
delay in the triple helix formation (Engel & Prockop 1991, Raghunath et al. 1994).
The gene FKBP10 and Bruck syndrome
The mutations in the FKBP10 gene, connected to BS1, include a homozygous 1
bp duplication, a 35 bp deletion and two different compound heterozygous
mutations, all of them located in the coding region of FKBP10, while no
mutations have been found from the PLOD2 gene in these patients (Kelley et al.
2011). FKBP10 is a member of immunophilins that are known to function as
intracellular receptors capable of binding and mediating the action of certain
immunosuppressant drugs (Coss et al. 1995; for a review, see Göthel & Marahiel
51

1999). Like other immunophilins, FKBP10 is suggested to have some
peptidylprolyl cis-trans isomerase activity, which can be important in the folding
of collagens and tropoelastin monomers (Coss et al. 1995, Davis et al. 1998,
Ishikawa et al. 2008). Interestingly, mutations in another FK506 binding protein,
FKBP14 (FK506 binding protein 14) are reported to cause a new variant of EDS,
a disorder where patients share many clinical features particularly of EDS VIA
and Ullrich congenital muscular dystrophy. In contrast to EDS VIA, these patients
have normal urinary LP/HP ratio. (Baumann et al. 2012).
Like LHs, FKBP10 is located in the lumen of the ER and it acts as a
chaperone of folded and unfolded collagen (Myllylä et al. 1991, Kellokumpu et al.
1994, Davis et al. 1998, Patterson et al. 2000, Ishikawa et al. 2008). It is also
suggested to have a role in the chaperoning of tropoelastin (Davis et al. 1998,
Patterson et al. 2000). Studies with FKBP10 and collagen together have
suggested that FKBP10 could function as a complex with other proteins to
process collagen (Zeng et al. 1998, Ishikawa et al. 2008). Collagens from patients
with a mutation in FKBP10 are not overmodified (Alanay et al. 2010). This was
observed also with an OI patient who had a mutation in SERPINH1 (Christiansen
et al. 2010), when typically in OI the overmodification is quite common (Marini
et al. 2010). In contrast to the other FKBP family members that are expressed
ubiquitously, FKBP10 mRNA is detected in the mouse lung, spleen, heart, brain
and testis (Coss et al. 1995). Another study shows that the expression of FKBP10
in adult mice is absent or minimal, but in developing mice the expression was
seen in the smooth muscle layer of the airways and pulmonary vessels and at the
mRNA level in tissues such as aorta and kidney, indicating developmental
regulation of the protein (Patterson et al. 2000). However, the possible connection
between FKBP10 and LH2 is still unknown.
Treatment of Bruck syndrome
There is no treatment for BS. However, bisphosphonate cyclic pamidronate has
been used as a pharmacotherapy against bone fragility (Ha-Vinh et al. 2004,
Mokete et al. 2005, Andiran et al. 2008, Yapicioğlu et al. 2009, Shaheen et al.
2010). Bisphosphonates are suggested to function as inhibitors of bone resorption
by interfering with the function of osteoclasts (for a review, see Russell 2006). In
a reported BS patient the treatment affected fracture incidence and bone mineral
density (Andiran et al. 2008). However, a long-term follow-up study of a BS
52

patient, not receiveing treatment, revealed that the prevalence of fractures
decreases after maturation of the bone (Mokete et al. 2005).
2.5.3 Fibrotic diseases and cancer
Fibrotic diseases
Inflammatory reaction in the tissue is a normal response to damage caused by
numerous factors from bacterial infections to mechanical damage. The
programmed phases of inflammation protect the tissue and enable recovery.
Sometimes impaired regulation of the inflammatory reaction causing a chronic
inflammation can lead to fibrosis, which is characterized by excessive
accumulation of ECM and especially collagen. (Rubin & Farber 1999,
Trojanowska et al. 1998, Hunzelmann & Brinckmann 2011). In soft tissues
collagen cross-links are normally formed from unhydroxylated telopeptide lysines,
but in fibrotic lesions the cross-linking pattern of collagen is altered (Brinckmann
et al. 1996, Knott & Bailey 1998, Brinckmann et al. 1999, Brinckmann et al.
2001). The increased amount of pyridinoline cross-links (see Fig. 3) found in
fibrotic tissue is caused by overhydroxylation of the telopeptide lysine residues
accompanied by the LO reaction (van der Slot et al. 2003, van der Slot et al. 2004,
Levental et al. 2009, Saito & Marumo 2010). The resulting collagenous structures
are highly resistant to normal remodelling and degenerative procedures such as
degradation by MMPs (Trojanowska et al. 1998). Collagen matrices having an
increased amount of pyridinoline cross-links are shown to be less susceptible to
MMP-1 degradation, for instance (van der Slot-Verhoeven et al. 2005). As
LH2(long) is the only LH isoform significantly upregulated in fibrosis and it is
known to be a telopeptide LH, it has been suggested to be responsible for
overhydroxylation of the telopeptide lysines and thus increased pyridinoline
cross-link formation (van der Slot et al. 2003, van der Slot et al. 2004, Takaluoma
et al. 2007). Although overexpression of all LH isoforms is shown to increase
collagen synthesis (Mercer et al. 2003), overexpression of only LH2 was shown
to both increase the collagen synthesis and change the cross-linking pattern (Wu
et al. 2006). Thus, LH2(long) is an important factor in fibrosis and prevention of
its upregulation together with LO is important from the clinical point of view. The
role of LH2(short) seems to be insignificant in fibrotic conditions (van der Slot et
al. 2003, van der Slot et al. 2004). Although fibrosis and increased pyridinoline
53

cross-links are connected, direct connection of this phenomenon to the resistance
of collagenous structures to remodeling is not confirmed.
As a consequence of the inflammatory response the fibrotic process is
connected to many inflammatory cytokines and growth factors (Trojanowska et al.
1998, Hunzelmann & Brinckmann 2011). TGF-β is an important and common
inflammatory factor and it is shown to increase the expression of LH2(long) and
thus the formation of pyridinoline cross-links (Trojanowska et al. 1998, van der
Slot et al. 2005, Hunzelmann & Brinckmann 2011). In addition, TGF-β is known
to increase generally the expression of ECM proteins and inhibitors of MMPs
(Hunzelmann & Brinckmann 2011). Besides TGF-β, other cytokines such as
interleukin 4 and activin A are also shown to induce LH2(long) expression (van
der Slot et al. 2005, Brinckmann et al. 2005). However, the effect of at least TGFβ
is presumably not direct, but it is mediated by other profibrotic factors that are
rapidly upregulated via TGF-β induction (van der Slot et al. 2005). Thus,
knowledge about factors contributing to increased expression of the LH2(long)
form, including knowledge about the inflammatory response and alternative
splicing, are important in the prevention of fibrosis.
Treatment of fibrosis
Certain organophosphate compounds inhibit the function of LH and have been
earlier suggested to be valuable for testing as antifibrotic drugs (Samimi & Last
2001). An antihypertension drug minoxidil has been suggested to be a potential
drug for fibrosis as it decreases LH activity (Murad & Pinnell 1987, Hautala et al.
1992b). However, minoxidil has consequently been found to be ineffective as it
decreases LH1 expression rather than LH2 or LH3 expression, and thus changes
the LP/HP ratio but does not decrease the total amount of pyridinoline cross-links
(Zuurmond et al. 2005).
Fibrosis, hypoxia and cancer
In the chronic phase of inflammation, fibrosis is accompanied by hypoxia (Rubin
& Farber 1999, Hunzelmann & Brinckmann 2011). In hypoxic conditions C-P4H,
LH1 and LH2, but not LH3, have been shown to be upregulated leading to
increased accumulation of collagen (Lal et al. 2001, Denko et al. 2003, Hofbauer
et al. 2003, Scheurer et al. 2004, Brinckmann et al. 2005, van Vlimmeren et al.
2010), while the expression of collagen polypeptides themselves has been
54

eported to be either upregulated, downregulated or not affected (Hofbauer et al.
2003, Scheurer et al. 2004, Brinckmann et al. 2005, van Vlimmeren et al. 2010).
Hypoxia is characteristic also in cancer, and ECM remodeling and changes
in tissue stiffness are also characteristic in tumor formation (for reviews, see
Butcher et al. 2009, Semenza 2011). Moreover, a fibrotic tumor predicts
malignancy (Colpaert et al. 2003, Levental et al. 2009). Hypoxia-inducible factor
1 (HIF-1) is a transcription factor that has been reported to influence cancer
progression (Semenza 2011). Mouse and human PLOD1 and PLOD2 genes have
been shown to have a putative hypoxia responsive element in their promoter
region suggesting that the regulation of these genes may be mediated by HIF
(Hofbauer et al. 2003, Scheurer et al. 2004, Brinckmann et al. 2005). In fact,
PLOD2 expression is elevated in cancer cells (Lal et al. 2001, Dong et al. 2005,
Evens et al. 2010, Rajkumar et al. 2011, Noda et al. 2012), and the elevation is
reported to be associated with increased HIF-1α gene expression (Evens et al.
2010). Upregulation of several collagen-modifying enzymes including LO in
hypoxia could indicate their important role in the ECM remodeling during early
phases of tumor angiogenesis, for instance (Denko et al. 2003, Scheurer et al.
2004, Levental et al. 2009). HIF-1α-induced upregulation of PLOD2 in cancer is
suggested to be associated with poor prognosis (Dong et al. 2005).
2.5.4 A disease caused by mutations in LH3 and LH3 null mice
LH3 was for a long time the only LH isoform that was not connected to any
human diseases, although mice having an inactivated Plod3 gene were known to
die during embryogenesis (Rautavuoma et al. 2004, Ruotsalainen et al. 2006).
Recently, however, LH3 mutations were connected to two different connective
tissue disorders. A patient with severe symptoms overlapping several connective
tissue disorders was shown to have mutations in the PLOD3 gene (Salo et al.
2008). In addition, one epidermolysis bullosa family was connected to PLOD3
mutations (Savolainen et al. 1981, Risteli et al. 2009).
Human disorders connected to LH3
As indicated earlier, some connective tissue disorders can be detected based on
the collagen turnover products in urine or serum samples (Knott & Bailey 1998,
Eyre & Wu 2005, Saito & Marumo 2010). The first patient with a defective LH3
enzyme was also detected via cross-link analysis from urine. A glycosylated
55

derivate of the pyridinoline cross-link was absent in the sample and the amount of
glucosylgalactosyl hydroxylysine was decreased. (Salo et al. 2008). As LH3 is
known to have collagen glycosylation activity (Heikkinen et al. 2000,
Rautavuoma et al. 2002, Wang et al. 2002a, Rautavuoma et al. 2004,
Ruotsalainen et al. 2006, Salo et al. 2006a) besides the two GT25 family
members that possess collagen GT activity (Schegg et al. 2009), the PLOD3 gene
of the patient was analyzed and found to have a compound heterozygote mutation
(Salo et al. 2008). One of the mutations was located in the region responsible for
the glycosylation activities and the other in the LH activity region (Pirskanen et al.
1996, Heikkinen et al. 2000, Rautavuoma et al. 2002, Wang et al. 2002a, Wang et
al. 2002b, Salo et al. 2008). However, despite the locations of the mutations, both
reduced all three activities i.e. LH, GT and GGT. The mutations lead to a
decreased concentration of the LH3 protein and reduced LH and glycosylation
activities, which are suggested to be the molecular basis for the disorder by
affecting collagen synthesis. (Salo et al. 2008). Based on these findings and the
fact that LH3 is known to be essential for the development of the mouse
(Rautavuoma et al. 2004, Ruotsalainen et al. 2006), it is not surprising that the
patient has severe symptoms such as osteopenia, joint contractures, arterial
ruptures, skin blistering, deafness and myopia that are all connected to many
different collagen disorders (Salo et al. 2008). In addition, the reported patient
had a stillborn sibling who was reported to have several abnormalities. According
to autopsy, the intrauterine death at 28 weeks of gestation was suggested to be
caused by cerebral vascular rupture. (Salo et al. 2008).
In addition to the LH3 compound heterozygous patient, a large family
diagnosed to have an epidermolysis bullosa simplex was discovered to transcribe
only one LH3 allele (Savolainen et al. 1981, Risteli et al. 2009). The patients
were reported to develop blisters extremely easily and blisters were frequently
observed particularly in the hands and feet, which are regularly under mechanical
stress. The healing occurred without scarring. The first blisters were detected at
birth or shortly after that, but the symptoms usually became milder with age.
(Savolainen et al. 1981). The decreased amount of LH3 in these patients led to
decreased glycosylation activity, which in turn was suggested to affect the
organization of the ECM (Savolainen et al. 1981, Risteli et al. 2009). Four
polymorphic changes in the non-coding region of PLOD3 were detected, but the
reason for the allelic silencing remained unknown. However, as a heterozygous
deficiency it demonstrated the importance of the correct amount of LH3 activity
in humans. (Risteli et al. 2009).
56

LH3 knock-out mice
The currently unnamed, severe ECM disorder caused by LH3 mutations, reveals
the importance of this enzyme for humans, but LH3 has been demonstrated to be
essential also in the mouse (Rautavuoma et al. 2004, Ruotsalainen et al. 2006).
Mouse embryos having an inactivated Plod3 gene die during early development
at E9.5 (Rautavuoma et al. 2004, Ruotsalainen et al. 2006). The deficiency of
LH3 was found to be critical for the synthesis of type IV collagen and thus for the
stability of basement membranes (Rautavuoma et al. 2004, Ruotsalainen et al.
2006). In addition to the defect in the embryos, a possible role of abnormal
function of placental vasculature or the Reichert’s membrane (RM) in the
developmental retardation was not excluded (Rautavuoma et al. 2004,
Ruotsalainen et al. 2006).
The embryonic lethality of LH3 knock-out mice was determined to be caused
by the deficient glycosylation activity rather than LH activity of the LH3 enzyme
(Rautavuoma et al. 2004, Ruotsalainen et al. 2006). At least the GGT activity in
mouse LH3 is critical for mouse development, indicating that LH3 may be the
main molecule responsible for GGT activity in mouse (Ruotsalainen et al. 2006).
Studies with LH3 knock-out embryos and fibroblasts suggested that the deficient
glycosylation of type IV collagen leads to delayed processing, intracellular
accumulation, and delayed secretion as well as increased intracellular degradation
of type IV collagen (Rautavuoma et al. 2004, Sipilä et al. 2007). In addition, the
lack of glycosylation activity but not LH activity of LH3 seemed to prevent the
assembly and thus secretion of type VI collagen. Although the lack of LH activity
of LH3 does not prevent the secretion of type VI collagen, it affects the
distribution of type IV and VI collagens in the examined adult mouse tissues.
(Sipilä et al. 2007). Recent studies with cells have revealed that the impaired
glycosylation activity of LH3 affects fibrillogenesis of the bone type I collagen as
well. Furthermore, GGT activity is suggested to be the main function of LH3 in
osteoblasts. (Sricholpech et al. 2011).
2.5.5 Inactivation of LH in other animal models
The nematode Caenorhabditis elegans and the fly Drosophila melanogaster have
only a single LH, while the zebrafish Danio rerio, like other vertebrates, has all
three forms including both LH2 splicing variants (Norman & Moerman 2000,
Schneider & Granato 2006, Schneider & Granato 2007, Bunt et al. 2010). The
57

single C. elegans LH possesses both lysyl hydroxylase and collagen
glucosyltransferase activity and thus resembles the vertebrate LH3 (Wang et al.
2002a, Wang et al. 2002b). The D. melanogaster LH also contains the residues
that have been shown to be necessary for the glycosyltransferase activity of
vertebrate LH3 (Bunt et al. 2010).
In C. elegans and D. melanogaster the absence of the only LH leads to
lethality during embryogenesis (Norman & Moerman 2000, Bunt et al. 2010). As
in LH3 knock-out mice (Rautavuoma et al. 2004, Sipilä et al. 2007), the enzyme
is required in the processing and secretion of type IV collagen in these organisms
(Norman & Moerman 2000, Bunt et al. 2010). Similarly, LH3 is essential for
collagen secretion in D. rerio (Schneider & Granato 2006).
58

3 Outlines of the present study
Collagens are the main structural components of the ECM in the human body and
LH isoenzymes have an important role in their synthesis. Mutations in the
PLOD1 and PLOD2 genes are known to cause the rare and severe connective
tissue disorders EDS VIA and BS2 in humans, respectively, emphasizing the
importance of these enzymes in a healthy individual. Moreover, earlier studies on
the lack of LH3 in mouse had revealed novel information of its importance during
the embryonic development. Later, LH3 mutations have also been connected to
human disorders. The Plod3 knock-out mouse line demonstrated that LH1 and
LH2 are not able to compensate for the lack of LH3 activity, but otherwise the
exact in vivo roles of LH1 and LH2 have remained mostly elusive. Thus, we
wanted to generate Plod1 and Plod2 knock-out mouse lines to study the
expression patterns and in vivo roles of these isoenzymes, as well as potential
compensatory roles of the other LH isoenzymes. Moreover, the mouse lines were
also expected to serve as the first animal models for EDS VIA and BS, enabling
the study of tissues that were impossible and unethical to study from the human
patients.
The first BS mutation was linked to chromosome 17, but the pathological
mechanism remained unknown. Later, three different mutations were found in the
PLOD2 gene located in chromosome 3. The BS form connected to LH2 mutations
was thereafter called BS2. Although studies of BS2 patients provided clinical
support for the theory that LH2 is the telopeptide LH, the mutations did not
change any of the catalytically critical amino acids of the enzyme and the
molecular mechanism of the disorder remained unknown. We therefore decided to
produce the BS2 mutant and wild-type human LH2(long) polypeptides in insect
cells, purify and analyze them to understand the molecular pathology of the BS2
syndrome in more detail.
59

4 Materials and methods
The materials and methods used in this thesis are summarized in the table below.
Detailed information with references can be found in the original articles I-III.
Table 2. Methods.
Level Method Used in
DNA
Cloning techniques
I
PCR
I, II, III
Southern Blot analysis
I, III
Site-directed mutagenesis
II
RNA
RNA isolation
I, III
RT-PCR
I, III
Quantitative real-time PCR
III
Protein Expression of recombinant proteins in insect cells
II
Purification and characterization of recombinant proteins II
SDS-PAGE, non-denaturing PAGE and Western blotting
LH activity assay using 2-oxo[1- 14 C]glutarate
LH activity assay with [ 14 II
II
C]lysine-labelled protocollagen
I
Metal chelate affinity chromatography
II
Gel filtration
II
N-terminal sequencing
II
Amino acid analysis
I, III
Circular dichroism spectroscopy II
Cell and tissue Cell culture I, II
Preparation and staining of paraffin sections I, III
Immunohistochemistry and immunofluorescence
I, III
Transmission electron microscopy
I, III
Collagen cross-link analysis
I, III
Other Generation of mouse lines I, III
61

5 Results
5.1 LH1 knock-out mice have increased risk of lethal arterial
ruptures (I)
5.1.1 Generation of LH1 null mice and LH1 expression in adult
mouse tissues
Mouse lines lacking LH1 activity were generated to study the role of LH1 in vivo
and to generate a mouse model for EDS VIA. The Plod1 gene was disrupted by a
homologous recombination of a targeting construct to exon 2 (Figure 1A in I).
The in-frame insertion of the LacZNeo cassette led to the deletion of exons 3–6.
Two mouse lines having the same phenotype were generated from ES cells and
backcrossed into the C57BL/6 mouse line. RT-PCR downstream from the site of
recombination confirmed the inactivation of the Plod1 gene (Figure 1C in I). The
genotypes of the offspring obtained by cross-breeding of the heterozygous mice
followed the Mendelian inheritance, but both null lines lacked sufficiently strong
β-galactosidase staining in tissues to study the expression of the Plod1 gene (data
not shown). Therefore, an additional LH1 null mouse line with a gene-trap
insertion cassette containing a β-galactosidase reporter in intron 12 of the Plod1
gene was generated and backcrossed into the C57BL/6 mouse line (not shown).
The LH1 gene-trap mouse line with functional β-galactosidase expression in
tissues had identical phenotypic characteristics with the two other Plod1 -/- mouse
lines. The gene-trap mouse line was used for analysis of Plod1 expression by Xgal
staining.
X-gal staining, representing the Plod1 promoter-driven β-galactosidase
expression, was localized to cells that produce fibrillar collagens, such as
fibroblasts in the lung, skin and tendon (Figure 2A–B and 2D in I). In skin the
papilla of hair follicles and arrector pili muscles were also stained (Figure 2B in I).
In cornea the staining was seen in keratocytes (Figure 2F in I). Intense staining in
Plod1 -/- mouse tissues was observed in the heart, where the strongest staining was
seen in atrial auricles (Figure 2C in I), while in the heart muscle the staining was
weaker (not shown). In the skeletal muscle the staining was observed mostly in
the epimysium (not shown). Moreover, strong LH1 expression was observed in
the bone and cartilage, where the staining was seen in osteoblasts and the
periosteum and in chondrocytes (Figure 2E in I).
63

5.1.2 LH1 null mice have no kyphoscoliosis but are hypotonic and
have increased mortality due to arterial ruptures
LH1 null mice were fertile and most of them had a normal lifespan. The mice
were not observed to suffer from kyphoscoliosis, which is a characteristic feature
of the EDS VIA syndrome (Steinmann et al. 2002), but they were passive and
hypotonic during handling and they also appeared to be slower than wild-type
mice. In walking tests on a rod or on the wall of a plastic mouse cage, LH1 null
mice had notable difficulties to move normally and they got tired easily, while the
wild-type littermates had no difficulties in running fast (Figure 3 in I). These
features were more distinct when the mice became older and their weight
increased.
Although most of the Plod1 -/- mice had a normal lifespan, 9% of the females
and 17% of the males were found dead in their cages in the morning. Premature
deaths occurred suddenly during the most active night-time. In almost all of these
mice autopsy revealed a hemorrhage in the thoracic or abdominal cavity (Table 1
in I). In addition, histological analyses showed aortic ruptures (Figure 4C in I),
and in one case there were also signs of a previous non-lethal injury (Figure 4D in
I). The ruptures were seen in the tunica media, between the external layers of the
elastic lamellae, while the elastic lamellae themselves were normal (Figure 4C–D
and 4F in I). Healthy Plod1 -/- mice showed no distinct histological changes in the
aorta when compared to the wild-type mice (Figure 4A–B in I). X-gal staining
was observed in the cells of the tunica media and adventitia of the aorta in the
LH1 gene-trap mouse line (Figure 4B in I). These cells appeared to be less
organized in LH1 null mice. Echocardiographic analysis indicated that the
structure and function of the heart is normal (not shown). However, transmission
electron microscopy (TEM) analyses of healthy Plod1 -/- mouse aortic walls
revealed degenerative changes in smooth muscle cells, such as vacuolization of
the cells and swollen mitochondria (Figure 4F in I). Some completely degenerated
cells were also seen (Figure 4F in I).
5.1.3 Abnormal collagen fibrils in LH1 null mice
In addition to kyphoscoliosis, typical characteristics of EDS VIA include
hyperextensible, fragile skin and abnormal scarring, for instance (Steinmann et al.
2002). The Plod1 -/- mice appeared to have normal skin, and no hyperextensibility
or abnormal scarring was observed. Histological analyses of the skin were also
64

normal (not shown). However, TEM analysis from the skin revealed alterations in
collagen fibrils. The mean collagen fibril diameter in the skin of Plod1 -/- mice was
increased when compared to that of wild-type (91 nm and 62 nm, respectively)
and also the variation between fibril diameters was larger in LH1 null mice
(Figure 5E–H in I). Similar observations were made from aorta samples, where
mean fibril diameter in the Plod1 -/- samples was 67 nm and in the wild-type
samples 44 nm (Figure 5A–D in I). In addition to larger variability in the fibril
diameters, LH1 null samples were also observed to have fibrils with irregular
shape (Figure 5B and 5F in I). In the aorta, also the diameter within a fibril varied
in the Plod1 -/- samples (not shown) and the null skin samples contained degrading
fibrils (Figure 5F in I). Collagen fibrils in the cornea were normal (not shown).
5.1.4 Decreased LH activity in LH1 null mice affects the
hydroxylysine content in various tissues differently
One diagnostic marker of EDS VIA is decreased LH activity in cultured skin
fibroblasts (Steinmann et al. 2002). Thus, LH activity was determined from
homogenized skin samples of the LH1 null mice and also from homogenized
aorta, which was found to be a structurally weak tissue in the mutant mice as well
as in EDS VIA patients (Steinmann et al. 2002). The method where [ 14 C]-lysine
labelled non-hydroxylated protocollagen is used as a substrate (Kivirikko &
Myllylä 1982) determines total LH activity in samples comprising the activities of
LH1, LH2 and LH3. LH activity in the skin samples of Plod1 -/- mice was
decreased to 36% of that in the wild-type samples (Table 2 in I). No reports are
available about the LH activity in the aorta of EDS VIA patients. It was thus quite
surprising that the residual LH activity in the Plod1 -/- aorta was as high as 44% of
that in the wild-type sample (Table 2 in I). The activity in the skin of the LH1
heterozygous mice was about 55% and in the aorta about 60% of that in the wildtype
samples (Table 2 in I).
Total hydroxylysine content determined by reverse-phase high-pressure liquid
chromatography (HPLC) was decreased in all the Plod1 -/- tissues studied when
compared to the wild-type tissues (Table 3 in I). The hydroxylysine content,
expressed as per triple helix, varies markedly between tissues as each tissue
contains different amounts of different collagen types and different collagen types
contain variable amounts of hydroxylated lysines. The highest amount of
hydroxylysine in wild-type samples was observed in the lung (55 residues / triple
helix) (Table 3 in I). Also in the Plod1 -/- lung samples the level was relatively high,
65

eing about 85% of that of the wild-type (Table 3 in I). The lowest amount of
hydroxylysine in the wild-type samples was detected in the skin with 21 residues
per triple helix (Table 3 in I). In Plod1 -/- skin the corresponding value was only
22% of that of the wild-type, being the most dramatically affected tissue studied
(Table 3 in I). The hydroxylysine content was markedly decreased also in the
Plod1 -/- tail tendon and cornea, being 24% and 30% of the wild-type values,
respectively, while the aorta, femur and heart auricle were less affected
(hydroxylysine content 63%, 75% and 69%, respectively, of the wild-type values)
(Table 3 in I).
5.1.5 Abnormal hydroxylation of triple helical lysine residues affects
the cross-linking pattern in the LH1 null mouse tissues
As changes in the hydroxylation of certain lysine residues in collagens affect
cross-link formation and as in EDS VIA patients the cross-linking pattern is
changed (Steinmann et al. 2002), mature cross-links were studied from wild-type
and Plod1 -/- mice. Similar to the hydroxylysine levels, also the LP and HP content
varies markedly between tissues (Table 4 in I). The highest levels of both mature
cross-links in the wild-type mouse were observed in the aorta, the amounts of HP
and LP being 0.71 and 0.26 residues per triple helix, respectively (Table 4 in I).
The amount of HP was decreased in all LH1 null tissues studied, and the most
dramatic change was seen in the aorta, where the HP content was 28% of that of
the wild-type sample (Table 4 in I). The lowest HP and LP levels in wild-type
tissues were observed in the tail tendon (0.051 and 0.006 residues / triple helix,
respectively) and cornea (0.051 and 0.016 residues / triple helix) (Table 4 in I). In
the cornea of LH1 null mice the HP amount was only 33% of that of the wild-type,
but in the tail tendon the relative amount was 59% (Table 4 in I). In Plod1 -/- lung
the amount of HP was decreased markedly (34% of the wild-type), while the
femur was slightly less affected (47% of the wild-type) (Table 4 in I). Although
only certain hydroxylysine residues in collagens participate in the cross-link
formation, it was interesting that there was no correlation between the changes in
hydroxylysine and HP levels (Tables 3 and 4 in I). For example, in the tail tendon
the hydroxylysine level was 24% of the wild-type tissue, but the amount of HP
was as high as 59% of the wild-type value and in the lung the corresponding
percentages were 86% and 34%, respectively. In human skin the levels of mature
LP and HP cross-links are generally low (Knott & Bailey 1998, Avery & Bailey
2005) and they were not detectable in the mouse skin either.
66

As in the EDS VIA patients, the LP level in the Plod1 -/- mouse tissues was
increased in comparison to the corresponding wild-type tissues (Table 4 in I). The
increase varied from about 5-fold to about 60-fold when compared to the wildtype
(Table 4 in I). As a consequence of the decreased HP levels and increased LP
levels, the LP/HP ratio was also increased dramatically in the LH1 null tissues
(Table 4 in I). In wild-type tissues the LP/HP ratio varied between 0.12 and 0.39,
whereas the values in the Plod1 -/- mouse tissues ranged from 3.8 to 12.5, the
smallest increase being observed in the aorta and the highest in the tail tendon
(Table 4 in I). Also the total amount of pyridinoline cross-links (HP + LP) was
increased in all LH1 null mouse tissues studied (Table 4 in I). The increase varied
between tissues, being lowest in the aorta (about 140% of the wild-type) and
highest in the tail tendon, where the rise was about 720% of the wild-type sample.
5.2 Analysis of the molecular level consequences of Bruck
syndrome mutations by the use of recombinant LH2 proteins (II)
5.2.1 Expression of wild-type and recombinant LH2(long)
polypeptides with Bruck syndrome mutations in insect cells
and their purification
BS2 is caused by mutations in LH2 (van der Slot et al. 2003). To study the cause
of BS2 at the molecular level, the currently known point mutations leading to
LH2 amino acid substitutions R594H, G597V, and T604I were introduced into the
pAcgp67A-LH2(long) baculovirus transfer vector by site-directed mutagenesis
(Figure 1 in II) (Krol et al. 1996, Rautavuoma et al. 2002). The baculoviral signal
peptide gb67 directs the synthesized protein for secretion into the culture medium
and the N-terminal histidine tag enables selective purification. Wild-type and the
three BS mutant LH2(long) polypeptides were expressed in High five insect cells
and the secreted recombinant proteins were purified from the culture medium 72
h after infection by metal chelate affinity chromatography followed by histidine
elution.
As the final yield of the purified LH2 mutant polypeptides R594H and
G597V was low, we decided to analyze the behavior of the mutant polypeptides
during purification by taking aliquots from different purification steps for analysis
by SDS-PAGE (Figure 2 in II). Wild-type and T604I LH2(long) bound effectively
to the affinity column, whereas a significant amount of the mutants R594H and
67

G597V were present in the flow-through samples (Figure 2A–B in II). In addition,
the bound R594H and G597V LH2(long) polypeptides detached easily from the
column during washing steps indicating that the binding mediated by the histidine
tag is weak (Figure 2C–D in II).
The R594H and G597V mutant polypeptides that remained bound to the
affinity column after the washing procedure eluted immediately after the addition
of the elution buffer, whereas the T604I mutant and wild-type polypeptides
required a longer presence of histidine to detach from the column (Figure 3 in II).
These results indicate that in the T604I mutant and wild-type LH2(long) the
histidine tag is readily exposed, while the R594H and G597V mutations may
change the structure of the protein so that the tag becomes partly hidden.
5.2.2 The Bruck syndrome mutant R594H and G597V LH2(long)
polypeptides have problems in dimerization and are prone to
degradation
To analyze the purified recombinant wild-type and mutant LH2(long) proteins,
fractions containing the purest samples were pooled, concentrated and the
histidine was removed by PD-10 columns. Analyses by SDS-PAGE under
reducing conditions and nondenaturing PAGE revealed that in contrast to the
wild-type and T604I mutant polypeptides, the R594H and G597V mutants
coeluted with additional polypeptides (Figure 4A and 4C in II). An anti-LH
antibody identified some of these additional polypeptides in Western blot analysis
(Figure 4B and 4D in II). LHs are dimeric proteins (Kivirikko et al. 1992,
Rautavuoma et al. 2002) that dissociate into a monomeric form in reducing
SDS-PAGE (Rautavuoma et al. 2002). Western blot analysis showed that the
R594H and G597V mutants formed aggregates that were not dissociated into
monomers in SDS-PAGE like the wild-type and T604I polypeptides (Figure 4B in
II). The antibody recognized also two other bands in the R594H and G597V
samples with a molecular weight smaller than that of the LH2 monomer,
indicating that these two mutant forms were prone to degradation (Figure 4B in
II). Identification of the other coeluting bands by N-terminal sequencing was
unsuccessful and thus these polypeptides were regarded as impurities (data not
shown).
As indicated earlier, LH1 is shown to be a homodimeric protein in tissues
(Kivirikko et al. 1992) and also recombinant LH isoforms produced in insect cells
are shown to form dimers (Rautavuoma et al. 2002). Analysis of the mutant
68

ecombinant polypeptides by nondenaturing PAGE revealed that the T604I
mutant assembled into dimers normally when compared to the wild-type, while
oligomerization of the R594H and G597V mutants was abnormal (Figure 4C–D
in II). The R594H and G597V polypeptides formed oligomers that had either a
slower or faster mobility than the dimers formed by the wild-type or T604I
mutant LH2(long) polypeptides, and they also formed aggregates that were
unable to enter the gel properly (Figure 4C–D in II).
5.2.3 Folding of the R594H and G597V mutant LH2(long)
polypeptides is impaired and their proteolytic sensitivity is
increased
To understand the consequences of the BS point mutations on the folding of LH2,
the wild-type and mutant LH2(long) polypeptides were analyzed by far-UV
circular dichroism (CD) spectroscopy. The CD spectra of the wild-type and T604I
mutant LH2(long) were essentially identical and typical for folded, mainly
α-helical proteins with a positive maximum at 193 nm and negative minima at
208 and 222 nm (Figure 5 in II). In accordance with the purification, SDS-PAGE,
native PAGE and Western analyses, marked changes in the folding of the mutants
R594H and G597V were observed. The R594H and G597V LH2(long)
polypeptides both had only one negative minimum at about 207 nm (Figure 5 in
II), indicating a dramatic change in the folding of these two mutant forms.
As the results from the above analyses indicated impaired structural quality
of two of the mutant LH2(long) polypeptides, the proteolytic sensitivity of wildtype
and mutant forms were studied by thermolysin digestion. LH polypeptides
have been previously shown to have two protease-sensitive regions that are
digested by thermolysin generating three domains with molecular weights of 16,
30 and 37 kDa (Rautavuoma et al. 2002). Digestion of the wild-type and T604I
mutant polypeptides produced these three thermolysin-resistant domains with
molecular weights similar to those reported earlier (Rautavuoma et al. 2002)
(Figure 6 in II), but digestion of the R594H and G597V mutants led to more
extensive proteolysis, supporting the above results that folding of these mutants is
not normal (Figure 6 in II).
69

5.2.4 All Bruck syndrome mutations affect the activity of LH2
To understand the effect of the BS2 point mutations on the activity of LH2, LH
activity of the recombinant wild-type and mutant enzymes was analyzed by an
assay based on the hydroxylation-coupled decarboxylation of
2-oxo[1- 14 C]glutarate with a synthetic peptide (Ile-Lys-Gly)3 as a substrate
(Kivirikko & Myllylä 1982). The wild-type LH2(long) was used in the activity
assays as an internal control.
The results of the assays showed that the activity of all the three mutant LH2
enzymes is decreased. Under the assay conditions used, typical activity values
obtained with the wild-type LH2(long) were about 6000 dpm, whereas the
activity of the T604I mutant was only about 500 dpm and those of the R594H and
G597V mutants less than 300 dpm. Although the activity of the T604I mutant was
decreased to less than 10% of that of the wild-type, the Km value of this mutant
for the peptide substrate was normal (Table 1 in II). However, changes in the Km
values for the reaction cosubstrates may also have an effect on the enzyme
activity. The Km values of the T604I mutant for ascorbate and Fe 2+ were normal
when compared to the wild-type, whereas the Km for 2-oxoglutarate was
increased by about 10-fold (Table 1 in II). When 2-oxoglutarate was used in a
saturating concentration in the reaction, the activity of the T604I mutant increased
slightly, being about 30% of the wild-type (data not shown). The Km values of the
R594H and G597V mutants for the peptide substrate were markedly increased
when compared to the wild-type (Table 1 in II). Because of the folding problems
and extremely high Km values for the substrate, the catalytic properties of the
R594H and G597V mutants were not studied further.
LH2(long) is reported to hydroxylate the telopeptide lysines in collagens
(Uzawa et al. 1999, Mercer et al. 2003, van der Slot et al. 2003, Pornprasertsuk et
al. 2004, van der Slot et al. 2004, Takaluoma et al. 2007). Previously it was
shown that LH2(long) is the only LH isoenzyme capable of hydroxylating the N
telopeptide lysine (Lys 9 ) in a full-length pro-α1(I) chain of type I collagen when
coexpressed in insect cells (Takaluoma et al. 2007). As impaired 2-oxoglutarate
binding was the only characteristic that clearly distinguished the T604I mutant
LH2(long) from the wild-type, we tested whether the mutant is nevertheless able
to hydroxylate the telopeptide Lys 9 of type I collagen in insect cell coexpression
experiments. Full-length proα1(I) chains were coexpressed with C-P4H and wildtype
or T604I mutant LH2(long). The produced recombinant type I procollagen
homotrimers were digested with pepsin and the triple helical collagen I
70

homotrimers were further purified and analyzed by N-terminal sequencing. The
wild-type LH2(long) hydroxylated the telopeptide Lys 9 as described previously
(Takaluoma et al. 2007), whereas the hydroxylysine peak was absent in the
reverse-phase HPLC profile of the collagen I homotrimer purified after
coexpression with the T604I mutant LH2(long) (Figure 7A–B in II). Thus, the
T604I mutant LH2(long) was not able to hydroxylate the telopeptide Lys 9 residue
during coexpression, revealing the importance of proper binding of reaction
cosubstrates in the LH reaction.
5.3 Lysyl hydroxylase 2 is essential for the embryonic
development of mouse (III)
5.3.1 LH2 null mice die after embryonic day 10.5
To study the in vivo role of LH2, we generated three different mouse lines
carrying a gene-trap insertion cassette with a β-galactosidase reporter in the Plod2
gene that disrupts the gene function (the LH2:KST067 gene-trap targeting
strategy is shown in Figure 1A in III). All three lines showed the same main
phenotype i.e. embryonic lethality, and thus we chose one mouse line originating
from the targeted ES cell line LH2:KST067 for further analysis. The lack of the
production of LH2 mRNA with a sequence from exon 5 onwards was verified by
RT-PCR (Figure 1B in III). The heterozygous Plod2 wt/gt (wt stands for a wild-type
and gt for a gene-trapped allele) mice appeared to be healthy and fertile and had a
normal lifespan, but genotyping (Figure 1C in III) revealed that intercrosses of the
heterozygous mice yielded only wild-type and heterozygous offspring. Analysis
of the embryos at different ages from heterozygous breedings revealed that the
Plod2 gt/gt embryos developed seemingly normally until E10.5, when the LH2 null
embryos constituted 23% of the population (n = 165 and 724, respectively) (Table
2 in III). Thus, the amounts of the Plod2 gt/gt embryos at E10.5 were close to the
expected Mendelian ratio. At E11.5 the proportion of LH2 null embryos was
about 14% (n = 326) and at E12.5 only 6% (n = 85) (Table 2 in III). After E12.5
living Plod2 gt/gt embryos were not observed. Older embryos between E11.5–12.5
were collected when the backcrossing was still ongoing. Thus, the given
percentages could be even lower in the backcrossed lines.
The Plod2 gt/gt embryos developed apparently normally until E10.5 but died
suddenly after this. The size and appearance of the LH2 null embryos at 10.5 was
71

seemingly normal when compared to the wild-type and heterozygous littermates
(Figure 2A in III), but pooled blood was frequently observed inside the fetal
membranes (Figure 2B–C in III) and sometimes possibly between the visceral
yolk sac and amnion (data not shown), while the heart was still beating. After the
first observation of the accumulated blood, 21% of the Plod2 gt/gt embryos and 2%
of the Plod2 wt/gt embryos were found to have this phenomenon, while wild-type
embryos were all normal (Table 3 in III).
5.3.2 LH2 is expressed ubiquitously in the E10.5 embryos and in
extraembryonic tissues and its lack seems to lead to certain
changes in collagen expression
X-Gal staining, representing the Plod2 promoter-driven β-galactosidase
expression was extensive in E10.5 Plod2 gt/gt embryos (Figure 3A–D in III).
Staining was observed ubiquitously in the mesenchymal tissue (Figure 3A in III).
The expression was intense also in somites, the auditory vesicle and amnion, for
instance (Figure 3A–D in III). In placenta, X-gal staining was observed in the
maternal decidua and on the embryonic side in the chorionic plate and labyrinth
layer, where the expression was seen in the mesenchymal tissue surrounding
embryonic vessels (Figure 3E–F in III). Parietal endoderm cells producing the
Reichert’s membrane, the parietal layer of the yolk sac, were strongly stained in
Plod2 gt/gt embryos, but the expression was not seen in the membrane (Figure 4B
in III).
It has been shown earlier that the collagen α1(IV)2α2(IV) null mice die
between E10.5–E11.5 because of deficient basement membrane assembly and
finally of the failure of the Reichert's membrane (Pöschl et al. 2004). No obvious
differences between the Plod2 gt/gt and wild-type embryos were seen in light
microscopy or TEM analysis of the Reichert’s membrane or in the type IV
collagen immunofluorescence staining (Figure 4A–F in III). However, according
to TEM analysis, we were not able to exclude a weakened adhesion of the
Reichert’s membrane to placenta in Plod2 gt/gt embryos (data not shown) and also
the normal detachment of parietal endoderm cells was sometimes questioned
(Figure 4D in III).
In other extraembryonic tissues, pronounced X-gal staining was observed in
the mesodermal part of the visceral yolk sac, but the endodermal side was
unstained (Figure 5B–C in III). Morphological analysis of the visceral yolk sac of
Plod2 gt/gt embryos revealed that some areas of the mesodermal part of the tissue
72

appeared loose in comparison to wild-type (Figure 5A–C in III). The loosened
mesodermal tissue was also surrounding vitelline vessels, and thus it cannot be
excluded that in some of the hemorrhage cases this site may have collapsed
followed by a rupture in the vitelline vessel that could have led to the bleeding
between the yolk sac and amnion. In immunofluorescence staining of
extraembryonic tissues, type III and type V collagens showed differential
expression patterns (Figure 6A–D in III). Type III collagen was seen in the
endodermal side of the visceral yolk sac (Figure 6A–B in III), while type V
collagen was observed in the mesodermal part of the tissue (Figure 6C–D in III).
In the Plod2 gt/gt yolk sac the staining of both collagen types seems weaker than
that of the Plod2 wt/wt tissue although the distribution was normal (Figure 6A–D in
III). Staining of type I collagen was seen in the mesodermal part of the visceral
yolk sac, but differences between genotypes were not observed (data not shown).
Type I, IV and V collagens but not type III collagen were expressed also in the
chorionic plate of the placenta, but no obvious differences were seen between the
Plod2 gt/gt and Plod2 wt/wt embryos (data not shown).
As the phenotype of the mouse line indicated a cardiovascular defect,
anti-CD31 staining was performed to study the structure of large vessels of the
embryos, but the endothelium of the aorta appeared intact (Figure 7A–B in III).
However, TEM analysis revealed that in the Plod2 gt/gt aorta the overall structure
of the endothelium seemed looser when compared to Plod2 wt/wt embryos (Figure
7C–D in III). The mesenchymal cell network under the endothelium was sparse
and the cells appeared elongated and showed numerous thin protrusions in
Plod2 gt/gt embryos, while in the Plod2 wt/wt embryos the mesenchymal cells were
larger and they were located more closely to the endothelium. LH2 is expressed in
the mesodermal tissue surrounding the aorta (Figure 3A in III). As blood pressure
increases during development this area must sustain higher mechanical loads and
provide support for the large vessels. Thus, collagen fibrils were studied by TEM
from this mesodermal area and some fibrils were seen in both Plod2 wt/wt and
Plod2 gt/gt samples (Figure 7E–F in III).
5.3.3 Expression of the Plod1 and Plod3 genes is not upregulated in
the Plod2 gt/gt embryos
To study whether inactivation of the Plod2 gene leads to upregulation of Plod1
and Plod3 gene expression, we performed Q-PCR and RT-PCR analyses from
10.5 days old embryos. Plod2 expression followed the expected levels with about
73

50% expression in the Plod2 wt/gt embryos relative to wild-type, whereas
expression was not detected in the Plod2 gt/gt embryos (Figure 8A–B in III). The
LH mRNA levels were determined from single embryos and results varied
markedly between individuals in the same genotype group. At E10.5 the
development of embryos is fast and even minor age differences may cause large
differences in the mRNA levels. According to our LH1 null mouse study, the
other two LH isoenzymes are suggested to at least partially compensate for the
LH1 deficiency (Article I in this thesis). However, the mRNA levels of Plod1 and
Plod3 genes in the Plod2 wt/gt and Plod2 gt/gt embryos did not show any significant
increase, thus suggesting that they cannot compensate for the decreased LH2 level
(Figure 8A–B in III). In any case, the LH1 and LH3 could not compensate for the
telopeptide LH activity of LH2.
5.3.4 LH2 is widely expressed in adult Plod2 wt/gt tissues
Plod2 was expressed in almost all the adult tissues studied. In the bone, X-gal
staining was observed in the endosteum, periosteum and round osteocytes (Figure
9A–B in III). The pattern was similar in both bone types analyzed, the femur and
calvaria. Staining was also intense in the cartilage (Figure 9C in III), tendon (data
not shown), and in the smooth muscle layers of the uterus (Figure 9D in III) and
intestine (not shown). In blood vessels LH2 was expressed moderately in
endothelial cells (Figure 9E–F in III). In the tunica media the expression was
localized to smooth muscle cells located between the elastic lamellae (Figure 9E–
F in III). X-gal staining was also seen in the cells of tunica adventitia (Figure 9E–
F in III). In the eye, LH2 was expressed in the anterior lens epithelium as well as
in the ganglion cell layer and in the inner nuclear layer of the retina (data not
shown). In the skin the expression was strongest in the hair bulbs (data not
shown).
5.3.5 Changes in lysyl hydroxylation and cross-linking of collagen in
adult Plod2 wt/gt mice
In BS2 patients the collagen cross-linking is altered due to abnormal function of
LH2, therefore we wanted to study the collagen cross-linking from E10.5
embryos. However, maturation of collagen cross-links requires time. Therefore,
the collagen cross-link analysis from embryos turned out to be unsuccessful as the
mature cross-links had not yet formed at E10.5 (data not shown). We thus decided
74

to study whether any differences in collagen content, collagen hydroxylations,
and cross-linking can be seen in adult heterozygous Plod2 wt/gt mouse tissues
relative to wild-type.
The amounts of collagen and 4-hydroxyproline were unchanged in all adult
Plod2 wt/gt tissues studied when compared to wild-type control mice (Figure 10A–
B in III). The highest collagen levels in wild-type tissues were observed in the tail
tendon and skin, where the collagen content was 58% and 52% of the dry weight
of the tissue, respectively (Figure 10B in III).
The total hydroxylysine content, expressed per triple helix, varies markedly
between tissues as collagen types contain variable amounts of hydroxylated
lysines and tissues contain different amounts of different collagen types.
LH2(long) is responsible for the hydroxylation of telopeptide lysines that
participate in cross-link formation and there are a maximum of two of these
lysines per collagen molecule (Knott & Bailey 1998, van der Slot et al. 2003,
Takaluoma et al. 2007). The triple helical area contains additional lysine residues
that can be hydroxylated most probably by all three LH isoenzymes (Takaluoma
et al. 2007). Two of these lysine residues are involved in cross-links (Knott &
Bailey 1998). Thus, changes in the hydroxylation levels of the telopeptide lysines
do not dramatically affect the total lysine hydroxylation level. In addition, the
Plod2 wt/gt mice have one functional Plod2 allele left. The highest amount of total
hydroxylysine in wild-type samples was observed in the lung (51 residues / triple
helix) and the lowest in the skin (18 residues / triple helix) (Figure 10D in III). In
the Plod2 wt/gt mouse tissues the total hydroxylysine content was changed
significantly only in the calvaria and femur, which also normally contain a large
amount of pyridinoline cross-links. The amount of hydroxylysine was decreased
to 85% in calvaria and to 80% in femur when compared to the corresponding
wild-type tissues (Figure 10D in III).
As changes in the hydroxylation of the telopeptide lysine residues in
collagens affect the cross-links in BS patients (Bank et al. 1999), mature
pyridinoline cross-links were studied from the wild-type and Plod2 wt/gt mice. Total
hydroxylysine levels but also the LP and HP contents varied markedly between
different tissues (Figure 10D–F in III). The highest levels of both mature crosslinks
in wild-type mouse were observed in the lung (LP 0.57 and HP 0.51 residues
/ triple helix) and aorta (LP 0.12 and HP 0.38 residues / triple helix) (Figure 10E–
F in III), although the total amount of collagen was low in these tissues (Figure
10B in III). The lowest levels of mature cross-links in wild-type were detected in
the tail tendon (LP 0.0015 and HP 0.0078 residues / triple helix) (Figure 10E–F in
75

III). The change in the LP content between Plod2 wt/wt and Plod2 wt/gt was
significant only in the calvaria, where the level was decreased in the Plod2 wt/gt
sample to 51% relative to the wild-type (Figure 10E in III). In contrast, the HP
levels in the Plod2 wt/gt mouse tissues were markedly more decreased. The
decrease in the HP levels was significant in all the Plod2 wt/gt tissues studied except
the aorta (Figure 10F in III), which was most severely affected in the Plod1 -/-
mouse (Article I of this thesis). The most dramatic change in the HP was seen in
the calvaria, where the HP content was 33% (reduced from 0.1 to 0.04 residues /
triple helix) of the wild-type. In addition to calvaria, the skin had only 45%
(reduced from 0.01 to 0.005 residues / triple helix) and femur 65% (reduced from
0.4 to 0.26 residues / triple helix) of the wild-type HP levels. Slightly minor
changes were observed in the tail tendon (70% of the wild-type value of 0.008
residues / triple helix) and lung (83% of the wild-type value of 0.51 residues /
triple helix). Interestingly, in general the amount of HP cross-links was markedly
more affected than the LP cross-links in the Plod2 wt/gt tissues, although the only
difference between the LP and HP is in the hydroxylation state of the helical
lysine. However, LH2 is suggested to have also helical lysyl hydroxylase activity
(Takaluoma et al. 2007, Article I in this thesis).
As long-lived proteins, collagens are susceptible to non-enzymatic, unwanted
cross-links involving reducing sugars (Saito & Marumo 2010). As lysine residues
are participating in these advanced glycation end-products (AGEs), the effect of
reduced lysyl hydroxylation on the amount of pentosidine, one of the AGEs, was
also studied in the Plod2 wt/gt tissues. Some changes were seen in the femur, for
instance, where the pentosidine level was slightly increased in the Plod2 wt/gt mice,
but the results were not statistically significant (Figure 10C in III).
76

6 Discussion
6.1 LH1 null mice provide a tool to study tissue-specific
consequences of the lack of LH1 activity
We have generated a Plod1 -/- mouse line to study the function of LH1 in vivo. In
addition, we expected that the mouse line could also serve as a disease model for
the human EDS VIA. This is the first reported animal model for a defect in the
gene coding for LH1, and the created mouse line enables for the first time
detailed examination of the effects of the lack of this enzyme in various organs
and tissues which is difficult, impossible, or unethical to do in the case of the
EDS VIA patients.
EDS VIA patients suffer already at birth from kyphoscoliosis, muscular
hypotonia and joint laxity, for instance (Steinmann et al. 2002). The LH1 null
mice are viable, their breeding is normal, and they are not observed to suffer from
kyphoscoliosis. However, they do have muscular hypotonia which becomes more
obvious when the mice are older and have increased weight. Although the Plod1 -/-
mice appear to be flaccid during handling and have muscle weakness and gait
abnormalities in situations where strength and coordination is needed, they do not
have an obvious skeletal phenotype like the EDS VIA patients. This can be due to
differences in the size and anatomy of the mouse and human as well as the
upright posture of humans that exposes the skeleton to forces different from those
in the mice. In EDS VIA patients the kyphoscoliosis is also progressive, which is
thought to be due to a supplementary effect of the muscle hypotonia and joint
laxity as the vertebral bodies, for instance, are shown to be normal (Steinmann et
al. 2002). Muscular hypotonia, which is connected to a basic ECM-forming
enzyme, is an interesting phenotype which could also be exploited in studies of
other muscular disorders and especially the role of connective tissue in these
diseases. Mutations in highly glycosylated and thus hydroxylated type VI
collagen cause Ullrich congenital muscular dystrophy and the milder Bethlem
myopathy (for a review, see Lampe & Bushby 2005). In addition to muscular
dystrophy, the disorders are characterized by joint hyperlaxity and contractures. A
null mouse for type VI collagen develops a muscular phenotype with variation in
the muscle fiber diameter and fiber necrosis (Bonaldo et al. 1998). Interestingly,
similar to the Plod1 -/- mouse a null mouse line for type VI collagen has recently
been shown also to possess abnormalities in collagen fibril diameter and fibril
77

assembly in tendons (Izu et al. 2011). Moreover, the mechanical properties of
tendons were decreased. In contrast to Plod1 -/- mice, however, the collagen fibril
diameter in the collagen VI null mice was generally decreased and not increased
(Izu et al. 2011).
EDS VIA patients have soft, hyperextensible skin that is prone to abnormal
scarring. In addition, the patients bruise easily. The skin of the Plod1 -/- mice
appeared to be normal, but detailed analysis revealed that the collagen fibrils were
abnormal. The diameter of the fibrils varied in the Plod1 -/- skin samples and their
shape was in part irregular when compared to the wild-type skin. In addition to
cross-link formation, hydroxylysine residues are known to function as attachment
sites for carbohydrate residues, which in turn affect the fibril diameter and
fibrillogenesis of collagen molecules (Myllyharju 2005), supporting these
findings for the Plod1 -/- skin collagens. The hydroxylysine content in EDS VIA
skin is reduced typically to about 5% of that of healthy controls (Steinmann et al.
2002), whereas in Plod1 -/- mouse skin the hydroxylysine content was over 20%.
In addition, the residual LH activity in Plod1 -/- mouse skin homogenates was
quite high, 35% of the wild-type, while in EDS VIA patient fibroblasts the values
vary from 5% to 25% of the controls (Yeowell & Walker 2000). Thus, the lack of
an obvious skin phenotype in the Plod1 -/- mice is likely to be because of the
higher LH activity and hydroxylysine levels in the mutant mouse skin. These
differences might also be due to possible variable expressions of the LH
isoenzymes in the Plod1 -/- mouse and EDS VIA skin.
A life-threatening complication in EDS VIA patients is an arterial rupture.
The arteries of the Plod1 -/- mice were also found to be structurally weak as some
of the LH1 null mice died during the night because of aortic rupture. This feature
was almost twice as prevalent in male as in female mice. In histological analysis
the aortic wall of apparently healthy Plod1 -/- mice appeared to be generally
normal. Moreover, heart function was also normal. In contrast, in necropsy
samples of the animals that had died suddenly and had hemorrhages in the
thoracic or abdominal cavity, dissections in the aorta were observed. As in the
skin, detailed analysis of the aorta by TEM revealed abnormal collagen fibrils and
also degrading fibrils even in the apparently healthy looking Plod1 -/- mice.
Degenerative changes were also observed in the smooth muscle cells of the
Plod1 -/- aortas. These observations are interesting because the hydroxylysine level
(63% of wild-type) and residual LH activity (44% of wild-type) in the Plod1 -/-
aorta are in fact higher than in the skin (22% and 36%, respectively), which
appeared to be asymptomatic. However, mechanical requirements in the aorta are
78

high due to high blood pressure and the weakened structures in the tissue are
probably predisposing it to ruptures.
In addition to the skin and aorta, the hydroxylysine content was also
decreased in other Plod1 -/- tissues studied. It is suggested that LH1 is mainly
responsible for the hydroxylation of lysine residues in the triple helical region of
collagens, while LH2(long) exclusively hydroxylates the telopeptides (Eyre et al.
2002, van der Slot et al. 2003, Takaluoma et al. 2007). The exact role of LH3 in
lysyl hydroxylation is still unknown as for example the embryonic lethality of
LH3 null mice has been shown to be due to the lack of its collagen
glycosyltransferase and not LH activity (Rautavuoma et al. 2004, Ruotsalainen et
al. 2006). Although the Plod1 -/- mice do not have functional LH1, in certain
tissues, such as the lung, the hydroxylysine level is as high as 86% of the wildtype,
while in the skin the level was only 22%. The differential hydroxylation
content in tissues most likely reflects tissue specificities of the LH isoenzymes,
differences in their amounts in different tissues, and probably also differential
compensatory mechanisms of the LHs. At the mRNA level all LH isoforms are
expressed in almost all mouse and human tissues analyzed (Valtavaara et al. 1997,
Valtavaara et al. 1998, Passoja et al. 1998a, Ruotsalainen et al. 1999, Yeowell &
Walker 1999a, Rautavuoma et al. 2004). At the protein level the data is still
limited. Our data also clearly suggest that in addition to LH1, the other two
isoforms, LH2 and LH3, are able to hydroxylate helical lysines to variable extents
in different tissues in vivo.
The mechanical strength of tissues depends in part on the type and amount of
the existing collagen cross-links. The type of the cross-link, in turn, depends on
the hydroxylation state of certain triple helical and telopeptide lysines. Two
different routes exist for cross-link formation. The hydroxyallysine pathway leads
to the formation of pyridinoline cross-links HP and LP that predominate in tissues
such as bone and cartilage. LP is formed between two telopeptide hydroxylysine
molecules and a helical lysine, whereas in the HP cross-link all participating
lysines are hydroxylated. The allysine pathway, which leads to mechanically
weaker bonds, is common in the skin and cornea, for instance. (Knott & Bailey
1998, Eyre & Wu 2005). As the cross-linking profile in tissues of EDS VIA
patients is changed, they have increased urinary LP levels and decreased HP
levels (Pasquali et al. 1994, Pasquali et al. 1995, Steinmann et al. 1995). As was
expected, the amount of HP cross-links was decreased in all the Plod1 -/- mouse
tissues studied, while the amount of LP cross-links was increased. This supports
the earlier suggestion that LH1 is the main LH that hydroxylates helical lysines.
79

The total amount of HP + LP was increased in the Plod1 -/- mice because of
increased LP that could indicate a response to the changed cross-linking state.
Both pyridinolines require telopeptide lysine hydroxylation and thus, at least
LH2(long) upregulation could be behind this response. However, the maturation
of cross-links to LP and HP requires time and it is not known whether the rate is
the same for both pyridinoline cross-links; this may also influence the results.
Surprisingly, there was no correlation between the overall hydroxylation level and
the amount of cross-links. Although the total hydroxylysine amount in the aorta of
the Plod1 -/- mouse was rather high (63% of wild-type), the amount of HP was
only 28% of that of the wild-type, while in the tail tendon the total hydroxylysine
content was low (24% of wild-type), but the HP value was still 59% of the control,
for instance. These results clearly show that LH2 and/or LH3 are able to
hydroxylate the triple helical cross-link site lysines to some extent in various
tissues when LH1 is not functional. Moreover, the results demonstrate again the
complex roles and interplay between the LH isoenzymes in tissues.
Changes in lysyl hydroxylation not only influence the mechanical strength of
tissues via cross-links but also individual collagen fibrils as was discussed earlier.
The diameter and shape of collagen molecules in the Plod1 -/- skin and aorta were
changed, but not in cornea. Many factors may contribute to these characteristics,
but currently they are not well known. However, glycosylation of hydroxylysine
residues is known to affect fibrillogenesis and fibril diameter (Torre-Blanco et al.
1992, Notbohm et al. 1999, Sricholpech et al. 2011). In the Plod1 -/- mice the
amount of hydroxylysine residues available for glycosylation is decreased, which
is likely to affect the glycosylation level as well. Moreover, collagen I fibrils
contain small amounts of type III and V collagens, which in turn are shown to
regulate the thickness of fibrils (Kielty & Grant 2002). EDS patients that have a
vascular or classical form of the syndrome have mutations in these minor
collagens and they have abnormal collagen I fibrils (Steinmann et al. 2002). The
null mouse lines for these collagen types are also shown to have abnormal
collagen fibrils (Liu et al. 1997, Wenstrup et al. 2006), as well as mouse lines
lacking small leucine-rich proteoglycans such as biglycan, decorin, fibromodulin
and lumican (Danielson et al. 1997, Chakravarti et al. 1998, Svensson et al. 1999,
Corsi et al. 2002). Thus, it is possible that deficient lysyl hydroxylation and
glycosylation of fibrillar collagens in the Plod1 -/- mice may interfere with proper
binding of other proteins, which in turn may lead to abnormal collagen fibrils.
Moreover, the Plod1 -/- mice show that the abnormal collagen fibrils observed in
80

different EDS types can also be caused by a defective collagen-modifying
enzyme and not just by mutations in the collagen molecules themselves.
6.2 Mutations causing Bruck syndrome lead to decreased LH2
activity, explaining the symptoms of diagnosed Bruck
syndrome type 2 patients
BS has been linked to lysyl hydroxylation of telopeptides because of the
decreased amount of pyridinoline cross-links in bone collagen (Bank et al. 1999).
Three different mutations found from the PLOD2 gene led to the suggestion that
these mutations cause BS2 and that the LH2 is the telopeptide LH (van der Slot et
al. 2003, Ha-Vinh et al. 2004). As these mutations do not change any of the
amino acids that are known to be catalytically critical, the molecular mechanism
causing the disorder has been unknown. In this study we produced the BS mutant
and wild-type human LH2(long) polypeptides in insect cells and purified and
analyzed them to understand the molecular pathology of the syndrome.
Our data show that all three known BS2 mutations lead to marked reduction
of LH2 activity. The most significant decrease in the activity was observed in the
mutants R594H and G597V, whose activity was below 5% of that of the wild-type
enzyme. These mutations most probably affect proper folding of the polypeptides,
which in turn may also affect their oligomerization. These effects are likely to
lead to the marked decrease in the enzyme activity. All the mutated amino acids
found in LH2 in the BS2 patients, Arg 594 , Gly 597 and Thr 604 , are conserved in the
human and mouse LH isoenzymes (Valtavaara et al. 1998, Passoja et al. 1998a,
Ruotsalainen et al. 1999). LH monomers are shown to consist of three domains.
The C-terminal domain of all three isoenzymes is responsible for the LH activity
and the N-terminal domain in LH3 possesses glycosyltransferase activity, but the
function of the middle domain is unknown. (Rautavuoma et al. 2002). The
residues Arg 594 and Gly 597 in LH2(long) are located in the region that separates
the middle and C-terminal domains, which is known to be a protease-sensitive
region. Moreover, C-terminal recombinant LH1 and LH3 fragments beginning
with the residue corresponding to Lys 589 in LH2(long) have been shown to be
fully active enzymes (Rautavuoma et al. 2002). Thus, it is likely that the amino
acids Arg 594 and Gly 597 are part of an unstructured domain boundary that may
function as a linker. Changes in this area may affect its flexibility and the folding
capability of the flanking domains.
81

Folding of the mutant T604I appears to be normal, but the activity in standard
conditions was decreased to under 10% of that of the wild-type LH2(long). The
change in the activity level was caused by a 10-fold increase in the Km for
2-oxoglutarate, whereas binding efficiency of the peptide substrate or the other
cosubstrates of the reaction was not affected. The maximum activity of the T604I
obtained under a saturating 2-oxoglutarate concentration was about 30% of that of
the wild-type. The increased Km value for 2-oxoglutarate indicates that binding of
this cosubstrate is abnormal in the T604I mutant. Moreover, the T604I mutant
was not able to hydroxylate the N-terminal telopeptide lysine in recombinant type
I procollagen chains in cellulo. LHs belong to the superfamily of 2-oxoglutarate
dioxygenases, which have a catalytic site with a common structure formed by an
8-stranded β-helix core fold. A conserved basic residue in the eighth strand binds
the C-5 carboxylate moiety of 2-oxoglutarate and in LH2 the residue is Arg 724 .
(Passoja et al. 1998b, Clifton et al. 2006). This residue is located near the second
Fe 2+ -binding histidine in all known 2-oxoglutarate dioxygenases except the factor
inhibiting HIF, an asparaginyl hydroxylase of HIF (Schofield et al. 2004). Certain
serine or threonine residues near the basic residue in the eighth strand provide an
additional interaction with the C-5 carboxylate of 2-oxoglutarate in many
2-oxoglutarate dioxygenases (Clifton et al. 2006, Koski et al. 2007). Mutation of
this serine residue in recombinant human LH1 is shown to reduce the enzyme
activity markedly, although the Km value for 2-oxoglutarate remained normal
(Passoja et al. 1998b). Unfortunately the three-dimensional structure of LHs is
not known and thus the exact structural role of Thr 604 in the binding of
2-oxoglutarate remains unknown.
The mutant enzymes were not totally inactive, which may be important for
patients carrying these mutations as during this study we observed that LH2 null
mice die early, during embryonic development (Article III in this thesis). On the
other hand, for example in EDS VIA LH1 is known to have alternative splicing
pathways to bypass or compensate mutations and to retain some enzyme activity
(Yeowell & Walker 2000). However, it is not known whether the residual activity
is enough to enable the survival of patients or is LH2 using similar alternative
splicing system in BS2 as LH1 in EDS VIA or whether the role of LH2 in human
development is not as critical as in mouse development. However, according to
published reports, two affected children carrying the BS mutation T604I deceased
prenatally and one child survived only for 2 years (van der Slot et al. 2003). Thus,
this mutation seems to be clinically the most severe form of the reported three
forms (van der Slot et al. 2003, Ha-Vinh et al. 2004). This is a bit surprising as
82

the effects of the T604I mutation in the recombinant LH2 were not as severe as
those of the Arg 594 , Gly 597 mutations.
The type and amount of collagen cross-links existing in the tissue determine
in part its mechanical strength as they stabilize the collagen fibrils. Cross-links
are formed via two different routes, the allysine and hydroxyallysine routes. The
hydroxylation state of the telopeptide cross-link lysine residues determines the
route and in the hydroxyallysine route these lysines are hydroxylated, while in the
allysine route they are not. The mature pyridinoline cross-links HP and LP, which
predominate in tissues requiring strength such as bone and cartilage, are formed
via the hydroxyallysine route and thus require the presence of hydroxylysine in
the telopeptide cross-link sites. (Knott & Bailey 1998, Eyre & Wu 2005). BS
patients have a highly abnormal cross-linking profile in the bone collagen. The
amount of LP and HP is dramatically decreased, while the amount of allysinederived
cross-links is increased. (Bank et al. 1999). Thus, it can be deduced that
the BS2 mutations decrease the LH2 activity to an extent that results in changes in
the cross-linking profile, which causes the clinical consequences observed in the
patients.
6.3 Inactivation of mouse Plod2 leads to lethality during early
embryonic development
Mutations in the PLOD1 gene lead to LH1 deficiency and EDS VIA, which is a
severe but not lethal connective tissue disorder. The mouse model for this human
syndrome shows a milder phenotype, but both have an increased risk of arterial
rupture, for instance (Article I in this thesis). Mice lacking the LH3 enzyme die
during embryonic development due to abnormal processing of type IV collagen
leading to basement membrane fragmentation (Rautavuoma et al. 2004,
Ruotsalainen et al. 2006). Recently one patient was shown to have compound
heterozygote mutations in the PLOD3 gene (Salo et al. 2008). The syndrome is
severe with symptoms overlapping several connective tissue disorders and the
patient also has a stillborn sibling. Mutations in the gene encoding LH2 cause
BS2, where patients have severe skeletal problems, but it is usually not lethal. In
this study we have generated a novel LH2 null mouse line from an ES cell line, in
which the Plod2 gene is disrupted by a gene-trap insertion cassette.
As the human BS2 is not lethal, it was surprising that LH2 null mice died
during embryonic development, after E10.5. However, our previous in vitro
studies showed that recombinant LH2 enzymes carrying the mutations identified
83

in BS2 patients are not totally inactive and, thus, the residual LH2 activity may be
indispensable for the patients (Article II in this thesis). LH2 null embryos
developed seemingly normally until E10.5, but died after this without any
obvious signs of growth retardation. However, about a fifth of the null embryos
were observed to have accumulated blood inside the fetal membranes while the
heart was still contracting. This could indicate problems in the cardiovascular
system or the placenta.
BS2 patients are not reported to have direct cardiovascular symptoms.
However, in Plod2 gt/gt embryos some changes were seen in the endothelium of
aorta, which could affect the structural resistance of the tissue. The cause of death
of all BS patients is not available, however. Two children having a BS2 mutation
deceased prenatally and one child survived for 2 years (van der Slot et al. 2003).
Because of the limited amount of published clinical data about these patients, the
possible role of the placenta, for instance, cannot be excluded. The placenta of
mouse and human are fairly different, but they also have considerable similarities
(reviewed in Rossant & Cross 2001). Although the interest in the development of
placenta and its defects is increasing, knowledge about it is still extremely limited
and far behind the knowledge of embryonic development. As a tissue originating
from a differentiating embryo, it is affected by the same genetic defects as the
embryo itself (Rossant & Cross 2001). The mutant embryo may die before the
phenotype would be apparent if the gene is required in the development and
function of the placenta. LH2 is widely expressed in the embryo and also in the
extraembryonic tissues. On the embryonic side of the placenta LH2 is expressed
in the chorionic plate and mesenchymal tissue surrounding embryonic vessels,
where the maternal and fetal blood are in close proximity to each other and the
tissue between the compartments must be thin but firm. Moreover, the role of the
placenta sets also certain other structural requirements for the tissue, such as
substantial blood flow in the chorionic plate-attached umbilical cord (Inman &
Downs 2007). The cause of death of the null embryos could also be a collapse in
these tissue structures leading to hemorrhage. The major fibril-forming collagens
in the placenta are type I, III and V collagens (Iwahashi et al. 1996, Arai &
Nishiyama 2007), but there is a limited amount of data available about the
collagen types in the chorionic plate at different developmental stages. Type I, IV
and V collagens, but not type III collagen, were expressed in the chorionic plate
of the E10.5 embryos, but clear differences were not observed between the
Plod2 wt/wt and Plod2 gt/gt tissues. Interestingly, immunofluorescence staining of
collagens III and V but not type I was slightly weaker at least in the yolk sac of
84

the Plod2 gt/gt embryos than in the wild-type. The size of the mouse placenta is
increasing markedly between E10–E14 (Arai & Nishiyama 2007), and thus the
tissue must come up with the fast-changing mechanical requirements. Thus,
mechanically inadequate collagenous structures may break down during the fast
development. As LH2 is required for the formation of appropriate ECM, it may be
necessary for the mechanical integrity of the placenta.
Rodents have a Reichert’s membrane (RM) in the placenta, between the
parietal endoderm cells and trophoblast cells, that is lacking in humans (Gardner
1983). The membrane contains high levels of type IV collagen and collagen
α1(IV)2α2(IV) knock-out mice are reported to die between E10.5–E11.5 because
of the failure of this membrane (Pöschl et al. 2004). Moreover, LH3 knock-out
mice die at E9.5 because of abnormal collagen IV synthesis. The parietal
endoderm cells producing the RM as well as the mesodermal part of the visceral
yolk sac have strong LH2 expression. Thus, we thought that the LH2 null mice
may also have a failure in the RM, but a clear defect was not found. Although no
clear abnormalities were observed, a possible role of the membrane failure in the
phenotype cannot be fully excluded. In addition, the mesodermal part of the
visceral yolk sac of Plod2 gt/gt embryos seemed to have sites that appeared loose.
These loosened parts were partly surrounding vitelline vessels and a collapse in
this site may also be the cause of vessel rupture leading to the observed bleeding
between the yolk sac and the amnion. Type III collagen was found to be expressed
in the endodermal part and types I and V in the mesodermal side of the yolk sac at
E10.5. Interestingly, as indicated earlier, immunofluorescence staining showed
that the amount of type V and III but not type I collagen may be slightly lower in
the yolk sac of the Plod2 gt/gt embryos. Moreover, collagen V knock-out mice are
shown to have a similar hemorrhage phenotype as the LH2 null mice at the time
of their death at E10 (Wenstrup et al. 2004). Generally this could indicate
inadequate support in the LH2 null tissues, especially for the blood vessels.
Abundant expression of type I collagen begins in mice at E12 and the Mov 13
mice that cannot synthesize the α1(I) collagen polypeptide die soon after this
thorugh rupture of the major blood vessels of the embryo (Schnieke et al. 1983,
Löhler et al. 1984). Type III collagen expression in mouse is even stronger than
that of type I between E8.5 – E12.5 (Niederreither et al. 1994). About 10% of
collagen III knock-out mice survive to adulthood, but have a much shorter
lifespan relative to the wild-type (Liu et al. 1997). The major cause of death in
collagen III knock-out mice is blood vessel ruptures. Interestingly, as indicated
above, collagen V knock-out mice are reported to have a similar hemorrhage
85

phenotype as the LH2 null mice (Wenstrup et al. 2004). Moreover, collagen V
expression increases markedly at E10.5 in mice (Roulet et al. 2007). Although the
site of vessel rupture or the exact cause for that remained unknown in the collagen
V null mice, type V collagen was shown to be required for proper collagen fibril
formation. Type V collagen participates as a quantitatively minor component to
the formation of type I collagen fibrils and the lack of it prevents the fibril
formation in the mesenchyme almost completely and the formed fibrils look
abnormal (Wenstrup et al. 2004). The phenotype of the collagen V null mouse
line is surprisingly severe, as the phenotype of the heterozygous mice resembles
the classical type of Ehlers-Danlos syndrome (Wenstrup et al. 2004). The
Plod2 gt/gt embryos were nevertheless able to synthesize collagen fibrils.
As a telopeptide LH, LH2 has a role in the synthesis of several collagen types.
Lack of the enzyme may not prevent the synthesis of collagen molecules as they
were seen in the extracellular space, but it may either affect the processing time,
quality and/or the strength of the assembled structures as indicated earlier.
Moreover, hydroxylysine residues at the triple helical regions of collagens serve
as attachment sites for carbohydrates and the exact contribution of LH2 to the
hydroxylation of these lysines is currently unknown. Thus, it is possible that lack
of LH2 may affect the glycosylation status of collagens. It is probable that the
lack of LH2 not only affects a specific collagen target but rather decreases the
strength of ECM structures in general by changing the cross-link type.
Underhydroxylation of the telopeptide lysines prevents the adequate formation of
trivalent, mature, pyridinoline cross-links LP and HP, but the telopeptide lysyl
hydroxylation also determines the type of immature divalent cross-link. At E10.5
mature collagen cross-links are not formed yet. Thus, divalent forms most
probably predominate at this developmental stage, suggesting the importance of
the immature forms of cross-links in developing tissues. Furthermore, our results
show that the compensatory effect of the other two LH isoforms is not significant
when LH2 activity is lacking. Thus, the embryonic lethality of LH2 null mice is
most probably caused by the absence of the telopeptide lysyl hydroxylase activity.
The dramatic consequence of the lack of one collagen-modifying enzyme is
interesting as, for instance, the type I collagen null Mov 13 mice and collagen V
knock-out mice die only a moment earlier or slightly later and the type III
collagen knock-out mice may survive even after birth (Schnieke et al. 1983, Liu
et al. 1997, Wenstrup et al. 2004).
BS patients have skeletal problems such as osteoporosis, fragile bones,
congenital joint contractures, progressive kyphoscoliosis and short stature. As
86

indicated earlier, unfortunately the E10.5 embryos are too young for the cross-link
analysis as maturation is still ongoing, and due to the early embryonic lethality
the skeletal consequences also remain unknown. However, it is interesting that the
LH2 heterozygous mice also have significant changes in the cross-link pattern,
particularly in bone, indicating that the amount of the LH2 needed in the process
is rather high. Moreover, reduction in the amount of pyridinoline cross-links was
more marked in the heterozygous calvaria than the femur. These bone types differ
from each other in their origin and mechanism of ossification, which in turn affect
their composition (Chung et al. 2004, van den Bos et al. 2008). The flat bone
calvaria originates from the ectodermal neural crest, while the femur, representing
a long bone, is of mesodermal origin (Chung et al. 2004). Flat bones are formed
by straight intramembranous ossification, where the cells differentiate directly
into bone-forming osteoblasts that produce a matrix that is particularly rich in
type I collagen. Long bones, in turn, are ossified via a cartilage model, by
endochondral ossification, which begins by a condensation of mesenchymal cells
to chondrocytes. (Kronenberg 2003). The function of the long bones is to provide
support for the body, whereas the role of calvaria, for instance, is more protective
than supportive. Our results indicate that a decrease in LH2 activity may be less
harmful in long bones. They are ossified via a cartilage model and LH2 may have
more time to hydroxylate the telopeptide lysines when cartilage is replaced slowly
by bone. The straight ossification process in the calvaria may in turn be too fast
for the reduced amount of LH2 to keep up with and thus the collagen crosslinking
is more affected. However, bone tissue is renewing constantly and an
interplay between the contributing components is important. The protein
composition in different bone types differs considerably, and for example the
collagen amount in flat bones is higher than in long bones and the collagens are
also more soluble (van den Bos et al. 2008). Interestingly, the levels of certain
growth factors such as TGF-β are higher in flat bones than in long bones
(Finkelman et al. 1994) and, furthermore, TGF-β is known to increase the
expression of LH2(long) and the formation of pyridinoline cross-links (van der
Slot et al. 2005). Thus, in addition to the time available for lysyl hydroxylation in
different ossification processes, changes in the interplay of the contributing
components of bone development may affect the cross-linking state of bones with
reduced amounts of LH2. Moreover, both pyridinoline cross-links require
telopeptide hydroxylysine, i.e. LH2(long), for their formation, and the difference
between the LP and HP cross-link is only in the hydroxylation state of the helical
lysine. Thus, it was interesting that the amount of HP was more severely affected
87

than the LP in the LH2 heterozygous mouse tissues. This indicates that LH2 is
also involved in the hydroxylation of the helical lysine residue participating in
cross-link formation.
88

7 Conclusions and future prospects
Today ECM is known to be a highly dynamic part of the human body and not just
a scaffold for cells and tissues. Functional characteristics of collagens, the most
common protein component of the ECM, are determined during collagen
biosynthesis. In addition to collagen gene mutations, failures in the enzymes
required for collagen biosynthesis and an imbalance of regulating factors, for
instance, can lead to clinical consequences. Besides rare hereditary connective
tissue disorders the role of ECM and its changes are now also emphasized in
common disease states such as fibrosis and cancer that cause a large economic
burden on the health care system.
In this study we generated an LH1 null mouse line, which is the first animal
model for human EDS VIA. Although the mice had generally milder symptoms,
the LH1 null mice and EDS VIA patients share the same life-threatening
phenotype i.e. risk of arterial ruptures. Structural weakness of arterial walls can
be explained by inadequate helical lysine hydroxylation and the accompanied
changes in the collagen cross-link pattern. Interestingly, in LH1 null mice the
correlation between the hydroxylysine content and the amount of HP cross-links,
which requires both telopeptide and helical hydroxylysine, was inconsistent. Thus,
in vivo the two other isoforms, LH2 and LH3, most probably can compensate for
the lack of LH1 activity, but the exact mechanism remains to be established.
Interestingly, in LH2 null mice we observed that LH1 and LH3 are not able to
compensate for at least the telopeptide LH activity. LH2 null mice died early
during development, which was surprising as the human disorder BS2 is
generally not lethal. However, our in vitro studies revealed that although the
known LH2 mutations in BS2 patients are destructive and able to explain the
clinical symptoms, they do not lead to complete inactivation of the enzyme. The
known mutations affect the folding of the enzyme or interfere with the binding of
a cosubstrate. Embyonic lethality of LH2 null mice raises the question whether a
total lack of LH2 would be lethal also in humans. The exact cause of death of
LH2 null mice remains unknown, but the results suggest that inadequate
telopeptide lysyl hydroxylation and thus formation of pyridinolin cross-links may
lead to a general weakness and collapse of tissue structures, particularly in the
vessel walls. Unfortunately, analysis of collagen cross-links from LH2 null mice
was impossible due to the early lethality, but the remarkable role of LH2 in crosslink
formation was seen with heterozygous mice that had surprisingly severe
changes in the cross-links, especially in the bone. Both mouse lines generated and
89

analyzed here emphasize the importance of LH1 and LH2 in proper cross-link
formation.
Although the LH2 gene-trap mouse line generated here has provided new
data about LH2 and emphasized its role in mouse development, the embryonic
lethal phenotype restricts studies of the contribution of LH2 to common
pathological situations such as fibrosis. It would be important also to know the
contribution of LH2 in tumorigenesis as the role of some other 2-oxoglutarate
dioxygenases in cancer is remarkable and because increased expression levels of
LH2 is connected to hypoxia and cancer. Thus, a conditional LH2 knock-out
mouse line could provide answers to some of these questions.
90

References
Akins ML, Luby-Phelps K, Bank RA & Mahendroo M (2011) Cervical softening during
pregnancy: regulated changes in collagen cross-linking and composition of
matricellular proteins in the mouse. Biol Reprod 84(5): 1053–1062.
Alanay Y, Avaygan H, Camacho N, Utine GE, Boduroglu K, Aktas D, Alikasifoglu M,
Tuncbilek E, Orhan D, Bakar FT, Zabel B, Superti-Furga A, Bruckner-Tuderman L,
Curry CJ, Pyott S, Byers PH, Eyre DR, Baldridge D, Lee B, Merrill AE, Davis EC,
Cohn DH, Akarsu N & Krakow D (2010) Mutations in the gene encoding the RER
protein FKBP65 cause autosomal-recessive osteogenesis imperfecta. Am J Hum
Genet 86(4):551–559. Erratum in Am J Hum Genet 87(4): 572–573.
Alanay Y & Krakow D (2010) Response to Shaheen et al. 2010. Am J Hum Genet 87(2):
308.
Andiran N, Alikasifoglu A, Alanay Y & Yordam N (2008) Cyclic pamidronate treatment
in Bruck syndrome: proposal of a new modality of treatment. Pediatr Int 50(6): 836–
838.
Arai KY & Nishiyama T (2007) Developmental changes in extracellular matrix messenger
RNAs in the mouse placenta during the second half of pregnancy: possible factors
involved in the regulation of placental extracellular matrix expression. Biol Reprod
77(6): 923–933.
Armstrong LC & Last JA (1995) Rat lysyl hydroxylase: molecular cloning, mRNA
distribution and expression in a baculovirus system. Biochim Biophys Acta 1264(1):
93–102.
Avery NC & Bailey AJ (2005) Enzymic and non-enzymic cross-linking mechanisms in
relation to turnover of collagen: relevance to aging and exercise. Scand J Med Sci
Sports 15(4): 231–240.
Badylak SF, Freytes DO & Gilbert TW (2009) Extracellular matrix as a biological scaffold
material: Structure and function. Acta Biomater 5(1): 1–13.
Bailey AJ (2002) Changes in bone collagen with age and disease. J Musculoskelet
Neuronal Interact 2(6): 529–531.
Bailey AJ, Paul RG & Knott L (1998) Mechanisms of maturation and ageing of collagen.
Mech Ageing Dev 106(1–2): 1–56.
Bank RA, Robins SP, Wijmenga C, Breslau-Siderius LJ, Bardoel AF, van der Sluijs HA,
Pruijs HE & TeKoppele JM (1999) Defective collagen crosslinking in bone, but not in
ligament or cartilage, in Bruck syndrome: indications for a bone-specific telopeptide
lysyl hydroxylase on chromosome 17. Proc Natl Acad Sci U S A 96(3): 1054–1058.
Baraniak AP, Chen JR & Garcia-Blanco MA (2006) Fox-2 mediates epithelial cell-specific
fibroblast growth factor receptor 2 exon choice. Mol Cell Biol 26(4): 1209–1222.
Barnes MJ, Constable BJ, Morton LF & Royce PM (1974) Age-related variations in
hydroxylation of lysine and proline in collagen. Biochem J 139(2): 461–468.
Basel D & Steiner RD (2009) Osteogenesis imperfecta: recent findings shed new light on
this once well-understood condition. Genet Med 11(6): 375–385.
91

Baumann M, Giunta C, Krabichler B, Rüschendorf F, Zoppi N, Colombi M, Bittner RE,
Quijano-Roy S, Muntoni F, Cirak S, Schreiber G, Zou Y, Hu Y, Romero NB, Carlier
RY, Amberger A, Deutschmann A, Straub V, Rohrbach M, Steinmann B, Rostásy K,
Karall D, Bönnemann CG, Zschocke J & Fauth C (2012) Mutations in FKBP14 Cause
a Variant of Ehlers-Danlos Syndrome with Progressive Kyphoscoliosis, Myopathy,
and Hearing Loss. Am J Hum Genet 90: 1–16.
Berg C, Geipel A, Noack F, Smrcek J, Krapp M, Germer U, Bender G & Gembruch U
(2005) Prenatal diagnosis of Bruck syndrome. Prenat Diagn 25(7): 535–538.
Blacksin MF, Pletcher BA & David M (1998) Osteogenesis imperfecta with joint
contractures: bruck syndrome. Pediatr Radiol 28(2): 117–119.
Bonaldo P, Braghetta P, Zanetti M, Piccolo S, Volpin D & Bressan GM (1998) Collagen
VI deficiency induces early onset myopathy in the mouse: an animal model for
Bethlem myopathy. Hum Mol Genet 7(13): 2135–2140.
Brady AF & Patton MA (1997) Osteogenesis imperfecta with arthrogryposis multiplex
congenita (Bruck syndrome)--evidence for possible autosomal recessive inheritance.
Clin Dysmorphol 6(4): 329–336.
Brenner RE, Vetter U, Stöss H, Müller PK & Teller WM (1993) Defective collagen fibril
formation and mineralization in osteogenesis imperfecta with congenital joint
contractures (Bruck syndrome). Eur J Pediatr 152(6): 505–508.
Breslau-Siderius EJ, Engelbert RH, Pals G & van der Sluijs JA (1998) Bruck syndrome: a
rare combination of bone fragility and multiple congenital joint contractures. J Pediatr
Orthop B 7(1): 35–38.
Brinckmann J, Açil Y, Feshchenko S, Katzer E, Brenner R, Kulozik A & Kügler S (1998)
Ehlers-Danlos syndrome type VI: lysyl hydroxylase deficiency due to a novel point
mutation (W612C). Arch Dermatol Res 290(4): 181–186.
Brinckmann J, Açil Y, Tronnier M, Notbohm H, Bätge B, Schmeller W, Koch MH, Müller
PK & Wolff HH (1996) Altered x-ray diffraction pattern is accompanied by a change
in the mode of cross-link formation in lipodermatosclerosis. J Invest Dermatol 107(4):
589–592.
Brinckmann J, Kim S, Wu J, Reinhardt DP, Batmunkh C, Metzen E, Notbohm H, Bank
RA, Krieg T & Hunzelmann N (2005) Interleukin 4 and prolonged hypoxia induce a
higher gene expression of lysyl hydroxylase 2 and an altered cross-link pattern:
important pathogenetic steps in early and late stage of systemic scleroderma? Matrix
Biol 24(7): 459–468.
Brinckmann J, Neess CM, Gaber Y, Sobhi H, Notbohm H, Hunzelmann N, Fietzek PP,
Müller PK, Risteli J, Gebker R & Scharffetter-Kochanek K (2001) Different pattern of
collagen cross-links in two sclerotic skin diseases: lipodermatosclerosis and
circumscribed scleroderma. J Invest Dermatol 117(2): 269–273.
Brinckmann J, Notbohm H, Tronnier M, Açil Y, Fietzek PP, Schmeller W, Müller PK &
Bätge B (1999) Overhydroxylation of lysyl residues is the initial step for altered
collagen cross-links and fibril architecture in fibrotic skin. J Invest Dermatol 113(4):
617–621.
92

Brodsky B & Persikov AV (2005) Molecular structure of the collagen triple helix. Adv
Protein Chem 70: 301–339.
Bunt S, Denholm B & Skaer H (2010) Characterisation of the Drosophila procollagen lysyl
hydroxylase, dPlod. Gene Expr Patterns 11(1–2): 72–78.
Butcher DT, Alliston T & Weaver VM (2009) A tense situation: forcing tumour
progression. Nat Rev Cancer 9(2): 108–122.
Canty EG & Kadler KE (2002) Collagen fibril biosynthesis in tendon: a review and recent
insights. Comp Biochem Physiol A Mol Integr Physiol 133(4): 979–985.
Canty EG & Kadler KE (2005) Procollagen trafficking, processing and fibrillogenesis. J
Cell Sci 118(Pt 7): 1341–1353.
Chakravarti S, Magnuson T, Lass JH, Jepsen KJ, LaMantia C & Carroll H (1998) Lumican
regulates collagen fibril assembly: skin fragility and corneal opacity in the absence of
lumican. J Cell Biol 141(5): 1277–1286.
Chang B, Chen Y, Zhao Y & Bruick RK (2007) JMJD6 is a histone arginine demethylase.
Science 318(5849): 444–447.
Christiansen HE, Schwarze U, Pyott SM, AlSwaid A, Al Balwi M, Alrasheed S, Pepin MG,
Weis MA, Eyre DR & Byers PH (2010) Homozygosity for a missense mutation in
SERPINH1, which encodes the collagen chaperone protein HSP47, results in severe
recessive osteogenesis imperfecta. Am J Hum Genet 86(3): 389–398.
Chu ML, Conway D, Pan TC, Baldwin C, Mann K, Deutzmann R & Timpl R (1988)
Amino acid sequence of the triple-helical domain of human collagen type VI. J Biol
Chem 263(35): 18601–18606.
Chung UI, Kawaguchi H, Takato T & Nakamura K (2004) Distinct osteogenic
mechanisms of bones of distinct origins. J Orthop Sci 9(4): 410–414.
Clifton IJ, McDonough MA, Ehrismann D, Kershaw NJ, Granatino N & Schofield CJ
(2006) Structural studies on 2-oxoglutarate oxygenases and related double-stranded
beta-helix fold proteins. J Inorg Biochem 100(4): 644–669.
Colpaert CG, Vermeulen PB, Fox SB, Harris AL, Dirix LY & Van Marck EA (2003) The
presence of a fibrotic focus in invasive breast carcinoma correlates with the expression
of carbonic anhydrase IX and is a marker of hypoxia and poor prognosis. Breast
Cancer Res Treat 81(2): 137–147.
Corsi A, Xu T, Chen XD, Boyde A, Liang J, Mankani M, Sommer B, Iozzo RV,
Eichstetter I, Robey PG, Bianco P & Young MF (2002) Phenotypic effects of biglycan
deficiency are linked to collagen fibril abnormalities, are synergized by decorin
deficiency, and mimic Ehlers-Danlos-like changes in bone and other connective
tissues. J Bone Miner Res 17(7): 1180–1189.
Coss MC, Winterstein D, Sowder RC 2 nd & Simek SL (1995) Molecular cloning, DNA
sequence analysis, and biochemical characterization of a novel 65-kDa FK506binding
protein (FKBP65). J Biol Chem 270(49): 29336–29341.
Daley WP, Peters SB & Larsen M (2008) Extracellular matrix dynamics in development
and regenerative medicine. J Cell Sci 121(Pt 3): 255–264.
93

Danielson KG, Baribault H, Holmes DF, Graham H, Kadler KE & Iozzo RV (1997)
Targeted disruption of decorin leads to abnormal collagen fibril morphology and skin
fragility. J Cell Biol 136(3): 729–743.
Datta V, Sinha A, Saili A & Nangia S (2005) Bruck syndrome. Indian J Pediatr 72(5):
441–442.
Davis EC, Broekelmann TJ, Ozawa Y & Mecham RP (1998) Identification of tropoelastin
as a ligand for the 65-kD FK506-binding protein, FKBP65, in the secretory pathway. J
Cell Biol 140(2): 295–303.
Deininger PL & Batzer MA (1999) Alu repeats and human disease. Mol Genet Metab
67(3): 183–193.
Del Gatto-Konczak F, Bourgeois CF, Le Guiner C, Kister L, Gesnel MC, Stévenin J &
Breathnach R (2000) The RNA-binding protein TIA-1 is a novel mammalian splicing
regulator acting through intron sequences adjacent to a 5' splice site. Mol Cell Biol
20(17): 6287–6299.
Dembure PP, Janko AR, Priest JH & Elsas LJ (1987) Ascorbate regulation of collagen
biosynthesis in Ehlers-Danlos syndrome, type VI. Metabolism 36(7): 687–691.
Denko NC, Fontana LA, Hudson KM, Sutphin PD, Raychaudhuri S, Altman R & Giaccia
AJ (2003) Investigating hypoxic tumor physiology through gene expression patterns.
Oncogene 22(37): 5907–5914.
Dong S, Nutt CL, Betensky RA, Stemmer-Rachamimov AO, Denko NC, Ligon KL,
Rowitch DH & Louis DN (2005) Histology-based expression profiling yields novel
prognostic markers in human glioblastoma. J Neuropathol Exp Neurol 64(11): 948–
955.
Engel J & Prockop DJ (1991) The zipper-like folding of collagen triple helices and the
effects of mutations that disrupt the zipper. Annu Rev Biophys Biophys Chem 20:
137–152.
Engler AJ, Sen S, Sweeney HL & Discher DE (2006) Matrix elasticity directs stem cell
lineage specification. Cell 126(4): 677–689.
Evens AM, Sehn LH, Farinha P, Nelson BP, Raji A, Lu Y, Brakman A, Parimi V, Winter
JN, Schumacker PT, Gascoyne RD & Gordon LI (2010) Hypoxia-inducible factor-1 α
expression predicts superior survival in patients with diffuse large B-cell lymphoma
treated with R-CHOP. J Clin Oncol 28(6): 1017–1024.
Eyre DR & Glimcher MJ (1972) Reducible crosslinks in hydroxylysine-deficient collagens
of a heritable disorder of connective tissue. Proc Natl Acad Sci U S A 69(9): 2594–
2598.
Eyre D, Shao P, Weis MA & Steinmann B (2002) The kyphoscoliotic type of Ehlers-
Danlos syndrome (type VI): differential effects on the hydroxylation of lysine in
collagens I and II revealed by analysis of cross-linked telopeptides from urine. Mol
Genet Metab 76(3): 211–216.
Eyre DR, Weis MA & Wu JJ (2010) Maturation of collagen Ketoimine cross-links by an
alternative mechanism to pyridinoline formation in cartilage. J Biol Chem 285(22):
16675–16682.
94

Eyre DR & Wu JJ (2005) Collagen cross-links. In: Brinckmann J, Notbohm H & Müller
PK (eds) Topics In Current Chemistry, Collagen: Primer in Structure, Processing and
Assembly. Springer-Verlag, Berlin Heidelberg, 247: 207–229.
Finkelman RD, Eason AL, Rakijian DR, Tutundzhyan Y & Hardesty RA (1994) Elevated
IGF-II and TGF-beta concentrations in human calvarial bone: potential mechanism for
increased graft survival and resistance to osteoporosis. Plast Reconstr Surg 93(4):
732–738.
Fitzgerald J, Rich C, Zhou FH & Hansen U (2008) Three novel collagen VI chains,
alpha4(VI), alpha5(VI), and alpha6(VI). J Biol Chem 283(29): 20170–20180.
Fleischmajer R, Perlish JS, Burgeson RE, Shaikh-Bahai F & Timpl R (1990) Type I and
type III collagen interactions during fibrillogenesis. Ann N Y Acad Sci 580: 161–175.
Förch P, Puig O, Kedersha N, Martínez C, Granneman S, Séraphin B, Anderson P &
Valcárcel J (2000) The apoptosis-promoting factor TIA-1 is a regulator of alternative
pre-mRNA splicing. Mol Cell 6(5): 1089–1098.
Gardner RL (1983) Origin and differentiation of extraembryonic tissues in the mouse. Int
Rev Exp Pathol 24: 63–133.
Gastinel LN, Bignon C, Misra AK, Hindsgaul O, Shaper JH & Joziasse DH (2001) Bovine
alpha1,3-galactosyltransferase catalytic domain structure and its relationship with
ABO histo-blood group and glycosphingolipid glycosyltransferases. EMBO J 20(4):
638–649.
Gelse K, Pöschl E & Aigner T (2003) Collagens--structure, function, and biosynthesis.
Adv Drug Deliv Rev 55(12): 1531–1546.
Gerriets JE, Curwin SL & Last JA (1993) Tendon hypertrophy is associated with increased
hydroxylation of nonhelical lysine residues at two specific cross-linking sites in type I
collagen. J Biol Chem 268(34): 25553–25560.
Giunta C, Elçioglu NH, Albrecht B, Eich G, Chambaz C, Janecke AR, Yeowell H, Weis M,
Eyre DR, Kraenzlin M & Steinmann B (2008) Spondylocheiro dysplastic form of the
Ehlers-Danlos syndrome--an autosomal-recessive entity caused by mutations in the
zinc transporter gene SLC39A13. Am J Hum Genet 82(6): 1290–1305.
Giunta C, Randolph A, Al-Gazali LI, Brunner HG, Kraenzlin ME & Steinmann B (2005b)
Nevo syndrome is allelic to the kyphoscoliotic type of the Ehlers-Danlos syndrome
(EDS VIA). Am J Med Genet A 133A(2): 158–164.
Giunta C, Randolph A & Steinmann B (2005a) Mutation analysis of the PLOD1 gene: an
efficient multistep approach to the molecular diagnosis of the kyphoscoliotic type of
Ehlers-Danlos syndrome (EDS VIA). Mol Genet Metab 86(1–2): 269–276.
Gordon MK & Hahn RA (2010) Collagens. Cell Tissue Res 339(1): 247–257.
Göthel SF & Marahiel MA (1999) Peptidyl-prolyl cis-trans isomerases, a superfamily of
ubiquitous folding catalysts. Cell Mol Life Sci 55(3): 423–436.
Ha VT, Marshall MK, Elsas LJ, Pinnell SR & Yeowell HN (1994) A patient with Ehlers-
Danlos syndrome type VI is a compound heterozygote for mutations in the lysyl
hydroxylase gene. J Clin Invest 93(4): 1716–1721.
95

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(2): 421–455.
Hanson DA & Eyre DR (1996) Molecular site specificity of pyridinoline and pyrrole crosslinks
in type I collagen of human bone. J Biol Chem 271(43): 26508–26516.
Hautala T, Byers MG, Eddy RL, Shows TB, Kivirikko KI & Myllylä R (1992a) Cloning of
human lysyl hydroxylase: complete cDNA-derived amino acid sequence and
assignment of the gene (PLOD) to chromosome 1p36.3----p36.2. Genomics 13(1):
62–69.
Hautala T, Heikkinen J, Kivirikko KI & Myllylä R (1992b) Minoxidil specifically
decreases the expression of lysine hydroxylase in cultured human skin fibroblasts.
Biochem J 283(Pt 1): 51–54.
Ha-Vinh R, Alanay Y, Bank RA, Campos-Xavier AB, Zankl A, Superti-Furga A & Bonafé
L (2004) Phenotypic and molecular characterization of Bruck syndrome (osteogenesis
imperfecta with contractures of the large joints) caused by a recessive mutation in
PLOD2. Am J Med Genet A 131(2): 115–120.
Heikkinen J, Hautala T, Kivirikko KI & Myllylä R (1994) Structure and expression of the
human lysyl hydroxylase gene (PLOD): introns 9 and 16 contain Alu sequences at the
sites of recombination in Ehlers-Danlos syndrome type VI patients. Genomics 24(3):
464–471.
Heikkinen J, Pousi B, Pope M & Myllylä R (1999) A null-mutated lysyl hydroxylase gene
in a compound heterozygote British patient with Ehlers-Danlos syndrome type VI.
Hum Mutat 14(4): 351.
Heikkinen J, Risteli M, Lampela O, Alavesa P, Karppinen M, Juffer AH & Myllylä R
(2011) Dimerization of human lysyl hydroxylase 3 (LH3) is mediated by the amino
acids 541–547. Matrix Biol 30(1): 27–33.
Heikkinen J, Risteli M, Wang C, Latvala J, Rossi M, Valtavaara M & Myllylä R (2000)
Lysyl hydroxylase 3 is a multifunctional protein possessing collagen
glucosyltransferase activity. J Biol Chem 275(46): 36158–36163.
Heikkinen J, Toppinen T, Yeowell H, Krieg T, Steinmann B, Kivirikko KI & Myllylä R
(1997) Duplication of seven exons in the lysyl hydroxylase gene is associated with
longer forms of a repetitive sequence within the gene and is a common cause for the
type VI variant of Ehlers-Danlos syndrome. Am J Hum Genet 60(1): 48–56.
Heise CT, Nicholls JR, Leamy CE & Wallis R (2000) Impaired secretion of rat mannosebinding
protein resulting from mutations in the collagen-like domain. J Immunol
165(3): 1403–1409.
Henkel W, Rauterberg J & Stirtz T (1976) Isolation of a crosslinked cyanogen-bromide
peptide from insoluble rabbit collagen. Tissue differences in hydroxylation and
glycosylation of the crosslink. Eur J Biochem 69(1): 223–231.
Hjalt TA, Amendt BA & Murray JC (2001) PITX2 regulates procollagen lysyl hydroxylase
(PLOD) gene expression: implications for the pathology of Rieger syndrome. J Cell
Biol 152(3): 545–552.
96

Hofbauer KH, Gess B, Lohaus C, Meyer HE, Katschinski D & Kurtz A (2003) Oxygen
tension regulates the expression of a group of procollagen hydroxylases. Eur J
Biochem 270(22): 4515–4522.
Hong X, Zang J, White J, Wang C, Pan CH, Zhao R, Murphy RC, Dai S, Henson P,
Kappler JW, Hagman J & Zhang G (2010) Interaction of JMJD6 with single-stranded
RNA. PNAS 107(33): 14568–14572.
Hunzelmann N & Brinckmann J (2011) What are the new milestones in the pathogenesis
of systemic sclerosis? Ann Rheum Dis 69 Suppl 1: i52–i56.
Hynes RO (2009) The extracellular matrix: not just pretty fibrils. Science 326(5957):
1216–1219.
Ihme A, Krieg T, Nerlich A, Feldmann U, Rauterberg J, Glanville RW, Edel G & Müller
PK (1984) Ehlers-Danlos syndrome type VI: collagen type specificity of defective
lysyl hydroxylation in various tissues. J Invest Dermatol 83(3): 161–165.
Inman KE & Downs KM (2007) The murine allantois: emerging paradigms in
development of the mammalian umbilical cord and its relation to the fetus. Genesis
45(5): 237–58.
Ishida Y & Nagata K (2011) Hsp47 as a collagen-specific molecular chaperone. Methods
Enzymol 499: 167–182.
Ishikawa Y, Vranka J, Wirz J, Nagata K & Bächinger HP (2008) The rough endoplasmic
reticulum-resident FK506-binding protein FKBP65 is a molecular chaperone that
interacts with collagens. J Biol Chem 283(46): 31584–31590.
Iwahashi M, Ooshima A & Nakano R (1996) Increase in the relative level of type V
collagen during development and ageing of the placenta. J Clin Pathol 49(11): 916–
919.
Izu Y, Ansorge HL, Zhang G, Soslowsky LJ, Bonaldo P, Chu ML & Birk DE (2011)
Dysfunctional tendon collagen fibrillogenesis in collagen VI null mice. Matrix Biol
30(1): 53–61.
Jarisch A, Giunta C, Zielen S, König R & Steinmann B (1998) Sibs affected with both
Ehlers-Danlos syndrome type IV and cystic fibrosis. Am J Med Genet 78(5): 455–460.
Järveläinen H, Sainio A, Koulu M, Wight TN & Penttinen R (2009) Extracellular matrix
molecules: potential targets in pharmacotherapy. Pharmacol Rev 61(2): 198–223.
Kadler KE, Baldock C, Bella J & Boot-Handford RP (2007) Collagens at a glance. J Cell
Sci 120(Pt 12): 1955–1958.
Kampranis SC & Tsichlis PN (2009) Histone demethylases and cancer. Adv Cancer Res
102: 103–169.
Kelley BP, Malfait F, Bonafe L, Baldridge D, Homan E, Symoens S, Willaert A, Elcioglu
N, Van Maldergem L, Verellen-Dumoulin C, Gillerot Y, Napierala D, Krakow D,
Beighton P, Superti-Furga A, De Paepe A & Lee B (2011) Mutations in FKBP10
cause recessive osteogenesis imperfecta and bruck syndrome. J Bone Miner Res 26(3):
666–672.
97

Kellokumpu S, Sormunen R, Heikkinen J & Myllylä R (1994) Lysyl hydroxylase, a
collagen processing enzyme, exemplifies a novel class of luminally-oriented
peripheral membrane proteins in the endoplasmic reticulum. J Biol Chem 269(48):
30524–30529.
Kielty CM & Grant ME (2002) The collagen family: structure, assembly, and organization
in the extracellular matrix. In: Royce PM & Steinmann B (eds) Connective Tissue and
Its Heritable Disorders. Molecular, Genetics and Medical Aspects. 2nd ed. Wiley-Liss,
Inc., New York, 159 –221.
Kivirikko KI & Myllylä R (1979) Collagen glycosyltransferases. Int Rev Connect Tissue
Res 8: 23–72.
Kivirikko KI & Myllylä R (1982) Posttranslational enzymes in the biosynthesis of collagen:
intracellular enzymes. Methods Enzymol 82(Pt A): 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 & Crabbe M
(eds) Post-Translational Modifications of Proteins. CRC Press, Inc., Boca Raton, 1–51.
Kivirikko KI & Pihlajaniemi T (1998) Collagen hydroxylases and the protein disulfide
isomerase subunit of prolyl-4-hydroxylase. Adv Enzymol Relat Areas Mol Biol 72:
325–398.
Knott L & Bailey AJ (1998) Collagen cross-links in mineralizing tissues: a review of their
chemistry, function, and clinical relevance. Bone 22(3): 181–187.
Knott L, Tarlton JF & Bailey AJ (1997) Chemistry of collagen cross-linking: biochemical
changes in collagen during the partial mineralization of turkey leg tendon. Biochem J
322(Pt 2): 535–542.
Koski MK, Hieta R, Böllner C, Kivirikko KI, Myllyharju J & Wierenga RK (2007) The
active site of an algal prolyl 4-hydroxylase has a large structural plasticity. J Biol
Chem 282(51): 37112–37123.
Krol BJ, Murad S, Walker LC, Marshall MK, Clark WL, Pinnell SR & Yeowell HN (1996)
The expression of a functional, secreted human lysyl hydroxylase in a baculovirus
system. J Invest Dermatol 106(1): 11–16.
Kronenberg HM (2003) Developmental regulation of the growth plate. Nature 423(6937):
332–336.
Lal A, Peters H, St Croix B, Haroon ZA, Dewhirst MW, Strausberg RL, Kaanders JH, van
der Kogel AJ & Riggins GJ (2001) Transcriptional response to hypoxia in human
tumors. J Natl Cancer Inst 93(17): 1337–1343.
Lampe AK & Bushby KM (2005) Collagen VI related muscle disorders. J Med Genet
42(9): 673–685.
Larraín J, Bachiller D, Lu B, Agius E, Piccolo S & De Robertis EM (2000) BMP-binding
modules in chordin: a model for signalling regulation in the extracellular space.
Development 127(4): 821–830.
Lauer-Fields JL, Malkar NB, Richet G, Drauz K & Fields GB (2003) Melanoma cell CD44
interaction with the alpha 1(IV)1263–1277 region from basement membrane collagen
is modulated by ligand glycosylation. J Biol Chem 278(16): 14321–14330.
98

Le Guiner C, Lejeune F, Galiana D, Kister L, Breathnach R, Stévenin J & Del Gatto-
Konczak F (2001) TIA-1 and TIAR activate splicing of alternative exons with weak 5'
splice sites followed by a U-rich stretch on their own pre-mRNAs. J Biol Chem
276(44): 40638–40646.
Lees JF, Tasab M & Bulleid NJ (1997) Identification of the molecular recognition
sequence which determines the type-specific assembly of procollagen. EMBO J 16(5):
908–916.
Leroy JG, Nuytinck L, De Paepe A, De Rammelaere M, Gillerot Y, Verloes A, Loeys B &
De Groote W (1998) Bruck syndrome: neonatal presentation and natural course in
three patients. Pediatr Radiol 28(10): 781–789.
Levental KR, Yu H, Kass L, Lakins JN, Egeblad M, Erler JT, Fong SF, Csiszar K, Giaccia
A, Weninger W, Yamauchi M, Gasser DL & Weaver VM (2009) Matrix crosslinking
forces tumor progression by enhancing integrin signaling. Cell 139(5): 891–906.
Liefhebber JM, Punt S, Spaan WJ & van Leeuwen HC (2010) The human collagen beta(1-
O)galactosyltransferase, GLT25D1, is a soluble endoplasmic reticulum localized
protein. BMC Cell Biol 11: 33.
Liu X, Wu H, Byrne M, Krane S & Jaenisch R (1997) Type III collagen is crucial for
collagen I fibrillogenesis and for normal cardiovascular development. Proc Natl Acad
Sci U S A 94(5): 1852–1856.
Löhler J, Timpl R & Jaenisch R (1984) Embryonic lethal mutation in mouse collagen I
gene causes rupture of blood vessels and is associated with erythropoietic and
mesenchymal cell death. Cell 38(2): 597–607.
Marini JC, Cabral WA & Barnes AM (2010) Null mutations in LEPRE1 and CRTAP
cause severe recessive osteogenesis imperfecta. Cell Tissue Res 339(1): 59–70.
McAlinden A, Havlioglu N, Liang L, Davies SR & Sandell LJ (2005) Alternative splicing
of type II procollagen exon 2 is regulated by the combination of a weak 5' splice site
and an adjacent intronic stem-loop cis element. J Biol Chem 280(38): 32700–32711.
McPherson E & Clemens M (1997) Bruck syndrome (osteogenesis imperfecta with
congenital joint contractures): review and report on the first North American case. Am
J Med Genet 70(1): 28–31.
Mercer DK, Nicol PF, Kimbembe C & Robins SP (2003) Identification, expression, and
tissue distribution of the three rat lysyl hydroxylase isoforms. Biochem Biophys Res
Commun 307(4): 803–809.
Mokete L, Robertson A, Viljoen D & Beighton P (2005) Bruck syndrome: congenital joint
contractures with bone fragility. J Orthop Sci 10(6): 641–646.
Morello R, Bertin TK, Chen Y, Hicks J, Tonachini L, Monticone M, Castagnola P, Rauch
F, Glorieux FH, Vranka J, Bächinger HP, Pace JM, Schwarze U, Byers PH, Weis M,
Fernandes RJ, Eyre DR, Yao Z, Boyce BF & Lee B (2006) CRTAP is required for
prolyl 3- hydroxylation and mutations cause recessive osteogenesis imperfecta. Cell
127(2): 291–304.
Murad S & Pinnell SR (1987) Suppression of fibroblast proliferation and lysyl hydroxylase
activity by minoxidil. J Biol Chem 262(25): 11973–11978.
99

Myers LK, Myllyharju J, Nokelainen M, Brand DD, Cremer MA, Stuart JM, Bodo M,
Kivirikko KI & Kang AH (2004) Relevance of posttranslational modifications for the
arthritogenicity of type II collagen. J Immunol 172(5): 2970–2975.
Myllyharju J (2005) Intracellular post-translational modifications of collagens. In:
Brinckmann J, Notbohm H & Müller PK (eds) Topics In Current Chemistry, Collagen:
Primer in Structure, Processing and Assembly. Springer-Verlag, Berlin Heidelberg,
247: 115–147.
Myllyharju J & Kivirikko KI (1997) Characterization of the iron- and 2-oxoglutaratebinding
sites of human prolyl 4-hydroxylase. EMBO J 16(6) :1173–1180.
Myllyharju J & Kivirikko KI (2001) Collagens and collagen-related diseases. Ann Med
33(1): 7–21.
Myllyharju J & Kivirikko KI (2004) Collagens, modifying enzymes and their mutations in
humans, flies and worms. Trends Genet 20(1): 33–43.
Myllylä R, Majamaa K, Günzler V, Hanauske-Abel HM & Kivirikko KI (1984) Ascorbate
is consumed stoichiometrically in the uncoupled reactions catalyzed by prolyl 4hydroxylase
and lysyl hydroxylase. J Biol Chem 259(9): 5403–5405.
Myllylä R, Pajunen L & Kivirikko KI (1988) Polyclonal and monoclonal antibodies to
human lysyl hydroxylase and studies on the molecular heterogeneity of the enzyme.
Biochem J 253(2): 489–496.
Myllylä R, Pihlajaniemi T, Pajunen L, Turpeenniemi-Hujanen T & Kivirikko KI (1991)
Molecular cloning of chick lysyl hydroxylase. Little homology in primary structure to
the two types of subunit of prolyl 4-hydroxylase. J Biol Chem 266(5): 2805–2810.
Mäki JM (2009) Lysyl oxidases in mammalian development and certain pathological
conditions. Histol Histopathol 24(5): 651–660.
Ng LJ, Tam PP & Cheah KS (1993) Preferential expression of alternatively spliced
mRNAs encoding type II procollagen with a cysteine-rich amino-propeptide in
differentiating cartilage and nonchondrogenic tissues during early mouse development.
Dev Biol 159(2): 403–417.
Niederreither K, D'Souza R, Metsäranta M, Eberspaecher H, Toman PD, Vuorio E & De
Crombrugghe B (1994) Coordinate patterns of expression of type I and III collagens
during mouse development. Matrix Biol 14(9): 705–713.
Noda T, Yamamoto H, Takemasa I, Yamada D, Uemura M, Wada H, Kobayashi S,
Marubashi S, Eguchi H, Tanemura M, Umeshita K, Doki Y, Mori M & Nagano H
(2012) PLOD2 induced under hypoxia is a novel prognostic factor for hepatocellular
carcinoma after curative resection. Liver Int 32(1): 110–118.
Norman KR & Moerman DG (2000) The let-268 locus of Caenorhabditis elegans encodes
a procollagen lysyl hydroxylase that is essential for type IV collagen secretion. Dev
Biol 227(2): 690–705.
Notbohm H, Nokelainen M, Myllyharju J, Fietzek PP, Müller PK & Kivirikko KI (1999)
Recombinant human type II collagens with low and high levels of hydroxylysine and
its glycosylated forms show marked differences in fibrillogenesis in vitro. J Biol
Chem 274(13): 8988–8992.
100

Parkin JD, San Antonio JD, Pedchenko V, Hudson B, Jensen ST & Savige J (2011)
Mapping structural landmarks, ligand binding sites and missense mutations to the
collagen IV heterotrimers predicts major functional domains, novel interactions and
variation in phenotypes in inherited diseases affecting basement membranes. Hum
Mutat 32(2): 127–143.
Pasquali M, Dembure PP, Still MJ & Elsas LJ (1994) Urinary pyridinium cross-links: a
noninvasive diagnostic test for Ehlers-Danlos syndrome type VI. N Engl J Med 331(2):
132–133.
Pasquali M, Still MJ, Dembure PP & Elsas LJ (1995) Pyridinium cross-links in heritable
disorders of collagen. Am J Hum Genet 57(6): 1508–1510.
Pasquali M, Still MJ, Vales T, Rosen RI, Evinger JD, Dembure PP, Longo N & Elsas LJ
(1997) Abnormal formation of collagen cross-links in skin fibroblasts cultured from
patients with Ehlers-Danlos syndrome type VI. Proc Assoc Am Physicians 109(1):
33–41.
Passoja K, Myllyharju J, Pirskanen A & Kivirikko KI (1998b) Identification of arginine-
700 as the residue that binds the C-5 carboxyl group of 2-oxoglutarate in human lysyl
hydroxylase 1. FEBS Lett 434(1–2): 145–148.
Passoja K, Rautavuoma K, Ala-Kokko L, Kosonen T & Kivirikko KI (1998a) Cloning and
characterization of a third human lysyl hydroxylase isoform. Proc Natl Acad Sci U S
A 95(18): 10482–10486.
Patterson CE, Schaub T, Coleman EJ & Davis EC (2000) Developmental regulation of
FKBP65. An ER-localized extracellular matrix binding-protein. Mol Biol Cell 11(11):
3925–3935.
Pinnell SR, Krane SM, Kenzora JE & Glimcher MJ (1972) A heritable disorder of
connective tissue. Hydroxylysine-deficient collagen disease. N Engl J Med 286(19):
1013–1020.
Pirskanen A, Kaimio AM, Myllylä R & Kivirikko KI (1996) Site-directed mutagenesis of
human lysyl hydroxylase expressed in insect cells. Identification of histidine residues
and an aspartic acid residue critical for catalytic activity. J Biol Chem 271(16): 9398–
9402.
Pornprasertsuk S, Duarte WR, Mochida Y & Yamauchi M (2004) Lysyl hydroxylase-2b
directs collagen cross-linking pathways in MC3T3-E1 cells. J Bone Miner Res 19(8):
1349–1355.
Pornprasertsuk S, Duarte WR, Mochida Y & Yamauchi M (2005) Overexpression of lysyl
hydroxylase-2b leads to defective collagen fibrillogenesis and matrix mineralization. J
Bone Miner Res 20(1): 81–87.
Pousi B, Hautala T, Heikkinen J, Pajunen L, Kivirikko KI & Myllylä R (1994) Alu-Alu
recombination results in a duplication of seven exons in the lysyl hydroxylase gene in
a patient with the type VI variant of Ehlers-Danlos syndrome. Am J Hum Genet 55(5):
899–906.
Pousi B, Heikkinen J, Schröter J, Pope M & Myllylä R (2000) A nonsense codon of exon
14 reduces lysyl hydroxylase mRNA and leads to aberrant RNA splicing in a patient
with Ehlers-Danlos syndrome type VI. Mutat Res 432(1–2): 33–37.
101

Pöschl E, Schlötzer-Schrehardt U, Brachvogel B, Saito K, Ninomiya Y & Mayer U (2004)
Collagen IV is essential for basement membrane stability but dispensable for initiation
of its assembly during early development. Development 131(7): 1619–1628.
Raghunath M, Bruckner P & Steinmann B (1994) Delayed triple helix formation of mutant
collagen from patients with osteogenesis imperfecta. J Mol Biol 236(3): 940–949.
Rajkumar T, Sabitha K, Vijayalakshmi N, Shirley S, Bose MV, Gopal G & Selvaluxmy G
(2011) Identification and validation of genes involved in cervical tumourigenesis.
BMC Cancer 11: 80.
Rautavuoma K, Passoja K, Helaakoski T & Kivirikko KI (2000) Complete exon-intron
organization of the gene for human lysyl hydroxylase 3 (LH3). Matrix Biol 19(1): 73–
79.
Rautavuoma K, Takaluoma K, Passoja K, Pirskanen A, Kvist AP, Kivirikko KI &
Myllyharju J (2002) Characterization of three fragments that constitute the monomers
of the human lysyl hydroxylase isoenzymes 1-3. The 30-kDa N-terminal fragment is
not required for lysyl hydroxylase activity. J Biol Chem 277(25): 23084–23091.
Rautavuoma K, Takaluoma K, Sormunen R, Myllyharju J, Kivirikko KI & Soininen R
(2004) Premature aggregation of type IV collagen and early lethality in lysyl
hydroxylase 3 null mice. Proc Natl Acad Sci U S A 101(39): 14120–14125.
Reiser K, McCormick RJ & Rucker RB (1992) Enzymatic and nonenzymatic cross-linking
of collagen and elastin. FASEB J 6(7): 2439–2449.
Ricard-Blum S (2011) The collagen family. Cold Spring Harb Perspect Biol 3(1): a004978.
Ricard-Blum S & Ruggiero F (2005) The collagen superfamily: from the extracellular
matrix to the cell membrane. Pathol Biol 53(7): 430–442.
Risteli M, Niemitalo O, Lankinen H, Juffer AH & Myllylä R (2004) Characterization of
collagenous peptides bound to lysyl hydroxylase isoforms. J Biol Chem 279(36):
37535–37543.
Risteli M, Ruotsalainen H, Salo AM, Sormunen R, Sipilä L, Baker NL, Lamandé SR,
Vimpari-Kauppinen L & Myllylä R (2009) Reduction of lysyl hydroxylase 3 causes
deleterious changes in the deposition and organization of extracellular matrix. J Biol
Chem 284(41): 28204–28211.
Robins SP (2007) Biochemistry and functional significance of collagen cross-linking.
Biochem Soc Trans 35(Pt 5): 849–852.
Rossant J & Cross JC (2001) Placental development: lessons from mouse mutants. Nat
Rev Genet 2(7): 538–548.
Roulet M, Ruggiero F, Karsenty G & LeGuellec D (2007) A comprehensive study of the
spatial and temporal expression of the col5a1 gene in mouse embryos: a clue for
understanding collagen V function in developing connective tissues. Cell Tissue Res
327(2): 323–332.
Royce PM & Barnes MJ (1985) Failure of highly purified lysyl hydroxylase to hydroxylate
lysyl residues in the non-helical regions of collagen. Biochem J 230(2): 475–480.
Rubin E & Farber JL (1999) Pathology. 3rd ed. Lippincott-Raven Publishers, Philadelphia,
37–75.
102

Ruotsalainen H, Sipilä L, Kerkelä E, Pospiech H & Myllylä R (1999) Characterization of
cDNAs for mouse lysyl hydroxylase 1, 2 and 3, their phylogenetic analysis and
tissuespecific expression in the mouse. Matrix Biol 18(3): 325–329.
Ruotsalainen H, Sipilä L, Vapola M, Sormunen R, Salo AM, Uitto L, Mercer DK, Robins
SP, Risteli M, Aszodi A, Fässler R & Myllylä R (2006) Glycosylation catalyzed by
lysyl hydroxylase 3 is essential for basement membranes. J Cell Sci 119(Pt 4): 625–
635.
Russell RG (2006) Bisphosphonates: from bench to bedside. Ann N Y Acad Sci 1068:
367–401.
Saito M & Marumo K (2010) Collagen cross-links as a determinant of bone quality: a
possible explanation for bone fragility in aging, osteoporosis, and diabetes mellitus.
Osteoporos Int 21(2): 195–214.
Salo AM, Cox H, Farndon P, Moss C, Grindulis H, Risteli M, Robins SP & Myllylä R
(2008) A connective tissue disorder caused by mutations of the lysyl hydroxylase 3
gene. Am J Hum Genet 83(4): 495–503.
Salo AM, Sipilä L, Sormunen R, Ruotsalainen H, Vainio S & Myllylä R (2006b) The lysyl
hydroxylase isoforms are widely expressed during mouse embryogenesis, but obtain
tissue- and cell-specific patterns in the adult. Matrix Biol 25(8): 475–483.
Salo AM, Wang C, Sipilä L, Sormunen R, Vapola M, Kervinen P, Ruotsalainen H,
Heikkinen J & Myllylä R (2006a) Lysyl hydroxylase 3 (LH3) modifies proteins in the
extracellular space, a novel mechanism for matrix remodeling. J Cell Physiol 207(3):
644–653.
Samimi A & Last JA (2001) Mechanism of inhibition of lysyl hydroxylase activity by the
organophosphates malathion and malaoxon. Toxicol Appl Pharmacol 176(3): 181–186.
Savolainen ER, Kero M, Pihlajaniemi T & Kivirikko KI (1981) Deficiency of
galactosylhydroxylysyl glucosyltransferase, an enzyme of collagen synthesis, in a
family with dominant epidermolysis bullosa simplex. N Engl J Med 304(4): 197–204.
Schegg B, Hülsmeier AJ, Rutschmann C, Maag C & Hennet T (2009) Core glycosylation
of collagen is initiated by two beta(1-O)galactosyltransferases. Mol Cell Biol 29(4):
943–952.
Scheurer SB, Rybak JN, Rösli C, Neri D & Elia G (2004) Modulation of gene expression
by hypoxia in human umbilical cord vein endothelial cells: A transcriptomic and
proteomic study. Proteomics 4(6): 1737–1760.
Schofield CJ & Ratcliffe PJ (2004) Oxygen sensing by HIF hydroxylases. Nat Rev Mol
Cell Biol 5(5): 343–354.
Schneider VA & Granato M (2006) The myotomal diwanka (lh3) glycosyltransferase and
type XVIII collagen are critical for motor growth cone migration. Neuron 50(5): 683–
695.
Schneider VA & Granato M (2007) Genomic structure and embryonic expression of
zebrafish lysyl hydroxylase 1 and lysyl hydroxylase 2. Matrix Biol 26(1): 12–19.
Schnieke A, Harbers K & Jaenisch R (1983) Embryonic lethal mutation in mice induced
by retrovirus insertion into the alpha 1(I) collagen gene. Nature 304(5924): 315–320.
103

Semenza GL (2011) Oxygen sensing, homeostasis, and disease. N Engl J Med 365(6):
537–547.
Seppinen L & Pihlajaniemi T (2011) The multiple functions of collagen XVIII in
development and disease. Matrix Biol 30(2): 83–92.
Seth P, Walker LC & Yeowell HN (2011) Identification of exonic cis-elements regulating
the alternative splicing of scleroderma-associated lysyl hydroxylase 2 mRNA. J Invest
Dermatol 131(2): 537–539.
Seth P & Yeowell HN (2010) Fox-2 protein regulates the alternative splicing of
scleroderma-associated lysyl hydroxylase 2 messenger RNA. Arthritis Rheum 62(4):
1167–1175.
Shaheen R, Al-Owain M, Sakati N, Alzayed ZS & Alkuraya FS (2010) FKBP10 and Bruck
syndrome: phenotypic heterogeneity or call for reclassification? Am J Hum Genet
87(2): 306–307. Erratum in Am J Hum Genet 87(4): 571.
Shoulders MD & Raines RT (2009) Collagen structure and stability. Annu Rev Biochem
78: 929–958.
Sipilä L, Ruotsalainen H, Sormunen R, Baker NL, Lamandé SR, Vapola M, Wang C, Sado
Y, Aszodi A & Myllylä R (2007) Secretion and assembly of type IV and VI collagens
depend on glycosylation of hydroxylysines. J Biol Chem 282(46): 33381–33388.
Sipilä L, Szatanik M, Vainionpää H, Ruotsalainen H, Myllylä R & Guénet JL (2000) The
genes encoding mouse lysyl hydroxylase isoforms map to chromosomes 4,5, and 9.
Mamm Genome 11(12): 1132–1134.
Sricholpech M, Perdivara I, Nagaoka H, Yokoyama M, Tomer KB & Yamauchi M (2011)
Lysyl hydroxylase 3 glucosylates galactosylhydroxylysine residues in type I collagen
in osteoblast culture. J Biol Chem 286(11): 8846–8856.
Steinmann B, Eyre DR & Shao P (1995) Urinary pyridinoline cross-links in Ehlers-Danlos
syndrome type VI. Am J Hum Genet 57(6): 1505–1508.
Steinmann B, Royce PM & Superti-Furga A (2002) The Ehlers-Danlos syndrome. In:
Royce PM & Steinmann B (eds) Connective tissue and its heritable disorders. 2nd ed.
Wiley-Liss, Inc., New York, 431–523.
Suokas M, Lampela O, Juffer AH, Myllylä R & Kellokumpu S (2003) Retrievalindependent
localization of lysyl hydroxylase in the endoplasmic reticulum via a
peptide fold in its iron-binding domain. Biochem J 370(Pt 3): 913–920.
Suokas M, Myllylä R & Kellokumpu S (2000) A single C-terminal peptide segment
mediates both membrane association and localization of lysyl hydroxylase in the
endoplasmic reticulum. J Biol Chem 275(23): 17863–17868.
Susic D (2007) Cross-link breakers as a new therapeutic approach to cardiovascular
disease. Biochem Soc Trans 35(Pt 5): 853–856.
Svensson L, Aszódi A, Reinholt FP, Fässler R, Heinegård D & Oldberg A (1999)
Fibromodulin-null mice have abnormal collagen fibrils, tissue organization, and
altered lumican deposition in tendon. J Biol Chem 274(14): 9636–9647.
Szpirer C, Szpirer J, Rivière M, Vanvooren P, Valtavaara M & Myllylä R (1997)
Localization of the gene encoding a novel isoform of lysyl hydroxylase. Mamm
Genome 8(9): 707–708.
104

Söderhäll C, Marenholz I, Kerscher T, Rüschendorf F, Esparza-Gordillo J, Worm M,
Gruber C, Mayr G, Albrecht M, Rohde K, Schulz H, Wahn U, Hubner N & Lee YA
(2007) Variants in a novel epidermal collagen gene (COL29A1) are associated with
atopic dermatitis. PLoS Biol 5(9): 1952–1961.
Takaluoma K, Lantto J & Myllyharju J (2007) Lysyl hydroxylase 2 is a specific
telopeptide hydroxylase, while all three isoenzymes hydroxylate collagenous
sequences. Matrix Biol 26(5): 396–403.
Torre-Blanco A, Adachi E, Hojima Y, Wootton JA, Minor RR & Prockop DJ (1992)
Temperature-induced post-translational over-modification of type I procollagen.
Effects of over-modification of the protein on the rate of cleavage by procollagen Nproteinase
and on self-assembly of collagen into fibrils. J Biol Chem 267(4): 2650–
2655.
Trojanowska M, LeRoy EC, Eckes B & Krieg T (1998) Pathogenesis of fibrosis: type 1
collagen and the skin. J Mol Med 76(3–4): 266–274.
Tsang KY, Cheung MC, Chan D & Cheah KS (2010) The developmental roles of the
extracellular matrix: beyond structure to regulation. Cell Tissue Res 339(1): 93–110.
Turpeenniemi-Hujanen TM, Puistola U & Kivirikko KI (1980) Isolation of lysyl
hydroxylase, an enzyme of collagen synthesis, from chick embryos as a homogeneous
protein. Biochem J 189(2): 247–253.
Turpeenniemi-Hujanen TM, Puistola U & Kivirikko KI (1981) Human lysyl hydroxylase:
purification to homogeneity, partial characterization and comparison of catalytic
properties with those of a mutant enzyme from Ehlers-Danlos syndrome type VI
fibroblasts. Coll Relat Res 1(4): 355–366.
Unligil UM, Zhou S, Yuwaraj S, Sarkar M, Schachter H & Rini JM (2000) X-ray crystal
structure of rabbit N-acetylglucosaminyltransferase I: catalytic mechanism and a new
protein superfamily. EMBO J 19(20): 5269–5280.
Uzawa K, Grzesik WJ, Nishiura T, Kuznetsov SA, Robey PG, Brenner DA & Yamauchi
M (1999) Differential expression of human lysyl hydroxylase genes, lysine
hydroxylation, and cross-linking of type I collagen during osteoblastic differentiation
in vitro. J Bone Miner Res 14(8): 1272–1280.
Uzawa K, Yeowell HN, Yamamoto K, Mochida Y, Tanzawa H & Yamauchi M (2003)
Lysine hydroxylation of collagen in a fibroblast cell culture system. Biochem Biophys
Res Commun 305(3): 484–487.
Valtavaara M, Papponen H, Pirttilä AM, Hiltunen K, Helander H & Myllylä R (1997)
Cloning and characterization of a novel human lysyl hydroxylase isoform highly
expressed in pancreas and muscle. J Biol Chem 272(11): 6831–6834.
Valtavaara M, Szpirer C, Szpirer J & Myllylä R (1998) Primary structure, tissue
distribution, and chromosomal localization of a novel isoform of lysyl hydroxylase
(lysyl hydroxylase 3). J Biol Chem 273(21): 12881–12886.
105

Vanacore RM, Friedman DB, Ham AJ, Sundaramoorthy M & Hudson BG (2005)
Identification of S-hydroxylysyl-methionine as the covalent cross-link of the
noncollagenous (NC1) hexamer of the alpha1alpha1alpha2 collagen IV network: a
role for the post-translational modification of lysine 211 to hydroxylysine 211 in
hexamer assembly. J Biol Chem 280(32): 29300–29310.
Vanacore R, Ham AJ, Voehler M, Sanders CR, Conrads TP, Veenstra TD, Sharpless KB,
Dawson PE & Hudson BG (2009) A sulfilimine bond identified in collagen IV.
Science 325(5945): 1230–1234.
Van Agtmael T & Bruckner-Tuderman L (2010) Basement membranes and human disease.
Cell Tissue Res 339(1): 167–188.
van den Bos T, Speijer D, Bank RA, Brömme D & Everts V (2008) Differences in matrix
composition between calvaria and long bone in mice suggest differences in
biomechanical properties and resorption: Special emphasis on collagen. Bone 43(3):
459–468.
van der Slot-Verhoeven AJ, van Dura EA, Attema J, Blauw B, Degroot J, Huizinga TW,
Zuurmond AM & Bank RA (2005) The type of collagen cross-link determines the
reversibility of experimental skin fibrosis. Biochim Biophys Acta 1740(1): 60–67.
van der Slot AJ, van Dura EA, de Wit EC, De Groot J, Huizinga TW, Bank RA &
Zuurmond AM (2005) Elevated formation of pyridinoline cross-links by profibrotic
cytokines is associated with enhanced lysyl hydroxylase 2b levels. Biochim Biophys
Acta 1741(1–2): 95–102.
van der Slot AJ, Zuurmond AM, Bardoel AF, Wijmenga C, Pruijs HE, Sillence DO,
Brinckmann J, Abraham DJ, Black CM, Verzijl N, DeGroot J, Hanemaaijer R,
TeKoppele JM, Huizinga TW & Bank RA (2003) Identification of PLOD2 as
telopeptide lysyl hydroxylase, an important enzyme in fibrosis. J Biol Chem 278(42):
40967–40972.
van der Slot AJ, Zuurmond AM, van den Bogaerdt AJ, Ulrich MM, Middelkoop E, Boers
W, Karel Ronday H, DeGroot J, Huizinga TW & Bank RA (2004) Increased
formation of pyridinoline cross-links due to higher telopeptide lysyl hydroxylase
levels is a general fibrotic phenomenon. Matrix Biol 23(4): 251–257.
van Vlimmeren MA, Driessen-Mol A, van den Broek M, Bouten CV & Baaijens FP (2010)
Controlling matrix formation and cross-linking by hypoxia in cardiovascular tissue
engineering. J Appl Physiol 109(5): 1483–1491.
Veit G, Kobbe B, Keene DR, Paulsson M, Koch M & Wagener R (2006) Collagen XXVIII,
a novel von Willebrand factor A domain-containing protein with many imperfections
in the collagenous domain. J Biol Chem 281(6): 3494–3504.
Viljoen D, Versfeld G & Beighton P (1989) Osteogenesis imperfecta with congenital joint
contractures (Bruck syndrome). Clin Genet 36(2): 122–126.
Vranka JA, Sakai LY & Bächinger HP (2004) Prolyl 3-hydroxylase 1, enzyme
characterization and identification of a novel family of enzymes. J Biol Chem 279(22):
23615–23621.
106

Walker LC, Marini JC, Grange DK, Filie J & Yeowell HN (1999) A patient with Ehlers-
Danlos syndrome type VI is homozygous for a premature termination codon in exon
14 of the lysyl hydroxylase 1 gene. Mol Genet Metab 67(1): 74–82.
Walker LC, Overstreet MA, Willing MC, Marini JC, Cabral WA, Pals G, Bristow J,
Atsawasuwan P, Yamauchi M & Yeowell HN (2004b) Heterogeneous basis of the
type VIB form of Ehlers-Danlos syndrome (EDS VIB) that is unrelated to decreased
collagen lysyl hydroxylation. Am J Med Genet A 131(2): 155–162.
Walker LC, Overstreet MA & Yeowell HN (2005) Tissue-specific expression and
regulation of the alternatively-spliced forms of lysyl hydroxylase 2 (LH2) in human
kidney cells and skin fibroblasts. Matrix Biol 23(8): 515–523.
Walker LC, Teebi AS, Marini JC, De Paepe A, Malfait F, Atsawasuwan P, Yamauchi M &
Yeowell HN (2004a) Decreased expression of lysyl hydroxylase 2 (LH2) in skin
fibroblasts from three Ehlers-Danlos patients does not result from mutations in either
the coding or proximal promoter region of the LH2 gene. Mol Genet Metab 83(4):
312–321.
Wang C, Kovanen V, Raudasoja P, Eskelinen S, Pospiech H & Myllylä R (2009) The
glycosyltransferase activities of lysyl hydroxylase 3 (LH3) in the extracellular space
are important for cell growth and viability. J Cell Mol Med 13(3): 508–521.
Wang C, Luosujärvi H, Heikkinen J, Risteli M, Uitto L & Myllylä R (2002a) The third
activity for lysyl hydroxylase 3: galactosylation of hydroxylysyl residues in collagens
in vitro. Matrix Biol 21(7): 559–566.
Wang C, Risteli M, Heikkinen J, Hussa AK, Uitto L & Myllyla R (2002b) Identification of
amino acids important for the catalytic activity of the collagen glucosyltransferase
associated with the multifunctional lysyl hydroxylase 3 (LH3). J Biol Chem 277(21):
18568–18573.
Wang C, Ristiluoma MM, Salo AM, Eskelinen S & Myllylä R (2012) Lysyl hydroxylase 3
is secreted from cells by two pathways. J Cell Physiol 227(2): 668–675.
Webby CJ, Wolf A, Gromak N, Dreger M, Kramer H, Kessler B, Nielsen ML, Schmitz C,
Butler DS, Yates JR 3rd, Delahunty CM, Hahn P, Lengeling A, Mann M, Proudfoot
NJ, Schofield CJ & Böttger A (2009) Jmjd6 catalyses lysyl-hydroxylation of U2AF65,
a protein associated with RNA splicing. Science 325(5936): 90–93.
Weis MA, Hudson DM, Kim L, Scott M, Wu JJ & Eyre DR (2010) Location of 3hydroxyproline
residues in collagen types I, II, III, and V/XI implies a role in fibril
supramolecular assembly. J Biol Chem 285(4): 2580–2590.
Wenstrup RJ, Florer JB, Brunskill EW, Bell SM, Chervoneva I & Birk DE (2004) Type V
collagen controls the initiation of collagen fibril assembly. J Biol Chem 279(51):
53331–53337.
Wenstrup RJ, Florer JB, Davidson JM, Phillips CL, Pfeiffer BJ, Menezes DW, Chervoneva
I & Birk DE (2006) Murine model of the Ehlers-Danlos syndrome. col5a1
haploinsufficiency disrupts collagen fibril assembly at multiple stages. J Biol Chem
281(18): 12888–12895.
Wenstrup RJ, Murad S & Pinnell SR (1989) Ehlers-Danlos syndrome type VI: clinical
manifestations of collagen lysyl hydroxylase deficiency. J Pediatr 115(3): 405–409.
107

Wilson R, Lees JF & Bulleid NJ (1998) Protein disulfide isomerase acts as a molecular
chaperone during the assembly of procollagen. J Biol Chem 273(16): 9637–9643.
Wu J, Reinhardt DP, Batmunkh C, Lindenmaier W, Far RK, Notbohm H, Hunzelmann N
& Brinckmann J (2006) Functional diversity of lysyl hydroxylase 2 in collagen
synthesis of human dermal fibroblasts. Exp Cell Res 312(18): 3485–3494.
Yamauchi M & Katz EP (1993) The post-translational chemistry and molecular packing of
mineralizing tendon collagens. Connect Tissue Res 29(2): 81–98.
Yamauchi M, Noyes C, Kuboki Y & Mechanic GL (1982) Collagen structural
microheterogeneity and a possible role for glycosylated hydroxylysine in type I
collagen. Proc Natl Acad Sci U S A 79(24): 7684–7688.
Yapicioğlu H, Ozcan K, Arikan O, Satar M, Narli N & Ozbek MH (2009) Bruck syndrome:
osteogenesis imperfecta and arthrogryposis multiplex congenita. Ann Trop Paediatr
29(2): 159–162.
Yeowell HN, Allen JD, Walker LC, Overstreet MA, Murad S & Thai SF (2000a) Deletion
of cysteine 369 in lysyl hydroxylase 1 eliminates enzyme activity and causes Ehlers-
Danlos syndrome type VI. Matrix Biol 19(1): 37–46.
Yeowell HN, Ha V, Clark WL, Marshall MK & Pinnell SR (1994) Sequence analysis of a
cDNA for lysyl hydroxylase isolated from human skin fibroblasts from a normal
donor: differences from human placental lysyl hydroxylase cDNA. J Invest Dermatol
102(3): 382–384.
Yeowell HN & Walker LC (1997) Ehlers-Danlos syndrome type VI results from a
nonsense mutation and a splice site-mediated exon-skipping mutation in the lysyl
hydroxylase gene. Proc Assoc Am Physicians 109(4): 383–396.
Yeowell HN & Walker LC (1999b) Prenatal exclusion of Ehlers-Danlos syndrome type VI
by mutational analysis. Proc Assoc Am Physicians 111(1): 57–62.
Yeowell HN & Walker LC (1999a) Tissue specificity of a new splice form of the human
lysyl hydroxylase 2 gene. Matrix Biol 18(2): 179–187.
Yeowell HN & Walker LC (2000) Mutations in the lysyl hydroxylase 1 gene that result in
enzyme deficiency and the clinical phenotype of Ehlers-Danlos syndrome type VI.
Mol Genet Metab 71(1–2): 212–224.
Yeowell HN, Walker LC, Farmer B, Heikkinen J & Myllyla R (2000b) Mutational analysis
of the lysyl hydroxylase 1 gene (PLOD) in six unrelated patients with Ehlers-Danlos
syndrome type VI: prenatal exclusion of this disorder in one family. Hum Mutat 16(1):
90.
Yeowell HN, Walker LC, Marshall MK, Murad S & Pinnell SR (1995) The mRNA and the
activity of lysyl hydroxylase are up-regulated by the administration of ascorbate and
hydralazine to human skin fibroblasts from a patient with Ehlers-Danlos syndrome
type VI. Arch Biochem Biophys 321(2): 510–516.
Yeowell HN, Walker LC, Mauger DM, Seth P & Garcia-Blanco MA (2009) TIA nuclear
proteins regulate the alternate splicing of lysyl hydroxylase 2. J Invest Dermatol
129(6): 1402–1411.
108

Yeowell HN, Walker LC, Murad S & Pinnell SR (1997) A common duplication in the
lysyl hydroxylase gene of patients with Ehlers Danlos syndrome type VI results in
preferential stimulation of lysyl hydroxylase activity and mRNA by hydralazine. Arch
Biochem Biophys 347(1): 126–131.
Zeng B, MacDonald JR, Bann JG, Beck K, Gambee JE, Boswell BA & Bächinger HP
(1998) Chicken FK506-binding protein, FKBP65, a member of the FKBP family of
peptidylprolyl cis-trans isomerases, is only partially inhibited by FK506. Biochem J
330(Pt 1): 109–114.
Zhu Y, Oganesian A, Keene DR & Sandell LJ (1999) Type IIA procollagen containing the
cysteine-rich amino propeptide is deposited in the extracellular matrix of
prechondrogenic tissue and binds to TGF-beta1 and BMP-2. J Cell Biol 144(5): 1069–
1080.
Zuurmond AM, van der Slot-Verhoeven AJ, van Dura EA, De Groot J & Bank RA (2005)
Minoxidil exerts different inhibitory effects on gene expression of lysyl hydroxylase 1,
2, and 3: implications for collagen cross-linking and treatment of fibrosis. Matrix Biol
24(4): 261–270.
109

110

Original articles
I Takaluoma K, Hyry M, Lantto J, Sormunen R, Bank RA, Kivirikko KI, Myllyharju J
& Soininen R (2007) Tissue-specific changes in the hydroxylysine content and crosslinks
of collagens and alterations in fibril morphology in lysyl hydroxylase 1 knockout
mice. J Biol Chem 282(9):6588–6596.
II Hyry M, Lantto J & Myllyharju J (2009) Missense mutations that cause Bruck
syndrome affect enzymatic activity, folding, and oligomerization of lysyl hydroxylase
2. J Biol Chem 284(45):30917–30924.
III Hyry M, Soininen R, Bank RA, Sormunen R, Miinalainen I, Talsta HL, Lantto J &
Myllyharju J (2012) Lack of lysyl hydroxylase 2 in mouse leads to early lethality
during embryonic development. Manuscript.
Reprinted with permission from American Society for Biochemistry and
Molecular Biology.
Original publications are not included in the electrical version of the dissertation.
111

112

ACTA UNIVERSITATIS OULUENSIS
SERIES D MEDICA
1143. Karjalainen, Minna (2011) Genetic predisposition to spontaneous preterm birth :
approaches to identify susceptibility genes
1144. Saaristo, Timo (2011) Assessment of risk and prevention of type 2 diabetes in
primary health care
1145. Vuononvirta, Tiina (2011) Etäterveydenhuollon käyttöönotto terveydenhuollon
verkostoissa
1146. Vanhala, Marja (2012) Lapsen ylipaino – riskitekijät, tunnistaminen ja elintavat
1147. Katisko, Jani (2012) Intraoperative imaging guided delineation and localization of
regions of surgical interest : Feasibility study
1148. Holmström, Anneli (2012) Etnografinen tutkimus natiivitutkimusten oppimisesta
röntgenhoitajaopiskelijoiden opinnoissa
1149. Ronkainen, Hanna-Leena (2012) Novel prognostic biomarkers for renal cell
carcinoma
1150. Murtomaa-Hautala, Mari (2012) Species-specific effects of dioxin exposure on
xenobiotic metabolism and hard tissue in voles
1151. Jauhola, Outi (2012) Henoch-Schönlein purpura in children
1152. Toljamo, Kari (2012) Gastric erosions – clinical significance and pathology : a
long-term follow-up study
1153. Meriläinen, Merja (2012) Tehohoitopotilaan hoitoympäristö : psyykkinen
elämänlaatu ja toipuminen
1154. Liisanantti, Janne (2012) Acute drug poisoning: outcome and factors affecting
outcome
1155. Markkanen, Piia (2012) Human δ opioid receptor : The effect of Phe27Cys
polymorphism, N-linked glycosylation and SERCA2b interaction on receptor
processing and trafficking
1156. Hannula, Annukka (2012) Imaging studies of the urinary tract in children with
acute urinary tract infection
1157. Korvala, Johanna (2012) In quest of genetic susceptibility to disorders manifesting
in fractures : Assessing the significance of genetic factors in femoral neck stress
fractures and childhood non-OI primary osteoporosis
1158. Takala, Heikki (2012) Biomarkers in esophageal cancer
Book orders:
Granum: Virtual book store
http://granum.uta.fi/granum/

UNIVERSITY OF OULU P.O.B. 7500 FI-90014 UNIVERSITY OF OULU FINLAND
A C T A U N I V E R S I T A T I S O U L U E N S I S
S E R I E S E D I T O R S
A
B
C
D
E
F
G
SCIENTIAE RERUM NATURALIUM
HUMANIORA
TECHNICA
MEDICA
Senior Assistant Jorma Arhippainen
SCIENTIAE RERUM SOCIALIUM
SCRIPTA ACADEMICA
OECONOMICA
Lecturer Santeri Palviainen
Professor Hannu Heusala
Professor Olli Vuolteenaho
Senior Researcher Eila Estola
Director Sinikka Eskelinen
Professor Jari Juga
EDITOR IN CHIEF
Professor Olli Vuolteenaho
PUBLICATIONS EDITOR
Publications Editor Kirsti Nurkkala
ISBN 978-951-42-9841-7 (Paperback)
ISBN 978-951-42-9842-4 (PDF)
ISSN 0355-3221 (Print)
ISSN 1796-2234 (Online)