Double Base Substitutions Research Articles

Mutational analysis of the bglH catabolite-responsive element ( cre) in Lactobacillus plantarum

Catabolite repression of the Lactobacillus plantarum bglH gene is mediated by the catabolite control protein A (CcpA). Here we show that the binding site for the protein is a catabolite-responsive element ( cre) sequence overlapping the start site of transcription. Two single and one double base substitutions in the cre sequence caused derepression of a plasmid-borne bglH–cat transcriptional fusion in L. plantarum cells grown on glucose. Analysis of the single mutations indicates that the G and C nucleotide residues in positions 2 and 13, respectively, of the 14-bp cre sequence are required for catabolite repression.

PATCHED and p53 gene alterations in sporadic and hereditary basal cell cancer.

It is widely accepted that disruption of the hedgehog-patched pathway is a key event in development of basal cell cancer. In addition to patched gene alterations, p53 gene mutations are also frequent in basal cell cancer. We determined loss of heterozygosity in the patched and p53 loci as well as sequencing the p53 gene in tumors both from sporadic and hereditary cases. A total of 70 microdissected samples from tumor and adjacent skin were subjected to PCR followed by fragment analysis and DNA sequencing. We found allelic loss in the patched locus in 6/8 sporadic basal cell cancer and 17/19 hereditary tumors. All sporadic and 7/20 hereditary tumors showed p53 gene mutations. Loss of heterozygosity in the p53 locus was rare in both groups. The p53 mutations detected in hereditary tumors included rare single nucleotide deletions and unusual double-base substitutions compared to the typical ultraviolet light induced missense mutations found in sporadic tumors. Careful microdissection of individual tumors revealed genetically linked subclones with different p53 and/or patched genotype providing an insight on time sequence of genetic events. The high frequency and co-existence of genetic alterations in the patched and p53 genes suggest that both these genes are important in the development of basal cell cancer.

Error rate and specificity of human and murine DNA polymerase η

We describe here the error specificity of mammalian DNA polymerase η (pol η), an enzyme that performs translesion DNA synthesis and may participate in somatic hypermutation of immunoglobulin genes. Both mouse and human pol η lack intrinsic proofreading exonuclease activity and both copy undamaged DNA inaccurately. Analysis of more than 1500 single-base substitutions by human pol η indicates that error rates for all 12 mismatches are high and variable depending on the composition and symmetry of the mismatch and its location. pol η also generates tandem base substitutions at an unprecedented rate, and kinetic analysis indicates that it extends a tandem double mismatch about as efficiently as other replicative enzymes extend single-base mismatches. This ability to use an aberrant primer terminus and the high rate of single and double-base substitutions support the idea that pol η may forego strict shape complementarity in order to facilitate highly efficient lesion bypass. Relaxed discrimination is further indicated by pol η infidelity for a wide variety of nucleotide deletion and addition errors. The nature and location of these errors suggest that some may be initiated by strand slippage, while others result from additional mechanisms.

XPA protein alters the specificity of ultraviolet light–induced mutagenesis in vitro

Studies of ultraviolet (UV) light mutagenesis have demonstrated mutations at common sites in the target genes of shuttle vector plasmids replicated in cultured cells or by cellular extracts. The reasons for the specific pattern of mutagenesis are largely unknown. We have examined the specificity of UV-induced mutagenesis by replicating plasmid pLS189, irradiated with 40 J/m(2) UVC or unirradiated, in either xeroderma pigmentosum group A (XP-A) or HeLa cellular extracts. The XP-A extract displayed slightly lower replication ability, but produced a higher mutant frequency, compared to that of HeLa extract. Use of irradiated plasmid inhibited replication by an average of 63% and increased the mutant frequency by an average of 16.7-fold. Analysis of mutation spectra revealed nonrandom patterns of mutagenesis that differed significantly between HeLa and XP-A extracts. In comparison to HeLa extract, replication in XP-A extract resulted in lower frequencies of GC --> AT transitions and tandem double-base substitutions, and a higher frequency of deletions. Replication in HeLa extract produced hotspots at positions 100, 108, and 156 that were not produced by XP-A extract. Furthermore, XP-A extract produced hotspots at positions 124, 133, and 164, sites not characteristic of previous UV-induced mutagenesis studies using XPA-expressing cells. Addition of purified XPA protein to reactions containing XP-A extract altered each of these parameters, including loss of the hotspots at positions 124 and 133, to yield a more HeLa-like spectrum. These results indicate a previously uncharacterized role of the XPA protein in influencing the specificity of UV-induced mutagenesis during DNA replication.

Suppression of tumor necrosis factor (TNF-α) gene expression by prostaglandin E2. Role of early growth response protein-1 (Egr-1)

We examined the mechanism by which PGE2suppresses the expression of TNF-α in human macrophages and synovial fibroblasts. Prostaglandin E2increased, in a time and dose-dependent (EC5075±15ng/ml, mean±SD) fashion, the expression and synthesis of Egr-1/Krox24 as judged by Northern blotting and electrophoretic mobility gel-shift analysis, respectively. In human macrophagic THP-1 cells, rhIL-17 increased promoter activity by 7.6±0.35-fold over controls, an effect that was abrogated in a dose-dependent fashion by coincubations with PGE2(IC5025±4ng/ml). An intact Egr-1/Krox-24 enhancer sequence in the TNF-α promoter region was essential for the latter PGE2-dependent inhibitory effect as double base substitutions (GC→TT) in the sequence curtailed promoter response to PGE2. Overexpression of two dominant negative Egr-1/Krox-24 constructs in THP-1 cells considerably diminished the inhibitory effects of PGE2on rhIL-17-induced TNF-α mRNA expression. We conclude that PGE2inhibits induced TNF-α expression in target cells through an Egr-1/Krox-24 mediated signaling process.

Base substitution spectra of nalidixylate resistant mutations induced by monochromatic soft X and 60Co gamma-rays in Bacillus subtilis spores.

Bacillus subtilis spores were exposed to three types of photons, monochromatic soft X-rays with the energy corresponding to the absorption peak of phosphorus K-shell electron (2,153 eV) and with the slightly lower energy (2,147 eV), and 60Co gamma-rays. From the irradiated spores, 233 mutants exhibiting nalidixic acid resistance were isolated, and together with 94 spontaneous mutants, the sequence changes in the 5'-terminal region of the gyrA gene coding for DNA gyrase subunit A were determined. Among eighteen alleles of the gyrA mutations, eight were single-base substitutions, nine were tandem double-base substitutions, and one was a double substitution skipping a middle base pair. About 6% of the radiation-induced mutations were tandem double-base substitutions, whereas none was observed among the spontaneous ones. Among spontaneous mutations, A:T and G:C pairs were equally subjected to mutations, whereas the substitutions from G:C pairs and those to A:T pairs predominated among those induced with soft X-rays. The peak-energy X-rays were more effective in killing and causing mutations than the low-energy X-rays, however, there seemed no base-change events uniquely attributable to phosphorus K-shell absorption.

Role of nucleotide sequences of loxP spacer region in Cre-mediated recombination.

The Cre recombinase mediates precise site-specific recombination between a pair of loxP sequences through an intermediate containing Holiday junction. The recombination junction in the loxP sequence is located within the asymmetric 8-nucleotide spacer region. To examine the role of each nucleotide sequence of the spacer region in the recombination process, we synthesized a complete set of 24 loxP spacer mutants with single-base substitutions and 30 loxP spacer mutants with double-base substitutions. Each synthesized loxP mutant was ligated at both ends of a linear DNA or to one end of a DNA-containing wild-type loxP at the other end and their recombination efficiencies were analyzed with an in vitro system. The sequence identity of the right two nucleotides and left four nucleotides in the central six bases of the spacer region was found to be essential for formation and resolution, respectively, of the intermediate product. Furthermore, even when homology was maintained, the recombination efficiencies were lower than that of wild-type loxP and varied among mutants. Based on this knowledge, we identified two loxP mutants with double-base substitutions, mutants 5171 and 2272, which recombine efficiently with an identical mutant but not with the other mutant or wild-type loxP.

Neocarzinostatin-induced mutations at the hprt locus in exponentially growing CHO cells, compared with spontaneous mutations

Spontaneous mutations and neocarzinostatin-induced mutations were investigated in the hypoxanthine-guanine phosphoribosyltransferase (hprt) gene in exponentially growing Chinese hamster ovary cells. Neocarzinostatin (NCS) treatment caused an 4.5-fold increase in mutation frequency. Analysis by multiplex polymerase chain reaction and sequencing of hprt cDNA revealed that spontaneous mutations in this system were characterized by a relatively high rate of large deletions and double-base substitutions, and a low rate of splice mutations compared with data reported in fibroblastic cell lines. NCS hardly affected this spectrum of spontaneous mutations, which indicates the rare incidence of NCS-specific change in the exponential growth phase. This is in contrast to aprt gene mutations reported in plateau phase cells in which base substitutions occur preferentially at sites affected by NCS. These results suggest that differences in the loci assayed or in the processes involved in mammalian mutagenesis in the exponential growth phase and the plateau phase may be the source of the different results.

Mutagenic Specificity of Ultraviolet Light in the tonB Gene on the Chromosome of Escherichia coli uvrA Cells

We have analyzed the DNA sequence changes in a total of 60 ultraviolet-induced mutations in the endogenous tonB gene of Escherichia coli uvrA strain. Of the mutations 82% were base substitutions among which G:C→A:T transition predominated. Three GG→AA tandem double-base substitutions, which are thought to originate from UV damage, were also observed. The sites where base substitutions occurred were correlated with sequences of adjacent pyrimidines, indicating mutation-targeted UV photoproducts. G:C→A:T transition in the tonB gene mutation can be exclusively observed at either the 3′ side of the TC site which is on the template for the lagging strand of DNA replication or the 5′/3′ sides of the CC site on the template for the leading strand. We hypothesize that this extreme strand specificity is due to a difference in fidelity of DNA replication of the leading and the lagging strand.

Improvement of thermal stability of proteins by evolutionary molecular engineering.

Thermal stability of proteins can be improved by evolutionary molecular engineering, without the knowledge about 3D structures of the proteins nor the principle that governs the protein stability. The method will provide a way of improving the thermal stability of industrially important proteins. It can also be used as a tool to study protein stability. There are a substantial number of examples now. Our recent study suggested that it is easy to obtain thermo-stabilized proteins by single-base substitution. Although, it requires further improvement of the method to obtain mutants with double-base substitution within a codon.

Molecular analysis of Salmonella hisG428 ochre revertants for rapid characterization of mutational specificity.

The Salmonella typhimurium tester strains TA104 and TA102 were developed primarily to aid in the detection of oxidative mutagens and other agents that react preferentially with AT base pairs. Reversion of prototrophy of strains harboring the hisG428 ochre allele can occur by (i) any of seven single base substitutions or (ii) several tandem double base substitutions at the ochre codon, (iii) in-frame deletions removing all or part of the ochre codon or (iv) mutations at several distinct tRNA extragenic suppressor loci. We have used allele-specific oligonucleotide probes and DNA sequence analysis to characterize 625 revertants of strain TA104 (hisG428, rfa, DeltauvrB/pKM101) arising spontaneously or after treatment with methyl methane-sulfonate, glyoxal, streptonigrin or angelicin with UVA radiation. The reversion profiles obtained from these analyses distinguished readily each of the mutagen-treated populations from one another and from spontaneously derived revertants. Both GC and AT base pair-specific revertants were observed. Molecular analyses of S. typhimurium hisG428 revertants permitted rapid assessment of base pair substitution specificity of mutagens, especially the detection of AT base pair substitutions not recovered in strains carrying the complementary hisG46 allele.

Analysis of mutations induced by replication of UV‐damaged plasmid DNA in hela cell extracts

We have used an SV40-based shuttle vector, pZ189, to investigate the capacity of HeLa cell extracts to reproduce the in vivo process of mutation fixation. We showed previously that when UV-irradiated pZ189 is replicated in these extracts, bypass of UV photoproducts occurs, resulting in base substitution mutations in the supF gene of the vector. Here we report the DNA sequence characterization of a collection of 60 of these UV-induced mutants. Most of the mutations observed are single or tandem double base substitutions at dipyrimidine sites; of these, approximately 90% are G:C-->A:T transitions. Mutations are observed predominantly at a few sites, in particular at positions 155 and 156 in the supF sequence. No dramatic differences in the mutation spectrum were observed when the orientation of the supF gene was reversed with respect to the SV40 origin of replication, suggesting that mutation fixation occurs similarly on both the leading and the lagging strands for DNA replication. Generally, the mutational hot spots observed in vitro are at the same sites as those observed when UV-irradiated pZ189 was passaged in human or monkey cells in culture. Thus, it appears that the replication and mutagenesis of UV-damaged templates in HeLa cell extracts accurately reflects these processes in the intact cell.

Polymerase chain reaction-directed DNA sequencing of bleomycin-induced "nondeletion"-type, 6-thioguanine-resistant mutants in Chinese hamster ovary cell derivative AS52: effects of an inhibitor and a mimic of superoxide dismutase.

Bleomycin-induced, 6-thioguanine-resistant, "non deletion" mutants pretreated with or without either TRIEN (triethylenetetramine), a superoxide dismutase (SOD) inhibitor, or TEMPOL (4-hydroxy-2,2,6,6-tetramethylpiperidine-1-oxyl), a SOD mimic, were analyzed by polymerase chain reaction (PCR)-directed DNA sequencing in a Chinese hamster ovary (CHO) cell derivative, AS52. Among the 23 bleomycin-induced mutants, six have 3-bp 5'-TGA-3' deletions in the region of 366-371, five have single-base deletions, seven have base substitutions, three have insertions, and two have possible translocations. Among the 16 bleomycin-induced mutants pretreated with TRIEN, six have the 5'-TGA-3' deletion (366-371), two have single-base deletions, one has a 13-bp deletion, four have single-base substitutions, one has a double-base substitution, and two have insertions. Among the 17 bleomycin-induced mutants pretreated with TEMPOL, six have the same TGA deletions, two have single-base deletions, two have single-base insertions, four have single-base substitutions, one mutant has a 12-bp deletion, one has a 13-bp deletion, and one mutant shows no detectable change in its coding region in the DNA sequence. A possible shift from a ROS-mediated mutational spectrum to a spontaneous mutational spectrum by TRIEN further indicates that reactive oxygen species play an important role in bleomycin mutagenesis in mammalian cells.

Salmonella typhimurium strain TA100 differentiates several classes of carcinogens and mutagens by base substitution specificity

The mutational specificity of N-methylnitrosourea (MNU), nitrosoguanidine (MNNG), methyl methanesulfonate (MMS), sodium azide (NaN3), 4-nitroquinoline oxide (4NQO), benzo[a]pyrene (BP), nitrofurantoin (NF), aflatoxin B1 (AFB1), adriamycin (ADM) and UVA-activated angelicin in Salmonella typhimurium strain TA100 has been examined using allele-specific oligonucleotide hybridization and DNA sequence analyses. These ten mutagens produced five unique classes of reversion spectra, distinct from spontaneous, or the previously characterized 5-azacytidine, ultraviolet light (UV), 8-methoxypsoralen plus UVA (PUVA) and 60Co-induced mutation spectra. For example, 90% of MNU and MNNG-induced mutations in strain TA100 revertants were G:C-->A:T transitions with the majority (82%) occurring in the first position of the CCC codon. In contrast, NaN3 preferentially induced G:C-->A:T transitions at the second codon position (78%). Although MMS, NQO, BP, NF, ADM and AFB1 induced primarily G:C-->T:A transversions (73-86%), these mutagens fall into two classes based on site preference: NF and AFB1 yielded almost exclusively position two transversions (69-78%) whereas ADM, NQO, BP and MMS exhibited a two-fold preference for site 2 over site 1 (on average 52% versus 22%). Angelicin photomutagenesis resulted in the recovery of G:C-->A:T and G:C-->T:A mutations at both codon positions in roughly equal proportions (approximately 20-25% each). Approximately 1% of the mutagen-induced revertants occurred via extragenic tRNA suppressor mutations, while 1% were multiple (usually tandem double) base substitutions. Ultraviolet mutagenesis experiments demonstrated that tandem base substitutions are promoted by pKM101-encoded mucAB gene products. A comparison of the mutagenic specificity derived for several carcinogens in hisG46 with the responses of several eukaryotic gene targets (e.g. HPRT, aprt, supF) revealed a high concordance between these targets. Thus, the Salmonella hisG46 locus provides a rapid, simple system for determining base substitution specificity and for studying mechanisms of mutagenesis.

Adenovirus VARNA1 gene B block promoter element sequences required for transcription and for interaction with transcription factors.

We constructed mutants with a deletion of either half of the 18 base-pair B block palindrome in the VARNA1 gene, mutants with different intra-palindromic spacings, a complete set of mutants with single base substitutions, and mutants with double and triple base substitutions in the palindrome. The transcription efficiencies of these mutants were determined in human KB cell-free cytoplasmic S100 extracts. The relative competing strength of each mutant, as determined by a sequential competition experiment, was used to assess each mutant's ability to sequester factors into formation of a stable preinitiation complex. The ability of each mutant to assemble transcriptionally active preinitiation complexes was also determined by direct transcription of the isolated complexes. Finally, the ability of each mutant to interact with the transcription factor(s) TFIIIC and form a distinct gel-resolved complex was also determined. From the results of the above assays, we concluded that the two seemingly identical halves of the palindrome did not contribute equally to transcription, or to assembly of the functional preinitiation complex, nor to interaction with TFIIIC. The anterior half (B1) of the B block palindrome, which is proximal to the A block promoter element, played a stronger role in transcription and in assembly of the functional preinitiation complex than the posterior half (B2) of the palindrome. Consistent with this observation, the point mutations in four base-pairs, GTTC, from +60 to +63 in the anterior half of the B block palindrome, has the most severe effect on transcription. In contrast, we showed that the central sequence and the posterior half (B2) played a stronger role than the anterior half (B1) of the B block palindrome in the interaction of the promoter with TFIIIC. This was corroborated by the observation that base substitutions in the central four base-pair sequence of the palindrome, TCGA, from +62 to +65, had the most severe effect on interaction with TFIIIC, and that mutations in most of the sequences in the posterior half of the B block palindrome had more drastic effects than mutations in the anterior half of the palindrome in this interaction. Furthermore, the spacing between the two halves of the B block palindrome had a drastic effect on the overall transcription efficiency and the interaction of the promoter with TFIIIC, suggesting that the interaction between the two halves of the B block palindrome is not only essential, but also synergistic for the interaction with TFIIIC as well as the assembly of a transcriptionally active preinitiation complex and efficient transcription.

The requirement for a 5' stem-loop structure in brome mosaic virus replication supports a new model for viral positive-strand RNA initiation

Sequences with strong similarity to internal control regions 1 and 2 (ICR1 and -2; A and B boxes) of tRNA genes are found at the 5' termini of the genomic RNAs of brome mosaic virus (BMV) and other plant viruses. The functionality of these motifs was studied by introducing point mutations into the ICR2-like sequence of pRNA-2 M/S, a BMV RNA-2 deletion mutant that replicates in the presence of RNAs-1 and -2 but does not encode a functional viral protein. The accumulation of positive-strand progeny from pRNAs bearing single and double base substitutions was 70 to 91% lower than that of the wild type, while the addition of single bases at position 8 of this motif reduced replication by 80%. These dramatic decreases in positive-strand synthesis paralleled decreases in transcription seen (C. Traboni, G. Ciliberto, and R. Cortese, Cell 36:179-187, 1984) from a tRNAPro gene containing similar mutations, suggesting comparable functions for the ICR regions in protein factor binding and demonstrating that a wild-type composition of the virus ICR2-like motif is required for proper RNA replication. Substitutions introduced at bases surrounding the ICR2 motif yielded levels of pRNA replication that differed, depending on the maintenance of a putative 5' stem-loop structure in the positive strand of the viral genome. Mutations disrupting this positive-strand stem-loop while maintaining the integrity of the complementary negative-strand structure reduced pRNA replication by 85 to 97%. In contrast, disruption of the negative-strand structure while maintaining the positive-strand stem-loop did not reduce pRNA replication. Similar positive-strand structures can be predicted to form from 5' sequences of cucumber mosaic virus (strain Q) and cowpea chlorotic mottle virus RNAs-1 and -2, supporting the concept that such structures comprise an integral part of virus genomic positive-strand promoters. The requirement of a stem-loop structure present on the positive-strand provided the basis for a new model describing how these sequence and structural elements act in the production of virus positive-strand RNA.

Mutational specificity of chromium(VI) compounds in the hprt locus of Chinese hamster ovary-K1 cells.

Chromium(VI) compounds exert their genotoxicity and mutagenicity by complex metabolic reducing pathways that generate a variety of reactive forms of chromium and free radicals. To investigate the molecular nature of chromium-induced mutations, we characterized the entire coding region of the hypoxanthine (guanine) phosphoribosyltransferase (hprt) gene of 27 independent mutants derived from chromium(VI) oxide (CrO3)-treated Chinese hamster ovary-K1 cells, by direct sequencing of PCR-amplified cDNA. Among these mutants, 10 consisted of single base substitutions, five contained two base substitutions, one had four base substitutions, six were splicing mutations, and five exhibited single base pair insertions or deletions. All of the base substitutions and most of the frameshift mutations observed were located at A/T-rich sequences. More than 90% of the base substitutions (22/24) occurred in A.T base pairs. Among them, T-->A and T-->G transversions (18/22) predominated. The mutational hotspots for single and double base substitutions were the 3' thymidine of 5'PuT and thymidines of 5'ATTT sequences respectively. This mutational specificity was also observed in CHO-K1 cells treated with two other chromium(VI) compounds, namely K2Cr2O7 and PbCrO4. Strand bias was noticed in chromium mutagenicity, since 77% of T base substitutions occurred on the non-transcribed strand. This highly sequence-specific mutation spectrum suggests that a particular form of chromium may directly interact with DNA at these hotspot sequences.

In Vivo Analysis of the Saccharomyces cerevisiae Centromere CDEIII Sequence: Requirements for Mitotic Chromosome Segregation

In the yeast Saccharomyces cerevisiae, the complete information needed in cis to specify a fully functional mitotic and meiotic centromere is contained within 120 bp arranged in the three conserved centromeric (CEN) DNA elements CDEI, -II, and -III. The 25-bp CDEIII is most important for faithful chromosome segregation. We have constructed single- and double-base substitutions in all highly conserved residues and one nonconserved residue of this element and analyzed the mitotic in vivo function of the mutated CEN DNAs, using an artificial chromosome. The effects of the mutations on chromosome segregation vary between wild-type-like activity (chromosome loss rate of 4.8 x 10(-4)) and a complete loss of CEN function. Data obtained by saturation mutagenesis of the palindromic core sequence suggest asymmetric involvement of the palindromic half-sites in mitotic CEN function. The poor CEN activity of certain single mutations could be improved by introducing an additional single mutation. These second-site suppressors can be found at conserved and nonconserved positions in CDEIII. Our suppression data are discussed in the context of natural CDEIII sequence variations found in the CEN sequences of different yeast chromosomes.

In vivo analysis of the Saccharomyces cerevisiae centromere CDEIII sequence: requirements for mitotic chromosome segregation.

In the yeast Saccharomyces cerevisiae, the complete information needed in cis to specify a fully functional mitotic and meiotic centromere is contained within 120 bp arranged in the three conserved centromeric (CEN) DNA elements CDEI, -II, and -III. The 25-bp CDEIII is most important for faithful chromosome segregation. We have constructed single- and double-base substitutions in all highly conserved residues and one nonconserved residue of this element and analyzed the mitotic in vivo function of the mutated CEN DNAs, using an artificial chromosome. The effects of the mutations on chromosome segregation vary between wild-type-like activity (chromosome loss rate of 4.8 x 10(-4)) and a complete loss of CEN function. Data obtained by saturation mutagenesis of the palindromic core sequence suggest asymmetric involvement of the palindromic half-sites in mitotic CEN function. The poor CEN activity of certain single mutations could be improved by introducing an additional single mutation. These second-site suppressors can be found at conserved and nonconserved positions in CDEIII. Our suppression data are discussed in the context of natural CDEIII sequence variations found in the CEN sequences of different yeast chromosomes.

High-efficiency oligonucleotide-directed plasmid mutagenesis.

A number of single- and double-base substitutions have been introduced into either the polylinker region or the lacZ gene in the plasmid vector pUC19. The efficiencies of these changes upon transfection of TG-1 bacterial cells were generally 70-80%. A strategy has been devised by which the wild-type DNA can be selectively destroyed. It is primarily based on the resistance of phosphorothioate internucleotide linkages to some restriction enzymes. A mismatch oligonucleotide is introduced into a gapped region and the gap is filled using three deoxynucleoside 5'-triphosphates and one deoxynucleoside 5'-[alpha-thio]triphosphate. Reaction with a restriction enzyme that is unable to hydrolyze phosphorothioates ensures that the DNA containing the mismatch oligonucleotide is only nicked. Concomitantly, the DNA that does not contain the desired mutation is linearized. Subsequent reactions with an exonuclease and DNA polymerase I yield mutant homoduplex DNA for transfection.