Repair system for noncanonical purines in Escherichia coli

Abstract

Exposure of Escherichia coli strains deficient in molybdopterin biosynthesis (moa) to the purine base N-6-hydroxylaminopurine (HAP) is mutagenic and toxic. We show that moa mutants exposed to HAP also exhibit elevated mutagenesis, a hyperrecombination phenotype, and increased SOS induction. The E. coli rdgB gene encodes a protein homologous to a deoxyribonucleotide triphosphate pyrophosphatase from Methanococcus jannaschii that shows a preference for purine base analogs. moa rdgB mutants are extremely sensitive to killing by HAP and exhibit increased mutagenesis, recombination, and SOS induction upon HAP exposure. Disruption of the endonuclease V gene, nfi, rescues the HAP sensitivity displayed by moa and moa rdgB mutants and reduces the level of recombination and SOS induction, but it increases the level of mutagenesis. Our results suggest that endonuclease V incision of DNA containing HAP leads to increased recombination and SOS induction and even cell death. Double-strand break repair mutants display an increase in HAP sensitivity, which can be reversed by an nfi mutation. This suggests that cell killing may result from an increase in double-strand breaks generated when replication forks encounter endonuclease V-nicked DNA. We propose a pathway for the removal of HAP from purine pools, from deoxynucleotide triphosphate pools, and from DNA, and we suggest a general model for excluding purine base analogs from DNA. The system for HAP removal consists of a molybdoenzyme, thought to detoxify HAP, a deoxyribonucleotide triphosphate pyrophosphatase that removes noncanonical deoxyribonucleotide triphosphates from replication precursor pools, and an endonuclease that initiates the removal of HAP from DNA.

FIG. 1.

Relative HAP sensitivity of E. coli strains mutant for suspected purine base analog-metabolizing enzymes. Mid-log-phase cells were exposed to various concentrations of HAP for 1 h at 37°C. Data were recorded as percent survival at the various HAP concentrations. Each data point represents one bacterial culture. Symbols: ○, CSH106 (wild type); •, NEB9 (moa); □, NEB10 (moa rdgB); ▪, NEB19 (moa nfi); ▵, NEB41 (moa rdgB nfi); ▴, NEB1 (rdgB).

FIG. 2.

(A) HAP-induced mutagenesis frequencies of E. coli strains mutant for suspected purine base analog-metabolizing enzymes. Mid-log-phase cells were exposed to various concentrations of HAP for 1 h at 37°C. Results for wild type and rdgB nfi were indistinguishable. (B) HAP-induced mutagenesis frequencies of NEB10 (moa rdgB) at very low HAP concentrations. Each data point represents one bacterial culture. Symbols: ○, CSH106 (wild type); •, NEB9 (moa); □, NEB10 (moa rdgB); ▪, NEB19 (moa nfi); ▵, NEB41 (moa rdgB nfi); ▴, NEB32 (rdgB nfi).

FIG. 3.

Recombination frequencies of E. coli strains mutant for suspected purine base analog-metabolizing enzymes upon HAP exposure. Early-log-phase cells were exposed to various concentrations of HAP for 1 h at 37°C. Each data point represents one bacterial culture. Symbols: ○, NEB123 (wild type); •, NEB124 (moa); □, NEB126 (moa rdgB); ▪, NEB127 (moa nfi); ▵, NEB128 (moa rdgB nfi); ▴, NEB125 (rdgB); ▿, NEB129 (rdgB nfi).

FIG. 4.

(A) SOS induction of E. coli strains mutant for suspected purine base analog-metabolizing enzymes upon HAP exposure. Mid-log-phase cells were exposed to various concentrations of HAP for 1 h at 37°C. (B) Expanded view of induction in three strains. Each data point represents one bacterial culture. Symbols: ○, NO120 (wild type); •, NEB137 (moa); □, NEB139 (moa rdgB); ▪, NEB138 (moa nfi); ▵, NEB140 (moa rdgB nfi); ▴, NEB134 (rdgB); ▿, NEB135 (rdgB nfi).

FIG. 5.

Gradient plate test for HAP sensitivity. The length of a line of cell growth is a measure of the strain's resistance to HAP (MIC). (A) The agar (50 ml) contained a total of 50 μl of a 5-mg/ml HAP solution distributed in a linear gradient increasing from left to right. The strains used (top to bottom) were NEB21, NEB122, NEB152, NEB117, NEB118, NEB55, NEB119, and NEB120. (B) Same experiments as shown in panel A, except that 100 μl of a 5-mg/ml HAP solution was used. (C) The agar (50 ml) contained a total of 250 μl of a 5-mg/ml HAP solution distributed in a linear gradient increasing from left to right. The strains used (top to bottom) were NEB21, NEB117, JB43, NEB118, JB48, NEB55, JB42, NEB119, and JB44. (D) The agar (50 ml) contained a total of 250 μl of a 5-mg/ml HAP solution distributed in a linear gradient increasing from left to right. The strains used (top to bottom) were NEB122, NEB152, JB46, NEB120, JB47, and AB1157.

FIG. 6.

Model for excluding HAP and endogenous purine base analogs from DNA. Symbols: H = HAP, N = any base.