M. F. PERUTZ 



revealing the arc of high vector density at 5 A from the origin (S) and 

 the maxima along the length of the rods. The main rod is surrounded 

 by six others at a distance of 10-5 A, arranged at the corners of a 

 regular hexagon, which can be identified with the rods ' A ' and ' B ' 

 in the sections. 



It will be shown elsewhere in this volume (see p. 176) that this is 

 the type of vector structure to be expected from a set of parallel chain 

 molecules. According to the relationship between vector structure 

 and molecular structure outlined in Figure 10 of that paper (p. 177), 

 the arrangement of the rods suggests that the haemoglobin molecule 

 contains chains parallel to X with a molecular pattern repeating at 

 5 A intervals along the chain direction and a distance of 10-11 A 

 between neighbouring chains. This conclusion emerges from the 

 vector structure alone and is quite independent of other evidence, 

 comprising both chemical and x-ray data, which can be used to carry 

 the interpretation further. The detailed argument was given in // and 

 need not be repeated here. I showed there that the molecular chains 

 whose presence is indicated by the vector structure can be none other 

 than the polypeptide chains themselves. The repeat of the molecular 

 pattern at intervals of 5 A along the chain direction implies that the 

 polypeptide chains are folded, since in a fully extended chain the 

 amino-acid residues repeat at intervals of 3-4 A and not of 5 A ; 

 hence the 5 A vector would appear to represent the distance between 

 atoms which are spaced two or more amino-acid residues apart. 

 According to Figure 1 the chains would have to run parallel to the 

 base of the cylindrical haemoglobin molecule. These chains would 

 probably be arranged in the form of four layers, to conform with the 

 layered arrangement of the vector peaks in Figure 6, and the layered 

 structure of the haemoglobin molecule previously deduced from one- 

 dimensional Fourier projections (see / ; also this volume, Figure 1 

 p. 137). 



Arguing purely from considerations of packing there should be 

 20 such chains in the haemoglobin molecule. R. R. Porter and 

 F. Sanger (p. 121), on the other hand, have shown the horse haemo- 

 globin molecule to contain six terminal a-amino groups. Hence the 

 20 chains cannot be independent, but must be combined into six 

 bigger chains folded backwards and forwards in long zig-zags. 

 Alternatively there might be six open chains together with a number 

 of closed rings. 



The foregoing conclusions are summarized in Figure 8 which is an 

 idealized drawing showing the type of chain configuration and packing 

 which would be compatible with the Patterson synthesis, (a) shows 

 a polypeptide chain with a pattern repeating at intervals of 5 A in the 



142 



