40 J. D. BERNAL 



tually permanent by the formation of covalent links, as in the production of 

 fibrinogen. The most important is the so-called globular fibrous (g ^ f ) trans- 

 formation by which small protein molecules can be rapidly and reversibly aggre- 

 gated into fibres, a phenomenon most carefully studied in the case of insulin. 

 Such transformations seem to play a considerable part in intracellular mechanics, 

 notably in the formation and disappearance of the mitotic spindle (tactoid). 



The achievement of higher degrees of regular configuration must have de- 

 pended to a certain extent on geometrical accidents, that is, it was a secondary 

 consequence of a polymerization process involving primarily only the individual 

 in the chain. For steric reasons any polynucleotide would tend to adopt a 

 quasi-spiral structure with the flat purine and pyrimidine groups piled on top 

 of each other, but nothing in the monomers could have preconditioned the pos- 

 sibility of the snug fitting of two identical polymers into a double spiral. Indeed 

 out of many possible nucleosides produced by variations in their biosynthesis, 

 those containing desoxyribose may have been selected on account of the close 

 fitting and consequently low energy of their association in pairs. How important 

 is this low energy is shown by the work of Rich [8] in which two complementary 

 synthetic polyribose nucleotides, one with purine and the other with pyrimidine 

 groups, each separately irregularly coiled, came together in vitro out of the 

 quasi-infinity of possible configurations into a close double spiral similar to 

 that of natural DNA. The stability and specificity of this fit we may well imagine 

 led, though possibly somewhat late in biopoesis, to its use as the chief carrier of 

 specificity in sexual reproduction. 



The complementary, but not symmetrical, relation of proteins and nucleic 

 acids probably belongs to this stage of biopoesis. A particular nucleic acid can 

 promote the synthesis of one specific protein molecule, as is shown in the be- 

 haviour in vivo of viruses of both ribonucleic acid (RNA) and DNA type, and as 

 Straub [9] has shown also in the production of the enzyme amylase in vitro using 

 liver nucleic acid. On the other hand a specific protein enzyme can promote 

 the syntheses of many if not all nucleic acids, as has been shown by Ochoa [10]. 

 Which particular nucleic acid is formed may depend on the nature of the nucleic 

 acid model already present. The story is only beginning to be unfolded and we 

 shall no doubt hear more about it at this Symposium; however, enough seems 

 to be estabHshed already to mark out the first appearance of the nucleic acid- 

 protein, mutual synthesis relation as a decisive stage in biopoesis. Subsequent 

 stages would involve the coming together of a set of different systems of this 

 character to provide a comprehensive self-regenerating metabohc system. 



The association of protein and nucleic acid also provides a new type of phy- 

 sical body in which a core of nucleic acid or nucleoprotein is surrounded by a 

 shell of protein molecule synthesized by the core itself. This was first described 

 for the plant (RNA) viruses by Franklin [11] and Caspar [12] and probably also 

 holds for the RNA-containing microsomes in the cytoplasm. Though it has not 

 been so convincingly demonstrated, it also holds for the DNA-containing chro- 

 matin bodies in bacteria and for the chromosomes in higher organisms. It does 

 not follow, however, that the actual production of protein or nucleic acids occurs 

 in these bodies in the condition in which these have been observed. Rather 



