BRIEF DESCRIPTION OF THE STRUCTURE OF COLLAGEN PROPOSED BY DR. LINUS PAULING

(P. C. Corns 3-8-67

Intro.

Collagen is a protein found mainly in connective tissues, although it may occur to a greater or lesser extent in other parts of the living body.

It has excited a good deal of interest in the field of protein structure because of its rather unusual amino acid content, containing exceptionally high quantities of Proline and Hydroxy Proline, and a (?) pleasing 'coiled coil' type tertiary structure, consisting of three polypeptide backbones coiled around one another.

At the present time three are these are two basic structures proposed, not of Ramachandran1 and two closely related structures from Crick and Rich2, both of which consist of three polypeptide chains coiled round one another, and which are merely compatible with a skewed (?) X-ray differentiation pattern. However, both forms accept the assumption, made by Bader and Pullam(?)3, not all the peptide residues(?) are trans/ However, this assumption in s felt to be false as the assumption was made on (?) from satisfactory evidence gained from I.R. (?), and also, of course, the presence of the carbon ring in proline and hydroxy proline present an excellent means of subsidizing the cis configuration. It is for these reasons that this new structure is proposed.

p.r.o.

Experimental evidence has shown that the Glycine-Proline-Hydroxy Proline sequence is very common in collagen, and to a good approximation it may be stated that every (?) in Glycine. Because of this common sequence the models made all assumed that it was the only one to occur, although slight distortions in collagen may be explained by the presence of other amino acids.

p.r.o.

The presence of the nitrogen atom with its low pair of electrons given the [hand-drawn image] bond 40% double bond character (?) by the process of delocalization [hand-drawn image] and (?) cis and trans structures may be considered.

When such delocalization occurs from the nitrogen in the proline as hydroxy-proline, an excellent case may be made for the stabilization o;f the cis configuration by the very presence of the carbon ring. Accordingly, for the unorganized primary structure, a trans-cis-trans configuration has been proposed, although because in the models built to date as differentiation has been made between proteins and hydroxy proline the configuration of the sequence may equally well lead trans-trans-cis, although there is doubt in the writers' mind as to the inductive affect of the hydroxy group on the double bond (?) of the hydroxy proline residue, if any.

It is found that in the stated configuration of the sequence, the backbone may be twisted into a left-hand (?) helix with approximately 3 residues (?) per helix to give the correct translation of ~8.6 Å.

Furthermore, this model proves to be satisfactory on consideration of the van der Waal contact distances. To begin with (?) polystyrene spheres of radius 1 Å were inserted, upon which to represent hydrogen atoms it was found that there were no short contacts involving protons. Also certain other contact distances were checked and found to be acceptable. These distances are given immediately below in table 1 where the rotations used are the same as those employed by Ramakrishnan and Ramachandran4.

TABLE OF CONTACT DISTANCES

FROM TO OBSERVED DISTANCE MIN. DISTANCE ALLOWED

C1 3.8 Å

CB C1 4.5 Å 3.20 Å

O1 5.0 Å 2.80 Å

C'1 3.7 Å 2.95 Å

C'2 C1 3.8 Å

O1 5.0 (?) Å

C2 3.0 Å 2.80 Å

CB N2 3.7 Å 2.9 Å

C3 5.0 Å 3.20 Å

Tertiary Structure

As with Crick and Rich, it is found that the three polypeptide backbones only form one systematic set of hydrogen bonds each, although they may be constructed in two different ways giving rise to two closely related structures.

Structure 1

In this case the hydrogen bonds are formed both from and to the glycine residue [hand-drawn image]

When this structure was assembled several features were noted. to begin with the hydrogen bonds were at the same level forming an equilateral triangle perpendicular to and symmetrically about the fisher anion (?), thereby giving the structure its required 3-fold axis of symmetry. Also the model required very little alteration to give the required pitch (residue height) of 2.86 Å. Finally it was observed that the polypeptide backbones had naturally begun to coil around one another, although at this stage the degree of coiling was not sufficient to give the required amount. ( 2th/3 in 10 residues). This means that the 10th residue of one chain should lie directly above the zeroth residue of an adjacent chain. It was observed that the coiling of the 3 polypeptide backbones was in a right-handed manner which is observed.

N.B. This model locks the 3-fold screw axis that is a feature of the Crick and Rich structures.

Structure II

In this case the hydrogen bonds were formed from the carbonyl of the residue formed between glycine and a ring nitrogen, and to the nitrogen of the glycine.

By forming hydrogen bonds in such a manner it is found that a screw axis is formed. However, on closer examination of the model it was found that there were a number of serious short contacts, between the δ c of the amino acid directly above glycine in the chain and the β c of the amino acid directly below it. This distance was found to be on average 2.5 Å, which is at least 0.5 Å too short, without even considering the interaction of the protons on those carbon atoms.

The structure may be worth considering in parts (?) of the protein fiber where proline and hydroxyproline do not occur to such a great extent. As the nitrogen of the glycine may be linked to a mixed carbonyl group by means of the hydrogen bond it was felt that a third structure should be attempted. However, it was soon found that such an attempt was bound to fail as the hydrogen bonds are formed solely about one peptide bond, so that they are, in fact, parallel. Hence when the chains are joined the hydrogen bonds have to bend through 120º, and the chains are squashed together.

Such hydrogen bond formation is compatible with the "pleated sheet" type structure of other proteins.

References

1. Ramachandran - Bose 70th Birthday - Pt II

2. Crick and Rich - J. Mol. Biol. 3, P. 483.

3. Badger and Pullin - J. Chem. Phys. 22, 1142

4. Ramakrishnan and Ramachandran - Stereochemical Criteria for Polypeptide and Proton Chain Conformations pt II.

Reading

Selection and There (?) Research. (Resent Advances in)

Ed. G. Stainsby (?) Publisher: Pergammon (?) Press.