]'.. D. Davis: Studies r.n Nutritionally Deficient Bacterial Muta 



47 



stitutions of carboxyl, amino, and phenolic groups, 

 were tested for their capacity to substitute for the 

 multiple requirements of the aromatic-less mutants. 

 None of these was effective. Similarly cyclohexane 

 carboxylic acid, cyclohexanol, and inositol were not 

 utilized. Shikimic acid, however (suggested and fur- 

 nished by R. Stanier) was used by aromatic-less 

 mutants with quadruple, but not with triple or double 

 requirements. 



C 

 HC CH 2 



HOHC CHOH 



\/ 



CHOH 



shikimic acid 



It therefore appears that the precursor of the benzene 

 ring is at least partly saturated. These mutants do not 

 use quinic acid, in which the double bond of shikimic 

 acid is hydrated. 



It will be noted that Fig. 4 contains an unknown 

 "end-product", presumably aromatic, labelled Y. The 

 evidence for the existence of this postulated compound 

 is as follows, (a) The aromatic-less mutants with 

 quadruple requirements require more PABA than does 

 the PABA mutant, and their growth is accelerated by 

 high concentrations of PABA, suggesting that the 

 synthesis of PABA may be reversed, to yield a precursor 

 common to PABA and an unknown compound, (b) 

 Growth of the aromatic-less mutants on large amounts 

 of the four aromatic compounds is slow; similarly, 

 growth on shikimic acid alone is slow. But shikimic 

 acid plus the three amino acids yields much faster 

 growth, practically as fast as wild-type. These results 

 suggest that shikimic acid may occupy position B, 

 rather than A, serving rapidly as a precursor of Y plus 

 PABA (for which PABA alone serves only slowly), but 

 serving only slowly, by reversal of the normal process, 

 as precursor for the amino acids. 



The scheme of Fig. 4 must be considered quite 

 tentative. Certain phenomena are difficult to explain, 

 such as the fact that mutants with a double aromatic 

 requirement (tyrosine plus phenylalanine) heavily 

 feed the quadruples, while the triple neither feeds the 

 quadruples nor is fed by the doubles. One must there- 

 fore consider more seriously the possibility, always 

 theoretically present, that some of these multiple re- 

 quirements may depend on internal inhibition, as with 

 isoleucine and valine, rather than on block at an early 

 stage of synthesis. 



Partial back-mutants 



Almost all the mutants are detectably unstable; 

 that is to say, spontaneous reversions to nutritional 



independence (prototrophs) occur with a high enough 

 frequency (10- 7 -10 -8 ) to give rise after several days to 

 a few large colonies of back-mutants in a streak on a 

 medium which is sufficiently enriched to permit limited 

 growth (cf. Fig. 2). For this reason practically all of 

 our experiments, even on quantitative response, are 

 carried out on solid media ; in tubes of liquid media the 

 greater precision of measurement by turbidimetry is 

 accompanied by greater difficulty in distinguishing 

 growth of back-mutants from that of parent mutants. 

 Maintaining transfer cultures of mutants in liquid 

 media, however, has caused no difficulty provided 

 selection of back-mutants is avoided by the presence 

 of an excess of the nutrient requirement. 



Our study of back-mutants had a casual origin which 

 is pointed out here since it illustrates the possibility, 

 based on the uniform colonial growth of bacteria, of 

 encountering interesting phenomena by simple observa- 

 tions on plates. Two presumptive back-mutant colonies 

 from a PABA-less mutant were isolated and streaked 

 on minimal medium, along with wild-type, to verify 

 their nutritional independence. One of these grew as 

 rapidly as wild-type ; the other was by chance observed 

 at an early time (18 hrs.) to form slightly smaller 

 colonies than wild-type. Further study showed that 

 the slightly slow prototroph grew as rapidly as wild- 

 type in a medium supplemented with PABA; it ap- 

 parently had recovered the capacity to synthesize 

 PABA, but not rapidly enough to permit optimal 

 growth. 



Since PABA is one of the factors excreted by wild- 

 type in large amounts, this hypothesis was easily 

 tested. The two PABA back-mutants were compared 

 with wild-type for their output of PABA on minimal 

 medium, by pouring plates of minimal medium con- 

 taining few cells of the prototroph and many PABA- 

 requiring cells. (This technique is more sensitive to 

 slight differences in syntrophism than is the technique 

 of adjacent streaks.) As expected, the wild-type 

 colonies were surrounded by an extensive halo of 

 satellites, while the slow prototroph had none. This 

 fact confirmed the earlier conclusion that its growth 

 rate was limited by its rate of synthesis of PABA ; it 

 had none to spare. Unexpectedly, however, the other 

 back-mutant, previously indistinguishable from wild- 

 type, was surrounded by a smaller halo of satellites 

 than was wild-type. Apparently its recovered capacity 

 to synthesize PABA was not as great as that of wild- 

 type. 



Following this a dozen different back-mutant 

 strains, spontaneous and ultraviolet induced, were 

 isolated from a PABA mutant ; among these 6 different 

 rates of growth in the absence of PABA were recogniz- 

 ed. In addition, all were more susceptible than wild- 

 type to sulfonamide inhibition, which is not surprising 

 since all synthesized PABA at a lower rate than wild- 

 type. Some of the slower strains had their growth rate 



