the next base pair by a lesser amount. Destabilization was achieved by replacing an A- T base pair by a G C base pair. The base-pair substitution mutation rate ...
Proc. Nat. Acad. Sci. USA Vol. 68, No. 4, pp. 773-776, April 1971
The Influence of Neighboring Base Pairs upon Base-Pair Substitution Mutation Rates ROBERT E. KOCH Department of Microbiology, University of Illinois, Urbana, Ill. 61801
Communicated by Alexander Hollaender, January 28, 1971 ABSTRACT The 2-aminopurine-induced transition, A T -- G C, was studied at particular sites in bacteriophage T4 as a function of the nearby base-pair composition of the DNA. Changing a base pair changed the transition rate at the adjacent base pair up to 23-fold, and at the next base pair by a lesser amount. Destabilization was achieved by replacing an A T base pair by a G C base pair. -
The base-pair substitution mutation rate at a particular genetic site is often thought to depend not only upon the base pair at that site, but also upon the molecular environment determined by nearby base pairs. This idea seems to have originated as an explanation of the "hot spots" observed in the pioneering studies of mutability at the r1I locus of bacteriophage T4 (1, 2). The main difficulty in interpreting the spectra of forward- and reverse-mutation rates arises from indefinite knowledge of which degenerate codon occurs at a given DNA locus, and about what amino acids are acceptable at particular positions in the polypeptide. Determinations of the amino acid compositions of many revertants of amber mutations located in the T4 head-protein gene, however, revealed differing rates of specific base-pair substitution at different positions within the cistron (3). Furthermore, T4rII ochre mutants were induced by hydroxylamine (CAA UAA) at rates that varied by as much as 20-fold at different sites (4). The ultravioletinduced reversion of ochre mutants within the yeast iso-icytochrome c gene differs at different positions: although glutamine is acceptable at two particular positions, for instance, it was induced in 11 of 11 revertants at one, and in 0 of 11 at another (5). While strongly suggestive of neighboring-base effects, results of this type are also compatible with an alternative interpretation, namely, that mutation rates vary within a cistron because of extrinsic factors such as direction of gene replication, direction of transcription, proximity to control elements, and so on. In fact, only a small number of externally determined gradients of mutability would be necessary to explain the variations in specific base-pair substitution mutation rates thus far observed. A convincing demonstration that nearby base pairs do in fact influence base-pair substitution mutation rates therefore requires that a mutation rate be experimentally altered by a nearby base-pair substitution. RATIONALE The amber (UAG) and ochre (UAA) chain-terminating codons are mutationally interconvertible by a transition in the third position (Fig. 1). Mutation rates at the first two positions of such homologous codons can be measured, providing that the
phenotype corresponding to each possible genotype can be recognized with confidence. These measurements are possible using the T4rII system. First, the possible types of base-pair substitutions have been restricted by using the mutagen 2-aminopurine, which produces exclusively transitions. It is already clear that 2-aminopurine readily produces transitions, and the following evidence shows that it does not produce transversions at an appreciable rate (6). Revertants of mutants of both the tryptophan synthetase A protein of E8cherichia coli and the head protein of T4, when induced by 2-aminopurine and characterized by amino acid analysis, have been shown to arise by transversions no more often than expected from contamination by the spontaneous background. Furthermore, many T4rII mutants are reverted neither by proflavin (which induces frameshift mutations) nor by base analogues (including 2-aminopurine), but do revert spontaneously, and behave like point mutations in mapping experiments. Such mutants could exist only if 2aminopurine cannot revert transversions (or, less likely, if many transitions are not reverted by base analogues, and/or many frameshift mutants are not reverted by proflavin). Finally, unpublished experiments (L. Ripley, personal communication) in this laboratory have shown that 2-aminopurine-induced revertants of numerous T4rII mutants exhibit only one phenotype, whereas spontaneous revertants of the same mutants usually exhibit several phenotypes; the induced revertants, therefore, are exclusively transitions. Second, each possible revertant genotype has been identified, as follows. (a) Transition revertants from the first and
(glutamnine)
(glutamine)
CAG
CAA
("ochre") UAA
I
E
=Z> UAG ("amber")
if
UGA
UGG
("opal")
(tryptophan)
FIG. 1. The mRNA codon conversions corresponding to the transitions measured in this study. The ochre, amber, and opal codons produce chain termination and the mutant phenotype, unless suppressed. Glutamine and tryptophan are fairly acceptable amino acids at the position represented by rUV200.
773
Genetics: R. E. Koch
774
Proc. Nat. Acad. Sci. USA 68 (1971)
second positions of the ochre codon produce glutamine (UAA -a CAA) and "opal" (UAA
--
tryptophan and the glutamine revertants), whereas mutagenized stocks of rUV200, itself, produce only one of these plaque types (the glutamine revertant). On the opal suppressor, CAJ64, mutagenized stocks of rWV200 produce two distinct plaque types that are easily distinguished as the glutamine revertant and the opal convertant. All of these plaque types were further confirmed by plating on B cells. The glutamine revertants are virtually wild type. The tryptophan revertant is decidedly r-like on B cells at 370C, but is clearly distinct from the r phenotype at 320C. The opal convertant produces the completely mutant r phenotype. Plaque morphologies were routinely determined by adsorbing the particles before plating, and control plates were prepared containing predetermined mixtures of the genotypes to be scored. Mutagenesis was conducted in M9 medium containing 1 mg/ml of 2-aminopurine. Log-phase BB cells were infected with an average of 10 phage particles and were shaken at 370C. At 10 min, an additional 10 particles were added per cell. At 180 min, lysis was completed with chloroform. Total particles were assayed on BB cells, and revertants or convertants of the rII allele were assayed on the appropriate rIIrestrictive cell strain. Acridine-resistant mutants were assayed by plating with B cells on plates containing 0.05 ,g/mi of neutral acridine. The rI mutant frequency was measured on BB cells. Crosses were performed by infecting log-phase BB cells with five particles of each parent with shaking at 370C. An average of five additional particles were added at 6 min. The complexes were diluted at 10 min, and chloroform was added at 40 min or at 60 min to complete lysis. This procedure increases recombination frequencies somewhat compared to standard
UGA) codons, respectively.
The opal codon is recognized by the ability of the phage to grow on an rIu-restrictive host carrying an opal-specific suppressor. Providing that glutamine is acceptable, the glutamine revertant is recognized by its ability to grow on an rII-restrictive host. (b) Transition revertants from the first and second positions of the amber codon produce glutamine (UAG --
CAG)
and
tryptophan (UAG -- UGG) codons, respectively. revertants can often be distinguished phenotypi-
These two cally, as follows. Many T4rII mutants are sufficiently leaky to grow on the restrictive host, while producing a mutant or semi-mutant plaque morphology on certain permissive hosts. A codon was adopted for study at which 2-aminopurine produces two types of revertants of an amber mutant, one of which is virtually wild type while the other is semi-mutant. Since the glutamine revertant from the homologous ochre mutant is wild type, the wild-type revertant from the amber mutant must also be the glutamine revertant, and the semimutant amber revertant must therefore be the tryptophan revertant. MATERIALS AND METHODS
E. coli and bacteriophage T4B were used throughout. Most strains and methods have been described (2-4, 7-9); K38 is an rI-restrictive strain that is unusually restrictive for opal mutants, and was kindly provided by Dr. J. Speyer. The rI mutant whose reversion was studied is rUV200, which maps in the A6a2 segment. This mutant contains an ochre codon by the following criteria (L. Ripley, personal communication): it is suppressed exclusively by CA165, an rII-restrictive strain carrying an ochre suppressor; it is convertible by 2-aminopurine mutagenesis to an rII mutant that is suppressed by QA1, an rI-restrictive strain carrying an amber suppressor; it is also convertible to an rI mutant that is suppressed by CAJ64, an rII-restrictive strain carrying an opal suppressor; and finally, the amber and opal convertants, when crossed, produce semi-mutant recombinants at the expected low frequency of 10-6, whereas the original ochre mutant fails to recombine ( 3' direction, and that the neighboring base effects observed here propagate only in the direction of DNA synthesis and away from position 3, consider first the replication of the nontranscribed "TAPu" strand. Here, position 3 will be filled prior to the insertion of 2-aminopurine into position 1, and position 3 could affect the (rather small) probability that 2-aminopurine is actually inserted, but it could not affect the probability that the inserted base analogue acts mutagenically at some later replication. In the case of the replication of the transcribed "ATPy" strand, however, 2-aminopurine incorporation must already have occurred at position 2 before position 3 is filled. In this case, position 3 would have its effect directly on the mutation probability during some subsequent replication (not shown), since the base analogue at position 2 mispairs with C only after position 3 has been filled. Thus, the diminished effect at position 1 compared to position 2 could simply result from a mixed mechanism, rather than from a single polar mechanism. It is a pleasure to thank Miss Lynne Bartenstein for her expert technical assistance. I also thank Mrs. Lynn Ripley, who provided
776
Genetics: R. E. Koch
the pertinent phage stocks and for helpful discussions concerning the conception of this work. I am especially grateful to Dr. John W. Drake, who was constantly available for advice and who assisted in the preparation of this manuscript. This study was supported by grants E59 from the American Cancer Society, GB15139 from the National Science Foundation, and AI04886 from the National Institutes of Health.
1. Benzer, S., and E. Freese, Proc. Nat. Acad. Sci. USA, 44, 112 (1958). 2. Benzer, S., Proc. Nat. Acad. Sci. USA, 47, 403 (1961). 3. Stretton, A. 0. W., S. Kaplan, and S. Brenner, Cold Spring Harbor Symp. Quant. Biol., 31, 173 (1966). 4. Brenner, S., A. 0. W. Stretton, and S. Kaplan, Nature, 206, 994 (1965). 5. Sherman, F., J. W. Stewart, M. Cravens, F. L. X. Thomas, and N. Shipman, Genetics, 61, s55 (1969).
Proc. Nat. Acad. Sci. USA 68 (1971) 6. Drake, J. W., The Molecular Basis of Mutation (HoldenDay, Inc., San Francisco, Calif., 1970). 7. Brenner, S., and J. R. Beckwith, J. Mol. Biol., 13, 629
(1965). 8. Sambrook, J. F., D. P. Fan, and S. Brenner, Nature, 214, 452 (1967). 9. Drake, J. W., J. Mol. Biol., 6, 268 (1963). 10. Lindstrom, D. M., and J. W. Drake, Proc. Nat. Acad. Sci. USA, 65, 617 (1970). 11. McClain, W. H., FEBS Lett., 6, 99 (1970). 12. Rogan, E. G., and M. J. Bessman, J. Bacteriol., 103, 622 (1970). 13. Shugar, D., and W. Szer, J. Mol. Biol., 17, 174 (1966); Browne, D. T., J. Eisinger, and N. J. Leonard, J. Amer. Chem. Soc., 90, 7302 (1968). 14. Drake, J. W., E. F. Allen, S. A. Forsberg, R.-M. Preparata, and E. 0. Greening, Nature, 221, 1128 (1969).
