J. C. KENDREW 



available, but the value is known to vary little from protein to protein), 

 we may calculate the weight of the dry unit cell as 40,800 in molecular 

 weight units. This is 2-4 times the molecular weight of myoglobin 

 (16,850), so it may be deduced that each unit cell contains two protein 

 molecules, the surplus weight being accounted for by residual water 

 which always persists in protein crystals dried at room temperature. 

 The two molecules must be related to one another by the screw diad 

 axis of symmetry mentioned above. 



PROPOSED MOLECULAR MODEL FOR MYOGLOBIN 



In order to shorten the discussion it will be convenient to anticipate 

 an account of the Patterson projections of myoglobin by describing 

 the molecular model which appears to agree most closely with all the 

 available data ; as each projection is described in turn its correspond- 

 ence to the model will be indicated. 



M. F. Perutz has shown (see p. 137) that the haemoglobin molecule 

 has a four-layer structure, each layer consisting of a circular disk 

 8-8 A thick and 57 A in diameter. It is known that myoglobin has 

 about one quarter the molecular weight of haemoglobin, and that it 

 contains one haem group instead of four. The postulate for myoglobin 

 is that its molecule consists of a single layer of parallel polypeptide 

 chains, with dimensions approximately the same as one of the four 

 layers of haemoglobin. 



The first observation which suggested this hypothesis is the close 

 correspondence between the a and c dimensions of the myoglobin 

 unit cell and the diameter of the haemoglobin molecule : Figure 3a 

 shows the ac face of the cell and a haemoglobin disk superimposed. 

 This correspondence is suggestive since, although there is no absolute 

 compulsion for a molecule to crystallize with a unit cell whose dimen- 

 sions closely resemble those of the molecule, there is often in practice 

 a near agreement if the molecule has what may loosely be called a 

 reasonable shape. 



Since the unit cell contains two molecules it is necessary to fit in two 

 disks and to arrange them so that they are related by the screw diad. 

 This has been done in Figure 3b, which is shown in projection on to 

 the ab face in Figure 3c ; the arrows indicate the orientations of the 

 molecules. It will be seen that the two disks fit very conveniently into 

 the cell, with a layer of about 6-6 A of liquid of crystallization between 

 them. It should be mentioned that the data would equally fit a 

 ' staggered ' arrangement of the kind shown in Figure 3d. Furthermore 

 the disks need not necessarily be strictly coplanar, though the Patterson 

 projections indicate that they must be approximately so. 



152 



