<4 R. MARKHAM 



surprisingly scarce. In part this is due to the fact that accurate and reproduc- 

 ible methods for the analysis of amino acids have been developed compara- 

 tively recently, while much work was done before this by the tedious, and 

 relatively inaccurate, methods available before amino acid chromatography 

 had reached its present state of development. The best of the earlier methods 

 for the general analysis of amino acids was undoubtedly the microbiological 

 assay technique, which has been used extensively by Knight, and which 

 has produced much interesting comparative data between virus strains. 

 It is, however, a method which assays the amino acid in a hydrolyzate, using 

 the former as a growth factor in an otherwise complete medium for the 

 organism used for the assay, and is thus inevitably suspect, because the 

 hydrolyzate may contain inhibitory substances or even growth-promoting 

 substances. There are, in fact, indications that the method may overestimate 

 some amino acids and underestimate others in as complex a substance as a 

 virus. As it happens, analyses of tobacco mosaic virus, and indeed of the same 

 strain as used by Knight (1949), have been made for an entirely different 

 reason, and published by Newmark and Fraser (1956), and, using these and 

 Knight's data, it is possible to obtain a fairly reasonable and complete idea 

 of the actual amino acid composition of the virus. The analyses made by 

 the later workers were obtained by the Moore and Stein ion-exchange 

 chromatographic technique, which has been investigated most thoroughly 

 and is generally accepted as being reliable. There are, however, one or two 

 gaps in these data which are filled in from the microbiological assay data, 

 most of which are in good agreement with the ion-exchange results, but some 

 of which differ very appreciably from the latter. However, in this way one 

 can obtain a reasonably good idea of the amino acid composition as a whole. 



B. The Polypeptide Subunit 



Making the assumption that all the polypeptide chains are the same, and 

 that their length can be estimated from the information given earlier, one 

 may fit the amino acid analyses so that the chain contains integral numbers 

 of residues. If one does this, one finds that the chains have a molecular weight 

 of about 17,000 (the exact figure cannot be calculated directly from the data 

 because of the existence of amide groups in the molecule, and possibly of 

 cross linkages which involve eHmination of water). These chains contain about 

 145 amino acid residues. 



The chains contain 1 residue of cysteine. This is evidently on or near the 

 surface of the virus because it is reactive to substances such as methyl 

 mercuric chloride and iodine. The cysteine residue, presumably because of 

 its isolation, is not oxidized readily, and may be substituted by an iodine 

 atom to give an — S — I substituent. This type of grouping is usually very 



