STRUCTURAL AND CHEMICAL ARCHITECTURE OF HOST CELLS 133 



observed in many other types of mutation (Abrams, 1952; Cohen and Earner, 

 1954). 



The analysis of steps in tryptophan metabolism has also been most 

 instructive concerning the nature of the gene itself. Two apparently identical 

 mutants of Neurospora were found incapable of synthesizing tryptophan 

 desmolase, the enzyme condensing indole and serine to form tryptophan. 

 These mutations had occurred apparently at the same loci and were alleUc, 

 since the strains were functionally identical in all respects and crosses did not 

 result in the production of wild type strains. However a mutation at another 

 locus (a suppressor gene), which in itself did not restore the desmolase, did so 

 when present with one of the mutant genes but not the other. Thus, the 

 original apparently allehc mutants were structurally and functionally distinct 

 and two genes at different loci, each incapable of desmolase synthesis, could 

 effect synthesis when operating together (Yanofsky, 1952). Indeed, of five 

 independent suppressor mutations that have been studied for the ability to 

 restore desmolase synthesis, all have been different (Yanofsky and Bomier, 

 1955). 



This possibility of the dissection of structure within a gene locus has been 

 greatly extended in recent years by crossing-over techniques (Benzer, 1955; 

 Pontecorvo and Roper, 1956). Such techniques have revealed the existence 

 of units of recombination in Aspergillus, Drosophila, and phage, representing 

 fractions of 2 X 10"', 2.8 X 10~^, and 6 X 10-^ respectively, of the total 

 genetic maps of these organisms. In chemical terms, if these genetic units 

 consist exclusively of DNA in the form of a double-threaded hehx, if crossing 

 over involves nucleotide units, and if the probabihty of crossing over is the 

 same at all nucleotides, the number of nucleotide pairs in the smallest 

 recombination fractions consist of as few as 8 for Aspergillus, 216 for Droso- 

 phila and 12 for phage, whereas the site of mutation is sufficiently large to 

 contain on the order of lODO to 8000 nucleotide pairs. The possibility of the 

 existence of subunits of DNA molecules had been reahzed earher in the 

 discovery of an apparent crossing over between molecules of transforming 

 agents (DNA) in Pneumococcus (Taylor, 1949) and by evidence indicating 

 that units of transformation are sometimes linked within a single DNA 

 molecule (Hotchkiss and Marmur, 1954). 



Mcllwaine (1947) had observed that some coenzymes were present in 

 bacteria in such small amounts that it would be most efficient to have their 

 S3Tithesis occur on a belt fine comprised of single and unique enzyme mole- 

 cules. In such cases one might imagine that this belt fine of enzymes w^as 

 comprised of the linear array of genes of a chromosome, and that indeed the 

 genes might be the enzymes themselves. The demonstration of genetic 

 subunits by Pontecorvo and Roper began with the dissection of the genetic 

 control of vitamin synthesis in Aspergillus on the hypothesis that a close 



