Insights into mutagenesis using Escherichia coli chromosomal lacZ strains that enable detection of a wide spectrum of mutational events

Abstract

Strand misalignments at DNA repeats during replication are implicated in mutational hotspots. To study these events, we have generated strains carrying mutations in the Escherichia coli chromosomal lacZ gene that revert via deletion of a short duplicated sequence or by template switching within imperfect inverted repeat (quasipalindrome, QP) sequences. Using these strains, we demonstrate that mutation of the distal repeat of a quasipalindrome, with respect to replication fork movement, is about 10-fold higher than the proximal repeat, consistent with more common template switching on the leading strand. The leading strand bias was lost in the absence of exonucleases I and VII, suggesting that it results from more efficient suppression of template switching by 3' exonucleases targeted to the lagging strand. The loss of 3' exonucleases has no effect on strand misalignment at direct repeats to produce deletion. To compare these events to other mutations, we have reengineered reporters (designed by Cupples and Miller 1989) that detect specific base substitutions or frameshifts in lacZ with the reverting lacZ locus on the chromosome rather than an F' element. This set allows rapid screening of potential mutagens, environmental conditions, or genetic loci for effects on a broad set of mutational events. We found that hydroxyurea (HU), which depletes dNTP pools, slightly elevated templated mutations at inverted repeats but had no effect on deletions, simple frameshifts, or base substitutions. Mutations in nucleotide diphosphate kinase, ndk, significantly elevated simple mutations but had little effect on the templated class. Zebularine, a cytosine analog, elevated all classes.

Figure 1.—

Mutational reporter design. Design is based on Cupples and Miller (1989) and Cupples et al. (1990) for detection of specific base substitutions (A) and frameshift mutations (B) by lacZ reversion.

Figure 2.—

Mechanisms for mutagenesis in quasipalindromes involving replication template switching. (A) Sequence of an imperfect inverted repeat (boxed nucleotides will form base pairs in the quasipalindrome.) (B) Intramolecular template switching. Hairpin formation and templated mutation can occur by a template switch to copy one arm of the repeat, leading to perfection of the inverted repeat. A second template switch to resume normal replication produces the mutational event (shown in boldface type). Note that hairpin formation on leading and lagging strands leads to mutations on different sides of the inverted repeat.

Figure 3.—

Quasipalindrome (QP) mutational reporter design. Shown are hairpin structures for the 5′ sense strand. The orientation of lacZ is such as the replication fork proceeds in the antisense direction relative to lacZ, as shown by the arrow. Reporters QP3 and QP4 have a silent mutation, nucleotide A201G (G shown in boldface type), that strengthens the hairpin by 1 bp. A GC dinucleotide has been inserted at the 5′ side (QP3) or 3′ side (QP4) that shifts lacZ out of frame and generates a unique SacII site. Template-switch mutagenesis will remove the GC, generating an intact lacZ gene (LacZ*) and may also produce comutations G to C and T to A at nucleotides 189 and 187 for QP3 and nucleotides 218 and 220 for QP4. Second generation reporters QP5 and QP6 incorporate nucleotide T187A and G189C mutations (that produce a W63S amino acid change that does not interfere with lacZ function) and are mutated to Lac⁻ by insertion of a TCTC sequence, generating either an EarI site (QP5) or PvuII site (QP6), shown in boldface type. Templated reversion restores Lac⁺, with concomitant loss of the EarI or PvuII site.

Figure 4.—

Alternative hairpin structures that template reversion in the lacZ–QP3 reporter. Hairpin 1 represents the most frequent reversion event for QP3, generating a −GC frameshift (and loss of the SacII site) with comutations, nucleotides T187A and G189C. More infrequently, hairpin 2 formation templates a +CCGG mutation, restoring LacZ function without loss of the SacII site.

Figure 5.—

Colony papillation assay for reversion. Supplementation of growth medium with lactose allows revertants that occur during the growth of the colony to outgrow as blue papillae, in the presence of X-gal and IPTG. Shown are wild-type and Exo⁻ (xseA xonA) strains carrying reporters QP3, QP4, and QP5, where differences in reversion rate are clearly visible.

Figure 6.—

Deletion mutational reporter. An eleven nucleotide sequence was duplicated, inactivating lacZ. A deletion event, produced by “slippage” of the nascent strand on its template, gives rise to lacZ reversion.

Figure 7.—

Comparison of lacZ reversion rates on the chromosome vs. F′ lac plasmid. Reversion rates of specific mutation events measured on the chromosome (shaded bars) compared to the same alleles carried on F′ lac (solid bars, data from Miller et al. 2002) in wild-type strains (top) or mismatch repair defective, mutS strains (bottom).

Figure 8.—

Reversion rates for selected reporter strains by alterations in nucleotide pools. (A) Rates of an 11-bp deletion, quasipalindrome-associated mutation (QP5), AT to TA transversion, and +1G frameshift mutations in wild-type cells (bars with light shading) compared to ndk-deficient strains (bars with dark shading). (B) Effects of hydroxyurea on deletion of an 11-bp tandem repeat and reversion rates for the quasipalindrome reporter QP5 and GC→AT transition mutation and +1G frameshift. Mock-treated cells (bars with light shading) were compared to strains treated with 3 mM (bars with medium shading) and 4 mM HU (bars with dark shading). (C) Effects of zebularine on deletion of an 11-bp tandem repeat and reversion rates for the quasipalindrome reporter QP5. Mock-treated cells (bars with light shading) were compared to strains treated with 15 μg/ml (bars with medium shading) or 50 μg/ml zebularine (bars with dark shading).