312 



CHAPTER 23 



replicas. Many positions, however, show 

 growth on two replicas suggesting that 



these clones must have gained nutritional 

 independence at two loci. Finally, at 



many positions growth occurs on all six 

 replicas, each position of growth represent- 

 ing the occurrence of a complete prototroph 

 i I I. Bj B l\i C ). A study of these 

 clones on the replicas and on the master 

 plate shows that the changes involved 

 are transmissible and preadaptive. When 

 tested, such clones prove to be pure; that 

 is, the nutritional independence gained is not 

 attributable to any type of physical associa- 

 tion between two or more different auxo- 

 trophs. Since the findings that large num- 

 bers of clones grow on the replicas and 

 many are either complete prototrophs or 

 auxotrophs for only one nutrient cannot be 

 due to spontaneous mutation, they must be 

 attributed to some type of genetic recom- 

 bination. 



Could this genetic recombination be the 

 result of transformation? Recall that most 

 transformations involve single loci, and that 

 the frequency of double transformations is 

 much lower than the frequencies of single 

 transformations even when the loci are ap- 

 parently very close together, while double 

 recombinants are common in the present ex- 

 periment. In fact, certain recombinations for 

 two loci (for example T+L+BrB+Pa+C+) 

 occur more frequently than do recombina- 

 tions for these loci singly 



and 



T+L-BrB + Pa + C+ 



T-L + B 1 -B + Pa + C + . 



Transformation explains even less readily 

 the large number of triple recombinants — 

 the complete prototrophs. Nevertheless, 

 some additional experiments are performed 

 to test the transformation explanation. The 

 number of prototrophs obtained by recom- 

 bination is found to be uninfluenced by addi- 

 tion of DNase to the medium in which the 



bacteria are mixed or on which the bacteria 

 arc plated. No transformation occurs when 

 one culture of E. coli is exposed to filtrates 

 or autolysates of another. Therefore, the 

 genetic recombination detected in E. coli is 

 not due to transformation. 



In still another experiment, the base of 

 a U-tube is filled with sintered glass to sepa- 

 rate the arms; broth is then added, and each 

 of the two bacterial strains placed in differ- 

 ent arms. Since the sintered glass acts only 

 as a bacterial filter, the nutrient medium, 

 soluble substances, and small particles (in- 

 cluding viruses) can be shuttled back and 

 forth. Yet, when platings are subsequently 

 made from either arm, no recombinants are 

 found. Consequently, the recombinations 

 observed are not dependent upon a virus. 

 Moreover, after plating a mixture of three 

 different mutant lines of K12, we can ex- 

 plain all the different recombinants obtained 

 as resulting from recombination between any 

 two lines; no individuals obtained are re- 

 combinant with respect to markers in all 

 three lines. Apparently, this type of genetic- 

 recombination depends upon conjugation; 

 that is, it involves actual cell-to-cell contact 

 of bacteria in pairs and is, therefore, a sex- 

 ual process. 



The frequency of sexual recombination in 

 the first experiment discussed is only about 

 one per million cells (100 clones of recom- 

 binants per 100 million bacteria placed on 

 the master plate). At this point in the in- 

 vestigation, the rarity of such events makes 

 it fruitless to search for microscopic evidence 

 of bacterial mating. (The importance of a 

 new phenomenon should not be judged by 

 the frequency with which it occurs in experi- 

 ments first detecting it. Recall, for example, 

 that the quantity of DNA first synthesized 

 in vitro was infinitesimal compared with the 

 amount synthesized in later work; and the 

 rate of transformation observed initially was 

 very much smaller than the 10-25% rate 

 currently obtained with modified techniques.) 



