2014 Annual Science Report

University of Illinois at Urbana-Champaign Reporting  |  SEP 2013 – DEC 2014

Project 7 Control of Evolvability and Chromosomal Rearrangement by Stress

Project Summary

Evolution of the genome happens predominantly by gross chromosomal rearrangements (GCR), usually involving non-homologous recombination, and point mutation or single nucleotide variation (SNV). GCR re-assorts domains and regulatory functions, and frequently changes gene copy number, allowing further evolution. SNV changes coding sequences, modifying the properties of encoded macromolecules and their regulation. We are able to measure both SNV and GCR in the same assay in Escherichia coli, and have established that both are dramatically up-regulated in response to stress. Through many year’s work, we have learnt many details of these mechanisms, but some of the outstanding questions are crucial to understanding the significance of stress in evolution, notably the signaling and execution of the pathway from stress to mutation, and how the decision is made between SNV and GCR. This project aims to answer these questions.

4 Institutions
3 Teams
0 Publications
1 Field Site
Field Sites

Project Progress

We have screened for suppressors of the requirement for the RpoS stress-response in chromosomal rearrangement by using a transposon with outward-facing promoters that activates genes in the absence of stress. This could tell which stress-induced function is required for chromosomal rearrangement. Most of the suppressors isolated have activated parts of cryptic prophages: bacterial viruses that have lost their ability to lyse the host cell, and so the genetic material lives on in an inactive form in the host cell’s DNA. Four of these suppressors have resulted in the activation of alternative recombination through the RecET pathway. We are testing whether these have bypassed the requirement for stress, rather than suppressing it. The most interesting isolate at present has the transposon disrupting the gene dgt, which encodes a deoxyguanidine triphosphohydrolase. This protein regulates both deoxyguanidine triphosphate and thymidine triphosphate pools, and is stress-induced. Other work has shown that DNA nucleotide pool content modulates mutation rates, but the mechanism for this is unknown. Work on validating and explaining the effect of dgt is ongoing.

Nucleoid-associated proteins (NAPs) both modify the structure of bacterial chromatin and act as global regulators of gene expression. We have completed a study of the involvement of nucleoid-associated proteins in stress-induced mutation, finding that the effects of four NAPs are explained by factors other than their ability to modify chromatin structure. Two of the four modify the availability of oxygen radicals (see below), another regulates the expression of the DNA damage response, specifically the availability of trans-lesion polymerases, and the fourth is required for maintenance of the conjugative plasmid that carries the assay system. Thus we find no evidence from these experiments that chromatin remodeling regulates mutagenicity under stress.

We have established the unexpected result that both spontaneous mutation and chromosomal rearrangement require 8-oxoguanine incorporation into DNA. We already knew that double-strand breaks are required, but the requirement for oxidative DNA lesions is not via double-stand break formation. Using genetic control of responses to DNA damage via oxidation and using chemical scavengers of reactive oxygen, we show beyond doubt that there is a requirement for persistence of incorporated 8-oxoguanine. When we over-express enzymes that specifically excise 8-oxoguanine from DNA, point mutation formation is deficient. Hence the lesion must remain in the DNA to be effective, and so mutation is presumably formed by induction of trans-lesion synthesis. We have not yet shown that chromosomal rearrangement requires the lesion to remain in DNA, but are setting up a new assay to test that. Because reactive oxygen will not be present in all environments, we tested whether other DNA lesions could substitute for 8-oxoguanine. Suppressing reactive oxygen with thiourea, a condition that allows no mutation, we have shown that we can restore mutation with alkylating agents or with ultraviolet irradiation. Genetic tests indicate that these treatments restored the same stress-dependent pathway that we study. We have formulated a model in which base lesions in DNA are required to pause the highly processive replicative polymerase, allowing alternative error-prone polymerases to attain the active site on the replisome.

Figure 1
Figure 1 - Removal of reactive oxygen species (ROS) by activation of radical scavenging enzymes reduces stress-induced mutation. The graph shows the effect on mutation in the Lac assay of constitutive activation of the sox regulon encoding superoxide dismutases and the oxy regulon encoding catalases using mutations that activate these regulons constitutively. Both these scavenge oxygen radicals. We find that this effect applies to both point mutation and amplification (not shown). This confirms our previous report based on chemical ROS scavengers that ROS are required for stress-induced mutation. This plot and all similar plots show the mean and standard error of the mean of at least three experiments. WT: wild type.

Figure 2
Figure 2 - ROS are required for point mutagenesis, but not for the induction of double-strand breaks. Here we use a Tet assay to determine that over-expression of KatG, a catalase that breaks down hydroxyl radicals, causes a strong reduction in point mutation. The Tet assay does not measure gross chromosomal rearrangement. In the Tet assay, an enzymatically induced double-strand break on the E. coli chromosome induces stress-induced mutation in a tet gene nearby. This shows that the need for ROS for mutation operates on the chromosome as well as on the conjugative plasmid. At the same time, this experiment shows that the role of ROS is not the provision of the double-strand breaks shown previously to be required for stress-induced mutation, because providing enzymatic double-strand breaks does not overcome the inhibition caused by reduction in ROS (compare pVector with pKatG). DSB: double-strand break.

Figure 3
Figure 3 - Reactive oxygen species are not required for activation of the SOS DNA damage response. The only role of the DNA damage response in stress-induced mutation is to activate error-prone DNA polymerases, specifically DinB in this assay. Here we show that constitutive expression of DinB (DinBo/c) does not overcome the inhibitory effect of thiourea (TU) on stress-induced mutation and therefore does not function by inducing SOS.

Figure 4
Figure 4 - The function of ROS in stress-induced mutation is not the upregulation of the RpoS general stress-response regulator. RssB regulates RpoS negatively by marking it for destruction. In the absence of RssB, cells have high levels of RpoS at all stages of growth so that RpoS is present without stress induction. If ROS acted by inducing RpoS, we would see that deletion of rssB restored mutation when ROS have been removed by thiourea. The figure shows that deletion of rssB does not counter the effects of thiourea, showing that enhanced expression of RpoS is not sufficient to restore mutation when ROS are absent.

Figure 5
Figure 5 - The function of ROS is not to over-saturate mismatch repair. It was possible that the role of ROS in stress-induced mutation was to overload the mismatch repair system so that more mutations could persist. However, when we over-express MutL (MutLo/e), which performs the first and limiting step in mismatch repair, and treat with thiourea, we see that the effects are additive, showing that the two effects operate in different pathways and therefore that ROS do not function by saturation of mismatch repair.

Figure 6
Figure 6 - The role of ROS is to provide 8-oxo guanine in DNA. Using the Tet assay and over-expression of mutM and mutT from plasmids, we show that most point mutation stems from the persistence of damaged guanine incorporated into DNA from the nucleotide pool. MutT sanitizes the pool by breaking down 8-oxo guanine (8-oxoG), and over-expression of MutT causes a strong inhibition of stress-induced point mutation, showing that mutation is caused by incorporation into DNA of 8-oxodG from the nucleotide pool. MutM is the glycosylase that excises 8-oxoG from the DNA as the first step in base excision repair (BER). MutM is highly specific for 8-oxodG in DNA. We see that removal of 8-oxodG from DNA reduces mutation, showing that, to be effective, the base lesion must remain in the DNA.

Figure 7
Figure 7 - Other base lesions can substitute for 8-oxoG. Using a level of TU that prevents almost all mutation, we are able to restore stress-induced mutation by treatment with other mutagens. Both MMS and ultraviolet light (UV) induce base damage that is different from 8-oxoG, but these agents are able to lead to mutation like 8-oxoG. We ascertained that the restored mutation has the same characteristics and genetic requirements as conventional stress-induced mutation (not shown). This result generalizes the mechanism to other environments, including those that lack oxygen, because it can depend on base damage other than oxidative damage.