Collagen Structure and Stability

ANRV378-BI78-32 ARI 5 May 2009 15:11

Collagen Structureand StabilityMatthew D. Shoulders1 and Ronald T. Raines1,21Department of Chemistry and 2Department of Biochemistry, University of Wisconsin,Madison, Wisconsin 53706; email: [email protected]

Annu. Rev. Biochem. 2009. 78:92958

First published online as a Review in Advance onApril 3, 2009

The Annual Review of Biochemistry is online atbiochem.annualreviews.org

This articles doi:10.1146/annurev.biochem.77.032207.120833

Copyright c 2009 by Annual Reviews.All rights reserved

0066-4154/09/0707-0929$20.00

Key Words

biomaterials, extracellullar matrix, brillogenesis, proline,stereoelectronic effects, triple helix

AbstractCollagen is the most abundant protein in animals. This brous, struc-tural protein comprises a right-handed bundle of three parallel, left-handed polyproline II-type helices. Much progress has been made inelucidating the structure of collagen triple helices and the physicochem-ical basis for their stability. New evidence demonstrates that stereoelec-tronic effects and preorganization play a key role in that stability. Thebrillar structure of type I collagenthe prototypical collagen brilhas been revealed in detail. Articial collagen brils that display someproperties of natural collagen brils are now accessible using chemicalsynthesis and self-assembly. A rapidly emerging understanding of themechanical and structural properties of native collagen brils will guidefurther development of articial collagenous materials for biomedicineand nanotechnology.

929

Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

ECM: extracellularmatrix

PPII: polyproline IItype

Hyp: (2S,4R)-4-hydroxyproline

Tropocollagen (TC):the monomericcollagen triple helixafter proteolysis ofcollagen propeptides

Contents

INTRODUCTION . . . . . . . . . . . . . . . . . . 930STRUCTURE OF THE

COLLAGEN TRIPLE HELIX . . . . 930UNDERSTANDING

TRIPLE-HELIX STRUCTUREAND STABILITY . . . . . . . . . . . . . . . . . 934Interstrand Hydrogen Bonds . . . . . . . 935Glycine Substitutions . . . . . . . . . . . . . . 935Prolines in the Xaa

and Yaa Positions . . . . . . . . . . . . . . . 936Role of Hyp . . . . . . . . . . . . . . . . . . . . . . . 937Proline Derivatives in the Xaa

Position . . . . . . . . . . . . . . . . . . . . . . . . 940An n Interaction . . . . . . . . . . . . . . 941Hyp in the Xaa Position . . . . . . . . . . . . 942Heterotrimeric Synthetic

Triple Helices . . . . . . . . . . . . . . . . . . 942Nonproline Substitutions

in the Xaa and Yaa Positions . . . . . 944HIGHER-ORDER COLLAGEN

STRUCTURE . . . . . . . . . . . . . . . . . . . . 944Fibril Structure . . . . . . . . . . . . . . . . . . . . 944Nucleation and Modulation

of Collagen Fibrillogenesis . . . . . . 946MECHANICAL PROPERTIES

OF COLLAGEN FIBRILS . . . . . . . . 946COLLAGENOUS

BIOMATERIALS . . . . . . . . . . . . . . . . . 947Collagen via Chemical Synthesis . . . . 947Biological and Biomedical

Applications of SyntheticCollagen . . . . . . . . . . . . . . . . . . . . . . . 948

INTRODUCTION

Collagen is an abundant structural protein in allanimals. In humans, collagen comprises one-third of the total protein, accounts for three-quarters of the dry weight of skin, and is themost prevalent component of the extracellu-lar matrix (ECM). Twenty-eight different typesof collagen composed of at least 46 distinctpolypeptide chains have been identied in ver-tebrates, and many other proteins contain col-lagenous domains (1, 2). Remarkably, intact

collagen has been discovered in soft tissue of thefossilized bones of a 68 million-year-old Tyran-nosaurus rex fossil (3, 4), by far the oldest proteindetected to date. That discovery is, however,under challenge (5, 6).

The dening feature of collagen is an el-egant structural motif in which three parallelpolypeptide strands in a left-handed, polypro-line II-type (PPII) helical conformation coilabout each other with a one-residue staggerto form a right-handed triple helix (Figure 1).The tight packing of PPII helices within thetriple helix mandates that every third residuebe Gly, resulting in a repeating XaaYaaGlysequence, where Xaa and Yaa can be anyamino acid. This repeat occurs in all typesof collagen, although it is disrupted at cer-tain locations within the triple-helical domainof nonbrillar collagens (8). The amino acidsin the Xaa and Yaa positions of collagen areoften (2S )-proline (Pro, 28%) and (2S,4R)-4-hydroxyproline (Hyp, 38%), respectively.ProHypGly is the most common triplet(10.5%) in collagen (9). In animals, individualcollagen triple helices, known as tropocollagen(TC), assemble in a complex, hierarchical man-ner that ultimately leads to the macroscopicbers and networks observed in tissue, bone,and basement membranes (Figure 2).

The categories of collagen include the clas-sical brillar and network-forming collagens,the FACITs (bril-associated collagens with in-terrupted triple helices), MACITs (membrane-associated collagens with interrupted triplehelices), and MULTIPLEXINs (multipletriple-helix domains and interruptions). Colla-gen types, their distribution, composition, andpathology are listed in Table 1. It is notewor-thy that, although the three polypeptide chainsin the triple helix of each collagen type can beidentical, heterotrimeric triple helices are moreprevalent than are homotrimeric triple helices.

STRUCTURE OF THECOLLAGEN TRIPLE HELIX

In 1940, Astbury & Bell (11) proposed thatthe collagen molecule consists of a single

930 Shoulders Raines

Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

extended polypeptide chain with all amidebonds in the cis conformation. A signicant ad-vance was achieved when, in the same 1951 is-sue of the Proceedings of the National Academy ofSciences in which they put forth the correctstructures for the -helix and -sheet, Pauling& Corey (12) proposed a structure for colla-gen. In that structure, three polypeptide strandswere held together in a helical conformationby hydrogen bonds. Within each amino acidtriplet, those hydrogen bonds engaged fourof the six main chain heteroatoms, and theirformation required two of the three peptidebonds to be in the cis conformation. In 1954,Ramachandran & Kartha (13, 14) advanced astructure for the collagen triple helix on thebasis of ber diffraction data. Their structurewas a right-handed triple helix of three stag-gered, left-handed PPII helices with all peptidebonds in the trans conformation and two hy-drogen bonds within each triplet. In 1955, thisstructure was rened by Rich & Crick (1516)and by North and coworkers (17) to the triple-helical structure accepted today, which has asingle interstrand NH(Gly) O==C(Xaa) hydro-gen bond per triplet and a tenfold helical sym-metry with a 28.6-A axial repeat (10/3 helicalpitch) (Figure 1).

Fiber diffraction studies cannot reveal thestructure of collagen at atomic resolution.Exacerbating this difculty, the large size, in-solubility, repetitive sequence, and complex hi-erarchical structure of native collagen thwartmost biochemical and biophysical analyses.Hence, a reductionist approach using triple-helical, collagen-related peptides (CRPs) hasbeen employed extensively since the late 1960s(18).

In 1994, Berman and coworkers (19) re-ported the rst high-resolution crystal struc-ture of a triple-helical CRP (Figure 1a). Thisstructure conrmed the existence of inter-strand NH(Gly) O==C(Xaa) hydrogen bonds(Figure 1c,d ) and provided additional insights,including that CH(Gly/Yaa) O==C(Xaa/Gly) hy-drogen bonds could likewise stabilize the triplehelix (20). Using CRPs and X-ray crystal-lography, the structural impact of a single

N1

a b

N1c

N2

N3

Hyp

Gly

Pro

HypC2C1

C3

dN1 N2 N3

HypHyp

Hyp

ProC=O ProC=O

Pro

Gly

GlyNHHNGly

C1 C2 C3

HNGly

O=CPro

Hyp

Figure 1Overview of the collagen triple helix. (a) First high-resolution crystal structureof a collagen triple helix, formed from (ProHypGly)4(ProHypAla)(ProHypGly)5 [Protein Data Bank (PDB) entry 1cag] (19). (b) View down theaxis of a (ProProGly)10 triple helix [PDB entry 1k6f (7)] with the three strandsdepicted in space-lling, ball-and-stick, and ribbon representation. (c) Ball-and-stick image of a segment of collagen triple helix [PDB entry 1cag (19)],highlighting the ladder of interstrand hydrogen bonds. (d ) Stagger of the threestrands in the segment in panel c.

CRP: collagen-relatedpeptide

Gly Ala substitution was observed (19), theeffects of neighboring charged residues in atriple helix were analyzed (21), and a snapshotof the interaction of a triple-helical CRP withthe I domain of integrin 21 was obtained(Figure 3) (22).

Most X-ray crystallographic studies onCRPs have been performed on proline-richcollagenous sequences. All of the resultingstructures have a 7/2 helical pitch (20.0-A axialrepeat), in contrast to the 10/3 helical pitch(28.6-A axial repeat) predicted for naturalcollagen by ber diffraction (17). On thebasis of X-ray crystal structures of proline-rich

www.annualreviews.org Collagen Structure and Stability 931

Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

Figure 2Biosynthetic route to collagen bers (110), which are the major component of skin. Size and complexity areincreased by posttranslational modications and self-assembly. Oxidation of lysine side chains leads to thespontaneous formation of hydroxylysyl pyridinoline and lysyl pyridinoline cross-links.

CRPs, and in accordance with an early proposalregarding the helical pitch of natural triplehelices (23), Okuyama and coworkers (24)postulated that the correct average helicalpitch for natural collagen is 7/2. The generality

of this hypothesis is unclear, as few regions ofnatural collagen are as proline rich as the CRPsanalyzed by X-ray crystallography. The actualhelical pitch of collagen likely varies acrossthe domains and types of natural collagen.


N+

HN

O

HONH

O

HN

O

OH

N+

HN

O

HONH

O

HN

O

Hydroxylysyl pyridinoline Lysyl pyridinoline

0.8 nm

300 nmProtocollagen strand

N- and C-terminalpropeptides

P4H, P3HLysyl hydroxylase

Protein disulfide isomerase

12 nm

300 nmProcollagen triple helix

300 nmTropocollagen triple helix

ProcollagenN- and C-proteinases

ANRV378-BI78-32 ARI 5 May 2009 15:11

Table 1 Vertebrate collagensa

Type Class Composition Distributionb Pathologyc

I Fibrillar 1[I]22[I] Abundant and widespread: dermis,bone, tendon, ligament

OI, EhlersDanlos syndrome,osteoporosis

II Fibrillar 1[II]3 Cartilage, vitreous Osteoarthrosis, chondrodysplasiasIII Fibrillar 1[III]3 Skin, blood vessels, intestine Ehlers-Danlos syndrome, arterial

aneurysmsIV Network 1[IV]22[IV] Basement membranes Alport syndrome

3[IV]4[IV]5[IV]5[IV]26[IV]

V Fibrillar 1[V]3 Widespread: bone, dermis, cornea,placenta

Ehlers-Danlos syndrome1[V]22[V]

1[V]2[V]3[V]VI Network 1[VI]2[VI] 3[VI]d Widespread: bone, cartilage, cornea,

dermisBethlem myopathy

1[VI]2[VI] 4[VI]VII Anchoring brils 1[VII]22[VII] Dermis, bladder Epidermolysis bullosa acquisitaVIII Network 1[VIII]3 Widespread: dermis, brain, heart,

kidneyFuchs endothelia corneal dystrophy

2[VIII]31[VIII]22[VIII]

IX FACITe 1[IX]2[IX]3[IX] Cartilage, cornea, vitreous Osteoarthrosis, multiple epiphysealdysplasia

X Network 1[X]3 Cartilage ChondrodysplasiaXI Fibrillar 1[XI]2[XI]3[XI] Cartilage, intervertebral disc Chondrodysplasia, osteoarthrosisXII FACIT 1[XII]3 Dermis, tendon XIII MACIT Endothelial cells, dermis, eye, heart XIV FACIT 1[XIV]3 Widespread: bone, dermis, cartilage XV MULTIPLEXIN Capillaries, testis, kidney, heart XVI FACIT Dermis, kidney XVII MACIT 1[XVII]3 Hemidesmosomes in epithelia Generalized atrophic epidermolysis

bullosaXVIII MULTIPLEXIN Basement membrane, liver Knobloch syndromeXIX FACIT Basement membrane XX FACIT Cornea (chick) XXI FACIT Stomach, kidney XXII FACIT Tissue junctions XXIII MACIT Heart, retina XXIV Fibrillar Bone, cornea XXV MACIT Brain, heart, testis Amyloid formation?XXVI FACIT Testis, ovary XXVII Fibrillar Cartilage XXVIIIf Dermis, sciatic nerve Neurodegenerative disease?

aInformation is updated from References 1 and 32.bPartial listing of tissues in which the relevant collagen type occurs.cFor a discussion of the role of specic collagen types in human disease, see Reference 32.d4[VI], 5[VI], and 6[VI] chains were reported in 2008 (10), but the composition of triple helices containing 5[VI] and 6[VI] is unknown.eAbbreviations: FACIT, bril-associated collagen with interrupted triple helices; MACIT, membrane-associated collagen with interrupted triple helices;MULTIPLEXIN, multiple triple-helix domains and interruptions.fCollagen XXVIII was reported in 2006 (2).


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

a

b

c

}

Region where Ala residues replaceGly residues in the (XaaYaaGly) repeat

GluCo2+

Figure 3Snapshots of interesting crystal structures of collagen triple helices. (a) Impact of a GlyAla substitution onthe structure of a collagen triple helix formed from the collagen-related peptide (CRP) (ProHypGly)4(ProHypAla)(ProHypGly)5 [Protein Data Bank (PDB) entry 1cag (19)]. The Ala residues (red ) disturb thestructure. Mutations leading to such structural irregularities are common in osteogenesis imperfecta and canbe lethal. (b) Depiction of the effect of a single GluLysGly triplet on the packing of neighboringtriple-helical CRPs in crystalline (ProHypGly)4(GluLysGly)(ProHypGly)5 [PDB entry 1qsu (21)]. Theaxial stagger of the individual triple helices, which is presumably compelled by deleterious Coulombicinteractions between charged residues, is reminiscent of the D-periodic structure in collagen brils. Similarinteractions could contribute to the morphology of collagen brils. (c) Triple-helical CRP containing theintegrin-binding domain GFOGER in complex with the I domain of integrin 21 [PDB entry 1dzi (22)].The bend in the triple helix is thought to arise from the protein-protein interaction. A Glu residue in themiddle strand of the triple helix coordinates to cobalt(II) bound in the I domain of integrin 21.

Specically, the helical pitch could be 10/3 inproline-poor regions and 7/2 in proline-rich re-gions. This proposal is supported by the obser-vation that proline-poor regions within crys-talline CRPs occasionally display a 10/3 helicalpitch (25, 26). Variability in the triple-helicalpitch of native collagen could play a role in theinteraction of collagenous domains with otherbiomolecules (22, 2729).

UNDERSTANDINGTRIPLE-HELIX STRUCTUREAND STABILITY

The vital importance of collagen as a scaffoldfor animals demands a manifold of essentialcharacteristics. These characteristics includethermal stability, mechanical strength, and theability to engage in specic interactions with


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

other biomolecules. Understanding how suchproperties are derived from the fundamentalstructural unit of collagenthe triple helixnecessitates a comprehensive knowledge of themechanisms underlying triple-helix structureand stability.

Interstrand Hydrogen Bonds

The ubiquity of collagen makes the ladderof recurrent NH(Gly) O==C(Xaa) hydro-gen bonds that form within the triple helix(Figure 1c,d ) the most abundant amideamidehydrogen bond in kingdom Animalia. Replac-ing the YaaGly amide bond with an esterin a host-guest CRP (Figure 4a,b) enabledestimation of the strength of each amideamidehydrogen bond as G = 2.0 kcal/mol (30).Boryskina and coworkers (31) used a varietyof other experimental techniques to assess thissame parameter, estimating the strength ofeach amideamide hydrogen bond within apoly(GlyProPro) CRP as G = 1.8 kcal/moland within native collagen as G =1.4 kcal/mol.

Glycine Substitutions

Numerous collagen-related diseases are associ-ated with mutations in both triple-helical andnontriple-helical domains of various collagens(Table 1). These diseases have been reviewedin detail elsewhere (32) and are not discussedextensively herein.

The Gly residue in the XaaYaaGly repeat isinvariant in natural collagen, and favorable sub-stitutions are unknown in CRPs (33). A compu-tational study suggested that replacing the obli-gate Gly residues of collagen with d-alanine ord-serine would stabilize the triple helix (34) andthus that the Gly residues in collagen are surro-gates for nonnatural d-amino acids. Subsequentexperimental data demonstrated, however, thatthis notion was erroneous (35).

Many of the most damaging mutationsto collagen genes result in the replacementof a Gly residue within the triple helix(Figure 1c,d ). Both the identity of the amino

c

3S-Hyp

Amide bond

d

a b

Ester isostere

CH2

N

N

O

NH

O

F

N

NCH2

ON

O

FN

N

NCH2

O

O

F

O

O F

FO F

H

H

N

NO

H

O

NO

NH

O

N

NO

O

Xaa

XaaYaa

Yaa

N

OO

O

NO

NH

O

N

NO

O

Xaa

XaaYaa

Yaa

Figure 4Importance of interstrand hydrogen bonds for collagen triple-helix stability.(a) A segment of a (ProProGly)10 triple helix. (b) Comparison of the stability ofa triple helix formed from (ProProGly)4ProProOGly(ProProGly)5, whereinone ProGly amide bond is replaced with an ester, with that in panel a revealingthat each interstrand hydrogen bond contributes G = 2.0 kcal/mol totriple-helix stability (30). (c) Crystal structure of a triple helix formed from acollagen-related peptide that mimics a common sequence in type IV collagen,(GlyProHyp)3(3S-HypHypGly)2(GlyProHyp)4, showing that 3S-Hyp in theXaa position yields a prototypical collagen triple helix [PDB entry 2g66 (78)].(d ) (2S,3S )-3-Fluoroproline in the Xaa position destabilizes a collagen triplehelix, perhaps by withdrawing electron density from the proximal Xaa carbonyland thereby reducing the strength of the interstrand hydrogen bond (79).

OI: osteogenesisimperfecta

acid replacing Gly and the location of that sub-stitution can impact the pathology of, for ex-ample, osteogenesis imperfecta (OI) (33, 36).Substitutions for Gly in proline-rich portionsof the collagen sequence (Figure 3a) are far lessdisruptive than those in proline-poor regions, atestament to the importance of Pro derivativesfor triple-helix nucleation (37). In vivo, triplehelices fold in a C-terminalN-terminal man-ner (38). The time delay between disruption oftriple-helix folding by a Gly substitution andrenucleation of the folding process N-terminalto the substitution site is much shorter when


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

Protocollagen: thenonhydroxylated,nontriple-helical formof collagen prior to theaction of P4H, P3H,lysyl hydroxylase, andprotein disuldeisomerase

Preorganization: theextent to which hostsand guests areorganized for bindingprior to theircomplexation, therebyincreasing complexstability

P4H: prolyl4-hydroxylase

triple-helix nucleating, proline-rich sequencesare immediately N-terminal to the substitutionsite (37). Any delay in triple-helix folding resultsin overmodication of the protocollagen chains[in particular, inordinate hydroxylation of Lysresidues N-terminal to the Gly substitution andexcessive glycosylation of the resultant hydrox-ylysine residues (Figure 2)], thereby perturb-ing triple-helical structure and contributing tothe severity of OI (39). Thus, the severity ofOI correlates with the abundance of triple-helixnucleating, proline-rich sequences immediatelyN-terminal to the substitution site (36).

Prolines in the Xaa and Yaa Positions

In the strands of human collagen, 22% of allresidues are either Pro or Hyp (9). The abun-dance of these residues preorganizes the indi-vidual strands in a PPII conformation, therebydecreasing the entropic cost for collagen fold-ing (40). Despite their stabilizing properties,Pro derivatives also have certain deleteriousconsequences for triple-helix folding and sta-bility that partially offset their favorable effects.For example, Pro has a secondary amino groupand forms tertiary amides within a peptide orprotein. Tertiary amides have a signicant pop-ulation of both the trans and the cis isomers(Figure 5), whereas all peptide bonds in col-lagen are trans. Thus, before a (ProHypGly)nstrand can fold into a triple helix, all the cis

NO

O

Ktrans/cis

transcis

NO

O

Figure 5Pro cis-trans isomerization. Unlike otherproteinogenic amino acids, Pro forms tertiary amidebonds, resulting in a signicant population of the cisconformation.

peptide bonds must isomerize to trans. N-Methylalanine (an acyclic, tertiary amide miss-ing only C of Pro) decreases triple-helix stabil-ity when used to replace Pro or Hyp in CRPs,presumably because it lacks the preorganizationimposed by the pyrrolidine ring of Pro deriva-tives (41). In contrast, avoiding the issue of cis-trans isomerization altogether by replacing aGlyPro amide bond with a trans-locked alkeneisostere also results in a destabilized triple helix,despite leaving all interchain hydrogen bondsintact (42). Clearly, the factors dictating triple-helix structure and stability are intertwined in acomplex manner (vide infra).

Pro residues in the Yaa position of pro-tocollagen triplets are modied by prolyl 4-hydroxylase (P4H), a nonheme iron enzymethat catalyzes the posttranslational and stere-oselective hydroxylation of the unactivated-carbon of Pro residues in the Yaa positionof collagen sequences to form Hyp (Figure 6).P4H activity is required for the viability of

P4HFe(II)

O2

CO2

CO2

O

O2C

ProProGly ProHypGly

N

HN

O

O

ON

HH

N

HN

O

O

ON

HOH

O

OO2C

Figure 6Reaction catalyzed by prolyl 4-hydroxylase (P4H). Pro residues in the Yaa position of collagen strands areconverted into Hyp prior to triple-helix formation.


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

both the nematode Caenorhabditis elegans andthe mouse Mus musculus (43, 44). Thus, Hyp isessential for the formation of sound collagen invivo.

Role of Hyp

The hydroxylation of Pro residues in the Yaaposition of collagen increases dramatically thethermal stability of triple helices (Table 2).This stabilization occurs when the resultantHyp is in the Yaa position (45, 46) but notin the Xaa position, nor when the hydroxyl

Table 2 Values of Tm for triple-helical CRPs

(XaaYaaGly)n Tm (C)a References(ProFlpGly)7 45 (54)(ProHypGly)7 36 (54)(mepMepGly)7 36 (65)(pProGly)7 33 (68)(ProMepGly)7 29 (65)(ProClpGly)7 23 (61)(mepProGly)7 13 (65)(clpProGly)7 No helix (61)(ProProGly)7 No helix (79)(pFlpGly)7 No helix (79)(PropGly)7 No helix (54)(FlpProGly)7 No helix (68)(ProFlpGly)10 91 (53)(ProMopGly)10 70 (59)(HypHypGly)10 65 (85)(ProHypGly)10 6169 (53, 85)(pProGly)10 58 (69)(ProClpGly)10 52 (61)(clpProGly)10 33 (61)(ProProGly)10 3141 (53, 64)(pFlpGly)10 30 (94)(clpClpGly)10 No helix (61)(HypProGly)10 No helix (84)(ProhypGly)10 No helix (47)(FlpProGly)10 No helix (69)(ClpProGly)10 No helix (61)(hypProGly)10 No helix (47)

aValues of Tm depend on both CRP concentration andheating rate. Hence, detailed comparisons requireknowledge of experimental procedures.

hyp: (2S,4S )-4-hydroxyproline

Flp: (2S,4R)-4-uoroproline

flp: (2S,4S )-4-uoroproline

Stereoelectroniceffects: relationshipsbetween structure,conformation, energy,and reactivity thatresult from thealignment of lled orunlled electronicorbitals

group is installed in the 4S conguration as in(2S,4S )-4-hydroxyproline (hyp) (Table 2) (47,48). These ndings led to the proposal that the4R conguration of a prolyl hydroxyl group isprivileged in alone enabling the formation ofwater-mediated hydrogen bonds that stitch to-gether the folded triple helix (49). Indeed, suchwater bridges between Hyp and main chainheteroatoms were observed by Berman andcoworkers (19, 50) in their seminal X-ray crys-tallographic studies of CRPs. The frequency ofHyp in most natural collagen is, however, toolow to support an extensive network of waterbridges. For example, four or more repeatingtriads of XaaHypGly occur only twice in theamino acid sequence of human type I collagen.

The hypothesis that the water bridgesobserved in crystalline (ProHypGly)n triplehelices are meaningful was tested by replacingHyp residues in CRPs with (2S,4R)-4-uoroproline (Flp). As uoro groups do notform strong hydrogen bonds (51), waterbridges cannot play a major role in stabilizinga (ProFlpGly)10 triple helix. Nonetheless,(ProFlpGly)10 triple helices are hyperstable(Table 2) (52, 53). Accordingly, water bridgescannot be of fundamental importance fortriple-helix stability. How, then, does 4R-hydroxylation of Yaa-position Pro residuesstabilize the triple helix?

A gauche effect. Replacing Hyp in the Yaaposition with (2S,4S )-4-uoroproline (p),a diastereomer of Flp, prevents triple-helixformation (Table 2) (54). This discoverythat the stereochemistry of electronegativesubstituents at the 4-position of the Pro ringis important for the formation of stable triplehelices suggests that Flp and Hyp in the Yaaposition stabilize collagen via a stereoelectroniceffect, rather than a simple inductive effect(54). Pro and its derivatives prefer one of twomajor pyrrolidine ring puckers, which aretermed C-exo and C-endo (Figure 7). [Thering actually prefers two distinct twist, ratherthan envelope, conformations (55). As C

experiences a large out-of-plane displacementin the twisted rings, we refer to pyrrolidine ring


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

NO

R2

R1

NO

R2

R1

C-endo pucker C-exo pucker

Figure 7Ring conformations of Pro and Pro derivatives. TheC-endo conformation is favored strongly bystereoelectronic effects when R1 = H, R2 = F (p) orCl (clp), and by steric effects when R1 = Me (mep)or SH (mcp), R2 = H. The C-exo conformation isfavored strongly by stereoelectronic effects whenR1 = OH (Hyp), F (Flp), OMe (Mop), or Cl (Clp),R2 = H, and by steric effects when R1 = H, R2 = Me(Mep) or SH (Mcp). The C-endo:C-exo ratio is 2when R1 = R2 = H (56).

puckers simply as C-exo and C-endo.] Proitself has a slight preference for the C-endoring pucker (Table 3) (56). A key attributeof a 4R uoro group on Pro (as well as thenatural 4R hydroxyl group) is its impositionof a C-exo pucker on the pyrrolidine ringvia the gauche effect (Figure 8a,b) (5658).The C-exo ring pucker preorganizes the mainchain torsion angles (, Ci1NiC iCi; ,NiC iCiNi+1; and , C iCiNi+1Ci+1) to those in the Yaa position of a triplehelix (Table 4). Thus, 4R-hydroxylation ofPro residues in the Yaa position of collagenstabilizes the triple helix via a stereoelectroniceffect. Flp is more stabilizing than Hyp becauseuorine (F = 4.0) is more electronegativethan oxygen (O = 3.5), and a uoro group

Table 3 Conformation of Pro and its 4-substituted derivatives that prefer the Xaa position [ = 75, = 164 (7)] in acollagen triple helix

Residue(References)

Crystal structure

Ringpucker

(EendoEexo)a (kcal/mol)

(degrees)

(degrees)

Pro (54, 56, 57) 0.41 79b

76b

177b 4.6

mcp (66) 4.7

mep (65) 1.4

flp (54, 56, 68) 0.61 172a 2.5

clp (61)

aFrom DFT calculations. bValues of and (here, NiCiCiOi+1) are from the crystal structure of Ac-Pro-OMe, which has a cis peptide bond.cValues of Ktrans/cis (Figure 5) were determined in solution by NMR spectroscopy.

Ktrans/cisc

C-endo

C-endo

C-endo

C-endo

C-endo 2.2

3.7


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

(FF = 0.45) manifests a greater inductive effectthan does a hydroxyl group (FOH = 0.33).Thus, a 4R uoro group enforces the C-exoring pucker of a Pro derivative more stronglythan does a 4R hydroxyl group.

To probe further the role of Hyp in collagenstability, a (2S,4R)-4-methoxyproline residue(Mop) was incorporated into the Yaa positionof a (ProYaaGly)10 CRP (59). O-Methylationis perhaps the simplest possible covalent mod-ication of a Hyp residue and reduces the ex-tent of hydration without altering signicantlythe electron-withdrawing ability of the 4R sub-stituent. Accordingly, Mop and Hyp residueshave similar conformations (Table 4). Interest-ingly, reducing the hydration of (ProHypGly)10by methylation of Hyp residues enhances triple-helix stability signicantly (Table 2). Moreover,alkylation with functional groups larger thana methyl group does not necessarily perturbtriple-helix stability (60). Notably, (2S,4R)-4-chloroproline (Clp) residues also stabilize triplehelices in the Yaa position (Table 2) (61). LikeFlp, Clp has a strong preference for the C-exoring pucker, and a (ProClpGly)10 triple helixis therefore more stable than a (ProProGly)10triple helix. Thus, a plethora of data indicatethat the hydroxyl group of Hyp stabilizes col-lagen through a stereoelectronic effect. Waterbridges provide little (if any) net thermody-namic advantage to natural collagen (59).

Surprisingly, a host-guest CRP of theform AcGly(ProHypGly)3ProFlpGly(ProHypGly)4GlyNH2 actually forms a less stabletriple helix than does AcGly(ProHypGly)8GlyNH2 (62). In contrast, a host-guest CRPof the form (GlyProHyp)3GlyProFlpGlyValCysGlyAspLysGlyAsnProGlyTrpProGlyAlaPro(GlyProHyp)4-NH2 forms a morestable triple helix than one containing Hyprather than Flp (63). These results suggest thata uoro group might disrupt the hydrationinduced by a long string of Hyp residues.Kobayashi and coworkers (64) used differentialscanning calorimetry to demonstrate that(ProHypGly)10 triple helices are stabilized byenthalpy, whereas (ProFlpGly)10 triple helicesare stabilized by entropy. These ndings are

Stereoelectroniceffects

Gaucheeffect

C-exopucker

HighKtrans/cis

a

H

N HR1=EWG

HO

N

H

HR1=EWG

HOb

Gauche conformationC-exo pucker

c

Anti conformationC-endo pucker

NO

O

R1R2

n *interaction

Consequentpreorganization

Stabletriplehelix

d

Figure 8Stereoelectronic effects that stabilize the collagen triple helix. (a) A gaucheeffect and an n interaction preorganize main chain torsion angles andenhance triple-helix stability. (b) A gauche effect, elicited by an electron-withdrawing group (EWG) in the 4R position, stabilizes the C-exo ringpucker. (c) An n interaction stabilizes the trans isomer of the peptide bondbut is substantial only when Pro derivatives are in the C-exo ring pucker (e.g.,R1 = OH or F, R2 = H). (d ) Depiction of overlap between n and naturalbond orbitals (NBOView c) in a Pro residue with C-exo pucker.

consistent with Hyp decreasing the entropiccost for folding via main chain preorganizationbut increasing that cost by specic hydration.This interpretation is in accord with the sta-bility of (ProMopGly)10 triple helices arisingfrom a nearly equal contribution of enthalpyand entropy (59).

A steric effect. Electronegative substituentson Pro rings are not the only means of en-forcing an advantageous ring pucker. Pro ringpucker can also be dictated by steric effects,as in (2S,4S )-4-methylproline (Mep) (65) and(2S,4R)-4-mercaptoproline (Mpc) (Figure 7)(66). The 4-methyl substituent of Mep prefersthe pseudoequatorial orientation and thusenforces the C-exo ring pucker of Pro (analo-gous results are observed for Mpc) (Table 4).


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

Table 4 Conformation of 4-substituted derivatives of Pro that prefer the Yaa position [ = 60, = 152 (7)] in acollagen triple helix

Residue(References)

Crystal structure

Ringpucker

(EendoEexo)a (kcal/mol)

b(degrees)

b(degrees)

Mep (65) 1.7 62d

58

55

57

153d

148

141

7.4

Mop (59) 6.7

Flp (52, 54, 56, 57) 0.85

Hyp (46, 54, 57, 58) 151

148

6.1

Clp (61)

Mcp (66)

aFrom DFT calculations. bValues of and (here, NiCiCiOi+1) are from the crystal structure of Ac-Yaa-OMe.cValues of Ktrans/cis (Figure 5) were determined in solution by NMR spectroscopy. dM.D. Shoulders, I.A. Guzei & R.T. Raines, unpublished data.

Ktrans/cisc

C-exo

C-exo

C-exo

C-exo

C-exo

C-exo

5.4

5.4

56

6.7

Indeed, triple helices formed from(ProMepGly)7 have stability similar tothose formed from (ProHypGly)7 (Table 2)(65).

Proline Derivatives in the Xaa Position

The C-exo ring pucker of Pro residues inthe Yaa position enhances triple-helix stability.Likewise, the ring pucker of Pro in the Xaa posi-tion is important for triple-helix stability. Typi-cally, Pro residues in the Xaa position of biolog-ical collagen are not hydroxylated and usuallydisplay the C-endo ring pucker (67). By em-ploying C-substituents, both the gauche effectand steric effects can be availed to preorganize

the C-endo ring pucker (Figures 7 and 8). In-stallation of p, (2S,4S )-4-chloroproline (clp),or (2S,4R)-4-methylproline (mep) residues (allof which prefer the C-endo ring pucker)(Table 3) in the Xaa position of collagen isstabilizing relative to Pro, but installation ofFlp, Clp, or Hyp (which prefer the C-exo ringpucker) is destabilizing (Table 2) (61, 65, 6870). These results suggest that preorganizingthe C-endo ring pucker in the Xaa position ofCRPs stabilizes triple helices. This conclusionis reasonable because Pro derivatives with a C-endo ring pucker have and main chain tor-sion angles similar to those observed in the Xaaposition of triple helices (Table 3).


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

Notably, replacing Pro in the Xaa position of(ProProGly)10 with hyp, a Pro derivative that,like p and clp, should prefer the C-endo ringpucker owing to the gauche effect, yields CRPsthat do not form triple helices (Table 2) (47).This anomalous result for hyp in the Xaa posi-tion could be attributable to deleterious hydra-tion, idiosyncratic conformational preferencesof hyp residues, or both (71).

Type IV collagen, which is the primary com-ponent of basement membranes, has a high in-cidence of (2S,3S )-3-hydroxyproline (3S-Hyp)in the Xaa position (72). This modication ispresent in some other collagen types and in in-vertebrate collagens. 3S-Hyp, which prefers aC-endo ring pucker (73), is introduced almostexclusively within ProHypGly triplets via post-translational modication of individual colla-gen strands by prolyl 3-hydroxylase (P3H),which is distinct from P4H (74). A recessiveform of OI is associated with a P3H de-ciency (75, 76). Certain mutations to the geneencoding cartilage-associated protein, a P3H-helper protein, prevent 3S-hydroxylation of1(I)Pro986 as well as 3S-hydroxylation ofsome other Xaa-position Pro residues, result-ing in a phenotype nearly identical to clas-sical OI. The underlying basis for the im-portance of 3S-hydroxylation of 1(I)Pro986is unclear but could involve lower rates oftriple-helix secretion (76). Replacing Pro with3S-Hyp in the Xaa position of CRPs can en-hance triple-helix stability slightly (73, 77). Acrystal structure of a triple helix containing3S-Hyp substitutions reveals the maintenanceof the prototypical triple-helix structure andthe absence of unfavorable steric interactions(Figure 4c) (78). In contrast, replacing 3S-Hyp with (2S,3S )-3-uoroproline destabilizesa triple helix markedly, possibly owing to athrough-bond inductive effect that diminishesthe ability of its main chain oxygen to accept ahydrogen bond (Figure 4d ) (79).

An n InteractionA general principle in the design of CRPsis that Pro residues with either a C-endo or

C-exo ring pucker will stabilize triple he-lices in the Xaa and Yaa positions, respectively(Tables 24). Appropriate ring pucker, en-forced by a stereoelectronic or steric effect, pre-organizes the and torsion angles to thoserequired for triple-helix formation.

Intriguingly, the stability of a (pProGly)7or (clpProGly)10 triple helix is signicantly lessthan that of a (ProFlpGly)7 or (ProClpGly)10triple helix, respectively (Table 2) (61, 68).Likewise, a (mepProGly)7 triple helix isless stable than a (ProMepGly)7 triple helix(Table 2) (65). Two factors contribute to thelower stability of triple helices formed fromCRPs with stabilizing Pro derivatives substi-tuted in the Xaa position rather than the Yaaposition. First, a C-endo ring pucker is alreadyfavored in Pro (56); p, clp, and mep merelyenhance that preference (Table 3). In contrast,Flp, Clp, Hyp, and Mep have the more dramaticeffect of reversing the preferred ring pucker ofPro, thereby alleviating the entropic penaltyfor triple-helix formation to a greater extent(Table 4). Second, Flp, Clp, and Mep in theYaa position cause favorable preorganization ofall three main chain torsion angles (, , and) (Table 4). In contrast, p, clp, and mep havea low probability of adopting a trans peptidebond ( = 180) (54, 61, 65) relative to Pro(Table 3), thereby mitigating the benet ac-crued from proper preorganization of and .Notably, 13C-NMR studies on collagen in vitroshow that 16% of GlyPro bonds in unfoldedcollagen are in the cis conformation, whereasonly 8% of XaaHyp bonds in unfolded col-lagen are cis, an observation that conrms theeffect of C-substitution on the conformationof the preceding peptide bond (80).

How does the effect of a 4-X substituenton Pro ring pucker inuence the peptide bondisomerization equilibrium constant (Ktrans/cis)(Figure 5 and Tables 3 and 4)? The ex-planation stems from another stereoelectroniceffect: an n interaction (56, 81). In ann interaction, the oxygen of a peptide bond(Oi1) donates electron density from its lonepairs into the antibonding orbital of the car-bonyl in the subsequent peptide bond (Ci ==Oi)


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

(Figure 8c,d ). The C-exo ring pucker ofa Pro residue provides a more favorableOi1. . .Ci ==Oi distance and angle for an ninteraction than does the C-endo pucker (56).Importantly, Ktrans/cis for peptidyl prolyl amidebonds is determined by the pyrrolidine ringpucker and is not generally affected by theidentity of substituents in the 4-position of thepyrrolidine ring (82). Because an n interac-tion can occur only if the peptide bond contain-ing Oi1 is trans, the n interaction has animpact on the value of Ktrans/cis for main chainswith appropriate torsion angles (Table 4).Thus, imposing a C-exo pucker on a pyrroli-dine ring in the Yaa position of a CRP preorga-nizes not only the and angles for triple-helixformation, but also the angle. Indeed, a singlen interaction can stabilize the trans confor-mation by G = 0.7 kcal/mol (81, 83).

Hyp in the Xaa Position

In the Xaa position, a Pro residue with a C-endo pucker generally stabilizes a triple helix,whereas one with a C-exo pucker destabilizes atriple helix. For example, (HypProGly)n triplehelices are far less stable than (ProProGly)ntriple helices (Table 2) (84) because Hypprefers the C-exo ring pucker and thus preor-ganizes the and torsion angles improperlyfor the Xaa position of a collagen triple helix(Table 4). Surprisingly, (HypHypGly)10 triplehelices are actually slightly more stable than(ProHypGly)10 triple helices (Table 2) (85, 86)despite the Hyp residues in the Xaa positionof (HypHypGly)10 displaying the C-exo ringpucker in the triple helix (87, 88). It is notewor-thy that crystal structures of (HypHypGly)10show that the main chain torsion angles in theXaa position of a (HypHypGly)n triple helix ad-just to accommodate a C-exo ring pucker inthat position (87, 88).

The nding that Hyp can stabilize triple he-lices in the Xaa position in a context-dependentmanner was presaged in a study by Gruskinand coworkers (89) on the global substitutionof Hyp for Pro in recombinant type I col-lagen polypeptides that formed stable triple

helices. Notably, Hyp is found in the Xaaposition of some invertebrate collagens (90)and can be acceptable in CRPs in which theYaa position residue is not Pro (86, 91, 92).Berisio and coworkers (93) have suggested that(HypHypGly)10 triple helices might be hyper-stable owing to interstrand dipole-dipole in-teractions between proximal COH bonds ofadjacent Hyp residues. Kobayashi and cowork-ers (87) have proposed that the stability of(HypHypGly)10 triple helices is attributable tothe high hydration level of the peptide chainsin the single-coil state prior to triple-helix for-mation, which could reduce the entropic costof water bridge formation. A combination ofthese factors is likely to be responsible for thisanomaly.

Heterotrimeric SyntheticTriple Helices

Both p and Flp greatly enhance triple-helixstability when in the Xaa and Yaa position,respectively. Nonetheless, (pFlpGly)n formsmuch less stable triple helices than does(ProProGly)n (Table 2) (79, 94). In such ahelix, the uorine atoms of p and Flp residuesin alternating strands would be proximal, andthe CF dipoles would interact unfavorably(Figure 9a) (79). These negative steric andelectronic interactions presumably compro-mise triple-helix stability despite appropriatepreorganization of main chain torsion angles.This hypothesis was conrmed by two otherndings. First, a (clpClpGly)10 triple helix doesnot even form at 4C, whereas a (pFlpGly)10triple helix has Tm = 30C (Table 2) (61, 94).The steric clash between chlorine atoms ofopposing clp and Clp residues is exacerbatedby the large size of chlorine relative to uo-rine (Figure 9b). Second, (mepMepGly)7forms more stable triple helices than doeither of the corresponding mono-substitutedCRPs, (mepProGly)7 and (ProMepGly)7(Table 2). The 4-methyl groups protrude radi-ally from the triple helix (Figure 9c) and thuscannot interact detrimentally with each other(65).


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

a b c

d

flpFlp

clpClp

mep Mep

Unstable

Stable

(ProHypGly)10: neutral strand

(AspHypGly)10: anionic strand

(ProLysGly)10: cationic strand

+

+

Figure 9Heterotrimeric synthetic collagen triple helices. (ac) Steric approach. Space-lling models of triple-helixsegments constructed from the structure of a (ProHypGly)n triple helix [PDB entry 1cag (19)] with theprogram SYBYL (Tripos, St. Louis, MO). In panel a, rFF = 2.4 A in a (pFlpGly)n triple helix (79). In panelb, rClCl = 1.9 A (61) in a (clpClpGly)n triple helix. In panel c, the methyl groups in a (mepMepGly)n triplehelix are radial and distal. (d ) Coulombic approach. Favorable Coulombic interactions drive the preferentialassembly of triple helices having a 1:1:1 ratio of (ProLysGly)10:(AspHypGly)10:(ProHypGly)10 (96).

The steric and stereoelectronic effectson triple-helix stability manifested in the(pFlpGly)7 CRP provided, for the rst time, ameans to generate noncovalently linked, het-erotrimeric triple helices with dened stoi-chiometry. Analysis of triple-helix cross sec-tions suggested a triple helix composed of(pFlpGly)7:(ProProGly)7 in either a 1:2 or2:1 ratio could be stable, as the presenceof some Pro residues in the Xaa and Yaapositions would eliminate deleterious stericinteractions between uorine residues in op-posing strands. A (pFlpGly)7:(ProProGly)7ratio of 2:1 yielded the most stable triple he-lices, thereby demonstrating the rst instance of

heterotrimeric assembly of triple helices withcontrolled stoichiometry (79) and suggestingthe possibility of developing a code for triple-helix assembly along the lines of the Watson-Crick code for DNA assembly.

Gauba & Hartgerink (95) developedan alternative strategy that employsCoulombic interactions to guide the as-sembly of heterotrimeric triple helices.They observed that a 1:1:1 mixture of(ProArgGly)10:(GluHypGly)10:(ProHypGly)10produces triple helices containing one neg-atively charged, one positively charged, andone neutral CRP. Intriguingly, a (ProLysGly)10:(AspHypGly)10:(ProHypGly)10 triple helix has


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

Self-assembly: aprocess in whichspecic, localinteractions betweendisorderedcomponents lead to anorganized structure,without externaldirection

Collagenfibrillogenesis: theprocess oftropocollagenmonomers assemblinginto mature brils

a Tm value similar to that of a (ProHypGly)10homotrimer, even though Asp and Lys areknown to destabilize signicantly the triplehelix relative to Pro and Hyp (Figure 9d ). Thisnding demonstrates the utility of Coulom-bic interactions for stabilizing triple helices(96).

Synthetic collagen heterotrimers are appeal-ing mimics of natural collagen strands, as mostcollagen types are themselves heterotrimers(Table 1). Gauba & Hartgerink (97) employedtheir Coulombic approach to generate mim-ics of type I collagen variants that lead to OI.Specically, they studied the stability of triple-helical heterotrimers containing one, two, orthree GlySer substitutions. They observedthat a GlySer substitution in only one or twochains is not as debilitating for triple-helix sta-bility and folding as is a GlySer substitutionin all three chains.

AVOIDING AGGREGATION

Long, unfolded polypeptides have an innate tendency to formaggregates (145), such as the amyloid brils implicated in neu-rodegenerative diseases. Interestingly, despite their long lengthand slow folding, protocollagen strands are not known to aggre-gate. Amyloid brils and other aggregates are composed largelyof -sheets (146). Pro and Gly are the two amino acid residueswith the lowest propensity to form a -sheet (147, 148), and Glyresidues are known explicitly to reduce protein aggregation rates(149).

We propose that the prevalence of Pro and Gly residues inprotocollagen is necessary to avert the formation of harmful ag-gregates. This proposal is supported by the remarkably highPro/Gly content of other brous, structural proteins in plantsand animals, such as elastin, extensin, glycine-rich proteins, andproline-rich proteins. Molecular dynamics simulations of elastinpolypeptides likewise support this proposal, as a minimum thresh-old of Pro/Gly content must be attained to realize elastomers in-stead of amyloid brils (150). Apparently, the molecular evolutionof collagen and other brous, structural proteins has availed Proand Gly residues to avoid -sheet formation and the consequentformation of harmful aggregates.

Nonproline Substitutions in the Xaaand Yaa Positions

Brodsky and coworkers (9) determined thefrequency of occurrence of all possible tripep-tides in a set of brillar and nonbrillar col-lagen sequences. Only a few of the 400 possi-ble triplets formed from the 20 natural aminoacids are observed with any frequency in col-lagen. Additionally, they have examined ex-haustively the incorporation of all 20 commonamino acids in the Xaa and Yaa positions ofCRPs using a host-guest model system whereina single XaaYaaGly triplet is placed withina (ProProGly)n or (ProHypGly)n CRP (98).These host-guest studies revealed a correlationbetween the propensity of a particular residue toadopt a PPII conformation and its contributionto triple-helix stability (98). Notably, Arg in theYaa position confers triple-helix stability similarto Hyp (99). The aromatic amino acid residuesTrp, Phe, and Tyr are all strongly destabilizingto the triple helix (98), although the structuralbasis for this destabilization is unclear. Brodskyand coworkers (100) used their data on host-guest CRPs to develop an algorithm that en-ables a priori calculation of the effect of Xaaand Yaa substitutions on triple-helix stability.

HIGHER-ORDER COLLAGENSTRUCTURE

In vivo collagen has a hierarchical struc-ture (Figure 2). Individual TC monomersself-assemble into the macromolecular bersthat are essential components of tissues andbones. The self-assembly processes involved incollagen brillogenesis are of enormous impor-tance to ECM pathology and proper animal de-velopment (see the sidebar Avoiding Aggre-gation for a discussion of how collagen self-assembly might be directed away from delete-rious protein aggregates).

Fibril Structure

There are many classes of collagenous struc-tures in the ECM, including brils, networks,


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

and transmembrane collagenous domains. Forbrevity, we focus here on brils composed pri-marily of type I collagen.

TC monomers of type I collagen have theunique property of actually being unstable atbody temperature (101); that is, the random coilconformation is the preferred one. How can sta-ble tissue structures form from an unstable pro-tein? The answer must be that collagen brillo-genesis has a stabilizing effect on triple helices.Moreover, the assembly of strong macromolec-ular structures is essential to enable collagen tosupport stress in one, two, and three dimensions(102). The importance of collagen brillogen-esis is underscored by the conclusion of Kadlerand coworkers (103) that the fundamental prin-ciples underlying the formation of some typesof modern collagen brils were established atleast 500 Mya.

Collagen brillogenesis in situ occurs via as-sembly of intermediate-sized bril segments,called microbrils (Figure 2) (104). Thus, thereare two important issues for understandingthe molecular structure of the collagen bril.First, what is the arrangement of individual TCmonomers within the microbril? Second, whatis the arrangement of the individual microb-rils within the collagen bril? These questionshave proven difcult to answer, as individualnatural microbrils are not isolable and thelarge size and insolubility of mature collagenbrils prevent the use of standard structure-determination techniques.

Collagen brils formed mainly from typeI collagen (all brous tissues except cartilage)and brils formed largely from type II collagen(cartilage) have slightly different structures. Al-though we focus solely on type I collagen b-rils, recent data have enabled the determinationof thin cartilage bril structure to intermedi-ate resolution (4 nm). This structure suggeststhat cartilage collagen brils have a 10 + 4 het-erotypic microbril structuremeaning thatthe bril surface presents ten equally spaced mi-crobrils and that there are four equally spacedmicrobrils in the core of the bril (105).

Fibrils of type I collagen in tendon are upto 1 cm in length (106) and up to 500 nm

D-Periodicity: theaxial stagger ofadjacent tropocollagenmolecules by adistance, D, which isthe sum of gap andoverlap regions

in diameter. An individual triple helix in typeI collagen is

ANRV378-BI78-32 ARI 5 May 2009 15:11

Collagentelopeptides: N- andC-terminal 11- to26-residuenontriple-helicaldomains oftropocollagen strandsinvolved inbrillogenesis andcross-linking

Procollagen: thehydroxylated form ofcollagen prior tocollagen propeptidecleavage

Collagenpropeptides: N- andC-terminalnontriple-helicaldomains of collagenstrands that directtriple-helix foldingprior to brillogenesis

right-handed microbrils that interdigitatewith neighboring microbrilsleading to aspiral-like structure for the mature collagen b-ril (113). Their model advances the provoca-tive idea that the collagen bril is a networked,nanoscale ropean idea also suggested by theAFM studies of Bozec and coworkers (111).

Orgel and coworkers determined the axiallocation of the N- and C-terminal collagentelopeptides and found that neighboringtelopeptides within a TC monomer interactwith each other and are cross-linked covalentlysubsequent to the action of lysyl oxidase(114). The cross-links can be both within andbetween microbers. Intriguingly, the super-twisted nature of the collagen microbril ismaintained through the nonhelical telopeptideregions (113).

This new model of the bril of type I colla-gen explains the failure of previous researchersto isolate individual collagen microbrils fromtissue samples: The microbrils interdigitateand cross-link, thus preventing separation fromeach other in an intact form. The new modelalso justies the observation that TC in b-rils is far more resistant to collagen proteolysisby matrix metalloproteinase 1 (MMP1) than ismonomeric TC; the collagen bril protects re-gions vulnerable to proteolysis by MMP1. Pro-teolysis of the C-terminal telopeptide of TCin a bril is required before MMP1 can gainaccess to the cleavage site of a TC monomer(115).

Nucleation and Modulationof Collagen Fibrillogenesis

Collagen brillogenesis requires completionof two stages of self-assembly: nucleation andber growth. Collagen brillogenesis beginsonly after procollagen N- and C-proteinasescleave the collagen propeptides at each triple-helix terminus to generate TC monomers. TheC-terminal propeptides are essential for propertriple-helix formation but prevent brilloge-nesis (116). After cleavage of the propeptides,TC monomers are composed of a lengthytriple-helical domain consisting of a repeating

XaaYaaGly sequence anked by short,nontriple-helical telopeptides (Figure 2).

The C-terminal telopeptides of TC areimportant for initiating proper brillogenesis.Prockop and Fertala (117) suggested that colla-gen self-assembly into brils is driven by the in-teraction of C-terminal telopeptides with spe-cic binding sites on triple-helical monomers.The addition of synthetic telopeptide mimicscan inhibit collagen brillogenesis, presumablyby preventing the interaction between collagentelopeptides and TC monomers. Triple heliceslacking the telopeptides can, however, assem-ble into brils with proper morphology (118).Thus, collagen telopeptides could acceleratebril assembly and establish the proper regis-ter within microbrils and brils but might notbe essential for brillogenesis.

Collagen telopeptides have a second rolein stabilizing mature collagen brils. Lys sidechains in the telopeptides are cross-linked sub-sequent to bril assembly, forminghydroxylysylpyridinoline andlysyl pyridinoline cross-linksbetween Lys and hydroxylysine residues withthe aid of oxidase (Figure 2) (119). The cross-linking process endows mature collagen brilswith strength and stability, but is not involved infibrillogenesis. Thus, although collagen telopep-tides might not be essential for nucleatingcollagen brillogenesis, their absence greatlyweakens the mature bril owing to the lack ofcross-links within and between triple helices(119).

MECHANICAL PROPERTIESOF COLLAGEN FIBRILS

The hierarchical nature of collagen structuretheoretically enables evaluation of the me-chanical properties of collagen at varying lev-els of structural complexity, including the TCmonomer, individual collagen brils, and col-lagen bers. Perhaps the most direct measuresof the mechanical properties of collagen havebeen obtained by studying TC monomers andbrils formed from type I collagen. Researchershave employed various biophysical and theo-retical techniques over the past 20 years, and


*Erratum

*This PDF amended on (21 Apr. 2010): See explanation at http://arjournals.annualreviews.org/errata/biochem

Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

recent advances in AFM methodology have en-abled more rened evaluations.

In 2006, Buehler estimated the fracturestrength of a TC monomer to be 11 GPa, whichis signicantly greater than that of a collagenbril (0.5 GPa) (102). This difference is rea-sonable, given that fracture of a TC monomerrequires unraveling of the triple helix and ul-timately breaking of covalent bonds, whereasfracture of a bril does not necessarily requirethe disruption of covalent bonds. For compar-ison, the tensile strength of collagen in tendonis estimated to be 100 MPa (120).

The Youngs modulus of a TC monomer isE = 67 GPa (102, 121), whereas AFM mea-surements show that dehydrated brils of typeI collagen from bovine Achilles tendon (122)and rat tail tendon (123) have E 5 GPa andE 11 GPa, respectively. Because collagen b-rils are anisotropic, the shear modulus (which isa measure of rigidity) is also an important mea-sure of the strength of a collagen bril. In 2008,AFM revealed that dehydrated brils of typeI collagen from bovine Achilles tendon haveG = 33 MPa (124). Hydration of these b-rils reduced their shear modulus signicantly,whereas carbodiimide-mediated cross-linkingincreased their shear modulus. It is notewor-thy that a certain level of cross-linking is favor-able for the mechanical properties of collagenbrils, but excessive cross-linking results in ex-tremely brittle collagen brils (102), a commonsymptom of aging.

An analysis by Buehler (102) of the me-chanical properties of collagen brils sug-gests that nature has selected a length for theTC monomer that maximizes the robustnessof the assembled collagen bril via efcientenergy dissipation. Simulations indicate thatTC monomers either longer or shorter than300 nm (which is the length of a type I col-lagen triple helix) would form collagen brilswith less favorable mechanical properties.

COLLAGENOUS BIOMATERIALS

Research on the structure and stability of colla-gen triple helices has focused on blunt-ended

triple helices composed of (XaaYaaGly)n10CRPs. These short triple helices, althoughvaluable for studies directed at understandingthe physicochemical basis of triple-helix struc-ture and stability, are not useful for many po-tential biomaterial applications because of theirsmall size, which does not approach the scale ofnatural collagen bers (Figure 2).

Bovine collagen is readily available and use-ful for some biomedical purposes, but it suffersfrom heterogeneity, potential immunogenic-ity, and loss of structural integrity during theisolation process. An efcient recombinant orsynthetic source of collagen could avoid thesecomplications. The heterologous productionof collagen is made problematic by the dif-culty of incorporating posttranslational modi-cations, such as that leading to the essentialHyp residues (Figure 6), and by the needto use complex expression systems (125).These challenges underscore the need for syn-thetic sources of collagen-like proteins andbrils.

Collagen via Chemical Synthesis

Early approaches to long synthetic collagentriple helices relied on the condensation (126,127) or native chemical ligation of shortCRPs (127). Interestingly, concentrated aque-ous solutions of (ProHypGly)10 self-assembleinto highly branched brils (128). Brodskyand coworkers (129) have shown that therate of (ProHypGly)10 self-assembly and themorphology of the resultant brils are se-quence dependent. CRPs containing a singleProAla or ProLeu substitution displayslower self-assembly; bril morphology can bemodied by a GlySer substitution, or pre-vented by a single GlyAla substitution orglobal HypPro substitutions. Regardless, thehigher-order structures formed by the self-assembly of (ProHypGly)10 and related CRPsdo not resemble natural collagen brils.

Long collagen triple helices have been pre-pared by using a design that takes advantageof the intrinsic propensity of individual CRPstrands to form triple helices. Specically, a


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

cystine knot within short collagen fragmentswas utilized to set the register of individualcollagen strands such that short, sticky endspreorganized for further triple-helix formationwere displayed at the end of each triple-helical,monomeric segment (Figure 10a) (130, 131).Self-assembly of these short, triple-helix frag-ments was then mediated by association of thesticky ends, resulting in collagen assemblies aslong as 400 nmsignicantly longer than nat-ural TC monomers (131). Koide and cowork-ers (132) used this system to prepare tunablecollagen-like gels with potential biomaterialapplications.

Maryanoff and coworkers (133) developedanother approach to long triple helices, onethat relied on the predilection of electron-richphenyl rings of C-terminal phenylalanineresidues installed in a short CRP to engagein -stacking interactions preferentially withelectron-poor pentauorophenyl rings ofN-terminal pentauorophenylalanine residues(Figure 10b). Their strategy producedmicrometer-scale triple-helical bers. This-stacking approach has been used to gen-erate thrombogenic collagen-like brils forapplications in biomedicine (134). In addition,attachment of gold nanoparticles to thesebrils and subsequent electroless silver platingyielded collagen-based nanowires that conductelectricity (135).

Przybyla & Chmielewski (136) used metal-triggered self-assembly to obtain collagen b-rils from a CRP. A single Hyp residue inAc-(ProHypGly)9-NH2 was replaced with abipyridyl-modied Lys residue. Addition ofFe(II) to a solution of this CRP triggered self-assembly into morphologically diverse brils ofup to 5 m in length with a mean radius of0.5 m.

A major advance in the development of syn-thetic CRP assemblies with improved similarityto collagen brils was reported by Chaikof andcoworkers (137). They synthesized a CRP withthe sequence (ProArgGly)4(ProHypGly)4(GluHypGly)4 and observed self-assembly insolution into brils 34 m in length and 1215 nm in diameter. Upon heating the peptide

solution to 75C for 40 min and then cooling toroom temperature, they observed thicker b-rils (70 nm in diameter). Importantly, thesebrils exhibited two key characteristics of natu-ral collagen brils. First, the brils displayedtapered tips at their terminia feature ob-served in type I collagen bers and thought tobe important for ber growth (138). Second,Chaikof and coworkers observed D-periodicstructure in synthetic collagen brils, with D 18 nm. The self-assembly process presumablyrelies on Coulombic interactions and hydrogenbonds between charged Arg and Glu residuesin individual, axially staggered triple helices(Figure 10c).

The methodologies described above enablethe creation of long, triple-helical, collagen-like brils. Despite major advances, syntheticcollagen-mimetic brils still lack many of thecharacteristics of higher-order collagen struc-tures. In addition, the mechanical properties ofsynthetic collagenous materials have not beenstudied to date. Synthetic collagens that closelymimic the length, girth, patterns, mechanicalproperties, and complexity of natural collagenbrils remain to be developed, but rapidprogress in the past few years engenders greatoptimism.

Biological and BiomedicalApplications of Synthetic Collagen

Relatively few CRPs have been tested as bioma-terials. Goodman and coworkers (139) showedthat peptoid-containing CRPs have a notableability to bind to epithelial cells and broblasts,particularly when displayed on a surface. CRPsare also useful for inducing platelet aggrega-tion, which can aid the wound-healing process(140, 141).

A key step toward utilizing collagenous bio-materials for therapeutic purposes is the de-velopment of CRPs that can either adhere toor bury themselves within biological collagen.Most efforts toward these objectives have re-lied on immobilization of CRPs on an unre-lated substance. Yu and coworkers (142) pre-pared CRP-functionalized gold nanoparticles


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

b

a

c

H2N

O

HN

O

NH O

HN

H2NNH2+

NH O

N

O

HN

O

O

O

O

4

|

SS

|

|

SS

|

(ProHypGly)5 (ProHypGly)3

(ProHypGly)3

(ProHypGly)3(ProHypGly)5

(ProHypGly)5

CysGly

CysGly

GlyCysCysGlySelf-assembly

Self-assembly

+

67 nm

D 17.9 nm

Self-assembly

Self-assembly

N

O

HN

O

NH O

HN

H2NNH2+

3

(ProHypGly)4

NH O

N

O

N

O

HN

O

H

H

HH

H

OO

F

F

FF

F

H3N+

OH

10

N CN C

N C

OH

Figure 10Strategies for the self-assembly of long, synthetic collagen triple helices and brils. (a) Disulde bondsenforce a strand register with sticky ends that self-assemble (131). (b) Stacking interactions betweenelectron-poor pentauorophenyl rings and electron-rich phenyl rings lead to self-assembly (133, 134).(c) Coulombic forces between cationic and anionic blocks encourage self-assembly. TEM image of aresulting ber shows D-periodicity with D = 17.9 nm (137). Natural type I collagen has D = 67 nm.


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

and demonstrated binding of the gold nanopar-ticles to the gap region of natural collagen.Maryanoff and coworkers found that CRPsdisplayed on latex nanoparticles can stimu-late human platelet aggregation with a potencysimilar to that of type I collagen (140). In animportant extension of this work, they demon-strated that triple-helical brils obtained viaaromatic interactions had a similar level of

thrombogenic activity to the CRPs immobi-lized on latex nanoparticles (134). Finally, sin-gle strands of CRPs and polyethylene glycol-conjugated CRPs bind to collagen lms evenwithout immobilization on nanoparticles (143)and are of potential use in collagen imaging(144) and wound-healing applications. The fu-ture of these approaches appears to be especiallybright.

SUMMARY POINTS

1. High-resolution crystal structures and modern biophysical approaches have enabled de-tailed study of the structure and stability of collagen triple helices. The ladder of hydrogenbonds observed in these crystal structures is essential for holding the triple helix together,and its absence in natural collagen leads to a variety of pathological conditions.

2. Stereoelectronic effects impart signicant structural stability to collagen by preorganizingindividual polypeptides for triple-helix formation. For example, Hyp in the Yaa positionstabilizes the triple helix via a stereoelectronic effect. Stereoelectronic effects are alsoimportant for the structure and stability of numerous other peptides and proteins.

3. Posttranslational modications to protocollagen are of fundamental importance to thesynthesis of a stable ECM. These modications include hydroxylation and cross-linkingreactions.

4. Collagen brillogenesis is an essential process for the formation of macromolecularbiological scaffolds. Relatively high-resolution models of type I and type II collagen b-rils are now available and, for type I collagen, show that collagen brils can be describedas nanoscale ropes.

5. Simple means to synthesize long collagen triple helices and brils have become apparent.The resultant materials are poised for use in biomedicine and nanotechnology.

FUTURE ISSUES

1. The factors that affect triple-helix stability for Pro derivatives in the Yaa position are nowclear. In comparison, the Xaa position is understood only poorly. What, for example, isthe physicochemical basis for the anomalous effects of hyp and Hyp on triple-helixstability in the Xaa position?

2. The current understanding of triple-helix structure and stability derives from analysesof triple-helical CRPs. Do these analyses provide insight on the stability and mechanicalproperties of natural collagen brils?

3. What functionalities in natural collagen are important for proper bril formation inthe ECM? How might diseases stemming from improper bril formation be subject totherapeutic intervention?


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

4. Can improved methods be developed to synthesize long collagen triple helices and rele-vant mimics of complex, hierarchical collagen assemblies?

5. What are the molecular structures of nonbrillar collagen assemblies? How are thoseassemblies formed in vivo?

6. Natural collagens appear to engage many other proteins and biomolecules. Which ones?How? Can those interactions be manipulated to treat disease?

7. How can synthetic collagen-based biomaterials lead to expeditious therapies?

DISCLOSURE STATEMENT

The authors are not aware of any afliations, memberships, funding, or nancial holdings thatmight be perceived as affecting the objectivity of this review.

ACKNOWLEDGMENTS

The authors acknowledge Dr. Jeet Kalia for critical reading of the manuscript and Amit Choudharyfor creating Figure 8d. M.D.S. was supported by graduate fellowships from the Departmentof Homeland Security and the Division of Medicinal Chemistry, American Chemical Society.Collagen research in our laboratory is supported by Grant AR044276 (NIH).

LITERATURE CITED

1. Brinckmann J. 2005. Collagens at a glance. Top. Curr. Chem. 247:162. Veit G, Kobbe B, Keene DR, Paulsson M, Koch M, Wagener R. 2006. Collagen XXVIII, a novel von

Willebrand factor A domain-containing protein with many imperfections in the collagenous domain.J. Biol. Chem. 281:3494504

3. Schweitzer MH, Suo Z, Avci R, Asara JM, Allen MA, et al. 2007. Analyses of soft tissue from Tyrannosaurusrex suggest the presence of protein. Science 316:27780

4. Asara JM, Schweitzer MH, Freimark LM, Phillips M, Cantley LC. 2007. Protein sequences frommastodon and Tyrannosaurus rex revealed by mass spectrometry. Science 316:28085

5. Buckley M, Walker A, Ho SYW, Yang Y, Smith C, et al. 2008. Comment on Protein sequences frommastodon and Tyrannosaurus rex revealed by mass spectrometry. Science 319:33

6. Pevzner PA, Kim S, Ng J. 2008. Comment on Protein sequences from mastodon and Tyrannosaurus rexrevealed by mass spectrometry. Science 321:1040

7. Berisio R, Vitagliano L, Mazzarella L, Zagari A. 2002. Crystal structure of the collagen triple helix model[(Pro-Pro-Gly)10]3. Protein Sci. 11:26270

8. Brazel D, Oberbaumer I, Dieringer H, Babel W, Glanville RW, et al. 1987. Completion of the aminoacid sequence of the 1 chain of human basement membrane collagen (type IV) reveals 21 nontripletinterruptions located within the collagenous domain. Eur. J. Biochem. 168:52936

9. Ramshaw JAM, Shah NK, Brodsky B. 1998. Gly-X-Y tripeptide frequencies in collagen: a context forhost-guest triple-helical peptides. J. Struct. Biol. 122:8691

10. Fitzgerald J, Rich C, Zhou FH, Hansen U. 2008. Three novel collagen VI chains, 4(VI), 5(VI), and6(VI). J. Biol. Chem. 283:2017080

11. Astbury WT, Bell FO. 1940. The molecular structure of the bers of the collagen group. Nature 145:42122

12. Pauling L, Corey RB. 1951. The structure of brous proteins of the collagen-gelatin group. Proc. Natl.Acad. Sci. USA 37:27281


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

13. Ramachandran GN, Kartha G. 1954. Structure of collagen. Nature 174:2697014. Ramachandran GN, Kartha G. 1955. Structure of collagen. Nature 176:5939515. Rich A, Crick FHC. 1955. The structure of collagen. Nature 176:9151616. Rich A, Crick FHC. 1961. The molecular structure of collagen. J. Mol. Biol. 3:48350617. Cowan PM, McGavin S, North ACT. 1955. The polypeptide chain conguration of collagen. Nature

176:10626418. Fields GB, Prockop DJ. 1996. Perspectives on the synthesis and application of triple-helical, collagen-

model peptides. Biopolymers 40:3455719. First high-resolution (1.9-A)crystal structure of acollagen triple helix,formed from CRPs.

19. Bella J, Eaton M, Brodsky B, Berman HM. 1994. Crystal and molecular structure of a collagen-like peptide at 1.9 A resolution. Science 266:7581

20. Bella J, Berman HM. 1996. Crystallographic evidence for CHO==C hydrogen bonds in a collagentriple helix. J. Mol. Biol. 264:73442

21. Kramer RZ, Venugopal MG, Bella J, Mayville P, Brodsky B, Berman HM. 2000. Staggered molecularpacking in crystals of a collagen-like peptide with a single charged pair. J. Mol. Biol. 301:1191205

22. Emsley J, Knight CG, Farndale RW, Barnes MJ, Liddington RC. 2000. Structural basis of collagenrecognition by integrin 21. Cell 101:4756

23. Cohen C, Bear RS. 1953. Helical polypeptide chain conguration in collagen. J. Am. Chem. Soc. 75:278384

24. Okuyama K, Xu X, Iguchi M, Noguchi K. 2006. Revision of collagen molecular structure. Biopolymers84:18191

25. Kramer RZ, Bella J, Mayville P, Brodsky B, Berman HM. 1999. Sequence dependent conformationalvariations of collagen triple-helical structure. Nat. Struct. Biol. 6:45457

26. Boudko S, Engel J, Okuyama K, Mizuno K, Bachinger HP, Schumacher MA. 2008. Crystal structure ofhuman type III collagen G991-G1032 cystine knot-containing peptide shows both 7/2 and 10/3 triplehelical symmetries. J. Biol. Chem. 283:3258089

27. Sweeney SM, Guy CA, Fields GB, San Antonio JD. 1998. Dening the domains of type I collageninvolved in heparin-binding and endothelial tube formation. Proc. Natl. Acad. Sci. USA 95:727580

28. Di Lullo GA, Sweeney SM, Korkko J, Ala-Kokko L, San Antonio JD. 2002. Mapping the ligand-bindingsites and disease-associated mutations on the most abundant protein in the human, type I collagen.J. Biol. Chem. 277:422331

29. Thorough analysisof the cell interactionand matrix interactiondomains of the collagenfibrils.

29. Sweeney SM, Orgel JP, Fertala A, McAuliffe JD, Turner KR, et al. 2008. Candidate cell andmatrix interaction domains on the collagen fibril, the predominant protein of vertebrates. J. Biol.Chem. 283:2118797

30. Jenkins CL, Vasbinder MM, Miller SJ, Raines RT. 2005. Peptide bond isosteres: ester or (E )-alkene inthe backbone of the collagen triple helix. Org. Lett. 7:261922

31. Boryskina OP, Bolbukh TV, Semenov MA, Gasan AI, Maleev VY. 2007. Energies of peptide-peptide andpeptide-water hydrogen bonds in collagen: evidences from infrared spectroscopy, quartz piezogravimetry,and differential scanning calorimetry. J. Mol. Struct. 827:110

32. Myllyharju J, Kivirikko KI. 2001. Collagens and collagen-related diseases. Ann. Med. 33:72133. Beck K, Chan VC, Shenoy N, Kirkpatrick A, Ramshaw JAM, Brodsky B. 2000. Destabilization of

osteogenesis imperfecta collagen-like model peptides correlates with the identity of the residue replacingglycine. Proc. Natl. Acad. Sci. USA 97:427378

34. Tsai MI-H, Xu Y, Dannenberg JJ. 2005. Completely geometrically optimized DFT/ONIOM triple-helical collagen-like structures containing the ProProGly, ProProAla, ProProDAla, and ProProDSertriads. J. Am. Chem. Soc. 127:1413031

35. Horng J-C, Kotch FW, Raines RT. 2007. Is glycine a surrogate for a d-amino acid in the collagen triplehelix? Protein Sci. 16:20815

36. Bodian DL, Madhan B, Brodsky B, Klein TE. 2008. Predicting the clinical lethality of osteogenesisimperfecta from collagen glycine mutations. Biochemistry 47:542432

37. Hyde TJ, Bryan MA, Brodsky B, Baum J. 2006. Sequence dependence of renucleation after a Glymutation in model collagen peptides. J. Biol. Chem. 281:3693743


Ann

u. R

ev. B

ioch

em. 2

009.

78:9

29-9

58. D

ownl

oade

d fro

m w

ww

.annu

alre

view

s.org

A

cces

s pro

vide

d by

Uni

vers

idad

Nac

iona

l Aut

onom

a de

Mex

ico

on 0

3/26

/15.

For

per

sona

l use

onl

y.

ANRV378-BI78-32 ARI 5 May 2009 15:11

38. Khoshnoodi J, Cartailler J-P, Alvares K, Veis A, Hudson BG. 2006. Molecular recognition in the assemblyof collagens: Terminal noncollagenous domains are key recognition modules in the formation of triple-helical protomers. J. Biol. Chem. 281:3811721

39. Raghunath M, Bruckner P, Steinmann B. 1994. Delayed triple helix formation of mutant collagen frompatients with osteogenesis imperfecta. J. Mol. Biol. 236:94049

40. Cram DJ. 1988. The design of molecular hosts, guests, and their complexes. Science 240:7606741. Kersteen EA, Raines RT. 2001. Contribution of tertiary amides to the conformational stability of collagen

triple helices. Biopolymers 59:242842. Nan D, Wang XJ, Etzkorn FA. 2008. The effect of a trans-locked GlyPro alkene isostere on collagen

triple helix stability. J. Am. Chem. Soc. 130:53969743. Demonstration that4R-hydroxylation of Proresidues in the Yaaposition of collagenstrands is required foranimal life.

43. Friedman L, Higgin JJ, Moulder G, Barstead R, Raines RT, Kimble J. 2000. Prolyl 4-hydroxylaseis required for viability and morphogenesis in Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA97:473641

44. Holster T, Pakkanen O, Soininen R, Sormunen R, Nokelainen M, et al. 2007. Loss of assembly of themain basement membrane collagen, type IV, but not bril-forming collagens and embryonic death incollagen prolyl 4-hydroxylase I null mice. J. Biol. Chem. 282:251219

45. Berg RA, Prockop DJ. 1973. The thermal transition of a nonhydroxylated form of collagen. Evidencefor a role for hydroxyproline in stabilizing the triple helix of collagen. Biochem. Biophys. Res. Commun.52:11520

46. Sakakibara S, Inouye K, Shudo K, Kishida Y, Kobayashi Y, Prockop DJ. 1973. Synthesis of (ProHypGly)n of dened molecular weights. Evidence for the stabilization of collagen triple helix by hydrox-yproline. Biochim. Biophys. Acta 303:198202

47. Inouye K, Sakakibara S, Prockop DJ. 1976. Effects of the stereo-conguration of the hydroxyl group in4-hydroxyproline on the triple-helical structures formed by homogenous peptides resembling collagen.Biochim. Biophys. Acta 420:13341

48. 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 81:22533

49. Suzuki E, Fraser RDB, MacRae TP. 1980. Role of hydroxyproline in the stabilization of the collagenmolecule via water molecules. Int. J. Biol. Macromol. 2:5456

50. Bella J, Brodsky B, Berman HM. 1995. Hydration structure of a collagen peptide. Structure 3:89390651. Dunitz JD, Taylor R. 1997. Organic uorine hardly ever accepts hydrogen bonds. Chem. Eur. J. 3:8998

52. Overturned thelong-standinghypothesis that waterbridges are importantfor the structure andstability of the collagentriple helix.

52. Holmgren SK, Taylor KM, Bretscher LE, Raines RT. 1998. Code for collagens stability deci-phered. Nature 392:66667

53. Holmgren SK, Bretsch

Collagen Structure and Stability

Documents

Transcript of Collagen Structure and Stability