The collagen superfamily from the extracellular matrix to cell membrane, kolagen2
[ Pobierz całość w formacie PDF ]
Pathologie Biologie 53 (2005) 430–442
The collagen superfamily:
from the extracellular matrix to the cell membrane
La superfamille des collagènes :
de la matrice extracellulaire à la membrane cellulaire
Sylvie Ricard-Blum *, Florence Ruggiero
Institut de Biologie et Chimie des Protéines, UMR 5086 CNRS UCBL, IFR128 Biosciences Gerland, 7, passage du Vercors, 69367 Lyon cedex 7, France
Received 26 November 2004; accepted 10 December 2004
Available online 20 January 2005
Abstract
The collagen superfamily is highly complex and shows a remarkable diversity in molecular and supramolecular organization, tissue dis-
tribution and function. However, all its members share a common structural feature, the presence of at least one triple-helical domain, which
corresponds to a number of (Gly–X–Y)
n
repeats (X being frequently proline and Y hydroxyproline) in the amino acid sequence. Several
sub-families have been determined according to sequence homologies and to similarities in the structural organization and supramolecular
assembly. In the present review, we focus on the newly described fibrillar collagens, fibrillar-associated collagens with interrupted triple helix,
membrane collagens and multiplexins. Recent advances in the characterization of proteins containing triple-helical domains but not referred
to as collagens are also discussed.
© 2005 Elsevier SAS. All rights reserved.
Résumé
La superfamille des collagènes est particulièrement complexe et présente une remarquable diversité tant au niveau de l’organisation molécu-
laire et supra-moléculaire qu’à celui de la distribution tissulaire et des fonctions biologiques. Néanmoins, tous ses membres partagent une
caractéristique structurale commune, la présence dans leur structure d’un domaine en triple hélice qui correspond à la répétition du triplet
(Gly–X–Y)
n
(X étant fréquemment la proline et Y l’hydroxyproline) dans la séquence d’acides aminés. Les collagènes sont classés en
plusieurs sous-familles déterminées en fonction des homologies de séquence et de similitudes au niveau de l’architecture modulaire des
chaînes polypeptidiques et de leur assemblage supramoléculaire. L’essentiel de cette revue est consacré aux collagènes nouvellement identi-
fiés qui sont des collagènes fibrillaires, des FACITs (associés aux collagènes fibrillaires), les collagènes membranaires et les multiplexines.
Cette revue inclut également une présentation des protéines possédant au moins un domaine en triple hélice mais non référencés pour l’instant
comme étant des collagènes.
© 2005 Elsevier SAS. All rights reserved.
Keywords:
Collagen; Collagen-related diseases; Extracellular domains; Modular proteins
Mots clés :
Collagène ; Domaines extracellulaires ; Protéines modulaires ; Pathologies liées au collagène
Abbreviations:
BMP, bone morphogenetic protein; COL, collagenous domain; CRR, cystein-rich repeat domain; EMI, N-terminal cysteine-rich domain
found in the Emu family; Emu, Emilin and multimerin; Erk, extracellular-signal-regulated kinase; FAK, focal adhesion kinase; FACIT, fibril-associated colla-
gen with interrupted triple helix; FN3, fibronectin type III repeat; MARCO, macrophage receptor with collagenous structure, NC, non collagenous domain;
SR-Ai, SR-AII, scavenger receptors type A; TGF-b, transforming growth factor b; TSPN, thrombospondin N-terminal-like domain; vWA, von Willlebrand
factor A-like domain.
* Corresponding author.
E-mail address:
s.ricard-blum@ibcp.fr (S. Ricard-Blum).
0369-8114/$ - see front matter © 2005 Elsevier SAS. All rights reserved.
doi:10.1016/j.patbio.2004.12.024
S. Ricard-Blum, F. Ruggiero / Pathologie Biologie 53 (2005) 430–442
431
1. Introduction
extracellular matrix proteins
and are thus modular pro-
teins. Since many collagen chains were initially character-
ized by partial sequencing, missing the sequences encoding
the N-terminus of the protein, investigators in the field num-
bered these domains starting from the C-terminus. However,
the reverse order is also found in the literature for collagens
VII, XIII, XV, XVIII, XXII
XXIII
XXV
and XXVI
In this review, we will focus on recent advances in the field
of collagens (
. For a comprehensive coverage of the
structure and functions of well-known collagens, the reader
is referred to previous reviews on fibrillar collagens
unconventional collagens including collagen VII, network-
forming collagens (VI, VIII and X) and the fibril-associated
collagen with interrupted triple helix (FACITs) (IX, XII, XIV,
XVI and XIX)
and to recent reviews on the collagen super-
family
Collagens are trimeric molecules composed of three
polypeptide a chains, which contain the sequence repeat
(G–X–Y)
n
, X being frequently proline and Y hydroxypro-
line. These repeats allow the formation of a triple helix, which
is the characteristic structural feature of the collagen super-
family. Each member of the collagen family contains at least
one triple-helical domain (COL), which is located in the extra-
cellular matrix, and most collagens are able to form supramo-
lecular aggregates. However, some molecules such as emi-
lins, which do contain a collagenous domain and are
associated with extracellular supramolecular assemblies (elas-
tic fibers), are surprisingly not classified as collagens.
Besides triple-helical domains, collagens contain non
triple-helical (NC) domains, used as building blocks by other
Table 1
Swiss Prot or TrEMBL accession number (
are provided for non fibrillar a chains. Entries are provided for human proteins,
unless otherwise stated. For accession number of other collagen chains and collagen-like chains, the reader is refreed to previous reviews
Les numéros d’accès Swiss Prot ou TrEMBL
sont fournis pour les chaînes alpha non fibrillaires. Les entrées sont fournies
pour les protéines humaines, à moins qu’il ne soit indiqué autrement. Pour les numéros d’accès d’autres chaînes de collagène et les chaînes collagène-like, le
lecteur doit se référer à des revues antérieures
Localization
Collagen type
Supramolecular assembly
Basement membrane associated
collagens
Collagen IV
a1(IV) P02462
a2(IV) P08572
a3(IV) Q01955
a4(IV) P53420
a5(IV) P29400
a6(IV) Q14031
Collagen VIII
Short chain collagen, hexagonal network
a1(VIII) P27658
a2(VIII) P25067
Collagen XV
Multiplexin, supramolecular assembly unknown
a1(XV) P39059
Collagen XVIII
Multiplexin, supramolecular assembly unknown
a1(XVIII) Q14035 TrEMBL
Membrane collagens
Collagen XIII
Type II membrane protein
a1(XIII)
Collagen XVII (BP180: 180 kDa Bullous Pemphigoid antigen 2) Type II membrane protein
a1(XVII) Q9NQK9
Q9UMD9
Collagen XXIII
Type II membrane protein
a1(XIII) Q86Y22 TrEMBL
Collagen XXV (CLAC/P) (AMY protein)
Type II membrane protein
a1(XXV) Q99MQ5 TrEMBL (Murine)
Ubiquitous
Collagen VI
Beaded filaments
a1(VI) P12109
a2(VI) P12110
a3(VI) P12111
Dermo-epidermal junction
Collagen VII
Anchoring fibrils
a1(VII) Q02388
Hypertrophic cartilage
Collagen X
Short chain collagen, hexagonal network
a1(X) Q03692
Ovary, testis
Collagen XXVI (Emu2 protein, Emu-domain containing protein
2)
Supramolecular assembly unknown
a1(XXVI) Q96A83
432
S. Ricard-Blum, F. Ruggiero / Pathologie Biologie 53 (2005) 430–442
Fig. 1
.
Ultrastuctural analysis of collagen fibrils. A, negative staining revealing the periodic striation of the fibrils. B, rotary shadowing showing collagen IX
molecules that decorates the surface of cartilaginous fibrils. C, atomic force microscopy enlightening the microfibrils that composed large fibers (Courtesy of
J.-M. Franc) (printed from Ref.
with permission).
Fig. 1. Etude ultrastructurale des fibres de collagène. A, coloration négative montrant les striations périodiques des fibrilles ; B, ombrage tournant montrant les
molécules de collagène IX décorant la surface des fibrilles du cartilage ; C, microscopie à force atomique montrant les mirofibrilles formant les larges fibres
(communiqué par J.M. Franc) (reproduit de la Réf.
avec permission).
2. The fibrillar collagens and their associated collagens:
the FACITs
The fibril-forming or fibrillar collagens represent the first
members of the collagen superfamily to be discovered. They
represent the major products synthesized by connective tis-
sue cells (fibroblasts, osteoblasts and chondrocytes). The
fibrillar collagens comprised five members: the quantita-
tively major fibrillar collagens types I, II and III, and the minor
collagens types V and XI
Because of their relatively
simple structure (a large continuous triple helix flanked by
two terminal globular extensions) and because they represent
the first discovered members of the collagen superfamily,
these former collagens are also called the classical collagens.
They share the unique property to aggregate into highly-
ordered fibrils, which present a banded pattern at the ultra-
structural level
. The recent discovery of two addi-
tional novel members, collagens XXIV and XXVII, has
provided some renewal of interest for this subclass of col-
lagens. In the late eighties, it becomes clear that fibrils are
made with a mixture of major and minor fibrillar collagens
(collagens I, III and V in most connective tissues and col-
lagen II and XI in cartilage). The relative composition and
amounts of the different collagen types that co-assemble into
heterotypic fibrils govern the structure and organization of
the matrix network and consequently determine the biome-
chanical properties of the tissues (
. Soon after, a new
subclass of collagens characterized by the interruptions within
the triple helix domain was identified and it was shown that
they also interact with fibrils to form more complex alloys
than previously thought. Collagen IX was shown to covalently
interact with collagen II fibrils, whereas collagens XII and
XIV were localized at the surface of collagen I/III/V fibrils in
skin and tendon, respectively. Based on their common chain
structure and their fibril association, the name of FACIT was
given to this subclass of collagens. Recently, this collagen
subfamily has expanded to eight members including the novel
collagens XVI, XIX, XX, XXI and the most recently reported
collagen XXII
making the FACITs the largest subclass of
the collagen superfamily.
Fig. 2
.
The relative composition and amount of the different collagens
influence the matrix organization of connective tissues. The non oriented
thin fibrils present in cartilage are composed of collagen II/XI and IX. Ten-
don contains parallel large fibril bundles, which are essentially made of col-
lagen I with small amounts of collagen III and V. Orthogonal array of thin
and regulated fibrils in cornea contains unusual high amounts of collagen V.
The heterogenous size of fibrils found in dermis are composed of a mixture
of collagens I and III with small amounts of collagen V.
Fig. 2. La composition relative et la proportion des différents collagènes
influencent l’organisation de la matrice des tissus conjonctifs. Les fibres fines,
non-orientées du cartilage sont composées des collagènes II/XI et IX. Les
tendons contiennent de larges faisceaux fibrillaires à orientation parallèle,
composés essentiellement de collagène I avec de faibles quantités de colla-
gènes III et V. Les arrangements orthogonaux des fines et régulières fibrilles
de la cornée contiennent une proportion élevée de collagène V. Les fibres
hétérogènes du derme sont composées d’un mélange de collagène I et III
avec de faibles quantités de collagène V.
2.1. The fibril-forming collagens
The classical fibrillar collagen chains can associate into
homotrimers (collagens II and III), heterotrimers (collagen
XI) or both (collagens I and V) depending on tissue localiza-
tion. Collagen V molecular forms are the most heteroge-
neous among the fibrillar collagens. Besides the heterotrimer
[a1(V)]2a2(V) found in most tissues, several other different
chain associations exist for collagen V including hybrid mol-
ecules formed with both collagen V and XI chains
A
novel member of collagen V gene family referred to as a4(V)
S. Ricard-Blum, F. Ruggiero / Pathologie Biologie 53 (2005) 430–442
433
was recently described in the rat
There is a near identity
between the rat proa4(V) chain and the human proa3(V) chain
but the two chains differ by several features. Whereas the
a3(V) chain is present in placenta as the heterotrimer
a1(V)a2(V)a3(V), the a4(V)-containing molecule shows a
different heterotrimeric chain association [a1(V)]
2
a4(V).
Moreover, proa4(V) chain is solely synthesized by Schwann
cells and its expression is restricted to embryonic and early
postnatal peripheral nerves. The a4(V)-collagen was shown
to display domain-specific effects that regulate peripheral
nerves development
Collagen members of a given subclass share a common
chain structure. The fibril-forming collagens are all com-
posed of a large continuous triple helix (COL1) bordered by
the N- and C-terminal extensions called the N- and
C-propeptides respectively. The C-propeptide is referred to
as the NC1 domain, the N-propeptide is divided into sub-
domains: a short sequence (NC2) that links the major triple
helix to the minor one (COL2), and a globular N-terminal
end (NC3). Whereas the NC1 domains are well conserved
among the fibrillar collagen members, the NC3 domains show
significant variations. They vary in size (from about 200 resi-
dues to up to 500), in their primary structure and in the com-
position of their subdomains (
). The NC3 domains of
a1(I), a1(IIA), a1(III) and a2(V) chains all harbor a cystein-
rich repeat domain (CRR) and the a1(V), a3/a4(V), a1(XI),
a2(XI), a1(XXIV) and a1(XXVII) chains contain a throm-
bospondin N-terminal-like (TSPN) domain. It can also be sub-
jected to alternative splicing as shown for collagens II and XI
The CRRs are characterized by the presence of 10 cys-
teine residues in their sequence, which form disulfide bonds
The TSPN domain (~200 residues) was first described
in the N-terminal domain of the thrompobondin-1 but is fre-
quently encountered in the collagen superfamily members.
The fibrillar collagen V, XI, XXIV and XXVII, the FACITs
and the multiplexins harbor a TSPN domain in their
N-terminal extremities (
B). In fibrillar collagens, the
TSPN domain is adjacent to a domain called the variable
region (VR).
The fibrillar collagens are synthesized as precursors, the
procollagens, which are secreted in the extracellular space. It
was commonly admitted that procollagens require pro-
teolytic removal of the N- and C-terminal propeptides to be
fully processed into mature and functional molecules. How-
ever, it turned out that this holds true only for collagens I and
III. Actually, the NC3 domain can undergo partial proteolytic
cleavage after procollagen secretion or can be entirely retained
in the mature molecule as for the proa1(V) and proa2(V)
chains respectively. The significance of the persistence of the
N-propeptide in the mature molecule is still unknown. One
clearly established function of the collagen V and XI
N-propeptides is to contribute to the control of heterotypic
fibril growth by sterically limiting lateral molecule addition
(see Ref.
for review). This flexible domain, that persists
on the mature molecule, emerges at the heterotypic fibril sur-
face and thus, unlike the triple helix domain buried within
the fibrils, could interact with receptors at the cell surface. A
number of collagen V receptors have been already identified
(e.g.
but so far they all bind to the collagen V major
triple helix. However, the a4(V) N-propeptide was recently
shown to interact with the heparan sulfate transmembrane pro-
teoglycan syndecan 3, a finding important for its biological
function. Clearly, much work should be done to elucidate the
role of the fribrillar collagen N-propeptides. One possible
alternative is to approach the function of the N-propeptide by
analyzing the role of its subdomains as illustrated by recent
work.
The CRR domains are present in several copies in two
homologous proteins the
Xenopus
chordin and the
Droso-
phila
sog and bind respectively to bone morphogenetic pro-
tein (BMP)-4 and decapentaplegic, both members of the trans-
forming growth factor b (TGF-b) superfamily. The release of
these growth factors by xolloid metalloproteinase activity
establishes a gradient of available morphogens that play a
crucial role in dorsal-ventral patterning. Interestingly, BMP-
2 and TGF-b1 were shown to bind to the recombinant trim-
eric collagen IIA N-propeptide. This strongly suggests a role
of the collagen IIA N-propeptide in the regulation of growth
factor delivery during chondrogenesis
Conversely, mice
that harbored a deletion of the
col1a1
exon 2 encoding the
collagen I CRR-containing domain developed quite nor-
mally
This finding raises questions on the exact role of
the N-propeptide of collagen I, thought to be involved in vari-
ous collagen biogenesis events (molecular assembly, feed-
back regulation of procollagen synthesis and fibrillogenesis).
Little is known about the function of the TSPN domain of the
fibrillar collagens. The TSPN domain is released during pro-
collagen N-terminal maturation of collagens V and XI but
the biological relevance of this partial processing has not been
elucidated yet. However, a spontaneous mutation in human
COL5A1 gene that abolished the release of TSPN that nor-
mally occurred in collagen V maturation, disturbed normal
fibrillogenesis and caused the classical form of the Elhers–
Danlos syndrome
This finding attests for an important
function of this subdomain in fibrillogenesis.
With the completion of the genome sequencing, two novel
members of the fibrillar collagen group, numbered collagens
XXIV
(accession number in TrEMBL database Q7Z5L5)
and XXVII
(accession number in TrEMBL database
Q8IZC6) were identified. Like the classical fibril-forming col-
lagens, they both contain a long collagenous domain flanked
by globular N- and C- propeptides. Their large amino-
terminal domains (up to 550 residues) comprise two subdo-
mains, a TSPN domain and a VR as for collagens V and XI
(
A). Their C-propeptides contain eight cysteine resi-
dues that make possible homotrimeric association of these
collagens. Interestingly, some features of collagens XXIV and
XXVII are shared with invertebrate fibrillar collagens
The major triple helix is shorter (991–997 amino acids
depending on the collagen chains and species) and contains
imperfections in the (G–X–Y)
n
repeats. Mutations in fibril-
lar collagen genes that generate minor interruptions in the
434
S. Ricard-Blum, F. Ruggiero / Pathologie Biologie 53 (2005) 430–442
Fig. 3
.
Structure of the a chain of fibrillar collagens (A), FACITS (B) (printed from Ref.
with permission) and other collagens including membrane
collagens, multiplexins and collagen XXVI (C).
Fig. 3. Structure de la chaîne alpha des collagènes fibrillaires (A), des FACITS (B) (reproduite de la Ref.
avec permission) et d’autres collagènes comprenant
les collagènes des membranes, les multiplexines et le collagène XXVI (C).
major triple-helical domain are generally disease-causing.
However, such interruptions can also generate flexible regions
along the triple helix that favor molecular and cellular inter-
actions and protease clivage. Human collagen XXVII also
contains unexpected residues such as tryptophan and cys-
teine within the triple helix domain. Their significance is
unclear because they are not all conserved among species and
in the closely related collagen XXIV. Collagen XXIV expres-
sion is restricted to bone and cornea both containing hetero-
typic fibrils formed with collagens I and V, and it is not found
in collagen III-containing tissues such as skin, tendon and
vessels. Collagen XXVII is strongly expressed in cartilage
that contain the heterotypic II/XI fibrils. An interesting issue
with respect to the reduced length and the interruptions in
their triple helix domain is the question of their capacity to
form fibrils or to be incorporated into heterotypic fibrils with
classical fibrillar collagens.
and on the lately identified members of this subclass of col-
lagens.
Based on conserved structural features that characterize
the members of this family, five collagens from the last dozen
identified were recognized as members of the FACIT family:
(1) the presence of two highly conserved cysteine residues
separated by four amino acids at the NC1-COL1 junctions
and (2) the existence of two G–X–Y triplet imperfections
within the COL2 domain (3) a succession of triple-helical
domains connected by short non collagneous domains (4) the
presence of a large N-terminal domain that always exhibits a
TSPN subdomain next to the collagenous domain
).
Besides from these common criteria, the FACITs display
remarkable divergence in the size and composition of their
N-terminal domains and in the number of their collagenous
domains (from two for collagens type XII, XIV, XX, XXI to
10 for collagen XVI) (
B). Whereas TSPN represents
the sole module constituting the N-terminal NC domain of
collagens IX, XVI and XIX, in collagens XXI and XXII a
unique von Willlebrand factor A-like domains (vWA) placed
next to the TSPN domain. In collagens XII, XIV and XX,
vWA domains alternate with fibronectin type III repeats (FN3)
(
). In addition, in some species, alternative splicing of
collagen XIV, XII and XIX mRNAs occurring in the
N-terminal domains makes the structure and the function of
2.2. The fibrillar associated collagens with interrupted
triple helix (FACITs)
The structural aspects of the FACITs have been recently
reviewed
and more particularly, the cartilaginous col-
lagen IX, which was the first identified member of this sub-
class
This review will therefore focus on recent advances
[ Pobierz całość w formacie PDF ]