Cryptic mutants of bacteriophage T4.

HE apparent mutation rate depends upon a number of contingencies: the Tprimary error rate, the probability o€ repair, error avoidance by the apparatus of DNA replication, and the probability of detection. The probability of detection of those base pair substitutions which produce an amino acid substitution appears to be highly variable, and presumably reflects the sensitivity of polypeptide functions to small variations in primary structure. Fine-scale mapping has often revealed a highly nonrandom distribution of mutations among sites. In the T4rZZ system, for instance, 809 of the first 1609 independently isolated mutants were observed to map into only two sites (BENZER 1961). Genetic data indicate that the number of rZZ sites capable of mutation by base pair substitution is about 1700 (EDGAR et al. 1962; STAHL, EDGAR and STEINBERG 1964), and this estimate is supported by quasi-chemical measurements (GOLDBERG 1966). When frameshift mutations are excluded from consideration, however, fewer than 3% of these potential sites have been identified (FREESE 1959; BENZER 1961; ORGEL and BRENNER 1961; see Chapter 5 of DRAKE 1970). What then is the nature of the apparently immutable sites? Two possible answers deserve serious consideration: many sites are extremely weakly mutable, or else many amino acid substitutions are undetected in standard screening systems because they fail to exert significantly deleterious effects upon protein function. The first possibility is supported by the observation that TbZZ amber and ochre mutations, which are likely to be detected with uniformly high efficiencies, nevertheless appear at very different frequencies at different sites during the course of hydroxylamine mutagenesis (BRENNER, STRETTON and KAPLAN 1965). The second possibility is supported by the observation that a disproportionately large fraction of T4rZZ base pair substitution mutants contain amber codons, compared to the fraction expected from random base pair substitutions (BENZER and CHAMPE 1961; CHAMPE and BENZER 1962). The second possibility is also supported by the observation that approximately one-quarter of the beginning of the rZZB cistron is dispensable (DRAKE 1963). Furthermore, the rZZ function is overproduced by the (possibly trivial) criteria of the standard screening system, since many rZZ ochre mutants are well suppressed even though the efficiency of chain propagation is well below 10% (BRENNER and BECKWITH 1965). We have approached the problem of the missing sites by designing procedures to increase the efficiency of detection of leaky mutants. It was already clear that temperature-sensitive ( t s ) mutations of the rZZ region are readily obtained