what-when-how
In Depth Tutorials and Information
aggregation. In addition, very high triple-helix stabil-
ity and propensity within the N-anchor region
103
might
enhance triple helix re-nucleation on the N-terminal side
of closely located mutations. Three pronounced peaks
in the severity of α1(I) Gly substitutions (observed both
for all Gly substitutions pooled together and just for
Gly-to-Ser) might be explained by the lack of triple-helix
re-nucleation sites and a reduced number of HSP47 bind-
ing sites on the N-terminal side of these peaks.
might play such a role, but it is not clear yet whether
MIA3 is essential for trafficking of all collagens or pri-
marily type VII collagen.
111-113
In addition to distinct adaptor(s), trafficking of 300 nm-
long procollagen molecules requires larger than usual
(60-80 nm) vesicles. Mono-ubiquitylation of COPII pro-
tein SEC31 by CUL3-KLHL12 ubiquitin ligase complex
has been recently reported to be essential for formation of
such vesicles.
114
Suppression of this ubiquitylation impairs
export of type I procollagen from the ER, although this
process might be a general rather than a collagen-specific
regulator of the COPII-covered vesicle size.
Another interesting aspect of procollagen traffick-
ing is that N- and C-propeptide cleavage might begin
already in the secretory pathway inside the cell.
115
The
N-propeptide might be cleaved already in
cis
-Golgi
while C-propeptide cleavage appears to begin only in
post-Golgi compartments.
116
Nevertheless, a significant
fraction of procollagen molecules is secreted with intact
N- and C-propeptides, at least in cell culture.
34,115-117
These advances clarify some of the unique features
of procollagen trafficking, but they do not explain how
cells distinguish procollagen molecules with Gly sub-
stitutions and selectively retain them in the ER for sub-
sequent degradation. Such retention, which has been
reported in a number of studies,
89,118,119
might be a cru-
cial factor in excessive ER stress and osteoblast mal-
function in OI.
80
We hope that future studies of mutant
procollagen trafficking will address this problem.
PR
OCOLLAGEN TRAFFICKIN
G
Export of procollagen from the ER and its trafficking
through the cell also appear to involve distinct mecha-
nisms. In particular, secretion of incompletely folded
procollagen by cells that produce no or dysfunctional
HSP47 could indicate HSP47 involvement in control-
ling the quality of procollagen folding in the ER.
10,104
However, HSP47 cannot prevent the export of incom-
pletely folded chains from the ER by preferential bind-
ing to and tagging them like general ER chaperones
bind to and tag incompletely folded chains of other pro-
teins. Instead, HSP47 preferentially binds to and is co-
transported with natively folded procollagen from the
ER to Golgi.
81,82,105
Thereby, it is more likely to promote
the transport of native procollagen rather than inhibit
the transport of misfolded chains. Secretion of incom-
pletely folded procollagen by HSP47 knockout cells
might result simply from inability of these cells to fold
any procollagen molecules, some of which make their
way out of the cells. Furthermore, unfolded pro-α2(I)
chains seem to travel from the ER to Golgi in cells that
have no chaperone deficiencies. In particular, knockout
of pro-α1(I) chains prevents the association and fold-
ing of pro-α2(I) chains, resulting in translocation of
unfolded pro-α2(I) chains to Golgi where they are tar-
geted for lysosomal degradation via an unknown path-
way.
106
Transfection of cells with GFP-tagged pro-α2(I)
chains (most of which are not expected to incorporate
into folded procollagen trimers
107
) results in accumula-
tion of such chains in Golgi cisternae.
108
Other chaperones involved in the quality control
of procollagen chain folding in the ER include BiP and
PDI. BiP binding likely prevents the export of chains
with misfolded C-propeptides.
86-88
PDI binding seems
to prevent the export of molecules with large structural
defects, such as truncated chains.
109
Yet, the overall pic-
ture of procollagen folding quality control at export
from the ER is still largely unclear.
96
Like other proteins, procollagen is transported from
the ER to Golgi in COPII-covered vesicles,
110
but procol-
lagen might require distinct adaptor protein(s) for load-
ing into these vesicles. Studies of MIA3 knockout cells
and animals suggest that this ER membrane protein
ASSEMBLY AND FUNCTION OF
COLLAGEN FIBERS
Correctly folded collagen molecules are capable
of unassisted self-assembly into fibers once their
N- and C-propeptides are cleaved.
120
Yet, fiber assem-
bly (fibrillogenesis) is tightly regulated
in vivo
by a vari-
ety of molecules, e.g., it occurs on the outer cell surface
although N- and C-propeptide cleavage begins inside
the cell.
116,121
In bone, collagen fibers provide a template
for deposition of hydroxyapatite mineral, glue min-
eral crystals together, mechanically reinforce the tissue
(akin to steel bars in concrete), and bind other molecules
essential for normal bone matrix homeostasis and min-
eralization. Dysregulation of assembly, deposition, and
function of collagen fibers in bone might be caused by
(1) insufficient collagen synthesis; (2) abnormal osteo-
blast function; (3) abnormal collagen-collagen interac-
tions; and (4) abnormal collagen interactions with other
molecules or minerals. However, relative contributions
of these abnormalities are often difficult to distinguish.
For instance, insufficient collagen synthesis is consid-
ered to be responsible for the mild (type I) OI in patients
with null-allele
COL1A1
mutations (which prevent
