INFECTIVITY OF TOBACCO MOSAIC VIRUS 439 



some advantages. In general laboratory practice, groups of 6 and 12 half 

 leaves appear to have errors of about 50 and 25 %, but accuracy can be im- 

 proved when necessary by averaging many repeat assays (Loring, 1937). The 

 inclusion of a standard virus solution in all tests permits calculation of the 

 results in absolute terms, i.e., as percentage of the activity of the standard 

 TMV. Quantitative comparison of the results of several assays is thereby 

 greatly facilitated. 



As previously stated, 0.01 /zg./ml. of TMV usually produces a convenient 

 number of lesions. The nucleic acid comprises 5 % of the weight of the virus, 

 and if it represents the infectious component, then one might expect it to be 

 similarly infective at 0.0005 /xg. /ml. Actually, 200 to 2000 times as much 

 (0.1-1 fxg.) of various nucleic acid preparations is required to produce the 

 same number of lesions as is given by the 0.01 /itg./ml. standard virus. This 

 discrepancy can be partly obscured by comparing the infectivity of nucleic 

 acid and standard virus on the weight basis, when values of 10 % can be 

 arithmetically reached. The same discrepancy can also be used to conclude 

 that the infectivity of the nucleic acid is actually so low that it is obviously 

 due to contaminating, undegraded virus, and thus not worthy of serious 

 consideration. The following sections will show that neither of these trails of 

 aberrant logic is favored by the author. 



IV. Reconstitution of TMV 



New impetus was given to the interest in the mode of viral mfectivity 

 when it was shown, in 1955, that native virus protein, together with suitably 

 prepared nucleic acid, could form a co-aggregate consisting of rod-shaped 

 particles many of which were indistinguishable from the undegraded virus in 

 almost all respects (Fraenkel-Conrat and Williams, 1955). The discovery of 

 this reconstitution reaction preceded the realization that the nucleic acid was 

 infectious. The yields of reconstituted active virus were at first quite low; any 

 residual activity detected in either of the two components in separate assays, 

 usually at least two orders of magnitude lower, was regarded as a measure 

 of its contamination with virus. It is thus not surprising that these results 

 were first interpreted as indicating that viral uifectivity was a property only 

 of the complete 300 m/x particle, be it native or reconstituted in vitro. 



As further work showed that the low infectivity of the nucleic acid moiety 

 was actually an intrinsic property of the material (Fraenkel-Conrat, 1956; 

 Fraenkel-Conrat et al., 1957a; Gierer and Schramm, 1956) (see Section V) 

 the significance of the reconstitution seemed at first to be eclipsed by that 

 later finding. However, the quantitative difference between the infectivity 

 of the isolated nucleic acid and the reconstituted virus has remained the 

 same, as in the earliest experiments. Thus, combination of nucleic acid of rela- 

 tively high infectivity (i.e., 0.5 % of that contained m the standard TMV) 



