what-when-how
In Depth Tutorials and Information
single 3-hydroxyproline at residue 986 (numbered from
the start of the triple helix), towards the C-terminal end
of the triple helix, in the sequence Gly-3Hyp-4Hyp. The
role of 3-hydroxyproline in collagen type I remained a
mystery for more than 40 years until the enzyme pro-
lyl 3-hydroxylase 1 (P3H1) was purified and charac-
terized.
24,25
During the 1970s reports began to emerge
describing increased hydroxylysine, glucose and galac-
tose in collagen from OI patients.
26
These changes were
detected by amino acid and chemical analyses follow-
ing slow and labor-intensive purification of collagen
from tissues; however, the reason for these increased
post-translational modifications was still not known.
The advent of routine
s
odium
d
odecyl
s
ulfate
p
oly-
acrylamide
g
el
e
lectrophoresis (SDS-PAGE) combined
with biosynthetic labeling of collagen produced by
cultured fibroblasts significantly enhanced progress
towards understanding the molecular basis of OI.
As in many other disorders, the early mutations
characterized were major structural changes to the pro-
tein. The first, in a patient with perinatally lethal OI,
was a large deletion in the α1(I) triple helix, detected in
biosynthetically labeled procollagen, and localized by
digesting SDS-PAGE separated samples with cyanogen
bromide (CNBr) then separating the peptides on a sec-
ond SDS-PAGE gel.
27-30
Another early mutation, discov-
ered following the observation that patient fibroblasts
secreted only α1(I) homotrimers, was a homozygous
frameshift mutation that altered the sequence of the
C-terminal end of the pro-α2(I) chain and prevented it
assembling with pro-α1(I).
31,32
Identifying these muta-
tions established the central role of primary collagen
type I abnormalities in OI; however, researchers were
very aware that these were unusual mutations and
most patients had more subtle changes.
disrupted splicing leading to exon/multi-exon skip-
ping would be found to be common mechanisms result-
ing in presentations of OI, particularly those resulting
in progressively deforming OI type III and perinatally
lethal OI type II. It is of some interest that multi-exon
deletions in COL1A2 have been reported that result
in patients with classic OI with distinctly blue sclerae
phenotype with opalescent dentine with a low fracture
frequency.
37
Collagen gene transcription follows the
general rule that the spliceosome mechanism recog-
nizes the splice donor sequence…AG gt(a/g)agt and
splice acceptor sequence (t/c)n(t/c)agG (where the
capital letters represent the exonic sequence and lower
case the intronic sequences). When these sequences are
disrupted by mutations such as single nucleotide sub-
stitutions, deletions or insertions, there is potential to
generate exon skipping or activation of a cryptic splice
site and an abnormally spliced sequence which disrupts
the regular (Gly-X-Y) structure.
A series of detailed biochemical and gel electropho-
resis studies on the collagen produced by OI patient
fibroblasts appeared in the mid-1980s.
38-41
The key find-
ings of these experiments were that OI cells secreted
less collagen type I, there was increased intracellular
degradation, and the chains had increased post-trans-
lational lysine hydroxylation and glycosylation that
caused slow migration on SDS-PAGE. Two-dimensional
CNBr mapping, separating the peptides by charge in
the first dimension and molecular weight in the sec-
ond, showed that in some patients the increased lysine
modifications were found along the entire chains, while
in others the more C-terminal peptides were normal
and excess modifications were restricted to the more
N-terminal peptides.
38,39,41
The simplest explanation
to account for these observations was that underlying
mutations disrupted the Gly-X-Y sequence, slowing
triple helix folding and exposing the chains to post-
translational lysine modification enzymes for longer.
Peptides N-terminal to the mutation would thus con-
tain excess lysine modifications. Proline hydroxylation
was not increased because almost all available prolines
are already hydroxylated in normal collagen type I.
In addition, two-dimensional CNBr maps identified
peptides with charge changes in some patients
38,39
allowing the search for mutations to be narrowed, a
huge advantage in the pre-PCR and direct nucleotide
sequencing era.
The first glycine mutation identified42
42
was in a
patient with disulfide bonded α1(I) chains indicat-
ing the presence of a cysteine in the triple helix. Two-
dimensional SDS-PAGE CNBr maps localized the
introduced cysteine to the CB6 peptide.
43
To find the
mutation, Dan Cohn and colleagues cloned genomic
DNA into a λ-phage library, identified positive clones
with a
COL1A1
genomic probe, and subcloned and
MOLECULAR BIOLOGY OF COLLAGEN
GENES AND ITS RELEVANCE TO OI
RESEARCH
Successive discoveries in molecular biological meth-
odologies resulted in the cloning first of the gene for
the human procollagen α2 chain of type I collagen
(COLIA2)
33
and then the COLIA1 gene. It immediately
became apparent from studies of cDNA-gDNA hybrids
that the molecular organization of the fibrillar collagen
genes was particularly representative of the interrupted
nature of coding sequences (exons) with intervening
sequences (introns).
34
At the time this was only known
for a few mammalian genes.
35
What was even more
intriguing was that the modal exon size was a multiple
of 54bp.
36
The pro-α1 chain of type I collagen was found
to contain 52 exons and the pro-α2 chain 54 exons. Such
a structure might have predicted that mutations which
