46 R. MARKHAM 



were such as to deter all but the most optimistic workers. It is true that 

 colorimetric methods existed for the estimation for a number of amino acids, 

 such as tryptophan, arginine, and tyrosine. Total sulfur analysis, of course, 

 would give the cystine plus methionine, and so on. Microbiological assay, 

 using microorganisms having fastidious requirements for amino acids, have 

 been employed with some success, but these depend upon the measurement 

 of growth of the microorganism or the production of acid from carbohydrate 

 as a criterion of growth, and such methods are fraught with the dangers of 

 the production of growth inhibitors and/or stimulants during the hydrolysis 

 of the protein in order to produce the free amino acids for the assay. In 

 addition, the techniques for the hydrolysis of the proteins did not give 

 quantitative yields of the amino acids, particularly in the presence of 

 nucleic acid, which, apart from yielding glycine as a breakdown product 

 from the purine rings (Markham and J. D. Smith, 1949), may interfere with 

 the recovery of some amino acids because of its sugar content. 



In recent years, however, the techniques of amino acid analysis have been 

 greatly refined. Qualitative analysis may be performed either by paper 

 chromatography or by paper electrophoresis, the latter requiring as little 

 as half an hour's time to complete, and the ion exchange procedures developed 

 by Moore and Stein (1951, 1954) are now enabhng routine analyses to be 

 made on a few milligrams of total material — analyses which have adequate 

 precision for all normal purposes, and which can now be made by semi- 

 automatic apparatus. 



The analyses which have been performed so far on the plant viruses have 

 been more or less tentative, looking for the presence or absence of the common 

 amino acids, and attempting to find any striking peculiarities which make 

 the virus proteins stand out from other proteins. So far such studies have 

 been interesting, but are far from throwing much light on the way in which 

 viruses accomplish their object. What has been found is that similar viruses 

 may have very closely similar general amino acid compositions. It is quite 

 evident, however, that the mere routine analysis of the amino acid contents 

 of various viruses is not going to yield as much useful information as would 

 determination of the amino acid sequences in the virus as a whole. Now as 

 there are, as has already been mentioned, some hundreds of thousands of 

 amino acid residues in a single small plant virus, this task might a priori 

 seem to be an absurdly difficult one to midertake. However, it has been 

 realized for many years, largely owing to the work of Bernal and Fankuchen 

 (1941a,b), who were studying the structure of viruses with X-ray diffraction, 

 that the tobacco mosaic virus was built up of many similar, if not identical, 

 subunits, which were of the same order of size as small protein molecules. 

 The significance of this work was very largely overlooked imtil a few years 

 ago, when Harris and Knight first investigated the action of carboxypeptidase 



