Antimutator Role of DNA Glycosylase MutY in Pathogenic Neisseria Species

Abstract

Genome alterations due to horizontal gene transfer and stress constantly generate strain on the gene pool of Neisseria meningitidis, the causative agent of meningococcal (MC) disease. The DNA glycosylase MutY of the base excision repair pathway is involved in the protection against oxidative stress. MC MutY expressed in Escherichia coli exhibited base excision activity towards DNA substrates containing A:7,8-dihydro-8-oxo-2′-deoxyguanosine and A:C mismatches. Expression in E. coli fully suppressed the elevated spontaneous mutation rate found in the E. coli mutY mutant. An assessment of MutY activity in lysates of neisserial wild-type and mutY mutant strains showed that both MC and gonococcal (GC) MutY is expressed and active in vivo. Strikingly, MC and GC mutY mutants exhibited 60- to 140-fold and 20-fold increases in mutation rates, respectively, compared to the wild-type strains. Moreover, the differences in transitions and transversions in rpoB conferring rifampin resistance observed with the wild type and mutants demonstrated that the neisserial MutY enzyme works in preventing GC→AT transversions. These findings are important in the context of models linking mutator phenotypes of disease isolates to microbial fitness.

Infections caused by Neisseria meningitidis (the meningococcus; MC) and Neisseria gonorrhoeae (the gonococcus; GC), pathogenic members of the genus Neisseria, are associated with significant morbidity and mortality in their exclusive human host. MC and GC residing on mucosal surfaces are exposed to DNA-damaging agents from a potent immune system and also suffer genotoxic stress from endogenous sources, predisposing factors for mutations (29, 36). Mutator strains exhibit an increased spontaneous mutation rate compared to those commonly found in the corresponding wild-type species (17). Such a phenotype is often caused by heritable changes in components of the methyl-directed mismatch repair (MMR) pathway engaged in postreplication repair. However, Richardson et al. (41) demonstrated that only 39% of MC strains exhibiting elevated spontaneous mutation rates could be fully or partially complemented with wild-type mutS or mutL alleles and thus directly linked to defects in the MMR system. Conflicting evidence exists on the association of Dam methylase variants causing hypermutable neisserial strains with enhanced phase-variable capsule switching (4, 21, 40). Clearly, mechanisms other than MMR are implicated in MC mutator phenotypes. Associations between hypermutation and defects in MMR have not yet been reported in the close relative GC.

One of the most frequent forms of oxidative DNA damage is the oxidation product of guanine, 7,8-dihydro-8-oxo-2′-deoxyguanosine (8oxoG) (8). The base excision repair (BER) pathway is probably the cell's major line of defense against the deleterious effects of such DNA damage (45). BER involves the release of modified base residues from DNA by DNA glycosylases that leave abasic (AP) sites in the DNA. The AP site may be further cleaved by an AP-lyase activity inherent of many DNA glycosylases or by an AP endonuclease, leaving a strand break with a deoxyribose phosphate residue at the 3′ end or 5′ end, respectively. DNA glycosylases exist in all species so far investigated, confirming a conserved and important role for BER in protection against DNA damage. The DNA glycosylase MutY is an atypical glycosylase in the sense that it removes a normal base, adenine, from DNA when it is mispaired with 8oxoG, thereby preventing CG→AT transversions (32). 8oxoG mispairs are formed in vivo during DNA replication by two mechanisms, either incorporation of an adenine nucleotide opposite an 8oxoG derived from the direct oxidation in the template strand (8) or misincorporation of an 8oxoG that results from oxidation of GTP in the nucleotide pool (27). In Escherichia coli, MutY acts together with formamidopyrimidine DNA glycosylase (Fpg/MutM) and MutT, comprising the 8oxoG (GO) system, to prevent fixation of mutations caused by 8oxoG (30, 31). Fpg removes 8oxoG when paired with cytosine (48). MutT is a hydrolase that converts 8oxodGTP to 8oxodGMP when present in the nucleotide pool (27), thereby preventing 8oxoG from being misincorporated during replication. Inactivation of each of these genes individually in E. coli confers a mutator phenotype, and E. coli fpg, mutY, and mutT mutants have been reported as weak, moderate, and strong mutators, respectively (13, 30).

MutY belongs to a superfamily of DNA repair proteins hallmarked by a helix-hairpin-helix (HhH) motif involved in nonsequence-specific DNA binding (46). The HhH family includes other DNA glycosylases, such as endonuclease III (Nth) and DNA-3-methyladenine (AlkA). An evolutionary analysis of the HhH superfamily of DNA repair glycosylases performed by Denver et al. shows that MutY is present in most bacteria, many eukaryotes, and nearly 50% of the archaea investigated (9). The crystal structure of E. coli MutY has been solved (15, 19), and the gene has been cloned and characterized in a range of species, including mammals (23, 24, 28, 43). However, none of these species are directly comparable with MC.

In this work, we report the characterization of the neisserial mutY gene and of its gene product, which induces a hypermutable phenotype in both MC and GC when inactivated. The protein encoded by the mutY gene has been overexpressed, purified to homogeneity, and assessed for its activities and substrate specificity. Furthermore, functional phenotypes of MC and GC mutY null mutants were assessed.

MATERIALS AND METHODS

Bacterial strains, plasmids, and DNA manipulations.

The bacterial strains and plasmids employed in this study are listed in Table 1. The E. coli mutY mutant strain GBE943(DE3) was kindly provided by A. L. Lu and W. P. Fawcett, University of Maryland, Baltimore, Md. (24). The mutY gene from MC strains M1080 and H44/76 was amplified by PCR with primers TD5 and TD6 or with primers TD76 and TD124 (Table 2). The mutY-containing DNA fragments were cloned into pBluescript SK+ (pBSK+), creating plasmids pBSK-mutY M1080 and pBSK-mutYH44/76. The plasmids were transformed into E. coli ER2566 and GBE943(DE3) by standard methods. The mutY::Kan^r allele was constructed by inserting a kanamycin resistance gene cassette obtained from pUC4K (kindly provided by Knut Ivan Kristiansen, University of Oslo, Oslo, Norway) into the gene in pBSK+. The pBSK-mutYM1080-Kan^r was transformed into MC strains M1080 and H44/76 as well as GC strain FA1090 (50). The mutY-containing DNA fragment was also cloned into the expression vector pET28b, creating plasmid pET28b-mutYM1080. E. coli ER2566 was used for pET28b-mutYM1080 plasmid propagation.

TABLE 1.

Bacterial strains and plasmids used in this study

Plasmid or strain	Characteristic	Source or reference
Plasmid
pSY6	Contains point mutated copy of DNA gyrase, nalidixic acid resistant	47
pBluescript II SK/KS	General cloning vector, Ampr, abbreviated pBSK+	Stratagene
pET28b	Expression vector based on a T7 promoter-driven system, His tag, Apmr	Novagen
pUC4K	Drug resistance marker encoding aminoglycoside 3′-phosphotransferase	Amersham PB
pBSK-mutYH44/76	pBluescript harboring mutY from MC H44/76	This study
pBSK-mutYM1080	pBluescript harboring mutY from MC M1080	This study
pBSK-mutYM1080-KanR	pBluescript harboring mutY from MC M1080 with inserted kanamycin resistance marker	This study
pET28b-mutYM1080	pET28b harboring mutY from MC M1080	This study
Strain
E. coli
ER2566	Expression host with chromosomal copy of the T7 RNA polymerase gene	New England BioLabs
ER2566-pET28b-mutYM1080	ER2566 expressing Mc M1080 mutY from pET28b	This study
GBE791(DE3)	Wild type	24
GBE943(DE3)	1micA68::Tn10kan	24
GBE943(DE3)-pBSK-mutYH44/76	GBE943(DE3) expressing Mc H44/76 mutY from pBSK+	This study
GBE943(DE3)-pBSK-mutYM1080	GBE943(DE3) expressing Mc M1080 mutY from pBSK+	This study
N. meningitidis
M1080	Serogroup B, isolated in the United States in 1984	14
M1080 mutY mutant	M1080 strain with mutY::Kanr	This study
H44/76	Serogroup B, isolated in Norway in 1976. Currently the source for vaccine development in many countries	20
H44/76 mutY mutant	H44/76 strain with mutY::Kanr	This study
N. gonorrhoeae
FA1090	Wild-type clinical isolate, ATCC 700825, http://www.genome.ou.edu/gono.html	5
FA1090 mutY mutant	FA1090 strain with mutY::Kanr	This study

TABLE 2.

DNA sequences of primers used in this study.

Oligonucleotide	Sequence (5′-3′)	Source or reference
TD5	CCGAGAATACGGCCGTCGG	This study
TD6	GTTCTTGCTGCGGTACCTGC	This study
TD44	GCGCCTTCTCCTTCAACC	This study
TD45	GCAGAGTACGCGTTTGACG	This study
TD76	CGCCGCCTCGAGTTCTAACCTATTACCGTTTAA	This study
TD124	GAAAGGGCTCATGATATTAATGAACACACCCATCCCC	This study
TD135	ATTTCCAAAGGCTTGGGCA	This study
TD168	AAAAACTGTCCGAAGCCCAACAAAACTCT	33
TD169	ATATATTGGACGCGGTCGGGCGTT	33
TD170	TTGATTAATGCAAAACCTGT	33
TD171	TTCGGACCTTCAGGCGTTT	33
TD175	GCCAGCGGATGAGCCGTTCGGAGAAGGGGA	This study
TD176	CTGCACTAACCGCTAGTGCTAATAAGGTTTTCAT	This study
TD177	AATCACGCCACATGCCATACGTGGG	This study
1392	GGCGGCATGACCCCGAGGCCCATC	Eurogen
1393	GATGGGCCTCAGGGTCATGCCGCC	Eurogen
N248	GGCGGCATGACCC8oxoGGAGGCCCATC	Eurogen
T248	GATGGGCCTCCGGGTCATGCCGCC	Eurogen
PR286	GCTCATGCGCAGUCAGCCGTACTCG	Eurogen
I-A	CGAGTACGGCTGACTGCGCATGAGC	Eurogen

DNA isolation, cloning, and Southern hybridization were performed according to standard techniques. DNA sequencing of the mutY alleles was performed by using a Beckman Coulter CEQ 8000 Genetic Analyzer system (Beckman Instruments, Fullerton, Calif.) and an ABI BigDye Terminator v. 3.1 DNA sequencing kit (Applied Biosystems) with the primers listed in Table 2.

SMART RACE determination of mutY and sodC transcriptional start points.

The intergenic region between mutY and adh and the putative promoter region of sodC were analyzed with M. G. Reese's bacterial transcription promoter predictor, which is available at the Berkeley Drosophila Genome Project (http://www.fruitfly.org/seq_tools/promoter.html) (38). Total RNA from MC H44/76 and GC FA1090 was isolated by using a combination of TRIzol (Invitrogen, Carlsbad, Calif.) and RNeasy columns (QIAGEN, Hilden, Germany). The putative transcriptional start sites were experimentally determined by modifying the use of the BD SMART RACE cDNA amplificaton kit (BD Clontech, Franklin Lakes, N.J.). To allow amplification of cDNA from prokaryotes, gene-specific primers TD135 (mutY) and TD177 (sodC) (Table 2) were employed instead of the universal oligo(dT) primer in a switching mechanism at the 5′ end of RNA transcript (SMART) reaction. The 5′ rapid amplification of cDNA ends (RACE) was performed with primers TD175 (mutY) and TD176 (sodC) (Table 2). The complete 5′ sequences of the mutY and sodC cDNAs, and thus mutY and sodC transcriptional start points, were identified by DNA sequencing of the mutY and sodC RACE products.

Purification of the recombinant MC M1080 MutY protein.

E. coli strain ER2566 overexpressing MC MutY encoded by the plasmid pET28b-mutYM1080 was grown in Luria-Bertani medium (LB) containing 25 mM betaine, 0.5 M sorbitol, and ampicillin to a final concentration of 100 μg/ml at 37°C with shaking until the optical density at 600 nm was 0.5. The cells were transferred to 18°C for 30 min before the addition of 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG) and grown overnight. Cells were harvested and washed in lysis buffer (300 mM NaCl, 25 mM Tris, and 10 mM imidazole [pH 8.0]) before mechanical lysis in a French pressure cell press (SLM Aminco, Rochester, N.Y.). The cleared lysate was loaded onto a Ni-nitrilotriacetic acid agarose column (QIAGEN), washed with 300 mM NaCl and 25 mM Tris [pH 7.5], and eluted with 250 mM imidazole. Purified protein was dialyzed against 50 mM HEPES, 50 mM NaCl, and 1 mM EDTA [pH 8.0] overnight and loaded on a HiTrap SP-HP 1-ml column (Amersham Biosciences, Little Chalfont, United Kingdom). Proteins were eluted in a 0 to 100% gradient of buffer A (50 mM HEPES, 50 mM NaCl, and 5 mM β-mercaptoethanol) and buffer B (50 mM HEPES and 2 M NaCl).

Assays for base excision of 8oxoG:C, A:8oxoG, A:C, and AP:A containing duplex oligonucleotides.

Duplex DNA substrates containing a single 8oxoG:C, A:8oxoG, A:C, or AP:A base pair were generated by ³²P 5′-end labeling of oligonucleotides by using T4 polynucleotide kinase (New England BioLabs, Beverly, Mass.) as previously described (11). The oligonucleotide sequences of the DNA substrates are listed in Table 2. DNA glycosylase reactions were performed by mixing purified protein or whole cell extracts with DNA substrate in reaction buffer (70 mM MOPS [morpholinepropanesulfonic acid; pH 7.5], 1 mM EDTA, 1 mM dithiothreitol [DTT], 1 mM EDTA, and 5% glycerol) (1) and incubating at 37°C for 30 min, or other time intervals if appropriate, in a total volume of 10 μl. E. coli K-12 MutY (Trevigen, Gaithersburg, Md.) was included as a positive control. To cleave AP sites generated by the DNA glycosylase reaction, half of the reaction mix was treated with 0.5 M NaOH at 70°C for 10 min and neutralized with 0.5 M HCl. The products of the reactions were analyzed by 20% denaturing DNA sequencing gel and phosphorimaging.

Assays for alkylbase and formamidopyrimidine (FAPY) DNA glycosylase activities.

Generally, DNA glycosylase activity was assayed by mixing purified protein with substrate in a reaction buffer containing 70 mM MOPS [pH 7.5], 1 mM DTT, 1 mM EDTA, and 5% glycerol for 30 min at 37°C. N-[H3]-N-methyl-N′-nitrosourea (MNU; 1.5 Ci mmol⁻¹) was used to prepare alkylated calf thymus DNA (6,000 dpm mg of DNA⁻¹) (39) and poly(dG-dC) (12,000 dpm mg⁻¹) (2). Removal of bases was measured in a total reaction volume of 50 μl containing 7 mg of DNA substrate and 500 ng of enzyme.

Protein extracts of neisserial whole cells for in vivo DNA glycosylase assessment.

MC cells were harvested in phosphate-buffered saline, pelleted, incubated at −20°C overnight, and then thawed on ice and vortexed vigorously. Lysis was obtained by a combination of plasmolysis (84% sucrose in 10 mM EDTA) and lysozyme treatment (50 mM MOPS, 1 mM EDTA, 100 mM KCl, 1 mM DTT, 125 μg of lysozyme/ml) as previously described (44).

Complementation of E. coli GBE943(DE3) mutY::Kan^r.

E. coli GBE791 wild type, and GBE943(DE3) mutY::Kan^r harboring pBSK-mutY M1080, pBSK-mutYH44/76, or pBSK+ were grown overnight with shaking at 37°C in liquid LB containing 100 μg of ampicillin (Sigma, St. Louis, Mo.) if appropriate. The cells were inoculated on LB plates containing rifampin (Sigma) to a final concentration of 60 μg/ml and on LB plates containing ampicillin to a final concentration of 100 μg/ml or on plain LB plates. The ratio of rifampin-resistant cells to the total number of cells yielded the spontaneous mutation rate. The complementation experiments were repeated five times for each strain.

Determination of mutations in rpoB conferring rifampin resistance.

Rifampin-resistant single colonies were propagated overnight in 5% CO₂ at 34°C. Ten individual rifampin-resistant colonies (each) of MC H44/76 mutY mutants, the MC M1080 and GC FA1090 wild type, and mutY mutants and three rifampin-resistant colonies of the MC H44/76 wild type were analyzed. The 230-bp region of MC rpoB exhibiting mutations known to confer rifampin resistance (33) was PCR amplified and sequenced (primers TD168-TD171 listed in Table 2).

Assessment of neisserial spontaneous mutation rate.

Single colonies of MC and GC wild-type and mutY mutant strains grown for 24 to 36 h in 5% CO₂ at 34°C were transferred to 1.5 ml of liquid GC containing 1× Isovitalex (Becton Dickinson Biosciences, Oxford, United Kingdom) and incubated in 5% CO₂ at 34°C for 6 to 8 h. MC M1080 and GC FA1090 wild-type and mutY mutant strains were tumbled during the incubation period. After incubation, agglutinating MC and GC strains were exposed to 80-Hz sonication for 10 s in a sonicating water bath (Branson 2510, Branson, Danbury, Conn.) to achieve an even distribution of cells. All strains were plated on GC solid medium containing 3 μg of rifampin/ml or 1 μg of nalidixic acid/ml as well as plain GC solid medium. The ratio of rifampin-resistant cells to the total number of cells yielded the mutation rate. The assay was repeated 10 times for each strain.

MC sensitivity to alkylation and oxidizing agents.

Liquid cultures of MC H44/76 wild type and the mutY null mutant of even turbidity were inoculated on GC plates. Nonimpregnated paper disks manufactured by Becton Dickinson Microbiology Systems (Cockeysville, Md.) were saturated with 10, 20, and 30 mM hydrogen peroxide (Sigma) or menadione (Sigma); 50, 100, and 150 mM paraquat (Sigma); or 1, 2, and 3% methyl methanesulfonate (Sigma). The paper disks were placed on top of the agar plates inoculated with MC cell suspensions, preincubated for 10 min at room temperature, and then incubated in 5% CO₂ at 34°C for 20 h. The diameter of the inhibition zone was measured.

RESULTS

Characterization of the putative MC mutY gene and flanking regions

The putative MC mutY gene consists of 1,047 bp and the open reading frame (ORF), initiated by a TTG start codon often found in MC genes, contains the predicted DNA binding motif helix-hairpin-helix (HhH) in addition to a conserved catalytic residue and an iron-sulfur (4Fe-4S) cluster (24, 25, 46) (Fig. 1A). The mutY gene also contains three DNA uptake sequences, two of which are found as an inverted repeat in the 3′ end of the coding region, rendering mutY the MC DNA repair gene that exhibited the highest DUS density (6). The gene organization of the mutY flanking regions based on the annotated genome sequences of MC strains Z2491 and MC58 are identical (Fig. 1B) (37, 49). Interestingly, mutY was located in close vicinity to genes encoding two components of the antioxidant system: SodC, the neisserial Cu,Zn superoxide dismutase catalyzing the conversion of superoxide anion to hydrogen peroxide, and Adh, which encodes alcohol dehydrogenase, recently shown to participate in the defense against oxygen radicals (Fig. 1B).

FIG. 1.

Physical maps of the N. meningitidis (MC) mutY inter- and intragenic regions. (A) Physical map of the MC mutY open reading frame containing the DNA binding motifs HhH and 4Fe-4S-cluster (FeS), as well as the catalytic residue (C). The mutY gene also contains three DUS shown as striped boxes. Details concerning the cloning of the MC mutY gene and construction of MC and N. gonorrhoeae (GC) mutY mutants are given in Materials and Methods and Tables 1 and 2. The insertion site for the kanamycin resistance marker is depicted. (B) Physical map of the MC mutY flanking regions, which are identical in MC strains Z2491 and MC58. Interestingly, mutY is located in close vicinity to two genes encoding components of the antioxidant system, namely sodC (Cu,Zn superoxide dismutase; sodC) and adh (alcohol dehydrogenase; adh). The genes edd (phosphogluconate dehydratase) and eda (deoxyphosphogluconate aldolase) are involved in energy metabolism. Black, white, and striped arrows signify genes involved in DNA metabolism, energy metabolism, and cellular processes (detoxification), respectively.

To investigate a possible coregulation of mutY and its flanking genes, adh and sodC, the putative promoter regions of these three open reading frames were predicted. The putative mutY transcription start site producing the highest score (1.0) was confirmed by sequencing SMART RACE products of MC H44/76 and GC FA1090 (Fig. 2A) and was situated almost 350 bp upstream of the mutY TTG start codon. Such a long distance between the transcription site and start codon might suggest a possible misannotated intergenic region or mutY ORF in the MC and GC genome sequences. However, the sequence of the mutY-adh intergenic region in four neisserial genome sequences was verified (37, 49) (http://www.sanger.ac.uk/Projects/N_meningitidis/; http://www.genome.ou.edu/gono.html). Furthermore, the mutY ORF was manually confirmed by the National Center for Biotechnology Information's ORF finder (http://www.ncbi.nlm.nih.gov/gorf/gorf.html). Thus, the mutY intergenic region exhibits rather unusual properties. Prediction based on the upstream adh region produced three high-score (1.0, 0.99, and 0.98) potential transcription start sites in close vicinity to each other (Fig. 2A). Whereas mutY seemed to lack a ribosomal binding site, adh had a perfect consensus for this signature. An analysis of the sodC potential promoter region produced a predicted high score (1.0) for a putative transcriptional start site immediately downstream of the stop codon of the mutY reading frame (Fig. 2B). In addition, a transcriptional start site with a lower score (0.92) was identified closer to the sodC start codon. Since GC lacks sodC, only the transcriptional start site of MC H44/76 sodC could be verified by SMART RACE, producing a product corresponding to the transcriptional start site of the lower score (Fig. 2B). In addition, a consensus ribosomal binding site was identified for sodC. Accordingly, no evidence for cotranscription of mutY and adh or sodC was found.

FIG. 2.

The N. meningitidis (MC) mutY and sodC genes are not cotranscribed. (A) The putative promoter regions of mutY (upper strand) and adh (lower strand) running in opposite directions overlap. The putative mutY transcription site producing the highest score (1.0) was confirmed by sequencing SMART RACE products of MC H44/76 and N. gonorrhoeae (GC) FA1090. The potential adh transcription site was predicted by the production of three high-score (1.0, 0.99, and 0.98) putative sites in close vicinity to each other. The mutY and adh start codons are circled, and putative transcription start sites are boxed. Whereas mutY seems to lack a ribosomal binding site, adh has a perfect ribosomal binding site consensus (marked with black line). (B) The putative promoter region of sodC was predicted, yielding a high score (1.0) putative transcriptional start site (box 1) located right after the stop codon of the mutY reading frame (triangle). In addition, a lower score (0.92) transcriptional start site (box 2) was identified closer to the sodC start site (circled). Since GC lacks the sodC gene, the putative transcriptional start site of sodC could be verified by SMART RACE only in MC H44/76, which produced a product corresponding to the lower score transcriptional start site. A perfect consensus ribosomal binding site (RBS) for sodC is marked with a black line.

Domain conservation of MutY homologues.

To investigate the conservation of MutY in bacterial species, the deduced amino acid sequence of MutY homologues from a panel of human pathogens (Helicobacter pylori J99, Campylobacter jejuni, Salmonella enterica serovar Typhimurium, E. coli O:157, Yersinia pestis, Haemophilus influenzae KW20, Vibrio cholerae, N. meningitidis MC58, Streptococcus pyogenes, and Mycobacterium tuberculosis H37Rv) (downloaded from NCBI) were compared (data not shown). The MutY amino acid sequences from these phylogenetically diverse bacterial species exhibit a high degree of conservation, particularly in the DNA binding motifs and the catalytic essential residue (24, 25, 46); the HhH domain and the 4FE-4S-cluster as well as aspartic acid are present in nearly all alleles assessed. The overall deduced amino acid sequence of MC MutY exhibited 44% identity with E. coli MutY. Comparison of deduced MutY homologues in neisserial species revealed almost identical amino acid sequences in the MC strains M1080, Z2491, MC58, and FAM18 and in the GC strain FA1090 (data not shown).

Cloning, overexpression, and purification of the putative MC mutY gene.

The entire putative mutY gene from MC M1080 was cloned into pBSK+ and pET28b. Recombinant MC MutY protein was overexpressed from plasmid pET28b-mutYM1080 in E. coli ER2566 and purified under native conditions to approximately 98% purity. The recombinant MC MutY protein migrated at 40 kDa in sodium dodecyl sulfate-polyacrylamide gel electrophoresis, corresponding to the size predicted from genome sequence data.

Enzyme activity and substrate specificity of recombinant MC M1080 MutY.

MC MutY DNA specificity was investigated by interacting recombinant M1080 MutY with a panel of oligonucleotide substrates (Table 3). The preferred substrate of recognized MutY homologues (A:8oxoG) (25) as well as other DNA glycosylase substrates (C:8oxoG, C:A mismatch, AP site, alkylation damage, and FAPY residues) were tested. DNA nicking assays detected the base excision of the A:8oxoG (Fig. 3) and A:C substrates as expected (Table 3) (25, 51). Furthermore, time course analysis and various concentrations of substrate demonstrated similar activity profiles of the putative MC MutY and the E. coli K-12 MutY enzymes (Fig. 3). We could not identify any AP lyase activity associated with MC MutY since strand cleavage of A:8oxoG containing a DNA duplex took place only when NaOH was added and the AP substrate was not cleaved. Conflicting evidence concerning the monofunctional property of E. coli MutY exists (25); thus, MC MutY may possess AP-lyase activity under conditions different from those tested in this study. MC MutY exhibited no activity towards substrates containing C:8oxoG FAPY residues in a ³H-labeled poly(dG-dC) substrate or alkylated residues in calf thymus DNA treated with MNU (Table 3). Collectively, these results show that MC MutY has substrate specificity comparable to that of other MutY homologues.

TABLE 3.

Recombinant N. meningitidis M1080 MutY activity on substrates containing different DNA damages^a

Substrate tested	DNA damage category	Activity
A:8oxoG	Mismatch (MutY)	+
A:C	Mismatch (Mismatch repair, MutY)	(+)
8oxoG:C	Oxidized pyrimidine (Fpg)	−
AP substrate	AP site (AP lyase/endonuclease)	−
FAPY substrate	Oxidized pyrimidine (Fpg)	−
MNU substrate	Alkylated substrate (AlkA)	−

^aThe DNA repair proteins known to act on the different substrates are reported in brackets. +, high efficiency on a substrate; (+), low efficiency on a substrate; −, complete lack of activity on a substrate.

FIG. 3.

Recombinant N. meningitidis (Mc) MutY exhibit strong base excision of A:8oxoG. (A) Purified MutY from both MC M1080 and E. coli were tested for base excision towards A:8oxoG at various time intervals. Upper panel, MC M1080 MutY. Lower panel, E. coli K-12 MutY. Lanes 1, substrate; lanes 2, 15 s; lanes 3, 30 s; lanes 4, 45 s; lanes 5, 1 min; lanes 6, 2 min; lanes 7, 4 min; lanes 8, 8 min; lanes 9, 16 min. (B) Recombinant MutY from MC M1080 and E. coli K-12 was tested for base excision of different concentrations of the A:8oxoG substrate. Lanes 1, 6, and 11, 0.1× substrate; lanes 2, 7, and 12, 0.5× substrate; lanes 3, 8, and 13, 1.0× substrate; lanes 4, 8, and 14, 2.0× substrate; lanes 5, 10, and 15, 3.0× substrate. Lanes 1 to 5, substrate only; lanes 6 to 10, MC M1080 MutY; lanes 11 to 15, E. coli K-12 MutY.

Expression of MC and GC MutY in vivo.

To assess the in vivo activity of MC and GC MutY, whole-cell extracts of the wild type and mutY mutants were incubated with the A:8oxoG substrate. The protein extracts from MC and GC wild-type strains clearly excised the adenine in an A:8oxoG substrate in contrast to the protein extracts of mutY mutants (Fig. 4).

FIG. 4.

Neisserial MutY is expressed in vivo. Whole cell extracts of the N. meningitidis (MC) and N. gonorrhoeae (GC) wild type and mutY mutants were tested for activity in a DNA base excision assay using the A:8oxoG substrate. Cells were grown at 37°C in a 5% CO₂ atmosphere. Lane 1, A:8oxoG substrate only; lane 2, MC M1080 purified MutY, positive control (50 ng); lane 3, MC H44/76 wild type; lane 4, MC H44/76 mutY mutant; lane 5, MC M1080 wild type; lane 6, MC M1080 mutY mutant; lane 7, GC FA1090 wild type; lane 8, GC FA1090 mutY mutant.

MC mutY fully complements an E. coli mutY-deficient strain.

The mutY genes cloned from MC strains H44/76 and M1080 were transformed into E. coli GBE943(DE3) mutY::Kan^r to assess whether the putative MC MutY was able to complement a MutY-deficient E. coli strain that exhibited an increased spontaneous mutation rate (Table 4). Complementation of this MutY-defective E. coli with both MC M1080 and H44/76 mutY genes reduced the mutation frequency to wild-type levels or below, implying that MC MutY fully substituted for the lacking MutY functions in preventing spontaneous mutations in E. coli.

TABLE 4.

Spontaneous mutation rates of E. coli mutY mutant GBE943 (DE3) complemented with N. meningitidis H44/76 and M1080 mutY genes or plasmid pBSK+ assessed by rifampin resistance selection^a

Strain	No. of Rifr mutants per 108 CFU
	Median	(± 1 quartile)	Fold increase
GBE749 (wild type)	3.0	(2.5/3.6)	1.0
GBE943 (DE3)-pBSK	44.7	(24.2/53.8)	14.9
GBE943(DE3)-pBSK- mutYH44/76	2.4	(1.0/3.0)	0.8
GBE943 (DE3)-pBSK- mutYM1080	2.1	(1.5/3.9)	0.7

^aThe results are given as the medians of five independent measurements.

Differential repair of 8oxoG in wild-type and mutY mutant neisserial strains.

The occurrence of transitions and transversions in rifampin-resistant MC and GC wild-type and mutY mutant strains was assessed by DNA sequence analysis of the 230-bp fragment of the rpoB gene known to confer rifampin resistance in MC (Table 5). The mutations fell in distinct categories within the different strains sequenced; however, all wild-type strains consistently exhibited C→T transitions, while mutY mutant strains harbored C→A transversions as well as one G→T transversion (Table 5).

TABLE 5.

Position and transition/transversion in rpoB conferring rifampin resistance in N. meningitidis (MC) and N. gonorrhoeae (GC) wild-type and mutY mutant strains

Rifampin-resistant strains analyzed by rpoB sequencing	Nucleotide change conferring rifampin resistance	No. of strains with mutation/total no. tested
MC H44/76 wild type	C103→T	3/3
MC H44/76 mutY	C92→A	2/10
	C103→A	7/10
	C119→A	1/10
MC M1080 wild type	C103→T	10/10
MC M1080 mutY	C64→A	5/10
	C103→A	4/10
	C73→T	1/10
GC FA1090 wild type	C92→T	9/10
	C119→T	1/10
GC FA1090 mutY	C92→A	4/10
	C103→A	6/10

MC and GC mutY mutant strains exhibit elevated spontaneous mutation rates.

The mutation rate of MC and GC strains was assessed by rifampin resistance screening. As the baseline mutation rate of MC H44/76 was low, this important reference strain was chosen to investigate the in vivo significance of the putative MC MutY. The MC H44/76 mutY mutant showed a striking increase in the spontaneous mutation rate compared to the wild type when using both rifampin (63-fold increase) and nalidixic acid (140-fold increase) selection (Table 6). The results were confirmed by testing two additional MC H44/76 mutY mutant strains, which exhibited spontaneous mutation rates 60 to 70 times higher than those of the wild type in the rifampin assay (data not shown). The mutY phenotype was identified across different MC strains as the MC M1080 mutant showed a sixfold increase in spontaneous mutation rate compared to the wild-type strain (Table 6). However, the baseline mutation rate of the MC M1080 wild type was relatively high; thus, concurrent mutations in other genes may influence the spontaneous mutation rate found for the MC M1080 mutY mutant. To investigate the impact of GC MutY deficiency, an FA1090 mutY mutant strain was constructed and assayed for rifampin resistance. The GC FA1090 wild-type strain also had a baseline mutation rate which was higher than that of the MC H44/76 wild type; still, a 20-fold increase of the spontaneous mutation rate of the mutY compared to the wild type was found (Table 6).

TABLE 6.

Spontaneous mutation rates of N. meningitidis (MC) H44/76 and M1080 and N. gonorrhoea (GC) FA1090 wild-type and mutY mutant strains assessed by rifampin and nalidixic acid resistance^a

Strain	No. of Rifr mutants per 108 CFU			No. of Nalr mutants per 108 CFU
	Median	(± 1 quartile)	Fold increase	Median	(± 1 quartile)	Fold increase
Mc H44/76 wild type	<0.1	(0.0/0.3)	1	0.8	(0.6/0.9)	1
Mc H44/76 mutY	25.0	(18.1/35.0)	63b	100.5	(89.8/114.5)	138
Mc M1080 wild type	7.1	(3.5/8.9)	1	ND		ND
Mc M1080 mutY	44.1	(39.7/73.8)	6	ND		ND
Gc FA1090 wild type	0.4	(0.1/0.6)	1	ND		ND
Gc FA1090 mutY	9.3	(8.8/12.3)	23	ND		ND

^aThe results are given as the medians of 10 and 7 independent measurements, respectively, for Rif^r and Nal^r mutants. ND, not done.

^bThe median was reported as <0.1, so the average of 15 independent measurements (= 0.4) was used to calculate fold induction.

Defects in MC MutY do not affect survival under oxidative and alkylating stress.

Subjecting the MC wild-type and mutY mutant strains to oxidative and alkylating stress by exposure to hydrogen peroxide, menadione, paraquat, and methyl methanesulfonate yielded no obvious differences in survival rate under the conditions employed (data not shown).

DISCUSSION

We examined the role of MutY-initiated BER in MC defense against oxidative DNA damage. The nearly ubiquitous nature of MutY and the sequence conservation in human pathogens emphasize the significance of this enzyme.

MC MutY displayed base excision of A:8oxoG and A:C, fully complemented an E. coli mutY mutant strain, and was found to be functionally active in vivo. The main substrate of the E. coli MutY protein has been shown to be A:8oxoG (25). However, the E. coli paradigmatic processes do not necessarily convey the actions and interactions of DNA repair components in other microbial systems (12). MC MutY could be able to contribute to any of the DNA repair functions lacking in or distinguishing MC from E. coli. Recombinant MC M1080 MutY was thus tested for activity towards a panel of DNA glycosylase substrates as well as one alkylated substrate. We found that the MC MutY nicked A:8oxoG and A:C substrates, but not C:8oxoG, FAPY, and MNU substrates, suggesting a catalytic mechanism similar to that found for E. coli MutY.

The relative occurrence of transitions and transversions in rifampin-resistant MC and GC wild-type and mutY mutant strains, where all wild-type strains harbored C→T transitions, while mutY mutant strains exhibited C→A transversions in addition to one G→T transversion (Table 5), suggests disparate mechanisms of repair of 8oxoG-induced damage in wild-type and mutY-negative neisserial strains. Consistent with what has been shown for GO repair in E. coli (13, 32), repair of 8oxoG-induced damage and fixation of mutations in the MC and GC wild type and mutY mutants demonstrated that the MutY enzyme in the pathogenic Neisseria functions in preventing CG→AT transversions.

To further investigate the in vivo impact of neisserial MutY, the spontaneous mutation rates of the MC and GC wild type and mutY mutants were assessed. Strikingly, MC and GC mutY mutants showed a much higher spontaneous mutation rate than their wild-type strains (Table 6). The 60-fold increase in mutagenicity in the MC H44/76 mutY mutants compared to the wild type was confirmed by nalidixic acid resistance showing the mutY mutants to possess a 140-fold increase in spontaneous mutation rate compared to the wild-type strain (Table 6). The MC M1080 and GC FA1090 wild-type strains had higher baseline mutation rates than H44/76, possibly reflecting other concurrent mutations affecting the spontaneous mutation rate. This finding may explain the less-pronounced increase in the mutagenicity of these mutY mutants compared to H44/76 mutY mutants. Nevertheless, all of the neisserial mutY null mutants constructed exhibited spontaneous mutation rates comparable to the highest rates so far described for MC disease isolates with MMR-induced mutator activity (40, 41), indicating an essential impact of MutY in neisserial physiology and pathogenicity. An assessment of components likely to interact with MC MutY was therefore warranted.

The genes sodC and adh are located immediate to mutY (Fig. 1B) and encode components involved in the defense against reactive oxygen species. Since the majority of DNA lesions are caused by oxidative damage (7), an interaction between BER and the antioxidant system may be beneficial. However, no transcriptional coupling of mutY and the sodC or adh gene was detected in SMART RACE experiments (Fig. 2). However, this finding does not rule out the possibility that MutY, SodC, and Adh are coregulated or coact at other levels; i.e., transcription from one promoter might have an effect on the transcription from an adjacent promoter even when genes are organized in opposite directions (52), as are mutY and adh. On the other hand, subjecting the MC wild-type and mutY mutant strains to oxidative and alkylating stress yielded no obvious differences in survival rates under the conditions employed (data not shown), consistent with the finding of Gifford et al. that the mutY in E. coli is not induced by hydrogen peroxide or paraquat treatment (16).

Other DNA repair components expected to interact with MutY are the two GO partners engaged in the defense against 8oxoG-induced DNA damage (31): the DNA glycosylase Fpg and MutT. MC MutY exhibited no obvious functional overlap with Fpg substrates, as C:8oxoG and FAPY lesions have been demonstrated to be the preferred DNA damages repaired by this glycosylase (Table 3) (3, 34, 48). In E. coli MutT scavenges the nucleotide pool for 8oxoG (27). The MC genome harbors at least one MutT homolog (6), and the deduced amino acid sequence gives no indication of a nonfunctional protein. In conclusion, a seemingly fully functional GO system should not render to MutY an atypically large load of oxidative damage to correct.

MutS has been shown to interact with MutY physically and functionally in both E. coli (53) and human cells (18). Moreover, Richardson et al. (40, 41) found that mutS and mutL deficiencies only partly explain the high mutation rate of some MC isolates. As MutY represents a mismatch excision function in removing adenine opposite guanine (32) or cytosine (51) and previously has been shown to overlap with E. coli MMR in vivo (22), a coupling between MC BER and MMR could result in a hypermutator phenotype of mutY mutant strains.

In a larger perspective, MC genomes are highly recombinogenic due to the enormous amounts of repetitive DNA that could enable loss of MMR genes or other potentially important mutator genes. The MC mutY gene contains no less than three DNA uptake sequences (DUS) within its 1,047-bp ORF, resulting in the highest DUS density of all DNA repair genes detected. The MC mutS and mutL genes harbor four DUS each, although their respective ORFs are much longer than that of mutY (6). We hypothesize that the high DUS density could promote the reacquisition of these genes by transformation if they were damaged or deleted (6).

Considering the strong correlation between MC virulence and defects in MMR, BER, by way of MutY, could contribute to the pool of hypermutable neisserial strains. Such a link is further emphasized by the fact that Pseudomonas sp. mutY mutants have been found to exhibit a root-tip hypercolonizer phenotype (10) and that Oliver et al. (35), in the search for Pseudomonas aeruginosa mutators in cystic fibrosis patients, found four mutY mutants. Moreover, DNA glycosylase deficiencies have been associated with human cancer (42) and ageing (26). A search for defective or polymorphic variants of mutY in natural populations of clinically pathogenic MC strains that exhibit a mutator phenotype and are MMR competent is therefore warranted.