THE BIOCHEMISTRY OE PLANT VIRUSES 73 



atoms are presumably chelated, and it is difficult to visualize how they are 

 bound unless either there are two types of subunit, only one of which can 

 combine with lead, or, more probably, that the true submiit is paired, and 

 that the lead atom chelates groups from each of these true subunits. The 

 "chemical" subunit would then be half of the "physical" subunit. A very 

 convincing argument for the existence of such subunits was put forward 

 at one time, but was later withdrawn without comment (Franklin, 1955a). 



The idea of a true subunit of about 35,000 in molecular weight, and 

 composed of two similar subunits alternating would also fit in very nicely 

 with the data obtained with the closely related cucumber virus 4, which has 

 one tryptophan residue in a subunit of this size (Knight, 1949), and which 

 has a very similar X-ray diffraction pattern (Holmes and Franklin, 1958). 



Perhaps the most convincing evidence that polypeptide chains of molecular 

 weight 17,000-18,000 exist comes from the work of Harris and Knight (1955), 

 who observed that carboxypeptidase liberates one, and only one, amino 

 acid residue from a number of strains of tobacco mosaic virus, and that this 

 amino acid is threonine. Carboxypeptidase only attacks the C-terminal end 

 of polypeptide chains, and the action of the enzyme is stopped in this case 

 after the first link is broken, because the second one happens, quite by chance, 

 to be resistant, the terminal sequence being prolyl, alanyl, threonine, and 

 prolyl hnks are not broken by the enzyme. The number of threonine mole- 

 cules liberated indicates that there is one for every 17,000 molecular 

 weight. 



This quantity of C-terminal threonine has also been found by hydra- 

 zinolysis, although rather large corrections have to be made in this case 

 (Bramiitzer, 1955). It has also been reported that the N-terminal amino acid 

 was prolme, and that one N-termination was to be found for every 17,000 

 molecular weight (Schramm et al., 1956). As wiU be seen later, this is probably 

 not a true end group, but may result from an unusually specific chemical 

 hydrolysis, and so may also give an indication of the length of the poly- 

 peptide chains. 



The bulk of the evidence, therefore, is that the tobacco mosaic virus has 

 polypeptide chains of about 17 to 18 thousand in molecular weight, possibly 

 in pairs, all of which have similar, if not identical, structures, and that they 

 are arranged in some regular order to form a tubular molecule containing 

 a core of nucleic acid, and composed of about 2800 of such subunits. 



C. The Amino Acid Composition of the Protein of a Typical 

 Tobacco Mosaic Virus 



Although the tobacco mosaic virus has been mvestigated more intensively 

 than any other virus, quantitative data on the amino acid composition are 



