Conformational Effects of GlyXGly Interruptions in

the Collagen Triple Helix

Jordi Bella
1
, Jingsong Liu
1
, Rachel Kramer
1
, Barbara Brodsky
2
and Helen M. Berman
1

1
Department of Chemistry and
Chemical Biology, Rutgers,
The State University of New
Jersey, 610 Taylor Road,
Piscataway, NJ 08854-8087,
USA
2
Department of Biochemistry,
University of Medicine and
Dentistry of New Jersey,
Robert Wood Johnson Medical
School, Piscataway,
NJ 08854, USA
The collagen model peptide with sequence (ProHypGly)
4
ProGly(Pro
HypGly)
5
contains a central GlyProGly interruption in the consensus
collagen sequence. Its high-resolution crystal structure defines the
molecular consequences of such an interruption for the collagen triple-
helical conformation, and provides insight into possible structural and
biological roles of similar interruptions in the GlyXY repeating pattern
found in non-fibrillar collagens. The peptide (denoted as the Hyp minus
peptide or Hyp

) forms a rod-like triple helix structure without any bend or

kink, and crystallizes in a quasi-hexagonal lattice. The two ProHypGly
zones adopt the typical triple-helical collagen conformation with standard
Rich and Crick II hydrogen bonding topology. Notably, the central zone
containing the GlyProGly interruption deviates from the standard
structure in terms of hydrogen bonding topology, torsion angles, helical,
and superhelical parameters. These deviations are highly localized, such
that the standard features are regained within one to two residues on either
side. Conformational variations and high temperature factors seen for the
six chains of the asymmetric unit in the zone around the interruption point
to the presence of a local region of considerable plasticity and flexibility
embedded within two highly rigid and ordered standard triple-helical
segments. The structure suggests a role for GlyXGly interruptions as
defining regions of flexibility and molecular recognition in the otherwise
relatively uniform repeating collagen conformation.
2006 Elsevier Ltd. All rights reserved.
*Corresponding author
Keywords: collagen; peptide; crystal structure; collagen interruptions;
extracellular matrix
Introduction
Fifty years ago, three different research groups
proposed the triple helix conformation for collagen
on the basis of fiber diffraction analysis and model
building.
13
The original proposal interpreted the
fibre diffraction data with a left-handed triple helix
of ten residues in three turns (10
7
or left-handed
10/3 symmetry). Some twenty years later, the first
crystallographic study on a model collagen peptide
showed a left-handed triple helix of seven residues
in two turns,
4,5
which led to the proposal of a new
structural model for collagen.
6,7
More recent high
resolution crystallographic studies of synthetic
peptides with collagen-like sequences have pro-
vided a wealth of new information on collagen
structure, hydration, and helical symmetry that
could not be deduced from fiber diffraction analysis
of heterogeneous collagen samples
822
(see Brodsky
& Persikov,
23
for a review).
The primary structure of the collagen triple helix is
a conspicuous repetitive sequence GlyXY,
where the X and Y positions are often occupied by
proline (Pro, P) and its post-translational modifica-
tion 4-hydroxyproline (Hyp, O) respectively. Extra-
cellular matrix structural proteins containing one or
more collagen triple-helical motifs are generically
classified as collagens, and nearly 30 genetically
different types have been described so far in humans
J.B. and J.L. contributed equally to this work.
Present addresses: J. Bella, Wellcome Trust Centre for
Cell-Matrix Research, Faculty of Life Sciences, University
of Manchester, Manchester M13 9PT, United Kingdom;
J. Liu, Guangzhou Institute of Biomedicine and Health
(GIBH), Chinese Academy of Sciences, Guangzhou
(Canton) 510663, China.
E-mail address of the corresponding author:
berman@rcsb.rutgers.edu
doi:10.1016/j.jmb.2006.07.014 J. Mol. Biol. (2006) 362, 298311
0022-2836/$ - see front matter 2006 Elsevier Ltd. All rights reserved.
alone.
24
These collagens differ in the form of their
higher-order macromolecular assembly and in the
number and position of their triple-helical regions.
Notably they also show distinct degrees of tolerance
to breaks in the GlyXY repeating sequence. The
abundant fibrillar collagen types I, II, III, V and XI
have the strictest requirement for an uninterrupted
pattern: each contains a contiguous stretch of 1000
residues with a GlyXY repeating sequence.
Such regularity may be critical for the formation of
quarter-staggered fibrils with a 67 nm axial repeat
(D-period) that are the characteristic assemblies of
these collagens. Missense mutations in fibrillar
collagens where one Gly is changed to another
larger residue have been shown to lead to hereditary
connective tissue disorders.
25,26
In contrast, non-
fibrillar collagens typically contain one or more
breaks in the GlyXY repeating pattern under
normal, non-pathological conditions. These inter-
ruptions appear to be compatible with a variety of
supramolecular arrangements shown by these col-
lagens. For example, type IV collagen contains more
than 20 interruptions in the GlyXY pattern of
its 1350-residue triple-helical domain,
27,28
and forms
sheet-like networks that become the structural scaf-
fold of basement membranes.
Replacement of a GlyXY tripeptide by a GlyX
dipeptide in a collagen sequence creates a GlyX
Gly interruption site. Conceptually, this interruption
is equivalent to the deletion of a single amino acid
from the X or Y position in one triplet of the
repetitive GlyXY sequence. Such interruptions
are found in several collagens and may be associated
with specific structural functions. The 454-residue
triple helix of type VIII collagen includes eight of
such sites
29
and forms highly extended hexagonal
arrangements. The 59-residue triple helix of the
host-defense mannose binding lectin (MBL) contains
one GlyGlnGly site,
30
which is considered to
provide the unique kink in the collagen triple helix
necessary for the bouquet of flowers oligomeric
assembly of this protein.
31,32
Current understanding of the effect of breaks in
the GlyXY repeating sequence on collagen
conformation and their relation to supramolecular
structure and function is very limited. It appears
that tolerance of collagens to interruptions in their
GlyXY pattern is site specific, suggesting a
relationship to their higher-order structure or func-
tion. Additional Gly missense interruptions in
collagens that already tolerate several breaks in
their sequences result in pathological states.
33
Model collagen peptides offer a useful approach
for studying the effects of interruptions in the Gly
XY repeating sequence. Starting with the very
stable triple-helical peptide (ProHypGly)
10
, a
discontinuity in the form of a GlyAla substitution
was shown to decrease thermal stability.
34
The high-
resolution crystal structure of this peptide shows a
localized arrest in the helical twist rate and replace-
ment of direct hydrogen bonds by water-mediated
ones at the interruption site.
8
We have extended
these studies to the GlyXGly type interruption.
We designed a peptide, which we will refer to as
Hyp

, in which one central Hyp residue from a

stable (ProHypGly)
10
triple-helical peptide is
missing, resulting in the sequence (ProHyp
Gly)
4
ProGly(ProHypGly)
5
with a GlyPro
Gly interruption site. The absence of the Hyp
residue dramatically decreased the thermal stability
of the triple helix, from 60 C to 17 C in dilute acetic
acid,
34
showing a more severe destabilizing effect
than the GlyAla substitution. Here, we present the
2.0 X-ray crystal structure of the Hyp

peptide,
which provides for the first time a high-resolution
picture of the changes of local conformation of the
collagen triple helix in the vicinity of a GlyXGly
interruption.
Figure 1. Asymmetric unit of the Hyp

crystal struc-
ture. (a) Ribbon diagram of the two triple helices in the
asymmetric unit (two perpendicular views), shown in red
and blue respectively. The region coloured in yellow
corresponds to the central GlyProGlyPro sequence for
each peptide (central zone). Chains are labelled ABCfor the
first triple helix (red), and DEF for the second (blue). Labels
are placed at the N-terminal end of each chain. (b) Stick
representation of the two molecules (two perpendicular
views). Pro residues are coloured in gray, Hyp in blue and
Gly in red, except those from the central zone, which are
shown in yellowfor all three chains. The ABCtriple helix is
shown on the left and the DEF triple helix is shown on the
right. Residue numbering is as follows: chain A129, chain
B 3159, chain C 6186, chain D 91119, chain E 121149,
and chain F 151189 (labels for N-terminal proline residues
are shown). The central zones (in yellow) correspond to
residues 1215 (A), 4245 (B), 7275 (C), 102105 (D), 132
135 (E), and 162165 (F). The C-terminal triplet for chain C
(residues 8789) is not shown due to conformational
disorder. Figures were generated with SETOR.
52
299 Crystal Structure of an Interrupted Collagen
Results
Overall structure and main chain conformation
The Hyp

peptide crystallizes with the triple-

helical structure characteristic of collagen molecules.
Each asymmetric unit contains two antiparallel
triple helices related by a non-crystallographic
rotation of 178. The two triple helices are similar
in conformation but not identical. Therefore, there
are six non-equivalent chains in this structure,
designated as ABC in the first triple helix, and
DEF in the second (Figure 1). Both triple helices are
essentially straight, without any appreciable bend or
kink. Their length is about 3 shorter than expected
for a regular (ProHypGly)
10
triple helix, as a result
of the absence of a single Hyp residue. For the
purpose of conformational analysis it is convenient
to define a central zone which comprises the four
GlyProGlyPro residues in each chain that flank
the missing Hyp residues. This zone, shown in
yellow in Figure 1, will be considered separately
from the zones at either side in which several
consecutive ProHypGly triplets result in a regular
triple-helical conformation and standard hydrogen
bonding topology (see below). This regular con-
formation only unravels slightly at the terminal
ends, which show appreciable conformational dis-
order. The central zone shows well-defined electron
density (Figure 2) but exhibits a significant level of
flexibility and conformational variability between
the six individual chains (Figure 3).
Main-chain conformation and average tor-
sion angles of the ProHypGly zones (Table 1) are
very similar to those observed in the imino acid-rich
zones of previously determined collagen-like peptide
structures.
5,8,10,12,15,1719
Their helical symmetry
is close to 7
5
or left-handed 7/2 triple helix (Table
1), and a 7-fold symmetry can be easily noticed
when looking at the molecules along the helical axis
(Figure 1). The root-mean-square deviation between
the C

atoms of the ProHypGly zones and those

from an ideal 7
5
triple helix is 0.41 . The
Ramachandran diagram for the Hyp

peptide
shows most of the torsion angles clustered in the
region corresponding to the polyproline II helix.
However, six residues in the central zone and one
adjacent to it are outside this region (Figure 4). Two
Hyp residues with torsion angles in the -
helical region (Hyp58 and Hyp148; Figure 4)
illustrate the partial unraveling of the triple helix
at the C-terminal ends.
Side-chain conformation: imino acid ring
puckering
The ring puckering of Pro residues in the Hyp

structure is almost equally distributed between the

up (
1
<0) and down (
1
>0) conformations, with
24 conformers up and 35 conformers down (there is
no good density for residue Pro87). Pro residues in
the central zone are exactly split, six up and six
down, and follow a curious pattern: down-up (13
15), down-down (4345) and up-up (7375) in the
ABC triple helix, and down-down (103105),
down-up (133135) and up-up (163165) in the
DEF triple helix. Conversely, Hyp residues are
primarily found in the up conformation. Only four
out of the 53 Hyp residues with defined electron
density adopt the down conformation. Two of them
are at the terminal zones (Hyp92 and Hyp178), and
the other two are involved in main-chain to side-
chain intermolecular hydrogen bonding interac-
tions (see below).
Effects of the GlyProGly interruption in helix
supercoiling
The deletion of a single imino acid from the (Pro
HypGly)
10
sequence has important implications for
the overall conformation and supercoiling of the
collagen triple helix. This point will be illustrated
with helical nets, plots of cylindrical coordinates
versus Z, which are very useful for visualizing local
variations of parameters such as twist or height in a
helical molecule. Figure 5(a) shows the helical net
from an idealized +(ProHypGly)
10
triple helix,
Figure 2. Difference omit map
of the four residues from the central
zone in chain C (Gly72 to Pro75).
The map has been calculated with
CNS
50
using sigma-weighted F
obs

F
calc
coefficients, and has been con-
toured at 2. The Figure was gener-
ated with SETOR.
52
300 Crystal Structure of an Interrupted Collagen
with 7
5
symmetry and rise per residue of 2.87 .
Each individual chain adopts a right-handed 7
1
helix
with constant twist angle 51.4 (360/7) and height
8.61 (32.87 ). The =51.4 rotation and
Z =8.61 translation thus define a single helical
step. The path of the left-handed superhelix (blue
broken line in Figure 5(a)) shows constant super-
helical twist =102.8 (exactly 2, but left-
handed), and superhelical height Z=2.87 (Z/
3). These values define a single superhelical step in
a 7
5
superhelix. The vertical broken lines in light
blue in Figure 5(a) indicate that there is no helical
drift: any given residue in the helix has the same
cylindrical coordinate as the seventh residue ahead
or behind, both on the superhelix and on its own
helix. Thus, an ideal superhelix twists monotoni-
cally and all helical and superhelical steps are the
same.
A similar Z helical net is shown in Figure 5(b)
for the Pro residues of the ABC triple helix (Pro1
Pro86) in the structure of the Hyp

peptide. There are

obvious departures from helical ideality as a
consequence of the deletion of a single imino acid
in the (ProHypGly)
10
collagen sequence. The rate
of twist changes dramatically for each chain in
the central zone, the region around the Hyp deletion.
For example, the helical step between residues
Pro13 and Pro15 in chain A turns by =26 and
raises by Z=6.4 (as a result of the deletion of
one imino acid). For the B and C chains, departures
from a 7
1
helix are seen both at the interrupted
triplet and at the previous one: between Pro40
and Pro43 shortens to 19, whereas it gets
very long at the next triplet, (Pro43Pro45) =
78. Similarly, (Pro70Pro73) is very short (2)
while (Pro73Pro75) is again very long (87).
Figure 3. Conformational effects of the GlyProGly interruption. (a) Superposition of the C

traces from the six

chains of Hyp

. The main divergence occurs at the GlyProGlyPro central zone. (b) Variation with sequence of C

distances between the six chains after structural superposition, using chain B as reference. The GlyProGlyPro zone is
highlighted in yellow. (c) Average temperature factor plotted against sequence. The three sequences are shown staggered
to illustrate the correct alignment after the normal one-residue staggering in the collagen triple helix. The GlyProGly
Pro zone is highlighted in yellow. Amino acid one-letter codes used in this Figure: Pro, P; Hyp, O; Gly, G.
301 Crystal Structure of an Interrupted Collagen
All these Pro residues correspond to the central Gly
ProGlyPro region of the peptide, containing the
interruption. Values for between Pro residues in
the non-interrupted zones average 58(9) at the N-
terminal side and 54 (8) at the C-terminal side.
These values are consistent with an overall 7-fold
symmetry for the collagen triple helix (Table 1).
The effects on the superhelix (blue broken line)
appear to be more subtle. Only the superhelical step
between Pro13 and Pro43 results in a large rotation
angle =150. Short values occur at the
superhelical steps Pro40Pro70 (85) and Pro73
Pro15 (82). The rest of the superhelical steps
average =102, very close to the value for an
ideal 7
5
superhelix.
The Z plot allows visualization of the angular
relation between the two ProHypGly regions
flanking the central zones. Light blue lines in Figure
5 monitor the amount of helical drift. Such lines are
completely vertical for an ideal (ProHypGly)
10
triple helix (Figure 5(a)). The N-terminal side of the
Hyp

region has been fitted to corresponding

residues in a standard 7
5
helix. Therefore, a drift
towards the right indicates that the top part of the
helix (C-terminal side) has twisted too fast
compared to the standard 7
5
helix, and conversely
a drift to the left indicates that the top part has
twisted too slow. Drift lines in the Hyp

diagram
show some zigzag, which is more exaggerated
around the GlyProGlyPro zone. Remarkably,
however, the overall orientation of those lines is
mostly straight up, which means that the top part of
the helix has mostly regained its angular coherence
with a standard 7
5
helix. This is true even after the
conformational disruptions seen in the central,
interrupted region, and is mostly achieved through
compensating variations. For example in chain B the
long helical step Pro43Pro45 ( =78) follows
the short one Pro40Pro43 ( =19). The two
steps together add to 97, close to the 102 helical
rotation angle for a two-residue step in a standard 7
1
individual helix. As a result, Pro40 and Pro45 end up
pretty much aligned in the helical net with residues
Pro4 and Pro10 below (chain A), and Pro75 and
Pro81 above (chain C). Similar local variations
throughout the whole helix mostly compensate for
the local angular effect of the individual disruptions
caused by the deletion of a Hyp residue in each
chain. Analysis of the Z helical net for the DEF
triple helix (not shown) yields the same conclusions.
Hydrogen bonding
Hydrogen bonding topology and metrics in the
ProHypGly zones are consistent with the stan-
dard Rich and Crick II (RCII) and C

HO=C
hydrogen bonding patterns seen in other colla-
gen peptides (Table 2).
35
The standard hydrogen
bonding topology is interrupted in the central zone
(Figure 6). Substitution of a ProHypGly tripeptide
by a ProGly dipeptide affects the hydrogen bonding
environment of the Gly residue, but it does so in a
different way for each of the three chains. In the
leading chains, A and D, the NH amide groups
Table 1. Average conformational angles and helical
parameters for the Hyp

peptide, compared with those

from other triple-helical structures
Hyp
a
(POG)
10
b
EKG
c
GlyAla
d
Resolution () 2.0 1.26 1.75 1.9
A. Torsion angles ()
Pro 71.7 (5.2) 71.3 (1.4) 73.7 (3.0) 72.6 (7.6)
Pro 161.3 (8.0) 161.5 (1.1) 160.5 (5.9) 163.8 (8.8)
Hyp 60.8 (8.6) 56.9 (1.3) 59.7 (2.4) 59.6 (7.3)
Hyp 150.1 (13.2) 150.0 (1.1) 151.3 (4.0) 149.8 (8.8)
Gly 73.8 (8.1) 71.3 (1.6) 72.4 (3.7) 71.9 (9.6)
Gly 176.9 (6.3) 174.2 (1.2) 175.1 (4.2) 174.1 (11.9)
B. Helical parameters
e
Height () 8.5 (0.2) 8.5 8.5 (0.1) 8.4 (0.2)
Twist () 54 (10) 52 52 (7) 57 (8)
C. Superhelical parameters
e
Height () 2.8 (0.2) 2.8 2.8 (0.1) 2.8 (0.2)
Twist () 102 (11) 104 103 (8) 101 (11)
Standard deviations are given in parentheses.
a
Residues from the GlyProGlyPro zone not included in the
average calculations.
b
Kramer et al.
15
: standard deviations for helical parameters not
given.
c
Okuyama et al.
18
: peptide sequence (POG)
4
EKG(POG)
5
;
average conformational parameters include torsion angles
from the EKG triplets.
d
Bella et al.
8
: peptide sequence (POG)
4
POA(POG)
5
; average
conformational parameters include residues around the GlyAla
substitutions.
e
Average helical and superhelical parameters for the regular
zones of the different structures (interruptions in collagen
sequence and disordered terminal zones not included).
Figure 4. Ramachandran map of the Hyp

structure.
Pro residues are represented by black triangles, Hyp
residues by blue triangles, and Gly residues by red
triangles. Typical secondary structures are indicated
(, -helix; , -sheet; ppII, polyproline II and collagen).
Residues highlighted in yelloware fromthe GlyProGly
Pro zone and depart from the typical conformational
angles of the collagen triple helix. Two Hyp residues near
the C-terminal ends have helical conformational angles
(see the text).
302 Crystal Structure of an Interrupted Collagen
from Gly14 and Gly104 point towards the solvent
and thus lose the capability of inter-chain hydrogen
bonding. In the middle chains B and E, the NH
amide groups from Gly44 and Gly134 remain
pointing inwards, where they form non-standard
hydrogen bonds with carbonyl groups fromtrailing-
chain residues Gly72 and Gly162 respectively (Table
2). This new hydrogen bonding topology results
from the transition between the two sets of RCII
hydrogen bonds at either side of the interruption
(Figure 6). Finally, in the trailing chains C and F, the
NH amide groups from Gly74 and Gly164 revert to
Figure 5. Helical net plots for collagen triple helical molecules with their axes aligned along the Z polar coordinate
axis. (a) An ideal 7
5
triple helix. (b) The ABC triple helix in the Hyp

crystal structure. Squares represent the (, Z) polar

coordinates of the C

atoms from Pro residues. The origin of the angular coordinate is arbitrary. Red, yellow and green
lines represent the leading (A), intermediate (B) and trailing (C) chains, respectively. Broken blue lines show the path of
the superhelix, and cyan lines join residues that would be equivalent in an ideal 7
5
triple helix. Pro residues with
significant disruption in their helical twist are indicated (see the text).
303 Crystal Structure of an Interrupted Collagen
standard RCII hydrogen bonding topology. The
amide group of Gly72 points inside the triple helix
but does not have a suitable hydrogen-bonding
partner. Normal C

HO=C hydrogen bonding is

also disrupted by the interruption, with non-
standard C

HO=C hydrogen bonds appearing

in the central zone (Table 2). Other C=O groups in
the GlyProGlyPro zone form hydrogen bonds to
water molecules (not shown). These water mole-
cules, however, do not directly connect neighbour-
ing chains as do the interstitial water molecules
found in the substitution zone of the crystal
structure of the GlyAla peptide.
8,9
Intermolecular interactions
The Hyp

peptide molecules crystallize in antipar-

allel quasi-hexagonal packing. Triple helices pack
laterally in well-defined layers, and accumulation of
layers along the crystallographic c axis generates the
entire crystal (Figure 7). Within each layer, triple
helices align themselves such that the GlyProGly
Pro zones are at about the same vertical level. This
results inan arrangement without vertical staggering.
The axis-to-axis distances between neighboring
triple helices vary along the lattice direction,
between 12.2 and 14.5 (Figure 8(a)). The two
molecules at 12.2 are the closest so far in any
crystal structure of a collagen peptide, and consti-
tute the asymmetric unit shown in Figure 1. Four
direct hydrogen bonds occur between hydroxyl
groups of Hyp residues in one molecule and main-
chain carbonyl groups in the adjacent molecule (not
shown). This previously unobserved feature is
consistent with the close proximity between the
two molecules. One single case of HypOO
Hyp hydrogen bonding interaction is also seen
between triple helices separated by 13.7 , which is
consistent with interaxial distances in previously
reported HypHyp direct hydrogen bonds in the
structures of the EKG and (ProHypGly)
10
peptides.
13,14,18
One salt-bridge is observed across
layers, between the N-terminal head of one peptide
and the C-terminal tail of the other (not shown).
Hydration analysis
A significant number of ordered water molecules
are observed in the Hyp

crystal structure. Their

positions and hydrogen bonding patterns resemble
those previously described in detail in the struc-
ture of the GlyAla peptide,
9
and also seen in
other crystal structures of collagen peptides.
1018
There are fewer ordered water molecules in the
Hyp

structure than seen in other collagen peptide

structures rich in ProHypGly triplets. Thus,
there are only 2.3 water molecules per tripeptide
in the first hydration shell of the Hyp

structure,
compared with 4.1 per tripeptide in the (ProHyp
Gly)
10
and (ProHypGly)
11
structures,
18
or 3.7 per
tripeptide in the GlyAla structure.
9
The reduc-
tion in the number of ordered water molecules
may be due to several reasons, including the
disruption of the water network around the Gly
ProGlyPro central zone and the somewhat lower
resolution of this particular crystal structure
determination. Furthermore, the angular distribu-
tion of water molecules surrounding any given
triple helix in the crystal is unequal, closely
matching the variation of interhelical distance to
its closest neighbours (Figure 8). In particular, the
number of water molecules in the direction of the
shortest separation between triple helices (12.2 ,
angle =0 in Figure 8), is significantly lower than
the number of water molecules in other directions,
which suggests that close proximity between
collagen triple helices results in a smaller number
of intervening ordered water molecules.
Despite this decrease in the observed number of
first shell water molecules, several water bridges
incorporating second and third-shell water mole-
cules are still observed, with very similar geometries
to those reported.
9
Notably, the same intermolecular
water bridges are repeated at either side of the Gly
ProGlyPro central zones, as illustrated by two
examples in Figure 9. The two water motifs shown
connect the two antiparallel triple helices in the
asymmetric unit (those separated by 12.2 ). One
motif is the fusion of two 2 bridges (see Bella et al.
9
Table 2. Standard and unusual hydrogen bonding parameters in the Hyp

crystal structure (standard deviations in

parentheses)
Distances () Angles ()
NHO=C HO NO NHO HO=C NO=C
GlyPro standard 2.09 (0.17) 2.95 (0.15) 147(10) 155 (8) 165 (6)
Gly44Gly72 3.13 2.25 150 157 166
Gly134Gly162 2.76 2.02 132 152 160
C

HO=C H

O C

O C

O H

O=C C

O=C
HypPro standard 2.54 (0.19) 3.43 (0.20) 139 (5) 127 (6) 137 (6)
GlyPro standard 2.41 (0.21) 3.45 (0.19) 164 (7) 114 (8) 117 (7)
GlyGly standard 2.62 (0.21) 3.15 (0.16) 110 (9) 92 (8) 100 (7)
2.80 (0.20) 99 (8) 112 (7)
Gly42Gly12 2.07 3.15 176 114 114
Gly14Gly42 2.05 3.05 153 172 165
Gly134Gly162 2.43 3.08 118 143 156
Gly162Gly132 1.98 3.02 160 114 113
304 Crystal Structure of an Interrupted Collagen
for nomenclature), one from each triple helix (Figure
9(a) and (b)); the other is a cluster of two 2 bridges,
one from each triple helix, connected by several
additional waters (Figure 9(c) and (d)). Each of these
motifs has local 2-fold symmetry, which is perfectly
consistent with the antiparallel arrangement of the
triple helices that they connect. Each motif is
repeated twice, one at each side of the GlyPro
GlyPro central zone, and in almost exactly the same
angular orientation with respect to the respective
triple-helical axes. The identical positioning of these
water bridges is consistent with the observation
discussed above that the helical segments at either
side of the central zone are in almost perfect angular
register, but with a 2.8 shorter separation in the
vertical axis when compared with a triple helix with
ideal 7
5
symmetry.
Discussion
Interruptions of the characteristic GlyXY
repeating collagen sequence can be perfectly toler-
able in some collagens and have severe pathological
effects in others. Their respective structural signa-
tures are not well defined and may depend on the
nature, location and extent of the breaks. Here we
present the crystal structure of the Hyp

peptide at
2.0 resolution. This structure provides for the first
time a molecular picture of a collagen triple helix
accommodating a GlyXGly interruption in the
collagen repeating sequence GlyXY.
Conformation and hydrogen bonding are
perturbed in the GlyXGly interruption site
The central GlyProGlyPro zone shows a very
localized and subtle perturbation of the triple-helical
conformation, which rapidly reverts to normal in the
ProHypGly zones. Several residues in the central
zone adopt conformational angles outside the
polyproline II region (Figure 4), in contrast with
the situation seen in the structure of the GlyAla
peptide, in which conformational angles in the
interruption zone are similar to those of the rest of
the molecule.
8
High temperature factors and sig-
nificant variability between the individual chains
(Figure 3) suggest an increased conformational
flexibility in the central GlyProGlyPro zone.
The typical hydrogen bonding topology of col-
lagen is altered at the interruption site, which could
account in part for the large decrease in thermal
stability of the Hyp

peptide compared to that of

(ProHypGly)
10
. This latter peptide represents the
collagen triple helix with ideal hydrogen bonding,
where every available Gly NH group follows its
turn in forming a hydrogen bond with the corres-
ponding C=O group of the X residue from the next
Figure 6. Schematic represen-
tation of hydrogen bonding topo-
logy in the Hyp

structure around
the central GlyProGlyPro zone
(highlighted in yellow; residue col-
ouring as in the other Figures).
Green lines represent RCII hy-
drogen bonds and the single pur-
ple line represents the unusual
GlyGly hydrogen bonding pair-
ing (see the text). Only the DEF
triple helix is shown. Hydrogen
bonding in the ABC triple helix is
identical except for a missing
hydrogen bond between Gly72
and Pro13 (equivalent to Gly162
and Pro103 in this Figure). Chain D
is repeated at the right side to
provide a clearer description of
hydrogen bonds with chain F.
305 Crystal Structure of an Interrupted Collagen
chain with near-clockwork precision. With the
accompanying set of C

HO=C hydrogen bonds

the collagen triple helix becomes self-contained in
terms of hydrogen bonding and results in a highly
compact structure aroundthe helical axis.
35
Removal
of a single residue in the consensus repetitive
sequence sends this precise hydrogen bonding into
disarray and disrupts the triple-helical conformation
at the point of the interruption. At least one Gly NH
group per triple helix necessarily loses its hydrogen-
bonding partner as a result of the loss of consistency
between the three chains. Such perturbations will
contribute to a decrease in the free energy of
interaction between the three chains and conse-
quently will have a negative impact on the thermal
stability of the triple helix. It is interesting to note that
the T
m
value of the Hyp

peptide is substantially
lower than that of the GlyAla peptide, where a
network of interstitial waters interconnects the
peptide groups of the three chains around the three
GlyAla substitutions.
8,9
It would seem that main-
taining a consistent pattern of inter-chain hydrogen
bonds, even if mediated through water, is less
detrimental for the thermal stability of the collagen
triple helix than locally disorganizing that pattern.
The zones flanking the interruption site are
coaxial and show 7-fold symmetry
Despite the local disruption introduced by the
interruption, supercoiling of the Hyp

molecules
occurs in a way as to maintain the two ProHypGly
zones flanking the central zone in angular register
(Figure 5). Local changes in helical and superhelical
twist in the GlyProGlyPro zone quickly com-
pensate each other. An indication of the precision of
this alignment is the repetition of identical inter-
molecular water bridges at either side of the central
zone, at almost exactly the same angular orientation
(Figure 9). Optimization of lateral interactions as
molecules assemble into the crystal lattice may
favour this intramolecular alignment, compensating
for the local perturbations of helical twist at the
central zone.
Fibre diffraction patterns of collagen in tendons
have been classically interpreted with collagen tri-
ple helices with 10
7
or left-handed 10/3 sym-
metry.
36,37
The Hyp

structure follows the trend

observed in crystal structures of collagen peptides,
which show a clear preference for a more twisted
conformation with 7
5
or left-handed 7/2 symmetry
(Figure 5). This preference and a re-examination of
the collagen diffraction pattern has led Okuyama
and co-workers to revise the average molecular
structure of native collagen in terms of 7
5
sym-
metry.
6,7
A more relaxed conformation closer to the
10
7
symmetry is observed in the imino acid-free
triplets of the peptide with sequence (ProHyp
Gly)
3
IleThrGlyAlaArgGlyLeuAlaGly
(ProHypGly)
4
,
11,15
which suggests that variations
in sequence along the collagen molecules may re-
sult in a non-uniform twist along their length.
23
The inherent plasticity of the triple helix motif
documented here is striking. Collagen triple helices
seem to be able to tolerate significant local changes
in helical twist to respond to sequence variability,
Figure 7. Two lateral views of
the crystal packing in the Hyp

crystal structure: (a) projection

along the crystallographic a axis;
(b) projection along the ab crystal-
lographic direction. Triple-helical
molecules are shown as ribbons:
ABC chains in red, DEF chains in
blue, and central zones in yellow.
Notice the absence of staggering
and the adjacent positioning of the
central zones from neighbouring
molecules in the lattice.
306 Crystal Structure of an Interrupted Collagen
imino acid content, sequence disruption, and prob-
ably external influences like lateral assembly in
crystals. One can easily imagine a macroscopic
model for this plasticity in the form of a three-
stranded rope that can be twisted or relaxed locally.
Extending this model further, this rope may react
differently to torque forces along its length, with
perhaps local rigid spots at which further twisting
(or relaxing) may not be feasible.
Proline puckering distribution is altered in
interrupted collagen peptides
All collagen-peptide crystal structures published
to date show a clear predominance of the up con-
formation for the imino acid in the Yposition,
23
and
the same tendency is observed in the Hyp

structure. The preference of Hyp residues for the

up conformation is the basis of the propensity-based
hypothesis for the stabilization of the collagen triple
helix by hydroxyproline.
38
Unfavourable Hyp puck-
ering would explain the dramatic destabilization of
host-guest collagen peptides containing HypPro
Gly and ProalloHyp-Gly triplets,
39
although it
cannot account for the observed stability of the
peptide (HypHypGly)
9
in which Hyp residues in
the X position adopt the up conformation.
19
Pro re-
sidues do not show such a clear puckering pre-
ference in all crystal structures. In the structures of
peptides with uninterrupted collagen sequences,
Pro residues in the X position have a clear preference
for the down conformation.
10,12,13,15,17,18
In contrast,
Figure 8. Water distribution in the Hyp

structure as a function of intermolecular separation. (a) Top view of the six

closest neighbours around the ABC triple helix in the crystal lattice (ABC molecules shown in red; DEF molecules in blue).
Interaxial distances are shown. The two triple helices separated horizontally by 12.2 form the asymmetric unit in the
structure. (b) Projection of cylindrical coordinates for lattice water molecules around the ABC triple helix (centred at the
origin of coordinates and with its helical axis along the Z cylindrical axis). The origin of angular coordinates (=0) has
been defined at the direction of shortest interaxial separation between triple helices. Water molecules are shown between
cylindrical radius limits of 4 and 14 . (c) Histogram showing the distribution of water molecules in (b) as a function of
the orientation angle .
307 Crystal Structure of an Interrupted Collagen
the puckering of Pro residues in the Hyp

and
GlyAla structures showa more mixed distribution,
with a slight preference for the down conformation:
35 conformers down against 24 up in the Hyp

structure (including a 6:6 split in the Pro residues

from the central zone); 17 conformers down against
13 conformers up in the GlyAla structure.
8
It
is possible that this atypical puckering distribution
in Pro residues is a result of the adaptation of the
triple-helical conformation to interruptions. It is un-
known if such differences in Pro puckering would
have any significance for the thermal stability.
Implications for interruptions in collagen
proteins
The structure of the Hyp

peptide provides the

first look at the consequences of breaking the typical
GlyXY repeating pattern with the absence of
one residue in the Y position. Similar perturbations
of triple-helical structure and hydrogen bonding are
likely to occur around GlyXGly interruptions
found in the triple helix domains of non-fibrillar
collagens and host-defense proteins.
At low resolution, sequence interruptions in
triple-helical domains have been associated both
with flexible sites, as in type IV collagen,
40,41
and
permanent kinks as in the C1q molecule.
42
The high-
resolution crystal structure of the GlyAla peptide
shows a local unwinding and water-mediated
hydrogen bonds at the site of the substitution, but
no indication of a kink.
8
These interruptions will
alter the normal triple-helical structure in a highly
localized manner, with increased flexibility and loss
of conformational equivalence between the three
chains at the interruption site. A significant degree
of molecular plasticity could be expected at such
sites, which is likely to be biologically significant. It
has been suggested that the GlyGlnGly site in the
homotrimeric MBL results in a kink, on the basis of
electron microscopy of rotary shadowed samples
31
and its similarity to C1q.
42
Such a kink is not un-
expected at a flexible, conformationally variable site,
but the Hyp

structure demonstrates that it would

be wrong to assume that GlyXGly interruptions
necessarily result in kinks or bends on collagen triple
helices. Thus, the rod-like domain in the hexagonal
array of Descemet's membrane can be assigned to
the homotrimeric triple helix in type VIII collagen,
despite containing eight GlyXGly sites.
43
In addition to forming rod-like segments, triple-
helical domains in collagen proteins are involved in
self-association and binding to other molecules. A
collagen molecule harbouring a GlyXGly inter-
ruption can compensate for angular displacements,
as seen in the Hyp

structure, but cannot recover

the loss of molecular length resulting from having a
dipeptide rather than a tripeptide unit. Such an
interrupted collagen molecule therefore could not
be easily incorporated into long-range fibrillar
assemblies together with uninterrupted triple-heli-
cal molecules, since adjacent molecules wouldbe out
of register. Thus, a GlyXGly interruption could
serve to prevent the simple staggered self-associa-
tion seen in D-period fibrillar collagens. On the other
hand, because of the loss of axial coherence in the
triple helix, a GlyXGly site may provide an
axial registration marker in other supramolecular
Figure 9. Examples of intermo-
lecular water motifs between the
ABC and DEF triple helices in the
Hyp

crystal structure. (a) Fusion of

two 2 bridges, one fromeach triple
helix. The same motif is repeated at
either side of the central zone in the
same angular orientation. Residue
coloring scheme as in Figure 1(b),
water molecules shown as cyan
spheres and hydrogen bonds as
green sticks. (b) Detail of the water
motif shown in (a). Peptide atoms
are coloured by chemical type,
water molecules and hydrogen
bonds shown as in (a). (c) Cluster
of two 2 bridges connected by
additional water molecules. The
same motif is repeated at either
side of the central zone in the same
angular orientation. Colouring
scheme as in (a). (d) Detail of the
water motif shown in (c). Colouring
scheme as in (b). See Bella et al.
9
for
the nomenclature of the different
types of water bridges.
308 Crystal Structure of an Interrupted Collagen
associations, favouring in-register parallel packing
similar to that seen in the crystal lattice of the Hyp

peptide, where the GlyProGlyPro zones come

together (Figure 7). Alternatively, some GlyXGly
interruptions could provide recognition sites for
ligand binding. It is interesting to note that breaks in
a repeating sequence pattern are also observed in the
other supercoiled motif, coiled-coil -helices, where
stutter and stammer residues that break the repeat-
ing heptad pattern have been suggested to affect the
supercoiling and molecular recognition.
44
Materials and Methods
Crystallization and data collection
The Hyp

peptide was synthesized at the Protein

Microchemistry Laboratory at the Center for Advanced
Biotechnology and Medicine (CABM). Synthesis was
performed on an Applied Biosystem 430A peptide
synthesizer by step-wise solid phase procedures on
t-Boc-l-Gly PAM resin. Thin, plate-like crystals were
grown at 4 C using vapour diffusion methods. The best
crystals originated from drops containing initial concen-
trations of 5.4 mg/ml peptide, 10% (v/v) acetic acid and
15% (w/v) PEG 400, equilibrated against a reservoir
containing 30% PEG 400. The single specimen used for
X-ray diffraction experiments had grown to about
0.50 mm0.25 mm0.10 mm in size.
X-ray diffraction data were collected in a Rigaku RU200
rotating anode diffractometer (CuK radiation), with an
R-AXIS II area detector placed at a distance of 135 mm.
The crystal was mounted in a capillary and subsequently
vitrified under a nitrogen stream at 150 C.
Diffraction images were indexed and integrated using
DENZO and scaled and merged with SCALEPACK.
45
Space group, unit cell parameters and pertinent data
collection statistics are shown in Table 3. Due to crystal
and diffraction anisotropy the completeness of the data
was lower than in other crystal structures of collagen
peptides. As all three unit cell angles are close to 90, both
the unit cell dimensions and the space group were
reconfirmed by accurate individual diffraction measure-
ments in an Enraf Nonius CAD4 diffractometer. The unit
cell volume (about 27,000
3
) indicates that there are two
triple helices per asymmetric unit with a Matthews
coefficient of 1.8
3
/Da.
Structure determination and refinement
The structure was determined by molecular replace-
ment methods using AMoRe.
46
An idealized 7
5
triple helix
with the sequence (ProHypGly)
10
was constructed as
the parent model. Then the central Hyp residue was
deleted on each chain, and the two remaining fragments
were joined together with appropriate stereochemistry
using the graphics program CHAIN.
47
A rotation search
using this modified model yielded 14 different solutions
that could be grouped as two sets, each set of solutions
related by 7-fold symmetry. Rigid-body refinement of
these solutions with the translation vector (0,0,0) pro-
duced one top solution with an R
factor
of 53.7% and five
solutions with R-factors between 57% and 63%. The
remainder of the solutions gave much poorer results.
The first helix was then fixed at the origin using the top
rotation solution and a translation search was done for the
second helix using each of the next five highest rotation
solutions. Six translation solutions for the second helix
with an R
factor
of 53% were produced from the search.
Final rigid-body refinement was applied to both helices
using each of the top six solutions for the second helix. The
best solution had an R
factor
of 48% with the next best
solution having an R
factor
of 52.5%.
The model fromAMoRe was first subjected to rounds of
group rigid-body refinement. Different treatments were
applied for the different regions in the model during rigid
body refinement. For the central ProGly dipeptides each
residue was treated as one group, while for the rest of the
model each tripeptide was treated as one group. The
special treatment for the center residues was introduced to
reduce the bias fromthe initial model. After the rigid-body
refinement, the overall structure was confirmed correct by
the electron density map.
This model was subjected to several rounds of
refinement including positional refinement and simu-
lated annealing in X-PLOR
48
and manual rebuilding of
the residues in the GlyProGlyPro zone with the
molecular graphics program O.
49
CNS
50
was used to
refine the structure at the later stage with a maximum
likelihood target function. No sigma cutoff of the data
was applied and 10% of the data were excluded for the
calculation of the free-R value.
51
Bulk solvent and overall
anisotropic B-factor corrections were used in the refine-
ment. Two sets of non-crystallographic symmetry (NCS)
restraints were initially applied: a threefold NCS restraint
was applied to the three chains in each of the triple
helices; a twofold NCS restraint was applied to the two
triple helices in the asymmetric unit. Tight NCS restraints
were applied throughout the refinement until conver-
gence. Then the threefold NCS restraints were released
and several more cycles of refinement using CNS were
applied. The twofold NCS restraints were released
towards the end of the refinement before addition of
water molecules. Release of the NCS in the late stages of
the refinement did reveal differences in the central region
between the ABC and DEF triple helices as well as
between the three chains in each triple helix. Water
molecules were added to the model when they fit all the
following criteria: peaks of density in the F
o
F
c
difference
Table 3. Statistics for data collection and refinement
A. Data collection
Detector type R-AXIS II
Temperature (C) 150
Wavelength () 1.5418
Data range () 27.02.00
Unit cell axes () 14.15 23.77 82.21
Unit cell angles () 85.86 85.54 84.34
Space group P1
Unique reflections 5124
Overall completeness (%) 71.7
Last shell completeness (%) (2.032.00 ) 66.0
<I/> 14.0
R
merge
0.066
B. Refinement
Resolution range () 272.00
Final R
factor
0.217
Final R
free
0.289
Protein atoms 1078
Water molecules 249
Average B-factor (
2
) 15.9
Bond lengths r.m.s.d. () 0.008
Bond angles r.m.s.d. () 1.67
Improper angles r.m.s.d. () 2.36
309 Crystal Structure of an Interrupted Collagen
map higher than 2; spherical density in both 2F
o
F
c
and
F
o
F
c
maps; reasonable hydrogen-bonding coordination;
positional stability after refinement, and recovery of
electron density in simulated annealing omit maps
calculated after refinement. The final model includes
249 water molecules. The last three residues in chain C
(Pro87Hyp88Gly89) were not modeled due to the poor
electron density in that terminal region. Final refinement
statistics are shown in Table 3.
Acknowledgements
This work was supported by National Institutes of
Health grants GM21589 (to H.M.B.) and GM60048
(to B.B.).
References
1. Ramachandran, G. N. &Kartha, G. (1955). Structure of
collagen. Nature, 176, 593595.
2. Rich, A. & Crick, F. H. C. (1955). The structure of
collagen. Nature, 176, 915916.
3. Cowan, P. M., McGavin, S. & North, A. C. (1955). The
polypeptide chain configuration of collagen. Nature,
176, 10621064.
4. Okuyama, K., Tanaka, N., Ashida, T., Kakudo, M. &
Sakakibara, S. (1972). An X-ray study of the
synthetic polypeptide (Pro-Pro-Gly)
10
. J. Mol. Biol.
72, 571576.
5. Okuyama, K., Okuyama, K., Arnott, S., Takayanagi,
M. & Kakudo, M. (1981). Crystal and molecular
structure of a collagen-like polypeptide (Pro-Pro-
Gly)
10
. J. Mol. Biol. 152, 427443.
6. Okuyama, K., Takayanagi, M., Ashida, T. & Kakudo,
M. (1977). A new structural model for collagen. Polym.
J. 9, 341343.
7. Okuyama, K., Xu, X., Iguchi, M. & Noguchi, K. (2006).
Revision of collagen molecular structure. Biopolymers,
84, 181191.
8. Bella, J., Eaton, M., Brodsky, B. & Berman, H. M.
(1994). Crystal and molecular structure of a collagen-
like peptide at 1.9 resolution. Science, 266, 7581.
9. Bella, J., Brodsky, B. & Berman, H. M. (1995).
Hydration structure of a collagen peptide. Structure,
3, 893906.
10. Kramer, R. Z., Vitagliano, L., Bella, J., Berisio, R., Maz-
zarella, L., Brodsky, B., Zagari, A. & Berman, H. M.
(1998). X-ray crystallographic determination of a
collagen-like peptide with the repeating sequence
(Pro-Pro-Gly). J. Mol. Biol. 280, 623638.
11. Kramer, R. Z., Bella, J., Mayville, P., Brodsky, B. &
Berman, H. M. (1999). Sequence dependent conforma-
tional variations of collagen triple-helical structure.
Nature Struct. Biol. 6, 454457.
12. Nagarajan, V., Kamitori, S. & Okuyama, K. (1999).
Structure analysis of a collagen-model peptide with a
(Pro-Hyp-Gly) sequence repeat. J. Biochem. (Tokyo),
125, 310318.
13. Kramer, R. Z., Venugopal, M. G., Bella, J., Mayville, P.,
Brodsky, B. & Berman, H. M. (2000). Staggered
molecular packing in crystals of a collagen-like peptide
with a single charged pair. J. Mol. Biol. 301, 11911205.
14. Berisio, R., Vitagliano, L., Mazzarella, L. & Zagari, A.
(2000). Crystal structure of a collagen-like polypeptide
with repeating sequence Pro-Hyp-Gly at 1.4 reso-
lution: implications for collagen hydration. Biopoly-
mers, 56, 813.
15. Kramer, R. Z., Bella, J., Brodsky, B. & Berman, H. M.
(2001). The crystal and molecular structure of a
collagen-like peptide with a biologically relevant
sequence. J. Mol. Biol. 311, 131147.
16. Berisio, R., Vitagliano, L., Mazzarella, L. & Zagari, A.
(2002). Crystal structure of the collagen triple helix
model [(Pro-Pro-Gly)
10
]
3
. Protein Sci. 11, 262270.
17. Emsley, J., Knight, C. G., Farndale, R. W. & Barnes,
M. J. (2004). Structure of the integrin 21-binding
collagen peptide. J. Mol. Biol. 335, 10191028.
18. Okuyama, K., Hongo, C., Fukushima, R., Wu, G.,
Narita, H., Noguchi, K. et al. (2004). Crystal structures
of collagen model peptides with Pro-Hyp-Gly repeat-
ing sequence at 1.26 resolution: implications for
proline ring puckering. Biopolymers, 76, 367377.
19. Schumacher, M., Mizuno, K. &Bachinger, H. P. (2005).
The crystal structure of the collagen-like polypeptide
(glycyl-4(R)-hydroxyprolyl-4(R)-hydroxyprolyl)
9
at 1.55 resolution shows up-puckering of the
proline ring in the Xaa position. J. Biol. Chem. 280,
2039720403.
20. Kawahara, K., Nishi, Y., Nakamura, S., Uchiyama,
S., Nishiuchi, Y., Nakazawa, T. et al. (2005).
Effect of hydration on the stability of the col-
lagen-like triple-helical structure of [4(R)-hydroxy-
prolyl-4(R)-hydroxyprolylglycine]
10
. Biochemistry,
44, 1581215822.
21. Okuyama, K., Wu, G., Jiravanichanun, N., Hongo, C.
& Noguchi, K. (2006). Helical twists of collagen model
peptides. Biopolymers, 84, 421432.
22. Jiravanichanun, N., Nishino, N. & Okuyama, K.
(2006). Conformation of alloHyp in the Y position in
the host-guest peptide with the Pro-Pro-Gly sequence:
implication of the destabilization of (Pro-alloHyp-
Gly)
10
. Biopolymers, 81, 225233.
23. Brodsky, B. & Persikov, A. V. (2005). Molecular
structure of the collagen triple helix. Adv. Protein
Chem. 70, 301339.
24. Boot-Handford, R. P., Tuckwell, D. S., Plumb, D. A.,
Rock, C. F. & Poulsom, R. (2003). A novel and
highly conserved collagen (pro1(XXVII)) with a
unique expression pattern and unusual molecular
characteristics establishes a new clade within the
vertebrate fibrillar collagen family. J. Biol. Chem. 278,
3106731077.
25. Kuivaniemi, H., Tromp, G. & Prockop, D. J. (1991).
Mutations in collagen genes: causes of rare and some
common diseases in humans. FASEB J. 5, 20522060.
26. Prockop, D. J. & Kivirikko, K. I. (1995). Collagens:
molecular biology, diseases, and potentials for ther-
apy. Annu. Rev. Biochem. 64, 403434.
27. Brazel, D., Oberbaumer, I., Dieringer, H., Babel, W.,
Glanville, R. W., Deutzmann, R. & Kuhn, K. (1987).
Completion of the amino acid sequence of the 1
chain of human basement membrane collagen (type
IV) reveals 21 non-triplet interruptions located
within the collagenous domain. Eur. J. Biochem.
168, 529536.
28. Soininen, R., Haka-Risku, T., Prockop, D. J. &
Tryggvason, K. (1987). Complete primary structure
of the 1-chain of human basement membrane (type
IV) collagen. FEBS Letters, 225, 188194.
29. Yamaguchi, N., Benya, P. D., van der Rest, M. &
Ninomiya, Y. (1989). The cloning and sequencing of 1
(VIII) collagen cDNAs demonstrate that type VIII
collagen is a short chain collagen and contains triple-
310 Crystal Structure of an Interrupted Collagen
helical and carboxyl-terminal non-triple-helical
domains similar to those of type X collagen. J. Biol.
Chem. 264, 1602216029.
30. Oka, S., Itoh, N., Kawasaki, T. & Yamashina, I. (1987).
Primary structure of rat liver mannan-binding protein
deduced from its cDNA sequence. J. Biochem. (Tokyo),
101, 135144.
31. Lu, J., Wiedemann, H., Timpl, R. & Reid, K. B. (1993).
Similarity in structure between C1q and the collectins
as judged by electron microscopy. Behring Inst. Mitt.,
616.
32. Hoppe, H. J. & Reid, K. B. (1994). CollectinsSoluble
proteins containing collagenous regions and lectin
domainsand their roles in innate immunity. Protein
Sci. 3, 11431158.
33. Hudson, B. G., Tryggvason, K., Sundaramoorthy, M.
& Neilson, E. G. (2003). Alport's syndrome, Good-
pasture's syndrome, and Type iv Collagen. N. Engl. J.
Med. 348, 25432556.
34. Long, C. G., Braswell, E., Zhu, D., Apigo, J., Baum, J. &
Brodsky, B. (1993). Characterization of collagen-like
peptides containing interruptions in the repeating
Gly-X-Y sequence. Biochemistry, 32, 1168811695.
35. Bella, J. & Berman, H. M. (1996). Crystallographic
evidence for C-H

O=C hydrogen bonds in a

collagen triple helix. J. Mol. Biol. 264, 734742.
36. Rich, A. & Crick, F. H. (1961). The molecular structure
of collagen. J. Mol. Biol. 3, 483506.
37. Fraser, R. D., MacRae, T. P. & Suzuki, E. (1979). Chain
conformation in the collagen molecule. J. Mol. Biol.
129, 463481.
38. Vitagliano, L., Berisio, R., Mazzarella, L. & Zagari, A.
(2001). Structural bases of collagen stabilization
induced by proline hydroxylation. Biopolymers, 58,
459464.
39. Jiravanichanun, N., Hongo, C., Wu, G., Noguchi, K.,
Okuyama, K., Nishino, N. & Silva, T. (2005). Unex-
pected puckering of hydroxyproline in the guest
triplets, Hyp-Pro-Gly and Pro-alloHyp-Gly sand-
wiched between Pro-Pro-Gly sequence. Chem. Bio.
Chem. 6, 11841187.
40. Hofmann, H., Voss, T., Kuhn, K. & Engel, J. (1984).
Localization of flexible sites in thread-like molecules
from electron micrographs. Comparison of interstitial,
basement membrane and intima collagens. J. Mol. Biol.
172, 325343.
41. Siebold, B., Qian, R. A., Glanville, R. W., Hofmann,
H., Deutzmann, R. & Kuhn, K. (1987). Construction
of a model for the aggregation and cross-linking
region (7S domain) of type IV collagen based upon
an evaluation of the primary structure of the 1
and 2 chains in this region. Eur. J. Biochem. 168,
569575.
42. Kilchherr, E., Hofmann, H., Steigemann, W. &Engel, J.
(1985). Structural model of the collagen-like region of
C1q comprising the kink region and the fibre-like
packing of the six triple helices. J. Mol. Biol. 186,
403415.
43. Sawada, H., Konomi, H. & Hirosawa, K. (1990).
Characterization of the collagen in the hexagonal
lattice of Descemet's membrane: its relation to type
VIII collagen. J. Cell Biol. 110, 219227.
44. Oas, T. G. & Endow, S. A. (1994). Springs and hinges:
dynamic coiled coils and discontinuities. Trends
Biochem. Sci. 19, 5154.
45. Otwinowski, Z. & Minor, W. (1997). Processing of
X-ray diffraction data collected in oscillation mode.
Methods Enzymol. 276, 307326.
46. Navaza, J. (1994). AMoRe: An automated package for
molecular replacement. Acta Crystallog. sect. A, 50,
157163.
47. Sack, J. S. (1988). CHAINA crystallographic model-
ing program. J. Mol. Graph. 6, 224225.
48. Brnger, A. T. (1992). X-PLOR. Version 3.1. A system
for X-ray crystallography and NMR, 1st edit. Yale
University Press, New Haven, CT.
49. Jones, T. A., Zou, J.-Y., Cowan, S. W. & Kjeldgaard, M.
(1991). Improved methods for building protein
models in electron density maps and the location of
errors in these models. Acta Crystallog. sect. A, 47,
110119.
50. Brnger, A. T., Adams, P. D., Clore, G. M., DeLano,
W. L., Gros, P., Grosse-Kuntsleve, R. W. et al. (1998).
Crystallography and NMR system: a new software
suite for macromolecular structure determination.
Acta Crystallog. sect. D, 54, 905921.
51. Brnger, A. T. (1992). Free R value: a novel statistical
quantity for assessing the accuracy of crystal struc-
tures. Nature, 355, 472475.
52. Evans, S. V. (1993). SETOR: hardware lighted three-
dimensional solid model representation of macromo-
lecules. J. Mol. Graph. 11, 134138.
Edited by A. Klug
(Received 4 October 2005; received in revised form 9 May 2006; accepted 10 July 2006)
Available online 15 July 2006
311 Crystal Structure of an Interrupted Collagen