Proc. Natl. Acad. Sci. USA Vol. 79, pp. 7214-7217, December 1982 Biochemistry Local mutagenesis within deletion loops of DNA heteroduplexes (base substitutions /sodium bisulfite / plasmids) KEITH W. C. PEDEN AND DANIEL NATHANS Howard Hughes Medical Institute Laboratory, Department of Molecular Biology and Genetics, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205 Contributed by Daniel Nathans, September 7, 1982 ABSTRACT An efficient method has been developed to gen- erate base substitution mutations within deletion loops of DNA heteroduplexes. This method utilizes a heteroduplex formed be- tween a deletion mutant cloned in a plasmid vector and its wild- type counterpart from which two restriction sites had been re- moved from the vector. The heteroduplex is exposed to sodium bisulfite to deaminate cytosine residues in the single-stranded loop, and the mutagenized plasmid DNA is used to transform a strain of bacteria lacking the enzyme uracil N-glycosylase. Pooled progeny DNA is digested with the two restriction enzymes, whose sites had been mutated in the wild-type plasmid, to eliminate the original deletion mutant DNA. Point mutants with C-G-to-T-A transitions are obtained at high frequency after a second trans- formation. To test the feasibility of the approach, the tetracycline resistance gene of pBR322 was chosen as the target sequence. It was found that the proportion of tetracycline-sensitive transform- ants increased as both the size of the heteroduplex loop and the time of incubation with the mutagen increased and this varied from 20% up to 70%. Nucleotide sequence analysis of several tet- racycline-sensitive mutants confirmed that C-to-T transitions had been produced in the segment of DNA corresponding to the dele- tion loop. Functional analysis of cloned genes or DNA regulatory ele- ments often entails the in vitro construction of mutations by directed mutagenesis procedures (for review, see ref. 1), Re- gions of interest can be identified initially by enzymatic removal of nucleotides and by testing the resulting deletion mutants for biochemical or biological activity. Once a critical segment of DNA is localized-by this means, finer analysis requires the con- struction of base substitution mutations within those sequences defined by the deletions. For this purpose an efficient and gen- eral method has been developed for generating base substitu- tions within deletion loops of heteroduplexes formed by pairing DNA from a deletion mutant with its full-length counterpart. MATERIALS AND METHODS Bacterial Strains. MM294 (pro’, endoA™, thi, hsdR-, hsdM*; ref. 2) was obtained from Brown Murr, BD 1528 (thyA, met”, nadBF, ung-l, gal”, supE, supF, hsdR~, hsdM*) was obtained from Bruce Duncan, and GM4§8 (thr, leu”, thi-, tonA, gal-6, lacY1, lacZ4 dam-3 dem-6) was from Bernard Weiss. Transformation. Competent cells from MM294 and BD1528 were prepared by using a modified CaCl, procedure (3, 4). With pBR322 and its derivatives, this procedure gave efficiencies of 1-2 x 10” transformants per ug of DNA with MM294 and 1- 5 X 10° transformants per ug of DNA with BD1528. Trans- formants were plated onto L plates containing ampicillin (Ap; 100-250 ug/ml), and transformants were assayed for their tet- The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked “advertise- ment” in accordance with 18 U. S. C. §1734 solely to indicate this fact. 7214 racycline (Tc) resistance on L plates containing both Ap (100- 250 pg/ml) and Tc (5 wg/ml). Partial resistance to Tc was scored when there was normal growth on Ap plates but less growth on Tc plates. Preparation of Plasmid DNA and Construction of Mutants. Plasmid DNA was amplified overnight in the presence of chlor- amphenicol at 150 g/ml (5) and was extracted by using a mod- ified cleared lysate procedure (6) in which the stock lysozyme solution was 50 mg/ml and the lysing solution was altered to give final concentrations of 0.125% NaDodSO,/0.25% Triton X-100/25 mM Tris‘HC1/12.5 mM EDTA, pH 7.4. Centrifu- gation at 35,000 rpm at 4°C for 1 hr in a Beckman SW 41 Ti rotor removed cell debris and much of the chromosomal DNA. Plas- mid DNA was purified by banding twice in CsCl/ethidium bromide gradients, and the ethidium bromide was removed by three isobutanol extractions and one chloroform/isoamyl alco- hol, 24:1 (vol/vol), extraction. The DNA was concentrated by ethanol precipitation and dialyzed against 10 mM Tris‘HCI/1 mM EDTA, pH 8. Rapid DNA preparations were made ac- cording to the method of Holmes and Quigley (7). Deletion mutants were constructed in the Tc resistance gene of pBR322 (8). Plasmid DNA was digested to completion with Sal I and a variable number of nucleotides was removed by digestion at 0°C for 1-5 min with the exonuclease BAL-31 (Be- thesda Research Laboratories), and DNA at 5 units/ ug in 100 ql of 2) mM Tris‘HCl, pH 8.0/12 mM MgSO,/12 mM CaCl,/ 200 mM NaCl/1 mM EDTA. Reactions were stopped by ad- dition .f EDTA to 20 mM and NaDodSO, to 1%; the reaction mixture was extracted with phenol and the DNA was precipi- tated with ethanol. After ligation to circles with T4 DNA ligase, digestion with Sal I, and transformation of MM294, Tc-sensitive (°) transformants were screened for the presence of deletions by restriction enzyme analysis. DNA from three deletion mu- tants, pKP10, pK P11, pKP12, was prepared and their nucleo- tide sequences were determined (Table 1). To eliminate the Ava I and EcoRI sites in pBR322, plasmid DNA was digested to completion with Ava I and the four 5’ single-stranded nucleotides were repaired by incubation of 0°C for 15 min with 1-2 units of Micrococcus luteus DNA poly- merase I in the presence of the appropriate deoxynucleoside triphosphates at 20 4M each. The DNA was circularized with T4 DNA ligase, digested with Ava I, and Ap® transformants were tested for the loss of the Ava I site. One, pKP25, was di- gested with EcoRI and that site was removed in a similar way; pKP30 is both EcoRI® and Ava I® and Ap® and Tc®. Heteroduplex Formation and Mutagenesis. DNAs were lin- earized with Pst I (pKP10, pKP11, pKP12) or with BamHI (pKP30). DNA (1 yg) from a deletion mutant was mixed with DNA (1 yg) from pKP30 in a volume of 1,155 yl of H,O and was denatured by addition of 125 yl of 1 M NaOH and incu- Abbreviations: bp, base pair(s); ®, resistant; °, sensitive; Ap, ampicillin; Te, tetracycline. Biochemistry: Peden and Nathans Table 1. Properties and sequences of plasmids Extent of deletion, Plasmid § Phenotype Restriction sites base pairs (bp)* pKP30 Ap® Tc Ava I® EcoRI? — pKP10 Ap® Te® Ava I§ EcoRI® 49 (653-701) pKP11 Ap® TeS Ava I§ EcoRI® 125 (584-708) pKP12 Ap® Te® Ava I§ EcoRI® 327 (476-802) R resistant. * Numbers in parentheses refer to nucleotide positions of pBR322. bation for 15 min at room temperature. Annealing was accom- plished by addition of 320 yl of neutralizing solution [1 M Tris-HCl, pH 7.2/1 M HCL, 2:1 (vol/vol); ref. 9] and incubation at 60°C for 2 hr. Transfer RNA was added to 20 ug/ml and the nucleic acids were precipitated by addition of 2.5 vol of 95% ethanol and overnight incubation at - 20°C. The precipitates were collected by centrifugation, and the pellets were washed in 80% ethanol at room temperature, dried, and dissolved in 100 wl of 1 mM EDTA at pH 7.5. In the initial experiments the form II, heteroduplex molecules were purified by electropho- resis on 1.5% agarose gels and were extracted from the gel by the glass fiber filter method of Chen and Thomas (10), but in later experiments the mixture of linear and circular molecules was used directly for mutagenesis. Sodium bisulfite mutagenesis was carried out according to published procedures (11, 12). The final concentration of so- dium bisulfite was 2 M; control samples were incubated with 2M NaCl. After the final dialysis step, the DNA was used to transform BD1528 cells. DNA Sequence Analysis. Mutants were subjected to se- quence analysis by using the method of Maxam and Gilbert (13) after labeling the 3’ ends of restriction fragments with M. luteus DNA polymerase I and the appropriate labeled [a**P]dNTP and unlabeled dNTP. Cleavage products were resolved on 0.3-mm 8% and 20% polyacrylamide gels (14). RESULTS AND DISCUSSION To test the feasibility of producing point mutations in single- stranded heteroduplex loops, we chose to construct mutations in the Tc resistance gene of the bacterial plasmid pBR322, be- cause phenotypic changes in this gene could be readily scored. The overall method is summarized in Fig. 1. Heteroduplex molecules were prepared as described be- tween pKP30 (Ava I® EcoRI) and the deletion mutant deriv- atives (Table 1) of pBR322, pK P10 (49-bp deletion), pKP11 (125- bp deletion), and pKP12 (327-bp deletion). Heteroduplex for- mation was indicated by the appearance of molecules migrating at the position of form II molecules after agarose gel electro- phoresis and was estimated to be between 20% and 40% of the total DNA. In the initial experiments form II molecules were purified from agarose gels (10). Subsequent experiments (see below) showed that gel purification was not necessary and the mixed population of circular and linear molecules was muta- genized directly. Mutagenesis of the three purified heteroduplexes with 2 M sodium bisulfite was carried out as described (11, 12) for 1 hr and 2 hr. Control samples were incubated for 2 hr with 2 M NaCl. After dialysis the mutagenized and control DNA samples were used directly to transform BD1528 cells, a strain of Esch- erichia coli deficient in the enzyme uracil N-glycosylase (ung). Use of an ung” strain (BD 1528) was found to be obligatory, be- cause transformation of an ung™ strain with bisulfite-treated heteroduplex DNA resulted in the loss of the heteroduplex loop and the only Tc* mutants found contained the original deletion. Replication of heteroduplex molecules after transformation Proc. Natl. Acad. Sci. USA 79 (1982) 7215 EcorI*® “(eoo) re Ap oe) TcP / YA A Aval® JPsti | BamHI linear linear denature anneal CO + linears | sodium bisulphite mutagenesis | transformation of ung” bacteria + | AvaI + EcoRI | transformation Fic. 1. Summary of the mutagenesis procedure. The plasmid pKP10 is Ap® and carries a deletion rendering it Tc®, whereas pKP30 is both Ap® and Tc® as well as being resistant to restriction enzymes EcoRI and Ava 1. should result in each cell containing the two original plasmid species, assuming that the deletion loop is not excised prior to replication. To enrich for the full-length plasmid, DNA pre- pared from pooled progeny BD1528 transformants was digested with Ava I and EcoRI prior to a second transformation of MM294 cells. Because the Ava I and EcoRI sites had been eliminated in the full-length molecule (pKP30), digestion with these enzymes selectively fragmented the deletion mutant DNA. Because linear molecules with noncompatible ends transform about 1/100th to 1/1,000th as efficiently as circular molecules (unpublished data), the yield of deletion mutant DNA was greatly decreased. Individual Ap® MM294 colonies resulting from the second round of transformation were tested for their Tc resistance and the results are summarized in Table 2. With no sodium bisulfite treatment and no enzyme digestion before the second trans- formation, clones containing full-length plasmids (Tc®) occurred at frequencies varying from 36% with the 49-nucleotide loop down to 14% with the 327-nucleotide loop. If the DNA was digested with Ava I only, then the proportion of Te® transform- ants increased to >80% with all three loop sizes (data not shown). This value could be increased to >90% by digestion with both Ava I and EcoRI. In subsequent experiments in which heteroduplex molecules were not gel purified (Table 2), 98% of all Ap® colonies arising after transformation with DNA di- gested with Ava I and EcoRI were Tc*—i.e., only 2% carried the deletion in the Tc gene. The lower proportion of Tc® trans- formants when the heteroduplex molecules were gel purified was most likely due to nicking of the single-stranded loop during recovery from the gel. Sodium bisulfite mutagenesis increased the proportion of 7216 Biochemistry: Peden and Nathans Table 2. Phenotypes of transformants Proc. Natl. Acad. Sci. USA 79 (1982) Te sensitivity Heteroduplex Time of Ava I and Partially pKP30 with mutagenesis, min* EcoRI* Resistant resistant Sensitive Gel purified? pKP10 0 - 36 0 64 49-bp deletion 0 + 91 0 9 60 + 69 8 23 120 + 60 12 28 pKP11 0 - 22 0 78 125-bp deletion 0 + 91 0 9 60 + 64 13 23 120 + 52 18 30 pKP12 0 - 14 0 86 327-bp deletion 0 + 92 0 8 60 + 47 8 45 120 + 21 6 73 Mutagenized directly’ pKP11 0 - 26 0 74 125-bp deletion 0 + 98 0 2 60 + 83 6 11 120 + 63 13 24 *The times of incubation with 2 M sodium bisulfite were 60 and 120 min; 0 means that the DNA was incubated with 2 M NaCl. + Indicates whether the DNA was digested prior to the second transformation. +The heteroduplexes were gel purified. §The mixture of heteroduplexes and linears was mutagenized directly. Tc® colonies as both the time of exposure to the mutagen and the size of the deletion loop increased (Table 2). For example, with a loop of 327 nucleotides, 45% of the Ap® transformants were made Tc° with 1 hr of 2 M sodium bisulfite treatment, and this was increased to 73% after an additional hour of incubation. With a loop size of 49 nucleotides, ~30% of the Ap® transform- ants were rendered Tc® after a 2-hr incubation with 2 M sodium bisulfite. In addition to the Tc* phenotype, transformants that were partially resistant to Tc were found, as might be expected if the mutation decreased rather than eliminated the activity of the protein. As an alternative approach to the enrichment of full-length molecules over deleted molecules, the effect of methylation in dam* E. coli on the relative recovery of full-length molecules was investigated (15). Unmethylated plasmid DNA was ob- tained by propagation in the dam™ strain GM48. Heterodu- plexes were prepared between pKP30 and either methylated or unmethylated pKP11 (125-bp deletion). With no Ava I or EcoRI digestion prior to the second transformation, Tc resis- tance was found at a frequency of 37% with dam* /dam~ het- eroduplexes and at a frequency of 26% with dam* /dam* mol- ecules. If the DNA samples were first digested with Ava I and EcoRI, then the proportion of Tc® transformants from the dam* / dam* heteroduplex was increased to 98% (Table 2), whereas that from the dam* /dam™ heteroduplex was 100%. Therefore, the use of the dam methylation system in this procedure re- sulted at best in only a slight improvement in the yield of full- length molecules when the deletion loop is 125 nucleotides in size. [The difference between dam* /dam* (98%) and dam* / dam” (100%) suggests that correction of the mutations confer- ring Ava I and EcoRI resistance may have occurred after trans- formation with dam* /dam* heteroduplexes, and perhaps it was this correction rather than incomplete digestion that accounted for the difference. | To ascertain whether the Tc’ phenotypes were due to the production of base substitution mutations or due to the intro- duction of small deletions through bacterial correction mech- anisms, several mutants were subjected to sequence analysis. Ala LeuLeuLeuGlyCysPheLeuMetGlnGluSerHisLysGlyGluArgArgProMet ProLeuArgAlaPheAsnProVal SerSerPheArgTrpAlaArgGly Mutant GCG(18 bp) CTACTACTGGGCTGCTTCCTAATGCAGGAGTCGCATAAGGGAGAGCGTCGACCGATGCCCTTGAGAGCCTT CAACCCAGTCAGCTCCTTCCGGTGGGCGCGGGGC -- An- 2 A- 3 A A A- 4 TT T T 5 00 A~--A: A A A 6 -—- ---T T Leu SerTyr Ile Leu Glu Pro PheLeu Leu TrpTerm Fic. 2. Sequences of constructed point mutants. The nucleotide sequence of the Tc resistance gene and the deduced amino acid sequence of its protein for the relevant region is shown. The dashed lines indicate the nucleotides exposed to the mutagen in the six mutants analyzed: 49 nucleotides in mutants 1-4 and 125 nucleotides for mutants 5 and 6. Mutants 1, 2, 5, and 6 were incubated with sodium bisulfite for 1 hr, whereas mutants 3 and 4 were incubated for 2 hr. Term, chain termination. Biochemistry: Peden and Nathans Fig. 2 shows the sequences of six mutants, four produced by mutagenizing the 49-nucleotide loop, and two produced by mutagenizing the 125-nucleotide loop. Mutants 1, 2, 5, and 6 were obtained after 1 hr of sodium bisulfite treatment; mutants 3 and 4 were obtained after 2 hr of treatment. The results show that in all cases the changes found are those predicted by sodium bisulfite-induced C-to-T transitions in one or the other of the two strands, but C-to-T and G-to-A changes in the same molecule were not found. This result demonstrates that, as expected, each strand is available for mutagenesis. A comparison of the sequence changes in mutants 1 and 2 with mutants 3 and 4 indicates that the level of base substitution in- creases as the time of incubation is increased from 1 to 2 hr. Mutants 1 and 2 have only one change each, whereas mutants 3 and 4 have three changes and four changes, respectively. In the mutants thus far subjected to sequence analysis—the 6 shown in Fig. 2 and an additional 14 not shown—all of the base substitutions occurred within the heteroduplex loop. No changes were found up to 100 nucleotides on either side of the loop. Fig. 2 also shows that sodium bisulfite can react with cy- tosine residues within two and three nucleotides of the base of the loop. The mutants shown in Fig. 2 were selected either for their Tc sensitivity or, in the case of mutant 6, its partial Tc resistance. Therefore, silent base substitutions would not have been ob- served and the actual level of base substitution should be higher than that predicted from the loss of antibiotic resistance. In re- cent experiments it has been feasible to screen for mutants by DNA sequence analysis. In conclusion, it has been demonstrated that point mutations can be produced efficiently in deletion loops of plasmid het- eroduplexes. The method is generally applicable to any cloned DNA segment, although the choice of restriction sites to be Proc. Natl. Acad. Sci. USA 79 (1982) 7217 removed from the vector will be dictated by the sequence of the cloned fragment. We thank Phoebe Mounts for discussions and critical comments on the manuscript, Bruce Duncan for generously supplying BD1528, Ber- nard Weiss for suggesting the use of the dam methylation system, and Doris Wiczulis and Lily Cowan for typing the manuscript. This work was supported by Grant 5 PO1 CA16519 from the National Cancer Institute. 1. Shortle, D., DiMaio, D. & Nathans, D. (1981) Annu. Rev. Genet. 15, 265-294. 2. Backman, K., Ptashne, M. & Gilbert, W. (1976) Proc. Natl. Acad. Sci. USA 73, 4174-4178. Lederberg, E. M. & Cohen, S. N. (1974) J. Bacteriol. 119, 1072- 1074. Morrison, D. A. (1979) Methods Enzymol. 68, 326-331. Clewell, D. B. (1972) J. Bacteriol. 110, 667-676. Katz, L., Kingsbury, D. T. & Helinski, D. R. (1973) J. Bacteriol. 114, 577-591. Holmes, D. S. & Quigley, M. (1981) Anal. Biochem. 114, 193- 197. Bolivar, F., Rodriguez, R. L., Greene, P. J., Betlach, M. C., Heyneker, H. L., Boyer, H. W., Crosa, J. H. & Falkow, S. (1977) Gene 2, 95-113. 9. Lai, C.-J. & Nathans, D. (1975) Virology 66, 70-81. 10. Chen, C. W. & Thomas, C. A., Jr. (1980) Anal. Biochem. 101, 339-34]. ll. Shortle, D. & Nathans, D. (1978) Proc. Natl. Acad. Sci. USA 75, 2170-2174. 12. Giza, P. E., Schmit, D. M. & Murr, B. L. (1981) Gene 15, 331- 342. 13. Maxam, A. M. & Gilbert, W. (1980) Methods Enzymol. 65, 499- 560. 14. Sanger, F. & Coulson, A. R. (1978) FEBS Lett. 87, 107-110. 15. Meselson, M., Pukkila, P., Rykowski, M., Peterson, J., Radman, M., Wagner, R., Herman, G. & Modrich, P. (1980) J. Supramol. Struct. Suppl. 4, 311. CoN OOS