Lysyl hydroxylases 1 and 2 : Characterization of their in vivo ... - Oulun
Lysyl hydroxylases 1 and 2 : Characterization of their in vivo ... - Oulun
Lysyl hydroxylases 1 and 2 : Characterization of their in vivo ... - Oulun
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
ACTA<br />
Marjo Hyry<br />
OULU 2012<br />
D 1159<br />
UNIVERSITATIS OULUENSIS<br />
D<br />
LYSYL HYDROXYLASES 1 AND 2<br />
CHARACTERIZATION OF THEIR IN VIVO ROLES<br />
IN MOUSE AND THE MOLECULAR LEVEL<br />
CONSEQUENCES OF THE LYSYL HYDROXYLASE 2<br />
MUTATIONS FOUND IN BRUCK SYNDROME<br />
UNIVERSITY OF OULU GRADUATE SCHOOL;<br />
UNIVERSITY OF OULU, FACULTY OF MEDICINE,<br />
INSTITUTE OF BIOMEDICINE,<br />
DEPARTMENT OF MEDICAL BIOCHEMISTRY AND MOLECULAR BIOLOGY;<br />
BIOCENTER OULU;<br />
CENTER FOR CELL-MATRIX RESEARCH<br />
MEDICA
ACTA UNIVERSITATIS OULUENSIS<br />
D Medica 1159<br />
MARJO HYRY<br />
LYSYL HYDROXYLASES 1 AND 2<br />
<strong>Characterization</strong> <strong>of</strong> <strong>their</strong> <strong>in</strong> <strong>vivo</strong> roles <strong>in</strong> mouse <strong>and</strong> the<br />
molecular level consequences <strong>of</strong> the lysyl hydroxylase 2<br />
mutations found <strong>in</strong> Bruck syndrome<br />
Academic dissertation to be presented with the assent<br />
<strong>of</strong> the Doctoral Tra<strong>in</strong><strong>in</strong>g Committee <strong>of</strong> Health <strong>and</strong><br />
Biosciences <strong>of</strong> the University <strong>of</strong> Oulu for public defence<br />
<strong>in</strong> Auditorium F101 <strong>of</strong> the Department <strong>of</strong> Physiology<br />
(Aapistie 7), on 8 June 2012, at 10 a.m.<br />
UNIVERSITY OF OULU, OULU 2012
Copyright © 2012<br />
Acta Univ. Oul. D 1159, 2012<br />
Supervised by<br />
Pr<strong>of</strong>essor Johanna Myllyharju<br />
Reviewed by<br />
Pr<strong>of</strong>essor Heather Yeowell<br />
Pr<strong>of</strong>essor Helen Skaer<br />
ISBN 978-951-42-9841-7 (Paperback)<br />
ISBN 978-951-42-9842-4 (PDF)<br />
ISSN 0355-3221 (Pr<strong>in</strong>ted)<br />
ISSN 1796-2234 (Onl<strong>in</strong>e)<br />
Cover Design<br />
Raimo Ahonen<br />
JUVENES PRINT<br />
TAMPERE 2012
Hyry, Marjo, <strong>Lysyl</strong> <strong>hydroxylases</strong> 1 <strong>and</strong> 2. <strong>Characterization</strong> <strong>of</strong> <strong>their</strong> <strong>in</strong> <strong>vivo</strong> roles <strong>in</strong><br />
mouse <strong>and</strong> the molecular level consequences <strong>of</strong> the lysyl hydroxylase 2 mutations<br />
found <strong>in</strong> Bruck syndrome<br />
University <strong>of</strong> Oulu Graduate School; University <strong>of</strong> Oulu, Faculty <strong>of</strong> Medic<strong>in</strong>e, Institute <strong>of</strong><br />
Biomedic<strong>in</strong>e, Department <strong>of</strong> Medical Biochemistry <strong>and</strong> Molecular Biology; Biocenter Oulu;<br />
Center for Cell-Matrix Research, P.O. Box 5000, FI-90014 Oulu, F<strong>in</strong>l<strong>and</strong><br />
Acta Univ. Oul. D 1159, 2012<br />
Oulu, F<strong>in</strong>l<strong>and</strong><br />
Abstract<br />
The extracellular matrix is not just a scaffold for cells <strong>and</strong> tissues, but rather a dynamic part <strong>of</strong> the<br />
human body. Characteristics <strong>of</strong> collagens, the major prote<strong>in</strong> components <strong>of</strong> the extracellular<br />
matrix, are determ<strong>in</strong>ed already dur<strong>in</strong>g synthesis <strong>and</strong> mutations <strong>in</strong> genes encod<strong>in</strong>g collagens,<br />
unbalance <strong>of</strong> regulators or dysfunction <strong>of</strong> collagen modify<strong>in</strong>g enzymes, for <strong>in</strong>stance, can lead to<br />
severe cl<strong>in</strong>ical complications. Certa<strong>in</strong> hydroxylys<strong>in</strong>e residues formed by lysyl <strong>hydroxylases</strong> (LHs)<br />
function <strong>in</strong> collagens as precursors <strong>of</strong> collagen cross-l<strong>in</strong>ks that stabilize collagenous structures <strong>and</strong><br />
thereby tissues. In humans, a deficiency <strong>of</strong> LH1, which is known to hydroxylate lys<strong>in</strong>es <strong>in</strong> the<br />
helical regions <strong>of</strong> collagen polypeptides, causes Ehlers-Danlos syndrome VIA (EDS VIA). It is<br />
characterized e.g. by progressive kyphoscoliosis <strong>and</strong> hypermobile jo<strong>in</strong>ts. Mutations <strong>in</strong> LH2, which<br />
is known to hydroxylate lys<strong>in</strong>es <strong>in</strong> the telopeptides <strong>of</strong> collagen polypeptides, cause Bruck<br />
syndrome type 2 (BS2). BS2 patients suffer from fragile bones <strong>and</strong> congenital jo<strong>in</strong>t contractures,<br />
for <strong>in</strong>stance, but the syndrome is usually not lethal.<br />
In this work we have generated <strong>and</strong> analyzed genetically modified LH1 <strong>and</strong> LH2 null mouse<br />
l<strong>in</strong>es to study the <strong>in</strong> <strong>vivo</strong> functions <strong>and</strong> roles <strong>of</strong> these enzymes. Analyses concentrated also on<br />
collagen cross-l<strong>in</strong>ks that were determ<strong>in</strong>ed from several null or heterozygous mouse tissues. In the<br />
present work we also studied the effects <strong>of</strong> known BS2 mutations on recomb<strong>in</strong>ant human LH2<br />
polypeptides to underst<strong>and</strong> the molecular pathology <strong>of</strong> the syndrome.<br />
As an animal model for human EDS VIA, LH1 null mice had certa<strong>in</strong> characteristics typical for<br />
EDS VIA, such as muscular hypotonia, but generally the symptoms were milder. Like EDS VIA<br />
patients, the mice have an <strong>in</strong>creased risk <strong>of</strong> arterial ruptures <strong>and</strong> ultrastructural changes can be seen<br />
<strong>in</strong> the wall <strong>of</strong> the aorta, expla<strong>in</strong>ed by <strong>in</strong>adequate helical lys<strong>in</strong>e hydroxylation accompanied by a<br />
changed cross-l<strong>in</strong>k<strong>in</strong>g state <strong>of</strong> tissues. Similarly, analysis <strong>of</strong> the LH2 null mouse l<strong>in</strong>e demonstrated<br />
the importance <strong>of</strong> the enzyme <strong>in</strong> cross-l<strong>in</strong>k formation. We showed that even a reduced amount <strong>of</strong><br />
LH2 <strong>in</strong> adult mice changes the cross-l<strong>in</strong>k<strong>in</strong>g pattern <strong>in</strong> tissues <strong>and</strong> a total lack <strong>of</strong> the enzyme leads<br />
to embryonic lethality. Furthermore, we demonstrated that LH2 is particularly important <strong>in</strong> tissue<br />
structures support<strong>in</strong>g blood vessels <strong>in</strong> the develop<strong>in</strong>g mouse embryo or <strong>in</strong> extraembryonic tissues.<br />
F<strong>in</strong>ally, our <strong>in</strong> vitro studies with recomb<strong>in</strong>ant human LH2 polypeptides revealed that the known<br />
BS2 mutations severely affect the activity <strong>of</strong> the enzyme thus expla<strong>in</strong><strong>in</strong>g the cl<strong>in</strong>ical symptoms <strong>of</strong><br />
the patients, but the mutations do not lead to a total <strong>in</strong>activation <strong>of</strong> the enzyme, which may be<br />
critical for the survival <strong>of</strong> patients.<br />
Keywords: Bruck syndrome, collagen, collagen cross-l<strong>in</strong>k, connective tissue disorder,<br />
Ehlers-Danlos syndrome type VIA, lysyl hydroxylase
Hyry, Marjo, Lysyylihydroksylaasit 1 ja 2.<br />
<strong>Oulun</strong> yliopiston tutkijakoulu; <strong>Oulun</strong> yliopisto, Lääketieteell<strong>in</strong>en tiedekunta, Biolääketieteen<br />
laitos, Lääketieteell<strong>in</strong>en biokemia ja molekyylibiologia; Biocenter Oulu; Center for Cell-Matrix<br />
Research, PL 5000, 90014 Oulu<br />
Acta Univ. Oul. D 1159, 2012<br />
Oulu<br />
Tiivistelmä<br />
Solunulko<strong>in</strong>en matriksi ei ole a<strong>in</strong>oastaan soluja ja kudoksia tukeva rakenne, vaan se on dynaam<strong>in</strong>en<br />
osa ihmiskehoa. Kollageenien, solunulkoisen matriks<strong>in</strong> yleisimpien protei<strong>in</strong>ien om<strong>in</strong>aisuudet<br />
määräytyvät jo kollageenien synteesivaiheessa ja mutaatiot kollageeneja koodittavissa<br />
geeneissä, säätelytekijöiden epätasapa<strong>in</strong>o tai esimerkiksi kollageeneja muokkaavien entsyymien<br />
toim<strong>in</strong>tahäiriöt voivat johtaa vaikeisi<strong>in</strong> kli<strong>in</strong>isi<strong>in</strong> komplikaatioih<strong>in</strong>. Tietyt lysyylihydroksylaasien<br />
(LH) muodostamat hydroksilysi<strong>in</strong>itähteet toimivat kollageeneissa kollageeniristisidosten esiaste<strong>in</strong>a.<br />
Ristisidokset vakauttavat kollageenirakenteita ja siten myös kudoksia. LH1 hydroksyloi<br />
lysi<strong>in</strong>ejä kollageenipolypeptidien kolmoiskierteisellä alueella ja ihmisellä entsyym<strong>in</strong> puutos<br />
aiheuttaa tyyp<strong>in</strong> VIA Ehlers-Danlos<strong>in</strong> syndrooman (EDS VIA), jossa potilailla on esimerkiksi<br />
etenevää kyfoskolioosia ja yliliikkuvat nivelet. Mutaatiot LH2-entsyymissä, joka hydroksyloi<br />
lysi<strong>in</strong>ejä kollageenipolypeptidien telopeptidialueilla, aiheuttavat tyyp<strong>in</strong> 2 Bruck<strong>in</strong> syndrooman<br />
(BS2). BS2-potilaat kärsivät mm. luiden hauraudesta ja niveljäykkyydestä, mutta syndrooma ei<br />
yleensä ole letaali.<br />
Tässä työssä loimme ja analysoimme geneettisesti muunnellut LH1 ja LH2 hiiril<strong>in</strong>jat, joiden<br />
kyse<strong>in</strong>en LH-geeniaktiivisuus on hiljennetty. L<strong>in</strong>jojen avulla halusimme tutkia näiden entsyymien<br />
toim<strong>in</strong>taa ja merkitystä <strong>in</strong> <strong>vivo</strong>. Analyysit keskittyivät myös kollageeniristisidoksi<strong>in</strong>, joita tutkitti<strong>in</strong><br />
useista poistogeenisten tai heterotsygoottisten hiirten kudoksista. Ymmärtääksemme<br />
BS2:n molekyylipatologiaa, tutkimme tässä työssä myös tunnettujen BS2-mutaatioiden vaikutuksia<br />
ihmisen LH2-rekomb<strong>in</strong>anttiprotei<strong>in</strong>issa.<br />
EDS VIA:n elä<strong>in</strong>mall<strong>in</strong>a LH1 poistogeenisillä hiirillä on joitak<strong>in</strong> om<strong>in</strong>aisuuksia, kuten lihashypotonia,<br />
jotka ovat tyypillisiä EDS VIA:lle, mutta yleisesti oireet ovat lievempiä. Kuten EDS<br />
VIA-potilailla, hiirillä on kohonnut valtimoiden repeytymisriski ja aortan se<strong>in</strong>ämän ultrarakenteessa<br />
voidaank<strong>in</strong> havaita muutoksia. Oireita voidaan selittää riittämättömällä kollageenien kolmoiskierteisen<br />
alueen lysi<strong>in</strong>ien hydroksylaatiolla, joka muuttaa kollageenien ristisidostilaa<br />
kudoksissa. Myös LH2-hiiril<strong>in</strong>jan analyso<strong>in</strong>ti osoitti kyseisen entsyym<strong>in</strong> tärkeyden ristisidosten<br />
muodostamisessa. Jo alentunut LH2:n määrä aikuisissa hiirissä muuttaa kudosten kollageeniristisidoksia<br />
ja täydell<strong>in</strong>en entsyym<strong>in</strong> puuttum<strong>in</strong>en johtaa sikiön kuolemaan. Lisäksi osoitimme,<br />
että LH2 on erityisen tärkeä kudosrakenteissa, jotka tukevat kehittyvän hiiren sikiön tai sikiön<br />
ulkopuolisten kudosten verisuonia. In vitro-tutkimukset ihmisen LH2-rekomb<strong>in</strong>anttiprotei<strong>in</strong>illa<br />
paljastivat, että tunnetut BS2-mutaatiot vaikuttavat erittä<strong>in</strong> haitallisesti entsyym<strong>in</strong> toim<strong>in</strong>taan,<br />
mikä selittää potilaiden kli<strong>in</strong>iset oireet, mutta mutaatiot eivät kuitenkaan aiheuta entsyym<strong>in</strong> täydellistä<br />
<strong>in</strong>aktivaatiota, mikä voi olla kriittistä potilaiden selviytymisen kannalta.<br />
Asiasanat: Bruck<strong>in</strong> syndrooma, kollageeni, kollageeniristisidos, lysyylihydroksylaasi,<br />
sidekudossairaus, tyyp<strong>in</strong> VIA Ehlers-Danlos<strong>in</strong> syndrooma
Acknowledgements<br />
This research was carried out at the Institute <strong>of</strong> Biomedic<strong>in</strong>e, Department <strong>of</strong><br />
Medical Biochemistry <strong>and</strong> Molecular Biology, University <strong>of</strong> Oulu, dur<strong>in</strong>g years<br />
2005–2012.<br />
I wish to express my deepest gratitude to my excellent supervisor, Pr<strong>of</strong>essor<br />
Johanna Myllyharju, who gave me the opportunity to start my scientific career<br />
<strong>and</strong> prepare my thesis <strong>in</strong> her group. I have been privileged to have her as a<br />
supervisor as she has provided all the guidance, support, trust <strong>and</strong> encouragement<br />
that the project required. Academy Pr<strong>of</strong>essor Emeritus Kari Kivirikko deserves<br />
my most s<strong>in</strong>cere thanks for his amaz<strong>in</strong>g attitude towards science <strong>and</strong> for his<br />
endless optimism, which truly gives <strong>in</strong>spiration <strong>and</strong> basis for the scientific career<br />
as well as life itself. I want to thank warmly Docent Raija So<strong>in</strong><strong>in</strong>en for her<br />
enthusiasm, trust, encouragement, <strong>and</strong> all the effort <strong>in</strong> the mouse projects.<br />
I am grateful to Pr<strong>of</strong>essor Emeritus Heather Yeowell <strong>and</strong> Pr<strong>of</strong>essor Helen<br />
Skaer for <strong>their</strong> careful revision <strong>of</strong> this thesis <strong>and</strong> valuable comments on it. M.Sc.<br />
S<strong>and</strong>ra Hänn<strong>in</strong>en is acknowledged for her expert revision <strong>of</strong> the language.<br />
We are not able to do science on our own. I want to express my respect <strong>and</strong><br />
thanks to other key persons, Pr<strong>of</strong>essor Emeritus Ilmo Hass<strong>in</strong>en, Pr<strong>of</strong>essor Ta<strong>in</strong>a<br />
Pihlajaniemi, Pr<strong>of</strong>essor Seppo Va<strong>in</strong>io, Pr<strong>of</strong>essor Leena Ala-Kokko, Docent M<strong>in</strong>na<br />
Männikkö, Docent Peppi Karpp<strong>in</strong>en, Dr. Aki Mann<strong>in</strong>en <strong>and</strong> Dr. Lauri Eklund, <strong>of</strong><br />
the Department <strong>of</strong> Medical Biochemistry <strong>and</strong> Molecular Biology for provid<strong>in</strong>g<br />
excellent research facilities <strong>and</strong> atmosphere to make science.<br />
I want to thank all my co-authors for the pleasant collaboration <strong>and</strong> <strong>their</strong><br />
essential contribution to this work. I want to express my deepest thanks to Dr.<br />
Kati Takaluoma for giv<strong>in</strong>g me h<strong>and</strong>s-on <strong>in</strong>troduction to the world <strong>of</strong> collagens<br />
<strong>and</strong> science itself. I also want to thank her for supervis<strong>in</strong>g my M.Sc. project <strong>and</strong><br />
for enjoyable moments we have had together. M.D. Juha Lantto <strong>and</strong> M.D. Hanna-<br />
Leena Talsta are thanked for <strong>their</strong> contribution to these projects <strong>and</strong> nice moments<br />
<strong>in</strong> the <strong>of</strong>fice. I wish to express my gratitude to Pr<strong>of</strong>essor Ruud Bank for his help<br />
with cross-l<strong>in</strong>ks <strong>and</strong> Docent Raija Sormunen <strong>and</strong> Dr. Ilkka Mi<strong>in</strong>ala<strong>in</strong>en for <strong>their</strong><br />
expertise <strong>and</strong> help with the electron microscope.<br />
This work would have not been possible without the outst<strong>and</strong><strong>in</strong>g technical<br />
assistance <strong>of</strong> Liisa Äijälä dur<strong>in</strong>g all these years start<strong>in</strong>g from lab basics to a<br />
novice. I also want to thank the whole staff at the department for creat<strong>in</strong>g the<br />
great atmosphere to work <strong>and</strong> for <strong>their</strong> given assistance along the way.<br />
Particularly Auli K<strong>in</strong>nunen, Pertti Vuokila, Risto Helm<strong>in</strong>en <strong>and</strong> Seppo<br />
7
Lähdesmäki deserve my warmest acknowledgements for <strong>their</strong> pr<strong>of</strong>essional <strong>and</strong><br />
k<strong>in</strong>d help. I am also really grateful for the people <strong>in</strong> the Biocenter Oulu<br />
Transgenic core facility <strong>and</strong> the Laboratory Animal Center for the excellent<br />
collaboration <strong>and</strong> good care <strong>of</strong> our mice.<br />
All former <strong>and</strong> present members <strong>of</strong> our research group are heartily thanked<br />
for be<strong>in</strong>g such wonderful colleagues. It has been a privilege to work with you <strong>and</strong><br />
share all those memorable sunny <strong>and</strong> cloudy days dur<strong>in</strong>g these years. You are the<br />
ones who have made our group so <strong>in</strong>spir<strong>in</strong>g, supportive <strong>and</strong> warm. I couldn’t ask<br />
for more. I also wish to thank Johanna Korvala for shar<strong>in</strong>g the unforgettable time<br />
prepar<strong>in</strong>g the theses <strong>and</strong> organiz<strong>in</strong>g the dissertation party.<br />
All <strong>of</strong> this would not be possible without the s<strong>in</strong>cere <strong>and</strong> endless support <strong>of</strong><br />
family <strong>and</strong> friends. I s<strong>in</strong>cerely want you to know that you have always brought<br />
out the best <strong>in</strong> me. It’s not easy to f<strong>in</strong>d words to thank all <strong>of</strong> my friends who have<br />
been with me dur<strong>in</strong>g all the good days but also dur<strong>in</strong>g those days when I have<br />
needed you the most. Every day I wonder with smile on my face how lucky I am<br />
to have all <strong>of</strong> you <strong>in</strong> my life. And f<strong>in</strong>ally, I owe my deepest gratitude to my family,<br />
my parents <strong>and</strong> sisters <strong>and</strong> <strong>their</strong> families. Without your unconditional love,<br />
support <strong>and</strong> encouragement throughout all my life none <strong>of</strong> this would be possible.<br />
I love you.<br />
This work was supported by the Biocenter Oulu <strong>and</strong> the Emil Aaltonen<br />
Foundation.<br />
Oulu, April 2012 Marjo Hyry<br />
8
Abbreviations<br />
AGE advanced glycation end-product<br />
bp base pair<br />
BS Bruck syndrome<br />
C carboxy<br />
CD circular dichroism<br />
cDNA complementary DNA<br />
C-P4H collagen prolyl 4-hydroxylase<br />
CRTAP cartilage-associated prote<strong>in</strong><br />
Da Dalton(s)<br />
deH-HLNL dehydro-hydroxylys<strong>in</strong>onorleuc<strong>in</strong>e<br />
deH-LNL dehydro-lys<strong>in</strong>onorleuc<strong>in</strong>e<br />
E embryonic day<br />
ECM extracellular matrix<br />
EDS Ehlers-Danlos syndrome<br />
ER endoplasmic reticulum<br />
FACIT fibril-associated collagens with <strong>in</strong>terrupted triple helices<br />
FKBP10 FK506 b<strong>in</strong>d<strong>in</strong>g prote<strong>in</strong> 10<br />
GGT hydroxylysyl glucosyltransferase<br />
GT hydroxylysyl galactosyltransferase<br />
HIF hypoxia-<strong>in</strong>ducible factor<br />
HLKNL hydroxylys<strong>in</strong>o-5-ketonorleuc<strong>in</strong>e<br />
HP hydroxylysyl pyrid<strong>in</strong>ol<strong>in</strong>e<br />
HPLC reverse-phase high-pressure liquid chromatography<br />
Hyl hydroxylys<strong>in</strong>e<br />
JMJD6 Jumonji doma<strong>in</strong>-6 prote<strong>in</strong><br />
Km<br />
Michaelis-Menten constant<br />
LH lysyl hydroxylase<br />
LKNL lys<strong>in</strong>o-5-ketonorleuc<strong>in</strong>e<br />
LO lysyl oxidase<br />
LP lysyl pyrid<strong>in</strong>ol<strong>in</strong>e<br />
MMP matrix metalloprote<strong>in</strong>ase<br />
mRNA messenger RNA<br />
N am<strong>in</strong>o<br />
NC noncollagenous<br />
OI osteogenesis imperfecta<br />
9
P3H prolyl 3-hydroxylase<br />
PCR polymerase cha<strong>in</strong> reaction<br />
PDI prote<strong>in</strong> disulfide isomerase<br />
PLOD procollagen-lys<strong>in</strong>e 2-oxoglutarate 5-dioxygenase; human gene<br />
Plod procollagen-lys<strong>in</strong>e 2-oxoglutarate 5-dioxygenase; mouse gene<br />
Q-PCR real-time quantitative PCR<br />
RM Reichert’s membrane<br />
TEM transmission electron microscopy<br />
TGF-β transform<strong>in</strong>g growth factor-β<br />
TIA T-cell-restricted <strong>in</strong>tracellular antigen<br />
X, <strong>in</strong> -Gly-X-Y- any am<strong>in</strong>o acid<br />
Y, <strong>in</strong> -Gly-X-Y- any am<strong>in</strong>o acid<br />
10
List <strong>of</strong> orig<strong>in</strong>al articles<br />
This thesis is based on the follow<strong>in</strong>g articles, which are referred to <strong>in</strong> the text by<br />
<strong>their</strong> Roman numerals:<br />
I Takaluoma K, Hyry M, Lantto J, Sormunen R, Bank RA, Kivirikko KI, Myllyharju J<br />
& So<strong>in</strong><strong>in</strong>en R (2007) Tissue-specific changes <strong>in</strong> the hydroxylys<strong>in</strong>e content <strong>and</strong> crossl<strong>in</strong>ks<br />
<strong>of</strong> collagens <strong>and</strong> alterations <strong>in</strong> fibril morphology <strong>in</strong> lysyl hydroxylase 1 knockout<br />
mice. J Biol Chem 282(9): 6588–6596.<br />
II Hyry M, Lantto J & Myllyharju J (2009) Missense mutations that cause Bruck<br />
syndrome affect enzymatic activity, fold<strong>in</strong>g, <strong>and</strong> oligomerization <strong>of</strong> lysyl hydroxylase<br />
2. J Biol Chem 284(45): 30917–30924.<br />
III Hyry M, So<strong>in</strong><strong>in</strong>en R, Bank RA, Sormunen R, Mi<strong>in</strong>ala<strong>in</strong>en I, Talsta HL, Lantto J &<br />
Myllyharju J (2012) Lack <strong>of</strong> lysyl hydroxylase 2 <strong>in</strong> mouse leads to early lethality<br />
dur<strong>in</strong>g embryonic development. Manuscript.<br />
11
Contents<br />
Abstract<br />
Tiivistelmä<br />
Acknowledgements 7<br />
Acknowledgements 7<br />
Abbreviations 9<br />
List <strong>of</strong> orig<strong>in</strong>al articles 11<br />
Contents 13<br />
1 Introduction 17<br />
2 Review <strong>of</strong> the literature 19<br />
2.1 Extracellular matrix ................................................................................. 19<br />
2.1.1 The molecular multiplicity <strong>of</strong> the extracellular matrix <strong>and</strong><br />
its connection to disorders ............................................................ 19<br />
2.2 Collagens ................................................................................................. 20<br />
2.2.1 The collagen family ...................................................................... 21<br />
2.2.2 Collagen biosynthesis ................................................................... 24<br />
2.3 Collagen hydroxylys<strong>in</strong>es <strong>and</strong> <strong>their</strong> function ............................................ 28<br />
2.3.1 Glycosylation <strong>of</strong> hydroxylys<strong>in</strong>es .................................................. 29<br />
2.3.2 Collagen cross-l<strong>in</strong>ks ..................................................................... 31<br />
2.4 <strong>Lysyl</strong> <strong>hydroxylases</strong> .................................................................................. 36<br />
2.4.1 <strong>Lysyl</strong> hydroxylase isoenzymes ..................................................... 37<br />
2.4.2 Molecular <strong>and</strong> catalytic properties ............................................... 42<br />
2.5 Human diseases related to lysyl <strong>hydroxylases</strong> <strong>and</strong> animal models ......... 46<br />
2.5.1 Ehlers-Danlos type VI .................................................................. 46<br />
2.5.2 Bruck syndrome 1 <strong>and</strong> 2 ............................................................... 49<br />
2.5.3 Fibrotic diseases <strong>and</strong> cancer ......................................................... 53<br />
2.5.4 A disease caused by mutations <strong>in</strong> LH3 <strong>and</strong> LH3 null mice .......... 55<br />
2.5.5 Inactivation <strong>of</strong> LH <strong>in</strong> other animal models ................................... 57<br />
3 Outl<strong>in</strong>es <strong>of</strong> the present study 59<br />
4 Materials <strong>and</strong> methods 61<br />
5 Results 63<br />
5.1 LH1 knock-out mice have <strong>in</strong>creased risk <strong>of</strong> lethal arterial<br />
ruptures (I) .............................................................................................. 63<br />
5.1.1 Generation <strong>of</strong> LH1 null mice <strong>and</strong> LH1 expression <strong>in</strong> adult<br />
mouse tissues ................................................................................ 63<br />
13
5.1.2 LH1 null mice have no kyphoscoliosis but are hypotonic<br />
<strong>and</strong> have <strong>in</strong>creased mortality due to arterial ruptures ................... 64<br />
5.1.3 Abnormal collagen fibrils <strong>in</strong> LH1 null mice ................................. 64<br />
5.1.4 Decreased LH activity <strong>in</strong> LH1 null mice affects the<br />
hydroxylys<strong>in</strong>e content <strong>in</strong> various tissues differently .................... 65<br />
5.1.5 Abnormal hydroxylation <strong>of</strong> triple helical lys<strong>in</strong>e residues<br />
affects the cross-l<strong>in</strong>k<strong>in</strong>g pattern <strong>in</strong> the LH1 null mouse<br />
tissues ........................................................................................... 66<br />
5.2 Analysis <strong>of</strong> the molecular level consequences <strong>of</strong> Bruck<br />
syndrome mutations by the use <strong>of</strong> recomb<strong>in</strong>ant LH2 prote<strong>in</strong>s (II) .......... 67<br />
5.2.1 Expression <strong>of</strong> wild-type <strong>and</strong> recomb<strong>in</strong>ant LH2(long)<br />
polypeptides with Bruck syndrome mutations <strong>in</strong> <strong>in</strong>sect<br />
cells <strong>and</strong> <strong>their</strong> purification ............................................................ 67<br />
5.2.2 The Bruck syndrome mutant R594H <strong>and</strong> G597V<br />
LH2(long) polypeptides have problems <strong>in</strong> dimerization<br />
<strong>and</strong> are prone to degradation ........................................................ 68<br />
5.2.3 Fold<strong>in</strong>g <strong>of</strong> the R594H <strong>and</strong> G597V mutant LH2(long)<br />
polypeptides is impaired <strong>and</strong> <strong>their</strong> proteolytic sensitivity is<br />
<strong>in</strong>creased ....................................................................................... 69<br />
5.2.4 All Bruck syndrome mutations affect the activity <strong>of</strong> LH2 ........... 70<br />
5.3 <strong>Lysyl</strong> hydroxylase 2 is essential for the embryonic development<br />
<strong>of</strong> mouse (III) .......................................................................................... 71<br />
5.3.1 LH2 null mice die after embryonic day 10.5 ................................ 71<br />
5.3.2 LH2 is expressed ubiquitously <strong>in</strong> the E10.5 embryos <strong>and</strong><br />
<strong>in</strong> extraembryonic tissues <strong>and</strong> its lack seems to lead to<br />
certa<strong>in</strong> changes <strong>in</strong> collagen expression ......................................... 72<br />
5.3.3 Expression <strong>of</strong> the Plod1 <strong>and</strong> Plod3 genes is not<br />
upregulated <strong>in</strong> the Plod2 gt/gt embryos ........................................... 73<br />
5.3.4 LH2 is widely expressed <strong>in</strong> adult Plod2 wt/gt tissues ...................... 74<br />
5.3.5 Changes <strong>in</strong> lysyl hydroxylation <strong>and</strong> cross-l<strong>in</strong>k<strong>in</strong>g <strong>of</strong><br />
collagen <strong>in</strong> adult Plod2 wt/gt mice ................................................... 74<br />
6 Discussion 77<br />
6.1 LH1 null mice provide a tool to study tissue-specific<br />
consequences <strong>of</strong> the lack <strong>of</strong> LH1 activity ............................................... 77<br />
6.2 Mutations caus<strong>in</strong>g Bruck syndrome lead to decreased LH2<br />
activity, expla<strong>in</strong><strong>in</strong>g the symptoms <strong>of</strong> diagnosed Bruck syndrome<br />
type 2 patients ......................................................................................... 81<br />
14
6.3 Inactivation <strong>of</strong> mouse Plod2 leads to lethality dur<strong>in</strong>g early<br />
embryonic development .......................................................................... 83<br />
7 Conclusions <strong>and</strong> future prospects 89<br />
References 91<br />
Orig<strong>in</strong>al articles 111<br />
15
1 Introduction<br />
Collagens are the most abundant prote<strong>in</strong>s <strong>in</strong> the human body <strong>and</strong> comprise the<br />
major structural backbone <strong>in</strong> the extracellular matrix (ECM) <strong>of</strong> tissues. Several<br />
collagen-specific enzymes are required <strong>in</strong> collagen biosynthesis, <strong>in</strong>clud<strong>in</strong>g lysyl<br />
<strong>hydroxylases</strong> (LHs). Hence, it is natural that changes <strong>in</strong> collagens or enzymes<br />
<strong>in</strong>volved <strong>in</strong> collagen synthesis can lead to severe consequences. Collagen-related<br />
changes <strong>in</strong> the ECM cause many common conditions such as fibrosis, but they<br />
also cause rare disorders because <strong>of</strong> <strong>in</strong>herited mutations <strong>in</strong> ECM genes, such as<br />
osteogenesis imperfecta (OI) or the rarer Bruck syndrome (BS).<br />
Humans <strong>and</strong> mice have three different LH is<strong>of</strong>orms i.e. LH1, LH2 <strong>and</strong> LH3.<br />
Mutations <strong>in</strong> all three LH is<strong>of</strong>orms are known to cause severe connective tissue<br />
disorders <strong>in</strong> humans demonstrat<strong>in</strong>g <strong>their</strong> importance. LH1 mutations have long<br />
been known to cause the Ehlers-Danlos syndrome VIA (EDS VIA), which is<br />
characterized e.g. by progressive kyphoscoliosis <strong>and</strong> hypermobile jo<strong>in</strong>ts. The<br />
discovery <strong>of</strong> LH2 was soon accompanied by identification <strong>of</strong> mutations <strong>in</strong> LH2<br />
that cause the BS type 2. BS2 patients suffer from symptoms like fragile bones<br />
<strong>and</strong> congenital jo<strong>in</strong>t contractures. Recently, LH3 mutations have also been<br />
connected to two human disorders, <strong>and</strong> some years earlier a mouse LH3 knockout<br />
was observed to be embryonically lethal, suggest<strong>in</strong>g the importance <strong>of</strong> LHs<br />
also <strong>in</strong> mice.<br />
Hydroxylys<strong>in</strong>e residues formed by LHs function <strong>in</strong> collagens as attachment<br />
sites for carbohydrate residues <strong>and</strong> certa<strong>in</strong> residues participate <strong>in</strong> collagen crossl<strong>in</strong>ks.<br />
Moreover, the hydroxylation state <strong>of</strong> the lys<strong>in</strong>es at the cross-l<strong>in</strong>k sites<br />
determ<strong>in</strong>es the cross-l<strong>in</strong>k<strong>in</strong>g pattern <strong>and</strong> thus partly the mechanical properties <strong>of</strong><br />
tissues. Although <strong>in</strong> vitro studies have revealed much <strong>in</strong>formation about LHs, the<br />
knowledge <strong>of</strong> <strong>in</strong> <strong>vivo</strong> functions <strong>of</strong> LH1 <strong>and</strong> LH2 has been restricted mostly to<br />
analyses <strong>of</strong> human patients. Our aim <strong>in</strong> this work was to generate LH1 <strong>and</strong> LH2<br />
null mouse l<strong>in</strong>es to obta<strong>in</strong> more detailed <strong>in</strong>formation <strong>of</strong> the <strong>in</strong> <strong>vivo</strong> functions <strong>of</strong><br />
these enzymes. The mouse l<strong>in</strong>es were also expected to provide more <strong>in</strong>formation<br />
about the EDS VIA <strong>and</strong> BS2 syndroms at the molecular level <strong>and</strong> about the<br />
<strong>in</strong>tricate cross-l<strong>in</strong>k formation. In addition, we studied the effects <strong>of</strong> the known<br />
BS2 mutations on recomb<strong>in</strong>ant human LH2 polypeptides to underst<strong>and</strong> the<br />
molecular pathology <strong>of</strong> the syndrome. Our results show that the LH1 null mouse<br />
l<strong>in</strong>e is valuable <strong>in</strong> the analysis <strong>of</strong> the <strong>in</strong> <strong>vivo</strong> function <strong>of</strong> LH1, as well as allow<strong>in</strong>g<br />
better underst<strong>and</strong><strong>in</strong>g <strong>of</strong> the pathology <strong>of</strong> EDS VIA. Surpris<strong>in</strong>gly, our work<br />
demonstrates that LH2 is essential for the embryonic development <strong>of</strong> the mouse.<br />
17
Our data also show that the mutations found <strong>in</strong> human BS2 patients decrease the<br />
enzyme activity <strong>in</strong> vitro by two different mechanisms, but do not lead to complete<br />
LH2 <strong>in</strong>activation, which is likely to expla<strong>in</strong> the survival <strong>of</strong> the BS2 patients. Both<br />
mouse l<strong>in</strong>es also provide novel <strong>in</strong>formation on the roles <strong>of</strong> LH1 <strong>and</strong> LH2 <strong>in</strong><br />
determ<strong>in</strong><strong>in</strong>g the collagen cross-l<strong>in</strong>k<strong>in</strong>g pattern <strong>in</strong> tissues.<br />
18
2 Review <strong>of</strong> the literature<br />
2.1 Extracellular matrix<br />
The extracellular matrix (ECM) is a complex environment that surrounds cells<br />
<strong>and</strong> enables the multicellularity <strong>of</strong> organisms. ECM consists <strong>of</strong> molecules that are<br />
secreted by the adjacent cells <strong>and</strong> thus the environment is tailored for the<br />
requirements <strong>of</strong> the cells <strong>and</strong> the tissue. ECM functions as a scaffold provid<strong>in</strong>g<br />
strength <strong>and</strong> support aga<strong>in</strong>st physical stress, but it also has an essential role <strong>in</strong> cell<br />
behavior. It <strong>in</strong>teracts with signal<strong>in</strong>g molecules <strong>and</strong> growth factors, which regulate<br />
cellular events such as determ<strong>in</strong>ation, differentiation, proliferation, migration <strong>and</strong><br />
survival. The ECM is highly dynamic <strong>and</strong> it undergoes constant remodel<strong>in</strong>g<br />
dur<strong>in</strong>g life <strong>and</strong> particularly <strong>in</strong> many disease states. Failures <strong>in</strong> the renewal process<br />
underlie many pathological states like fibrosis, atherosclerosis <strong>and</strong> tumorigenesis.<br />
(For reviews, see Kielty & Grant 2002, Daley et al. 2008, Butcher et al. 2009,<br />
Tsang et al. 2010).<br />
2.1.1 The molecular multiplicity <strong>of</strong> the extracellular matrix <strong>and</strong> its<br />
connection to disorders<br />
The ma<strong>in</strong> components <strong>of</strong> the ECM are collagens <strong>and</strong> noncollagenous prote<strong>in</strong>s<br />
such as proteoglycans, glycoprote<strong>in</strong>s <strong>and</strong> elast<strong>in</strong>s, but the different mechanical<br />
requirements <strong>of</strong> tissues determ<strong>in</strong>e the ratios <strong>of</strong> these prote<strong>in</strong>s. Thus, the molecular<br />
diversity <strong>in</strong> the ECM is overwhelm<strong>in</strong>g <strong>and</strong> these structurally very different<br />
molecules are themselves responsible for the dynamic nature <strong>of</strong> the ECM. Many<br />
ECM prote<strong>in</strong>s are large <strong>and</strong> complex consist<strong>in</strong>g <strong>of</strong> several <strong>in</strong>dividually folded<br />
doma<strong>in</strong>s that are highly conserved between species. The synthesis, organization<br />
<strong>and</strong> <strong>in</strong>terplay <strong>of</strong> the ECM components are strict. The synthesis is controlled by<br />
growth factors such as transform<strong>in</strong>g growth factor-β (TGF-β) <strong>and</strong> the degradation<br />
by proteases like cyste<strong>in</strong>e <strong>and</strong> ser<strong>in</strong>e proteases <strong>and</strong> matrix metalloprote<strong>in</strong>ases<br />
(MMPs). It is clear that the activity <strong>of</strong> these molecules must be under precise<br />
control <strong>and</strong> misregulation can easily lead to severe consequences. (For reviews,<br />
see Daley et al. 2008, Badylak et al. 2009, Butcher et al. 2009, Hynes 2009,<br />
Järvelä<strong>in</strong>en et al. 2009, Tsang et al. 2010). Recently it was suggested that <strong>in</strong><br />
addition to signal<strong>in</strong>g molecules <strong>and</strong> growth factors, basic features such as the<br />
19
stiffness <strong>of</strong> the ECM also affect differentiation <strong>and</strong> cell behavior (Engler et al.<br />
2006, Butcher et al. 2009).<br />
One m<strong>in</strong>imal <strong>in</strong>herited or acquired change <strong>in</strong> one ECM component can lead<br />
to physicochemical <strong>and</strong> thus to mechanical <strong>and</strong> behavioral changes <strong>in</strong> the tissue or<br />
develop<strong>in</strong>g <strong>in</strong>dividual followed by a possible disorder or even lethality<br />
(Järvelä<strong>in</strong>en et al. 2009, Levental et al. 2009). The rare Ehlers-Danlos syndrome<br />
(EDS) type VIA is a severe connective tissue disorder <strong>and</strong> an example <strong>of</strong> an<br />
<strong>in</strong>herited monogenic disorder. It is caused by mutations <strong>in</strong> one <strong>of</strong> the collagenmodify<strong>in</strong>g<br />
enzymes, lysyl hydroxylase 1 (LH1, see 2.2.2 <strong>and</strong> 2.4). It is<br />
characterized by progressive kyphoscoliosis, hypermobile jo<strong>in</strong>ts <strong>and</strong> muscle<br />
hypotonia, <strong>and</strong> patients have an <strong>in</strong>creased risk <strong>of</strong> arterial ruptures. (For reviews,<br />
see Yeowell & Walker 2000, Ste<strong>in</strong>mann et al. 2002). ECM changes are dist<strong>in</strong>ct<br />
also <strong>in</strong> common diseases that can be acquired or caused by a comb<strong>in</strong>ation <strong>of</strong><br />
acquired <strong>and</strong> genetic factors, not forgett<strong>in</strong>g the <strong>in</strong>teractions between the ECM <strong>and</strong><br />
growth factors <strong>and</strong> other signal<strong>in</strong>g molecules. In coronary heart disease, for<br />
<strong>in</strong>stance, the ECM components participate <strong>in</strong> the formation <strong>of</strong> the atherosclerotic<br />
plaque. Changes <strong>in</strong> the ECM composition have been connected also to<br />
carc<strong>in</strong>ogenesis <strong>and</strong> metastasis <strong>of</strong> the tumor cells. On the other h<strong>and</strong>, probably all<br />
diseases affect the quality <strong>and</strong>/or quantity <strong>of</strong> the ECM. Thus, exact knowledge<br />
about <strong>in</strong>dividual ECM molecules, <strong>their</strong> function <strong>and</strong> <strong>in</strong>terplay with growth factors<br />
<strong>and</strong> signal<strong>in</strong>g molecules is beneficial not only for the drug development <strong>and</strong><br />
treatment <strong>of</strong> ECM disorders, but it can be utilized also <strong>in</strong> the treatment <strong>of</strong> a wide<br />
variety <strong>of</strong> other diseases. A deficiency <strong>in</strong> structural ECM prote<strong>in</strong>s is challeng<strong>in</strong>g<br />
to treat, but the regulators provide a potential tool when <strong>their</strong> function is known.<br />
(Hynes 2009, Järvelä<strong>in</strong>en et al. 2009, Levental et al. 2009, Van Agtmael &<br />
Bruckner-Tuderman 2010).<br />
2.2 Collagens<br />
Collagens are the major <strong>in</strong>soluble fibrous prote<strong>in</strong>s <strong>in</strong> the ECM <strong>and</strong> the most<br />
abundant prote<strong>in</strong>s <strong>in</strong> the human body. Thus, collagens have a remarkable role <strong>in</strong><br />
the ma<strong>in</strong>tenance <strong>of</strong> various structures as well as <strong>in</strong> other essential ECM activities.<br />
Collagens are <strong>in</strong>volved <strong>in</strong> many disorders with numerous mutations <strong>in</strong> collagen<br />
genes beh<strong>in</strong>d them. Collagen biosynthesis requires specific enzymes <strong>and</strong><br />
mutations <strong>in</strong> these enzymes <strong>in</strong>crease the number <strong>of</strong> collagen-related disorders.<br />
(For reviews, see Myllyharju & Kivirikko 2001, Myllyharju & Kivirikko 2004,<br />
20
Gelse et al. 2003, Van Agtmael & Bruckner-Tuderman 2010, Mar<strong>in</strong>i et al. 2010,<br />
Park<strong>in</strong> et al. 2011).<br />
2.2.1 The collagen family<br />
Humans have at least 28 different collagen types, which consist <strong>of</strong> over 40<br />
different α cha<strong>in</strong>s (Myllyharju & Kivirikko 2004, Gordon & Hahn 2010). The<br />
heterogeneity <strong>of</strong> the collagen family members is also <strong>in</strong>creased by alternative<br />
splic<strong>in</strong>g <strong>of</strong> the primary transcripts cod<strong>in</strong>g for collagen polypeptides (Myllyharju<br />
& Kivirikko 2001, McAl<strong>in</strong>den et al. 2005, Sepp<strong>in</strong>en & Pihlajaniemi 2011). All<br />
collagen molecules consist <strong>of</strong> three identical or different polypeptides called α<br />
cha<strong>in</strong>s, which are wrapped around each other to form a triple helix. The triple<br />
helical structure forms as a result <strong>of</strong> the repeat<strong>in</strong>g Gly-X-Y triplet, which<br />
represents a collagenous doma<strong>in</strong>. As the smallest am<strong>in</strong>o acid, the glyc<strong>in</strong>e residue<br />
is positioned <strong>in</strong> the central axis <strong>of</strong> the triple helix <strong>and</strong> it is the only am<strong>in</strong>o acid that<br />
fits <strong>in</strong>to that restricted space. In fact, the most harmful mutations <strong>in</strong> the collagen<br />
genes are <strong>of</strong>ten those that replace the glyc<strong>in</strong>e residue <strong>in</strong> the triple helical doma<strong>in</strong>.<br />
In the collagenous (Gly-X-Y)n sequence, the X <strong>and</strong> Y can be any am<strong>in</strong>o acid<br />
except glyc<strong>in</strong>e, but <strong>of</strong>ten prol<strong>in</strong>e is <strong>in</strong> the X position <strong>and</strong> 4-hydroxyprol<strong>in</strong>e <strong>in</strong> the<br />
Y position. (Brodsky & Persikov 2005, Shoulders & Ra<strong>in</strong>es 2009, Gordon &<br />
Hahn 2010).<br />
All collagen molecules have at least one collagenous doma<strong>in</strong> <strong>in</strong> <strong>their</strong> structure,<br />
but the amount <strong>and</strong> lengths <strong>of</strong> these parts vary between collagen types. Over 90%<br />
<strong>of</strong> type I collagen for <strong>in</strong>stance is collagenous, while collagen XII conta<strong>in</strong>s only<br />
10% <strong>of</strong> triple helix. Besides collagenous doma<strong>in</strong>s, collagens conta<strong>in</strong> also<br />
noncollagenous (NC) parts which <strong>of</strong>ten have important functions. These doma<strong>in</strong>s<br />
may be <strong>in</strong>volved <strong>in</strong> the b<strong>in</strong>d<strong>in</strong>g <strong>of</strong> molecules or <strong>in</strong> the assembly <strong>of</strong> supramolecular<br />
structures, or they may have <strong>their</strong> own functions when released from the collagen,<br />
an example be<strong>in</strong>g endostat<strong>in</strong>. Endostat<strong>in</strong> is cleaved from the C-term<strong>in</strong>al NC1<br />
doma<strong>in</strong> <strong>of</strong> type XVIII collagen <strong>and</strong> is able to <strong>in</strong>hibit angiogenesis <strong>and</strong> tumor<br />
growth. Similar doma<strong>in</strong>s called rest<strong>in</strong> <strong>and</strong> arresten are also present <strong>in</strong> collagens<br />
XV <strong>and</strong> IV, respectively. (Myllyharju & Kivirikko 2001, Gelse et al. 2003,<br />
Ricard-Blum & Ruggiero 2005, Ricard-Blum 2011). Collagenous doma<strong>in</strong>s are not<br />
unique to collagens. There are several prote<strong>in</strong>s such as adiponect<strong>in</strong>, collect<strong>in</strong>,<br />
ficol<strong>in</strong> <strong>and</strong> macrophage receptor that conta<strong>in</strong> a collagen-like doma<strong>in</strong> <strong>in</strong> <strong>their</strong><br />
structure, but they are not called collagens. (Myllyharju & Kivirikko 2004,<br />
Ricard-Blum 2011).<br />
21
Collagens can be divided <strong>in</strong>to fibrillar <strong>and</strong> non-fibrillar collagens <strong>and</strong> further<br />
<strong>in</strong>to several subfamilies (see Fig. 1). The collagen superfamily consists <strong>of</strong> the<br />
fibril-form<strong>in</strong>g collagens such as type I, II, III <strong>and</strong> V collagens, fibril-associated<br />
collagens with <strong>in</strong>terrupted triple helices (FACITs) such as type IX <strong>and</strong> XII<br />
collagens, collagens that form hexagonal networks (type VIII <strong>and</strong> X collagens),<br />
network-form<strong>in</strong>g collagens (type IV collagens), collagens that form beaded<br />
filaments (type VI collagen), collagens that form anchor<strong>in</strong>g fibrils (type VII<br />
collagen), endostat<strong>in</strong>-produc<strong>in</strong>g collagens (type XV <strong>and</strong> XVIII collagens) <strong>and</strong><br />
transmembrane collagens such as collagens XIII <strong>and</strong> XXV. (For reviews, see<br />
Myllyharju & Kivirikko 2004, Kadler et al. 2007, Gordon & Hahn 2010). Type<br />
XXVIII collagen is the most recently identified collagen <strong>in</strong> the superfamily. It<br />
cannot be placed clearly <strong>in</strong>to any collagen subfamily mentioned above as it has<br />
properties typical <strong>of</strong> more than one <strong>of</strong> those groups. (Veit et al. 2006). Type XXIX<br />
collagen, reported <strong>in</strong> 2007, is likely a member <strong>of</strong> the collagen VI family rather<br />
than a new collagen type (Söderhäll et al. 2007, Fitzgerald et al. 2008).<br />
22
a. Fibril-form<strong>in</strong>g collagens, types I, II, III, V, XI, XXIV <strong>and</strong> XXVII<br />
Genes: COL1A1, COL1A2; COL2A1; COL3A1; COL5A1, COL5A2, COL5A3,<br />
COL5A4; COL11A1, COL11A2; COL24A1; COL27A1<br />
Triple helical region<br />
N pro-<br />
C propeptidepeptide<br />
300 nm<br />
100 nm<br />
c. Collagens form<strong>in</strong>g hexagonal networks, types VIII <strong>and</strong> X<br />
Genes: COL8A1, COL8A2; COL10A1<br />
VIII<br />
100 nm<br />
Fig. 1. Members <strong>of</strong> the collagen superfamily <strong>and</strong> <strong>their</strong> known supramolecular<br />
assemblies. Repr<strong>in</strong>ted from Myllyharju & Kivirikko 2004 with permission <strong>of</strong> the<br />
publisher.<br />
X<br />
Fibrillar collagens<br />
100 nm<br />
d. The family <strong>of</strong> type IV collagens<br />
e. Type VI collagen form<strong>in</strong>g beaded filaments<br />
Genes: COL4A1, COL4A2, COL4A3, COL4A4, COL4A5, COL4A6 Genes: COL6A1, COL6A2, COL6A3<br />
7S<br />
Dimer<br />
VI<br />
Dimer Tetramer Beaded filament<br />
100 nm<br />
100 nm<br />
Dimer<br />
200 nm<br />
Anchor<strong>in</strong>g fibril<br />
Basement<br />
membrane<br />
7S<br />
Tetramer<br />
200 nm<br />
Anchor<strong>in</strong>g<br />
plaque<br />
b. FACIT <strong>and</strong> related collagens, types IX, XII,<br />
XIV, XVI, XIX, XX, XXI, XXII <strong>and</strong> XXVI<br />
Genes: COL9A1, COL9A2, COL9A3; COL12A1; COL14A1;<br />
COL16A1; COL19A1; COL20A1; COL21A1; COL22A1;<br />
COL26A1<br />
IX<br />
Type II fibril<br />
GAG<br />
GAG<br />
XII <strong>and</strong> XIV<br />
100 nm 100 nm<br />
ficol<strong>in</strong> ectodysplas<strong>in</strong><br />
Fibril-form<strong>in</strong>g (fibrillar) collagens form the most abundant subfamily <strong>of</strong> collagens<br />
<strong>and</strong> this group <strong>in</strong>cludes seven different collagen types. Quantitatively the most<br />
common types are type I, II <strong>and</strong> III collagens, whereas types V, XI, XXIV <strong>and</strong><br />
XXVII are m<strong>in</strong>or forms. (Gelse et al. 2003, Myllyharju & Kivirikko 2004,<br />
Ricard-Blum & Ruggiero 2005). The fibril diameter <strong>of</strong> the collagen molecules<br />
100 nm<br />
f. Type VII collagen form<strong>in</strong>g anchor<strong>in</strong>g fibrils<br />
Gene: COL7A1<br />
Genes: COL13A1; COL17A1; COL23A1;<br />
VII<br />
COL25A1<br />
100 nm<br />
Type I fibril<br />
Genes: COL15A1; COL18A1<br />
GAG<br />
23
varies from 12 nm to over 500 nm <strong>and</strong> depends on the tissue <strong>and</strong> the stage <strong>of</strong><br />
development (Kadler et al. 2007). Tissues that must withst<strong>and</strong> high stress levels<br />
conta<strong>in</strong> a higher proportion <strong>of</strong> collagen fibers with a large diameter. Thus,<br />
diameter <strong>in</strong>creases at the expense <strong>of</strong> flexibility. (Bailey et al. 1998). Fibrillar<br />
collagens, exclud<strong>in</strong>g types XXIV <strong>and</strong> XXVII, share an about 1000 am<strong>in</strong>o acid<br />
residues long doma<strong>in</strong> which has a perfect collagenous (Gly-X-Y)n sequence<br />
without any <strong>in</strong>terruptions (Gordon & Hahn 2010). In addition, they all have N-<br />
<strong>and</strong> C-term<strong>in</strong>al propeptides, which are cleaved before fibril formation, <strong>and</strong> may<br />
have also some smaller NC doma<strong>in</strong>s. However, the cleavage <strong>of</strong> the N propeptide<br />
is not always essential. It can also be partially cleaved or entirely <strong>in</strong>tact such as <strong>in</strong><br />
the early embryonic form <strong>of</strong> type II collagen (IIA). The role <strong>of</strong> N propeptide <strong>in</strong><br />
the mature collagen is not known, but the globular N-term<strong>in</strong>al end <strong>of</strong> the N<br />
propeptides may have b<strong>in</strong>d<strong>in</strong>g functions. One common structure <strong>in</strong> the globular<br />
N-term<strong>in</strong>al end is a cyste<strong>in</strong>e-rich doma<strong>in</strong>, which is shown to be able to b<strong>in</strong>d<br />
growth factors such as certa<strong>in</strong> bone morphogenetic prote<strong>in</strong>s <strong>and</strong> TGF-β1.<br />
(Fleischmajer et al. 1990, Ng et al. 1993, Zhu et al. 1999, Larraín et al. 2000<br />
Ricard-Blum & Ruggiero 2005).<br />
Fibrillar collagens provide stability <strong>and</strong> <strong>in</strong>tegrity to the tissues <strong>and</strong> they are<br />
ubiquitously found <strong>in</strong> tissues such as bone, cartilage <strong>and</strong> sk<strong>in</strong>. Type I collagen is<br />
usually a heterotrimer consist<strong>in</strong>g <strong>of</strong> two α1(I) cha<strong>in</strong>s <strong>and</strong> one α2(I) cha<strong>in</strong>. It forms<br />
about 90% <strong>of</strong> the organic mass <strong>of</strong> bone <strong>and</strong> together with type V collagen it is<br />
responsible for the structural backbone <strong>of</strong> bone. Type I collagen is also the major<br />
collagen type <strong>in</strong> many other tissues such as tendon, sk<strong>in</strong> <strong>and</strong> ligaments. Although<br />
type I collagen is the most common collagen type <strong>in</strong> many tissues, the tissues can<br />
be mechanically very different. In sk<strong>in</strong>, for example, type I collagen forms<br />
compounds with type III collagen provid<strong>in</strong>g tensile strength for the tissue, <strong>and</strong> <strong>in</strong><br />
bone, load bear<strong>in</strong>g ability as well as tensional <strong>and</strong> torsional properties are<br />
accompanied by calcification. (Gelse et al. 2003). The stability <strong>of</strong> the collagenous<br />
structures is also affected by the type <strong>of</strong> cross-l<strong>in</strong>ks predom<strong>in</strong>at<strong>in</strong>g between<br />
collagen molecules (Knott & Bailey 1998, Bank et al. 1999).<br />
2.2.2 Collagen biosynthesis<br />
Collagen biosynthesis from messenger RNA (mRNA) to an active prote<strong>in</strong> is a<br />
complex process characterized by many co- <strong>and</strong> post-translational modifications<br />
that are unique only for collagens <strong>and</strong> prote<strong>in</strong>s conta<strong>in</strong><strong>in</strong>g collagen-like doma<strong>in</strong>s<br />
(see Fig. 2). The biosynthesis <strong>of</strong> fibril-form<strong>in</strong>g collagens is best known <strong>and</strong> thus<br />
24
used <strong>of</strong>ten as an example <strong>of</strong> collagen synthesis <strong>in</strong> general. Collagen biosynthesis<br />
can be divided to <strong>in</strong>tracellular <strong>and</strong> extracellular stages. Intracellular steps <strong>in</strong>clude<br />
the synthesis <strong>and</strong> modification <strong>of</strong> a procollagen molecule from a gene transcript,<br />
whereas extracellular steps convert procollagen molecules to collagen molecules<br />
<strong>and</strong> furthermore to more extensive f<strong>in</strong>al structures. (Kielty & Grant 2002, Gelse et<br />
al. 2003).<br />
Cytoplasm<br />
Nucleus<br />
Extracellular<br />
space<br />
Endoplasmic<br />
reticulum<br />
Fig. 2. Ma<strong>in</strong> steps <strong>in</strong> the biosynthesis <strong>of</strong> fibril-form<strong>in</strong>g collagens. Modified from<br />
Myllyharju & Kivirikko 2004.<br />
-OH<br />
-OH<br />
Intracellular co- <strong>and</strong> post-translational modifications<br />
-OH<br />
OH<br />
-OH<br />
-OH<br />
-O-Gal-Glc<br />
-<br />
OH OH OH<br />
OH OH<br />
-<br />
-<br />
-<br />
-<br />
-<br />
-<br />
O-Gal-Glc<br />
-OH<br />
-OH<br />
-O-Gal<br />
-OH<br />
S<br />
S<br />
S<br />
S<br />
Synthesis & co- <strong>and</strong><br />
post-translational modifications<br />
Collagen prolyl 4-hydroxylase<br />
-OH Prolyl 3-hydroxylase<br />
<strong>Lysyl</strong> hydroxylase<br />
-O-Gal-Glc<br />
-O-Gal<br />
Collagen glycosyltransferases<br />
Assembly <strong>of</strong> 3 procollagen cha<strong>in</strong>s &<br />
disulfide bond<strong>in</strong>g<br />
Prote<strong>in</strong> disulfide isomerase<br />
Assembly <strong>of</strong> triple helix<br />
Cleavage <strong>of</strong> propeptides<br />
N <strong>and</strong> C prote<strong>in</strong>ases<br />
Assembly <strong>in</strong>to collagen fibrils &<br />
cross-l<strong>in</strong>k formation<br />
<strong>Lysyl</strong> oxidase<br />
Collagen polypeptide cha<strong>in</strong>s are translated <strong>in</strong> the rough endoplasmic reticulum<br />
(ER) as pre-pro-α-cha<strong>in</strong>s. The grow<strong>in</strong>g polypeptide cha<strong>in</strong> is directed <strong>in</strong>to the<br />
lumen <strong>of</strong> the ER accord<strong>in</strong>g to an N-term<strong>in</strong>al signal sequence. The signal sequence<br />
is removed after translocation <strong>in</strong>to the lumen by a signal peptidase, <strong>and</strong> the<br />
produced pro-α-cha<strong>in</strong> is immediately available for further co- <strong>and</strong> posttranslational<br />
modifications. These modifications are performed with<strong>in</strong> the<br />
cisternae <strong>of</strong> ER <strong>and</strong> they consist ma<strong>in</strong>ly <strong>of</strong> hydroxylations <strong>and</strong> glycosylations.<br />
Collagen prolyl 4-<strong>hydroxylases</strong> (C-P4Hs), prolyl 3-<strong>hydroxylases</strong> (P3Hs) <strong>and</strong> lysyl<br />
25
<strong>hydroxylases</strong> (LHs) are members <strong>of</strong> the 2-oxoglutarate dioxygenase family, <strong>and</strong><br />
they catalyze hydroxylations <strong>of</strong> certa<strong>in</strong> prol<strong>in</strong>e <strong>and</strong> lys<strong>in</strong>e residues <strong>in</strong> procollagen<br />
cha<strong>in</strong>s to 4-hydroxyprol<strong>in</strong>e, 3-hydroxyprol<strong>in</strong>e <strong>and</strong> hydroxylys<strong>in</strong>e, respectively<br />
(see Fig. 2). Glycosylations <strong>of</strong> some <strong>of</strong> the hydroxylysyl residues to<br />
galactosylhydroxylys<strong>in</strong>e <strong>and</strong> glucosylgalactosyl-hydroxylys<strong>in</strong>e are catalyzed by<br />
collagen galactosyltransferase <strong>and</strong> glugosyltransferase, respectively. (Kielty &<br />
Grant 2002, Canty & Kadler 2005, Myllyharju 2005).<br />
4-Hydroxyprol<strong>in</strong>e residues are essential for the thermal stability <strong>of</strong> the<br />
collagen triple helix. Collagen α-cha<strong>in</strong>s that are not properly hydroxylated by C-<br />
P4H do not hold together <strong>in</strong> a triple helix conformation <strong>in</strong> physiological<br />
conditions. (Kivirikko & Pihlajaniemi 1998, Myllyharju & Kivirikko 2001). The<br />
role <strong>of</strong> 3-hydroxyprol<strong>in</strong>e was unknown for a long time until a decreased prolyl 3hydroxylation<br />
<strong>in</strong> CRTAP (cartilage-associated prote<strong>in</strong>) deficiency was shown to<br />
be connected to osteogenesis imperfecta (OI) (Morello et al. 2006, Mar<strong>in</strong>i et al.<br />
2010). 3-Hydroxyprol<strong>in</strong>e residues are now suggested to have a role <strong>in</strong> the<br />
assembly <strong>of</strong> collagen fibrils (see below, Weis et al. 2010). Hydroxylys<strong>in</strong>e residues<br />
are needed for the formation <strong>of</strong> covalent collagen cross-l<strong>in</strong>ks <strong>and</strong> they serve as<br />
sites for glycosylation (Kielty & Grant 2002). In addition to the collagenous<br />
doma<strong>in</strong>, hydroxylys<strong>in</strong>e residues are found <strong>in</strong> the short noncollagenous telopeptide<br />
regions at the ends <strong>of</strong> the collagen molecule (Kivirikko et al. 1992). LHs <strong>and</strong><br />
hydroxylys<strong>in</strong>es will be discussed <strong>in</strong> more detail later (see 2.3 <strong>and</strong> 2.4).<br />
C-P4Hs, P3Hs <strong>and</strong> LHs require the c<strong>of</strong>actors Fe 2+ , 2-oxoglutarate, molecular<br />
oxygen, <strong>and</strong> ascorbate (vitam<strong>in</strong> C) for the enzymatic reaction, <strong>and</strong> the activity <strong>of</strong><br />
these enzymes is required for the formation <strong>of</strong> functionally normal collagens<br />
(Myllyharju & Kivirikko 2004, Mar<strong>in</strong>i et al. 2010). In humans the lack <strong>of</strong> vitam<strong>in</strong><br />
C, for <strong>in</strong>stance, leads to <strong>in</strong>adequate hydroxylations <strong>of</strong> collagens, the reta<strong>in</strong><strong>in</strong>g <strong>of</strong><br />
collagen cha<strong>in</strong>s with<strong>in</strong> the ER, <strong>and</strong> thus to poor connective tissue renewal which<br />
is characteristic <strong>of</strong> scurvy, for <strong>in</strong>stance (Canty & Kadler 2005). Intracellular<br />
modifications <strong>of</strong> collagen also consist <strong>of</strong> reactions that are common for many<br />
noncollagenous prote<strong>in</strong>s, such as glycosylation <strong>of</strong> certa<strong>in</strong> asparag<strong>in</strong>e residues <strong>and</strong><br />
formation <strong>of</strong> disulfide bonds (Myllyharju & Kivirikko 2004, Kielty & Grant 2002,<br />
Gelse et al. 2003).<br />
Trimerization <strong>and</strong> triple helix formation <strong>of</strong> collagens<br />
Dur<strong>in</strong>g post-translational modifications three newly synthesized pro-α-cha<strong>in</strong>s<br />
connect via <strong>their</strong> C propeptide areas to form an attachment po<strong>in</strong>t, the nucleus. The<br />
26
association is guided accord<strong>in</strong>g to specific recognition sequences. (Lees et al.<br />
1997, Kielty & Grant 2002). The trimeric structure is stabilized by formation <strong>of</strong><br />
<strong>in</strong>tercha<strong>in</strong> disulfide bonds between the C propeptides. This reaction is catalyzed<br />
by the multifunctional prote<strong>in</strong> disulfide isomerase (PDI). (Kielty & Grant 2002,<br />
Myllyharju & Kivirikko 2004). PDI also has another stabiliz<strong>in</strong>g feature when it<br />
acts as a chaperone <strong>and</strong> b<strong>in</strong>ds to newly synthesized collagen polypeptide cha<strong>in</strong>s<br />
prevent<strong>in</strong>g <strong>their</strong> aggregation (Wilson et al. 1998, Myllyharju & Kivirikko 2001).<br />
Collagen biosynthesis also requires other molecular chaperones such as the<br />
essential collagen-specific chaperone Hsp47. Hsp47 <strong>in</strong>hibits procollagen<br />
aggregation by b<strong>in</strong>d<strong>in</strong>g procollagen <strong>in</strong> the ER <strong>and</strong> facilitat<strong>in</strong>g triple helix<br />
formation. (Canty & Kadler 2005, Ishida & Nagata 2011).<br />
Triple helix formation beg<strong>in</strong>s when C propeptides have associated <strong>and</strong> the<br />
nucleus is formed, the cis peptide bonds <strong>in</strong>volv<strong>in</strong>g prol<strong>in</strong>e are converted to the<br />
trans form by the peptidyl-prolyl cis-trans isomerase, <strong>and</strong> when each pro-α-cha<strong>in</strong><br />
conta<strong>in</strong>s about 100 4-hydroxyprol<strong>in</strong>e residues. Triple helix formation propagates<br />
from the C term<strong>in</strong>us to the N term<strong>in</strong>us like a zipper. (Myllyharju & Kivirikko<br />
2001, Canty & Kadler 2002, Kielty & Grant 2002). Generally the triple helix<br />
structure prevents further hydroxylations, but one <strong>of</strong> the three LH isoenzymes,<br />
LH3, has been recently reported to modify also native triple helical collagen<br />
molecules <strong>in</strong> the extracellular space (Kielty & Grant 2002, Salo et al. 2006a).<br />
Although the enzymatic modifications <strong>of</strong> collagen polypeptides described above<br />
are necessary, overmodification is harmful. A delay <strong>in</strong> triple helix formation can<br />
lead to <strong>in</strong>tracellular overmodification <strong>of</strong> the collagen polypeptides followed by a<br />
perturbation <strong>of</strong> the f<strong>in</strong>al triple helix structure (Torre-Blanco et al. 1992,<br />
Raghunath et al. 1994). The triple helical procollagen molecules are transported<br />
from the ER to the extracellular space across the Golgi stacks probably via<br />
cisternal maturation (Myllyharju & Kivirikko 2001, Kielty & Grant 2002, Canty<br />
& Kadler 2005).<br />
Extracellular stage<br />
The assembly <strong>of</strong> collagen molecules <strong>in</strong>to fibrils is thought to be an extracellular<br />
step <strong>in</strong> collagen biosynthesis. However, the assembly may beg<strong>in</strong> already dur<strong>in</strong>g<br />
the transportation from the ER. (Canty & Kadler 2005). A key po<strong>in</strong>t <strong>in</strong> the fibril<br />
formation is the cleavage <strong>of</strong> the globular N <strong>and</strong> C propeptides <strong>of</strong> the procollagen<br />
molecules. This is carried out by two specific metalloprote<strong>in</strong>ases, a procollagen N<br />
prote<strong>in</strong>ase <strong>and</strong> a procollagen C prote<strong>in</strong>ase, respectively. The produced collagen<br />
27
molecules then assemble spontaneously <strong>in</strong>to a fibrillar structure <strong>in</strong> the case <strong>of</strong><br />
fibril-form<strong>in</strong>g collagens. F<strong>in</strong>ally, collagen cross-l<strong>in</strong>ks are formed with<strong>in</strong> <strong>and</strong><br />
between collagen molecules to stabilize the formed structures. (Myllyharju &<br />
Kivirikko 2001, Kielty & Grant 2002, Canty & Kadler 2005). This f<strong>in</strong>al crossl<strong>in</strong>k<strong>in</strong>g<br />
step beg<strong>in</strong>s with an oxidative deam<strong>in</strong>ation <strong>of</strong> the telopeptide lys<strong>in</strong>e or<br />
hydroxylys<strong>in</strong>e residue by lysyl oxidase (LO), which is followed by certa<strong>in</strong><br />
spontaneous reactions (Knott & Bailey 1998, Mäki 2009). In tissues the fibrils are<br />
usually organized further <strong>in</strong>to higher-order three-dimensional structures (Canty &<br />
Kadler 2005).<br />
Specific features <strong>in</strong> the biosynthesis <strong>of</strong> non-fibrillar collagens<br />
Although the basic features <strong>of</strong> the biosynthesis <strong>of</strong> non-fibrillar collagens are<br />
identical to those described above with fibrillar collagens, dist<strong>in</strong>ct differences also<br />
exist. F<strong>in</strong>al macromolecular assemblies determ<strong>in</strong>e the direction <strong>of</strong> the synthesis.<br />
Some collagen types are subjected to additional modifications like additions <strong>of</strong><br />
glycosam<strong>in</strong>oglycan side cha<strong>in</strong>s. It is also assumed that the triple helix <strong>of</strong><br />
transmembrane collagens may be propagated from the N to C term<strong>in</strong>us. Many<br />
collagens reta<strong>in</strong> an N- or C-term<strong>in</strong>al noncollagenous doma<strong>in</strong> or both <strong>in</strong> <strong>their</strong><br />
structures, <strong>and</strong> thus these doma<strong>in</strong>s are not cleaved dur<strong>in</strong>g extracellular<br />
modifications. (Myllyharju & Kivirikko 2001, Kielty & Grant 2002).<br />
2.3 Collagen hydroxylys<strong>in</strong>es <strong>and</strong> <strong>their</strong> function<br />
Hydroxylys<strong>in</strong>e (Hyl) residues are usually found <strong>in</strong> the Y position <strong>of</strong> the repeat<strong>in</strong>g<br />
X-Y-Gly triplet sequence <strong>in</strong> collagens <strong>and</strong> the collagenous doma<strong>in</strong>s <strong>of</strong> several<br />
noncollagenous prote<strong>in</strong>s. The repeat<strong>in</strong>g Gly-X-Y sequences are written here as<br />
X-Y-Gly due to the hydroxylation properties <strong>of</strong> collagen <strong>hydroxylases</strong>, the<br />
m<strong>in</strong>imum sequence requirement for hydroxylation by e.g. LH be<strong>in</strong>g X-Lys-Gly.<br />
Hydroxylys<strong>in</strong>e residues are also found <strong>in</strong> X-Hyl-Ser <strong>and</strong> X-Hyl-Ala sequences <strong>in</strong><br />
the telopeptide regions, i.e. <strong>in</strong> the ends <strong>of</strong> the α cha<strong>in</strong>s <strong>of</strong> some fibril-form<strong>in</strong>g<br />
collagens. (Kivirikko et al. 1992, Myllyharju 2005). A hydroxylys<strong>in</strong>e residue has<br />
been discovered <strong>in</strong> an X-Hyl-Ala sequence also <strong>in</strong> the NC1 doma<strong>in</strong> <strong>of</strong> type IV<br />
collagen, where it was suggested to participate <strong>in</strong> the formation <strong>of</strong> an Shydroxylysyl-methion<strong>in</strong>e<br />
cross-l<strong>in</strong>k, a sulfilim<strong>in</strong>e bond, with a particular<br />
methion<strong>in</strong>e residue (Vanacore et al. 2005, Vanacore et al. 2009). In addition to<br />
collagenous polypeptides, a s<strong>in</strong>gle hydroxylys<strong>in</strong>e residue is known to be present<br />
28
e.g. <strong>in</strong> anglerfish somatostat<strong>in</strong>-28 (Kivirikko et al. 1992). Recently hydroxylys<strong>in</strong>e<br />
residues were found <strong>in</strong> histone peptide fragments <strong>and</strong> <strong>in</strong> prote<strong>in</strong>s contribut<strong>in</strong>g to<br />
RNA splic<strong>in</strong>g, where the hydroxylation reaction is suggested to be catalyzed by<br />
the Jumonji doma<strong>in</strong>-6 prote<strong>in</strong> (JMJD6) <strong>and</strong> not by LHs (Chang et al. 2007,<br />
Webby et al. 2009).<br />
The lysyl hydroxylation is usually not complete <strong>and</strong> some lys<strong>in</strong>e residues<br />
located <strong>in</strong> the Y position <strong>of</strong> X-Y-Gly triplets or <strong>in</strong> telopeptides rema<strong>in</strong><br />
unhydroxylated. The hydroxylys<strong>in</strong>e content <strong>of</strong> collagens depends on the collagen<br />
type <strong>and</strong> tissue as well as the physiological <strong>and</strong> pathological state <strong>of</strong> the tissue.<br />
(Kivirikko et al. 1992, Myllyharju 2005). Generally the highest hydroxylation<br />
content is observed <strong>in</strong> type IV collagen, where almost 90% <strong>of</strong> lys<strong>in</strong>e residues <strong>in</strong><br />
the helical area are hydroxylated, whereas <strong>in</strong> collagen III, for <strong>in</strong>stance, the value<br />
is less than 20% (Kivirikko et al. 1992). The significance <strong>of</strong> hydroxylys<strong>in</strong>e<br />
residues can be seen <strong>in</strong> various diseases. The extent <strong>of</strong> hydroxylation is decreased<br />
<strong>in</strong> Bruck Syndrome (Bank et al. 1999, van der Slot et al. 2003) <strong>and</strong> EDS VIA<br />
(Ste<strong>in</strong>mann et al. 2002) <strong>and</strong> <strong>in</strong>creased <strong>in</strong> fibrotic diseases (Br<strong>in</strong>ckmann et al. 1996,<br />
Br<strong>in</strong>ckmann et al. 1999, van der Slot et al. 2003).<br />
Hydroxylys<strong>in</strong>e residues have at least two major functions <strong>in</strong> collagen<br />
molecules. They act as attachment sites for either galactose or glucosylgalactose<br />
carbohydrate residues. However, the purpose <strong>of</strong> these glycosylations is not fully<br />
understood. Hydroxylys<strong>in</strong>e residues are also essential <strong>in</strong> the formation <strong>of</strong><br />
<strong>in</strong>termolecular stabiliz<strong>in</strong>g covalent collagen cross-l<strong>in</strong>ks <strong>and</strong> thus <strong>in</strong> support<strong>in</strong>g the<br />
structure <strong>of</strong> the collagen molecule <strong>and</strong> its supramolecular assemblies. (Kivirikko<br />
& Pihlajaniemi 1998, Kielty & Grant 2002).<br />
2.3.1 Glycosylation <strong>of</strong> hydroxylys<strong>in</strong>es<br />
Carbohydrate units l<strong>in</strong>ked to hydroxylys<strong>in</strong>e residues <strong>of</strong> collagen molecules are<br />
either monosaccharide galactoses or disaccharide glucosylgalactoses (Myllyharju<br />
2005). As the hydroxylation <strong>of</strong> lys<strong>in</strong>e residues varies between collagen types,<br />
tissue distribution <strong>and</strong> physiological state, similarly the extent <strong>of</strong> the<br />
hydroxylys<strong>in</strong>e glycosylation as well as the ratio <strong>of</strong> galactose <strong>and</strong><br />
glucosylgalactose vary (Kivirikko & Myllylä 1979, Kivirikko et al. 1992). Type<br />
IV <strong>and</strong> VI collagens conta<strong>in</strong> the highest amount <strong>of</strong> glycosylated hydroxylys<strong>in</strong>es<br />
with almost complete glycosylation (Kivirikko & Myllylä 1979, Chu et al. 1988),<br />
<strong>and</strong> recently the <strong>in</strong>adequate glycosylation <strong>of</strong> these collagen types was reported to<br />
prevent <strong>their</strong> normal assembly <strong>and</strong> secretion (Sipilä et al. 2007).<br />
29
Two different transferase activities are required dur<strong>in</strong>g the two glycosylation<br />
steps. In the first step the formation <strong>of</strong> a bond between the hydroxylys<strong>in</strong>e residue<br />
<strong>and</strong> galactose moiety is catalyzed by hydroxylysyl galactosyltransferase (GT).<br />
Some <strong>of</strong> the formed galactosylhydroxylys<strong>in</strong>es are further modified by galactosyl<br />
hydroxylysyl glucosyltransferase (GGT) to form a glucosylgalactosyl<br />
hydroxylys<strong>in</strong>e residue. (Myllyharju 2005). LH3, but not LH1 or LH2, is now<br />
known to have both collagen glycosyltransferase activities <strong>in</strong> addition to its LH<br />
activity (Heikk<strong>in</strong>en et al. 2000, Rautavuoma et al. 2002, Wang et al. 2002a). It<br />
was speculated earlier that <strong>in</strong> addition to LH3 there are likely to be additional<br />
collagen glycosyltransferases <strong>in</strong> tissues, especially as the GT activity <strong>of</strong> LH3 is<br />
low (Rautavuoma et al. 2002, Rautavuoma et al. 2004) <strong>and</strong> two<br />
glycosyltransferase GLT25 family members, GLT25 doma<strong>in</strong> 1 <strong>and</strong> 2 (GLT25D1<br />
<strong>and</strong> GLT25D2), were recently observed to have strong collagen GT activity<br />
(Schegg et al. 2009). GLT25D1 is localized <strong>in</strong> the ER like LHs (Liefhebber et al.<br />
2010). It has been suggested that the GLT25D1 could be responsible for the GT<br />
reactions for example <strong>in</strong> bone type I collagen (Sricholpech et al. 2011).<br />
The function <strong>of</strong> glycosylated hydroxylys<strong>in</strong>es<br />
Although the hydroxylys<strong>in</strong>e glycosylation <strong>of</strong> collagens was observed decades ago,<br />
its exact biological role is still not fully understood (Kivirikko & Myllylä 1979).<br />
However, studies <strong>of</strong> type I <strong>and</strong> II collagens have revealed that the carbohydrate<br />
residues may affect the collagen fibril diameter <strong>and</strong> collagen fibrillogenesis<br />
(Torre-Blanco et al. 1992, Notbohm et al. 1999, Br<strong>in</strong>ckmann et al. 1999,<br />
Sricholpech et al. 2011), <strong>and</strong> an older study <strong>of</strong> type I collagen suggests that the<br />
carbohydrate moieties could prevent excess cross-l<strong>in</strong>k formation (Yamauchi et al.<br />
1982). Nevertheless, glycosylated hydroxylys<strong>in</strong>es have been shown to participate<br />
<strong>in</strong> the formation <strong>of</strong> cross-l<strong>in</strong>ks (Henkel et al. 1976, Hanson & Eyre 1996,<br />
Br<strong>in</strong>ckmann et al. 1999). The importance <strong>of</strong> glycosylated hydroxylys<strong>in</strong>e residues<br />
is also demonstrated by studies with LH3 knock-out mice. The mice die dur<strong>in</strong>g<br />
embryonic development due to lack <strong>of</strong> hydroxylys<strong>in</strong>e glycosylation activity <strong>and</strong><br />
subsequent impaired production <strong>of</strong> type IV collagen (Rautavuoma et al. 2004,<br />
Ruotsala<strong>in</strong>en et al. 2006). As <strong>in</strong>dicated earlier, the under-glycosylation <strong>of</strong> highly<br />
glycosylated collagen types IV <strong>and</strong> VI affected <strong>their</strong> assembly <strong>and</strong> secretion<br />
(Sipilä et al. 2007). Cellular studies have also shown that glycosylation catalyzed<br />
by LH3 is important for the organization <strong>of</strong> the ECM as well as cell growth <strong>and</strong><br />
viability (Risteli et al. 2009, Wang et al. 2009). Mutation studies <strong>of</strong> mannose-<br />
30
<strong>in</strong>d<strong>in</strong>g lect<strong>in</strong>s that conta<strong>in</strong> collagenous doma<strong>in</strong>s have also revealed that the<br />
glycosylation <strong>of</strong> hydroxylys<strong>in</strong>es is important <strong>in</strong> the fold<strong>in</strong>g <strong>and</strong> secretion <strong>of</strong> the<br />
prote<strong>in</strong> (Heise et al. 2000). However, just as collagens have numerous different<br />
functions, the role <strong>of</strong> hydroxylys<strong>in</strong>e-l<strong>in</strong>ked carbohydrates is presumably more<br />
diverse than discussed above. For example, at least <strong>in</strong> type II collagen<br />
glycosylated hydroxylys<strong>in</strong>e residues can act as antigens <strong>and</strong> <strong>in</strong>duce autoimmune<br />
responses such as autoimmune arthritis (Myers et al. 2004), <strong>and</strong> they are also<br />
suggested to affect the b<strong>in</strong>d<strong>in</strong>g <strong>of</strong> type IV collagen to the melanoma receptor<br />
(Lauer-Fields et al. 2003).<br />
2.3.2 Collagen cross-l<strong>in</strong>ks<br />
Collagen cross-l<strong>in</strong>ks are essential <strong>in</strong> the determ<strong>in</strong>ation <strong>of</strong> the physical <strong>and</strong><br />
chemical properties <strong>of</strong> collagens <strong>and</strong> <strong>their</strong> significance is demonstrated <strong>in</strong> many<br />
diseases (Bank et al. 1999, Br<strong>in</strong>ckmann et al. 1999, Ste<strong>in</strong>mann et al. 2002, Eyre<br />
& Wu 2005, Levental et al. 2009). Cross-l<strong>in</strong>ks can be formed between similar<br />
collagen molecules but also between different collagen types (Eyre & Wu 2005).<br />
Two op<strong>in</strong>ions exist whether one cross-l<strong>in</strong>k is formed between two or three<br />
collagen molecules (Bailey et al. 1998). Cross-l<strong>in</strong>ks can be separated <strong>in</strong>to<br />
enzyme-mediated or non-enzymatic accord<strong>in</strong>g to the <strong>in</strong>itiation <strong>of</strong> the cross-l<strong>in</strong>k<br />
formation (Reiser et al. 1992). Undesirable, occasional sugar-related cross-l<strong>in</strong>ks<br />
derived from the non-enzymatic pathway are connected to ag<strong>in</strong>g <strong>and</strong> diabetes, for<br />
<strong>in</strong>stance, while the enzymatic cross-l<strong>in</strong>k<strong>in</strong>g <strong>in</strong>itiated by LO is essential for proper<br />
tissue function (Reiser et al. 1992, Eyre & Wu 2005, Rob<strong>in</strong>s 2007, Mäki 2009).<br />
At least fibril-form<strong>in</strong>g collagen types I, II, III, V <strong>and</strong> XI as well as certa<strong>in</strong><br />
other collagen types are cross-l<strong>in</strong>ked by a mechanism requir<strong>in</strong>g LO enzyme. The<br />
formation <strong>of</strong> these cross-l<strong>in</strong>ks can follow two basic pathways, the lys<strong>in</strong>e aldehyde<br />
(allys<strong>in</strong>e) or hydroxylys<strong>in</strong>e aldehyde (hydroxyallys<strong>in</strong>e) pathway, where certa<strong>in</strong><br />
lys<strong>in</strong>e or hydroxylys<strong>in</strong>e residues, respectively, function as precursors. (Knott &<br />
Bailey 1998, Eyre & Wu 2005). At least types I, II <strong>and</strong> III <strong>of</strong> the fibrillar<br />
collagens have four lys<strong>in</strong>e residues that participate <strong>in</strong> the cross-l<strong>in</strong>k formation.<br />
Two <strong>of</strong> the residues are located <strong>in</strong> the triple helical area <strong>and</strong> two <strong>in</strong> the telopeptide<br />
regions, one residue <strong>in</strong> each telopeptide. (Knott & Bailey 1998). The<br />
alternatively-spliced form <strong>of</strong> LH2, LH2(long), is the LH that hydroxylates the<br />
telopeptide lys<strong>in</strong>es <strong>in</strong> collagens, while LH1 is ma<strong>in</strong>ly responsible for helical<br />
lys<strong>in</strong>e hydroxylation (Uzawa et al. 1999, Eyre et al. 2002, Mercer et al. 2003, van<br />
der Slot et al. 2003, Pornprasertsuk et al. 2004, Takaluoma et al. 2007). The<br />
31
hydroxylation <strong>of</strong> lys<strong>in</strong>e residues <strong>in</strong> the telopeptide area determ<strong>in</strong>es whether the<br />
allys<strong>in</strong>e or hydroxyallys<strong>in</strong>e pathway is utilized for cross-l<strong>in</strong>k<strong>in</strong>g, the<br />
hydroxylation state <strong>of</strong> the helical lys<strong>in</strong>e determ<strong>in</strong>es the f<strong>in</strong>al form for cross-l<strong>in</strong>k,<br />
<strong>and</strong> LO completes the procedure (Knott & Bailey 1998, Eyre & Wu 2005). Thus,<br />
LHs <strong>and</strong> LO are the regulatory enzymes <strong>in</strong> the cross-l<strong>in</strong>k<strong>in</strong>g pathway. Many<br />
general <strong>and</strong> connective tissue-specific factors such as TGF-β, connective tissue<br />
growth factor, tumor necrosis factor-α, vitam<strong>in</strong> B6, <strong>and</strong> the active form <strong>of</strong> vitam<strong>in</strong><br />
D regulate the activity <strong>of</strong> these enzymes <strong>and</strong> thus also cross-l<strong>in</strong>k<strong>in</strong>g. (Saito &<br />
Marumo 2010). The cross-l<strong>in</strong>k<strong>in</strong>g pathways <strong>in</strong>itiated by LO are illustrated <strong>in</strong> Fig.<br />
3.<br />
Fig. 3. <strong>Lysyl</strong> oxidase-mediated collagen cross-l<strong>in</strong>k<strong>in</strong>g. Modified from Eyre & Wu 2005.<br />
32<br />
Divalent<br />
cross-l<strong>in</strong>ks<br />
Mature<br />
cross-l<strong>in</strong>ks<br />
Allys<strong>in</strong>e pathway<br />
+ hLys<br />
LO<br />
deH-LNL<br />
tLys<br />
Allys<strong>in</strong>e<br />
Aldim<strong>in</strong>es<br />
+ hHyl<br />
deH-HLNL<br />
+ hHis<br />
HHL<br />
?<br />
Hydroxyallys<strong>in</strong>e pathway<br />
<strong>Lysyl</strong><br />
pyrrole<br />
+ hLys<br />
+ hHyl<br />
LKNL HLKNL<br />
Ketoim<strong>in</strong>es<br />
LP<br />
tHyl<br />
LO<br />
Hydroxyallys<strong>in</strong>e<br />
+ Hydroxyallys<strong>in</strong>e /<br />
HLKNL<br />
HP<br />
Hydroxylysyl<br />
pyrrole
<strong>Lysyl</strong> oxidase-mediated cross-l<strong>in</strong>k<strong>in</strong>g<br />
Enzymatic cross-l<strong>in</strong>k<strong>in</strong>g beg<strong>in</strong>s with an oxidative deam<strong>in</strong>ation <strong>of</strong> a telopeptide<br />
lys<strong>in</strong>e or hydroxylys<strong>in</strong>e residue by LO (Knott & Bailey 1998). The formed<br />
aldehydes allys<strong>in</strong>e or hydroxyallys<strong>in</strong>e, respectively, are further modified<br />
spontaneously dur<strong>in</strong>g subsequent steps <strong>of</strong> the two separate pathways (Knott &<br />
Bailey 1998, Eyre & Wu 2005) (see Fig. 3).<br />
In allys<strong>in</strong>e pathway the telopeptide lys<strong>in</strong>e residue reacts with a triple helical<br />
lys<strong>in</strong>e or hydroxylys<strong>in</strong>e residue yield<strong>in</strong>g a divalent cross-l<strong>in</strong>k dehydrolys<strong>in</strong>onorleuc<strong>in</strong>e<br />
(deH-LNL) or dehydro-hydroxylys<strong>in</strong>onorleuc<strong>in</strong>e (deH-HLNL),<br />
respectively (Knott & Bailey 1998). The deH-HLNL <strong>of</strong> these immature aldim<strong>in</strong>es<br />
can react further with a helical histid<strong>in</strong>e residue to form a mature, trifunctional<br />
cross-l<strong>in</strong>k histid<strong>in</strong>ohydroxylys<strong>in</strong>onorleuc<strong>in</strong>e (HHL), which is common <strong>in</strong> the sk<strong>in</strong><br />
for <strong>in</strong>stance (Eyre & Wu 2005, Rob<strong>in</strong>s 2007).<br />
In hydroxyallys<strong>in</strong>e pathway the telopeptide hydroxylys<strong>in</strong>e residue can react<br />
with a triple helical lys<strong>in</strong>e or hydroxylys<strong>in</strong>e residue to form a<br />
lys<strong>in</strong>o-5-ketonorleuc<strong>in</strong>e (LKNL) or a hydroxylys<strong>in</strong>o-5-ketonorleuc<strong>in</strong>e (HLKNL)<br />
cross-l<strong>in</strong>k, respectively (Knott & Bailey 1998). Depend<strong>in</strong>g on the theory, these<br />
ketoim<strong>in</strong>es are further <strong>in</strong>corporated with a lys<strong>in</strong>e aldehyde or a hydroxylys<strong>in</strong>e<br />
aldehyde or two ketoim<strong>in</strong>es are condensed to form mature pyrid<strong>in</strong>ol<strong>in</strong>e <strong>and</strong><br />
pyrrole cross-l<strong>in</strong>ks (Knott & Bailey 1998, Eyre & Wu 2005, Rob<strong>in</strong>s 2007). Thus,<br />
the pyrid<strong>in</strong>ol<strong>in</strong>es lysylpyrid<strong>in</strong>ol<strong>in</strong>e (LP) <strong>and</strong> hydroxylysylpyrid<strong>in</strong>ol<strong>in</strong>e (HP) are<br />
formed from the reaction between LKNL or HLKNL, respectively, <strong>and</strong><br />
hydroxyallys<strong>in</strong>e or another HLKNL (Knott & Bailey 1998, Eyre & Wu 2005).<br />
Both cross-l<strong>in</strong>ks are common <strong>in</strong> stiff tissues, but LP is predom<strong>in</strong>ant <strong>in</strong> calcified<br />
tissues <strong>and</strong> HP <strong>in</strong> cartilage, for <strong>in</strong>stance. <strong>Lysyl</strong> pyrrole <strong>and</strong> hydroxylysyl pyrrole<br />
are formed <strong>in</strong> turn from LKNL or HLKNL, respectively, <strong>and</strong> allys<strong>in</strong>e or<br />
deH-HLNL. (Bailey et al. 1998, Knott & Bailey 1998, Eyre & Wu 2005).<br />
Recently, a new maturation product, named arg<strong>in</strong>ol<strong>in</strong>e, was detected from the<br />
cartilage collagens. The arg<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>k is formed from the ketoim<strong>in</strong>e by an<br />
oxidation reaction followed by an addition <strong>of</strong> free arg<strong>in</strong><strong>in</strong>e. It is a functionally<br />
divalent but nonreducible cross-l<strong>in</strong>k. About half <strong>of</strong> the mature cross-l<strong>in</strong>ks <strong>in</strong> the<br />
type II collagen are arg<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks <strong>in</strong>stead <strong>of</strong> HP. (Eyre et al. 2010). A<br />
hydroxylys<strong>in</strong>e residue is also <strong>in</strong>volved <strong>in</strong> the covalent sulfilim<strong>in</strong>e bond found <strong>in</strong><br />
collagen IV. This cross-l<strong>in</strong>k is formed between a hydroxylys<strong>in</strong>e <strong>and</strong> a methion<strong>in</strong>e.<br />
(Vanacore et al. 2005, Vanacore et al. 2009).<br />
33
Cross-l<strong>in</strong>ks <strong>in</strong> tissues<br />
The cross-l<strong>in</strong>k<strong>in</strong>g pattern as well as the degree <strong>of</strong> it seem to be more tissuespecific<br />
than collagen type-specific (Knott & Bailey 1998, Eyre & Wu 2005).<br />
Cross-l<strong>in</strong>ks <strong>in</strong> sk<strong>in</strong> <strong>and</strong> other s<strong>of</strong>t tissues are ma<strong>in</strong>ly products from the allys<strong>in</strong>e<br />
pathway, whereas cross-l<strong>in</strong>ks derived from the hydroxyallys<strong>in</strong>e pathway<br />
predom<strong>in</strong>ate <strong>in</strong> tissues that require strength <strong>and</strong> stiffness such as bone, dent<strong>in</strong>,<br />
cartilage <strong>and</strong> tendon (Knott & Bailey 1998, Kielty & Grant 2002, Eyre & Wu<br />
2005). In addition, physiological processes affect the cross-l<strong>in</strong>k<strong>in</strong>g state as seen <strong>in</strong><br />
the relatively fast cervix remodell<strong>in</strong>g dur<strong>in</strong>g pregnancy, for <strong>in</strong>stance (Ak<strong>in</strong>s et al.<br />
2011). In sk<strong>in</strong> the level <strong>of</strong> lys<strong>in</strong>e hydroxylation is low <strong>and</strong> the ma<strong>in</strong> form <strong>of</strong> the<br />
divalent cross-l<strong>in</strong>k is deH-HLNL, which can mature later to HHL (Knott & Bailey<br />
1998, Avery & Bailey 2005, Rob<strong>in</strong>s 2007). Despite the high physical<br />
requirements <strong>of</strong> bone, only a small amount <strong>of</strong> mature pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks<br />
exist <strong>in</strong> bone collagens, but the exist<strong>in</strong>g cross-l<strong>in</strong>ks are formed from lys<strong>in</strong>es that<br />
are hydroxylated, i.e. ketoim<strong>in</strong>es <strong>and</strong> pyrid<strong>in</strong>ol<strong>in</strong>es (Knott & Bailey 1998). This is<br />
suggested to be caused by the <strong>in</strong>creased proportion <strong>of</strong> pyrrole cross-l<strong>in</strong>ks (Bailey<br />
et al. 1998). In addition, bone tissue is under constant remodell<strong>in</strong>g with the halflife<br />
<strong>of</strong> collagen be<strong>in</strong>g low when compared to the slowly renew<strong>in</strong>g cartilage, where<br />
lys<strong>in</strong>e residues are highly hydroxylated <strong>and</strong> collagen cross-l<strong>in</strong>ks well matured<br />
(Knott & Bailey 1998, Avery & Bailey 2005, Eyre et al. 2010). The half-life <strong>of</strong><br />
type II collagen predom<strong>in</strong>at<strong>in</strong>g <strong>in</strong> cartilage is estimated to be about 100 years<br />
(Kielty & Grant 2002, Avery & Bailey 2005).<br />
The cross-l<strong>in</strong>k<strong>in</strong>g pattern with a large amount <strong>of</strong> divalent cross-l<strong>in</strong>ks <strong>in</strong> bone<br />
is suggested to have an important role <strong>in</strong> the correct m<strong>in</strong>eralization (Yamauchi &<br />
Katz 1993, Knott et al. 1997, Bank et al. 1999, Pornprasertsuk et al. 2005), but<br />
m<strong>in</strong>eralization is also suggested to affect the maturation <strong>of</strong> collagen cross-l<strong>in</strong>ks<br />
(Knott & Bailey 1998). The nucleation <strong>of</strong> calcium apatite crystals beg<strong>in</strong>s <strong>in</strong> areas<br />
adjacent to cross-l<strong>in</strong>k<strong>in</strong>g sites, for <strong>in</strong>stance (Yamauchi & Katz 1993). In addition,<br />
the level <strong>of</strong> telopeptide hydroxylation <strong>and</strong> cross-l<strong>in</strong>k<strong>in</strong>g may also affect the<br />
collagen fibril diameter (Br<strong>in</strong>ckmann et al. 1999, Bailey 2002, Pornprasertsuk et<br />
al. 2005). Studies <strong>of</strong> osteoporotic bones have revealed over-hydroxylated<br />
collagens lead<strong>in</strong>g to <strong>in</strong>creased immature cross-l<strong>in</strong>k<strong>in</strong>g, f<strong>in</strong>er collagen fibrils,<br />
<strong>in</strong>creased collagen synthesis as well as degradation, <strong>and</strong> reduced calcification,<br />
which <strong>in</strong> turn lead to the <strong>in</strong>creased fragility <strong>of</strong> bone (Bailey 2002, Pornprasertsuk<br />
et al. 2005). Recently, decreased levels <strong>of</strong> LH2, mature HP <strong>and</strong> LP cross-l<strong>in</strong>ks <strong>and</strong><br />
two matrix prote<strong>in</strong>s <strong>in</strong> mouse cervical tissue were accompanied by <strong>in</strong>creased<br />
34
collagen fibril diameter (Ak<strong>in</strong>s et al. 2011). In addition to enzymatic collagen<br />
cross-l<strong>in</strong>ks, the role <strong>of</strong> non-enzymatic, sugar related cross-l<strong>in</strong>ks is also<br />
emphasized <strong>in</strong> the determ<strong>in</strong>ation <strong>of</strong> bone quality (see below) (Saito & Marumo<br />
2010).<br />
Cross-l<strong>in</strong>ks can be determ<strong>in</strong>ed from tissue samples, but the analysis <strong>of</strong><br />
collagen turnover products from ur<strong>in</strong>e <strong>and</strong> blood is also a useful method to detect<br />
some connective tissue disorders (Knott & Bailey 1998, Eyre & Wu 2005, Saito<br />
& Marumo 2010). Pyrid<strong>in</strong>ol<strong>in</strong>es <strong>and</strong> ketoim<strong>in</strong>es orig<strong>in</strong>ate ma<strong>in</strong>ly from<br />
m<strong>in</strong>eralized tissues <strong>and</strong> are thus valuable markers <strong>of</strong> the quality <strong>of</strong> these tissues<br />
<strong>and</strong> ongo<strong>in</strong>g collagen turnover <strong>in</strong> them (Knott & Bailey 1998). However, the<br />
analysis is not always that straightforward <strong>and</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks <strong>in</strong> samples<br />
are not always from bone collagen. For example, grow<strong>in</strong>g children get higher<br />
values due to the breakdown <strong>of</strong> m<strong>in</strong>eralized cartilage <strong>in</strong> growth plates rather than<br />
bone. (Eyre & Wu 2005).<br />
Age-related, non-enzymatic cross-l<strong>in</strong>ks<br />
Lys<strong>in</strong>e residues are also <strong>in</strong>volved <strong>in</strong> the non-enzymatic cross-l<strong>in</strong>ks between<br />
collagen molecules <strong>in</strong> advanced glycation end-products (AGEs), such as<br />
glucosepane <strong>and</strong> pentosid<strong>in</strong>e (Saito & Marumo 2010). AGEs are formed<br />
spontaneously by a reaction between reduc<strong>in</strong>g sugars <strong>and</strong> prote<strong>in</strong>s <strong>of</strong> the body<br />
(Susic 2007, Saito & Marumo 2010). The formation requires time so that it affects<br />
long-lived prote<strong>in</strong>s <strong>of</strong> the body, such as collagen (Susic 2007). As collagens<br />
mature after enzymatic cross-l<strong>in</strong>k<strong>in</strong>g, the metabolic turnover decreases <strong>and</strong> allows<br />
for non-enzymatic cross-l<strong>in</strong>k<strong>in</strong>g (Bailey et al. 1998). In contrast to the enzymemediated<br />
cross-l<strong>in</strong>ks that are a prerequisite for normal tissue function, AGEmediated<br />
cross-l<strong>in</strong>ks are connected to ag<strong>in</strong>g <strong>and</strong> they are harmful for tissues<br />
(Rob<strong>in</strong>s 2007). These cross-l<strong>in</strong>ks are suggested to have a role <strong>in</strong> diabetes mellitus,<br />
cardiovascular diseases <strong>and</strong> osteoporosis, for <strong>in</strong>stance (Susic 2007, Saito &<br />
Marumo 2010). They <strong>in</strong>crease the stiffness <strong>of</strong> the tissue <strong>and</strong> affect the resistance<br />
<strong>of</strong> the tissue to normal enzymatic degradation (Susic 2007). In bone AGEs make<br />
collagen fibers brittle <strong>and</strong> they may also affect the activity <strong>of</strong> osteoblasts <strong>and</strong><br />
osteoclasts. AGE-related cross-l<strong>in</strong>ks can be determ<strong>in</strong>ed from tissue samples <strong>and</strong>,<br />
just as turnover products <strong>of</strong> enzymatic cross-l<strong>in</strong>ks, also e.g. pentosid<strong>in</strong>e can be<br />
analyzed from serum or ur<strong>in</strong>e samples. Serum <strong>and</strong> ur<strong>in</strong>ary levels <strong>of</strong> pentosid<strong>in</strong>e<br />
do not directly reflect the content <strong>in</strong> bone but may correlate to fracture risk. (Saito<br />
& Marumo 2010).<br />
35
2.4 <strong>Lysyl</strong> <strong>hydroxylases</strong><br />
Hydroxylys<strong>in</strong>e residues <strong>in</strong> polypeptides are derived from lys<strong>in</strong>e residues by<br />
enzymatic reaction <strong>and</strong> not from free hydroxylys<strong>in</strong>es (Kivirikko & Pihlajaniemi<br />
1998). The reaction <strong>in</strong>volves an oxidative decarboxylation <strong>of</strong> 2-oxoglutarate that<br />
is catalyzed by LH, which was first purified to homogeneity from chick embryos<br />
<strong>and</strong> human placental tissue <strong>and</strong> cloned from chick (Turpeenniemi-Hujanen et al.<br />
1980, Turpeenniemi-Hujanen et al. 1981, Myllylä et al. 1991, Kivirikko &<br />
Pihlajaniemi 1998). At present, three LH is<strong>of</strong>orms, LH1, LH2 <strong>and</strong> LH3, have<br />
been characterized from vertebrates <strong>in</strong>clud<strong>in</strong>g human (Hautala et al. 1992a,<br />
Valtavaara et al. 1997, Valtavaara et al. 1998, Passoja et al. 1998a), mouse<br />
(Ruotsala<strong>in</strong>en et al. 1999), rat (Armstrong & Last 1995, Mercer et al. 2003) <strong>and</strong><br />
zebrafish (Schneider & Granato 2006, Schneider & Granato 2007). However,<br />
only a s<strong>in</strong>gle LH is found <strong>in</strong> the nematode Caenorhabditis elegans (Norman &<br />
Moerman 2000) <strong>and</strong> the fly Drosophila melanogaster (Bunt et al. 2010). LHs<br />
together with C-P4Hs <strong>and</strong> P3Hs belong to the 2-oxoglutarate-dependent<br />
dioxygenase family <strong>and</strong> the genes encod<strong>in</strong>g LHs are called PLODs<br />
(protocollagen-lys<strong>in</strong>e, 2-oxoglutarate, 5-dioxygenase) (Kivirikko & Pihlajaniemi<br />
1998, Kielty & Grant 2002).<br />
Recently a Jumonji doma<strong>in</strong>-6 prote<strong>in</strong> (JMJD6) was also reported to have<br />
lysyl hydroxylation activity (Webby et al. 2009). Like collagen-modify<strong>in</strong>g<br />
<strong>hydroxylases</strong>, JMJD6 is suggested to be an iron- <strong>and</strong> 2-oxoglutarate-dependent<br />
dioxygenase that belongs to the subfamily <strong>of</strong> Jumonji C doma<strong>in</strong>-conta<strong>in</strong><strong>in</strong>g<br />
<strong>hydroxylases</strong> (Chang et al. 2007). These enzymes <strong>in</strong> turn are connected to<br />
diseases such as cancer (for a review, see Kampranis & Tsichlis 2009). The latest<br />
data suggest that JMJD6 hydroxylates lys<strong>in</strong>e residues <strong>in</strong> histones <strong>and</strong> <strong>in</strong> prote<strong>in</strong>s<br />
contribut<strong>in</strong>g to RNA splic<strong>in</strong>g (Webby et al. 2009) <strong>and</strong> that the enzyme may also<br />
b<strong>in</strong>d <strong>and</strong> modify s<strong>in</strong>gle-str<strong>and</strong>ed RNA (Hong et al. 2010).<br />
LHs are localized to the lumen <strong>of</strong> the ER (Myllylä et al. 1991, Kellokumpu et<br />
al. 1994) even though they do not conta<strong>in</strong> any retention sequence typical for<br />
prote<strong>in</strong>s rema<strong>in</strong><strong>in</strong>g <strong>in</strong> the ER (Myllylä et al. 1991, Hautala et al. 1992a). In<br />
addition, any cycl<strong>in</strong>g or retrieval from post-ER compartments is not <strong>in</strong>volved<br />
(Suokas et al. 2003). LH1 <strong>and</strong> LH3 have been shown to be associated with the ER<br />
membrane (Kellokumpu et al. 1994, Salo et al. 2006a). LH1 conta<strong>in</strong>s a dist<strong>in</strong>ct Cterm<strong>in</strong>al<br />
segment that is suggested to be important both for its localization <strong>in</strong>to<br />
the lumen <strong>of</strong> the ER as well as its association with ER membranes (Kellokumpu<br />
et al. 1994, Suokas et al. 2000, Suokas et al. 2003). Thus, the retention with<strong>in</strong> the<br />
36
ER <strong>and</strong> the association with the ER membrane are found to be coupled<br />
phenomena as both characteristics are due to the C-term<strong>in</strong>al segment (Suokas et<br />
al. 2000). This C-term<strong>in</strong>al region shows high sequence homology <strong>in</strong> all three LH<br />
is<strong>of</strong>orms (Valtavaara et al. 1998, Passoja et al. 1998a), <strong>and</strong> it is assumed that all<br />
three LH isoenzymes utilize the same mechanism <strong>in</strong> <strong>their</strong> localization to the ER<br />
(Suokas et al. 2000, Suokas et al. 2003), which is also supported by the latest data<br />
concern<strong>in</strong>g LH3 (Wang et al. 2012). In addition to the ER, the active form <strong>of</strong> LH3<br />
is found <strong>in</strong> the extracellular space <strong>and</strong> the localization seems to have tissuespecificity.<br />
The <strong>in</strong>tracellular form is preferred at least <strong>in</strong> the kidney, whereas the<br />
extracellular form is prevalent <strong>in</strong> the liver. In addition, LH3 is detected <strong>in</strong> the<br />
serum. (Salo et al. 2006a).<br />
2.4.1 <strong>Lysyl</strong> hydroxylase isoenzymes<br />
The existence <strong>of</strong> LH isoenzymes <strong>and</strong> especially separate forms for the<br />
hydroxylation <strong>of</strong> the helical <strong>and</strong> telopeptide regions was debated for a long time<br />
(Barnes et al. 1974, Ihme et al. 1984, Royce & Barnes 1985, Gerriets et al. 1993).<br />
The presence <strong>of</strong> hydroxylys<strong>in</strong>e residues <strong>in</strong> dist<strong>in</strong>ct am<strong>in</strong>o acid sequences <strong>in</strong><br />
helical <strong>and</strong> telopeptide regions, for <strong>in</strong>stance, led to the suggestion that different<br />
forms <strong>of</strong> LHs may exist (Barnes et al. 1974, Royce & Barnes 1985). Moreover,<br />
EDS VIA patients were reported to have immense variability <strong>in</strong> the lys<strong>in</strong>e<br />
hydroxylation levels between various tissues <strong>and</strong> collagen types, values rang<strong>in</strong>g<br />
from total lack <strong>of</strong> hydroxylation to normal levels (Ihme et al. 1984), <strong>and</strong> highly<br />
purified LH was reported to be able to hydroxylate essentially completely the<br />
triple helical lys<strong>in</strong>es <strong>in</strong> type I protocollagen, while telopeptide lys<strong>in</strong>es were<br />
unhydroxylated (Royce & Barnes 1985). F<strong>in</strong>ally, two new is<strong>of</strong>orms, LH2<br />
(Valtavaara et al. 1997) <strong>and</strong> LH3 (Valtavaara et al. 1998, Passoja et al. 1998a),<br />
were identified <strong>and</strong> later LH2 was recognized to have two different splic<strong>in</strong>g forms<br />
(Yeowell & Walker 1999a). PLOD1, PLOD2 <strong>and</strong> PLOD3, the genes encod<strong>in</strong>g<br />
human LH1, LH2 <strong>and</strong> LH3 (Hautala et al. 1992a, Szpirer et al. 1997, Valtavaara<br />
et al. 1998), respectively, as well as the correspond<strong>in</strong>g mouse genes (Plod1, Plod2<br />
<strong>and</strong> Plod3) are located <strong>in</strong> different chromosomes (Sipilä et al. 2000).<br />
Accord<strong>in</strong>g to a phylogenetic study, it has been suggested that the mouse<br />
Plod1 <strong>and</strong> Plod2 is<strong>of</strong>orms are derived from one gene by two gene duplications,<br />
Plod3 be<strong>in</strong>g the ancestral gene (Ruotsala<strong>in</strong>en et al. 1999). At the am<strong>in</strong>o acid level<br />
all three human LH is<strong>of</strong>orms share about 50% identity (Valtavaara et al. 1998). If<br />
<strong>in</strong>dividual is<strong>of</strong>orms are compared (LH1-LH2, LH1-LH3, LH2-LH3) the identities<br />
37
are even higher, rang<strong>in</strong>g from about 60% to over 75% <strong>in</strong> both human <strong>and</strong> mouse<br />
(Valtavaara et al. 1997, Valtavaara et al. 1998, Passoja et al. 1998a, Ruotsala<strong>in</strong>en<br />
et al. 1999). Between species, <strong>in</strong>dividual mouse LHs are about 90% identical to<br />
the correspond<strong>in</strong>g human LHs at the am<strong>in</strong>o acid level (Ruotsala<strong>in</strong>en et al. 1999).<br />
The most concerved areas are located <strong>in</strong> the C-term<strong>in</strong>al region, where the am<strong>in</strong>o<br />
acids essential for LH activity are located, <strong>and</strong> <strong>in</strong> the central region <strong>of</strong> the<br />
molecule (Pirskanen et al. 1996, Valtavaara et al. 1997, Passoja et al. 1998a,<br />
Passoja et al. 1998b, Valtavaara et al. 1998, Rautavuoma et al. 2002).<br />
<strong>Lysyl</strong> hydroxylase 1<br />
LH1 was the first LH isoenzyme cloned from chick <strong>and</strong> human <strong>and</strong> it was named<br />
LH before another is<strong>of</strong>orm, LH2, was discovered (Hautala et al. 1992a,<br />
Valtavaara et al. 1997). The gene cod<strong>in</strong>g for human LH1, PLOD1, is located on<br />
chromosome 1 <strong>and</strong> it conta<strong>in</strong>s 19 exons (Hautala et al. 1992a, Heikk<strong>in</strong>en et al.<br />
1994). PLOD1 encodes a polypeptide <strong>of</strong> 727 am<strong>in</strong>o acids <strong>in</strong>clud<strong>in</strong>g a signal<br />
peptide <strong>of</strong> 18 residues (Hautala et al. 1992a). In mouse the gene is localized to<br />
chromosome 4 (Sipilä et al. 2000). Interest<strong>in</strong>gly, <strong>in</strong>trons 9 <strong>and</strong> 16 <strong>of</strong> PLOD1 show<br />
extensive homology because <strong>of</strong> many Alu sequences (Heikk<strong>in</strong>en et al. 1994).<br />
These sequences are suggested to be <strong>in</strong>volved <strong>in</strong> many human gene<br />
rearrangements <strong>and</strong> therefore <strong>in</strong> some hereditary diseases (De<strong>in</strong><strong>in</strong>ger & Batzer<br />
1999). In the case <strong>of</strong> PLOD1, Alu sequences have been connected to one <strong>of</strong> the<br />
most common mutations caus<strong>in</strong>g EDS VIA (Heikk<strong>in</strong>en et al. 1994, Pousi et al.<br />
1994, Heikk<strong>in</strong>en et al. 1997, De<strong>in</strong><strong>in</strong>ger & Batzer 1999).<br />
<strong>Lysyl</strong> hydroxylase 2<br />
The human PLOD2 gene cod<strong>in</strong>g for LH2 is localized to chromosome 3 (Szpirer et<br />
al. 1997) <strong>and</strong> the correspond<strong>in</strong>g mouse gene Plod2 to chromosome 9 (Sipilä et al.<br />
2000). Mutations <strong>in</strong> the PLOD2 gene cause Bruck syndrome (BS) type 2 (see<br />
2.5.2., van der Slot et al. 2003, Ha-V<strong>in</strong>h et al. 2004). LH2 is the only LH is<strong>of</strong>orm<br />
that is known to have different splic<strong>in</strong>g variants (Yeowell & Walker 1999a). The<br />
two different forms are named LH2(short) <strong>and</strong> LH2(long) or LH2a <strong>and</strong> LH2b,<br />
respectively (Yeowell & Walker 1999a, Walker et al. 2005). LH2(long) was<br />
characterized two years later than the short form (Valtavaara et al. 1997, Yeowell<br />
& Walker 1999a). In the short form exon 13A between exons 13 <strong>and</strong> 14 is spliced<br />
out, yield<strong>in</strong>g a product <strong>of</strong> 737 am<strong>in</strong>o acids that <strong>in</strong>cludes an N-term<strong>in</strong>al signal<br />
38
peptide (Valtavaara et al. 1997, Yeowell & Walker 1999a). Exon 13A is 63 bp<br />
long correspond<strong>in</strong>g to 21 am<strong>in</strong>o acids, <strong>and</strong> thus the long form is 758 am<strong>in</strong>o acids<br />
long (Yeowell & Walker 1999a). The am<strong>in</strong>o acids encoded by the additional exon<br />
13A <strong>in</strong> LH2(long) are suggested to change the peptide b<strong>in</strong>d<strong>in</strong>g properties <strong>of</strong> LH2<br />
(Risteli et al. 2004).<br />
The expression <strong>of</strong> LH2 splice variants shows tissue specificity (Yeowell &<br />
Walker 1999a, Salo et al. 2006b). In humans LH2(long) is expressed <strong>in</strong> all tissues<br />
studied, while the short form is expressed <strong>in</strong> certa<strong>in</strong> tissues <strong>and</strong> only together with<br />
the long form (Yeowell & Walker 1999a, Walker et al. 2005). The LH2 splic<strong>in</strong>g<br />
pattern is suggested to be connected to the specificity for collagen telopeptide<br />
lys<strong>in</strong>e hydroxylation (Mercer et al. 2003, van der Slot et al. 2003). LH2, <strong>and</strong><br />
particularly LH2(long), is reported to be the collagen telopeptide LH responsible<br />
for the <strong>in</strong>itiation <strong>of</strong> the formation <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks (Uzawa et al. 1999,<br />
Mercer et al. 2003, van der Slot et al. 2003, Pornprasertsuk et al. 2004, van der<br />
Slot et al. 2004, Takaluoma et al. 2007). The role <strong>of</strong> the short form is unknown. In<br />
fibrotic diseases such as systemic sclerosis, the splic<strong>in</strong>g <strong>of</strong> LH2 is changed so that<br />
LH2(long) expression is <strong>in</strong>creased (van der Slot et al. 2003, van der Slot et al.<br />
2004). Accord<strong>in</strong>gly, overaccumulation <strong>of</strong> collagen <strong>in</strong> fibrotic conditions is<br />
accompanied by overhydroxylation <strong>of</strong> the collagen telopeptides <strong>and</strong> an <strong>in</strong>creased<br />
amount <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks (Trojanowska et al. 1998, Br<strong>in</strong>ckmann et al.<br />
1999, Br<strong>in</strong>ckmann et al. 2001, van der Slot et al. 2004).<br />
The prevalence <strong>and</strong> cl<strong>in</strong>ical significance <strong>of</strong> fibrosis have led to <strong>in</strong>creas<strong>in</strong>g<br />
<strong>in</strong>vestigation <strong>of</strong> the factors regulat<strong>in</strong>g splic<strong>in</strong>g <strong>of</strong> the LH2 mRNA. Early studies<br />
<strong>of</strong> LH2 splic<strong>in</strong>g with a prote<strong>in</strong> synthesis <strong>in</strong>hibitor cycloheximide demonstrated<br />
that the splic<strong>in</strong>g pattern can be changed by affect<strong>in</strong>g the synthesis <strong>of</strong> a prote<strong>in</strong><br />
factor necessary for the splic<strong>in</strong>g (Walker et al. 2005). Phylogenetic <strong>and</strong> sequence<br />
analyses have revealed that <strong>in</strong>trons flank<strong>in</strong>g exon 13A <strong>of</strong> the LH2 transcript <strong>and</strong><br />
exon 13A itself conta<strong>in</strong> certa<strong>in</strong> regulatory splic<strong>in</strong>g sequences (Yeowell & Walker<br />
1999a, Yeowell et al. 2009, Seth & Yeowell 2010, Seth et al. 2011). A branch<br />
po<strong>in</strong>t sequence, suggested to participate <strong>in</strong> the exon <strong>in</strong>clusion, is followed by a<br />
polypyrimid<strong>in</strong>e tract that could contribute to the exclusion <strong>of</strong> exon 13A (Yeowell<br />
& Walker 1999a, Yeowell et al. 2009). In addition, a sequence similar to a U-rich<br />
element that is known to function as a b<strong>in</strong>d<strong>in</strong>g site for TIA (T-cell-restricted<br />
<strong>in</strong>tracellular antigen) prote<strong>in</strong>s is detected downstream <strong>of</strong> exon 13A (Förch et al.<br />
2000, Del Gatto-Konczak et al. 2000, Le Gu<strong>in</strong>er et al. 2001, Yeowell et al. 2009).<br />
The b<strong>in</strong>d<strong>in</strong>g <strong>of</strong> TIA prote<strong>in</strong>s to U-rich elements has been connected to the<br />
<strong>in</strong>clusion <strong>of</strong> alternatively spliced exons <strong>in</strong> general (Förch et al. 2000, Del Gatto-<br />
39
Konczak et al. 2000, Le Gu<strong>in</strong>er et al. 2001). Actually, two TIA prote<strong>in</strong>s, TIA-1<br />
<strong>and</strong> TIAL1, are shown to be essential for efficient <strong>in</strong>clusion <strong>of</strong> the alternatively<br />
spliced exon 13A <strong>in</strong> LH2(long) mRNA <strong>and</strong> thus for the regulation <strong>of</strong> LH2(long)<br />
expression (Yeowell et al. 2009). The <strong>in</strong>tron upstream <strong>of</strong> exon 13A also conta<strong>in</strong>s<br />
highly conserved motifs that b<strong>in</strong>d Fox prote<strong>in</strong>s (Baraniak et al. 2006, Seth &<br />
Yeowell 2010). These prote<strong>in</strong>s are known to regulate alternative splic<strong>in</strong>g<br />
(Baraniak et al. 2006), <strong>and</strong> <strong>in</strong> the splic<strong>in</strong>g <strong>of</strong> LH2 Fox-2 has also been shown to<br />
be required for the <strong>in</strong>clusion <strong>of</strong> exon 13A <strong>and</strong> thus production <strong>of</strong> the LH2(long)<br />
mRNA (Seth & Yeowell 2010). Recently exon 13A itself was also reported to<br />
conta<strong>in</strong> elements affect<strong>in</strong>g the alternative splic<strong>in</strong>g. Exon 13A is suggested to<br />
conta<strong>in</strong> at least one enhancer <strong>and</strong> two silencer elements <strong>of</strong> exonic splic<strong>in</strong>g that<br />
could b<strong>in</strong>d regulatory prote<strong>in</strong>s. (Seth et al. 2011).<br />
<strong>Lysyl</strong> hydroxylase 3<br />
The PLOD3 gene encod<strong>in</strong>g human LH3 is located on chromosome 7 (Valtavaara<br />
et al. 1998). The gene conta<strong>in</strong>s 19 exons <strong>and</strong> the size <strong>of</strong> the LH3 polypeptide is<br />
738 am<strong>in</strong>o acids <strong>in</strong>clud<strong>in</strong>g a 24-residue signal peptide (Valtavaara et al. 1998,<br />
Rautavuoma et al. 2000). The gene conta<strong>in</strong>s Alu sequences <strong>in</strong> the <strong>in</strong>trons like that<br />
<strong>of</strong> LH1, but the Alu sequences are shorter (Rautavuoma et al. 2000). In mouse the<br />
gene is localized to chromosome 5 (Sipilä et al. 2000). LH3 is a multifunctional<br />
enzyme possess<strong>in</strong>g GT <strong>and</strong> GGT activities <strong>in</strong> addition to its LH activity<br />
(Heikk<strong>in</strong>en et al. 2000, Rautavuoma et al. 2002, Wang et al. 2002a, Rautavuoma<br />
et al. 2004, Ruotsala<strong>in</strong>en et al. 2006, Salo et al. 2006a).<br />
The glycosyltransferase activities <strong>of</strong> LH3 are localized <strong>in</strong> the N-term<strong>in</strong>us <strong>in</strong><br />
contrast to the LH activity which exists <strong>in</strong> the C-term<strong>in</strong>us <strong>of</strong> the enzyme<br />
(Pirskanen et al. 1996, Passoja et al. 1998b, Heikk<strong>in</strong>en et al. 2000, Rautavuoma<br />
et al. 2002, Wang et al. 2002a, Wang et al. 2002b). LH3 differs from other LH<br />
is<strong>of</strong>orms also <strong>in</strong> its cellular localization. In addition to its location <strong>in</strong> the lumen <strong>of</strong><br />
the ER, it is found as an active enzyme <strong>in</strong> the extracellular space <strong>and</strong> <strong>in</strong><br />
association with collagenous prote<strong>in</strong>s (Kellokumpu et al. 1994, Salo et al. 2006a).<br />
Thus, <strong>in</strong> addition to the general conception that LHs modify collagen<br />
polypeptides only <strong>in</strong> the ER lumen before the formation <strong>of</strong> the triple helical<br />
structure, it has been suggested that LH3 could modify also native collagen<br />
molecules <strong>in</strong> the extracellular space, which furthermore <strong>in</strong>dicates a possible role<br />
for LH3 <strong>in</strong> ECM remodel<strong>in</strong>g (Kielty & Grant 2002, Salo et al. 2006a). Actually,<br />
cellular studies have suggested that glycosylation activity <strong>in</strong> the extracellular<br />
40
space could dramatically affect cell morphology <strong>and</strong> cell behavior, although it<br />
depends on the cell type (Wang et al. 2009). However, LH3 is suggested to prefer<br />
collagen polypeptide cha<strong>in</strong>s rather than native triple helical collagen molecules as<br />
substrates (Sricholpech et al. 2011). The enzyme is reported to be secreted from<br />
cells via two pathways by us<strong>in</strong>g a Golgi complex or by bypass<strong>in</strong>g it (Wang et al.<br />
2012). LH3 is also found from serum samples <strong>in</strong>dicat<strong>in</strong>g that the GGT activity<br />
present <strong>in</strong> serum could be derived from LH3 (Salo et al. 2006a).<br />
Expression <strong>of</strong> different lysyl hydroxylase is<strong>of</strong>orms<br />
Expression levels <strong>of</strong> LH isoenzymes <strong>in</strong> human <strong>and</strong> mouse have been studied at<br />
the mRNA <strong>and</strong> prote<strong>in</strong> levels. LHs are widely expressed <strong>in</strong> different tissues <strong>and</strong><br />
developmental stages. Straightforward conclusions about any specific LH<br />
isoenzyme expression patterns are essentially impossible as the results vary<br />
markedly between different studies. The expression patterns <strong>of</strong> LH is<strong>of</strong>orms <strong>in</strong><br />
certa<strong>in</strong> adult human <strong>and</strong> mouse tissues are shown <strong>in</strong> Table 1.<br />
Table 1. Expression <strong>of</strong> lysyl hydroxylase isoenzymes <strong>in</strong> selected adult human <strong>and</strong><br />
mouse tissues.<br />
Tissue LH1 LH2 LH3<br />
human mouse human mouse human mouse<br />
Blood vessels yes 2 n.d. yes 7 n.d. n.d. yes 9<br />
Bra<strong>in</strong> yes 1 yes 6,8 yes 3,7 yes 6,8,10 yes 4,5 yes 6,10<br />
Cartilage yes 2 n.d. yes 7 n.d. n.d. yes 9<br />
Heart yes 1 yes 6,8 yes 3 yes 6,8,10 yes 4,5 yes 6<br />
Kidney n.d. no 8 /yes 6 yes 3,7 yes 6,8,10 yes 4,5 yes 6,9,10<br />
Liver yes 1 yes 6,8 yes 3,7 no 6,8,10 yes 4,5 yes 6,9,10<br />
Lung yes 1,2 yes 6,8 yes 3,7 no 8 /yes 6,10 yes 4,5 yes 6,9,10<br />
Pancreas n.d. n.d. yes 3 yes 10 yes 4,5 yes 10<br />
Placenta yes 1,2 n.d. yes 3,7 n.d. yes 4,5 yes 9<br />
Skeletal muscle yes 1,8 yes 6,8 yes 3,8 no 8 /yes 6,10 yes 4,5 yes 6,10<br />
1 (Heikk<strong>in</strong>en et al. 1994); 2 (Yeowell et al. 1994); 3 (Valtavaara et al. 1997); 4 (Passoja et al. 1998a);<br />
5 (Valtavaara et al. 1998); 6 (Ruotsala<strong>in</strong>en et al. 1999); 7 (Yeowell & Walker 1999a); 8 (Hjalt et al. 2001);<br />
9 (Rautavuoma et al. 2004); 10 (Salo et al. 2006b); n.d. = not determ<strong>in</strong>ed.<br />
In rat tissues mRNA expression <strong>of</strong> all LH is<strong>of</strong>orms <strong>in</strong>clud<strong>in</strong>g both LH2 splic<strong>in</strong>g<br />
variants was detected <strong>in</strong> almost all tissue types exam<strong>in</strong>ed, <strong>in</strong>dicat<strong>in</strong>g that LHs do<br />
not have tissue-specific expression <strong>in</strong> the rat (Mercer et al. 2003). In humans,<br />
both LH2 variants are expressed <strong>in</strong> the cartilage, frontal lobe, kidney, liver, spleen<br />
41
<strong>and</strong> placenta (Yeowell & Walker 1999a). LH2(long) is expressed <strong>in</strong> all tissues<br />
studied so far <strong>and</strong> it is the ma<strong>in</strong> transcript <strong>in</strong> all tissues studied except the kidney<br />
<strong>and</strong> spleen (Yeowell & Walker 1999a). In mouse the short form <strong>of</strong> LH2 is<br />
predom<strong>in</strong>ant <strong>in</strong> the testis, bra<strong>in</strong> <strong>and</strong> kidney (Salo et al. 2006b). Thus, the<br />
expression <strong>of</strong> the two human <strong>and</strong> mouse LH2 splic<strong>in</strong>g variants is suggested to be<br />
tissue-specific (Yeowell & Walker 1999a, Salo et al. 2006b).<br />
All LH is<strong>of</strong>orms have been shown to be expressed with similar tissue<br />
distributions <strong>in</strong> the mouse embryos from embryonic day (E) 7 onwards, the first<br />
time po<strong>in</strong>t analyzed. LH1 expression was high throughout the analyzed period,<br />
E7-E17, but LH2 was expressed strongly only at E7 <strong>and</strong> rema<strong>in</strong>ed low after that.<br />
LH3 was highly expressed at E7 <strong>and</strong> E15, the expression level be<strong>in</strong>g lower at the<br />
other time po<strong>in</strong>ts. (Salo et al. 2006b). Earlier data from LH3 null mice with the<br />
Plod3 gene disrupted with a β-galactosidase reporter demonstrated <strong>in</strong>tense X-gal<br />
sta<strong>in</strong><strong>in</strong>g <strong>in</strong> heterozygous <strong>and</strong> null embryos at E8.5-9.5, the sta<strong>in</strong><strong>in</strong>g be<strong>in</strong>g<br />
localized also to placental blood vessels (Rautavuoma et al. 2004). At E9.5 <strong>and</strong><br />
E11.5 all LH is<strong>of</strong>orms are expressed <strong>in</strong> the dorsal aorta, somites, tail bud, bra<strong>in</strong><br />
<strong>and</strong> at sites <strong>of</strong> the develop<strong>in</strong>g maxilla <strong>and</strong> m<strong>and</strong>ible <strong>of</strong> branchial arch as well as <strong>in</strong><br />
the limb bud at E11.5. Expression <strong>of</strong> the two LH2 splic<strong>in</strong>g forms is suggested to<br />
be developmentally regulated dur<strong>in</strong>g mouse embryogenesis as the short form<br />
predom<strong>in</strong>ates until E11.5 <strong>and</strong> the long form after that. (Salo et al. 2006b).<br />
However, <strong>in</strong> humans, only the long form is present <strong>in</strong> fetal sk<strong>in</strong> cell l<strong>in</strong>es, but both<br />
forms are seen <strong>in</strong> embryonic kidney cells (Yeowell & Walker 1999a, Walker et al.<br />
2005).<br />
2.4.2 Molecular <strong>and</strong> catalytic properties<br />
An active form <strong>of</strong> LHs is a homodimer, the size <strong>of</strong> one monomer be<strong>in</strong>g about<br />
85000 Da (Turpeenniemi-Hujanen et al. 1980, Turpeenniemi-Hujanen et al. 1981,<br />
Myllylä et al. 1988, Rautavuoma et al. 2002). The dimerization <strong>of</strong> human LH3<br />
has been reported to be dependent on am<strong>in</strong>o acids 541-547, but am<strong>in</strong>o acids from<br />
other sites may also contribute (Heikk<strong>in</strong>en et al. 2011). However, dimerization<br />
seems not to be a prerequisite for the gycosylation activity <strong>of</strong> LH3 (Heikk<strong>in</strong>en et<br />
al. 2011). LHs consist <strong>of</strong> three doma<strong>in</strong>s (Rautavuoma et al. 2002) <strong>and</strong> the Cterm<strong>in</strong>al<br />
doma<strong>in</strong> conta<strong>in</strong>s essential am<strong>in</strong>o acids required for the LH activity<br />
(Pirskanen et al. 1996, Passoja et al. 1998b). The glycosylation activity <strong>of</strong> LH3 is<br />
located <strong>in</strong> its N-term<strong>in</strong>al doma<strong>in</strong> (Rautavuoma et al. 2002, Wang et al. 2002a,<br />
Wang et al. 2002b).<br />
42
Reaction mechanism <strong>and</strong> catalytically important am<strong>in</strong>o acids<br />
As a member <strong>of</strong> the 2-oxoglutarate dioxygenase family, LHs, as well as C-P4Hs<br />
<strong>and</strong> P3Hs, require Fe 2+ , 2-oxoglutarate, molecular oxygen <strong>and</strong> ascorbate for the<br />
enzymatic reaction, <strong>and</strong> the reaction mechanism <strong>of</strong> all these collagen<br />
<strong>hydroxylases</strong> resembles each other. Fe 2+ , 2-oxoglutarate, O2 <strong>and</strong> the peptide<br />
substrate are bound <strong>in</strong> this order dur<strong>in</strong>g the enzyme reaction. Reaction products<br />
are released <strong>in</strong> sequence as well, although Fe 2+ rema<strong>in</strong>s usually unreleased<br />
between catalytic cycles. The substrate is released first, followed by CO2 <strong>and</strong><br />
succ<strong>in</strong>ate. (Myllyharju 2005). A simplified representation <strong>of</strong> the hydroxylation <strong>of</strong><br />
a lys<strong>in</strong>e by LH is shown <strong>in</strong> Fig. 4.<br />
NH CH C<br />
CH2 CH2 CH2 O<br />
CH NH<br />
2 2<br />
HOOC<br />
NH CH<br />
O<br />
C<br />
COOH<br />
+<br />
HOOC<br />
O<br />
O2 Fe<br />
Ascorbate<br />
CH2 CH2 HC OH<br />
+<br />
HOOC<br />
CH2NH2 2+<br />
+ +<br />
Lys<strong>in</strong>e 2-Oxoglutarate<br />
Hydroxylys<strong>in</strong>e Succ<strong>in</strong>ate<br />
Fig. 4. A schematic representation <strong>of</strong> the hydroxylation <strong>of</strong> lys<strong>in</strong>e by lysyl hydroxylase.<br />
In the lysyl hydroxylation reaction, molecular oxygen b<strong>in</strong>ds to the Fe 2+ <strong>in</strong> the<br />
catalytic site <strong>of</strong> LH. One oxygen molecule <strong>of</strong> the O2 is transferred to the hydroxyl<br />
group formed <strong>in</strong> the lys<strong>in</strong>e, while another oxygen atom is <strong>in</strong>corporated <strong>in</strong>to the<br />
succ<strong>in</strong>ate produced <strong>in</strong> the reaction. (Kivirikko et al. 1992, Kivirikko &<br />
Pihlajaniemi 1998). Ascorbate is essential for lysyl hydroxylation, although it is<br />
not consumed stoichiometrically <strong>in</strong> the complete hydroxylation reaction. Collagen<br />
<strong>hydroxylases</strong> can perform many catalytic cycles <strong>in</strong> the absence <strong>of</strong> ascorbate, but it<br />
is needed as an obligatory element to reactivate the enzymes. This is because the<br />
enzymes occasionally catalyze uncoupled decarboxylation <strong>of</strong> 2-oxoglutarate,<br />
where ascorbate is consumed stoichiometrically. (Myllylä et al. 1984).<br />
The substrate b<strong>in</strong>d<strong>in</strong>g site <strong>of</strong> LH is suggested to be an open, hydrophilic<br />
pocket where acidic am<strong>in</strong>o acids have a significant role <strong>in</strong> b<strong>in</strong>d<strong>in</strong>g (Risteli et al.<br />
2004). At the am<strong>in</strong>o acid level <strong>in</strong>dividual human <strong>and</strong> mouse LH is<strong>of</strong>orms are<br />
about 60–75% identical, <strong>and</strong> when all is<strong>of</strong>orms are compared together the identity<br />
is about 50% (Valtavaara et al. 1997, Valtavaara et al. 1998, Passoja et al. 1998a<br />
Ruotsala<strong>in</strong>en et al. 1999). The C-term<strong>in</strong>al region is especially highly conserved <strong>in</strong><br />
LHs, conta<strong>in</strong><strong>in</strong>g several critical am<strong>in</strong>o acids that are found <strong>in</strong> other 2-oxoglutarate<br />
CO 2<br />
43
dioxygenase family members as well (Pirskanen et al. 1996, Myllyharju &<br />
Kivirikko 1997, Valtavaara et al. 1997, Passoja et al. 1998a, Passoja et al. 1998b,<br />
Valtavaara et al. 1998, Vranka et al. 2004). It is assumed that Fe 2+ is bound to LH<br />
with three l<strong>in</strong>kages (Kivirikko & Pihlajaniemi 1998) via the am<strong>in</strong>o acids histid<strong>in</strong>e<br />
638, aspartate 640 <strong>and</strong> histid<strong>in</strong>e 690 <strong>in</strong> processed human LH1 (Pirskanen et al.<br />
1996). Many divalent cations, such as Zn 2+ , function as competitive <strong>in</strong>hibitors<br />
with respect to Fe 2+ (Kivirikko & Pihlajaniemi 1998). In addition, arg<strong>in</strong><strong>in</strong>e 700<br />
(<strong>in</strong> processed human LH1) is a positively charged residue that ionically b<strong>in</strong>ds the<br />
C5 carboxyl group <strong>of</strong> 2-oxoglutarate (Hanauske-Abel & Günzer 1982, Passoja et<br />
al. 1998b). LHs also conta<strong>in</strong> several conserved cyste<strong>in</strong>e residues <strong>and</strong> seven <strong>of</strong><br />
them are shown to be important for the enzyme activity <strong>of</strong> LH1 (Yeowell et al.<br />
2000a).<br />
There is only a limited amount <strong>of</strong> data available about catalytically important<br />
am<strong>in</strong>o acids for the glycosyltransferase activity <strong>of</strong> LH3. However, <strong>in</strong>dications<br />
about critical am<strong>in</strong>o acids have been obta<strong>in</strong>ed by compar<strong>in</strong>g the cDNA sequences<br />
<strong>of</strong> the human <strong>and</strong> mouse LH3s to the only LH <strong>of</strong> Caenorhabditis elegans, which<br />
is reported to possess both LH <strong>and</strong> glycosylation activity (Norman & Moerman<br />
2000, Wang et al. 2002a, Wang et al. 2002b). 29 am<strong>in</strong>o acids are conserved <strong>in</strong><br />
these sequences <strong>of</strong> which cyste<strong>in</strong>e 144 <strong>and</strong> leuc<strong>in</strong>e 208 <strong>in</strong> the human sequence<br />
were shown to be important for the LH3 GGT activity <strong>in</strong> all three species (Wang<br />
et al. 2002b). The cyste<strong>in</strong>e but not the leuc<strong>in</strong>e residue is also important for the GT<br />
activity <strong>of</strong> the enzyme (Wang et al. 2002a). In addition, a conserved motif at<br />
am<strong>in</strong>o acids 187–191 <strong>in</strong> the human sequence conta<strong>in</strong><strong>in</strong>g important aspartate<br />
residues is suggested to be a potential b<strong>in</strong>d<strong>in</strong>g site for Mn 2+ <strong>in</strong> the GGT active site<br />
(Wang et al. 2002a, Wang et al. 2002b). The aspartate-conta<strong>in</strong><strong>in</strong>g short site may<br />
represent a DxD motif that is commonly found <strong>in</strong> many glycosyltransferases<br />
(Unligil et al. 2000, Gast<strong>in</strong>el et al. 2001, Wang et al. 2002b).<br />
Substrate specificity<br />
The hydroxylys<strong>in</strong>e content <strong>of</strong> collagens varies between collagen types, tissues <strong>and</strong><br />
normal physiological states <strong>of</strong> the tissue but also <strong>in</strong> certa<strong>in</strong> disorders. These<br />
observations have <strong>in</strong>dicated some k<strong>in</strong>d <strong>of</strong> substrate specificity as well as tissue<br />
specificity <strong>of</strong> the LH isoenzymes. (Ihme et al. 1984, Kivirikko et al. 1992, Bank<br />
et al. 1999, Eyre et al. 2002). The m<strong>in</strong>imal sequence requirement for the function<br />
<strong>of</strong> LHs is the X-Lys-Gly triplet (Kivirikko et al. 1992). All is<strong>of</strong>orms have been<br />
shown to hydroxylate synthetic peptides represent<strong>in</strong>g collagenous doma<strong>in</strong>s <strong>of</strong><br />
44
type I <strong>and</strong> IV collagens (Takaluoma et al. 2007) <strong>and</strong> they seem not to have strict<br />
sequence requirements (Risteli et al. 2004, Takaluoma et al. 2007). In addition, all<br />
isoenzymes are able to hydroxylate lys<strong>in</strong>es <strong>in</strong> the triple helical doma<strong>in</strong> <strong>of</strong><br />
recomb<strong>in</strong>ant type I procollagen produced <strong>in</strong> <strong>in</strong>sect cells, but LH2 is the only<br />
is<strong>of</strong>orm capable <strong>of</strong> hydroxylat<strong>in</strong>g the telopeptide lys<strong>in</strong>es <strong>in</strong> this system<br />
(Takaluoma et al. 2007). However, overexpression studies with LH2 <strong>in</strong> fibroblasts<br />
demonstrated that LH2 may not be able to hydroxylate lys<strong>in</strong>es <strong>in</strong> the triple helical<br />
area (Wu et al. 2006). Hydroxylys<strong>in</strong>e residues <strong>in</strong> collagen telopeptides are found<br />
<strong>in</strong> X-Hyl-Ser <strong>and</strong> X-Hyl-Ala sequences, which differ from the collagenous<br />
X-Hyl-Gly sequence (Myllyharju 2005). However, these X-Hyl-Ser <strong>and</strong><br />
X-Hyl-Ala sequences seem not to be the only requirement for the hydroxylation<br />
by LH2 as <strong>in</strong> <strong>in</strong> vitro studies the enzyme was not able to hydroxylate a short<br />
synthetic peptide represent<strong>in</strong>g the type I collagen telopeptide, but rather required<br />
a longer substrate as demonstrated <strong>in</strong> the <strong>in</strong>sect cell coexpression experiments<br />
with a procollagen I cha<strong>in</strong> (Takaluoma et al. 2007).<br />
Despite the lack <strong>of</strong> a strict substrate specificity, several observations suggest<br />
preferential roles for each LH isoenzyme. LH2 hydroxylates the telopeptide<br />
lys<strong>in</strong>es, but may also hydroxylate helical lys<strong>in</strong>es (Uzawa et al. 1999, Mercer et al.<br />
2003, van der Slot et al. 2003, Pornprasertsuk et al. 2004, Takaluoma et al. 2007),<br />
LH1 hydroxylates helical lys<strong>in</strong>es (Eyre et al. 2002, Ste<strong>in</strong>mann et al. 2002) <strong>and</strong><br />
LH3 glycosylates hydroxylys<strong>in</strong>es (Heikk<strong>in</strong>en et al. 2000, Rautavuoma et al. 2002,<br />
Wang et al. 2002a, Rautavuoma et al. 2004, Ruotsala<strong>in</strong>en et al. 2006, Salo et al.<br />
2006a), but its exact role <strong>in</strong> lysyl hydroxylation is not known. LH3 is capable <strong>of</strong><br />
hydroxylat<strong>in</strong>g helical lys<strong>in</strong>es <strong>of</strong> type I procollagen at least <strong>in</strong> a recomb<strong>in</strong>ant<br />
overexpression system (Takaluoma et al. 2007), <strong>and</strong> it is suggested to be<br />
responsible for the hydroxylation <strong>of</strong> lys<strong>in</strong>e residues that are further glycosylated<br />
by the same enzyme (Sipilä et al. 2007). LH1 has been suggested to have some<br />
telopeptide hydroxylation activity (Mercer et al. 2003, Walker et al. 2004a), but<br />
opposite results have also been obta<strong>in</strong>ed (Royce & Barnes 1985, Takaluoma et al.<br />
2007). In addition, analysis <strong>of</strong> EDS VIA patients with null LH1 mutations have<br />
revealed some hydroxylation <strong>of</strong> collagenous doma<strong>in</strong>s <strong>in</strong>dicat<strong>in</strong>g that LH2 or LH3<br />
or both could function as helical LHs as well (Eyre et al. 2002).<br />
LH3 is the only isoenzyme that is reported as likely to be able to modify also<br />
native collagen molecules <strong>in</strong> the extracellular space (Salo et al. 2006a). However,<br />
<strong>in</strong> vitro studies on sk<strong>in</strong> <strong>and</strong> tendon collagens have suggested that the native type I<br />
collagen is not the most appropriate substrate for LH3, which rather prefers<br />
collagen polypeptide cha<strong>in</strong>s (Sricholpech et al. 2011).<br />
45
2.5 Human diseases related to lysyl <strong>hydroxylases</strong> <strong>and</strong> animal<br />
models<br />
The kyphoscoliotic form <strong>of</strong> EDS (EDS VI) was for a long time the only disorder<br />
connected to LHs. Studies <strong>of</strong> the syndrome have provided important <strong>in</strong>formation<br />
about the function <strong>of</strong> LHs <strong>and</strong> the importance <strong>of</strong> lysyl hydroxylation <strong>in</strong> collagens<br />
(Yeowell & Walker 2000, Ste<strong>in</strong>mann et al. 2002). Currently mutations are found<br />
<strong>in</strong> all LH is<strong>of</strong>orms lead<strong>in</strong>g to severe connective tissue disorders (Yeowell &<br />
Walker 2000, van der Slot et al. 2003, Salo et al. 2008). In addition,<br />
overexpression <strong>of</strong> the long form <strong>of</strong> LH2 is connected to fibrosis (van der Slot et al.<br />
2003, van der Slot et al. 2004).<br />
2.5.1 Ehlers-Danlos type VI<br />
The kyphoscoliotic type <strong>of</strong> EDS or EDS VI is an autosomal recessive, <strong>in</strong>heritable<br />
connective tissue disorder. EDS VI patients are cl<strong>in</strong>ically characterized by<br />
articular hypermobility, kyphoscoliosis, hyperextensible sk<strong>in</strong>, bad scarr<strong>in</strong>g, easy<br />
bruis<strong>in</strong>g, <strong>and</strong> muscular hypotonia. (Yeowell & Walker 2000, Ste<strong>in</strong>mann et al.<br />
2002). Patients suffer also from ophthalmological complications <strong>and</strong> have an<br />
<strong>in</strong>creased risk <strong>of</strong> spontaneous arterial ruptures (Wenstrup et al. 1989, Ste<strong>in</strong>mann<br />
et al. 2002). Two different forms <strong>of</strong> EDS VI have been identified. In EDS VIA the<br />
biochemical basis for this disease is a deficiency <strong>of</strong> LH1 activity, caused by<br />
mutations <strong>in</strong> the PLOD1 gene. Therefore, the amount <strong>of</strong> triple helical<br />
hydroxylys<strong>in</strong>e is decreased <strong>and</strong> thus collagen cross-l<strong>in</strong>k<strong>in</strong>g is changed (see Figs. 3<br />
<strong>and</strong> 5). (Yeowell & Walker 2000, Ste<strong>in</strong>mann et al. 2002). EDS VIB patients have<br />
the same phenotype as EDS VIA patients, but the LH activity is normal<br />
(Ste<strong>in</strong>mann et al. 2002, Walker et al. 2004b) <strong>and</strong> currently the molecular<br />
pathology <strong>of</strong> this form is unknown. Moreover, a mutation <strong>in</strong> the z<strong>in</strong>c transporter<br />
gene was recently reported to cause EDS VI-like symptoms with abnormal<br />
collagen hydroxylation <strong>and</strong> ur<strong>in</strong>ary pyrid<strong>in</strong>ol<strong>in</strong>e levels but with normal LH <strong>and</strong><br />
C-P4H activities <strong>in</strong> cell extracts (Giunta et al. 2008).<br />
Genetic background <strong>of</strong> Ehlers-Danlos syndrome type VIA<br />
Various different mutations <strong>in</strong> the PLOD1 gene have been identified from EDS<br />
VIA patients. Mutations consist <strong>of</strong> a s<strong>in</strong>gle base pair (bp) mutation to large<br />
duplications <strong>and</strong> deletions. (Yeowell & Walker 2000, Ste<strong>in</strong>mann et al. 2002). The<br />
46
most common mutation connected to Alu-Alu recomb<strong>in</strong>ation is a large duplication<br />
<strong>of</strong> seven exons (Pousi et al. 1994, Heikk<strong>in</strong>en et al. 1997, Br<strong>in</strong>ckmann et al. 1998,<br />
Yeowell et al. 2000b, Giunta et al. 2005a). Another predom<strong>in</strong>ant form is a<br />
nonsense mutation <strong>in</strong> exon 14 (Yeowell & Walker 1997, Walker et al. 1999,<br />
Yeowell & Walker 1999b, Pousi et al. 2000, Yeowell et al. 2000b). Several<br />
compound heterozygote mutations have been reported as well (Yeowell & Walker<br />
2000). The first found compound heterozygote patient also confirmed the<br />
autosomal recessive <strong>in</strong>heritance <strong>of</strong> the syndrome (Ha et al. 1994). LH1 has<br />
several alternative splic<strong>in</strong>g pathways that can either bypass or compensate for<br />
mutations <strong>and</strong> thus cells may susta<strong>in</strong> partial LH1 activity (Yeowell & Walker<br />
2000). Accord<strong>in</strong>g to one study small amounts <strong>of</strong> differentially spliced LH1<br />
transcripts are also detected from normal cells (Heikk<strong>in</strong>en et al. 1999), although<br />
LH2 is considered to be the only LH is<strong>of</strong>orm hav<strong>in</strong>g splic<strong>in</strong>g variants (Yeowell &<br />
Walker 1999a).<br />
Molecular biology<br />
The LH activity <strong>in</strong> patients with EDS VIA is markedly reduced (Yeowell &<br />
Walker 2000, Ste<strong>in</strong>mann et al. 2002). However, activity levels vary significantly<br />
between tissues, <strong>in</strong>dividuals <strong>and</strong> even between assays performed (Ste<strong>in</strong>mann et al.<br />
2002). There are differences <strong>in</strong> hydroxylation levels also between collagen types<br />
as type I <strong>and</strong> III collagens are severely affected, while collagen types II <strong>and</strong> V are<br />
basically normally hydroxylated <strong>in</strong> EDS VIA patients, generally (Ihme et al. 1984,<br />
Eyre et al. 2002). Although the collagen is normally secreted from cells, the<br />
decreased hydroxylation <strong>and</strong> thus glycosylation <strong>of</strong> collagens can be seen as<br />
<strong>in</strong>creased electrophoretic mobility <strong>in</strong> SDS-PAGE analysis (Jarish et al. 1998,<br />
Walker et al. 1999, Giunta et al. 2005a, Giunta et al. 2005b). In addition, the<br />
solubility rate <strong>of</strong> tissue collagens <strong>in</strong> denatur<strong>in</strong>g solvents is suggested to be related<br />
to the level <strong>of</strong> hydroxylation (Ste<strong>in</strong>mann et al. 2002).<br />
The degree <strong>of</strong> collagen hydroxylation is suggested to correlate with the<br />
cl<strong>in</strong>ical status <strong>of</strong> the EDS VIA patient (Ihme et al. 1984). The sk<strong>in</strong> for <strong>in</strong>stance is<br />
severely affected <strong>in</strong> the patients, while <strong>in</strong> cartilage the consequences are milder<br />
(Eyre & Glimcher 1972, P<strong>in</strong>nell et al. 1972, Eyre et al. 2002). The LH activity<br />
determ<strong>in</strong>ed from sk<strong>in</strong> fibroblasts <strong>of</strong> patients is usually less than 25% <strong>of</strong> controls<br />
(Yeowell & Walker 2000, Ste<strong>in</strong>mann et al. 2002). In EDS VIA patients the lys<strong>in</strong>e<br />
residues <strong>in</strong> collagen telopeptides are normally hydroxylated, while the lys<strong>in</strong>es <strong>in</strong><br />
the triple helical area are underhydroxylated, affect<strong>in</strong>g the collagen cross-l<strong>in</strong>k<strong>in</strong>g<br />
47
<strong>and</strong> thus the characteristics <strong>of</strong> collagenous structures (see Fig. 5) (Ste<strong>in</strong>mann et al.<br />
2002, Eyre & Wu 2005). Although the pyrid<strong>in</strong>ol<strong>in</strong>e levels <strong>in</strong> sk<strong>in</strong> are generally<br />
low, <strong>in</strong> normal sk<strong>in</strong> the mature HP cross-l<strong>in</strong>ks are suggested to be four times more<br />
prevalent than LP cross-l<strong>in</strong>ks, whereas <strong>in</strong> EDS VIA patients HP cross-l<strong>in</strong>ks are<br />
absent (see Figs. 3 <strong>and</strong> 5) (Pasquali et al. 1995, Pasquali et al. 1997, Knott &<br />
Bailey 1998). The changed excretion pattern <strong>of</strong> LP <strong>and</strong> HP cross-l<strong>in</strong>ks derived<br />
from degraded collagens especially from tissues such as bone <strong>in</strong>creases the LP/HP<br />
ratio <strong>in</strong> the ur<strong>in</strong>e, which can be determ<strong>in</strong>ed <strong>and</strong> used as a diagnostic marker<br />
(Pasquali et al. 1994, Pasquali et al. 1995, Ste<strong>in</strong>mann et al. 1995).<br />
In EDS VIB patients the hydroxylation level <strong>of</strong> helical lys<strong>in</strong>es <strong>in</strong> sk<strong>in</strong><br />
fibroblasts is decreased almost as much as <strong>in</strong> EDS VIA (Ste<strong>in</strong>mann et al. 2002,<br />
Uzawa et al. 2003). Interest<strong>in</strong>gly, some <strong>of</strong> the EDS VIB patients are reported to<br />
have decreased LH2 or LH3 mRNA levels (Walker et al. 2004b).<br />
Diagnosis <strong>and</strong> treatment <strong>of</strong> Ehlers-Danlos syndrome type VIA<br />
Dur<strong>in</strong>g childhood EDS VIA patients are usually suspected to have a<br />
neuromuscular disease because <strong>of</strong> the hypotonia <strong>and</strong> delayed gross motor<br />
development, <strong>and</strong> the correct diagnosis is thus <strong>of</strong>ten obta<strong>in</strong>ed late (Ste<strong>in</strong>mann et<br />
al. 2002, Giunta et al. 2005a). In addition to cl<strong>in</strong>ical symptoms, the disease can be<br />
diagnosed by measur<strong>in</strong>g LH activity <strong>in</strong> cultured fibroblasts or LP/HP ratio <strong>in</strong> the<br />
ur<strong>in</strong>e, as <strong>in</strong>dicated earlier (Pasquali et al. 1994, Pasquali et al. 1995, Ste<strong>in</strong>mann et<br />
al. 1995). One study has <strong>in</strong>troduced a method to confirm the diagnosis from RNA<br />
<strong>and</strong> DNA samples (Giunta et al. 2005a) <strong>and</strong> successful prenatal exclusions <strong>of</strong> the<br />
syndrome are also reported (Yeowell & Walker 1999b, Yeowell et al. 2000b).<br />
Although the syndrome is not treatable, some patients have had some benefit<br />
from pharmacological doses <strong>of</strong> ascorbate (vitam<strong>in</strong> C) (Dembure et al. 1987). It<br />
has been shown to <strong>in</strong>crease the collagen synthesis <strong>and</strong> LH activity both <strong>in</strong> normal<br />
<strong>and</strong> EDS VIA fibroblasts (Dembure et al. 1987, Yeowell et al. 1995, Pasquali et<br />
al. 1997, Yeowell et al. 1997). In addition, ascorbate is shown to affect also the<br />
cross-l<strong>in</strong>k content <strong>in</strong> fibroblasts (Pasquali et al. 1997).<br />
48
Normal<br />
EDS VIA<br />
hLys<br />
hHyl<br />
tHyl<br />
tHyl<br />
Fig. 5. Simplified illustration <strong>of</strong> collagen cross-l<strong>in</strong>k<strong>in</strong>g <strong>in</strong> a stiff tissue, such as bone, <strong>in</strong><br />
a normal situation <strong>and</strong> <strong>in</strong> Ehlers-Danlos syndrome type VIA <strong>and</strong> Bruck syndrome.<br />
2.5.2 Bruck syndrome 1 <strong>and</strong> 2<br />
tHyl<br />
tHyl<br />
tLys<br />
tLys<br />
hHyl<br />
Bruck syndrome (BS) is an extremely rare, autosomal recessive connective tissue<br />
disorder with features typical for osteogenesis imperfecta (OI). About 30 patients<br />
have been reported to suffer from BS. BS patients typically have osteoporosis,<br />
fragile bones, congenital contractures <strong>of</strong> jo<strong>in</strong>ts, progressive kyphoscoliosis <strong>and</strong><br />
short stature, but they usually have normal dent<strong>in</strong>ogenesis <strong>and</strong> no hear<strong>in</strong>g loss.<br />
BS can be best dist<strong>in</strong>guished from OI by the presence <strong>of</strong> congenital jo<strong>in</strong>t<br />
contractures <strong>and</strong> the absence <strong>of</strong> hear<strong>in</strong>g loss. (Viljoen et al. 1989, Brenner et al.<br />
1993, McPherson & Clemens 1997, Brady & Patton 1997, Breslau-Siderius et al.<br />
1998, Blacks<strong>in</strong> et al. 1998, Leroy et al. 1998, Bank et al. 1999, van der Slot et al.<br />
2003, Ha-V<strong>in</strong>h et al. 2004, Datta et al. 2005, Berg et al. 2005, Mokete et al. 2005,<br />
Yapicioğlu et al. 2009, Shaheen et al. 2010, Kelley et al. 2011). Typical features<br />
Bruck syndrome<br />
49
<strong>of</strong> OI <strong>in</strong>clude <strong>in</strong>creased bone fragility <strong>and</strong> short stature, impaired hear<strong>in</strong>g,<br />
dent<strong>in</strong>ogenesis imperfecta <strong>and</strong> hypermobile jo<strong>in</strong>ts (for a review, see Basel &<br />
Ste<strong>in</strong>er 2009). Most OI cases are caused by mutations <strong>in</strong> the genes cod<strong>in</strong>g for the<br />
pro-α1 or pro-α2 cha<strong>in</strong>s <strong>of</strong> type I collagen, but mutations are detected also <strong>in</strong><br />
members <strong>of</strong> the prote<strong>in</strong> complex responsible for collagen prolyl 3-hydroxylation<br />
<strong>and</strong> recently also from a SERPINH1 gene that encodes the collagen chaperone<br />
Hsp47 (Basel & Ste<strong>in</strong>er 2009, Christiansen et al. 2010).<br />
Genetic background<br />
BS is a genetically heterogeneous disease. Mutation analysis <strong>of</strong> BS families has<br />
led to two different l<strong>in</strong>kages. BS type 1 (BS1) is l<strong>in</strong>ked to chromosome 17 <strong>and</strong> BS<br />
type 2 (BS2) to the PLOD2 gene <strong>in</strong> chromosome 3. (Bank et al. 1999, van der<br />
Slot et al. 2003, Ha-V<strong>in</strong>h et al. 2004). Three different po<strong>in</strong>t mutations have been<br />
found so far from exon 17 <strong>of</strong> the PLOD2 gene that encodes LH2 (van der Slot et<br />
al. 2003, Ha-V<strong>in</strong>h et al. 2004), while the genetic background <strong>and</strong> the molecular<br />
mechanism <strong>of</strong> BS1 have been elusive for a long time. Recently, mutations <strong>in</strong><br />
FKBP10, a gene encod<strong>in</strong>g FK506 b<strong>in</strong>d<strong>in</strong>g prote<strong>in</strong> 10 (FKBP10, also known as<br />
FKBP65) were suggested to cause BS1 or a subtype <strong>of</strong> OI (Alanay et al. 2010,<br />
Alanay & Krakow, 2010, Shaheen et al. 2010, Kelley et al. 2011). The FKBP10<br />
gene is located <strong>in</strong> chromosome 17, but not at the location suggested earlier to<br />
cause BS1 (Bank et al. 1999, Kelley et al. 2011).<br />
Molecular biology<br />
BS subtypes can be dist<strong>in</strong>guished accord<strong>in</strong>g to the genetic defect (Bank et al.<br />
1999, van der Slot et al. 2003). BS2 is caused by a defective telopeptide LH,<br />
LH2(long) (Bank et al. 1999, van der Slot et al. 2003, Ha-V<strong>in</strong>h et al. 2004,<br />
Takaluoma et al. 2007), while the molecular basis <strong>of</strong> BS1 is still unknown. In BS<br />
patients the lys<strong>in</strong>e residues <strong>in</strong> the triple helical area <strong>of</strong> type I collagen are<br />
normally hydroxylated, while the telopeptide lys<strong>in</strong>es are underhydroxylated (see<br />
Fig. 5) (Bank et al. 1999). The hydroxylation state <strong>of</strong> the telopeptide lys<strong>in</strong>es<br />
affects the cross-l<strong>in</strong>k<strong>in</strong>g pr<strong>of</strong>ile <strong>of</strong> collagen molecules (Eyre & Wu 2005). In the<br />
bone <strong>of</strong> the patients, for <strong>in</strong>stance, the amount <strong>of</strong> telopeptide hydroxylys<strong>in</strong>ederived<br />
mature pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks, HP <strong>and</strong> LP, <strong>in</strong> type I collagen is reduced<br />
(see Figs. 3 <strong>and</strong> 5). In contrast, the amount <strong>of</strong> lys<strong>in</strong>onorleuc<strong>in</strong>e, a derivative <strong>of</strong> an<br />
immature, difunctional cross-l<strong>in</strong>k orig<strong>in</strong>ated from unhydroxylated telopeptide<br />
50
lys<strong>in</strong>es via the allys<strong>in</strong>e route, is reported to be <strong>in</strong>creased 5-fold. Interest<strong>in</strong>gly,<br />
cross-l<strong>in</strong>k<strong>in</strong>g <strong>of</strong> type I collagen has tissue-specific features as <strong>in</strong> contrast to the<br />
bone, the type I collagen <strong>in</strong> the ligament <strong>of</strong> a BS1 patient is normally cross-l<strong>in</strong>ked<br />
<strong>and</strong> thus hydroxylated. (Bank et al. 1999). The decreased number <strong>of</strong> stable<br />
pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks <strong>and</strong> <strong>in</strong>creased collagen turnover <strong>in</strong> bone can be analyzed<br />
from ur<strong>in</strong>e samples (Ha-V<strong>in</strong>h et al. 2004).<br />
Morphological analyses <strong>of</strong> severely affected bone tissue <strong>of</strong> BS patients have<br />
revealed wormian bones <strong>in</strong> the skull, cystic changes at old fracture sites <strong>and</strong> a<br />
mixed callus as well as normal or immature m<strong>in</strong>eralization <strong>in</strong> bone <strong>and</strong> osteoids<br />
that vary <strong>in</strong> thickness (Viljoen et al. 1989, Brenner et al. 1993, McPherson &<br />
Clemens 1997, Brady & Patton 1997, Blacks<strong>in</strong> et al. 1998, Leroy et al. 1998,<br />
Ha-V<strong>in</strong>h et al. 2004, Datta et al. 2005, Mokete et al. 2005). At the cell level<br />
osteoblasts have swollen mitochondria <strong>and</strong> dilated ER <strong>and</strong> the diameter <strong>of</strong><br />
collagen fibrils <strong>in</strong> osteoids vary (Brenner et al. 1993). The cross-l<strong>in</strong>k<strong>in</strong>g pattern is<br />
suggested to have an important role <strong>in</strong> the correct m<strong>in</strong>eralization <strong>of</strong> calcified<br />
tissues (Yamauchi & Katz 1993, Knott et al. 1997, Bank et al. 1999,<br />
Pornprasertsuk et al. 2005). Unsurpris<strong>in</strong>gly, the m<strong>in</strong>eral content <strong>in</strong> BS bone is<br />
shown to be reduced <strong>and</strong> the size <strong>of</strong> hydroxyapatite crystals is <strong>in</strong>creased (Brenner<br />
et al. 1993). Thus, it is suggested that the fragility <strong>of</strong> BS bones is likely to be due<br />
to a changed cross-l<strong>in</strong>k<strong>in</strong>g pattern together with <strong>in</strong>adequate m<strong>in</strong>eralization (Bank<br />
et al. 1999). Biochemical studies show also <strong>in</strong>creased extractability <strong>of</strong> type I<br />
collagen <strong>in</strong> BS bone tissue (Brenner et al. 1993). However, the amounts <strong>of</strong> type I<br />
<strong>and</strong> III collagens are reported to be normal <strong>and</strong> also <strong>their</strong> structure <strong>and</strong> synthesis<br />
are seem<strong>in</strong>gly normal (Brenner et al. 1993, McPherson & Clemens 1997,<br />
Breslau-Siderius et al. 1998, Leroy et al. 1998, Bank et al. 1999, Kelley et al.<br />
2011), which differs from OI where collagens are usually overmodified due to a<br />
delay <strong>in</strong> the triple helix formation (Engel & Prockop 1991, Raghunath et al. 1994).<br />
The gene FKBP10 <strong>and</strong> Bruck syndrome<br />
The mutations <strong>in</strong> the FKBP10 gene, connected to BS1, <strong>in</strong>clude a homozygous 1<br />
bp duplication, a 35 bp deletion <strong>and</strong> two different compound heterozygous<br />
mutations, all <strong>of</strong> them located <strong>in</strong> the cod<strong>in</strong>g region <strong>of</strong> FKBP10, while no<br />
mutations have been found from the PLOD2 gene <strong>in</strong> these patients (Kelley et al.<br />
2011). FKBP10 is a member <strong>of</strong> immunophil<strong>in</strong>s that are known to function as<br />
<strong>in</strong>tracellular receptors capable <strong>of</strong> b<strong>in</strong>d<strong>in</strong>g <strong>and</strong> mediat<strong>in</strong>g the action <strong>of</strong> certa<strong>in</strong><br />
immunosuppressant drugs (Coss et al. 1995; for a review, see Göthel & Marahiel<br />
51
1999). Like other immunophil<strong>in</strong>s, FKBP10 is suggested to have some<br />
peptidylprolyl cis-trans isomerase activity, which can be important <strong>in</strong> the fold<strong>in</strong>g<br />
<strong>of</strong> collagens <strong>and</strong> tropoelast<strong>in</strong> monomers (Coss et al. 1995, Davis et al. 1998,<br />
Ishikawa et al. 2008). Interest<strong>in</strong>gly, mutations <strong>in</strong> another FK506 b<strong>in</strong>d<strong>in</strong>g prote<strong>in</strong>,<br />
FKBP14 (FK506 b<strong>in</strong>d<strong>in</strong>g prote<strong>in</strong> 14) are reported to cause a new variant <strong>of</strong> EDS,<br />
a disorder where patients share many cl<strong>in</strong>ical features particularly <strong>of</strong> EDS VIA<br />
<strong>and</strong> Ullrich congenital muscular dystrophy. In contrast to EDS VIA, these patients<br />
have normal ur<strong>in</strong>ary LP/HP ratio. (Baumann et al. 2012).<br />
Like LHs, FKBP10 is located <strong>in</strong> the lumen <strong>of</strong> the ER <strong>and</strong> it acts as a<br />
chaperone <strong>of</strong> folded <strong>and</strong> unfolded collagen (Myllylä et al. 1991, Kellokumpu et al.<br />
1994, Davis et al. 1998, Patterson et al. 2000, Ishikawa et al. 2008). It is also<br />
suggested to have a role <strong>in</strong> the chaperon<strong>in</strong>g <strong>of</strong> tropoelast<strong>in</strong> (Davis et al. 1998,<br />
Patterson et al. 2000). Studies with FKBP10 <strong>and</strong> collagen together have<br />
suggested that FKBP10 could function as a complex with other prote<strong>in</strong>s to<br />
process collagen (Zeng et al. 1998, Ishikawa et al. 2008). Collagens from patients<br />
with a mutation <strong>in</strong> FKBP10 are not overmodified (Alanay et al. 2010). This was<br />
observed also with an OI patient who had a mutation <strong>in</strong> SERPINH1 (Christiansen<br />
et al. 2010), when typically <strong>in</strong> OI the overmodification is quite common (Mar<strong>in</strong>i<br />
et al. 2010). In contrast to the other FKBP family members that are expressed<br />
ubiquitously, FKBP10 mRNA is detected <strong>in</strong> the mouse lung, spleen, heart, bra<strong>in</strong><br />
<strong>and</strong> testis (Coss et al. 1995). Another study shows that the expression <strong>of</strong> FKBP10<br />
<strong>in</strong> adult mice is absent or m<strong>in</strong>imal, but <strong>in</strong> develop<strong>in</strong>g mice the expression was<br />
seen <strong>in</strong> the smooth muscle layer <strong>of</strong> the airways <strong>and</strong> pulmonary vessels <strong>and</strong> at the<br />
mRNA level <strong>in</strong> tissues such as aorta <strong>and</strong> kidney, <strong>in</strong>dicat<strong>in</strong>g developmental<br />
regulation <strong>of</strong> the prote<strong>in</strong> (Patterson et al. 2000). However, the possible connection<br />
between FKBP10 <strong>and</strong> LH2 is still unknown.<br />
Treatment <strong>of</strong> Bruck syndrome<br />
There is no treatment for BS. However, bisphosphonate cyclic pamidronate has<br />
been used as a pharmacotherapy aga<strong>in</strong>st bone fragility (Ha-V<strong>in</strong>h et al. 2004,<br />
Mokete et al. 2005, Andiran et al. 2008, Yapicioğlu et al. 2009, Shaheen et al.<br />
2010). Bisphosphonates are suggested to function as <strong>in</strong>hibitors <strong>of</strong> bone resorption<br />
by <strong>in</strong>terfer<strong>in</strong>g with the function <strong>of</strong> osteoclasts (for a review, see Russell 2006). In<br />
a reported BS patient the treatment affected fracture <strong>in</strong>cidence <strong>and</strong> bone m<strong>in</strong>eral<br />
density (Andiran et al. 2008). However, a long-term follow-up study <strong>of</strong> a BS<br />
52
patient, not receive<strong>in</strong>g treatment, revealed that the prevalence <strong>of</strong> fractures<br />
decreases after maturation <strong>of</strong> the bone (Mokete et al. 2005).<br />
2.5.3 Fibrotic diseases <strong>and</strong> cancer<br />
Fibrotic diseases<br />
Inflammatory reaction <strong>in</strong> the tissue is a normal response to damage caused by<br />
numerous factors from bacterial <strong>in</strong>fections to mechanical damage. The<br />
programmed phases <strong>of</strong> <strong>in</strong>flammation protect the tissue <strong>and</strong> enable recovery.<br />
Sometimes impaired regulation <strong>of</strong> the <strong>in</strong>flammatory reaction caus<strong>in</strong>g a chronic<br />
<strong>in</strong>flammation can lead to fibrosis, which is characterized by excessive<br />
accumulation <strong>of</strong> ECM <strong>and</strong> especially collagen. (Rub<strong>in</strong> & Farber 1999,<br />
Trojanowska et al. 1998, Hunzelmann & Br<strong>in</strong>ckmann 2011). In s<strong>of</strong>t tissues<br />
collagen cross-l<strong>in</strong>ks are normally formed from unhydroxylated telopeptide lys<strong>in</strong>es,<br />
but <strong>in</strong> fibrotic lesions the cross-l<strong>in</strong>k<strong>in</strong>g pattern <strong>of</strong> collagen is altered (Br<strong>in</strong>ckmann<br />
et al. 1996, Knott & Bailey 1998, Br<strong>in</strong>ckmann et al. 1999, Br<strong>in</strong>ckmann et al.<br />
2001). The <strong>in</strong>creased amount <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks (see Fig. 3) found <strong>in</strong><br />
fibrotic tissue is caused by overhydroxylation <strong>of</strong> the telopeptide lys<strong>in</strong>e residues<br />
accompanied by the LO reaction (van der Slot et al. 2003, van der Slot et al. 2004,<br />
Levental et al. 2009, Saito & Marumo 2010). The result<strong>in</strong>g collagenous structures<br />
are highly resistant to normal remodell<strong>in</strong>g <strong>and</strong> degenerative procedures such as<br />
degradation by MMPs (Trojanowska et al. 1998). Collagen matrices hav<strong>in</strong>g an<br />
<strong>in</strong>creased amount <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks are shown to be less susceptible to<br />
MMP-1 degradation, for <strong>in</strong>stance (van der Slot-Verhoeven et al. 2005). As<br />
LH2(long) is the only LH is<strong>of</strong>orm significantly upregulated <strong>in</strong> fibrosis <strong>and</strong> it is<br />
known to be a telopeptide LH, it has been suggested to be responsible for<br />
overhydroxylation <strong>of</strong> the telopeptide lys<strong>in</strong>es <strong>and</strong> thus <strong>in</strong>creased pyrid<strong>in</strong>ol<strong>in</strong>e<br />
cross-l<strong>in</strong>k formation (van der Slot et al. 2003, van der Slot et al. 2004, Takaluoma<br />
et al. 2007). Although overexpression <strong>of</strong> all LH is<strong>of</strong>orms is shown to <strong>in</strong>crease<br />
collagen synthesis (Mercer et al. 2003), overexpression <strong>of</strong> only LH2 was shown<br />
to both <strong>in</strong>crease the collagen synthesis <strong>and</strong> change the cross-l<strong>in</strong>k<strong>in</strong>g pattern (Wu<br />
et al. 2006). Thus, LH2(long) is an important factor <strong>in</strong> fibrosis <strong>and</strong> prevention <strong>of</strong><br />
its upregulation together with LO is important from the cl<strong>in</strong>ical po<strong>in</strong>t <strong>of</strong> view. The<br />
role <strong>of</strong> LH2(short) seems to be <strong>in</strong>significant <strong>in</strong> fibrotic conditions (van der Slot et<br />
al. 2003, van der Slot et al. 2004). Although fibrosis <strong>and</strong> <strong>in</strong>creased pyrid<strong>in</strong>ol<strong>in</strong>e<br />
53
cross-l<strong>in</strong>ks are connected, direct connection <strong>of</strong> this phenomenon to the resistance<br />
<strong>of</strong> collagenous structures to remodel<strong>in</strong>g is not confirmed.<br />
As a consequence <strong>of</strong> the <strong>in</strong>flammatory response the fibrotic process is<br />
connected to many <strong>in</strong>flammatory cytok<strong>in</strong>es <strong>and</strong> growth factors (Trojanowska et al.<br />
1998, Hunzelmann & Br<strong>in</strong>ckmann 2011). TGF-β is an important <strong>and</strong> common<br />
<strong>in</strong>flammatory factor <strong>and</strong> it is shown to <strong>in</strong>crease the expression <strong>of</strong> LH2(long) <strong>and</strong><br />
thus the formation <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks (Trojanowska et al. 1998, van der<br />
Slot et al. 2005, Hunzelmann & Br<strong>in</strong>ckmann 2011). In addition, TGF-β is known<br />
to <strong>in</strong>crease generally the expression <strong>of</strong> ECM prote<strong>in</strong>s <strong>and</strong> <strong>in</strong>hibitors <strong>of</strong> MMPs<br />
(Hunzelmann & Br<strong>in</strong>ckmann 2011). Besides TGF-β, other cytok<strong>in</strong>es such as<br />
<strong>in</strong>terleuk<strong>in</strong> 4 <strong>and</strong> activ<strong>in</strong> A are also shown to <strong>in</strong>duce LH2(long) expression (van<br />
der Slot et al. 2005, Br<strong>in</strong>ckmann et al. 2005). However, the effect <strong>of</strong> at least TGFβ<br />
is presumably not direct, but it is mediated by other pr<strong>of</strong>ibrotic factors that are<br />
rapidly upregulated via TGF-β <strong>in</strong>duction (van der Slot et al. 2005). Thus,<br />
knowledge about factors contribut<strong>in</strong>g to <strong>in</strong>creased expression <strong>of</strong> the LH2(long)<br />
form, <strong>in</strong>clud<strong>in</strong>g knowledge about the <strong>in</strong>flammatory response <strong>and</strong> alternative<br />
splic<strong>in</strong>g, are important <strong>in</strong> the prevention <strong>of</strong> fibrosis.<br />
Treatment <strong>of</strong> fibrosis<br />
Certa<strong>in</strong> organophosphate compounds <strong>in</strong>hibit the function <strong>of</strong> LH <strong>and</strong> have been<br />
earlier suggested to be valuable for test<strong>in</strong>g as antifibrotic drugs (Samimi & Last<br />
2001). An antihypertension drug m<strong>in</strong>oxidil has been suggested to be a potential<br />
drug for fibrosis as it decreases LH activity (Murad & P<strong>in</strong>nell 1987, Hautala et al.<br />
1992b). However, m<strong>in</strong>oxidil has consequently been found to be <strong>in</strong>effective as it<br />
decreases LH1 expression rather than LH2 or LH3 expression, <strong>and</strong> thus changes<br />
the LP/HP ratio but does not decrease the total amount <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks<br />
(Zuurmond et al. 2005).<br />
Fibrosis, hypoxia <strong>and</strong> cancer<br />
In the chronic phase <strong>of</strong> <strong>in</strong>flammation, fibrosis is accompanied by hypoxia (Rub<strong>in</strong><br />
& Farber 1999, Hunzelmann & Br<strong>in</strong>ckmann 2011). In hypoxic conditions C-P4H,<br />
LH1 <strong>and</strong> LH2, but not LH3, have been shown to be upregulated lead<strong>in</strong>g to<br />
<strong>in</strong>creased accumulation <strong>of</strong> collagen (Lal et al. 2001, Denko et al. 2003, H<strong>of</strong>bauer<br />
et al. 2003, Scheurer et al. 2004, Br<strong>in</strong>ckmann et al. 2005, van Vlimmeren et al.<br />
2010), while the expression <strong>of</strong> collagen polypeptides themselves has been<br />
54
eported to be either upregulated, downregulated or not affected (H<strong>of</strong>bauer et al.<br />
2003, Scheurer et al. 2004, Br<strong>in</strong>ckmann et al. 2005, van Vlimmeren et al. 2010).<br />
Hypoxia is characteristic also <strong>in</strong> cancer, <strong>and</strong> ECM remodel<strong>in</strong>g <strong>and</strong> changes<br />
<strong>in</strong> tissue stiffness are also characteristic <strong>in</strong> tumor formation (for reviews, see<br />
Butcher et al. 2009, Semenza 2011). Moreover, a fibrotic tumor predicts<br />
malignancy (Colpaert et al. 2003, Levental et al. 2009). Hypoxia-<strong>in</strong>ducible factor<br />
1 (HIF-1) is a transcription factor that has been reported to <strong>in</strong>fluence cancer<br />
progression (Semenza 2011). Mouse <strong>and</strong> human PLOD1 <strong>and</strong> PLOD2 genes have<br />
been shown to have a putative hypoxia responsive element <strong>in</strong> <strong>their</strong> promoter<br />
region suggest<strong>in</strong>g that the regulation <strong>of</strong> these genes may be mediated by HIF<br />
(H<strong>of</strong>bauer et al. 2003, Scheurer et al. 2004, Br<strong>in</strong>ckmann et al. 2005). In fact,<br />
PLOD2 expression is elevated <strong>in</strong> cancer cells (Lal et al. 2001, Dong et al. 2005,<br />
Evens et al. 2010, Rajkumar et al. 2011, Noda et al. 2012), <strong>and</strong> the elevation is<br />
reported to be associated with <strong>in</strong>creased HIF-1α gene expression (Evens et al.<br />
2010). Upregulation <strong>of</strong> several collagen-modify<strong>in</strong>g enzymes <strong>in</strong>clud<strong>in</strong>g LO <strong>in</strong><br />
hypoxia could <strong>in</strong>dicate <strong>their</strong> important role <strong>in</strong> the ECM remodel<strong>in</strong>g dur<strong>in</strong>g early<br />
phases <strong>of</strong> tumor angiogenesis, for <strong>in</strong>stance (Denko et al. 2003, Scheurer et al.<br />
2004, Levental et al. 2009). HIF-1α-<strong>in</strong>duced upregulation <strong>of</strong> PLOD2 <strong>in</strong> cancer is<br />
suggested to be associated with poor prognosis (Dong et al. 2005).<br />
2.5.4 A disease caused by mutations <strong>in</strong> LH3 <strong>and</strong> LH3 null mice<br />
LH3 was for a long time the only LH is<strong>of</strong>orm that was not connected to any<br />
human diseases, although mice hav<strong>in</strong>g an <strong>in</strong>activated Plod3 gene were known to<br />
die dur<strong>in</strong>g embryogenesis (Rautavuoma et al. 2004, Ruotsala<strong>in</strong>en et al. 2006).<br />
Recently, however, LH3 mutations were connected to two different connective<br />
tissue disorders. A patient with severe symptoms overlapp<strong>in</strong>g several connective<br />
tissue disorders was shown to have mutations <strong>in</strong> the PLOD3 gene (Salo et al.<br />
2008). In addition, one epidermolysis bullosa family was connected to PLOD3<br />
mutations (Savola<strong>in</strong>en et al. 1981, Risteli et al. 2009).<br />
Human disorders connected to LH3<br />
As <strong>in</strong>dicated earlier, some connective tissue disorders can be detected based on<br />
the collagen turnover products <strong>in</strong> ur<strong>in</strong>e or serum samples (Knott & Bailey 1998,<br />
Eyre & Wu 2005, Saito & Marumo 2010). The first patient with a defective LH3<br />
enzyme was also detected via cross-l<strong>in</strong>k analysis from ur<strong>in</strong>e. A glycosylated<br />
55
derivate <strong>of</strong> the pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>k was absent <strong>in</strong> the sample <strong>and</strong> the amount <strong>of</strong><br />
glucosylgalactosyl hydroxylys<strong>in</strong>e was decreased. (Salo et al. 2008). As LH3 is<br />
known to have collagen glycosylation activity (Heikk<strong>in</strong>en et al. 2000,<br />
Rautavuoma et al. 2002, Wang et al. 2002a, Rautavuoma et al. 2004,<br />
Ruotsala<strong>in</strong>en et al. 2006, Salo et al. 2006a) besides the two GT25 family<br />
members that possess collagen GT activity (Schegg et al. 2009), the PLOD3 gene<br />
<strong>of</strong> the patient was analyzed <strong>and</strong> found to have a compound heterozygote mutation<br />
(Salo et al. 2008). One <strong>of</strong> the mutations was located <strong>in</strong> the region responsible for<br />
the glycosylation activities <strong>and</strong> the other <strong>in</strong> the LH activity region (Pirskanen et al.<br />
1996, Heikk<strong>in</strong>en et al. 2000, Rautavuoma et al. 2002, Wang et al. 2002a, Wang et<br />
al. 2002b, Salo et al. 2008). However, despite the locations <strong>of</strong> the mutations, both<br />
reduced all three activities i.e. LH, GT <strong>and</strong> GGT. The mutations lead to a<br />
decreased concentration <strong>of</strong> the LH3 prote<strong>in</strong> <strong>and</strong> reduced LH <strong>and</strong> glycosylation<br />
activities, which are suggested to be the molecular basis for the disorder by<br />
affect<strong>in</strong>g collagen synthesis. (Salo et al. 2008). Based on these f<strong>in</strong>d<strong>in</strong>gs <strong>and</strong> the<br />
fact that LH3 is known to be essential for the development <strong>of</strong> the mouse<br />
(Rautavuoma et al. 2004, Ruotsala<strong>in</strong>en et al. 2006), it is not surpris<strong>in</strong>g that the<br />
patient has severe symptoms such as osteopenia, jo<strong>in</strong>t contractures, arterial<br />
ruptures, sk<strong>in</strong> blister<strong>in</strong>g, deafness <strong>and</strong> myopia that are all connected to many<br />
different collagen disorders (Salo et al. 2008). In addition, the reported patient<br />
had a stillborn sibl<strong>in</strong>g who was reported to have several abnormalities. Accord<strong>in</strong>g<br />
to autopsy, the <strong>in</strong>trauter<strong>in</strong>e death at 28 weeks <strong>of</strong> gestation was suggested to be<br />
caused by cerebral vascular rupture. (Salo et al. 2008).<br />
In addition to the LH3 compound heterozygous patient, a large family<br />
diagnosed to have an epidermolysis bullosa simplex was discovered to transcribe<br />
only one LH3 allele (Savola<strong>in</strong>en et al. 1981, Risteli et al. 2009). The patients<br />
were reported to develop blisters extremely easily <strong>and</strong> blisters were frequently<br />
observed particularly <strong>in</strong> the h<strong>and</strong>s <strong>and</strong> feet, which are regularly under mechanical<br />
stress. The heal<strong>in</strong>g occurred without scarr<strong>in</strong>g. The first blisters were detected at<br />
birth or shortly after that, but the symptoms usually became milder with age.<br />
(Savola<strong>in</strong>en et al. 1981). The decreased amount <strong>of</strong> LH3 <strong>in</strong> these patients led to<br />
decreased glycosylation activity, which <strong>in</strong> turn was suggested to affect the<br />
organization <strong>of</strong> the ECM (Savola<strong>in</strong>en et al. 1981, Risteli et al. 2009). Four<br />
polymorphic changes <strong>in</strong> the non-cod<strong>in</strong>g region <strong>of</strong> PLOD3 were detected, but the<br />
reason for the allelic silenc<strong>in</strong>g rema<strong>in</strong>ed unknown. However, as a heterozygous<br />
deficiency it demonstrated the importance <strong>of</strong> the correct amount <strong>of</strong> LH3 activity<br />
<strong>in</strong> humans. (Risteli et al. 2009).<br />
56
LH3 knock-out mice<br />
The currently unnamed, severe ECM disorder caused by LH3 mutations, reveals<br />
the importance <strong>of</strong> this enzyme for humans, but LH3 has been demonstrated to be<br />
essential also <strong>in</strong> the mouse (Rautavuoma et al. 2004, Ruotsala<strong>in</strong>en et al. 2006).<br />
Mouse embryos hav<strong>in</strong>g an <strong>in</strong>activated Plod3 gene die dur<strong>in</strong>g early development<br />
at E9.5 (Rautavuoma et al. 2004, Ruotsala<strong>in</strong>en et al. 2006). The deficiency <strong>of</strong><br />
LH3 was found to be critical for the synthesis <strong>of</strong> type IV collagen <strong>and</strong> thus for the<br />
stability <strong>of</strong> basement membranes (Rautavuoma et al. 2004, Ruotsala<strong>in</strong>en et al.<br />
2006). In addition to the defect <strong>in</strong> the embryos, a possible role <strong>of</strong> abnormal<br />
function <strong>of</strong> placental vasculature or the Reichert’s membrane (RM) <strong>in</strong> the<br />
developmental retardation was not excluded (Rautavuoma et al. 2004,<br />
Ruotsala<strong>in</strong>en et al. 2006).<br />
The embryonic lethality <strong>of</strong> LH3 knock-out mice was determ<strong>in</strong>ed to be caused<br />
by the deficient glycosylation activity rather than LH activity <strong>of</strong> the LH3 enzyme<br />
(Rautavuoma et al. 2004, Ruotsala<strong>in</strong>en et al. 2006). At least the GGT activity <strong>in</strong><br />
mouse LH3 is critical for mouse development, <strong>in</strong>dicat<strong>in</strong>g that LH3 may be the<br />
ma<strong>in</strong> molecule responsible for GGT activity <strong>in</strong> mouse (Ruotsala<strong>in</strong>en et al. 2006).<br />
Studies with LH3 knock-out embryos <strong>and</strong> fibroblasts suggested that the deficient<br />
glycosylation <strong>of</strong> type IV collagen leads to delayed process<strong>in</strong>g, <strong>in</strong>tracellular<br />
accumulation, <strong>and</strong> delayed secretion as well as <strong>in</strong>creased <strong>in</strong>tracellular degradation<br />
<strong>of</strong> type IV collagen (Rautavuoma et al. 2004, Sipilä et al. 2007). In addition, the<br />
lack <strong>of</strong> glycosylation activity but not LH activity <strong>of</strong> LH3 seemed to prevent the<br />
assembly <strong>and</strong> thus secretion <strong>of</strong> type VI collagen. Although the lack <strong>of</strong> LH activity<br />
<strong>of</strong> LH3 does not prevent the secretion <strong>of</strong> type VI collagen, it affects the<br />
distribution <strong>of</strong> type IV <strong>and</strong> VI collagens <strong>in</strong> the exam<strong>in</strong>ed adult mouse tissues.<br />
(Sipilä et al. 2007). Recent studies with cells have revealed that the impaired<br />
glycosylation activity <strong>of</strong> LH3 affects fibrillogenesis <strong>of</strong> the bone type I collagen as<br />
well. Furthermore, GGT activity is suggested to be the ma<strong>in</strong> function <strong>of</strong> LH3 <strong>in</strong><br />
osteoblasts. (Sricholpech et al. 2011).<br />
2.5.5 Inactivation <strong>of</strong> LH <strong>in</strong> other animal models<br />
The nematode Caenorhabditis elegans <strong>and</strong> the fly Drosophila melanogaster have<br />
only a s<strong>in</strong>gle LH, while the zebrafish Danio rerio, like other vertebrates, has all<br />
three forms <strong>in</strong>clud<strong>in</strong>g both LH2 splic<strong>in</strong>g variants (Norman & Moerman 2000,<br />
Schneider & Granato 2006, Schneider & Granato 2007, Bunt et al. 2010). The<br />
57
s<strong>in</strong>gle C. elegans LH possesses both lysyl hydroxylase <strong>and</strong> collagen<br />
glucosyltransferase activity <strong>and</strong> thus resembles the vertebrate LH3 (Wang et al.<br />
2002a, Wang et al. 2002b). The D. melanogaster LH also conta<strong>in</strong>s the residues<br />
that have been shown to be necessary for the glycosyltransferase activity <strong>of</strong><br />
vertebrate LH3 (Bunt et al. 2010).<br />
In C. elegans <strong>and</strong> D. melanogaster the absence <strong>of</strong> the only LH leads to<br />
lethality dur<strong>in</strong>g embryogenesis (Norman & Moerman 2000, Bunt et al. 2010). As<br />
<strong>in</strong> LH3 knock-out mice (Rautavuoma et al. 2004, Sipilä et al. 2007), the enzyme<br />
is required <strong>in</strong> the process<strong>in</strong>g <strong>and</strong> secretion <strong>of</strong> type IV collagen <strong>in</strong> these organisms<br />
(Norman & Moerman 2000, Bunt et al. 2010). Similarly, LH3 is essential for<br />
collagen secretion <strong>in</strong> D. rerio (Schneider & Granato 2006).<br />
58
3 Outl<strong>in</strong>es <strong>of</strong> the present study<br />
Collagens are the ma<strong>in</strong> structural components <strong>of</strong> the ECM <strong>in</strong> the human body <strong>and</strong><br />
LH isoenzymes have an important role <strong>in</strong> <strong>their</strong> synthesis. Mutations <strong>in</strong> the<br />
PLOD1 <strong>and</strong> PLOD2 genes are known to cause the rare <strong>and</strong> severe connective<br />
tissue disorders EDS VIA <strong>and</strong> BS2 <strong>in</strong> humans, respectively, emphasiz<strong>in</strong>g the<br />
importance <strong>of</strong> these enzymes <strong>in</strong> a healthy <strong>in</strong>dividual. Moreover, earlier studies on<br />
the lack <strong>of</strong> LH3 <strong>in</strong> mouse had revealed novel <strong>in</strong>formation <strong>of</strong> its importance dur<strong>in</strong>g<br />
the embryonic development. Later, LH3 mutations have also been connected to<br />
human disorders. The Plod3 knock-out mouse l<strong>in</strong>e demonstrated that LH1 <strong>and</strong><br />
LH2 are not able to compensate for the lack <strong>of</strong> LH3 activity, but otherwise the<br />
exact <strong>in</strong> <strong>vivo</strong> roles <strong>of</strong> LH1 <strong>and</strong> LH2 have rema<strong>in</strong>ed mostly elusive. Thus, we<br />
wanted to generate Plod1 <strong>and</strong> Plod2 knock-out mouse l<strong>in</strong>es to study the<br />
expression patterns <strong>and</strong> <strong>in</strong> <strong>vivo</strong> roles <strong>of</strong> these isoenzymes, as well as potential<br />
compensatory roles <strong>of</strong> the other LH isoenzymes. Moreover, the mouse l<strong>in</strong>es were<br />
also expected to serve as the first animal models for EDS VIA <strong>and</strong> BS, enabl<strong>in</strong>g<br />
the study <strong>of</strong> tissues that were impossible <strong>and</strong> unethical to study from the human<br />
patients.<br />
The first BS mutation was l<strong>in</strong>ked to chromosome 17, but the pathological<br />
mechanism rema<strong>in</strong>ed unknown. Later, three different mutations were found <strong>in</strong> the<br />
PLOD2 gene located <strong>in</strong> chromosome 3. The BS form connected to LH2 mutations<br />
was thereafter called BS2. Although studies <strong>of</strong> BS2 patients provided cl<strong>in</strong>ical<br />
support for the theory that LH2 is the telopeptide LH, the mutations did not<br />
change any <strong>of</strong> the catalytically critical am<strong>in</strong>o acids <strong>of</strong> the enzyme <strong>and</strong> the<br />
molecular mechanism <strong>of</strong> the disorder rema<strong>in</strong>ed unknown. We therefore decided to<br />
produce the BS2 mutant <strong>and</strong> wild-type human LH2(long) polypeptides <strong>in</strong> <strong>in</strong>sect<br />
cells, purify <strong>and</strong> analyze them to underst<strong>and</strong> the molecular pathology <strong>of</strong> the BS2<br />
syndrome <strong>in</strong> more detail.<br />
59
4 Materials <strong>and</strong> methods<br />
The materials <strong>and</strong> methods used <strong>in</strong> this thesis are summarized <strong>in</strong> the table below.<br />
Detailed <strong>in</strong>formation with references can be found <strong>in</strong> the orig<strong>in</strong>al articles I-III.<br />
Table 2. Methods.<br />
Level Method Used <strong>in</strong><br />
DNA<br />
Clon<strong>in</strong>g techniques<br />
I<br />
PCR<br />
I, II, III<br />
Southern Blot analysis<br />
I, III<br />
Site-directed mutagenesis<br />
II<br />
RNA<br />
RNA isolation<br />
I, III<br />
RT-PCR<br />
I, III<br />
Quantitative real-time PCR<br />
III<br />
Prote<strong>in</strong> Expression <strong>of</strong> recomb<strong>in</strong>ant prote<strong>in</strong>s <strong>in</strong> <strong>in</strong>sect cells<br />
II<br />
Purification <strong>and</strong> characterization <strong>of</strong> recomb<strong>in</strong>ant prote<strong>in</strong>s II<br />
SDS-PAGE, non-denatur<strong>in</strong>g PAGE <strong>and</strong> Western blott<strong>in</strong>g<br />
LH activity assay us<strong>in</strong>g 2-oxo[1- 14 C]glutarate<br />
LH activity assay with [ 14 II<br />
II<br />
C]lys<strong>in</strong>e-labelled protocollagen<br />
I<br />
Metal chelate aff<strong>in</strong>ity chromatography<br />
II<br />
Gel filtration<br />
II<br />
N-term<strong>in</strong>al sequenc<strong>in</strong>g<br />
II<br />
Am<strong>in</strong>o acid analysis<br />
I, III<br />
Circular dichroism spectroscopy II<br />
Cell <strong>and</strong> tissue Cell culture I, II<br />
Preparation <strong>and</strong> sta<strong>in</strong><strong>in</strong>g <strong>of</strong> paraff<strong>in</strong> sections I, III<br />
Immunohistochemistry <strong>and</strong> immun<strong>of</strong>luorescence<br />
I, III<br />
Transmission electron microscopy<br />
I, III<br />
Collagen cross-l<strong>in</strong>k analysis<br />
I, III<br />
Other Generation <strong>of</strong> mouse l<strong>in</strong>es I, III<br />
61
5 Results<br />
5.1 LH1 knock-out mice have <strong>in</strong>creased risk <strong>of</strong> lethal arterial<br />
ruptures (I)<br />
5.1.1 Generation <strong>of</strong> LH1 null mice <strong>and</strong> LH1 expression <strong>in</strong> adult<br />
mouse tissues<br />
Mouse l<strong>in</strong>es lack<strong>in</strong>g LH1 activity were generated to study the role <strong>of</strong> LH1 <strong>in</strong> <strong>vivo</strong><br />
<strong>and</strong> to generate a mouse model for EDS VIA. The Plod1 gene was disrupted by a<br />
homologous recomb<strong>in</strong>ation <strong>of</strong> a target<strong>in</strong>g construct to exon 2 (Figure 1A <strong>in</strong> I).<br />
The <strong>in</strong>-frame <strong>in</strong>sertion <strong>of</strong> the LacZNeo cassette led to the deletion <strong>of</strong> exons 3–6.<br />
Two mouse l<strong>in</strong>es hav<strong>in</strong>g the same phenotype were generated from ES cells <strong>and</strong><br />
backcrossed <strong>in</strong>to the C57BL/6 mouse l<strong>in</strong>e. RT-PCR downstream from the site <strong>of</strong><br />
recomb<strong>in</strong>ation confirmed the <strong>in</strong>activation <strong>of</strong> the Plod1 gene (Figure 1C <strong>in</strong> I). The<br />
genotypes <strong>of</strong> the <strong>of</strong>fspr<strong>in</strong>g obta<strong>in</strong>ed by cross-breed<strong>in</strong>g <strong>of</strong> the heterozygous mice<br />
followed the Mendelian <strong>in</strong>heritance, but both null l<strong>in</strong>es lacked sufficiently strong<br />
β-galactosidase sta<strong>in</strong><strong>in</strong>g <strong>in</strong> tissues to study the expression <strong>of</strong> the Plod1 gene (data<br />
not shown). Therefore, an additional LH1 null mouse l<strong>in</strong>e with a gene-trap<br />
<strong>in</strong>sertion cassette conta<strong>in</strong><strong>in</strong>g a β-galactosidase reporter <strong>in</strong> <strong>in</strong>tron 12 <strong>of</strong> the Plod1<br />
gene was generated <strong>and</strong> backcrossed <strong>in</strong>to the C57BL/6 mouse l<strong>in</strong>e (not shown).<br />
The LH1 gene-trap mouse l<strong>in</strong>e with functional β-galactosidase expression <strong>in</strong><br />
tissues had identical phenotypic characteristics with the two other Plod1 -/- mouse<br />
l<strong>in</strong>es. The gene-trap mouse l<strong>in</strong>e was used for analysis <strong>of</strong> Plod1 expression by Xgal<br />
sta<strong>in</strong><strong>in</strong>g.<br />
X-gal sta<strong>in</strong><strong>in</strong>g, represent<strong>in</strong>g the Plod1 promoter-driven β-galactosidase<br />
expression, was localized to cells that produce fibrillar collagens, such as<br />
fibroblasts <strong>in</strong> the lung, sk<strong>in</strong> <strong>and</strong> tendon (Figure 2A–B <strong>and</strong> 2D <strong>in</strong> I). In sk<strong>in</strong> the<br />
papilla <strong>of</strong> hair follicles <strong>and</strong> arrector pili muscles were also sta<strong>in</strong>ed (Figure 2B <strong>in</strong> I).<br />
In cornea the sta<strong>in</strong><strong>in</strong>g was seen <strong>in</strong> keratocytes (Figure 2F <strong>in</strong> I). Intense sta<strong>in</strong><strong>in</strong>g <strong>in</strong><br />
Plod1 -/- mouse tissues was observed <strong>in</strong> the heart, where the strongest sta<strong>in</strong><strong>in</strong>g was<br />
seen <strong>in</strong> atrial auricles (Figure 2C <strong>in</strong> I), while <strong>in</strong> the heart muscle the sta<strong>in</strong><strong>in</strong>g was<br />
weaker (not shown). In the skeletal muscle the sta<strong>in</strong><strong>in</strong>g was observed mostly <strong>in</strong><br />
the epimysium (not shown). Moreover, strong LH1 expression was observed <strong>in</strong><br />
the bone <strong>and</strong> cartilage, where the sta<strong>in</strong><strong>in</strong>g was seen <strong>in</strong> osteoblasts <strong>and</strong> the<br />
periosteum <strong>and</strong> <strong>in</strong> chondrocytes (Figure 2E <strong>in</strong> I).<br />
63
5.1.2 LH1 null mice have no kyphoscoliosis but are hypotonic <strong>and</strong><br />
have <strong>in</strong>creased mortality due to arterial ruptures<br />
LH1 null mice were fertile <strong>and</strong> most <strong>of</strong> them had a normal lifespan. The mice<br />
were not observed to suffer from kyphoscoliosis, which is a characteristic feature<br />
<strong>of</strong> the EDS VIA syndrome (Ste<strong>in</strong>mann et al. 2002), but they were passive <strong>and</strong><br />
hypotonic dur<strong>in</strong>g h<strong>and</strong>l<strong>in</strong>g <strong>and</strong> they also appeared to be slower than wild-type<br />
mice. In walk<strong>in</strong>g tests on a rod or on the wall <strong>of</strong> a plastic mouse cage, LH1 null<br />
mice had notable difficulties to move normally <strong>and</strong> they got tired easily, while the<br />
wild-type littermates had no difficulties <strong>in</strong> runn<strong>in</strong>g fast (Figure 3 <strong>in</strong> I). These<br />
features were more dist<strong>in</strong>ct when the mice became older <strong>and</strong> <strong>their</strong> weight<br />
<strong>in</strong>creased.<br />
Although most <strong>of</strong> the Plod1 -/- mice had a normal lifespan, 9% <strong>of</strong> the females<br />
<strong>and</strong> 17% <strong>of</strong> the males were found dead <strong>in</strong> <strong>their</strong> cages <strong>in</strong> the morn<strong>in</strong>g. Premature<br />
deaths occurred suddenly dur<strong>in</strong>g the most active night-time. In almost all <strong>of</strong> these<br />
mice autopsy revealed a hemorrhage <strong>in</strong> the thoracic or abdom<strong>in</strong>al cavity (Table 1<br />
<strong>in</strong> I). In addition, histological analyses showed aortic ruptures (Figure 4C <strong>in</strong> I),<br />
<strong>and</strong> <strong>in</strong> one case there were also signs <strong>of</strong> a previous non-lethal <strong>in</strong>jury (Figure 4D <strong>in</strong><br />
I). The ruptures were seen <strong>in</strong> the tunica media, between the external layers <strong>of</strong> the<br />
elastic lamellae, while the elastic lamellae themselves were normal (Figure 4C–D<br />
<strong>and</strong> 4F <strong>in</strong> I). Healthy Plod1 -/- mice showed no dist<strong>in</strong>ct histological changes <strong>in</strong> the<br />
aorta when compared to the wild-type mice (Figure 4A–B <strong>in</strong> I). X-gal sta<strong>in</strong><strong>in</strong>g<br />
was observed <strong>in</strong> the cells <strong>of</strong> the tunica media <strong>and</strong> adventitia <strong>of</strong> the aorta <strong>in</strong> the<br />
LH1 gene-trap mouse l<strong>in</strong>e (Figure 4B <strong>in</strong> I). These cells appeared to be less<br />
organized <strong>in</strong> LH1 null mice. Echocardiographic analysis <strong>in</strong>dicated that the<br />
structure <strong>and</strong> function <strong>of</strong> the heart is normal (not shown). However, transmission<br />
electron microscopy (TEM) analyses <strong>of</strong> healthy Plod1 -/- mouse aortic walls<br />
revealed degenerative changes <strong>in</strong> smooth muscle cells, such as vacuolization <strong>of</strong><br />
the cells <strong>and</strong> swollen mitochondria (Figure 4F <strong>in</strong> I). Some completely degenerated<br />
cells were also seen (Figure 4F <strong>in</strong> I).<br />
5.1.3 Abnormal collagen fibrils <strong>in</strong> LH1 null mice<br />
In addition to kyphoscoliosis, typical characteristics <strong>of</strong> EDS VIA <strong>in</strong>clude<br />
hyperextensible, fragile sk<strong>in</strong> <strong>and</strong> abnormal scarr<strong>in</strong>g, for <strong>in</strong>stance (Ste<strong>in</strong>mann et al.<br />
2002). The Plod1 -/- mice appeared to have normal sk<strong>in</strong>, <strong>and</strong> no hyperextensibility<br />
or abnormal scarr<strong>in</strong>g was observed. Histological analyses <strong>of</strong> the sk<strong>in</strong> were also<br />
64
normal (not shown). However, TEM analysis from the sk<strong>in</strong> revealed alterations <strong>in</strong><br />
collagen fibrils. The mean collagen fibril diameter <strong>in</strong> the sk<strong>in</strong> <strong>of</strong> Plod1 -/- mice was<br />
<strong>in</strong>creased when compared to that <strong>of</strong> wild-type (91 nm <strong>and</strong> 62 nm, respectively)<br />
<strong>and</strong> also the variation between fibril diameters was larger <strong>in</strong> LH1 null mice<br />
(Figure 5E–H <strong>in</strong> I). Similar observations were made from aorta samples, where<br />
mean fibril diameter <strong>in</strong> the Plod1 -/- samples was 67 nm <strong>and</strong> <strong>in</strong> the wild-type<br />
samples 44 nm (Figure 5A–D <strong>in</strong> I). In addition to larger variability <strong>in</strong> the fibril<br />
diameters, LH1 null samples were also observed to have fibrils with irregular<br />
shape (Figure 5B <strong>and</strong> 5F <strong>in</strong> I). In the aorta, also the diameter with<strong>in</strong> a fibril varied<br />
<strong>in</strong> the Plod1 -/- samples (not shown) <strong>and</strong> the null sk<strong>in</strong> samples conta<strong>in</strong>ed degrad<strong>in</strong>g<br />
fibrils (Figure 5F <strong>in</strong> I). Collagen fibrils <strong>in</strong> the cornea were normal (not shown).<br />
5.1.4 Decreased LH activity <strong>in</strong> LH1 null mice affects the<br />
hydroxylys<strong>in</strong>e content <strong>in</strong> various tissues differently<br />
One diagnostic marker <strong>of</strong> EDS VIA is decreased LH activity <strong>in</strong> cultured sk<strong>in</strong><br />
fibroblasts (Ste<strong>in</strong>mann et al. 2002). Thus, LH activity was determ<strong>in</strong>ed from<br />
homogenized sk<strong>in</strong> samples <strong>of</strong> the LH1 null mice <strong>and</strong> also from homogenized<br />
aorta, which was found to be a structurally weak tissue <strong>in</strong> the mutant mice as well<br />
as <strong>in</strong> EDS VIA patients (Ste<strong>in</strong>mann et al. 2002). The method where [ 14 C]-lys<strong>in</strong>e<br />
labelled non-hydroxylated protocollagen is used as a substrate (Kivirikko &<br />
Myllylä 1982) determ<strong>in</strong>es total LH activity <strong>in</strong> samples compris<strong>in</strong>g the activities <strong>of</strong><br />
LH1, LH2 <strong>and</strong> LH3. LH activity <strong>in</strong> the sk<strong>in</strong> samples <strong>of</strong> Plod1 -/- mice was<br />
decreased to 36% <strong>of</strong> that <strong>in</strong> the wild-type samples (Table 2 <strong>in</strong> I). No reports are<br />
available about the LH activity <strong>in</strong> the aorta <strong>of</strong> EDS VIA patients. It was thus quite<br />
surpris<strong>in</strong>g that the residual LH activity <strong>in</strong> the Plod1 -/- aorta was as high as 44% <strong>of</strong><br />
that <strong>in</strong> the wild-type sample (Table 2 <strong>in</strong> I). The activity <strong>in</strong> the sk<strong>in</strong> <strong>of</strong> the LH1<br />
heterozygous mice was about 55% <strong>and</strong> <strong>in</strong> the aorta about 60% <strong>of</strong> that <strong>in</strong> the wildtype<br />
samples (Table 2 <strong>in</strong> I).<br />
Total hydroxylys<strong>in</strong>e content determ<strong>in</strong>ed by reverse-phase high-pressure liquid<br />
chromatography (HPLC) was decreased <strong>in</strong> all the Plod1 -/- tissues studied when<br />
compared to the wild-type tissues (Table 3 <strong>in</strong> I). The hydroxylys<strong>in</strong>e content,<br />
expressed as per triple helix, varies markedly between tissues as each tissue<br />
conta<strong>in</strong>s different amounts <strong>of</strong> different collagen types <strong>and</strong> different collagen types<br />
conta<strong>in</strong> variable amounts <strong>of</strong> hydroxylated lys<strong>in</strong>es. The highest amount <strong>of</strong><br />
hydroxylys<strong>in</strong>e <strong>in</strong> wild-type samples was observed <strong>in</strong> the lung (55 residues / triple<br />
helix) (Table 3 <strong>in</strong> I). Also <strong>in</strong> the Plod1 -/- lung samples the level was relatively high,<br />
65
e<strong>in</strong>g about 85% <strong>of</strong> that <strong>of</strong> the wild-type (Table 3 <strong>in</strong> I). The lowest amount <strong>of</strong><br />
hydroxylys<strong>in</strong>e <strong>in</strong> the wild-type samples was detected <strong>in</strong> the sk<strong>in</strong> with 21 residues<br />
per triple helix (Table 3 <strong>in</strong> I). In Plod1 -/- sk<strong>in</strong> the correspond<strong>in</strong>g value was only<br />
22% <strong>of</strong> that <strong>of</strong> the wild-type, be<strong>in</strong>g the most dramatically affected tissue studied<br />
(Table 3 <strong>in</strong> I). The hydroxylys<strong>in</strong>e content was markedly decreased also <strong>in</strong> the<br />
Plod1 -/- tail tendon <strong>and</strong> cornea, be<strong>in</strong>g 24% <strong>and</strong> 30% <strong>of</strong> the wild-type values,<br />
respectively, while the aorta, femur <strong>and</strong> heart auricle were less affected<br />
(hydroxylys<strong>in</strong>e content 63%, 75% <strong>and</strong> 69%, respectively, <strong>of</strong> the wild-type values)<br />
(Table 3 <strong>in</strong> I).<br />
5.1.5 Abnormal hydroxylation <strong>of</strong> triple helical lys<strong>in</strong>e residues affects<br />
the cross-l<strong>in</strong>k<strong>in</strong>g pattern <strong>in</strong> the LH1 null mouse tissues<br />
As changes <strong>in</strong> the hydroxylation <strong>of</strong> certa<strong>in</strong> lys<strong>in</strong>e residues <strong>in</strong> collagens affect<br />
cross-l<strong>in</strong>k formation <strong>and</strong> as <strong>in</strong> EDS VIA patients the cross-l<strong>in</strong>k<strong>in</strong>g pattern is<br />
changed (Ste<strong>in</strong>mann et al. 2002), mature cross-l<strong>in</strong>ks were studied from wild-type<br />
<strong>and</strong> Plod1 -/- mice. Similar to the hydroxylys<strong>in</strong>e levels, also the LP <strong>and</strong> HP content<br />
varies markedly between tissues (Table 4 <strong>in</strong> I). The highest levels <strong>of</strong> both mature<br />
cross-l<strong>in</strong>ks <strong>in</strong> the wild-type mouse were observed <strong>in</strong> the aorta, the amounts <strong>of</strong> HP<br />
<strong>and</strong> LP be<strong>in</strong>g 0.71 <strong>and</strong> 0.26 residues per triple helix, respectively (Table 4 <strong>in</strong> I).<br />
The amount <strong>of</strong> HP was decreased <strong>in</strong> all LH1 null tissues studied, <strong>and</strong> the most<br />
dramatic change was seen <strong>in</strong> the aorta, where the HP content was 28% <strong>of</strong> that <strong>of</strong><br />
the wild-type sample (Table 4 <strong>in</strong> I). The lowest HP <strong>and</strong> LP levels <strong>in</strong> wild-type<br />
tissues were observed <strong>in</strong> the tail tendon (0.051 <strong>and</strong> 0.006 residues / triple helix,<br />
respectively) <strong>and</strong> cornea (0.051 <strong>and</strong> 0.016 residues / triple helix) (Table 4 <strong>in</strong> I). In<br />
the cornea <strong>of</strong> LH1 null mice the HP amount was only 33% <strong>of</strong> that <strong>of</strong> the wild-type,<br />
but <strong>in</strong> the tail tendon the relative amount was 59% (Table 4 <strong>in</strong> I). In Plod1 -/- lung<br />
the amount <strong>of</strong> HP was decreased markedly (34% <strong>of</strong> the wild-type), while the<br />
femur was slightly less affected (47% <strong>of</strong> the wild-type) (Table 4 <strong>in</strong> I). Although<br />
only certa<strong>in</strong> hydroxylys<strong>in</strong>e residues <strong>in</strong> collagens participate <strong>in</strong> the cross-l<strong>in</strong>k<br />
formation, it was <strong>in</strong>terest<strong>in</strong>g that there was no correlation between the changes <strong>in</strong><br />
hydroxylys<strong>in</strong>e <strong>and</strong> HP levels (Tables 3 <strong>and</strong> 4 <strong>in</strong> I). For example, <strong>in</strong> the tail tendon<br />
the hydroxylys<strong>in</strong>e level was 24% <strong>of</strong> the wild-type tissue, but the amount <strong>of</strong> HP<br />
was as high as 59% <strong>of</strong> the wild-type value <strong>and</strong> <strong>in</strong> the lung the correspond<strong>in</strong>g<br />
percentages were 86% <strong>and</strong> 34%, respectively. In human sk<strong>in</strong> the levels <strong>of</strong> mature<br />
LP <strong>and</strong> HP cross-l<strong>in</strong>ks are generally low (Knott & Bailey 1998, Avery & Bailey<br />
2005) <strong>and</strong> they were not detectable <strong>in</strong> the mouse sk<strong>in</strong> either.<br />
66
As <strong>in</strong> the EDS VIA patients, the LP level <strong>in</strong> the Plod1 -/- mouse tissues was<br />
<strong>in</strong>creased <strong>in</strong> comparison to the correspond<strong>in</strong>g wild-type tissues (Table 4 <strong>in</strong> I). The<br />
<strong>in</strong>crease varied from about 5-fold to about 60-fold when compared to the wildtype<br />
(Table 4 <strong>in</strong> I). As a consequence <strong>of</strong> the decreased HP levels <strong>and</strong> <strong>in</strong>creased LP<br />
levels, the LP/HP ratio was also <strong>in</strong>creased dramatically <strong>in</strong> the LH1 null tissues<br />
(Table 4 <strong>in</strong> I). In wild-type tissues the LP/HP ratio varied between 0.12 <strong>and</strong> 0.39,<br />
whereas the values <strong>in</strong> the Plod1 -/- mouse tissues ranged from 3.8 to 12.5, the<br />
smallest <strong>in</strong>crease be<strong>in</strong>g observed <strong>in</strong> the aorta <strong>and</strong> the highest <strong>in</strong> the tail tendon<br />
(Table 4 <strong>in</strong> I). Also the total amount <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks (HP + LP) was<br />
<strong>in</strong>creased <strong>in</strong> all LH1 null mouse tissues studied (Table 4 <strong>in</strong> I). The <strong>in</strong>crease varied<br />
between tissues, be<strong>in</strong>g lowest <strong>in</strong> the aorta (about 140% <strong>of</strong> the wild-type) <strong>and</strong><br />
highest <strong>in</strong> the tail tendon, where the rise was about 720% <strong>of</strong> the wild-type sample.<br />
5.2 Analysis <strong>of</strong> the molecular level consequences <strong>of</strong> Bruck<br />
syndrome mutations by the use <strong>of</strong> recomb<strong>in</strong>ant LH2 prote<strong>in</strong>s (II)<br />
5.2.1 Expression <strong>of</strong> wild-type <strong>and</strong> recomb<strong>in</strong>ant LH2(long)<br />
polypeptides with Bruck syndrome mutations <strong>in</strong> <strong>in</strong>sect cells<br />
<strong>and</strong> <strong>their</strong> purification<br />
BS2 is caused by mutations <strong>in</strong> LH2 (van der Slot et al. 2003). To study the cause<br />
<strong>of</strong> BS2 at the molecular level, the currently known po<strong>in</strong>t mutations lead<strong>in</strong>g to<br />
LH2 am<strong>in</strong>o acid substitutions R594H, G597V, <strong>and</strong> T604I were <strong>in</strong>troduced <strong>in</strong>to the<br />
pAcgp67A-LH2(long) baculovirus transfer vector by site-directed mutagenesis<br />
(Figure 1 <strong>in</strong> II) (Krol et al. 1996, Rautavuoma et al. 2002). The baculoviral signal<br />
peptide gb67 directs the synthesized prote<strong>in</strong> for secretion <strong>in</strong>to the culture medium<br />
<strong>and</strong> the N-term<strong>in</strong>al histid<strong>in</strong>e tag enables selective purification. Wild-type <strong>and</strong> the<br />
three BS mutant LH2(long) polypeptides were expressed <strong>in</strong> High five <strong>in</strong>sect cells<br />
<strong>and</strong> the secreted recomb<strong>in</strong>ant prote<strong>in</strong>s were purified from the culture medium 72<br />
h after <strong>in</strong>fection by metal chelate aff<strong>in</strong>ity chromatography followed by histid<strong>in</strong>e<br />
elution.<br />
As the f<strong>in</strong>al yield <strong>of</strong> the purified LH2 mutant polypeptides R594H <strong>and</strong><br />
G597V was low, we decided to analyze the behavior <strong>of</strong> the mutant polypeptides<br />
dur<strong>in</strong>g purification by tak<strong>in</strong>g aliquots from different purification steps for analysis<br />
by SDS-PAGE (Figure 2 <strong>in</strong> II). Wild-type <strong>and</strong> T604I LH2(long) bound effectively<br />
to the aff<strong>in</strong>ity column, whereas a significant amount <strong>of</strong> the mutants R594H <strong>and</strong><br />
67
G597V were present <strong>in</strong> the flow-through samples (Figure 2A–B <strong>in</strong> II). In addition,<br />
the bound R594H <strong>and</strong> G597V LH2(long) polypeptides detached easily from the<br />
column dur<strong>in</strong>g wash<strong>in</strong>g steps <strong>in</strong>dicat<strong>in</strong>g that the b<strong>in</strong>d<strong>in</strong>g mediated by the histid<strong>in</strong>e<br />
tag is weak (Figure 2C–D <strong>in</strong> II).<br />
The R594H <strong>and</strong> G597V mutant polypeptides that rema<strong>in</strong>ed bound to the<br />
aff<strong>in</strong>ity column after the wash<strong>in</strong>g procedure eluted immediately after the addition<br />
<strong>of</strong> the elution buffer, whereas the T604I mutant <strong>and</strong> wild-type polypeptides<br />
required a longer presence <strong>of</strong> histid<strong>in</strong>e to detach from the column (Figure 3 <strong>in</strong> II).<br />
These results <strong>in</strong>dicate that <strong>in</strong> the T604I mutant <strong>and</strong> wild-type LH2(long) the<br />
histid<strong>in</strong>e tag is readily exposed, while the R594H <strong>and</strong> G597V mutations may<br />
change the structure <strong>of</strong> the prote<strong>in</strong> so that the tag becomes partly hidden.<br />
5.2.2 The Bruck syndrome mutant R594H <strong>and</strong> G597V LH2(long)<br />
polypeptides have problems <strong>in</strong> dimerization <strong>and</strong> are prone to<br />
degradation<br />
To analyze the purified recomb<strong>in</strong>ant wild-type <strong>and</strong> mutant LH2(long) prote<strong>in</strong>s,<br />
fractions conta<strong>in</strong><strong>in</strong>g the purest samples were pooled, concentrated <strong>and</strong> the<br />
histid<strong>in</strong>e was removed by PD-10 columns. Analyses by SDS-PAGE under<br />
reduc<strong>in</strong>g conditions <strong>and</strong> nondenatur<strong>in</strong>g PAGE revealed that <strong>in</strong> contrast to the<br />
wild-type <strong>and</strong> T604I mutant polypeptides, the R594H <strong>and</strong> G597V mutants<br />
coeluted with additional polypeptides (Figure 4A <strong>and</strong> 4C <strong>in</strong> II). An anti-LH<br />
antibody identified some <strong>of</strong> these additional polypeptides <strong>in</strong> Western blot analysis<br />
(Figure 4B <strong>and</strong> 4D <strong>in</strong> II). LHs are dimeric prote<strong>in</strong>s (Kivirikko et al. 1992,<br />
Rautavuoma et al. 2002) that dissociate <strong>in</strong>to a monomeric form <strong>in</strong> reduc<strong>in</strong>g<br />
SDS-PAGE (Rautavuoma et al. 2002). Western blot analysis showed that the<br />
R594H <strong>and</strong> G597V mutants formed aggregates that were not dissociated <strong>in</strong>to<br />
monomers <strong>in</strong> SDS-PAGE like the wild-type <strong>and</strong> T604I polypeptides (Figure 4B <strong>in</strong><br />
II). The antibody recognized also two other b<strong>and</strong>s <strong>in</strong> the R594H <strong>and</strong> G597V<br />
samples with a molecular weight smaller than that <strong>of</strong> the LH2 monomer,<br />
<strong>in</strong>dicat<strong>in</strong>g that these two mutant forms were prone to degradation (Figure 4B <strong>in</strong><br />
II). Identification <strong>of</strong> the other coelut<strong>in</strong>g b<strong>and</strong>s by N-term<strong>in</strong>al sequenc<strong>in</strong>g was<br />
unsuccessful <strong>and</strong> thus these polypeptides were regarded as impurities (data not<br />
shown).<br />
As <strong>in</strong>dicated earlier, LH1 is shown to be a homodimeric prote<strong>in</strong> <strong>in</strong> tissues<br />
(Kivirikko et al. 1992) <strong>and</strong> also recomb<strong>in</strong>ant LH is<strong>of</strong>orms produced <strong>in</strong> <strong>in</strong>sect cells<br />
are shown to form dimers (Rautavuoma et al. 2002). Analysis <strong>of</strong> the mutant<br />
68
ecomb<strong>in</strong>ant polypeptides by nondenatur<strong>in</strong>g PAGE revealed that the T604I<br />
mutant assembled <strong>in</strong>to dimers normally when compared to the wild-type, while<br />
oligomerization <strong>of</strong> the R594H <strong>and</strong> G597V mutants was abnormal (Figure 4C–D<br />
<strong>in</strong> II). The R594H <strong>and</strong> G597V polypeptides formed oligomers that had either a<br />
slower or faster mobility than the dimers formed by the wild-type or T604I<br />
mutant LH2(long) polypeptides, <strong>and</strong> they also formed aggregates that were<br />
unable to enter the gel properly (Figure 4C–D <strong>in</strong> II).<br />
5.2.3 Fold<strong>in</strong>g <strong>of</strong> the R594H <strong>and</strong> G597V mutant LH2(long)<br />
polypeptides is impaired <strong>and</strong> <strong>their</strong> proteolytic sensitivity is<br />
<strong>in</strong>creased<br />
To underst<strong>and</strong> the consequences <strong>of</strong> the BS po<strong>in</strong>t mutations on the fold<strong>in</strong>g <strong>of</strong> LH2,<br />
the wild-type <strong>and</strong> mutant LH2(long) polypeptides were analyzed by far-UV<br />
circular dichroism (CD) spectroscopy. The CD spectra <strong>of</strong> the wild-type <strong>and</strong> T604I<br />
mutant LH2(long) were essentially identical <strong>and</strong> typical for folded, ma<strong>in</strong>ly<br />
α-helical prote<strong>in</strong>s with a positive maximum at 193 nm <strong>and</strong> negative m<strong>in</strong>ima at<br />
208 <strong>and</strong> 222 nm (Figure 5 <strong>in</strong> II). In accordance with the purification, SDS-PAGE,<br />
native PAGE <strong>and</strong> Western analyses, marked changes <strong>in</strong> the fold<strong>in</strong>g <strong>of</strong> the mutants<br />
R594H <strong>and</strong> G597V were observed. The R594H <strong>and</strong> G597V LH2(long)<br />
polypeptides both had only one negative m<strong>in</strong>imum at about 207 nm (Figure 5 <strong>in</strong><br />
II), <strong>in</strong>dicat<strong>in</strong>g a dramatic change <strong>in</strong> the fold<strong>in</strong>g <strong>of</strong> these two mutant forms.<br />
As the results from the above analyses <strong>in</strong>dicated impaired structural quality<br />
<strong>of</strong> two <strong>of</strong> the mutant LH2(long) polypeptides, the proteolytic sensitivity <strong>of</strong> wildtype<br />
<strong>and</strong> mutant forms were studied by thermolys<strong>in</strong> digestion. LH polypeptides<br />
have been previously shown to have two protease-sensitive regions that are<br />
digested by thermolys<strong>in</strong> generat<strong>in</strong>g three doma<strong>in</strong>s with molecular weights <strong>of</strong> 16,<br />
30 <strong>and</strong> 37 kDa (Rautavuoma et al. 2002). Digestion <strong>of</strong> the wild-type <strong>and</strong> T604I<br />
mutant polypeptides produced these three thermolys<strong>in</strong>-resistant doma<strong>in</strong>s with<br />
molecular weights similar to those reported earlier (Rautavuoma et al. 2002)<br />
(Figure 6 <strong>in</strong> II), but digestion <strong>of</strong> the R594H <strong>and</strong> G597V mutants led to more<br />
extensive proteolysis, support<strong>in</strong>g the above results that fold<strong>in</strong>g <strong>of</strong> these mutants is<br />
not normal (Figure 6 <strong>in</strong> II).<br />
69
5.2.4 All Bruck syndrome mutations affect the activity <strong>of</strong> LH2<br />
To underst<strong>and</strong> the effect <strong>of</strong> the BS2 po<strong>in</strong>t mutations on the activity <strong>of</strong> LH2, LH<br />
activity <strong>of</strong> the recomb<strong>in</strong>ant wild-type <strong>and</strong> mutant enzymes was analyzed by an<br />
assay based on the hydroxylation-coupled decarboxylation <strong>of</strong><br />
2-oxo[1- 14 C]glutarate with a synthetic peptide (Ile-Lys-Gly)3 as a substrate<br />
(Kivirikko & Myllylä 1982). The wild-type LH2(long) was used <strong>in</strong> the activity<br />
assays as an <strong>in</strong>ternal control.<br />
The results <strong>of</strong> the assays showed that the activity <strong>of</strong> all the three mutant LH2<br />
enzymes is decreased. Under the assay conditions used, typical activity values<br />
obta<strong>in</strong>ed with the wild-type LH2(long) were about 6000 dpm, whereas the<br />
activity <strong>of</strong> the T604I mutant was only about 500 dpm <strong>and</strong> those <strong>of</strong> the R594H <strong>and</strong><br />
G597V mutants less than 300 dpm. Although the activity <strong>of</strong> the T604I mutant was<br />
decreased to less than 10% <strong>of</strong> that <strong>of</strong> the wild-type, the Km value <strong>of</strong> this mutant<br />
for the peptide substrate was normal (Table 1 <strong>in</strong> II). However, changes <strong>in</strong> the Km<br />
values for the reaction cosubstrates may also have an effect on the enzyme<br />
activity. The Km values <strong>of</strong> the T604I mutant for ascorbate <strong>and</strong> Fe 2+ were normal<br />
when compared to the wild-type, whereas the Km for 2-oxoglutarate was<br />
<strong>in</strong>creased by about 10-fold (Table 1 <strong>in</strong> II). When 2-oxoglutarate was used <strong>in</strong> a<br />
saturat<strong>in</strong>g concentration <strong>in</strong> the reaction, the activity <strong>of</strong> the T604I mutant <strong>in</strong>creased<br />
slightly, be<strong>in</strong>g about 30% <strong>of</strong> the wild-type (data not shown). The Km values <strong>of</strong> the<br />
R594H <strong>and</strong> G597V mutants for the peptide substrate were markedly <strong>in</strong>creased<br />
when compared to the wild-type (Table 1 <strong>in</strong> II). Because <strong>of</strong> the fold<strong>in</strong>g problems<br />
<strong>and</strong> extremely high Km values for the substrate, the catalytic properties <strong>of</strong> the<br />
R594H <strong>and</strong> G597V mutants were not studied further.<br />
LH2(long) is reported to hydroxylate the telopeptide lys<strong>in</strong>es <strong>in</strong> collagens<br />
(Uzawa et al. 1999, Mercer et al. 2003, van der Slot et al. 2003, Pornprasertsuk et<br />
al. 2004, van der Slot et al. 2004, Takaluoma et al. 2007). Previously it was<br />
shown that LH2(long) is the only LH isoenzyme capable <strong>of</strong> hydroxylat<strong>in</strong>g the N<br />
telopeptide lys<strong>in</strong>e (Lys 9 ) <strong>in</strong> a full-length pro-α1(I) cha<strong>in</strong> <strong>of</strong> type I collagen when<br />
coexpressed <strong>in</strong> <strong>in</strong>sect cells (Takaluoma et al. 2007). As impaired 2-oxoglutarate<br />
b<strong>in</strong>d<strong>in</strong>g was the only characteristic that clearly dist<strong>in</strong>guished the T604I mutant<br />
LH2(long) from the wild-type, we tested whether the mutant is nevertheless able<br />
to hydroxylate the telopeptide Lys 9 <strong>of</strong> type I collagen <strong>in</strong> <strong>in</strong>sect cell coexpression<br />
experiments. Full-length proα1(I) cha<strong>in</strong>s were coexpressed with C-P4H <strong>and</strong> wildtype<br />
or T604I mutant LH2(long). The produced recomb<strong>in</strong>ant type I procollagen<br />
homotrimers were digested with peps<strong>in</strong> <strong>and</strong> the triple helical collagen I<br />
70
homotrimers were further purified <strong>and</strong> analyzed by N-term<strong>in</strong>al sequenc<strong>in</strong>g. The<br />
wild-type LH2(long) hydroxylated the telopeptide Lys 9 as described previously<br />
(Takaluoma et al. 2007), whereas the hydroxylys<strong>in</strong>e peak was absent <strong>in</strong> the<br />
reverse-phase HPLC pr<strong>of</strong>ile <strong>of</strong> the collagen I homotrimer purified after<br />
coexpression with the T604I mutant LH2(long) (Figure 7A–B <strong>in</strong> II). Thus, the<br />
T604I mutant LH2(long) was not able to hydroxylate the telopeptide Lys 9 residue<br />
dur<strong>in</strong>g coexpression, reveal<strong>in</strong>g the importance <strong>of</strong> proper b<strong>in</strong>d<strong>in</strong>g <strong>of</strong> reaction<br />
cosubstrates <strong>in</strong> the LH reaction.<br />
5.3 <strong>Lysyl</strong> hydroxylase 2 is essential for the embryonic<br />
development <strong>of</strong> mouse (III)<br />
5.3.1 LH2 null mice die after embryonic day 10.5<br />
To study the <strong>in</strong> <strong>vivo</strong> role <strong>of</strong> LH2, we generated three different mouse l<strong>in</strong>es<br />
carry<strong>in</strong>g a gene-trap <strong>in</strong>sertion cassette with a β-galactosidase reporter <strong>in</strong> the Plod2<br />
gene that disrupts the gene function (the LH2:KST067 gene-trap target<strong>in</strong>g<br />
strategy is shown <strong>in</strong> Figure 1A <strong>in</strong> III). All three l<strong>in</strong>es showed the same ma<strong>in</strong><br />
phenotype i.e. embryonic lethality, <strong>and</strong> thus we chose one mouse l<strong>in</strong>e orig<strong>in</strong>at<strong>in</strong>g<br />
from the targeted ES cell l<strong>in</strong>e LH2:KST067 for further analysis. The lack <strong>of</strong> the<br />
production <strong>of</strong> LH2 mRNA with a sequence from exon 5 onwards was verified by<br />
RT-PCR (Figure 1B <strong>in</strong> III). The heterozygous Plod2 wt/gt (wt st<strong>and</strong>s for a wild-type<br />
<strong>and</strong> gt for a gene-trapped allele) mice appeared to be healthy <strong>and</strong> fertile <strong>and</strong> had a<br />
normal lifespan, but genotyp<strong>in</strong>g (Figure 1C <strong>in</strong> III) revealed that <strong>in</strong>tercrosses <strong>of</strong> the<br />
heterozygous mice yielded only wild-type <strong>and</strong> heterozygous <strong>of</strong>fspr<strong>in</strong>g. Analysis<br />
<strong>of</strong> the embryos at different ages from heterozygous breed<strong>in</strong>gs revealed that the<br />
Plod2 gt/gt embryos developed seem<strong>in</strong>gly normally until E10.5, when the LH2 null<br />
embryos constituted 23% <strong>of</strong> the population (n = 165 <strong>and</strong> 724, respectively) (Table<br />
2 <strong>in</strong> III). Thus, the amounts <strong>of</strong> the Plod2 gt/gt embryos at E10.5 were close to the<br />
expected Mendelian ratio. At E11.5 the proportion <strong>of</strong> LH2 null embryos was<br />
about 14% (n = 326) <strong>and</strong> at E12.5 only 6% (n = 85) (Table 2 <strong>in</strong> III). After E12.5<br />
liv<strong>in</strong>g Plod2 gt/gt embryos were not observed. Older embryos between E11.5–12.5<br />
were collected when the backcross<strong>in</strong>g was still ongo<strong>in</strong>g. Thus, the given<br />
percentages could be even lower <strong>in</strong> the backcrossed l<strong>in</strong>es.<br />
The Plod2 gt/gt embryos developed apparently normally until E10.5 but died<br />
suddenly after this. The size <strong>and</strong> appearance <strong>of</strong> the LH2 null embryos at 10.5 was<br />
71
seem<strong>in</strong>gly normal when compared to the wild-type <strong>and</strong> heterozygous littermates<br />
(Figure 2A <strong>in</strong> III), but pooled blood was frequently observed <strong>in</strong>side the fetal<br />
membranes (Figure 2B–C <strong>in</strong> III) <strong>and</strong> sometimes possibly between the visceral<br />
yolk sac <strong>and</strong> amnion (data not shown), while the heart was still beat<strong>in</strong>g. After the<br />
first observation <strong>of</strong> the accumulated blood, 21% <strong>of</strong> the Plod2 gt/gt embryos <strong>and</strong> 2%<br />
<strong>of</strong> the Plod2 wt/gt embryos were found to have this phenomenon, while wild-type<br />
embryos were all normal (Table 3 <strong>in</strong> III).<br />
5.3.2 LH2 is expressed ubiquitously <strong>in</strong> the E10.5 embryos <strong>and</strong> <strong>in</strong><br />
extraembryonic tissues <strong>and</strong> its lack seems to lead to certa<strong>in</strong><br />
changes <strong>in</strong> collagen expression<br />
X-Gal sta<strong>in</strong><strong>in</strong>g, represent<strong>in</strong>g the Plod2 promoter-driven β-galactosidase<br />
expression was extensive <strong>in</strong> E10.5 Plod2 gt/gt embryos (Figure 3A–D <strong>in</strong> III).<br />
Sta<strong>in</strong><strong>in</strong>g was observed ubiquitously <strong>in</strong> the mesenchymal tissue (Figure 3A <strong>in</strong> III).<br />
The expression was <strong>in</strong>tense also <strong>in</strong> somites, the auditory vesicle <strong>and</strong> amnion, for<br />
<strong>in</strong>stance (Figure 3A–D <strong>in</strong> III). In placenta, X-gal sta<strong>in</strong><strong>in</strong>g was observed <strong>in</strong> the<br />
maternal decidua <strong>and</strong> on the embryonic side <strong>in</strong> the chorionic plate <strong>and</strong> labyr<strong>in</strong>th<br />
layer, where the expression was seen <strong>in</strong> the mesenchymal tissue surround<strong>in</strong>g<br />
embryonic vessels (Figure 3E–F <strong>in</strong> III). Parietal endoderm cells produc<strong>in</strong>g the<br />
Reichert’s membrane, the parietal layer <strong>of</strong> the yolk sac, were strongly sta<strong>in</strong>ed <strong>in</strong><br />
Plod2 gt/gt embryos, but the expression was not seen <strong>in</strong> the membrane (Figure 4B<br />
<strong>in</strong> III).<br />
It has been shown earlier that the collagen α1(IV)2α2(IV) null mice die<br />
between E10.5–E11.5 because <strong>of</strong> deficient basement membrane assembly <strong>and</strong><br />
f<strong>in</strong>ally <strong>of</strong> the failure <strong>of</strong> the Reichert's membrane (Pöschl et al. 2004). No obvious<br />
differences between the Plod2 gt/gt <strong>and</strong> wild-type embryos were seen <strong>in</strong> light<br />
microscopy or TEM analysis <strong>of</strong> the Reichert’s membrane or <strong>in</strong> the type IV<br />
collagen immun<strong>of</strong>luorescence sta<strong>in</strong><strong>in</strong>g (Figure 4A–F <strong>in</strong> III). However, accord<strong>in</strong>g<br />
to TEM analysis, we were not able to exclude a weakened adhesion <strong>of</strong> the<br />
Reichert’s membrane to placenta <strong>in</strong> Plod2 gt/gt embryos (data not shown) <strong>and</strong> also<br />
the normal detachment <strong>of</strong> parietal endoderm cells was sometimes questioned<br />
(Figure 4D <strong>in</strong> III).<br />
In other extraembryonic tissues, pronounced X-gal sta<strong>in</strong><strong>in</strong>g was observed <strong>in</strong><br />
the mesodermal part <strong>of</strong> the visceral yolk sac, but the endodermal side was<br />
unsta<strong>in</strong>ed (Figure 5B–C <strong>in</strong> III). Morphological analysis <strong>of</strong> the visceral yolk sac <strong>of</strong><br />
Plod2 gt/gt embryos revealed that some areas <strong>of</strong> the mesodermal part <strong>of</strong> the tissue<br />
72
appeared loose <strong>in</strong> comparison to wild-type (Figure 5A–C <strong>in</strong> III). The loosened<br />
mesodermal tissue was also surround<strong>in</strong>g vitell<strong>in</strong>e vessels, <strong>and</strong> thus it cannot be<br />
excluded that <strong>in</strong> some <strong>of</strong> the hemorrhage cases this site may have collapsed<br />
followed by a rupture <strong>in</strong> the vitell<strong>in</strong>e vessel that could have led to the bleed<strong>in</strong>g<br />
between the yolk sac <strong>and</strong> amnion. In immun<strong>of</strong>luorescence sta<strong>in</strong><strong>in</strong>g <strong>of</strong><br />
extraembryonic tissues, type III <strong>and</strong> type V collagens showed differential<br />
expression patterns (Figure 6A–D <strong>in</strong> III). Type III collagen was seen <strong>in</strong> the<br />
endodermal side <strong>of</strong> the visceral yolk sac (Figure 6A–B <strong>in</strong> III), while type V<br />
collagen was observed <strong>in</strong> the mesodermal part <strong>of</strong> the tissue (Figure 6C–D <strong>in</strong> III).<br />
In the Plod2 gt/gt yolk sac the sta<strong>in</strong><strong>in</strong>g <strong>of</strong> both collagen types seems weaker than<br />
that <strong>of</strong> the Plod2 wt/wt tissue although the distribution was normal (Figure 6A–D <strong>in</strong><br />
III). Sta<strong>in</strong><strong>in</strong>g <strong>of</strong> type I collagen was seen <strong>in</strong> the mesodermal part <strong>of</strong> the visceral<br />
yolk sac, but differences between genotypes were not observed (data not shown).<br />
Type I, IV <strong>and</strong> V collagens but not type III collagen were expressed also <strong>in</strong> the<br />
chorionic plate <strong>of</strong> the placenta, but no obvious differences were seen between the<br />
Plod2 gt/gt <strong>and</strong> Plod2 wt/wt embryos (data not shown).<br />
As the phenotype <strong>of</strong> the mouse l<strong>in</strong>e <strong>in</strong>dicated a cardiovascular defect,<br />
anti-CD31 sta<strong>in</strong><strong>in</strong>g was performed to study the structure <strong>of</strong> large vessels <strong>of</strong> the<br />
embryos, but the endothelium <strong>of</strong> the aorta appeared <strong>in</strong>tact (Figure 7A–B <strong>in</strong> III).<br />
However, TEM analysis revealed that <strong>in</strong> the Plod2 gt/gt aorta the overall structure<br />
<strong>of</strong> the endothelium seemed looser when compared to Plod2 wt/wt embryos (Figure<br />
7C–D <strong>in</strong> III). The mesenchymal cell network under the endothelium was sparse<br />
<strong>and</strong> the cells appeared elongated <strong>and</strong> showed numerous th<strong>in</strong> protrusions <strong>in</strong><br />
Plod2 gt/gt embryos, while <strong>in</strong> the Plod2 wt/wt embryos the mesenchymal cells were<br />
larger <strong>and</strong> they were located more closely to the endothelium. LH2 is expressed <strong>in</strong><br />
the mesodermal tissue surround<strong>in</strong>g the aorta (Figure 3A <strong>in</strong> III). As blood pressure<br />
<strong>in</strong>creases dur<strong>in</strong>g development this area must susta<strong>in</strong> higher mechanical loads <strong>and</strong><br />
provide support for the large vessels. Thus, collagen fibrils were studied by TEM<br />
from this mesodermal area <strong>and</strong> some fibrils were seen <strong>in</strong> both Plod2 wt/wt <strong>and</strong><br />
Plod2 gt/gt samples (Figure 7E–F <strong>in</strong> III).<br />
5.3.3 Expression <strong>of</strong> the Plod1 <strong>and</strong> Plod3 genes is not upregulated <strong>in</strong><br />
the Plod2 gt/gt embryos<br />
To study whether <strong>in</strong>activation <strong>of</strong> the Plod2 gene leads to upregulation <strong>of</strong> Plod1<br />
<strong>and</strong> Plod3 gene expression, we performed Q-PCR <strong>and</strong> RT-PCR analyses from<br />
10.5 days old embryos. Plod2 expression followed the expected levels with about<br />
73
50% expression <strong>in</strong> the Plod2 wt/gt embryos relative to wild-type, whereas<br />
expression was not detected <strong>in</strong> the Plod2 gt/gt embryos (Figure 8A–B <strong>in</strong> III). The<br />
LH mRNA levels were determ<strong>in</strong>ed from s<strong>in</strong>gle embryos <strong>and</strong> results varied<br />
markedly between <strong>in</strong>dividuals <strong>in</strong> the same genotype group. At E10.5 the<br />
development <strong>of</strong> embryos is fast <strong>and</strong> even m<strong>in</strong>or age differences may cause large<br />
differences <strong>in</strong> the mRNA levels. Accord<strong>in</strong>g to our LH1 null mouse study, the<br />
other two LH isoenzymes are suggested to at least partially compensate for the<br />
LH1 deficiency (Article I <strong>in</strong> this thesis). However, the mRNA levels <strong>of</strong> Plod1 <strong>and</strong><br />
Plod3 genes <strong>in</strong> the Plod2 wt/gt <strong>and</strong> Plod2 gt/gt embryos did not show any significant<br />
<strong>in</strong>crease, thus suggest<strong>in</strong>g that they cannot compensate for the decreased LH2 level<br />
(Figure 8A–B <strong>in</strong> III). In any case, the LH1 <strong>and</strong> LH3 could not compensate for the<br />
telopeptide LH activity <strong>of</strong> LH2.<br />
5.3.4 LH2 is widely expressed <strong>in</strong> adult Plod2 wt/gt tissues<br />
Plod2 was expressed <strong>in</strong> almost all the adult tissues studied. In the bone, X-gal<br />
sta<strong>in</strong><strong>in</strong>g was observed <strong>in</strong> the endosteum, periosteum <strong>and</strong> round osteocytes (Figure<br />
9A–B <strong>in</strong> III). The pattern was similar <strong>in</strong> both bone types analyzed, the femur <strong>and</strong><br />
calvaria. Sta<strong>in</strong><strong>in</strong>g was also <strong>in</strong>tense <strong>in</strong> the cartilage (Figure 9C <strong>in</strong> III), tendon (data<br />
not shown), <strong>and</strong> <strong>in</strong> the smooth muscle layers <strong>of</strong> the uterus (Figure 9D <strong>in</strong> III) <strong>and</strong><br />
<strong>in</strong>test<strong>in</strong>e (not shown). In blood vessels LH2 was expressed moderately <strong>in</strong><br />
endothelial cells (Figure 9E–F <strong>in</strong> III). In the tunica media the expression was<br />
localized to smooth muscle cells located between the elastic lamellae (Figure 9E–<br />
F <strong>in</strong> III). X-gal sta<strong>in</strong><strong>in</strong>g was also seen <strong>in</strong> the cells <strong>of</strong> tunica adventitia (Figure 9E–<br />
F <strong>in</strong> III). In the eye, LH2 was expressed <strong>in</strong> the anterior lens epithelium as well as<br />
<strong>in</strong> the ganglion cell layer <strong>and</strong> <strong>in</strong> the <strong>in</strong>ner nuclear layer <strong>of</strong> the ret<strong>in</strong>a (data not<br />
shown). In the sk<strong>in</strong> the expression was strongest <strong>in</strong> the hair bulbs (data not<br />
shown).<br />
5.3.5 Changes <strong>in</strong> lysyl hydroxylation <strong>and</strong> cross-l<strong>in</strong>k<strong>in</strong>g <strong>of</strong> collagen <strong>in</strong><br />
adult Plod2 wt/gt mice<br />
In BS2 patients the collagen cross-l<strong>in</strong>k<strong>in</strong>g is altered due to abnormal function <strong>of</strong><br />
LH2, therefore we wanted to study the collagen cross-l<strong>in</strong>k<strong>in</strong>g from E10.5<br />
embryos. However, maturation <strong>of</strong> collagen cross-l<strong>in</strong>ks requires time. Therefore,<br />
the collagen cross-l<strong>in</strong>k analysis from embryos turned out to be unsuccessful as the<br />
mature cross-l<strong>in</strong>ks had not yet formed at E10.5 (data not shown). We thus decided<br />
74
to study whether any differences <strong>in</strong> collagen content, collagen hydroxylations,<br />
<strong>and</strong> cross-l<strong>in</strong>k<strong>in</strong>g can be seen <strong>in</strong> adult heterozygous Plod2 wt/gt mouse tissues<br />
relative to wild-type.<br />
The amounts <strong>of</strong> collagen <strong>and</strong> 4-hydroxyprol<strong>in</strong>e were unchanged <strong>in</strong> all adult<br />
Plod2 wt/gt tissues studied when compared to wild-type control mice (Figure 10A–<br />
B <strong>in</strong> III). The highest collagen levels <strong>in</strong> wild-type tissues were observed <strong>in</strong> the tail<br />
tendon <strong>and</strong> sk<strong>in</strong>, where the collagen content was 58% <strong>and</strong> 52% <strong>of</strong> the dry weight<br />
<strong>of</strong> the tissue, respectively (Figure 10B <strong>in</strong> III).<br />
The total hydroxylys<strong>in</strong>e content, expressed per triple helix, varies markedly<br />
between tissues as collagen types conta<strong>in</strong> variable amounts <strong>of</strong> hydroxylated<br />
lys<strong>in</strong>es <strong>and</strong> tissues conta<strong>in</strong> different amounts <strong>of</strong> different collagen types.<br />
LH2(long) is responsible for the hydroxylation <strong>of</strong> telopeptide lys<strong>in</strong>es that<br />
participate <strong>in</strong> cross-l<strong>in</strong>k formation <strong>and</strong> there are a maximum <strong>of</strong> two <strong>of</strong> these<br />
lys<strong>in</strong>es per collagen molecule (Knott & Bailey 1998, van der Slot et al. 2003,<br />
Takaluoma et al. 2007). The triple helical area conta<strong>in</strong>s additional lys<strong>in</strong>e residues<br />
that can be hydroxylated most probably by all three LH isoenzymes (Takaluoma<br />
et al. 2007). Two <strong>of</strong> these lys<strong>in</strong>e residues are <strong>in</strong>volved <strong>in</strong> cross-l<strong>in</strong>ks (Knott &<br />
Bailey 1998). Thus, changes <strong>in</strong> the hydroxylation levels <strong>of</strong> the telopeptide lys<strong>in</strong>es<br />
do not dramatically affect the total lys<strong>in</strong>e hydroxylation level. In addition, the<br />
Plod2 wt/gt mice have one functional Plod2 allele left. The highest amount <strong>of</strong> total<br />
hydroxylys<strong>in</strong>e <strong>in</strong> wild-type samples was observed <strong>in</strong> the lung (51 residues / triple<br />
helix) <strong>and</strong> the lowest <strong>in</strong> the sk<strong>in</strong> (18 residues / triple helix) (Figure 10D <strong>in</strong> III). In<br />
the Plod2 wt/gt mouse tissues the total hydroxylys<strong>in</strong>e content was changed<br />
significantly only <strong>in</strong> the calvaria <strong>and</strong> femur, which also normally conta<strong>in</strong> a large<br />
amount <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks. The amount <strong>of</strong> hydroxylys<strong>in</strong>e was decreased<br />
to 85% <strong>in</strong> calvaria <strong>and</strong> to 80% <strong>in</strong> femur when compared to the correspond<strong>in</strong>g<br />
wild-type tissues (Figure 10D <strong>in</strong> III).<br />
As changes <strong>in</strong> the hydroxylation <strong>of</strong> the telopeptide lys<strong>in</strong>e residues <strong>in</strong><br />
collagens affect the cross-l<strong>in</strong>ks <strong>in</strong> BS patients (Bank et al. 1999), mature<br />
pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks were studied from the wild-type <strong>and</strong> Plod2 wt/gt mice. Total<br />
hydroxylys<strong>in</strong>e levels but also the LP <strong>and</strong> HP contents varied markedly between<br />
different tissues (Figure 10D–F <strong>in</strong> III). The highest levels <strong>of</strong> both mature crossl<strong>in</strong>ks<br />
<strong>in</strong> wild-type mouse were observed <strong>in</strong> the lung (LP 0.57 <strong>and</strong> HP 0.51 residues<br />
/ triple helix) <strong>and</strong> aorta (LP 0.12 <strong>and</strong> HP 0.38 residues / triple helix) (Figure 10E–<br />
F <strong>in</strong> III), although the total amount <strong>of</strong> collagen was low <strong>in</strong> these tissues (Figure<br />
10B <strong>in</strong> III). The lowest levels <strong>of</strong> mature cross-l<strong>in</strong>ks <strong>in</strong> wild-type were detected <strong>in</strong><br />
the tail tendon (LP 0.0015 <strong>and</strong> HP 0.0078 residues / triple helix) (Figure 10E–F <strong>in</strong><br />
75
III). The change <strong>in</strong> the LP content between Plod2 wt/wt <strong>and</strong> Plod2 wt/gt was<br />
significant only <strong>in</strong> the calvaria, where the level was decreased <strong>in</strong> the Plod2 wt/gt<br />
sample to 51% relative to the wild-type (Figure 10E <strong>in</strong> III). In contrast, the HP<br />
levels <strong>in</strong> the Plod2 wt/gt mouse tissues were markedly more decreased. The<br />
decrease <strong>in</strong> the HP levels was significant <strong>in</strong> all the Plod2 wt/gt tissues studied except<br />
the aorta (Figure 10F <strong>in</strong> III), which was most severely affected <strong>in</strong> the Plod1 -/-<br />
mouse (Article I <strong>of</strong> this thesis). The most dramatic change <strong>in</strong> the HP was seen <strong>in</strong><br />
the calvaria, where the HP content was 33% (reduced from 0.1 to 0.04 residues /<br />
triple helix) <strong>of</strong> the wild-type. In addition to calvaria, the sk<strong>in</strong> had only 45%<br />
(reduced from 0.01 to 0.005 residues / triple helix) <strong>and</strong> femur 65% (reduced from<br />
0.4 to 0.26 residues / triple helix) <strong>of</strong> the wild-type HP levels. Slightly m<strong>in</strong>or<br />
changes were observed <strong>in</strong> the tail tendon (70% <strong>of</strong> the wild-type value <strong>of</strong> 0.008<br />
residues / triple helix) <strong>and</strong> lung (83% <strong>of</strong> the wild-type value <strong>of</strong> 0.51 residues /<br />
triple helix). Interest<strong>in</strong>gly, <strong>in</strong> general the amount <strong>of</strong> HP cross-l<strong>in</strong>ks was markedly<br />
more affected than the LP cross-l<strong>in</strong>ks <strong>in</strong> the Plod2 wt/gt tissues, although the only<br />
difference between the LP <strong>and</strong> HP is <strong>in</strong> the hydroxylation state <strong>of</strong> the helical<br />
lys<strong>in</strong>e. However, LH2 is suggested to have also helical lysyl hydroxylase activity<br />
(Takaluoma et al. 2007, Article I <strong>in</strong> this thesis).<br />
As long-lived prote<strong>in</strong>s, collagens are susceptible to non-enzymatic, unwanted<br />
cross-l<strong>in</strong>ks <strong>in</strong>volv<strong>in</strong>g reduc<strong>in</strong>g sugars (Saito & Marumo 2010). As lys<strong>in</strong>e residues<br />
are participat<strong>in</strong>g <strong>in</strong> these advanced glycation end-products (AGEs), the effect <strong>of</strong><br />
reduced lysyl hydroxylation on the amount <strong>of</strong> pentosid<strong>in</strong>e, one <strong>of</strong> the AGEs, was<br />
also studied <strong>in</strong> the Plod2 wt/gt tissues. Some changes were seen <strong>in</strong> the femur, for<br />
<strong>in</strong>stance, where the pentosid<strong>in</strong>e level was slightly <strong>in</strong>creased <strong>in</strong> the Plod2 wt/gt mice,<br />
but the results were not statistically significant (Figure 10C <strong>in</strong> III).<br />
76
6 Discussion<br />
6.1 LH1 null mice provide a tool to study tissue-specific<br />
consequences <strong>of</strong> the lack <strong>of</strong> LH1 activity<br />
We have generated a Plod1 -/- mouse l<strong>in</strong>e to study the function <strong>of</strong> LH1 <strong>in</strong> <strong>vivo</strong>. In<br />
addition, we expected that the mouse l<strong>in</strong>e could also serve as a disease model for<br />
the human EDS VIA. This is the first reported animal model for a defect <strong>in</strong> the<br />
gene cod<strong>in</strong>g for LH1, <strong>and</strong> the created mouse l<strong>in</strong>e enables for the first time<br />
detailed exam<strong>in</strong>ation <strong>of</strong> the effects <strong>of</strong> the lack <strong>of</strong> this enzyme <strong>in</strong> various organs<br />
<strong>and</strong> tissues which is difficult, impossible, or unethical to do <strong>in</strong> the case <strong>of</strong> the<br />
EDS VIA patients.<br />
EDS VIA patients suffer already at birth from kyphoscoliosis, muscular<br />
hypotonia <strong>and</strong> jo<strong>in</strong>t laxity, for <strong>in</strong>stance (Ste<strong>in</strong>mann et al. 2002). The LH1 null<br />
mice are viable, <strong>their</strong> breed<strong>in</strong>g is normal, <strong>and</strong> they are not observed to suffer from<br />
kyphoscoliosis. However, they do have muscular hypotonia which becomes more<br />
obvious when the mice are older <strong>and</strong> have <strong>in</strong>creased weight. Although the Plod1 -/-<br />
mice appear to be flaccid dur<strong>in</strong>g h<strong>and</strong>l<strong>in</strong>g <strong>and</strong> have muscle weakness <strong>and</strong> gait<br />
abnormalities <strong>in</strong> situations where strength <strong>and</strong> coord<strong>in</strong>ation is needed, they do not<br />
have an obvious skeletal phenotype like the EDS VIA patients. This can be due to<br />
differences <strong>in</strong> the size <strong>and</strong> anatomy <strong>of</strong> the mouse <strong>and</strong> human as well as the<br />
upright posture <strong>of</strong> humans that exposes the skeleton to forces different from those<br />
<strong>in</strong> the mice. In EDS VIA patients the kyphoscoliosis is also progressive, which is<br />
thought to be due to a supplementary effect <strong>of</strong> the muscle hypotonia <strong>and</strong> jo<strong>in</strong>t<br />
laxity as the vertebral bodies, for <strong>in</strong>stance, are shown to be normal (Ste<strong>in</strong>mann et<br />
al. 2002). Muscular hypotonia, which is connected to a basic ECM-form<strong>in</strong>g<br />
enzyme, is an <strong>in</strong>terest<strong>in</strong>g phenotype which could also be exploited <strong>in</strong> studies <strong>of</strong><br />
other muscular disorders <strong>and</strong> especially the role <strong>of</strong> connective tissue <strong>in</strong> these<br />
diseases. Mutations <strong>in</strong> highly glycosylated <strong>and</strong> thus hydroxylated type VI<br />
collagen cause Ullrich congenital muscular dystrophy <strong>and</strong> the milder Bethlem<br />
myopathy (for a review, see Lampe & Bushby 2005). In addition to muscular<br />
dystrophy, the disorders are characterized by jo<strong>in</strong>t hyperlaxity <strong>and</strong> contractures. A<br />
null mouse for type VI collagen develops a muscular phenotype with variation <strong>in</strong><br />
the muscle fiber diameter <strong>and</strong> fiber necrosis (Bonaldo et al. 1998). Interest<strong>in</strong>gly,<br />
similar to the Plod1 -/- mouse a null mouse l<strong>in</strong>e for type VI collagen has recently<br />
been shown also to possess abnormalities <strong>in</strong> collagen fibril diameter <strong>and</strong> fibril<br />
77
assembly <strong>in</strong> tendons (Izu et al. 2011). Moreover, the mechanical properties <strong>of</strong><br />
tendons were decreased. In contrast to Plod1 -/- mice, however, the collagen fibril<br />
diameter <strong>in</strong> the collagen VI null mice was generally decreased <strong>and</strong> not <strong>in</strong>creased<br />
(Izu et al. 2011).<br />
EDS VIA patients have s<strong>of</strong>t, hyperextensible sk<strong>in</strong> that is prone to abnormal<br />
scarr<strong>in</strong>g. In addition, the patients bruise easily. The sk<strong>in</strong> <strong>of</strong> the Plod1 -/- mice<br />
appeared to be normal, but detailed analysis revealed that the collagen fibrils were<br />
abnormal. The diameter <strong>of</strong> the fibrils varied <strong>in</strong> the Plod1 -/- sk<strong>in</strong> samples <strong>and</strong> <strong>their</strong><br />
shape was <strong>in</strong> part irregular when compared to the wild-type sk<strong>in</strong>. In addition to<br />
cross-l<strong>in</strong>k formation, hydroxylys<strong>in</strong>e residues are known to function as attachment<br />
sites for carbohydrate residues, which <strong>in</strong> turn affect the fibril diameter <strong>and</strong><br />
fibrillogenesis <strong>of</strong> collagen molecules (Myllyharju 2005), support<strong>in</strong>g these<br />
f<strong>in</strong>d<strong>in</strong>gs for the Plod1 -/- sk<strong>in</strong> collagens. The hydroxylys<strong>in</strong>e content <strong>in</strong> EDS VIA<br />
sk<strong>in</strong> is reduced typically to about 5% <strong>of</strong> that <strong>of</strong> healthy controls (Ste<strong>in</strong>mann et al.<br />
2002), whereas <strong>in</strong> Plod1 -/- mouse sk<strong>in</strong> the hydroxylys<strong>in</strong>e content was over 20%.<br />
In addition, the residual LH activity <strong>in</strong> Plod1 -/- mouse sk<strong>in</strong> homogenates was<br />
quite high, 35% <strong>of</strong> the wild-type, while <strong>in</strong> EDS VIA patient fibroblasts the values<br />
vary from 5% to 25% <strong>of</strong> the controls (Yeowell & Walker 2000). Thus, the lack <strong>of</strong><br />
an obvious sk<strong>in</strong> phenotype <strong>in</strong> the Plod1 -/- mice is likely to be because <strong>of</strong> the<br />
higher LH activity <strong>and</strong> hydroxylys<strong>in</strong>e levels <strong>in</strong> the mutant mouse sk<strong>in</strong>. These<br />
differences might also be due to possible variable expressions <strong>of</strong> the LH<br />
isoenzymes <strong>in</strong> the Plod1 -/- mouse <strong>and</strong> EDS VIA sk<strong>in</strong>.<br />
A life-threaten<strong>in</strong>g complication <strong>in</strong> EDS VIA patients is an arterial rupture.<br />
The arteries <strong>of</strong> the Plod1 -/- mice were also found to be structurally weak as some<br />
<strong>of</strong> the LH1 null mice died dur<strong>in</strong>g the night because <strong>of</strong> aortic rupture. This feature<br />
was almost twice as prevalent <strong>in</strong> male as <strong>in</strong> female mice. In histological analysis<br />
the aortic wall <strong>of</strong> apparently healthy Plod1 -/- mice appeared to be generally<br />
normal. Moreover, heart function was also normal. In contrast, <strong>in</strong> necropsy<br />
samples <strong>of</strong> the animals that had died suddenly <strong>and</strong> had hemorrhages <strong>in</strong> the<br />
thoracic or abdom<strong>in</strong>al cavity, dissections <strong>in</strong> the aorta were observed. As <strong>in</strong> the<br />
sk<strong>in</strong>, detailed analysis <strong>of</strong> the aorta by TEM revealed abnormal collagen fibrils <strong>and</strong><br />
also degrad<strong>in</strong>g fibrils even <strong>in</strong> the apparently healthy look<strong>in</strong>g Plod1 -/- mice.<br />
Degenerative changes were also observed <strong>in</strong> the smooth muscle cells <strong>of</strong> the<br />
Plod1 -/- aortas. These observations are <strong>in</strong>terest<strong>in</strong>g because the hydroxylys<strong>in</strong>e level<br />
(63% <strong>of</strong> wild-type) <strong>and</strong> residual LH activity (44% <strong>of</strong> wild-type) <strong>in</strong> the Plod1 -/-<br />
aorta are <strong>in</strong> fact higher than <strong>in</strong> the sk<strong>in</strong> (22% <strong>and</strong> 36%, respectively), which<br />
appeared to be asymptomatic. However, mechanical requirements <strong>in</strong> the aorta are<br />
78
high due to high blood pressure <strong>and</strong> the weakened structures <strong>in</strong> the tissue are<br />
probably predispos<strong>in</strong>g it to ruptures.<br />
In addition to the sk<strong>in</strong> <strong>and</strong> aorta, the hydroxylys<strong>in</strong>e content was also<br />
decreased <strong>in</strong> other Plod1 -/- tissues studied. It is suggested that LH1 is ma<strong>in</strong>ly<br />
responsible for the hydroxylation <strong>of</strong> lys<strong>in</strong>e residues <strong>in</strong> the triple helical region <strong>of</strong><br />
collagens, while LH2(long) exclusively hydroxylates the telopeptides (Eyre et al.<br />
2002, van der Slot et al. 2003, Takaluoma et al. 2007). The exact role <strong>of</strong> LH3 <strong>in</strong><br />
lysyl hydroxylation is still unknown as for example the embryonic lethality <strong>of</strong><br />
LH3 null mice has been shown to be due to the lack <strong>of</strong> its collagen<br />
glycosyltransferase <strong>and</strong> not LH activity (Rautavuoma et al. 2004, Ruotsala<strong>in</strong>en et<br />
al. 2006). Although the Plod1 -/- mice do not have functional LH1, <strong>in</strong> certa<strong>in</strong><br />
tissues, such as the lung, the hydroxylys<strong>in</strong>e level is as high as 86% <strong>of</strong> the wildtype,<br />
while <strong>in</strong> the sk<strong>in</strong> the level was only 22%. The differential hydroxylation<br />
content <strong>in</strong> tissues most likely reflects tissue specificities <strong>of</strong> the LH isoenzymes,<br />
differences <strong>in</strong> <strong>their</strong> amounts <strong>in</strong> different tissues, <strong>and</strong> probably also differential<br />
compensatory mechanisms <strong>of</strong> the LHs. At the mRNA level all LH is<strong>of</strong>orms are<br />
expressed <strong>in</strong> almost all mouse <strong>and</strong> human tissues analyzed (Valtavaara et al. 1997,<br />
Valtavaara et al. 1998, Passoja et al. 1998a, Ruotsala<strong>in</strong>en et al. 1999, Yeowell &<br />
Walker 1999a, Rautavuoma et al. 2004). At the prote<strong>in</strong> level the data is still<br />
limited. Our data also clearly suggest that <strong>in</strong> addition to LH1, the other two<br />
is<strong>of</strong>orms, LH2 <strong>and</strong> LH3, are able to hydroxylate helical lys<strong>in</strong>es to variable extents<br />
<strong>in</strong> different tissues <strong>in</strong> <strong>vivo</strong>.<br />
The mechanical strength <strong>of</strong> tissues depends <strong>in</strong> part on the type <strong>and</strong> amount <strong>of</strong><br />
the exist<strong>in</strong>g collagen cross-l<strong>in</strong>ks. The type <strong>of</strong> the cross-l<strong>in</strong>k, <strong>in</strong> turn, depends on<br />
the hydroxylation state <strong>of</strong> certa<strong>in</strong> triple helical <strong>and</strong> telopeptide lys<strong>in</strong>es. Two<br />
different routes exist for cross-l<strong>in</strong>k formation. The hydroxyallys<strong>in</strong>e pathway leads<br />
to the formation <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks HP <strong>and</strong> LP that predom<strong>in</strong>ate <strong>in</strong> tissues<br />
such as bone <strong>and</strong> cartilage. LP is formed between two telopeptide hydroxylys<strong>in</strong>e<br />
molecules <strong>and</strong> a helical lys<strong>in</strong>e, whereas <strong>in</strong> the HP cross-l<strong>in</strong>k all participat<strong>in</strong>g<br />
lys<strong>in</strong>es are hydroxylated. The allys<strong>in</strong>e pathway, which leads to mechanically<br />
weaker bonds, is common <strong>in</strong> the sk<strong>in</strong> <strong>and</strong> cornea, for <strong>in</strong>stance. (Knott & Bailey<br />
1998, Eyre & Wu 2005). As the cross-l<strong>in</strong>k<strong>in</strong>g pr<strong>of</strong>ile <strong>in</strong> tissues <strong>of</strong> EDS VIA<br />
patients is changed, they have <strong>in</strong>creased ur<strong>in</strong>ary LP levels <strong>and</strong> decreased HP<br />
levels (Pasquali et al. 1994, Pasquali et al. 1995, Ste<strong>in</strong>mann et al. 1995). As was<br />
expected, the amount <strong>of</strong> HP cross-l<strong>in</strong>ks was decreased <strong>in</strong> all the Plod1 -/- mouse<br />
tissues studied, while the amount <strong>of</strong> LP cross-l<strong>in</strong>ks was <strong>in</strong>creased. This supports<br />
the earlier suggestion that LH1 is the ma<strong>in</strong> LH that hydroxylates helical lys<strong>in</strong>es.<br />
79
The total amount <strong>of</strong> HP + LP was <strong>in</strong>creased <strong>in</strong> the Plod1 -/- mice because <strong>of</strong><br />
<strong>in</strong>creased LP that could <strong>in</strong>dicate a response to the changed cross-l<strong>in</strong>k<strong>in</strong>g state.<br />
Both pyrid<strong>in</strong>ol<strong>in</strong>es require telopeptide lys<strong>in</strong>e hydroxylation <strong>and</strong> thus, at least<br />
LH2(long) upregulation could be beh<strong>in</strong>d this response. However, the maturation<br />
<strong>of</strong> cross-l<strong>in</strong>ks to LP <strong>and</strong> HP requires time <strong>and</strong> it is not known whether the rate is<br />
the same for both pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks; this may also <strong>in</strong>fluence the results.<br />
Surpris<strong>in</strong>gly, there was no correlation between the overall hydroxylation level <strong>and</strong><br />
the amount <strong>of</strong> cross-l<strong>in</strong>ks. Although the total hydroxylys<strong>in</strong>e amount <strong>in</strong> the aorta <strong>of</strong><br />
the Plod1 -/- mouse was rather high (63% <strong>of</strong> wild-type), the amount <strong>of</strong> HP was<br />
only 28% <strong>of</strong> that <strong>of</strong> the wild-type, while <strong>in</strong> the tail tendon the total hydroxylys<strong>in</strong>e<br />
content was low (24% <strong>of</strong> wild-type), but the HP value was still 59% <strong>of</strong> the control,<br />
for <strong>in</strong>stance. These results clearly show that LH2 <strong>and</strong>/or LH3 are able to<br />
hydroxylate the triple helical cross-l<strong>in</strong>k site lys<strong>in</strong>es to some extent <strong>in</strong> various<br />
tissues when LH1 is not functional. Moreover, the results demonstrate aga<strong>in</strong> the<br />
complex roles <strong>and</strong> <strong>in</strong>terplay between the LH isoenzymes <strong>in</strong> tissues.<br />
Changes <strong>in</strong> lysyl hydroxylation not only <strong>in</strong>fluence the mechanical strength <strong>of</strong><br />
tissues via cross-l<strong>in</strong>ks but also <strong>in</strong>dividual collagen fibrils as was discussed earlier.<br />
The diameter <strong>and</strong> shape <strong>of</strong> collagen molecules <strong>in</strong> the Plod1 -/- sk<strong>in</strong> <strong>and</strong> aorta were<br />
changed, but not <strong>in</strong> cornea. Many factors may contribute to these characteristics,<br />
but currently they are not well known. However, glycosylation <strong>of</strong> hydroxylys<strong>in</strong>e<br />
residues is known to affect fibrillogenesis <strong>and</strong> fibril diameter (Torre-Blanco et al.<br />
1992, Notbohm et al. 1999, Sricholpech et al. 2011). In the Plod1 -/- mice the<br />
amount <strong>of</strong> hydroxylys<strong>in</strong>e residues available for glycosylation is decreased, which<br />
is likely to affect the glycosylation level as well. Moreover, collagen I fibrils<br />
conta<strong>in</strong> small amounts <strong>of</strong> type III <strong>and</strong> V collagens, which <strong>in</strong> turn are shown to<br />
regulate the thickness <strong>of</strong> fibrils (Kielty & Grant 2002). EDS patients that have a<br />
vascular or classical form <strong>of</strong> the syndrome have mutations <strong>in</strong> these m<strong>in</strong>or<br />
collagens <strong>and</strong> they have abnormal collagen I fibrils (Ste<strong>in</strong>mann et al. 2002). The<br />
null mouse l<strong>in</strong>es for these collagen types are also shown to have abnormal<br />
collagen fibrils (Liu et al. 1997, Wenstrup et al. 2006), as well as mouse l<strong>in</strong>es<br />
lack<strong>in</strong>g small leuc<strong>in</strong>e-rich proteoglycans such as biglycan, decor<strong>in</strong>, fibromodul<strong>in</strong><br />
<strong>and</strong> lumican (Danielson et al. 1997, Chakravarti et al. 1998, Svensson et al. 1999,<br />
Corsi et al. 2002). Thus, it is possible that deficient lysyl hydroxylation <strong>and</strong><br />
glycosylation <strong>of</strong> fibrillar collagens <strong>in</strong> the Plod1 -/- mice may <strong>in</strong>terfere with proper<br />
b<strong>in</strong>d<strong>in</strong>g <strong>of</strong> other prote<strong>in</strong>s, which <strong>in</strong> turn may lead to abnormal collagen fibrils.<br />
Moreover, the Plod1 -/- mice show that the abnormal collagen fibrils observed <strong>in</strong><br />
80
different EDS types can also be caused by a defective collagen-modify<strong>in</strong>g<br />
enzyme <strong>and</strong> not just by mutations <strong>in</strong> the collagen molecules themselves.<br />
6.2 Mutations caus<strong>in</strong>g Bruck syndrome lead to decreased LH2<br />
activity, expla<strong>in</strong><strong>in</strong>g the symptoms <strong>of</strong> diagnosed Bruck<br />
syndrome type 2 patients<br />
BS has been l<strong>in</strong>ked to lysyl hydroxylation <strong>of</strong> telopeptides because <strong>of</strong> the<br />
decreased amount <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks <strong>in</strong> bone collagen (Bank et al. 1999).<br />
Three different mutations found from the PLOD2 gene led to the suggestion that<br />
these mutations cause BS2 <strong>and</strong> that the LH2 is the telopeptide LH (van der Slot et<br />
al. 2003, Ha-V<strong>in</strong>h et al. 2004). As these mutations do not change any <strong>of</strong> the<br />
am<strong>in</strong>o acids that are known to be catalytically critical, the molecular mechanism<br />
caus<strong>in</strong>g the disorder has been unknown. In this study we produced the BS mutant<br />
<strong>and</strong> wild-type human LH2(long) polypeptides <strong>in</strong> <strong>in</strong>sect cells <strong>and</strong> purified <strong>and</strong><br />
analyzed them to underst<strong>and</strong> the molecular pathology <strong>of</strong> the syndrome.<br />
Our data show that all three known BS2 mutations lead to marked reduction<br />
<strong>of</strong> LH2 activity. The most significant decrease <strong>in</strong> the activity was observed <strong>in</strong> the<br />
mutants R594H <strong>and</strong> G597V, whose activity was below 5% <strong>of</strong> that <strong>of</strong> the wild-type<br />
enzyme. These mutations most probably affect proper fold<strong>in</strong>g <strong>of</strong> the polypeptides,<br />
which <strong>in</strong> turn may also affect <strong>their</strong> oligomerization. These effects are likely to<br />
lead to the marked decrease <strong>in</strong> the enzyme activity. All the mutated am<strong>in</strong>o acids<br />
found <strong>in</strong> LH2 <strong>in</strong> the BS2 patients, Arg 594 , Gly 597 <strong>and</strong> Thr 604 , are conserved <strong>in</strong> the<br />
human <strong>and</strong> mouse LH isoenzymes (Valtavaara et al. 1998, Passoja et al. 1998a,<br />
Ruotsala<strong>in</strong>en et al. 1999). LH monomers are shown to consist <strong>of</strong> three doma<strong>in</strong>s.<br />
The C-term<strong>in</strong>al doma<strong>in</strong> <strong>of</strong> all three isoenzymes is responsible for the LH activity<br />
<strong>and</strong> the N-term<strong>in</strong>al doma<strong>in</strong> <strong>in</strong> LH3 possesses glycosyltransferase activity, but the<br />
function <strong>of</strong> the middle doma<strong>in</strong> is unknown. (Rautavuoma et al. 2002). The<br />
residues Arg 594 <strong>and</strong> Gly 597 <strong>in</strong> LH2(long) are located <strong>in</strong> the region that separates<br />
the middle <strong>and</strong> C-term<strong>in</strong>al doma<strong>in</strong>s, which is known to be a protease-sensitive<br />
region. Moreover, C-term<strong>in</strong>al recomb<strong>in</strong>ant LH1 <strong>and</strong> LH3 fragments beg<strong>in</strong>n<strong>in</strong>g<br />
with the residue correspond<strong>in</strong>g to Lys 589 <strong>in</strong> LH2(long) have been shown to be<br />
fully active enzymes (Rautavuoma et al. 2002). Thus, it is likely that the am<strong>in</strong>o<br />
acids Arg 594 <strong>and</strong> Gly 597 are part <strong>of</strong> an unstructured doma<strong>in</strong> boundary that may<br />
function as a l<strong>in</strong>ker. Changes <strong>in</strong> this area may affect its flexibility <strong>and</strong> the fold<strong>in</strong>g<br />
capability <strong>of</strong> the flank<strong>in</strong>g doma<strong>in</strong>s.<br />
81
Fold<strong>in</strong>g <strong>of</strong> the mutant T604I appears to be normal, but the activity <strong>in</strong> st<strong>and</strong>ard<br />
conditions was decreased to under 10% <strong>of</strong> that <strong>of</strong> the wild-type LH2(long). The<br />
change <strong>in</strong> the activity level was caused by a 10-fold <strong>in</strong>crease <strong>in</strong> the Km for<br />
2-oxoglutarate, whereas b<strong>in</strong>d<strong>in</strong>g efficiency <strong>of</strong> the peptide substrate or the other<br />
cosubstrates <strong>of</strong> the reaction was not affected. The maximum activity <strong>of</strong> the T604I<br />
obta<strong>in</strong>ed under a saturat<strong>in</strong>g 2-oxoglutarate concentration was about 30% <strong>of</strong> that <strong>of</strong><br />
the wild-type. The <strong>in</strong>creased Km value for 2-oxoglutarate <strong>in</strong>dicates that b<strong>in</strong>d<strong>in</strong>g <strong>of</strong><br />
this cosubstrate is abnormal <strong>in</strong> the T604I mutant. Moreover, the T604I mutant<br />
was not able to hydroxylate the N-term<strong>in</strong>al telopeptide lys<strong>in</strong>e <strong>in</strong> recomb<strong>in</strong>ant type<br />
I procollagen cha<strong>in</strong>s <strong>in</strong> cellulo. LHs belong to the superfamily <strong>of</strong> 2-oxoglutarate<br />
dioxygenases, which have a catalytic site with a common structure formed by an<br />
8-str<strong>and</strong>ed β-helix core fold. A conserved basic residue <strong>in</strong> the eighth str<strong>and</strong> b<strong>in</strong>ds<br />
the C-5 carboxylate moiety <strong>of</strong> 2-oxoglutarate <strong>and</strong> <strong>in</strong> LH2 the residue is Arg 724 .<br />
(Passoja et al. 1998b, Clifton et al. 2006). This residue is located near the second<br />
Fe 2+ -b<strong>in</strong>d<strong>in</strong>g histid<strong>in</strong>e <strong>in</strong> all known 2-oxoglutarate dioxygenases except the factor<br />
<strong>in</strong>hibit<strong>in</strong>g HIF, an asparag<strong>in</strong>yl hydroxylase <strong>of</strong> HIF (Sch<strong>of</strong>ield et al. 2004). Certa<strong>in</strong><br />
ser<strong>in</strong>e or threon<strong>in</strong>e residues near the basic residue <strong>in</strong> the eighth str<strong>and</strong> provide an<br />
additional <strong>in</strong>teraction with the C-5 carboxylate <strong>of</strong> 2-oxoglutarate <strong>in</strong> many<br />
2-oxoglutarate dioxygenases (Clifton et al. 2006, Koski et al. 2007). Mutation <strong>of</strong><br />
this ser<strong>in</strong>e residue <strong>in</strong> recomb<strong>in</strong>ant human LH1 is shown to reduce the enzyme<br />
activity markedly, although the Km value for 2-oxoglutarate rema<strong>in</strong>ed normal<br />
(Passoja et al. 1998b). Unfortunately the three-dimensional structure <strong>of</strong> LHs is<br />
not known <strong>and</strong> thus the exact structural role <strong>of</strong> Thr 604 <strong>in</strong> the b<strong>in</strong>d<strong>in</strong>g <strong>of</strong><br />
2-oxoglutarate rema<strong>in</strong>s unknown.<br />
The mutant enzymes were not totally <strong>in</strong>active, which may be important for<br />
patients carry<strong>in</strong>g these mutations as dur<strong>in</strong>g this study we observed that LH2 null<br />
mice die early, dur<strong>in</strong>g embryonic development (Article III <strong>in</strong> this thesis). On the<br />
other h<strong>and</strong>, for example <strong>in</strong> EDS VIA LH1 is known to have alternative splic<strong>in</strong>g<br />
pathways to bypass or compensate mutations <strong>and</strong> to reta<strong>in</strong> some enzyme activity<br />
(Yeowell & Walker 2000). However, it is not known whether the residual activity<br />
is enough to enable the survival <strong>of</strong> patients or is LH2 us<strong>in</strong>g similar alternative<br />
splic<strong>in</strong>g system <strong>in</strong> BS2 as LH1 <strong>in</strong> EDS VIA or whether the role <strong>of</strong> LH2 <strong>in</strong> human<br />
development is not as critical as <strong>in</strong> mouse development. However, accord<strong>in</strong>g to<br />
published reports, two affected children carry<strong>in</strong>g the BS mutation T604I deceased<br />
prenatally <strong>and</strong> one child survived only for 2 years (van der Slot et al. 2003). Thus,<br />
this mutation seems to be cl<strong>in</strong>ically the most severe form <strong>of</strong> the reported three<br />
forms (van der Slot et al. 2003, Ha-V<strong>in</strong>h et al. 2004). This is a bit surpris<strong>in</strong>g as<br />
82
the effects <strong>of</strong> the T604I mutation <strong>in</strong> the recomb<strong>in</strong>ant LH2 were not as severe as<br />
those <strong>of</strong> the Arg 594 , Gly 597 mutations.<br />
The type <strong>and</strong> amount <strong>of</strong> collagen cross-l<strong>in</strong>ks exist<strong>in</strong>g <strong>in</strong> the tissue determ<strong>in</strong>e<br />
<strong>in</strong> part its mechanical strength as they stabilize the collagen fibrils. Cross-l<strong>in</strong>ks<br />
are formed via two different routes, the allys<strong>in</strong>e <strong>and</strong> hydroxyallys<strong>in</strong>e routes. The<br />
hydroxylation state <strong>of</strong> the telopeptide cross-l<strong>in</strong>k lys<strong>in</strong>e residues determ<strong>in</strong>es the<br />
route <strong>and</strong> <strong>in</strong> the hydroxyallys<strong>in</strong>e route these lys<strong>in</strong>es are hydroxylated, while <strong>in</strong> the<br />
allys<strong>in</strong>e route they are not. The mature pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks HP <strong>and</strong> LP, which<br />
predom<strong>in</strong>ate <strong>in</strong> tissues requir<strong>in</strong>g strength such as bone <strong>and</strong> cartilage, are formed<br />
via the hydroxyallys<strong>in</strong>e route <strong>and</strong> thus require the presence <strong>of</strong> hydroxylys<strong>in</strong>e <strong>in</strong><br />
the telopeptide cross-l<strong>in</strong>k sites. (Knott & Bailey 1998, Eyre & Wu 2005). BS<br />
patients have a highly abnormal cross-l<strong>in</strong>k<strong>in</strong>g pr<strong>of</strong>ile <strong>in</strong> the bone collagen. The<br />
amount <strong>of</strong> LP <strong>and</strong> HP is dramatically decreased, while the amount <strong>of</strong> allys<strong>in</strong>ederived<br />
cross-l<strong>in</strong>ks is <strong>in</strong>creased. (Bank et al. 1999). Thus, it can be deduced that<br />
the BS2 mutations decrease the LH2 activity to an extent that results <strong>in</strong> changes <strong>in</strong><br />
the cross-l<strong>in</strong>k<strong>in</strong>g pr<strong>of</strong>ile, which causes the cl<strong>in</strong>ical consequences observed <strong>in</strong> the<br />
patients.<br />
6.3 Inactivation <strong>of</strong> mouse Plod2 leads to lethality dur<strong>in</strong>g early<br />
embryonic development<br />
Mutations <strong>in</strong> the PLOD1 gene lead to LH1 deficiency <strong>and</strong> EDS VIA, which is a<br />
severe but not lethal connective tissue disorder. The mouse model for this human<br />
syndrome shows a milder phenotype, but both have an <strong>in</strong>creased risk <strong>of</strong> arterial<br />
rupture, for <strong>in</strong>stance (Article I <strong>in</strong> this thesis). Mice lack<strong>in</strong>g the LH3 enzyme die<br />
dur<strong>in</strong>g embryonic development due to abnormal process<strong>in</strong>g <strong>of</strong> type IV collagen<br />
lead<strong>in</strong>g to basement membrane fragmentation (Rautavuoma et al. 2004,<br />
Ruotsala<strong>in</strong>en et al. 2006). Recently one patient was shown to have compound<br />
heterozygote mutations <strong>in</strong> the PLOD3 gene (Salo et al. 2008). The syndrome is<br />
severe with symptoms overlapp<strong>in</strong>g several connective tissue disorders <strong>and</strong> the<br />
patient also has a stillborn sibl<strong>in</strong>g. Mutations <strong>in</strong> the gene encod<strong>in</strong>g LH2 cause<br />
BS2, where patients have severe skeletal problems, but it is usually not lethal. In<br />
this study we have generated a novel LH2 null mouse l<strong>in</strong>e from an ES cell l<strong>in</strong>e, <strong>in</strong><br />
which the Plod2 gene is disrupted by a gene-trap <strong>in</strong>sertion cassette.<br />
As the human BS2 is not lethal, it was surpris<strong>in</strong>g that LH2 null mice died<br />
dur<strong>in</strong>g embryonic development, after E10.5. However, our previous <strong>in</strong> vitro<br />
studies showed that recomb<strong>in</strong>ant LH2 enzymes carry<strong>in</strong>g the mutations identified<br />
83
<strong>in</strong> BS2 patients are not totally <strong>in</strong>active <strong>and</strong>, thus, the residual LH2 activity may be<br />
<strong>in</strong>dispensable for the patients (Article II <strong>in</strong> this thesis). LH2 null embryos<br />
developed seem<strong>in</strong>gly normally until E10.5, but died after this without any<br />
obvious signs <strong>of</strong> growth retardation. However, about a fifth <strong>of</strong> the null embryos<br />
were observed to have accumulated blood <strong>in</strong>side the fetal membranes while the<br />
heart was still contract<strong>in</strong>g. This could <strong>in</strong>dicate problems <strong>in</strong> the cardiovascular<br />
system or the placenta.<br />
BS2 patients are not reported to have direct cardiovascular symptoms.<br />
However, <strong>in</strong> Plod2 gt/gt embryos some changes were seen <strong>in</strong> the endothelium <strong>of</strong><br />
aorta, which could affect the structural resistance <strong>of</strong> the tissue. The cause <strong>of</strong> death<br />
<strong>of</strong> all BS patients is not available, however. Two children hav<strong>in</strong>g a BS2 mutation<br />
deceased prenatally <strong>and</strong> one child survived for 2 years (van der Slot et al. 2003).<br />
Because <strong>of</strong> the limited amount <strong>of</strong> published cl<strong>in</strong>ical data about these patients, the<br />
possible role <strong>of</strong> the placenta, for <strong>in</strong>stance, cannot be excluded. The placenta <strong>of</strong><br />
mouse <strong>and</strong> human are fairly different, but they also have considerable similarities<br />
(reviewed <strong>in</strong> Rossant & Cross 2001). Although the <strong>in</strong>terest <strong>in</strong> the development <strong>of</strong><br />
placenta <strong>and</strong> its defects is <strong>in</strong>creas<strong>in</strong>g, knowledge about it is still extremely limited<br />
<strong>and</strong> far beh<strong>in</strong>d the knowledge <strong>of</strong> embryonic development. As a tissue orig<strong>in</strong>at<strong>in</strong>g<br />
from a differentiat<strong>in</strong>g embryo, it is affected by the same genetic defects as the<br />
embryo itself (Rossant & Cross 2001). The mutant embryo may die before the<br />
phenotype would be apparent if the gene is required <strong>in</strong> the development <strong>and</strong><br />
function <strong>of</strong> the placenta. LH2 is widely expressed <strong>in</strong> the embryo <strong>and</strong> also <strong>in</strong> the<br />
extraembryonic tissues. On the embryonic side <strong>of</strong> the placenta LH2 is expressed<br />
<strong>in</strong> the chorionic plate <strong>and</strong> mesenchymal tissue surround<strong>in</strong>g embryonic vessels,<br />
where the maternal <strong>and</strong> fetal blood are <strong>in</strong> close proximity to each other <strong>and</strong> the<br />
tissue between the compartments must be th<strong>in</strong> but firm. Moreover, the role <strong>of</strong> the<br />
placenta sets also certa<strong>in</strong> other structural requirements for the tissue, such as<br />
substantial blood flow <strong>in</strong> the chorionic plate-attached umbilical cord (Inman &<br />
Downs 2007). The cause <strong>of</strong> death <strong>of</strong> the null embryos could also be a collapse <strong>in</strong><br />
these tissue structures lead<strong>in</strong>g to hemorrhage. The major fibril-form<strong>in</strong>g collagens<br />
<strong>in</strong> the placenta are type I, III <strong>and</strong> V collagens (Iwahashi et al. 1996, Arai &<br />
Nishiyama 2007), but there is a limited amount <strong>of</strong> data available about the<br />
collagen types <strong>in</strong> the chorionic plate at different developmental stages. Type I, IV<br />
<strong>and</strong> V collagens, but not type III collagen, were expressed <strong>in</strong> the chorionic plate<br />
<strong>of</strong> the E10.5 embryos, but clear differences were not observed between the<br />
Plod2 wt/wt <strong>and</strong> Plod2 gt/gt tissues. Interest<strong>in</strong>gly, immun<strong>of</strong>luorescence sta<strong>in</strong><strong>in</strong>g <strong>of</strong><br />
collagens III <strong>and</strong> V but not type I was slightly weaker at least <strong>in</strong> the yolk sac <strong>of</strong><br />
84
the Plod2 gt/gt embryos than <strong>in</strong> the wild-type. The size <strong>of</strong> the mouse placenta is<br />
<strong>in</strong>creas<strong>in</strong>g markedly between E10–E14 (Arai & Nishiyama 2007), <strong>and</strong> thus the<br />
tissue must come up with the fast-chang<strong>in</strong>g mechanical requirements. Thus,<br />
mechanically <strong>in</strong>adequate collagenous structures may break down dur<strong>in</strong>g the fast<br />
development. As LH2 is required for the formation <strong>of</strong> appropriate ECM, it may be<br />
necessary for the mechanical <strong>in</strong>tegrity <strong>of</strong> the placenta.<br />
Rodents have a Reichert’s membrane (RM) <strong>in</strong> the placenta, between the<br />
parietal endoderm cells <strong>and</strong> trophoblast cells, that is lack<strong>in</strong>g <strong>in</strong> humans (Gardner<br />
1983). The membrane conta<strong>in</strong>s high levels <strong>of</strong> type IV collagen <strong>and</strong> collagen<br />
α1(IV)2α2(IV) knock-out mice are reported to die between E10.5–E11.5 because<br />
<strong>of</strong> the failure <strong>of</strong> this membrane (Pöschl et al. 2004). Moreover, LH3 knock-out<br />
mice die at E9.5 because <strong>of</strong> abnormal collagen IV synthesis. The parietal<br />
endoderm cells produc<strong>in</strong>g the RM as well as the mesodermal part <strong>of</strong> the visceral<br />
yolk sac have strong LH2 expression. Thus, we thought that the LH2 null mice<br />
may also have a failure <strong>in</strong> the RM, but a clear defect was not found. Although no<br />
clear abnormalities were observed, a possible role <strong>of</strong> the membrane failure <strong>in</strong> the<br />
phenotype cannot be fully excluded. In addition, the mesodermal part <strong>of</strong> the<br />
visceral yolk sac <strong>of</strong> Plod2 gt/gt embryos seemed to have sites that appeared loose.<br />
These loosened parts were partly surround<strong>in</strong>g vitell<strong>in</strong>e vessels <strong>and</strong> a collapse <strong>in</strong><br />
this site may also be the cause <strong>of</strong> vessel rupture lead<strong>in</strong>g to the observed bleed<strong>in</strong>g<br />
between the yolk sac <strong>and</strong> the amnion. Type III collagen was found to be expressed<br />
<strong>in</strong> the endodermal part <strong>and</strong> types I <strong>and</strong> V <strong>in</strong> the mesodermal side <strong>of</strong> the yolk sac at<br />
E10.5. Interest<strong>in</strong>gly, as <strong>in</strong>dicated earlier, immun<strong>of</strong>luorescence sta<strong>in</strong><strong>in</strong>g showed<br />
that the amount <strong>of</strong> type V <strong>and</strong> III but not type I collagen may be slightly lower <strong>in</strong><br />
the yolk sac <strong>of</strong> the Plod2 gt/gt embryos. Moreover, collagen V knock-out mice are<br />
shown to have a similar hemorrhage phenotype as the LH2 null mice at the time<br />
<strong>of</strong> <strong>their</strong> death at E10 (Wenstrup et al. 2004). Generally this could <strong>in</strong>dicate<br />
<strong>in</strong>adequate support <strong>in</strong> the LH2 null tissues, especially for the blood vessels.<br />
Abundant expression <strong>of</strong> type I collagen beg<strong>in</strong>s <strong>in</strong> mice at E12 <strong>and</strong> the Mov 13<br />
mice that cannot synthesize the α1(I) collagen polypeptide die soon after this<br />
thorugh rupture <strong>of</strong> the major blood vessels <strong>of</strong> the embryo (Schnieke et al. 1983,<br />
Löhler et al. 1984). Type III collagen expression <strong>in</strong> mouse is even stronger than<br />
that <strong>of</strong> type I between E8.5 – E12.5 (Niederreither et al. 1994). About 10% <strong>of</strong><br />
collagen III knock-out mice survive to adulthood, but have a much shorter<br />
lifespan relative to the wild-type (Liu et al. 1997). The major cause <strong>of</strong> death <strong>in</strong><br />
collagen III knock-out mice is blood vessel ruptures. Interest<strong>in</strong>gly, as <strong>in</strong>dicated<br />
above, collagen V knock-out mice are reported to have a similar hemorrhage<br />
85
phenotype as the LH2 null mice (Wenstrup et al. 2004). Moreover, collagen V<br />
expression <strong>in</strong>creases markedly at E10.5 <strong>in</strong> mice (Roulet et al. 2007). Although the<br />
site <strong>of</strong> vessel rupture or the exact cause for that rema<strong>in</strong>ed unknown <strong>in</strong> the collagen<br />
V null mice, type V collagen was shown to be required for proper collagen fibril<br />
formation. Type V collagen participates as a quantitatively m<strong>in</strong>or component to<br />
the formation <strong>of</strong> type I collagen fibrils <strong>and</strong> the lack <strong>of</strong> it prevents the fibril<br />
formation <strong>in</strong> the mesenchyme almost completely <strong>and</strong> the formed fibrils look<br />
abnormal (Wenstrup et al. 2004). The phenotype <strong>of</strong> the collagen V null mouse<br />
l<strong>in</strong>e is surpris<strong>in</strong>gly severe, as the phenotype <strong>of</strong> the heterozygous mice resembles<br />
the classical type <strong>of</strong> Ehlers-Danlos syndrome (Wenstrup et al. 2004). The<br />
Plod2 gt/gt embryos were nevertheless able to synthesize collagen fibrils.<br />
As a telopeptide LH, LH2 has a role <strong>in</strong> the synthesis <strong>of</strong> several collagen types.<br />
Lack <strong>of</strong> the enzyme may not prevent the synthesis <strong>of</strong> collagen molecules as they<br />
were seen <strong>in</strong> the extracellular space, but it may either affect the process<strong>in</strong>g time,<br />
quality <strong>and</strong>/or the strength <strong>of</strong> the assembled structures as <strong>in</strong>dicated earlier.<br />
Moreover, hydroxylys<strong>in</strong>e residues at the triple helical regions <strong>of</strong> collagens serve<br />
as attachment sites for carbohydrates <strong>and</strong> the exact contribution <strong>of</strong> LH2 to the<br />
hydroxylation <strong>of</strong> these lys<strong>in</strong>es is currently unknown. Thus, it is possible that lack<br />
<strong>of</strong> LH2 may affect the glycosylation status <strong>of</strong> collagens. It is probable that the<br />
lack <strong>of</strong> LH2 not only affects a specific collagen target but rather decreases the<br />
strength <strong>of</strong> ECM structures <strong>in</strong> general by chang<strong>in</strong>g the cross-l<strong>in</strong>k type.<br />
Underhydroxylation <strong>of</strong> the telopeptide lys<strong>in</strong>es prevents the adequate formation <strong>of</strong><br />
trivalent, mature, pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks LP <strong>and</strong> HP, but the telopeptide lysyl<br />
hydroxylation also determ<strong>in</strong>es the type <strong>of</strong> immature divalent cross-l<strong>in</strong>k. At E10.5<br />
mature collagen cross-l<strong>in</strong>ks are not formed yet. Thus, divalent forms most<br />
probably predom<strong>in</strong>ate at this developmental stage, suggest<strong>in</strong>g the importance <strong>of</strong><br />
the immature forms <strong>of</strong> cross-l<strong>in</strong>ks <strong>in</strong> develop<strong>in</strong>g tissues. Furthermore, our results<br />
show that the compensatory effect <strong>of</strong> the other two LH is<strong>of</strong>orms is not significant<br />
when LH2 activity is lack<strong>in</strong>g. Thus, the embryonic lethality <strong>of</strong> LH2 null mice is<br />
most probably caused by the absence <strong>of</strong> the telopeptide lysyl hydroxylase activity.<br />
The dramatic consequence <strong>of</strong> the lack <strong>of</strong> one collagen-modify<strong>in</strong>g enzyme is<br />
<strong>in</strong>terest<strong>in</strong>g as, for <strong>in</strong>stance, the type I collagen null Mov 13 mice <strong>and</strong> collagen V<br />
knock-out mice die only a moment earlier or slightly later <strong>and</strong> the type III<br />
collagen knock-out mice may survive even after birth (Schnieke et al. 1983, Liu<br />
et al. 1997, Wenstrup et al. 2004).<br />
BS patients have skeletal problems such as osteoporosis, fragile bones,<br />
congenital jo<strong>in</strong>t contractures, progressive kyphoscoliosis <strong>and</strong> short stature. As<br />
86
<strong>in</strong>dicated earlier, unfortunately the E10.5 embryos are too young for the cross-l<strong>in</strong>k<br />
analysis as maturation is still ongo<strong>in</strong>g, <strong>and</strong> due to the early embryonic lethality<br />
the skeletal consequences also rema<strong>in</strong> unknown. However, it is <strong>in</strong>terest<strong>in</strong>g that the<br />
LH2 heterozygous mice also have significant changes <strong>in</strong> the cross-l<strong>in</strong>k pattern,<br />
particularly <strong>in</strong> bone, <strong>in</strong>dicat<strong>in</strong>g that the amount <strong>of</strong> the LH2 needed <strong>in</strong> the process<br />
is rather high. Moreover, reduction <strong>in</strong> the amount <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks was<br />
more marked <strong>in</strong> the heterozygous calvaria than the femur. These bone types differ<br />
from each other <strong>in</strong> <strong>their</strong> orig<strong>in</strong> <strong>and</strong> mechanism <strong>of</strong> ossification, which <strong>in</strong> turn affect<br />
<strong>their</strong> composition (Chung et al. 2004, van den Bos et al. 2008). The flat bone<br />
calvaria orig<strong>in</strong>ates from the ectodermal neural crest, while the femur, represent<strong>in</strong>g<br />
a long bone, is <strong>of</strong> mesodermal orig<strong>in</strong> (Chung et al. 2004). Flat bones are formed<br />
by straight <strong>in</strong>tramembranous ossification, where the cells differentiate directly<br />
<strong>in</strong>to bone-form<strong>in</strong>g osteoblasts that produce a matrix that is particularly rich <strong>in</strong><br />
type I collagen. Long bones, <strong>in</strong> turn, are ossified via a cartilage model, by<br />
endochondral ossification, which beg<strong>in</strong>s by a condensation <strong>of</strong> mesenchymal cells<br />
to chondrocytes. (Kronenberg 2003). The function <strong>of</strong> the long bones is to provide<br />
support for the body, whereas the role <strong>of</strong> calvaria, for <strong>in</strong>stance, is more protective<br />
than supportive. Our results <strong>in</strong>dicate that a decrease <strong>in</strong> LH2 activity may be less<br />
harmful <strong>in</strong> long bones. They are ossified via a cartilage model <strong>and</strong> LH2 may have<br />
more time to hydroxylate the telopeptide lys<strong>in</strong>es when cartilage is replaced slowly<br />
by bone. The straight ossification process <strong>in</strong> the calvaria may <strong>in</strong> turn be too fast<br />
for the reduced amount <strong>of</strong> LH2 to keep up with <strong>and</strong> thus the collagen crossl<strong>in</strong>k<strong>in</strong>g<br />
is more affected. However, bone tissue is renew<strong>in</strong>g constantly <strong>and</strong> an<br />
<strong>in</strong>terplay between the contribut<strong>in</strong>g components is important. The prote<strong>in</strong><br />
composition <strong>in</strong> different bone types differs considerably, <strong>and</strong> for example the<br />
collagen amount <strong>in</strong> flat bones is higher than <strong>in</strong> long bones <strong>and</strong> the collagens are<br />
also more soluble (van den Bos et al. 2008). Interest<strong>in</strong>gly, the levels <strong>of</strong> certa<strong>in</strong><br />
growth factors such as TGF-β are higher <strong>in</strong> flat bones than <strong>in</strong> long bones<br />
(F<strong>in</strong>kelman et al. 1994) <strong>and</strong>, furthermore, TGF-β is known to <strong>in</strong>crease the<br />
expression <strong>of</strong> LH2(long) <strong>and</strong> the formation <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks (van der<br />
Slot et al. 2005). Thus, <strong>in</strong> addition to the time available for lysyl hydroxylation <strong>in</strong><br />
different ossification processes, changes <strong>in</strong> the <strong>in</strong>terplay <strong>of</strong> the contribut<strong>in</strong>g<br />
components <strong>of</strong> bone development may affect the cross-l<strong>in</strong>k<strong>in</strong>g state <strong>of</strong> bones with<br />
reduced amounts <strong>of</strong> LH2. Moreover, both pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks require<br />
telopeptide hydroxylys<strong>in</strong>e, i.e. LH2(long), for <strong>their</strong> formation, <strong>and</strong> the difference<br />
between the LP <strong>and</strong> HP cross-l<strong>in</strong>k is only <strong>in</strong> the hydroxylation state <strong>of</strong> the helical<br />
lys<strong>in</strong>e. Thus, it was <strong>in</strong>terest<strong>in</strong>g that the amount <strong>of</strong> HP was more severely affected<br />
87
than the LP <strong>in</strong> the LH2 heterozygous mouse tissues. This <strong>in</strong>dicates that LH2 is<br />
also <strong>in</strong>volved <strong>in</strong> the hydroxylation <strong>of</strong> the helical lys<strong>in</strong>e residue participat<strong>in</strong>g <strong>in</strong><br />
cross-l<strong>in</strong>k formation.<br />
88
7 Conclusions <strong>and</strong> future prospects<br />
Today ECM is known to be a highly dynamic part <strong>of</strong> the human body <strong>and</strong> not just<br />
a scaffold for cells <strong>and</strong> tissues. Functional characteristics <strong>of</strong> collagens, the most<br />
common prote<strong>in</strong> component <strong>of</strong> the ECM, are determ<strong>in</strong>ed dur<strong>in</strong>g collagen<br />
biosynthesis. In addition to collagen gene mutations, failures <strong>in</strong> the enzymes<br />
required for collagen biosynthesis <strong>and</strong> an imbalance <strong>of</strong> regulat<strong>in</strong>g factors, for<br />
<strong>in</strong>stance, can lead to cl<strong>in</strong>ical consequences. Besides rare hereditary connective<br />
tissue disorders the role <strong>of</strong> ECM <strong>and</strong> its changes are now also emphasized <strong>in</strong><br />
common disease states such as fibrosis <strong>and</strong> cancer that cause a large economic<br />
burden on the health care system.<br />
In this study we generated an LH1 null mouse l<strong>in</strong>e, which is the first animal<br />
model for human EDS VIA. Although the mice had generally milder symptoms,<br />
the LH1 null mice <strong>and</strong> EDS VIA patients share the same life-threaten<strong>in</strong>g<br />
phenotype i.e. risk <strong>of</strong> arterial ruptures. Structural weakness <strong>of</strong> arterial walls can<br />
be expla<strong>in</strong>ed by <strong>in</strong>adequate helical lys<strong>in</strong>e hydroxylation <strong>and</strong> the accompanied<br />
changes <strong>in</strong> the collagen cross-l<strong>in</strong>k pattern. Interest<strong>in</strong>gly, <strong>in</strong> LH1 null mice the<br />
correlation between the hydroxylys<strong>in</strong>e content <strong>and</strong> the amount <strong>of</strong> HP cross-l<strong>in</strong>ks,<br />
which requires both telopeptide <strong>and</strong> helical hydroxylys<strong>in</strong>e, was <strong>in</strong>consistent. Thus,<br />
<strong>in</strong> <strong>vivo</strong> the two other is<strong>of</strong>orms, LH2 <strong>and</strong> LH3, most probably can compensate for<br />
the lack <strong>of</strong> LH1 activity, but the exact mechanism rema<strong>in</strong>s to be established.<br />
Interest<strong>in</strong>gly, <strong>in</strong> LH2 null mice we observed that LH1 <strong>and</strong> LH3 are not able to<br />
compensate for at least the telopeptide LH activity. LH2 null mice died early<br />
dur<strong>in</strong>g development, which was surpris<strong>in</strong>g as the human disorder BS2 is<br />
generally not lethal. However, our <strong>in</strong> vitro studies revealed that although the<br />
known LH2 mutations <strong>in</strong> BS2 patients are destructive <strong>and</strong> able to expla<strong>in</strong> the<br />
cl<strong>in</strong>ical symptoms, they do not lead to complete <strong>in</strong>activation <strong>of</strong> the enzyme. The<br />
known mutations affect the fold<strong>in</strong>g <strong>of</strong> the enzyme or <strong>in</strong>terfere with the b<strong>in</strong>d<strong>in</strong>g <strong>of</strong><br />
a cosubstrate. Embyonic lethality <strong>of</strong> LH2 null mice raises the question whether a<br />
total lack <strong>of</strong> LH2 would be lethal also <strong>in</strong> humans. The exact cause <strong>of</strong> death <strong>of</strong><br />
LH2 null mice rema<strong>in</strong>s unknown, but the results suggest that <strong>in</strong>adequate<br />
telopeptide lysyl hydroxylation <strong>and</strong> thus formation <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong> cross-l<strong>in</strong>ks may<br />
lead to a general weakness <strong>and</strong> collapse <strong>of</strong> tissue structures, particularly <strong>in</strong> the<br />
vessel walls. Unfortunately, analysis <strong>of</strong> collagen cross-l<strong>in</strong>ks from LH2 null mice<br />
was impossible due to the early lethality, but the remarkable role <strong>of</strong> LH2 <strong>in</strong> crossl<strong>in</strong>k<br />
formation was seen with heterozygous mice that had surpris<strong>in</strong>gly severe<br />
changes <strong>in</strong> the cross-l<strong>in</strong>ks, especially <strong>in</strong> the bone. Both mouse l<strong>in</strong>es generated <strong>and</strong><br />
89
analyzed here emphasize the importance <strong>of</strong> LH1 <strong>and</strong> LH2 <strong>in</strong> proper cross-l<strong>in</strong>k<br />
formation.<br />
Although the LH2 gene-trap mouse l<strong>in</strong>e generated here has provided new<br />
data about LH2 <strong>and</strong> emphasized its role <strong>in</strong> mouse development, the embryonic<br />
lethal phenotype restricts studies <strong>of</strong> the contribution <strong>of</strong> LH2 to common<br />
pathological situations such as fibrosis. It would be important also to know the<br />
contribution <strong>of</strong> LH2 <strong>in</strong> tumorigenesis as the role <strong>of</strong> some other 2-oxoglutarate<br />
dioxygenases <strong>in</strong> cancer is remarkable <strong>and</strong> because <strong>in</strong>creased expression levels <strong>of</strong><br />
LH2 is connected to hypoxia <strong>and</strong> cancer. Thus, a conditional LH2 knock-out<br />
mouse l<strong>in</strong>e could provide answers to some <strong>of</strong> these questions.<br />
90
References<br />
Ak<strong>in</strong>s ML, Luby-Phelps K, Bank RA & Mahendroo M (2011) Cervical s<strong>of</strong>ten<strong>in</strong>g dur<strong>in</strong>g<br />
pregnancy: regulated changes <strong>in</strong> collagen cross-l<strong>in</strong>k<strong>in</strong>g <strong>and</strong> composition <strong>of</strong><br />
matricellular prote<strong>in</strong>s <strong>in</strong> the mouse. Biol Reprod 84(5): 1053–1062.<br />
Alanay Y, Avaygan H, Camacho N, Ut<strong>in</strong>e GE, Boduroglu K, Aktas D, Alikasifoglu M,<br />
Tuncbilek E, Orhan D, Bakar FT, Zabel B, Superti-Furga A, Bruckner-Tuderman L,<br />
Curry CJ, Pyott S, Byers PH, Eyre DR, Baldridge D, Lee B, Merrill AE, Davis EC,<br />
Cohn DH, Akarsu N & Krakow D (2010) Mutations <strong>in</strong> the gene encod<strong>in</strong>g the RER<br />
prote<strong>in</strong> FKBP65 cause autosomal-recessive osteogenesis imperfecta. Am J Hum<br />
Genet 86(4):551–559. Erratum <strong>in</strong> Am J Hum Genet 87(4): 572–573.<br />
Alanay Y & Krakow D (2010) Response to Shaheen et al. 2010. Am J Hum Genet 87(2):<br />
308.<br />
Andiran N, Alikasifoglu A, Alanay Y & Yordam N (2008) Cyclic pamidronate treatment<br />
<strong>in</strong> Bruck syndrome: proposal <strong>of</strong> a new modality <strong>of</strong> treatment. Pediatr Int 50(6): 836–<br />
838.<br />
Arai KY & Nishiyama T (2007) Developmental changes <strong>in</strong> extracellular matrix messenger<br />
RNAs <strong>in</strong> the mouse placenta dur<strong>in</strong>g the second half <strong>of</strong> pregnancy: possible factors<br />
<strong>in</strong>volved <strong>in</strong> the regulation <strong>of</strong> placental extracellular matrix expression. Biol Reprod<br />
77(6): 923–933.<br />
Armstrong LC & Last JA (1995) Rat lysyl hydroxylase: molecular clon<strong>in</strong>g, mRNA<br />
distribution <strong>and</strong> expression <strong>in</strong> a baculovirus system. Biochim Biophys Acta 1264(1):<br />
93–102.<br />
Avery NC & Bailey AJ (2005) Enzymic <strong>and</strong> non-enzymic cross-l<strong>in</strong>k<strong>in</strong>g mechanisms <strong>in</strong><br />
relation to turnover <strong>of</strong> collagen: relevance to ag<strong>in</strong>g <strong>and</strong> exercise. Sc<strong>and</strong> J Med Sci<br />
Sports 15(4): 231–240.<br />
Badylak SF, Freytes DO & Gilbert TW (2009) Extracellular matrix as a biological scaffold<br />
material: Structure <strong>and</strong> function. Acta Biomater 5(1): 1–13.<br />
Bailey AJ (2002) Changes <strong>in</strong> bone collagen with age <strong>and</strong> disease. J Musculoskelet<br />
Neuronal Interact 2(6): 529–531.<br />
Bailey AJ, Paul RG & Knott L (1998) Mechanisms <strong>of</strong> maturation <strong>and</strong> age<strong>in</strong>g <strong>of</strong> collagen.<br />
Mech Age<strong>in</strong>g Dev 106(1–2): 1–56.<br />
Bank RA, Rob<strong>in</strong>s SP, Wijmenga C, Breslau-Siderius LJ, Bardoel AF, van der Sluijs HA,<br />
Pruijs HE & TeKoppele JM (1999) Defective collagen crossl<strong>in</strong>k<strong>in</strong>g <strong>in</strong> bone, but not <strong>in</strong><br />
ligament or cartilage, <strong>in</strong> Bruck syndrome: <strong>in</strong>dications for a bone-specific telopeptide<br />
lysyl hydroxylase on chromosome 17. Proc Natl Acad Sci U S A 96(3): 1054–1058.<br />
Baraniak AP, Chen JR & Garcia-Blanco MA (2006) Fox-2 mediates epithelial cell-specific<br />
fibroblast growth factor receptor 2 exon choice. Mol Cell Biol 26(4): 1209–1222.<br />
Barnes MJ, Constable BJ, Morton LF & Royce PM (1974) Age-related variations <strong>in</strong><br />
hydroxylation <strong>of</strong> lys<strong>in</strong>e <strong>and</strong> prol<strong>in</strong>e <strong>in</strong> collagen. Biochem J 139(2): 461–468.<br />
Basel D & Ste<strong>in</strong>er RD (2009) Osteogenesis imperfecta: recent f<strong>in</strong>d<strong>in</strong>gs shed new light on<br />
this once well-understood condition. Genet Med 11(6): 375–385.<br />
91
Baumann M, Giunta C, Krabichler B, Rüschendorf F, Zoppi N, Colombi M, Bittner RE,<br />
Quijano-Roy S, Muntoni F, Cirak S, Schreiber G, Zou Y, Hu Y, Romero NB, Carlier<br />
RY, Amberger A, Deutschmann A, Straub V, Rohrbach M, Ste<strong>in</strong>mann B, Rostásy K,<br />
Karall D, Bönnemann CG, Zschocke J & Fauth C (2012) Mutations <strong>in</strong> FKBP14 Cause<br />
a Variant <strong>of</strong> Ehlers-Danlos Syndrome with Progressive Kyphoscoliosis, Myopathy,<br />
<strong>and</strong> Hear<strong>in</strong>g Loss. Am J Hum Genet 90: 1–16.<br />
Berg C, Geipel A, Noack F, Smrcek J, Krapp M, Germer U, Bender G & Gembruch U<br />
(2005) Prenatal diagnosis <strong>of</strong> Bruck syndrome. Prenat Diagn 25(7): 535–538.<br />
Blacks<strong>in</strong> MF, Pletcher BA & David M (1998) Osteogenesis imperfecta with jo<strong>in</strong>t<br />
contractures: bruck syndrome. Pediatr Radiol 28(2): 117–119.<br />
Bonaldo P, Braghetta P, Zanetti M, Piccolo S, Volp<strong>in</strong> D & Bressan GM (1998) Collagen<br />
VI deficiency <strong>in</strong>duces early onset myopathy <strong>in</strong> the mouse: an animal model for<br />
Bethlem myopathy. Hum Mol Genet 7(13): 2135–2140.<br />
Brady AF & Patton MA (1997) Osteogenesis imperfecta with arthrogryposis multiplex<br />
congenita (Bruck syndrome)--evidence for possible autosomal recessive <strong>in</strong>heritance.<br />
Cl<strong>in</strong> Dysmorphol 6(4): 329–336.<br />
Brenner RE, Vetter U, Stöss H, Müller PK & Teller WM (1993) Defective collagen fibril<br />
formation <strong>and</strong> m<strong>in</strong>eralization <strong>in</strong> osteogenesis imperfecta with congenital jo<strong>in</strong>t<br />
contractures (Bruck syndrome). Eur J Pediatr 152(6): 505–508.<br />
Breslau-Siderius EJ, Engelbert RH, Pals G & van der Sluijs JA (1998) Bruck syndrome: a<br />
rare comb<strong>in</strong>ation <strong>of</strong> bone fragility <strong>and</strong> multiple congenital jo<strong>in</strong>t contractures. J Pediatr<br />
Orthop B 7(1): 35–38.<br />
Br<strong>in</strong>ckmann J, Açil Y, Feshchenko S, Katzer E, Brenner R, Kulozik A & Kügler S (1998)<br />
Ehlers-Danlos syndrome type VI: lysyl hydroxylase deficiency due to a novel po<strong>in</strong>t<br />
mutation (W612C). Arch Dermatol Res 290(4): 181–186.<br />
Br<strong>in</strong>ckmann J, Açil Y, Tronnier M, Notbohm H, Bätge B, Schmeller W, Koch MH, Müller<br />
PK & Wolff HH (1996) Altered x-ray diffraction pattern is accompanied by a change<br />
<strong>in</strong> the mode <strong>of</strong> cross-l<strong>in</strong>k formation <strong>in</strong> lipodermatosclerosis. J Invest Dermatol 107(4):<br />
589–592.<br />
Br<strong>in</strong>ckmann J, Kim S, Wu J, Re<strong>in</strong>hardt DP, Batmunkh C, Metzen E, Notbohm H, Bank<br />
RA, Krieg T & Hunzelmann N (2005) Interleuk<strong>in</strong> 4 <strong>and</strong> prolonged hypoxia <strong>in</strong>duce a<br />
higher gene expression <strong>of</strong> lysyl hydroxylase 2 <strong>and</strong> an altered cross-l<strong>in</strong>k pattern:<br />
important pathogenetic steps <strong>in</strong> early <strong>and</strong> late stage <strong>of</strong> systemic scleroderma? Matrix<br />
Biol 24(7): 459–468.<br />
Br<strong>in</strong>ckmann J, Neess CM, Gaber Y, Sobhi H, Notbohm H, Hunzelmann N, Fietzek PP,<br />
Müller PK, Risteli J, Gebker R & Scharffetter-Kochanek K (2001) Different pattern <strong>of</strong><br />
collagen cross-l<strong>in</strong>ks <strong>in</strong> two sclerotic sk<strong>in</strong> diseases: lipodermatosclerosis <strong>and</strong><br />
circumscribed scleroderma. J Invest Dermatol 117(2): 269–273.<br />
Br<strong>in</strong>ckmann J, Notbohm H, Tronnier M, Açil Y, Fietzek PP, Schmeller W, Müller PK &<br />
Bätge B (1999) Overhydroxylation <strong>of</strong> lysyl residues is the <strong>in</strong>itial step for altered<br />
collagen cross-l<strong>in</strong>ks <strong>and</strong> fibril architecture <strong>in</strong> fibrotic sk<strong>in</strong>. J Invest Dermatol 113(4):<br />
617–621.<br />
92
Brodsky B & Persikov AV (2005) Molecular structure <strong>of</strong> the collagen triple helix. Adv<br />
Prote<strong>in</strong> Chem 70: 301–339.<br />
Bunt S, Denholm B & Skaer H (2010) Characterisation <strong>of</strong> the Drosophila procollagen lysyl<br />
hydroxylase, dPlod. Gene Expr Patterns 11(1–2): 72–78.<br />
Butcher DT, Alliston T & Weaver VM (2009) A tense situation: forc<strong>in</strong>g tumour<br />
progression. Nat Rev Cancer 9(2): 108–122.<br />
Canty EG & Kadler KE (2002) Collagen fibril biosynthesis <strong>in</strong> tendon: a review <strong>and</strong> recent<br />
<strong>in</strong>sights. Comp Biochem Physiol A Mol Integr Physiol 133(4): 979–985.<br />
Canty EG & Kadler KE (2005) Procollagen traffick<strong>in</strong>g, process<strong>in</strong>g <strong>and</strong> fibrillogenesis. J<br />
Cell Sci 118(Pt 7): 1341–1353.<br />
Chakravarti S, Magnuson T, Lass JH, Jepsen KJ, LaMantia C & Carroll H (1998) Lumican<br />
regulates collagen fibril assembly: sk<strong>in</strong> fragility <strong>and</strong> corneal opacity <strong>in</strong> the absence <strong>of</strong><br />
lumican. J Cell Biol 141(5): 1277–1286.<br />
Chang B, Chen Y, Zhao Y & Bruick RK (2007) JMJD6 is a histone arg<strong>in</strong><strong>in</strong>e demethylase.<br />
Science 318(5849): 444–447.<br />
Christiansen HE, Schwarze U, Pyott SM, AlSwaid A, Al Balwi M, Alrasheed S, Pep<strong>in</strong> MG,<br />
Weis MA, Eyre DR & Byers PH (2010) Homozygosity for a missense mutation <strong>in</strong><br />
SERPINH1, which encodes the collagen chaperone prote<strong>in</strong> HSP47, results <strong>in</strong> severe<br />
recessive osteogenesis imperfecta. Am J Hum Genet 86(3): 389–398.<br />
Chu ML, Conway D, Pan TC, Baldw<strong>in</strong> C, Mann K, Deutzmann R & Timpl R (1988)<br />
Am<strong>in</strong>o acid sequence <strong>of</strong> the triple-helical doma<strong>in</strong> <strong>of</strong> human collagen type VI. J Biol<br />
Chem 263(35): 18601–18606.<br />
Chung UI, Kawaguchi H, Takato T & Nakamura K (2004) Dist<strong>in</strong>ct osteogenic<br />
mechanisms <strong>of</strong> bones <strong>of</strong> dist<strong>in</strong>ct orig<strong>in</strong>s. J Orthop Sci 9(4): 410–414.<br />
Clifton IJ, McDonough MA, Ehrismann D, Kershaw NJ, Granat<strong>in</strong>o N & Sch<strong>of</strong>ield CJ<br />
(2006) Structural studies on 2-oxoglutarate oxygenases <strong>and</strong> related double-str<strong>and</strong>ed<br />
beta-helix fold prote<strong>in</strong>s. J Inorg Biochem 100(4): 644–669.<br />
Colpaert CG, Vermeulen PB, Fox SB, Harris AL, Dirix LY & Van Marck EA (2003) The<br />
presence <strong>of</strong> a fibrotic focus <strong>in</strong> <strong>in</strong>vasive breast carc<strong>in</strong>oma correlates with the expression<br />
<strong>of</strong> carbonic anhydrase IX <strong>and</strong> is a marker <strong>of</strong> hypoxia <strong>and</strong> poor prognosis. Breast<br />
Cancer Res Treat 81(2): 137–147.<br />
Corsi A, Xu T, Chen XD, Boyde A, Liang J, Mankani M, Sommer B, Iozzo RV,<br />
Eichstetter I, Robey PG, Bianco P & Young MF (2002) Phenotypic effects <strong>of</strong> biglycan<br />
deficiency are l<strong>in</strong>ked to collagen fibril abnormalities, are synergized by decor<strong>in</strong><br />
deficiency, <strong>and</strong> mimic Ehlers-Danlos-like changes <strong>in</strong> bone <strong>and</strong> other connective<br />
tissues. J Bone M<strong>in</strong>er Res 17(7): 1180–1189.<br />
Coss MC, W<strong>in</strong>terste<strong>in</strong> D, Sowder RC 2 nd & Simek SL (1995) Molecular clon<strong>in</strong>g, DNA<br />
sequence analysis, <strong>and</strong> biochemical characterization <strong>of</strong> a novel 65-kDa FK506b<strong>in</strong>d<strong>in</strong>g<br />
prote<strong>in</strong> (FKBP65). J Biol Chem 270(49): 29336–29341.<br />
Daley WP, Peters SB & Larsen M (2008) Extracellular matrix dynamics <strong>in</strong> development<br />
<strong>and</strong> regenerative medic<strong>in</strong>e. J Cell Sci 121(Pt 3): 255–264.<br />
93
Danielson KG, Baribault H, Holmes DF, Graham H, Kadler KE & Iozzo RV (1997)<br />
Targeted disruption <strong>of</strong> decor<strong>in</strong> leads to abnormal collagen fibril morphology <strong>and</strong> sk<strong>in</strong><br />
fragility. J Cell Biol 136(3): 729–743.<br />
Datta V, S<strong>in</strong>ha A, Saili A & Nangia S (2005) Bruck syndrome. Indian J Pediatr 72(5):<br />
441–442.<br />
Davis EC, Broekelmann TJ, Ozawa Y & Mecham RP (1998) Identification <strong>of</strong> tropoelast<strong>in</strong><br />
as a lig<strong>and</strong> for the 65-kD FK506-b<strong>in</strong>d<strong>in</strong>g prote<strong>in</strong>, FKBP65, <strong>in</strong> the secretory pathway. J<br />
Cell Biol 140(2): 295–303.<br />
De<strong>in</strong><strong>in</strong>ger PL & Batzer MA (1999) Alu repeats <strong>and</strong> human disease. Mol Genet Metab<br />
67(3): 183–193.<br />
Del Gatto-Konczak F, Bourgeois CF, Le Gu<strong>in</strong>er C, Kister L, Gesnel MC, Stéven<strong>in</strong> J &<br />
Breathnach R (2000) The RNA-b<strong>in</strong>d<strong>in</strong>g prote<strong>in</strong> TIA-1 is a novel mammalian splic<strong>in</strong>g<br />
regulator act<strong>in</strong>g through <strong>in</strong>tron sequences adjacent to a 5' splice site. Mol Cell Biol<br />
20(17): 6287–6299.<br />
Dembure PP, Janko AR, Priest JH & Elsas LJ (1987) Ascorbate regulation <strong>of</strong> collagen<br />
biosynthesis <strong>in</strong> Ehlers-Danlos syndrome, type VI. Metabolism 36(7): 687–691.<br />
Denko NC, Fontana LA, Hudson KM, Sutph<strong>in</strong> PD, Raychaudhuri S, Altman R & Giaccia<br />
AJ (2003) Investigat<strong>in</strong>g hypoxic tumor physiology through gene expression patterns.<br />
Oncogene 22(37): 5907–5914.<br />
Dong S, Nutt CL, Betensky RA, Stemmer-Rachamimov AO, Denko NC, Ligon KL,<br />
Rowitch DH & Louis DN (2005) Histology-based expression pr<strong>of</strong>il<strong>in</strong>g yields novel<br />
prognostic markers <strong>in</strong> human glioblastoma. J Neuropathol Exp Neurol 64(11): 948–<br />
955.<br />
Engel J & Prockop DJ (1991) The zipper-like fold<strong>in</strong>g <strong>of</strong> collagen triple helices <strong>and</strong> the<br />
effects <strong>of</strong> mutations that disrupt the zipper. Annu Rev Biophys Biophys Chem 20:<br />
137–152.<br />
Engler AJ, Sen S, Sweeney HL & Discher DE (2006) Matrix elasticity directs stem cell<br />
l<strong>in</strong>eage specification. Cell 126(4): 677–689.<br />
Evens AM, Sehn LH, Far<strong>in</strong>ha P, Nelson BP, Raji A, Lu Y, Brakman A, Parimi V, W<strong>in</strong>ter<br />
JN, Schumacker PT, Gascoyne RD & Gordon LI (2010) Hypoxia-<strong>in</strong>ducible factor-1 α<br />
expression predicts superior survival <strong>in</strong> patients with diffuse large B-cell lymphoma<br />
treated with R-CHOP. J Cl<strong>in</strong> Oncol 28(6): 1017–1024.<br />
Eyre DR & Glimcher MJ (1972) Reducible crossl<strong>in</strong>ks <strong>in</strong> hydroxylys<strong>in</strong>e-deficient collagens<br />
<strong>of</strong> a heritable disorder <strong>of</strong> connective tissue. Proc Natl Acad Sci U S A 69(9): 2594–<br />
2598.<br />
Eyre D, Shao P, Weis MA & Ste<strong>in</strong>mann B (2002) The kyphoscoliotic type <strong>of</strong> Ehlers-<br />
Danlos syndrome (type VI): differential effects on the hydroxylation <strong>of</strong> lys<strong>in</strong>e <strong>in</strong><br />
collagens I <strong>and</strong> II revealed by analysis <strong>of</strong> cross-l<strong>in</strong>ked telopeptides from ur<strong>in</strong>e. Mol<br />
Genet Metab 76(3): 211–216.<br />
Eyre DR, Weis MA & Wu JJ (2010) Maturation <strong>of</strong> collagen Ketoim<strong>in</strong>e cross-l<strong>in</strong>ks by an<br />
alternative mechanism to pyrid<strong>in</strong>ol<strong>in</strong>e formation <strong>in</strong> cartilage. J Biol Chem 285(22):<br />
16675–16682.<br />
94
Eyre DR & Wu JJ (2005) Collagen cross-l<strong>in</strong>ks. In: Br<strong>in</strong>ckmann J, Notbohm H & Müller<br />
PK (eds) Topics In Current Chemistry, Collagen: Primer <strong>in</strong> Structure, Process<strong>in</strong>g <strong>and</strong><br />
Assembly. Spr<strong>in</strong>ger-Verlag, Berl<strong>in</strong> Heidelberg, 247: 207–229.<br />
F<strong>in</strong>kelman RD, Eason AL, Rakijian DR, Tutundzhyan Y & Hardesty RA (1994) Elevated<br />
IGF-II <strong>and</strong> TGF-beta concentrations <strong>in</strong> human calvarial bone: potential mechanism for<br />
<strong>in</strong>creased graft survival <strong>and</strong> resistance to osteoporosis. Plast Reconstr Surg 93(4):<br />
732–738.<br />
Fitzgerald J, Rich C, Zhou FH & Hansen U (2008) Three novel collagen VI cha<strong>in</strong>s,<br />
alpha4(VI), alpha5(VI), <strong>and</strong> alpha6(VI). J Biol Chem 283(29): 20170–20180.<br />
Fleischmajer R, Perlish JS, Burgeson RE, Shaikh-Bahai F & Timpl R (1990) Type I <strong>and</strong><br />
type III collagen <strong>in</strong>teractions dur<strong>in</strong>g fibrillogenesis. Ann N Y Acad Sci 580: 161–175.<br />
Förch P, Puig O, Kedersha N, Martínez C, Granneman S, Séraph<strong>in</strong> B, Anderson P &<br />
Valcárcel J (2000) The apoptosis-promot<strong>in</strong>g factor TIA-1 is a regulator <strong>of</strong> alternative<br />
pre-mRNA splic<strong>in</strong>g. Mol Cell 6(5): 1089–1098.<br />
Gardner RL (1983) Orig<strong>in</strong> <strong>and</strong> differentiation <strong>of</strong> extraembryonic tissues <strong>in</strong> the mouse. Int<br />
Rev Exp Pathol 24: 63–133.<br />
Gast<strong>in</strong>el LN, Bignon C, Misra AK, H<strong>in</strong>dsgaul O, Shaper JH & Joziasse DH (2001) Bov<strong>in</strong>e<br />
alpha1,3-galactosyltransferase catalytic doma<strong>in</strong> structure <strong>and</strong> its relationship with<br />
ABO histo-blood group <strong>and</strong> glycosph<strong>in</strong>golipid glycosyltransferases. EMBO J 20(4):<br />
638–649.<br />
Gelse K, Pöschl E & Aigner T (2003) Collagens--structure, function, <strong>and</strong> biosynthesis.<br />
Adv Drug Deliv Rev 55(12): 1531–1546.<br />
Gerriets JE, Curw<strong>in</strong> SL & Last JA (1993) Tendon hypertrophy is associated with <strong>in</strong>creased<br />
hydroxylation <strong>of</strong> nonhelical lys<strong>in</strong>e residues at two specific cross-l<strong>in</strong>k<strong>in</strong>g sites <strong>in</strong> type I<br />
collagen. J Biol Chem 268(34): 25553–25560.<br />
Giunta C, Elçioglu NH, Albrecht B, Eich G, Chambaz C, Janecke AR, Yeowell H, Weis M,<br />
Eyre DR, Kraenzl<strong>in</strong> M & Ste<strong>in</strong>mann B (2008) Spondylocheiro dysplastic form <strong>of</strong> the<br />
Ehlers-Danlos syndrome--an autosomal-recessive entity caused by mutations <strong>in</strong> the<br />
z<strong>in</strong>c transporter gene SLC39A13. Am J Hum Genet 82(6): 1290–1305.<br />
Giunta C, R<strong>and</strong>olph A, Al-Gazali LI, Brunner HG, Kraenzl<strong>in</strong> ME & Ste<strong>in</strong>mann B (2005b)<br />
Nevo syndrome is allelic to the kyphoscoliotic type <strong>of</strong> the Ehlers-Danlos syndrome<br />
(EDS VIA). Am J Med Genet A 133A(2): 158–164.<br />
Giunta C, R<strong>and</strong>olph A & Ste<strong>in</strong>mann B (2005a) Mutation analysis <strong>of</strong> the PLOD1 gene: an<br />
efficient multistep approach to the molecular diagnosis <strong>of</strong> the kyphoscoliotic type <strong>of</strong><br />
Ehlers-Danlos syndrome (EDS VIA). Mol Genet Metab 86(1–2): 269–276.<br />
Gordon MK & Hahn RA (2010) Collagens. Cell Tissue Res 339(1): 247–257.<br />
Göthel SF & Marahiel MA (1999) Peptidyl-prolyl cis-trans isomerases, a superfamily <strong>of</strong><br />
ubiquitous fold<strong>in</strong>g catalysts. Cell Mol Life Sci 55(3): 423–436.<br />
Ha VT, Marshall MK, Elsas LJ, P<strong>in</strong>nell SR & Yeowell HN (1994) A patient with Ehlers-<br />
Danlos syndrome type VI is a compound heterozygote for mutations <strong>in</strong> the lysyl<br />
hydroxylase gene. J Cl<strong>in</strong> Invest 93(4): 1716–1721.<br />
95
Hanauske-Abel HM & Günzler V (1982) A stereochemical concept for the catalytic<br />
mechanism <strong>of</strong> prolylhydroxylase: applicability to classification <strong>and</strong> design <strong>of</strong><br />
<strong>in</strong>hibitors. J Theor Biol 94(2): 421–455.<br />
Hanson DA & Eyre DR (1996) Molecular site specificity <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e <strong>and</strong> pyrrole crossl<strong>in</strong>ks<br />
<strong>in</strong> type I collagen <strong>of</strong> human bone. J Biol Chem 271(43): 26508–26516.<br />
Hautala T, Byers MG, Eddy RL, Shows TB, Kivirikko KI & Myllylä R (1992a) Clon<strong>in</strong>g <strong>of</strong><br />
human lysyl hydroxylase: complete cDNA-derived am<strong>in</strong>o acid sequence <strong>and</strong><br />
assignment <strong>of</strong> the gene (PLOD) to chromosome 1p36.3----p36.2. Genomics 13(1):<br />
62–69.<br />
Hautala T, Heikk<strong>in</strong>en J, Kivirikko KI & Myllylä R (1992b) M<strong>in</strong>oxidil specifically<br />
decreases the expression <strong>of</strong> lys<strong>in</strong>e hydroxylase <strong>in</strong> cultured human sk<strong>in</strong> fibroblasts.<br />
Biochem J 283(Pt 1): 51–54.<br />
Ha-V<strong>in</strong>h R, Alanay Y, Bank RA, Campos-Xavier AB, Zankl A, Superti-Furga A & Bonafé<br />
L (2004) Phenotypic <strong>and</strong> molecular characterization <strong>of</strong> Bruck syndrome (osteogenesis<br />
imperfecta with contractures <strong>of</strong> the large jo<strong>in</strong>ts) caused by a recessive mutation <strong>in</strong><br />
PLOD2. Am J Med Genet A 131(2): 115–120.<br />
Heikk<strong>in</strong>en J, Hautala T, Kivirikko KI & Myllylä R (1994) Structure <strong>and</strong> expression <strong>of</strong> the<br />
human lysyl hydroxylase gene (PLOD): <strong>in</strong>trons 9 <strong>and</strong> 16 conta<strong>in</strong> Alu sequences at the<br />
sites <strong>of</strong> recomb<strong>in</strong>ation <strong>in</strong> Ehlers-Danlos syndrome type VI patients. Genomics 24(3):<br />
464–471.<br />
Heikk<strong>in</strong>en J, Pousi B, Pope M & Myllylä R (1999) A null-mutated lysyl hydroxylase gene<br />
<strong>in</strong> a compound heterozygote British patient with Ehlers-Danlos syndrome type VI.<br />
Hum Mutat 14(4): 351.<br />
Heikk<strong>in</strong>en J, Risteli M, Lampela O, Alavesa P, Karpp<strong>in</strong>en M, Juffer AH & Myllylä R<br />
(2011) Dimerization <strong>of</strong> human lysyl hydroxylase 3 (LH3) is mediated by the am<strong>in</strong>o<br />
acids 541–547. Matrix Biol 30(1): 27–33.<br />
Heikk<strong>in</strong>en J, Risteli M, Wang C, Latvala J, Rossi M, Valtavaara M & Myllylä R (2000)<br />
<strong>Lysyl</strong> hydroxylase 3 is a multifunctional prote<strong>in</strong> possess<strong>in</strong>g collagen<br />
glucosyltransferase activity. J Biol Chem 275(46): 36158–36163.<br />
Heikk<strong>in</strong>en J, Topp<strong>in</strong>en T, Yeowell H, Krieg T, Ste<strong>in</strong>mann B, Kivirikko KI & Myllylä R<br />
(1997) Duplication <strong>of</strong> seven exons <strong>in</strong> the lysyl hydroxylase gene is associated with<br />
longer forms <strong>of</strong> a repetitive sequence with<strong>in</strong> the gene <strong>and</strong> is a common cause for the<br />
type VI variant <strong>of</strong> Ehlers-Danlos syndrome. Am J Hum Genet 60(1): 48–56.<br />
Heise CT, Nicholls JR, Leamy CE & Wallis R (2000) Impaired secretion <strong>of</strong> rat mannoseb<strong>in</strong>d<strong>in</strong>g<br />
prote<strong>in</strong> result<strong>in</strong>g from mutations <strong>in</strong> the collagen-like doma<strong>in</strong>. J Immunol<br />
165(3): 1403–1409.<br />
Henkel W, Rauterberg J & Stirtz T (1976) Isolation <strong>of</strong> a crossl<strong>in</strong>ked cyanogen-bromide<br />
peptide from <strong>in</strong>soluble rabbit collagen. Tissue differences <strong>in</strong> hydroxylation <strong>and</strong><br />
glycosylation <strong>of</strong> the crossl<strong>in</strong>k. Eur J Biochem 69(1): 223–231.<br />
Hjalt TA, Amendt BA & Murray JC (2001) PITX2 regulates procollagen lysyl hydroxylase<br />
(PLOD) gene expression: implications for the pathology <strong>of</strong> Rieger syndrome. J Cell<br />
Biol 152(3): 545–552.<br />
96
H<strong>of</strong>bauer KH, Gess B, Lohaus C, Meyer HE, Katsch<strong>in</strong>ski D & Kurtz A (2003) Oxygen<br />
tension regulates the expression <strong>of</strong> a group <strong>of</strong> procollagen <strong>hydroxylases</strong>. Eur J<br />
Biochem 270(22): 4515–4522.<br />
Hong X, Zang J, White J, Wang C, Pan CH, Zhao R, Murphy RC, Dai S, Henson P,<br />
Kappler JW, Hagman J & Zhang G (2010) Interaction <strong>of</strong> JMJD6 with s<strong>in</strong>gle-str<strong>and</strong>ed<br />
RNA. PNAS 107(33): 14568–14572.<br />
Hunzelmann N & Br<strong>in</strong>ckmann J (2011) What are the new milestones <strong>in</strong> the pathogenesis<br />
<strong>of</strong> systemic sclerosis? Ann Rheum Dis 69 Suppl 1: i52–i56.<br />
Hynes RO (2009) The extracellular matrix: not just pretty fibrils. Science 326(5957):<br />
1216–1219.<br />
Ihme A, Krieg T, Nerlich A, Feldmann U, Rauterberg J, Glanville RW, Edel G & Müller<br />
PK (1984) Ehlers-Danlos syndrome type VI: collagen type specificity <strong>of</strong> defective<br />
lysyl hydroxylation <strong>in</strong> various tissues. J Invest Dermatol 83(3): 161–165.<br />
Inman KE & Downs KM (2007) The mur<strong>in</strong>e allantois: emerg<strong>in</strong>g paradigms <strong>in</strong><br />
development <strong>of</strong> the mammalian umbilical cord <strong>and</strong> its relation to the fetus. Genesis<br />
45(5): 237–58.<br />
Ishida Y & Nagata K (2011) Hsp47 as a collagen-specific molecular chaperone. Methods<br />
Enzymol 499: 167–182.<br />
Ishikawa Y, Vranka J, Wirz J, Nagata K & Bäch<strong>in</strong>ger HP (2008) The rough endoplasmic<br />
reticulum-resident FK506-b<strong>in</strong>d<strong>in</strong>g prote<strong>in</strong> FKBP65 is a molecular chaperone that<br />
<strong>in</strong>teracts with collagens. J Biol Chem 283(46): 31584–31590.<br />
Iwahashi M, Ooshima A & Nakano R (1996) Increase <strong>in</strong> the relative level <strong>of</strong> type V<br />
collagen dur<strong>in</strong>g development <strong>and</strong> age<strong>in</strong>g <strong>of</strong> the placenta. J Cl<strong>in</strong> Pathol 49(11): 916–<br />
919.<br />
Izu Y, Ansorge HL, Zhang G, Soslowsky LJ, Bonaldo P, Chu ML & Birk DE (2011)<br />
Dysfunctional tendon collagen fibrillogenesis <strong>in</strong> collagen VI null mice. Matrix Biol<br />
30(1): 53–61.<br />
Jarisch A, Giunta C, Zielen S, König R & Ste<strong>in</strong>mann B (1998) Sibs affected with both<br />
Ehlers-Danlos syndrome type IV <strong>and</strong> cystic fibrosis. Am J Med Genet 78(5): 455–460.<br />
Järvelä<strong>in</strong>en H, Sa<strong>in</strong>io A, Koulu M, Wight TN & Pentt<strong>in</strong>en R (2009) Extracellular matrix<br />
molecules: potential targets <strong>in</strong> pharmacotherapy. Pharmacol Rev 61(2): 198–223.<br />
Kadler KE, Baldock C, Bella J & Boot-H<strong>and</strong>ford RP (2007) Collagens at a glance. J Cell<br />
Sci 120(Pt 12): 1955–1958.<br />
Kampranis SC & Tsichlis PN (2009) Histone demethylases <strong>and</strong> cancer. Adv Cancer Res<br />
102: 103–169.<br />
Kelley BP, Malfait F, Bonafe L, Baldridge D, Homan E, Symoens S, Willaert A, Elcioglu<br />
N, Van Maldergem L, Verellen-Dumoul<strong>in</strong> C, Gillerot Y, Napierala D, Krakow D,<br />
Beighton P, Superti-Furga A, De Paepe A & Lee B (2011) Mutations <strong>in</strong> FKBP10<br />
cause recessive osteogenesis imperfecta <strong>and</strong> bruck syndrome. J Bone M<strong>in</strong>er Res 26(3):<br />
666–672.<br />
97
Kellokumpu S, Sormunen R, Heikk<strong>in</strong>en J & Myllylä R (1994) <strong>Lysyl</strong> hydroxylase, a<br />
collagen process<strong>in</strong>g enzyme, exemplifies a novel class <strong>of</strong> lum<strong>in</strong>ally-oriented<br />
peripheral membrane prote<strong>in</strong>s <strong>in</strong> the endoplasmic reticulum. J Biol Chem 269(48):<br />
30524–30529.<br />
Kielty CM & Grant ME (2002) The collagen family: structure, assembly, <strong>and</strong> organization<br />
<strong>in</strong> the extracellular matrix. In: Royce PM & Ste<strong>in</strong>mann B (eds) Connective Tissue <strong>and</strong><br />
Its Heritable Disorders. Molecular, Genetics <strong>and</strong> Medical Aspects. 2nd ed. Wiley-Liss,<br />
Inc., New York, 159 –221.<br />
Kivirikko KI & Myllylä R (1979) Collagen glycosyltransferases. Int Rev Connect Tissue<br />
Res 8: 23–72.<br />
Kivirikko KI & Myllylä R (1982) Posttranslational enzymes <strong>in</strong> the biosynthesis <strong>of</strong> collagen:<br />
<strong>in</strong>tracellular enzymes. Methods Enzymol 82(Pt A): 245–304.<br />
Kivirikko KI, Myllylä R & Pihlajaniemi T (1992) Hydroxylation <strong>of</strong> prol<strong>in</strong>e <strong>and</strong> lys<strong>in</strong>e<br />
residues <strong>in</strong> collagens <strong>and</strong> other animal <strong>and</strong> plant prote<strong>in</strong>s. In: Hard<strong>in</strong>g JJ & Crabbe M<br />
(eds) Post-Translational Modifications <strong>of</strong> Prote<strong>in</strong>s. CRC Press, Inc., Boca Raton, 1–51.<br />
Kivirikko KI & Pihlajaniemi T (1998) Collagen <strong>hydroxylases</strong> <strong>and</strong> the prote<strong>in</strong> disulfide<br />
isomerase subunit <strong>of</strong> prolyl-4-hydroxylase. Adv Enzymol Relat Areas Mol Biol 72:<br />
325–398.<br />
Knott L & Bailey AJ (1998) Collagen cross-l<strong>in</strong>ks <strong>in</strong> m<strong>in</strong>eraliz<strong>in</strong>g tissues: a review <strong>of</strong> <strong>their</strong><br />
chemistry, function, <strong>and</strong> cl<strong>in</strong>ical relevance. Bone 22(3): 181–187.<br />
Knott L, Tarlton JF & Bailey AJ (1997) Chemistry <strong>of</strong> collagen cross-l<strong>in</strong>k<strong>in</strong>g: biochemical<br />
changes <strong>in</strong> collagen dur<strong>in</strong>g the partial m<strong>in</strong>eralization <strong>of</strong> turkey leg tendon. Biochem J<br />
322(Pt 2): 535–542.<br />
Koski MK, Hieta R, Böllner C, Kivirikko KI, Myllyharju J & Wierenga RK (2007) The<br />
active site <strong>of</strong> an algal prolyl 4-hydroxylase has a large structural plasticity. J Biol<br />
Chem 282(51): 37112–37123.<br />
Krol BJ, Murad S, Walker LC, Marshall MK, Clark WL, P<strong>in</strong>nell SR & Yeowell HN (1996)<br />
The expression <strong>of</strong> a functional, secreted human lysyl hydroxylase <strong>in</strong> a baculovirus<br />
system. J Invest Dermatol 106(1): 11–16.<br />
Kronenberg HM (2003) Developmental regulation <strong>of</strong> the growth plate. Nature 423(6937):<br />
332–336.<br />
Lal A, Peters H, St Croix B, Haroon ZA, Dewhirst MW, Strausberg RL, Ka<strong>and</strong>ers JH, van<br />
der Kogel AJ & Rigg<strong>in</strong>s GJ (2001) Transcriptional response to hypoxia <strong>in</strong> human<br />
tumors. J Natl Cancer Inst 93(17): 1337–1343.<br />
Lampe AK & Bushby KM (2005) Collagen VI related muscle disorders. J Med Genet<br />
42(9): 673–685.<br />
Larraín J, Bachiller D, Lu B, Agius E, Piccolo S & De Robertis EM (2000) BMP-b<strong>in</strong>d<strong>in</strong>g<br />
modules <strong>in</strong> chord<strong>in</strong>: a model for signall<strong>in</strong>g regulation <strong>in</strong> the extracellular space.<br />
Development 127(4): 821–830.<br />
Lauer-Fields JL, Malkar NB, Richet G, Drauz K & Fields GB (2003) Melanoma cell CD44<br />
<strong>in</strong>teraction with the alpha 1(IV)1263–1277 region from basement membrane collagen<br />
is modulated by lig<strong>and</strong> glycosylation. J Biol Chem 278(16): 14321–14330.<br />
98
Le Gu<strong>in</strong>er C, Lejeune F, Galiana D, Kister L, Breathnach R, Stéven<strong>in</strong> J & Del Gatto-<br />
Konczak F (2001) TIA-1 <strong>and</strong> TIAR activate splic<strong>in</strong>g <strong>of</strong> alternative exons with weak 5'<br />
splice sites followed by a U-rich stretch on <strong>their</strong> own pre-mRNAs. J Biol Chem<br />
276(44): 40638–40646.<br />
Lees JF, Tasab M & Bulleid NJ (1997) Identification <strong>of</strong> the molecular recognition<br />
sequence which determ<strong>in</strong>es the type-specific assembly <strong>of</strong> procollagen. EMBO J 16(5):<br />
908–916.<br />
Leroy JG, Nuyt<strong>in</strong>ck L, De Paepe A, De Rammelaere M, Gillerot Y, Verloes A, Loeys B &<br />
De Groote W (1998) Bruck syndrome: neonatal presentation <strong>and</strong> natural course <strong>in</strong><br />
three patients. Pediatr Radiol 28(10): 781–789.<br />
Levental KR, Yu H, Kass L, Lak<strong>in</strong>s JN, Egeblad M, Erler JT, Fong SF, Csiszar K, Giaccia<br />
A, Wen<strong>in</strong>ger W, Yamauchi M, Gasser DL & Weaver VM (2009) Matrix crossl<strong>in</strong>k<strong>in</strong>g<br />
forces tumor progression by enhanc<strong>in</strong>g <strong>in</strong>tegr<strong>in</strong> signal<strong>in</strong>g. Cell 139(5): 891–906.<br />
Liefhebber JM, Punt S, Spaan WJ & van Leeuwen HC (2010) The human collagen beta(1-<br />
O)galactosyltransferase, GLT25D1, is a soluble endoplasmic reticulum localized<br />
prote<strong>in</strong>. BMC Cell Biol 11: 33.<br />
Liu X, Wu H, Byrne M, Krane S & Jaenisch R (1997) Type III collagen is crucial for<br />
collagen I fibrillogenesis <strong>and</strong> for normal cardiovascular development. Proc Natl Acad<br />
Sci U S A 94(5): 1852–1856.<br />
Löhler J, Timpl R & Jaenisch R (1984) Embryonic lethal mutation <strong>in</strong> mouse collagen I<br />
gene causes rupture <strong>of</strong> blood vessels <strong>and</strong> is associated with erythropoietic <strong>and</strong><br />
mesenchymal cell death. Cell 38(2): 597–607.<br />
Mar<strong>in</strong>i JC, Cabral WA & Barnes AM (2010) Null mutations <strong>in</strong> LEPRE1 <strong>and</strong> CRTAP<br />
cause severe recessive osteogenesis imperfecta. Cell Tissue Res 339(1): 59–70.<br />
McAl<strong>in</strong>den A, Havlioglu N, Liang L, Davies SR & S<strong>and</strong>ell LJ (2005) Alternative splic<strong>in</strong>g<br />
<strong>of</strong> type II procollagen exon 2 is regulated by the comb<strong>in</strong>ation <strong>of</strong> a weak 5' splice site<br />
<strong>and</strong> an adjacent <strong>in</strong>tronic stem-loop cis element. J Biol Chem 280(38): 32700–32711.<br />
McPherson E & Clemens M (1997) Bruck syndrome (osteogenesis imperfecta with<br />
congenital jo<strong>in</strong>t contractures): review <strong>and</strong> report on the first North American case. Am<br />
J Med Genet 70(1): 28–31.<br />
Mercer DK, Nicol PF, Kimbembe C & Rob<strong>in</strong>s SP (2003) Identification, expression, <strong>and</strong><br />
tissue distribution <strong>of</strong> the three rat lysyl hydroxylase is<strong>of</strong>orms. Biochem Biophys Res<br />
Commun 307(4): 803–809.<br />
Mokete L, Robertson A, Viljoen D & Beighton P (2005) Bruck syndrome: congenital jo<strong>in</strong>t<br />
contractures with bone fragility. J Orthop Sci 10(6): 641–646.<br />
Morello R, Bert<strong>in</strong> TK, Chen Y, Hicks J, Tonach<strong>in</strong>i L, Monticone M, Castagnola P, Rauch<br />
F, Glorieux FH, Vranka J, Bäch<strong>in</strong>ger HP, Pace JM, Schwarze U, Byers PH, Weis M,<br />
Fern<strong>and</strong>es RJ, Eyre DR, Yao Z, Boyce BF & Lee B (2006) CRTAP is required for<br />
prolyl 3- hydroxylation <strong>and</strong> mutations cause recessive osteogenesis imperfecta. Cell<br />
127(2): 291–304.<br />
Murad S & P<strong>in</strong>nell SR (1987) Suppression <strong>of</strong> fibroblast proliferation <strong>and</strong> lysyl hydroxylase<br />
activity by m<strong>in</strong>oxidil. J Biol Chem 262(25): 11973–11978.<br />
99
Myers LK, Myllyharju J, Nokela<strong>in</strong>en M, Br<strong>and</strong> DD, Cremer MA, Stuart JM, Bodo M,<br />
Kivirikko KI & Kang AH (2004) Relevance <strong>of</strong> posttranslational modifications for the<br />
arthritogenicity <strong>of</strong> type II collagen. J Immunol 172(5): 2970–2975.<br />
Myllyharju J (2005) Intracellular post-translational modifications <strong>of</strong> collagens. In:<br />
Br<strong>in</strong>ckmann J, Notbohm H & Müller PK (eds) Topics In Current Chemistry, Collagen:<br />
Primer <strong>in</strong> Structure, Process<strong>in</strong>g <strong>and</strong> Assembly. Spr<strong>in</strong>ger-Verlag, Berl<strong>in</strong> Heidelberg,<br />
247: 115–147.<br />
Myllyharju J & Kivirikko KI (1997) <strong>Characterization</strong> <strong>of</strong> the iron- <strong>and</strong> 2-oxoglutarateb<strong>in</strong>d<strong>in</strong>g<br />
sites <strong>of</strong> human prolyl 4-hydroxylase. EMBO J 16(6) :1173–1180.<br />
Myllyharju J & Kivirikko KI (2001) Collagens <strong>and</strong> collagen-related diseases. Ann Med<br />
33(1): 7–21.<br />
Myllyharju J & Kivirikko KI (2004) Collagens, modify<strong>in</strong>g enzymes <strong>and</strong> <strong>their</strong> mutations <strong>in</strong><br />
humans, flies <strong>and</strong> worms. Trends Genet 20(1): 33–43.<br />
Myllylä R, Majamaa K, Günzler V, Hanauske-Abel HM & Kivirikko KI (1984) Ascorbate<br />
is consumed stoichiometrically <strong>in</strong> the uncoupled reactions catalyzed by prolyl 4hydroxylase<br />
<strong>and</strong> lysyl hydroxylase. J Biol Chem 259(9): 5403–5405.<br />
Myllylä R, Pajunen L & Kivirikko KI (1988) Polyclonal <strong>and</strong> monoclonal antibodies to<br />
human lysyl hydroxylase <strong>and</strong> studies on the molecular heterogeneity <strong>of</strong> the enzyme.<br />
Biochem J 253(2): 489–496.<br />
Myllylä R, Pihlajaniemi T, Pajunen L, Turpeenniemi-Hujanen T & Kivirikko KI (1991)<br />
Molecular clon<strong>in</strong>g <strong>of</strong> chick lysyl hydroxylase. Little homology <strong>in</strong> primary structure to<br />
the two types <strong>of</strong> subunit <strong>of</strong> prolyl 4-hydroxylase. J Biol Chem 266(5): 2805–2810.<br />
Mäki JM (2009) <strong>Lysyl</strong> oxidases <strong>in</strong> mammalian development <strong>and</strong> certa<strong>in</strong> pathological<br />
conditions. Histol Histopathol 24(5): 651–660.<br />
Ng LJ, Tam PP & Cheah KS (1993) Preferential expression <strong>of</strong> alternatively spliced<br />
mRNAs encod<strong>in</strong>g type II procollagen with a cyste<strong>in</strong>e-rich am<strong>in</strong>o-propeptide <strong>in</strong><br />
differentiat<strong>in</strong>g cartilage <strong>and</strong> nonchondrogenic tissues dur<strong>in</strong>g early mouse development.<br />
Dev Biol 159(2): 403–417.<br />
Niederreither K, D'Souza R, Metsäranta M, Eberspaecher H, Toman PD, Vuorio E & De<br />
Crombrugghe B (1994) Coord<strong>in</strong>ate patterns <strong>of</strong> expression <strong>of</strong> type I <strong>and</strong> III collagens<br />
dur<strong>in</strong>g mouse development. Matrix Biol 14(9): 705–713.<br />
Noda T, Yamamoto H, Takemasa I, Yamada D, Uemura M, Wada H, Kobayashi S,<br />
Marubashi S, Eguchi H, Tanemura M, Umeshita K, Doki Y, Mori M & Nagano H<br />
(2012) PLOD2 <strong>in</strong>duced under hypoxia is a novel prognostic factor for hepatocellular<br />
carc<strong>in</strong>oma after curative resection. Liver Int 32(1): 110–118.<br />
Norman KR & Moerman DG (2000) The let-268 locus <strong>of</strong> Caenorhabditis elegans encodes<br />
a procollagen lysyl hydroxylase that is essential for type IV collagen secretion. Dev<br />
Biol 227(2): 690–705.<br />
Notbohm H, Nokela<strong>in</strong>en M, Myllyharju J, Fietzek PP, Müller PK & Kivirikko KI (1999)<br />
Recomb<strong>in</strong>ant human type II collagens with low <strong>and</strong> high levels <strong>of</strong> hydroxylys<strong>in</strong>e <strong>and</strong><br />
its glycosylated forms show marked differences <strong>in</strong> fibrillogenesis <strong>in</strong> vitro. J Biol<br />
Chem 274(13): 8988–8992.<br />
100
Park<strong>in</strong> JD, San Antonio JD, Pedchenko V, Hudson B, Jensen ST & Savige J (2011)<br />
Mapp<strong>in</strong>g structural l<strong>and</strong>marks, lig<strong>and</strong> b<strong>in</strong>d<strong>in</strong>g sites <strong>and</strong> missense mutations to the<br />
collagen IV heterotrimers predicts major functional doma<strong>in</strong>s, novel <strong>in</strong>teractions <strong>and</strong><br />
variation <strong>in</strong> phenotypes <strong>in</strong> <strong>in</strong>herited diseases affect<strong>in</strong>g basement membranes. Hum<br />
Mutat 32(2): 127–143.<br />
Pasquali M, Dembure PP, Still MJ & Elsas LJ (1994) Ur<strong>in</strong>ary pyrid<strong>in</strong>ium cross-l<strong>in</strong>ks: a<br />
non<strong>in</strong>vasive diagnostic test for Ehlers-Danlos syndrome type VI. N Engl J Med 331(2):<br />
132–133.<br />
Pasquali M, Still MJ, Dembure PP & Elsas LJ (1995) Pyrid<strong>in</strong>ium cross-l<strong>in</strong>ks <strong>in</strong> heritable<br />
disorders <strong>of</strong> collagen. Am J Hum Genet 57(6): 1508–1510.<br />
Pasquali M, Still MJ, Vales T, Rosen RI, Ev<strong>in</strong>ger JD, Dembure PP, Longo N & Elsas LJ<br />
(1997) Abnormal formation <strong>of</strong> collagen cross-l<strong>in</strong>ks <strong>in</strong> sk<strong>in</strong> fibroblasts cultured from<br />
patients with Ehlers-Danlos syndrome type VI. Proc Assoc Am Physicians 109(1):<br />
33–41.<br />
Passoja K, Myllyharju J, Pirskanen A & Kivirikko KI (1998b) Identification <strong>of</strong> arg<strong>in</strong><strong>in</strong>e-<br />
700 as the residue that b<strong>in</strong>ds the C-5 carboxyl group <strong>of</strong> 2-oxoglutarate <strong>in</strong> human lysyl<br />
hydroxylase 1. FEBS Lett 434(1–2): 145–148.<br />
Passoja K, Rautavuoma K, Ala-Kokko L, Kosonen T & Kivirikko KI (1998a) Clon<strong>in</strong>g <strong>and</strong><br />
characterization <strong>of</strong> a third human lysyl hydroxylase is<strong>of</strong>orm. Proc Natl Acad Sci U S<br />
A 95(18): 10482–10486.<br />
Patterson CE, Schaub T, Coleman EJ & Davis EC (2000) Developmental regulation <strong>of</strong><br />
FKBP65. An ER-localized extracellular matrix b<strong>in</strong>d<strong>in</strong>g-prote<strong>in</strong>. Mol Biol Cell 11(11):<br />
3925–3935.<br />
P<strong>in</strong>nell SR, Krane SM, Kenzora JE & Glimcher MJ (1972) A heritable disorder <strong>of</strong><br />
connective tissue. Hydroxylys<strong>in</strong>e-deficient collagen disease. N Engl J Med 286(19):<br />
1013–1020.<br />
Pirskanen A, Kaimio AM, Myllylä R & Kivirikko KI (1996) Site-directed mutagenesis <strong>of</strong><br />
human lysyl hydroxylase expressed <strong>in</strong> <strong>in</strong>sect cells. Identification <strong>of</strong> histid<strong>in</strong>e residues<br />
<strong>and</strong> an aspartic acid residue critical for catalytic activity. J Biol Chem 271(16): 9398–<br />
9402.<br />
Pornprasertsuk S, Duarte WR, Mochida Y & Yamauchi M (2004) <strong>Lysyl</strong> hydroxylase-2b<br />
directs collagen cross-l<strong>in</strong>k<strong>in</strong>g pathways <strong>in</strong> MC3T3-E1 cells. J Bone M<strong>in</strong>er Res 19(8):<br />
1349–1355.<br />
Pornprasertsuk S, Duarte WR, Mochida Y & Yamauchi M (2005) Overexpression <strong>of</strong> lysyl<br />
hydroxylase-2b leads to defective collagen fibrillogenesis <strong>and</strong> matrix m<strong>in</strong>eralization. J<br />
Bone M<strong>in</strong>er Res 20(1): 81–87.<br />
Pousi B, Hautala T, Heikk<strong>in</strong>en J, Pajunen L, Kivirikko KI & Myllylä R (1994) Alu-Alu<br />
recomb<strong>in</strong>ation results <strong>in</strong> a duplication <strong>of</strong> seven exons <strong>in</strong> the lysyl hydroxylase gene <strong>in</strong><br />
a patient with the type VI variant <strong>of</strong> Ehlers-Danlos syndrome. Am J Hum Genet 55(5):<br />
899–906.<br />
Pousi B, Heikk<strong>in</strong>en J, Schröter J, Pope M & Myllylä R (2000) A nonsense codon <strong>of</strong> exon<br />
14 reduces lysyl hydroxylase mRNA <strong>and</strong> leads to aberrant RNA splic<strong>in</strong>g <strong>in</strong> a patient<br />
with Ehlers-Danlos syndrome type VI. Mutat Res 432(1–2): 33–37.<br />
101
Pöschl E, Schlötzer-Schrehardt U, Brachvogel B, Saito K, N<strong>in</strong>omiya Y & Mayer U (2004)<br />
Collagen IV is essential for basement membrane stability but dispensable for <strong>in</strong>itiation<br />
<strong>of</strong> its assembly dur<strong>in</strong>g early development. Development 131(7): 1619–1628.<br />
Raghunath M, Bruckner P & Ste<strong>in</strong>mann B (1994) Delayed triple helix formation <strong>of</strong> mutant<br />
collagen from patients with osteogenesis imperfecta. J Mol Biol 236(3): 940–949.<br />
Rajkumar T, Sabitha K, Vijayalakshmi N, Shirley S, Bose MV, Gopal G & Selvaluxmy G<br />
(2011) Identification <strong>and</strong> validation <strong>of</strong> genes <strong>in</strong>volved <strong>in</strong> cervical tumourigenesis.<br />
BMC Cancer 11: 80.<br />
Rautavuoma K, Passoja K, Helaakoski T & Kivirikko KI (2000) Complete exon-<strong>in</strong>tron<br />
organization <strong>of</strong> the gene for human lysyl hydroxylase 3 (LH3). Matrix Biol 19(1): 73–<br />
79.<br />
Rautavuoma K, Takaluoma K, Passoja K, Pirskanen A, Kvist AP, Kivirikko KI &<br />
Myllyharju J (2002) <strong>Characterization</strong> <strong>of</strong> three fragments that constitute the monomers<br />
<strong>of</strong> the human lysyl hydroxylase isoenzymes 1-3. The 30-kDa N-term<strong>in</strong>al fragment is<br />
not required for lysyl hydroxylase activity. J Biol Chem 277(25): 23084–23091.<br />
Rautavuoma K, Takaluoma K, Sormunen R, Myllyharju J, Kivirikko KI & So<strong>in</strong><strong>in</strong>en R<br />
(2004) Premature aggregation <strong>of</strong> type IV collagen <strong>and</strong> early lethality <strong>in</strong> lysyl<br />
hydroxylase 3 null mice. Proc Natl Acad Sci U S A 101(39): 14120–14125.<br />
Reiser K, McCormick RJ & Rucker RB (1992) Enzymatic <strong>and</strong> nonenzymatic cross-l<strong>in</strong>k<strong>in</strong>g<br />
<strong>of</strong> collagen <strong>and</strong> elast<strong>in</strong>. FASEB J 6(7): 2439–2449.<br />
Ricard-Blum S (2011) The collagen family. Cold Spr<strong>in</strong>g Harb Perspect Biol 3(1): a004978.<br />
Ricard-Blum S & Ruggiero F (2005) The collagen superfamily: from the extracellular<br />
matrix to the cell membrane. Pathol Biol 53(7): 430–442.<br />
Risteli M, Niemitalo O, Lank<strong>in</strong>en H, Juffer AH & Myllylä R (2004) <strong>Characterization</strong> <strong>of</strong><br />
collagenous peptides bound to lysyl hydroxylase is<strong>of</strong>orms. J Biol Chem 279(36):<br />
37535–37543.<br />
Risteli M, Ruotsala<strong>in</strong>en H, Salo AM, Sormunen R, Sipilä L, Baker NL, Lam<strong>and</strong>é SR,<br />
Vimpari-Kaupp<strong>in</strong>en L & Myllylä R (2009) Reduction <strong>of</strong> lysyl hydroxylase 3 causes<br />
deleterious changes <strong>in</strong> the deposition <strong>and</strong> organization <strong>of</strong> extracellular matrix. J Biol<br />
Chem 284(41): 28204–28211.<br />
Rob<strong>in</strong>s SP (2007) Biochemistry <strong>and</strong> functional significance <strong>of</strong> collagen cross-l<strong>in</strong>k<strong>in</strong>g.<br />
Biochem Soc Trans 35(Pt 5): 849–852.<br />
Rossant J & Cross JC (2001) Placental development: lessons from mouse mutants. Nat<br />
Rev Genet 2(7): 538–548.<br />
Roulet M, Ruggiero F, Karsenty G & LeGuellec D (2007) A comprehensive study <strong>of</strong> the<br />
spatial <strong>and</strong> temporal expression <strong>of</strong> the col5a1 gene <strong>in</strong> mouse embryos: a clue for<br />
underst<strong>and</strong><strong>in</strong>g collagen V function <strong>in</strong> develop<strong>in</strong>g connective tissues. Cell Tissue Res<br />
327(2): 323–332.<br />
Royce PM & Barnes MJ (1985) Failure <strong>of</strong> highly purified lysyl hydroxylase to hydroxylate<br />
lysyl residues <strong>in</strong> the non-helical regions <strong>of</strong> collagen. Biochem J 230(2): 475–480.<br />
Rub<strong>in</strong> E & Farber JL (1999) Pathology. 3rd ed. Lipp<strong>in</strong>cott-Raven Publishers, Philadelphia,<br />
37–75.<br />
102
Ruotsala<strong>in</strong>en H, Sipilä L, Kerkelä E, Pospiech H & Myllylä R (1999) <strong>Characterization</strong> <strong>of</strong><br />
cDNAs for mouse lysyl hydroxylase 1, 2 <strong>and</strong> 3, <strong>their</strong> phylogenetic analysis <strong>and</strong><br />
tissuespecific expression <strong>in</strong> the mouse. Matrix Biol 18(3): 325–329.<br />
Ruotsala<strong>in</strong>en H, Sipilä L, Vapola M, Sormunen R, Salo AM, Uitto L, Mercer DK, Rob<strong>in</strong>s<br />
SP, Risteli M, Aszodi A, Fässler R & Myllylä R (2006) Glycosylation catalyzed by<br />
lysyl hydroxylase 3 is essential for basement membranes. J Cell Sci 119(Pt 4): 625–<br />
635.<br />
Russell RG (2006) Bisphosphonates: from bench to bedside. Ann N Y Acad Sci 1068:<br />
367–401.<br />
Saito M & Marumo K (2010) Collagen cross-l<strong>in</strong>ks as a determ<strong>in</strong>ant <strong>of</strong> bone quality: a<br />
possible explanation for bone fragility <strong>in</strong> ag<strong>in</strong>g, osteoporosis, <strong>and</strong> diabetes mellitus.<br />
Osteoporos Int 21(2): 195–214.<br />
Salo AM, Cox H, Farndon P, Moss C, Gr<strong>in</strong>dulis H, Risteli M, Rob<strong>in</strong>s SP & Myllylä R<br />
(2008) A connective tissue disorder caused by mutations <strong>of</strong> the lysyl hydroxylase 3<br />
gene. Am J Hum Genet 83(4): 495–503.<br />
Salo AM, Sipilä L, Sormunen R, Ruotsala<strong>in</strong>en H, Va<strong>in</strong>io S & Myllylä R (2006b) The lysyl<br />
hydroxylase is<strong>of</strong>orms are widely expressed dur<strong>in</strong>g mouse embryogenesis, but obta<strong>in</strong><br />
tissue- <strong>and</strong> cell-specific patterns <strong>in</strong> the adult. Matrix Biol 25(8): 475–483.<br />
Salo AM, Wang C, Sipilä L, Sormunen R, Vapola M, Kerv<strong>in</strong>en P, Ruotsala<strong>in</strong>en H,<br />
Heikk<strong>in</strong>en J & Myllylä R (2006a) <strong>Lysyl</strong> hydroxylase 3 (LH3) modifies prote<strong>in</strong>s <strong>in</strong> the<br />
extracellular space, a novel mechanism for matrix remodel<strong>in</strong>g. J Cell Physiol 207(3):<br />
644–653.<br />
Samimi A & Last JA (2001) Mechanism <strong>of</strong> <strong>in</strong>hibition <strong>of</strong> lysyl hydroxylase activity by the<br />
organophosphates malathion <strong>and</strong> malaoxon. Toxicol Appl Pharmacol 176(3): 181–186.<br />
Savola<strong>in</strong>en ER, Kero M, Pihlajaniemi T & Kivirikko KI (1981) Deficiency <strong>of</strong><br />
galactosylhydroxylysyl glucosyltransferase, an enzyme <strong>of</strong> collagen synthesis, <strong>in</strong> a<br />
family with dom<strong>in</strong>ant epidermolysis bullosa simplex. N Engl J Med 304(4): 197–204.<br />
Schegg B, Hülsmeier AJ, Rutschmann C, Maag C & Hennet T (2009) Core glycosylation<br />
<strong>of</strong> collagen is <strong>in</strong>itiated by two beta(1-O)galactosyltransferases. Mol Cell Biol 29(4):<br />
943–952.<br />
Scheurer SB, Rybak JN, Rösli C, Neri D & Elia G (2004) Modulation <strong>of</strong> gene expression<br />
by hypoxia <strong>in</strong> human umbilical cord ve<strong>in</strong> endothelial cells: A transcriptomic <strong>and</strong><br />
proteomic study. Proteomics 4(6): 1737–1760.<br />
Sch<strong>of</strong>ield CJ & Ratcliffe PJ (2004) Oxygen sens<strong>in</strong>g by HIF <strong>hydroxylases</strong>. Nat Rev Mol<br />
Cell Biol 5(5): 343–354.<br />
Schneider VA & Granato M (2006) The myotomal diwanka (lh3) glycosyltransferase <strong>and</strong><br />
type XVIII collagen are critical for motor growth cone migration. Neuron 50(5): 683–<br />
695.<br />
Schneider VA & Granato M (2007) Genomic structure <strong>and</strong> embryonic expression <strong>of</strong><br />
zebrafish lysyl hydroxylase 1 <strong>and</strong> lysyl hydroxylase 2. Matrix Biol 26(1): 12–19.<br />
Schnieke A, Harbers K & Jaenisch R (1983) Embryonic lethal mutation <strong>in</strong> mice <strong>in</strong>duced<br />
by retrovirus <strong>in</strong>sertion <strong>in</strong>to the alpha 1(I) collagen gene. Nature 304(5924): 315–320.<br />
103
Semenza GL (2011) Oxygen sens<strong>in</strong>g, homeostasis, <strong>and</strong> disease. N Engl J Med 365(6):<br />
537–547.<br />
Sepp<strong>in</strong>en L & Pihlajaniemi T (2011) The multiple functions <strong>of</strong> collagen XVIII <strong>in</strong><br />
development <strong>and</strong> disease. Matrix Biol 30(2): 83–92.<br />
Seth P, Walker LC & Yeowell HN (2011) Identification <strong>of</strong> exonic cis-elements regulat<strong>in</strong>g<br />
the alternative splic<strong>in</strong>g <strong>of</strong> scleroderma-associated lysyl hydroxylase 2 mRNA. J Invest<br />
Dermatol 131(2): 537–539.<br />
Seth P & Yeowell HN (2010) Fox-2 prote<strong>in</strong> regulates the alternative splic<strong>in</strong>g <strong>of</strong><br />
scleroderma-associated lysyl hydroxylase 2 messenger RNA. Arthritis Rheum 62(4):<br />
1167–1175.<br />
Shaheen R, Al-Owa<strong>in</strong> M, Sakati N, Alzayed ZS & Alkuraya FS (2010) FKBP10 <strong>and</strong> Bruck<br />
syndrome: phenotypic heterogeneity or call for reclassification? Am J Hum Genet<br />
87(2): 306–307. Erratum <strong>in</strong> Am J Hum Genet 87(4): 571.<br />
Shoulders MD & Ra<strong>in</strong>es RT (2009) Collagen structure <strong>and</strong> stability. Annu Rev Biochem<br />
78: 929–958.<br />
Sipilä L, Ruotsala<strong>in</strong>en H, Sormunen R, Baker NL, Lam<strong>and</strong>é SR, Vapola M, Wang C, Sado<br />
Y, Aszodi A & Myllylä R (2007) Secretion <strong>and</strong> assembly <strong>of</strong> type IV <strong>and</strong> VI collagens<br />
depend on glycosylation <strong>of</strong> hydroxylys<strong>in</strong>es. J Biol Chem 282(46): 33381–33388.<br />
Sipilä L, Szatanik M, Va<strong>in</strong>ionpää H, Ruotsala<strong>in</strong>en H, Myllylä R & Guénet JL (2000) The<br />
genes encod<strong>in</strong>g mouse lysyl hydroxylase is<strong>of</strong>orms map to chromosomes 4,5, <strong>and</strong> 9.<br />
Mamm Genome 11(12): 1132–1134.<br />
Sricholpech M, Perdivara I, Nagaoka H, Yokoyama M, Tomer KB & Yamauchi M (2011)<br />
<strong>Lysyl</strong> hydroxylase 3 glucosylates galactosylhydroxylys<strong>in</strong>e residues <strong>in</strong> type I collagen<br />
<strong>in</strong> osteoblast culture. J Biol Chem 286(11): 8846–8856.<br />
Ste<strong>in</strong>mann B, Eyre DR & Shao P (1995) Ur<strong>in</strong>ary pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks <strong>in</strong> Ehlers-Danlos<br />
syndrome type VI. Am J Hum Genet 57(6): 1505–1508.<br />
Ste<strong>in</strong>mann B, Royce PM & Superti-Furga A (2002) The Ehlers-Danlos syndrome. In:<br />
Royce PM & Ste<strong>in</strong>mann B (eds) Connective tissue <strong>and</strong> its heritable disorders. 2nd ed.<br />
Wiley-Liss, Inc., New York, 431–523.<br />
Suokas M, Lampela O, Juffer AH, Myllylä R & Kellokumpu S (2003) Retrieval<strong>in</strong>dependent<br />
localization <strong>of</strong> lysyl hydroxylase <strong>in</strong> the endoplasmic reticulum via a<br />
peptide fold <strong>in</strong> its iron-b<strong>in</strong>d<strong>in</strong>g doma<strong>in</strong>. Biochem J 370(Pt 3): 913–920.<br />
Suokas M, Myllylä R & Kellokumpu S (2000) A s<strong>in</strong>gle C-term<strong>in</strong>al peptide segment<br />
mediates both membrane association <strong>and</strong> localization <strong>of</strong> lysyl hydroxylase <strong>in</strong> the<br />
endoplasmic reticulum. J Biol Chem 275(23): 17863–17868.<br />
Susic D (2007) Cross-l<strong>in</strong>k breakers as a new therapeutic approach to cardiovascular<br />
disease. Biochem Soc Trans 35(Pt 5): 853–856.<br />
Svensson L, Aszódi A, Re<strong>in</strong>holt FP, Fässler R, He<strong>in</strong>egård D & Oldberg A (1999)<br />
Fibromodul<strong>in</strong>-null mice have abnormal collagen fibrils, tissue organization, <strong>and</strong><br />
altered lumican deposition <strong>in</strong> tendon. J Biol Chem 274(14): 9636–9647.<br />
Szpirer C, Szpirer J, Rivière M, Vanvooren P, Valtavaara M & Myllylä R (1997)<br />
Localization <strong>of</strong> the gene encod<strong>in</strong>g a novel is<strong>of</strong>orm <strong>of</strong> lysyl hydroxylase. Mamm<br />
Genome 8(9): 707–708.<br />
104
Söderhäll C, Marenholz I, Kerscher T, Rüschendorf F, Esparza-Gordillo J, Worm M,<br />
Gruber C, Mayr G, Albrecht M, Rohde K, Schulz H, Wahn U, Hubner N & Lee YA<br />
(2007) Variants <strong>in</strong> a novel epidermal collagen gene (COL29A1) are associated with<br />
atopic dermatitis. PLoS Biol 5(9): 1952–1961.<br />
Takaluoma K, Lantto J & Myllyharju J (2007) <strong>Lysyl</strong> hydroxylase 2 is a specific<br />
telopeptide hydroxylase, while all three isoenzymes hydroxylate collagenous<br />
sequences. Matrix Biol 26(5): 396–403.<br />
Torre-Blanco A, Adachi E, Hojima Y, Wootton JA, M<strong>in</strong>or RR & Prockop DJ (1992)<br />
Temperature-<strong>in</strong>duced post-translational over-modification <strong>of</strong> type I procollagen.<br />
Effects <strong>of</strong> over-modification <strong>of</strong> the prote<strong>in</strong> on the rate <strong>of</strong> cleavage by procollagen Nprote<strong>in</strong>ase<br />
<strong>and</strong> on self-assembly <strong>of</strong> collagen <strong>in</strong>to fibrils. J Biol Chem 267(4): 2650–<br />
2655.<br />
Trojanowska M, LeRoy EC, Eckes B & Krieg T (1998) Pathogenesis <strong>of</strong> fibrosis: type 1<br />
collagen <strong>and</strong> the sk<strong>in</strong>. J Mol Med 76(3–4): 266–274.<br />
Tsang KY, Cheung MC, Chan D & Cheah KS (2010) The developmental roles <strong>of</strong> the<br />
extracellular matrix: beyond structure to regulation. Cell Tissue Res 339(1): 93–110.<br />
Turpeenniemi-Hujanen TM, Puistola U & Kivirikko KI (1980) Isolation <strong>of</strong> lysyl<br />
hydroxylase, an enzyme <strong>of</strong> collagen synthesis, from chick embryos as a homogeneous<br />
prote<strong>in</strong>. Biochem J 189(2): 247–253.<br />
Turpeenniemi-Hujanen TM, Puistola U & Kivirikko KI (1981) Human lysyl hydroxylase:<br />
purification to homogeneity, partial characterization <strong>and</strong> comparison <strong>of</strong> catalytic<br />
properties with those <strong>of</strong> a mutant enzyme from Ehlers-Danlos syndrome type VI<br />
fibroblasts. Coll Relat Res 1(4): 355–366.<br />
Unligil UM, Zhou S, Yuwaraj S, Sarkar M, Schachter H & R<strong>in</strong>i JM (2000) X-ray crystal<br />
structure <strong>of</strong> rabbit N-acetylglucosam<strong>in</strong>yltransferase I: catalytic mechanism <strong>and</strong> a new<br />
prote<strong>in</strong> superfamily. EMBO J 19(20): 5269–5280.<br />
Uzawa K, Grzesik WJ, Nishiura T, Kuznetsov SA, Robey PG, Brenner DA & Yamauchi<br />
M (1999) Differential expression <strong>of</strong> human lysyl hydroxylase genes, lys<strong>in</strong>e<br />
hydroxylation, <strong>and</strong> cross-l<strong>in</strong>k<strong>in</strong>g <strong>of</strong> type I collagen dur<strong>in</strong>g osteoblastic differentiation<br />
<strong>in</strong> vitro. J Bone M<strong>in</strong>er Res 14(8): 1272–1280.<br />
Uzawa K, Yeowell HN, Yamamoto K, Mochida Y, Tanzawa H & Yamauchi M (2003)<br />
Lys<strong>in</strong>e hydroxylation <strong>of</strong> collagen <strong>in</strong> a fibroblast cell culture system. Biochem Biophys<br />
Res Commun 305(3): 484–487.<br />
Valtavaara M, Papponen H, Pirttilä AM, Hiltunen K, Hel<strong>and</strong>er H & Myllylä R (1997)<br />
Clon<strong>in</strong>g <strong>and</strong> characterization <strong>of</strong> a novel human lysyl hydroxylase is<strong>of</strong>orm highly<br />
expressed <strong>in</strong> pancreas <strong>and</strong> muscle. J Biol Chem 272(11): 6831–6834.<br />
Valtavaara M, Szpirer C, Szpirer J & Myllylä R (1998) Primary structure, tissue<br />
distribution, <strong>and</strong> chromosomal localization <strong>of</strong> a novel is<strong>of</strong>orm <strong>of</strong> lysyl hydroxylase<br />
(lysyl hydroxylase 3). J Biol Chem 273(21): 12881–12886.<br />
105
Vanacore RM, Friedman DB, Ham AJ, Sundaramoorthy M & Hudson BG (2005)<br />
Identification <strong>of</strong> S-hydroxylysyl-methion<strong>in</strong>e as the covalent cross-l<strong>in</strong>k <strong>of</strong> the<br />
noncollagenous (NC1) hexamer <strong>of</strong> the alpha1alpha1alpha2 collagen IV network: a<br />
role for the post-translational modification <strong>of</strong> lys<strong>in</strong>e 211 to hydroxylys<strong>in</strong>e 211 <strong>in</strong><br />
hexamer assembly. J Biol Chem 280(32): 29300–29310.<br />
Vanacore R, Ham AJ, Voehler M, S<strong>and</strong>ers CR, Conrads TP, Veenstra TD, Sharpless KB,<br />
Dawson PE & Hudson BG (2009) A sulfilim<strong>in</strong>e bond identified <strong>in</strong> collagen IV.<br />
Science 325(5945): 1230–1234.<br />
Van Agtmael T & Bruckner-Tuderman L (2010) Basement membranes <strong>and</strong> human disease.<br />
Cell Tissue Res 339(1): 167–188.<br />
van den Bos T, Speijer D, Bank RA, Brömme D & Everts V (2008) Differences <strong>in</strong> matrix<br />
composition between calvaria <strong>and</strong> long bone <strong>in</strong> mice suggest differences <strong>in</strong><br />
biomechanical properties <strong>and</strong> resorption: Special emphasis on collagen. Bone 43(3):<br />
459–468.<br />
van der Slot-Verhoeven AJ, van Dura EA, Attema J, Blauw B, Degroot J, Huiz<strong>in</strong>ga TW,<br />
Zuurmond AM & Bank RA (2005) The type <strong>of</strong> collagen cross-l<strong>in</strong>k determ<strong>in</strong>es the<br />
reversibility <strong>of</strong> experimental sk<strong>in</strong> fibrosis. Biochim Biophys Acta 1740(1): 60–67.<br />
van der Slot AJ, van Dura EA, de Wit EC, De Groot J, Huiz<strong>in</strong>ga TW, Bank RA &<br />
Zuurmond AM (2005) Elevated formation <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks by pr<strong>of</strong>ibrotic<br />
cytok<strong>in</strong>es is associated with enhanced lysyl hydroxylase 2b levels. Biochim Biophys<br />
Acta 1741(1–2): 95–102.<br />
van der Slot AJ, Zuurmond AM, Bardoel AF, Wijmenga C, Pruijs HE, Sillence DO,<br />
Br<strong>in</strong>ckmann J, Abraham DJ, Black CM, Verzijl N, DeGroot J, Hanemaaijer R,<br />
TeKoppele JM, Huiz<strong>in</strong>ga TW & Bank RA (2003) Identification <strong>of</strong> PLOD2 as<br />
telopeptide lysyl hydroxylase, an important enzyme <strong>in</strong> fibrosis. J Biol Chem 278(42):<br />
40967–40972.<br />
van der Slot AJ, Zuurmond AM, van den Bogaerdt AJ, Ulrich MM, Middelkoop E, Boers<br />
W, Karel Ronday H, DeGroot J, Huiz<strong>in</strong>ga TW & Bank RA (2004) Increased<br />
formation <strong>of</strong> pyrid<strong>in</strong>ol<strong>in</strong>e cross-l<strong>in</strong>ks due to higher telopeptide lysyl hydroxylase<br />
levels is a general fibrotic phenomenon. Matrix Biol 23(4): 251–257.<br />
van Vlimmeren MA, Driessen-Mol A, van den Broek M, Bouten CV & Baaijens FP (2010)<br />
Controll<strong>in</strong>g matrix formation <strong>and</strong> cross-l<strong>in</strong>k<strong>in</strong>g by hypoxia <strong>in</strong> cardiovascular tissue<br />
eng<strong>in</strong>eer<strong>in</strong>g. J Appl Physiol 109(5): 1483–1491.<br />
Veit G, Kobbe B, Keene DR, Paulsson M, Koch M & Wagener R (2006) Collagen XXVIII,<br />
a novel von Willebr<strong>and</strong> factor A doma<strong>in</strong>-conta<strong>in</strong><strong>in</strong>g prote<strong>in</strong> with many imperfections<br />
<strong>in</strong> the collagenous doma<strong>in</strong>. J Biol Chem 281(6): 3494–3504.<br />
Viljoen D, Versfeld G & Beighton P (1989) Osteogenesis imperfecta with congenital jo<strong>in</strong>t<br />
contractures (Bruck syndrome). Cl<strong>in</strong> Genet 36(2): 122–126.<br />
Vranka JA, Sakai LY & Bäch<strong>in</strong>ger HP (2004) Prolyl 3-hydroxylase 1, enzyme<br />
characterization <strong>and</strong> identification <strong>of</strong> a novel family <strong>of</strong> enzymes. J Biol Chem 279(22):<br />
23615–23621.<br />
106
Walker LC, Mar<strong>in</strong>i JC, Grange DK, Filie J & Yeowell HN (1999) A patient with Ehlers-<br />
Danlos syndrome type VI is homozygous for a premature term<strong>in</strong>ation codon <strong>in</strong> exon<br />
14 <strong>of</strong> the lysyl hydroxylase 1 gene. Mol Genet Metab 67(1): 74–82.<br />
Walker LC, Overstreet MA, Will<strong>in</strong>g MC, Mar<strong>in</strong>i JC, Cabral WA, Pals G, Bristow J,<br />
Atsawasuwan P, Yamauchi M & Yeowell HN (2004b) Heterogeneous basis <strong>of</strong> the<br />
type VIB form <strong>of</strong> Ehlers-Danlos syndrome (EDS VIB) that is unrelated to decreased<br />
collagen lysyl hydroxylation. Am J Med Genet A 131(2): 155–162.<br />
Walker LC, Overstreet MA & Yeowell HN (2005) Tissue-specific expression <strong>and</strong><br />
regulation <strong>of</strong> the alternatively-spliced forms <strong>of</strong> lysyl hydroxylase 2 (LH2) <strong>in</strong> human<br />
kidney cells <strong>and</strong> sk<strong>in</strong> fibroblasts. Matrix Biol 23(8): 515–523.<br />
Walker LC, Teebi AS, Mar<strong>in</strong>i JC, De Paepe A, Malfait F, Atsawasuwan P, Yamauchi M &<br />
Yeowell HN (2004a) Decreased expression <strong>of</strong> lysyl hydroxylase 2 (LH2) <strong>in</strong> sk<strong>in</strong><br />
fibroblasts from three Ehlers-Danlos patients does not result from mutations <strong>in</strong> either<br />
the cod<strong>in</strong>g or proximal promoter region <strong>of</strong> the LH2 gene. Mol Genet Metab 83(4):<br />
312–321.<br />
Wang C, Kovanen V, Raudasoja P, Eskel<strong>in</strong>en S, Pospiech H & Myllylä R (2009) The<br />
glycosyltransferase activities <strong>of</strong> lysyl hydroxylase 3 (LH3) <strong>in</strong> the extracellular space<br />
are important for cell growth <strong>and</strong> viability. J Cell Mol Med 13(3): 508–521.<br />
Wang C, Luosujärvi H, Heikk<strong>in</strong>en J, Risteli M, Uitto L & Myllylä R (2002a) The third<br />
activity for lysyl hydroxylase 3: galactosylation <strong>of</strong> hydroxylysyl residues <strong>in</strong> collagens<br />
<strong>in</strong> vitro. Matrix Biol 21(7): 559–566.<br />
Wang C, Risteli M, Heikk<strong>in</strong>en J, Hussa AK, Uitto L & Myllyla R (2002b) Identification <strong>of</strong><br />
am<strong>in</strong>o acids important for the catalytic activity <strong>of</strong> the collagen glucosyltransferase<br />
associated with the multifunctional lysyl hydroxylase 3 (LH3). J Biol Chem 277(21):<br />
18568–18573.<br />
Wang C, Ristiluoma MM, Salo AM, Eskel<strong>in</strong>en S & Myllylä R (2012) <strong>Lysyl</strong> hydroxylase 3<br />
is secreted from cells by two pathways. J Cell Physiol 227(2): 668–675.<br />
Webby CJ, Wolf A, Gromak N, Dreger M, Kramer H, Kessler B, Nielsen ML, Schmitz C,<br />
Butler DS, Yates JR 3rd, Delahunty CM, Hahn P, Lengel<strong>in</strong>g A, Mann M, Proudfoot<br />
NJ, Sch<strong>of</strong>ield CJ & Böttger A (2009) Jmjd6 catalyses lysyl-hydroxylation <strong>of</strong> U2AF65,<br />
a prote<strong>in</strong> associated with RNA splic<strong>in</strong>g. Science 325(5936): 90–93.<br />
Weis MA, Hudson DM, Kim L, Scott M, Wu JJ & Eyre DR (2010) Location <strong>of</strong> 3hydroxyprol<strong>in</strong>e<br />
residues <strong>in</strong> collagen types I, II, III, <strong>and</strong> V/XI implies a role <strong>in</strong> fibril<br />
supramolecular assembly. J Biol Chem 285(4): 2580–2590.<br />
Wenstrup RJ, Florer JB, Brunskill EW, Bell SM, Chervoneva I & Birk DE (2004) Type V<br />
collagen controls the <strong>in</strong>itiation <strong>of</strong> collagen fibril assembly. J Biol Chem 279(51):<br />
53331–53337.<br />
Wenstrup RJ, Florer JB, Davidson JM, Phillips CL, Pfeiffer BJ, Menezes DW, Chervoneva<br />
I & Birk DE (2006) Mur<strong>in</strong>e model <strong>of</strong> the Ehlers-Danlos syndrome. col5a1<br />
haplo<strong>in</strong>sufficiency disrupts collagen fibril assembly at multiple stages. J Biol Chem<br />
281(18): 12888–12895.<br />
Wenstrup RJ, Murad S & P<strong>in</strong>nell SR (1989) Ehlers-Danlos syndrome type VI: cl<strong>in</strong>ical<br />
manifestations <strong>of</strong> collagen lysyl hydroxylase deficiency. J Pediatr 115(3): 405–409.<br />
107
Wilson R, Lees JF & Bulleid NJ (1998) Prote<strong>in</strong> disulfide isomerase acts as a molecular<br />
chaperone dur<strong>in</strong>g the assembly <strong>of</strong> procollagen. J Biol Chem 273(16): 9637–9643.<br />
Wu J, Re<strong>in</strong>hardt DP, Batmunkh C, L<strong>in</strong>denmaier W, Far RK, Notbohm H, Hunzelmann N<br />
& Br<strong>in</strong>ckmann J (2006) Functional diversity <strong>of</strong> lysyl hydroxylase 2 <strong>in</strong> collagen<br />
synthesis <strong>of</strong> human dermal fibroblasts. Exp Cell Res 312(18): 3485–3494.<br />
Yamauchi M & Katz EP (1993) The post-translational chemistry <strong>and</strong> molecular pack<strong>in</strong>g <strong>of</strong><br />
m<strong>in</strong>eraliz<strong>in</strong>g tendon collagens. Connect Tissue Res 29(2): 81–98.<br />
Yamauchi M, Noyes C, Kuboki Y & Mechanic GL (1982) Collagen structural<br />
microheterogeneity <strong>and</strong> a possible role for glycosylated hydroxylys<strong>in</strong>e <strong>in</strong> type I<br />
collagen. Proc Natl Acad Sci U S A 79(24): 7684–7688.<br />
Yapicioğlu H, Ozcan K, Arikan O, Satar M, Narli N & Ozbek MH (2009) Bruck syndrome:<br />
osteogenesis imperfecta <strong>and</strong> arthrogryposis multiplex congenita. Ann Trop Paediatr<br />
29(2): 159–162.<br />
Yeowell HN, Allen JD, Walker LC, Overstreet MA, Murad S & Thai SF (2000a) Deletion<br />
<strong>of</strong> cyste<strong>in</strong>e 369 <strong>in</strong> lysyl hydroxylase 1 elim<strong>in</strong>ates enzyme activity <strong>and</strong> causes Ehlers-<br />
Danlos syndrome type VI. Matrix Biol 19(1): 37–46.<br />
Yeowell HN, Ha V, Clark WL, Marshall MK & P<strong>in</strong>nell SR (1994) Sequence analysis <strong>of</strong> a<br />
cDNA for lysyl hydroxylase isolated from human sk<strong>in</strong> fibroblasts from a normal<br />
donor: differences from human placental lysyl hydroxylase cDNA. J Invest Dermatol<br />
102(3): 382–384.<br />
Yeowell HN & Walker LC (1997) Ehlers-Danlos syndrome type VI results from a<br />
nonsense mutation <strong>and</strong> a splice site-mediated exon-skipp<strong>in</strong>g mutation <strong>in</strong> the lysyl<br />
hydroxylase gene. Proc Assoc Am Physicians 109(4): 383–396.<br />
Yeowell HN & Walker LC (1999b) Prenatal exclusion <strong>of</strong> Ehlers-Danlos syndrome type VI<br />
by mutational analysis. Proc Assoc Am Physicians 111(1): 57–62.<br />
Yeowell HN & Walker LC (1999a) Tissue specificity <strong>of</strong> a new splice form <strong>of</strong> the human<br />
lysyl hydroxylase 2 gene. Matrix Biol 18(2): 179–187.<br />
Yeowell HN & Walker LC (2000) Mutations <strong>in</strong> the lysyl hydroxylase 1 gene that result <strong>in</strong><br />
enzyme deficiency <strong>and</strong> the cl<strong>in</strong>ical phenotype <strong>of</strong> Ehlers-Danlos syndrome type VI.<br />
Mol Genet Metab 71(1–2): 212–224.<br />
Yeowell HN, Walker LC, Farmer B, Heikk<strong>in</strong>en J & Myllyla R (2000b) Mutational analysis<br />
<strong>of</strong> the lysyl hydroxylase 1 gene (PLOD) <strong>in</strong> six unrelated patients with Ehlers-Danlos<br />
syndrome type VI: prenatal exclusion <strong>of</strong> this disorder <strong>in</strong> one family. Hum Mutat 16(1):<br />
90.<br />
Yeowell HN, Walker LC, Marshall MK, Murad S & P<strong>in</strong>nell SR (1995) The mRNA <strong>and</strong> the<br />
activity <strong>of</strong> lysyl hydroxylase are up-regulated by the adm<strong>in</strong>istration <strong>of</strong> ascorbate <strong>and</strong><br />
hydralaz<strong>in</strong>e to human sk<strong>in</strong> fibroblasts from a patient with Ehlers-Danlos syndrome<br />
type VI. Arch Biochem Biophys 321(2): 510–516.<br />
Yeowell HN, Walker LC, Mauger DM, Seth P & Garcia-Blanco MA (2009) TIA nuclear<br />
prote<strong>in</strong>s regulate the alternate splic<strong>in</strong>g <strong>of</strong> lysyl hydroxylase 2. J Invest Dermatol<br />
129(6): 1402–1411.<br />
108
Yeowell HN, Walker LC, Murad S & P<strong>in</strong>nell SR (1997) A common duplication <strong>in</strong> the<br />
lysyl hydroxylase gene <strong>of</strong> patients with Ehlers Danlos syndrome type VI results <strong>in</strong><br />
preferential stimulation <strong>of</strong> lysyl hydroxylase activity <strong>and</strong> mRNA by hydralaz<strong>in</strong>e. Arch<br />
Biochem Biophys 347(1): 126–131.<br />
Zeng B, MacDonald JR, Bann JG, Beck K, Gambee JE, Boswell BA & Bäch<strong>in</strong>ger HP<br />
(1998) Chicken FK506-b<strong>in</strong>d<strong>in</strong>g prote<strong>in</strong>, FKBP65, a member <strong>of</strong> the FKBP family <strong>of</strong><br />
peptidylprolyl cis-trans isomerases, is only partially <strong>in</strong>hibited by FK506. Biochem J<br />
330(Pt 1): 109–114.<br />
Zhu Y, Oganesian A, Keene DR & S<strong>and</strong>ell LJ (1999) Type IIA procollagen conta<strong>in</strong><strong>in</strong>g the<br />
cyste<strong>in</strong>e-rich am<strong>in</strong>o propeptide is deposited <strong>in</strong> the extracellular matrix <strong>of</strong><br />
prechondrogenic tissue <strong>and</strong> b<strong>in</strong>ds to TGF-beta1 <strong>and</strong> BMP-2. J Cell Biol 144(5): 1069–<br />
1080.<br />
Zuurmond AM, van der Slot-Verhoeven AJ, van Dura EA, De Groot J & Bank RA (2005)<br />
M<strong>in</strong>oxidil exerts different <strong>in</strong>hibitory effects on gene expression <strong>of</strong> lysyl hydroxylase 1,<br />
2, <strong>and</strong> 3: implications for collagen cross-l<strong>in</strong>k<strong>in</strong>g <strong>and</strong> treatment <strong>of</strong> fibrosis. Matrix Biol<br />
24(4): 261–270.<br />
109
110
Orig<strong>in</strong>al articles<br />
I Takaluoma K, Hyry M, Lantto J, Sormunen R, Bank RA, Kivirikko KI, Myllyharju J<br />
& So<strong>in</strong><strong>in</strong>en R (2007) Tissue-specific changes <strong>in</strong> the hydroxylys<strong>in</strong>e content <strong>and</strong> crossl<strong>in</strong>ks<br />
<strong>of</strong> collagens <strong>and</strong> alterations <strong>in</strong> fibril morphology <strong>in</strong> lysyl hydroxylase 1 knockout<br />
mice. J Biol Chem 282(9):6588–6596.<br />
II Hyry M, Lantto J & Myllyharju J (2009) Missense mutations that cause Bruck<br />
syndrome affect enzymatic activity, fold<strong>in</strong>g, <strong>and</strong> oligomerization <strong>of</strong> lysyl hydroxylase<br />
2. J Biol Chem 284(45):30917–30924.<br />
III Hyry M, So<strong>in</strong><strong>in</strong>en R, Bank RA, Sormunen R, Mi<strong>in</strong>ala<strong>in</strong>en I, Talsta HL, Lantto J &<br />
Myllyharju J (2012) Lack <strong>of</strong> lysyl hydroxylase 2 <strong>in</strong> mouse leads to early lethality<br />
dur<strong>in</strong>g embryonic development. Manuscript.<br />
Repr<strong>in</strong>ted with permission from American Society for Biochemistry <strong>and</strong><br />
Molecular Biology.<br />
Orig<strong>in</strong>al publications are not <strong>in</strong>cluded <strong>in</strong> the electrical version <strong>of</strong> the dissertation.<br />
111
112
ACTA UNIVERSITATIS OULUENSIS<br />
SERIES D MEDICA<br />
1143. Karjala<strong>in</strong>en, M<strong>in</strong>na (2011) Genetic predisposition to spontaneous preterm birth :<br />
approaches to identify susceptibility genes<br />
1144. Saaristo, Timo (2011) Assessment <strong>of</strong> risk <strong>and</strong> prevention <strong>of</strong> type 2 diabetes <strong>in</strong><br />
primary health care<br />
1145. Vuononvirta, Ti<strong>in</strong>a (2011) Etäterveydenhuollon käyttöönotto terveydenhuollon<br />
verkostoissa<br />
1146. Vanhala, Marja (2012) Lapsen ylipa<strong>in</strong>o – riskitekijät, tunnistam<strong>in</strong>en ja el<strong>in</strong>tavat<br />
1147. Katisko, Jani (2012) Intraoperative imag<strong>in</strong>g guided del<strong>in</strong>eation <strong>and</strong> localization <strong>of</strong><br />
regions <strong>of</strong> surgical <strong>in</strong>terest : Feasibility study<br />
1148. Holmström, Anneli (2012) Etnograf<strong>in</strong>en tutkimus natiivitutkimusten oppimisesta<br />
röntgenhoitajaopiskelijoiden op<strong>in</strong>noissa<br />
1149. Ronka<strong>in</strong>en, Hanna-Leena (2012) Novel prognostic biomarkers for renal cell<br />
carc<strong>in</strong>oma<br />
1150. Murtomaa-Hautala, Mari (2012) Species-specific effects <strong>of</strong> diox<strong>in</strong> exposure on<br />
xenobiotic metabolism <strong>and</strong> hard tissue <strong>in</strong> voles<br />
1151. Jauhola, Outi (2012) Henoch-Schönle<strong>in</strong> purpura <strong>in</strong> children<br />
1152. Toljamo, Kari (2012) Gastric erosions – cl<strong>in</strong>ical significance <strong>and</strong> pathology : a<br />
long-term follow-up study<br />
1153. Merilä<strong>in</strong>en, Merja (2012) Tehohoitopotilaan hoitoympäristö : psyykk<strong>in</strong>en<br />
elämänlaatu ja toipum<strong>in</strong>en<br />
1154. Liisanantti, Janne (2012) Acute drug poison<strong>in</strong>g: outcome <strong>and</strong> factors affect<strong>in</strong>g<br />
outcome<br />
1155. Markkanen, Piia (2012) Human δ opioid receptor : The effect <strong>of</strong> Phe27Cys<br />
polymorphism, N-l<strong>in</strong>ked glycosylation <strong>and</strong> SERCA2b <strong>in</strong>teraction on receptor<br />
process<strong>in</strong>g <strong>and</strong> traffick<strong>in</strong>g<br />
1156. Hannula, Annukka (2012) Imag<strong>in</strong>g studies <strong>of</strong> the ur<strong>in</strong>ary tract <strong>in</strong> children with<br />
acute ur<strong>in</strong>ary tract <strong>in</strong>fection<br />
1157. Korvala, Johanna (2012) In quest <strong>of</strong> genetic susceptibility to disorders manifest<strong>in</strong>g<br />
<strong>in</strong> fractures : Assess<strong>in</strong>g the significance <strong>of</strong> genetic factors <strong>in</strong> femoral neck stress<br />
fractures <strong>and</strong> childhood non-OI primary osteoporosis<br />
1158. Takala, Heikki (2012) Biomarkers <strong>in</strong> esophageal cancer<br />
Book orders:<br />
Granum: Virtual book store<br />
http://granum.uta.fi/granum/
UNIVERSITY OF OULU P.O.B. 7500 FI-90014 UNIVERSITY OF OULU FINLAND<br />
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<br />
S E R I E S E D I T O R S<br />
A<br />
B<br />
C<br />
D<br />
E<br />
F<br />
G<br />
SCIENTIAE RERUM NATURALIUM<br />
HUMANIORA<br />
TECHNICA<br />
MEDICA<br />
Senior Assistant Jorma Arhippa<strong>in</strong>en<br />
SCIENTIAE RERUM SOCIALIUM<br />
SCRIPTA ACADEMICA<br />
OECONOMICA<br />
Lecturer Santeri Palvia<strong>in</strong>en<br />
Pr<strong>of</strong>essor Hannu Heusala<br />
Pr<strong>of</strong>essor Olli Vuolteenaho<br />
Senior Researcher Eila Estola<br />
Director S<strong>in</strong>ikka Eskel<strong>in</strong>en<br />
Pr<strong>of</strong>essor Jari Juga<br />
EDITOR IN CHIEF<br />
Pr<strong>of</strong>essor Olli Vuolteenaho<br />
PUBLICATIONS EDITOR<br />
Publications Editor Kirsti Nurkkala<br />
ISBN 978-951-42-9841-7 (Paperback)<br />
ISBN 978-951-42-9842-4 (PDF)<br />
ISSN 0355-3221 (Pr<strong>in</strong>t)<br />
ISSN 1796-2234 (Onl<strong>in</strong>e)