Formula De Cupido

The importance of proline residues in the structure, stability

and susceptibility to proteolytic degradation of collagens

Stephen M. Krane

Received: 16 November 2007 / Accepted: 7 February 2008 / Published online: 23 April 2008

_ Springer-Verlag 2008

Abstract Collagens are among proteins that undergo

several post-translational modifications, such as prolyl

hydroxylation, that occur during elongation of the nascent

chains in the endoplasmic reticulum. The major structural

collagens, types I, II and III, have large, uninterrupted

triple helices, comprising three polyproline II-like chains

supercoiled around a common axis. The structure has a

requirement for glycine, as every third residue, and is

stabilized by the high content of proline and 4-hydroxyproline

residues. Action of prolyl hydroxylases is critical.

Spontaneous or targeted genetic defects in prolyl hydroxylases

can be lethal or result in severe osteogenesis

imperfecta. Prolines, as determinants of substrate specificity

and susceptibility, also play a role in degradation of

collagen by collagenolytic matrix metalloproteinases

(MMPs). Targeted mutations in mice in the collagenase

cleavage domain have profound effects on collagen turnover

and the function of connective tissues. Prolines are

thus critical determinants of collagen structure and

function.

Keywords Prolylhydroxylases _ Collagen structure _

Osteogenesis imperfecta _ Collagenases _

Matrix metalloproteinases

Introduction

Collagens with uninterrupted triple helices, such as type I

collagen, are characterized biochemically by distinctive

amino acid composition, i.e., on a molar basis *33%

glycine, *10% proline, *10% 4-hydroxyproline and

*0.1% 3-hydroxyproline. This review will focus on the

unique roles of prolines and modified prolines in determining

the structure and biological stability as well as

proteolytic degradation of type I collagen.

Collagen structure, proline and 4-hydroxyproline

The structure of the collagens reflects the amino acid

composition. Thus, there is in the triple helical domain a

requirement for glycine (Gly) as every third residue and

the high content of prolines (Pro) stabilizes the polyproline-

II-like helices characteristic of collagen sequences

(Myllyharju and Kivirikko 2004). Hydroxylation of Pro in

the 4-position (4-Hyp, i.e., 4(R)-hydroxl-L-proline

(Schumacher et al. 2006) further stabilizes the helical

structure. This 4-hydroxylation takes place during elongation

of the nascent polypeptide chains in the endoplasmic

reticulum. 4-Hyp was isolated and its structure determined

at the turn of the last century (Fischer 1902). In an

extraordinary paper published over four decades later

(Stetten and Schoenheimer 1944), it was reasoned that,

‘‘perhaps a part of the hydroxyproline is metabolized by

way of proline’’. They then synthesized l (-) Pro where

‘‘the carbon skeleton was marked by stably bound deuterium

and the amino group by N15’’ and they fed the doubly

labeled Pro to rats and measured enrichment of the isotopes

in Hyp from carcass proteins. They concluded that 4-Hyp is

formed after incorporation of Pro into these proteins

S. M. Krane (&)

Department of Medicine, Harvard Medical School

and the Massachusetts General Hospital, Center for Immunology

and Inflammatory Diseases, Building 149, 13th Street,

Room 8301, Boston, MA 02129, USA

e-mail: krane.stephen@mgh.harvard.edu

123

Amino Acids (2008) 35:703–710

DOI 10.1007/s00726-008-0073-2

(collagens). We now know from amino acid and gene

sequencing that the helical region of these collagens

comprises uninterrupted repeats of Gly-X-Y tripeptides

(Piez 1984; Myllyharju and Kivirikko 2004). It has been

established by amino acid sequencing that whereas Pro is

found in either the -X- or -Y- position of the Gly-X-Y

tripeptide repeat, 4-Hyp is found only in the -Y- position.

4-Hyp is an abundant modification in types I and II collagens,

with * 85–90 residues/1000 amino acids (*40% of

the total Pro + 4-Hyp). Any substitutions for Gly residues

that result from mutations in the types I, II, III or V collagen

genes produce an unstable structure that results in

human diseases that primarily affect type I collagen in bone

(forms of ‘‘brittle bone disease’’, i.e., osteogenesis imperfecta),

type II collagen in growth plate or articular cartilage

(chondrodysplasias) or types III or V collagens in skin and

other ‘‘soft ‘‘connective tissues (forms of Ehlers-Danlos

syndrome) (Myllyharju and Kivirikko 2004; Byers 2000).

The triple helical structure of the fibril/fiber-forming

collagens thus reflects the high content of Pro and 4-Hyp

residues with a regular and uninterrupted repeat of Gly-XY

triplets. The denaturation temperature (TD) of collagen in

solution is usually measured as the melting temperature

transitition (Tm), and the TD of native, insoluble collagens

is usually measured as the shrinkage temperature (Ts). In

general, the stability of the helical structure is regulated by

the content of Pro plus 4-Hyp. These numbers have physiological

meaning. It has been shown (Piez 1984) that the

Tm of different animal collagens is proportional to

the upper limit of environmental temperature to which the

animal or tissue is exposed. For example, the Tm of skin

collagen from codfish that live in cold waters is *14_C,

whereas the Tm of hog intestine collagen and the cuticle

collagen of the parasite, Ascaris, which inhabits the hog

intestinal lumen, is *40_C. Furthermore, based on a

detailed analysis of these and other collagens from different

species (Josse and Harrington 1964), the best

correlation with TD is with total pyrrolidine content

(Pro + 4-Hyp) and not with either Pro or 4-Hyp content

alone. More recently, analyses have been made of the

refolding of thermally denatured model collagen-like

peptides that have Gly-X-Y triplet repeats with differing

residues in the -X- and -Y- positions (Ackerman et al.

1999). The results showed a strong dependence of the

folding rate on the identity of the ‘‘guest’’ triplet. For

example, all triplets of the form Gly-X-4-Hyp promoted

rapid folding of these model peptides, whereas triplets of

Gly-Pro-X and Gly-X-Y had much slower folding. In vivo,

there are additional mechanisms that direct an assembly of

these complex macromolecules (Khoshnoodi et al. 2006).

The component a chains have non-collagenous (NC)

domains at their N- and C-terminal ends. Type I collagen

comprises two a1 chains and one a2 chain in the triple

helix; it has now been shown that the a1 C-NC domain

contains sufficient information for directing homotrimer

assembly, whereas the a2 C-NC domain is required for

directing heterotrimer assembly (Khoshnoodi et al. 2006).

Disulfide bonding in the a2 C-NC domain is critical for this

process. Before the nascent chains fully fold, the prolyl

residues in the -Y- position are hydroxylated and the newly

formed 4-Hyp residues serve to stabilize the helix

(Myllyharju and Kivirikko 2004).

The enzymes that catalyze the formation of collagen

4-Hyp are the collagen prolyl 4-hydroxylases (P4Hs)

members of a family of 2-oxyglutarate and ferrous irondependent

dioxygenases (Myllyharju and Kivirikko 2004).

The P4Hs are a2b2 tetramers that localize in the endoplasmic

reticulum; the a subunits are the catalytic

hydroxylases whereas the chaperone, protein disulfide

isomerase (PDI), comprises the b subunits. There are three

isoforms of the a subunits in humans (Fig. 1). Spontaneous

mutations of collagen prolyl 4-hydroxylases have not yet

been reported and a targeted null mutation of one isoform

in mice is embryonic lethal (Myllyharju and Kivirikko

2004). The collagen prolyl 4-hydroxylases all require as

substrates unfolded collagen chains, molecular oxygen,

2-oxyglutarate, ferrous iron and ascorbate; substrate affinities

vary with the enzyme isoform. Other prolyl

4-hydroxylases function in regulating responses to hypoxia

by hydroxylating key prolyl residues in the sequence of the

hypoxia-inducible transcription factor, HIFa. The amino

acid sequence specificity that determines 4-hydroxylation

of prolyl residues in HIFa is different from that of the

collagen P4Hs (Kaelin 2005). The HIFa hydroxylases,

however, do have other substrate requirements that are

similar to those of the collagen prolyl 4-hydroxylases,

although they do not hydroxylate Pro residues in collagens,

have different affinities and do not complex with PDI b

subunits. Prolyl hydroxylation of HIFa alters binding to the

von Hippel Lindau tumor suppressor protein (pVHL) and

Fig. 1 Prolyl 4-Hydroxylases

704

...