DNA repair systems play a major role in maintaining the integrity of bacterial genomes. Neisseria meningitidis, a human pathogen capable of colonizing the human nasopharynx, possesses numerous DNA repair genes but lacks inducible DNA repair systems such as the SOS response, present in most bacteria species. We recently identified a set of genes upregulated by contact with host cells. An open reading frame having high homology with the small subunit of Escherichia coli exonuclease VII (xseB) belongs to this regulon. The increased sensitivity of a mutant in this coding sequence to UV irradiation, alkylating agent and nalidixic acid demonstrates the participation of this gene in meningococcal DNA repair. In addition, the upregulation of the transcription of this open reading frame upon interaction of N. meningitidis with host cells increased not only the bacterial ability to repair its DNA but also the rate of phase variation by frameshifting. Together these data demonstrate that N. meningitidis possesses an inducible DNA repair system that might be used by the bacteria to adapt to its niches when it is colonizing a new host.
Neisseria meningitidis (Nm) is a normal inhabitant of the human nasopharynx (Sim et al., 2000), thus stressing the role of the interaction with host cells in meningococcal life cycle. Occasionally Nm disseminates from this niche and is responsible for septicaemia and meningitis. Transmission of the meningococci from person to person occurs by direct contact via the respiratory route. Nm must therefore be able to adjust to various environmental conditions. Recently, a group of genes, co-ordinately upregulated when the bacteria interact with host cells, has been identified (Morelle et al., 2003). Upstream of these genes was found a sequence, present in 26 copies in Nm Z5463 genome, which is called REP2. For two of these genes, pilC1 and crgA, the REP2 repeat was shown to carry the transcription starting point which is upregulated during bacterial cell contact (Taha et al., 1996; Deghmane et al., 2000). Mutants in all the genes of this regulon were engineered and none, with the exception of the adhesin PilC1, was defective for bacterial adhesion to host cells. This suggested a role for this regulon in the adaptation to the environment that the bacteria encounter during the early stages of adhesion to host cells. Surprisingly among those genes upregulated during the initial step of adhesion is an open reading frame with a high degree of homology with the Escherichia coli small subunit of exonuclease VII (Exo VII), XseB, thus suggesting that some of the DNA repair mechanisms could be controlled by the interaction of meningococci with host cells. Single-strand exonucleases of E. coli are involved in methyl-directed mismatch repair and are proposed to be required for some forms of recombination repair termed post-replication repair reaction (Viswanathan and Lovett, 1998).
The repair of damaged DNA is a critical process in maintaining bacterial genome integrity. DNA repair mechanisms have been extensively studied in E. coli. In other species like Neisseria, the completion of the genome sequence has allowed direct comparisons of potential DNA repair genes to the E. coli paradigm (Kline et al., 2003). Neisseria have evolved several mechanisms to repair their DNA, such as a small-patch repair system (vsr), a nucleotide excision repair system (Campbell and Yasbin, 1984) (uvr), a mismatch repair system (mutL, mutS), and two recombinational repair systems, the RecBCD and the RecF-like pathway (Koomey et al., 1987; Mehr and Seifert, 1998; Hill, 2000; Stohl and Seifert, 2001; Skaar et al., 2002).
Neisseria lack other DNA repair systems such as photoreactivation (Campbell and Yasbin, 1979), and the two known inducible DNA repair systems: the adaptative response and the SOS response (Black et al., 1998). In E. coli, the adaptative response implicates genes whose products are implicated in repair of alkylation damage. Moreover, in most species like E. coli, the majority of the genes implicated in DNA repair are regulated by the SOS response and are induced after DNA damage (Radman, 1975).
Bacteria have to maintain the integrity of their genome but at the same time have to evolve to allow diversity of populations that are able to face the changing and adverse conditions of the environment. The variability of a genome can result from gene exchange by horizontal transfer, by recombination between two homologous DNA regions or by phase variation. Neisseria species have a high genome plasticity and possess a large number of genes undergoing phase variation. Slippage of DNA polymerase is in part responsible for the occurrence of frameshifting; recently, MutS and MutL have been shown to have a negative effect on frameshift frequences (Richardson and Stojiljkovic, 2001).
As Nm lacks classical inducible system repair pathways, the discovery that a gene having homologies with the E. coli xseB is part of a regulon induced by contact with cells prompted us to study its role in meningococcal DNA repair. In this work, we demonstrate that the upregulation of this gene during bacterial cell interaction not only modifies the ability of meningococci to respond to DNA damages but is also responsible for a small increase in phase variation (P < 0.05). Thus, it suggests that this gene may play a role in the adaptation of the bacteria to its environment when it is colonizing a new host.
Characterization of the transcription of meningococcal genes homologous to E. coli exoVII and recJ
In E. coli, XseB is one of the two subunits of Exo VII. The active enzyme is composed of one large XseA subunit and four small XseB subunits (Vales et al., 1982). A search of the Nm Z2491 genome revealed that open reading frame NMA1575 has 42% amino acid identity and 62% similarity with E. coli XseA, and open reading frame NMA2225 has 47% identity with E. coli XseB. In E. coli, several exonucleases with single-strand activity have a functional redundancy with Exo VII, i.e. recJ, exoI and exoX (Burdett et al., 2001; Viswanathan et al., 2001). This prompted us to look for homologous genes in the meningococcal genome. From the above, the annotated open reading frame NMA1052 of the genome of Z2491 was found to have 65% similarity with E. coli recJ. On the other hand, a careful search did not reveal any significant homology with exoI and exoX. This suggested that Nm has only two exonucleases with single-strand activity.
As mentioned above, NMA2225 is part of a regulon induced during the bacterial cell interactions (Morelle et al., 2003). All the genes of this regulon are preceded by a 150 bp repeat designated REP2. For two genes, pilC1 and crgA, the REP2 sequence has been shown to contain the transcriptional starting point (TSP) which is induced upon contact with host cells. Even though no REP2 sequence was found upstream of the Nm recJ and xseA, we first investigated the transcription of these genes when bacteria adhere to host cells by real-time polymerase chain reaction (PCR). Results are shown in Table 1. Neither Nm xseA nor recJ was regulated by the interaction of bacteria with cells. We then determined the TSP of NMA2225, the Nm xseB-like open reading frame, by primer extension in various conditions (Fig. 1). Two TSPs were located in the REP2 region, one at the 5′ end of the REP2 sequence which is constitutive and one at the 3′ end of REP2 which is only present in cell-associated bacteria and during the initial localized adhesion. It should be pointed out that these constitutive and inducible TSPs are localized in the REP2 sequence at a position similar to that of pilC1 (Taha et al., 1996). These data confirmed that the Nm xseB is specifically controlled by the interaction of the bacteria with host cells, and that as for pilC1 and crgA this regulation is carried by the REP2 sequence.
Table 1. Induction factors of the transcription of three exonuclease genes during interaction of Neisseria meningitidis with HUVECs.
Role of meningococcal Exo VII and RecJ in DNA repair
To assess the role of the open reading frames having homologies with E. coli Exo VII and RecJ in DNA repair, a set of isogenic Nm strains mutated in one of the open reading frames encoding these subunits was constructed (see Experimental procedures). A strain with a mutation in both xseB and recJ was engineered. Furthermore, to rule out a possible polar effect of the mutation in xseB, the gene ispA (homologous to a geranyltranstransferase) which is located downstream of this open reading frame was interrupted and the corresponding mutant used as control together with the parental strain.
The spontaneous mutation rate was measured as the number of bacteria able to become spontaneously resistant to rifampicin. This rate was identical in the wild-type strain and the strains carrying mutation in either xseB and/or recJ (data not shown). Furthermore, transformation and recombination were not affected by a mutation either in xseB and/or in recJ when assessed by the ability to acquire exogenous DNA (data not shown).
We then tested the consequence of these mutations in assays to assess the ability of a strain to repair its DNA: sensibility to nalidixic acid, UV irradiation and alkylating agents (EMS). These assays were performed as described in Experimental procedures. Results are reported in Fig. 2. Mutation in xseB or recJ increased the sensitivity both to nalidixic acid and to UV irradiation, but had little or no effect on bacteria grown in the presence of EMS. After 80 min of exposure to EMS, an xseB mutant behaved like a wild-type strain; on the other hand, the survival of a recJ derivative was reduced by one order of magnitude (P = 0.03). It should be pointed out that, as expected, the xseA mutant had a phenotype identical to that of the xseB strain. In addition, the strain having a mutation in ispA behaved like the wild-type strain (data not shown), thus ruling out a possible polar effect resulting from the insertion of the resistance cassette into xseB. It should be pointed out that in each of the assays the double recJ/xseB mutation had a more pronounced effect than that of a strain carrying a single mutation. For instance, when treated with EMS, the survival of the double mutant was reduced by two orders of magnitude (P = 0.02 after 80 min of exposure to EMS). These results suggest that these genes are redundant for some functions. Altogether these data confirmed that the Nm open reading frames homologous to the xseB, xseA and recJ E. coli have a function consistent with that of the corresponding E. coli single-strand exonucleases (Chase and Richardson, 1977; Viswanathan and Lovett, 1998).
Role of meningococcal host–cell interaction in DNA repair
As the transcription of xseB was induced by the interaction of Nm with cells, we then tested the ability of cell-associated bacteria to repair their DNA. Monolayers of human umbilical vein endothelial cells (HUVECs) were infected as described in Experimental procedures using either the wild-type strain or the xseB mutant. After 2 h of infection, at a time where the induction of xseB is maximal, monolayers were washed to remove non-adherent bacteria, and cells were harvested. The ability of cell-associated bacteria to repair their DNA, after UV irradiation or treatment with alkylating agent, was then tested and compared with that of control bacteria, i.e. bacteria grown in the same medium but without HUVECs. Results of these experiments are reported in Fig. 3. The ability to survive UV irradiation as well as alkylating agent was significantly increased in cell-associated bacteria compared with that of control bacteria. Furthermore, this upregulation disappeared in the xseB mutant. These results demonstrate that in Nm a DNA repair system is upregulated by the interaction of the bacteria with the cells and that this upregulation requires xseB
In order to confirm that the above results resulted from the increased expression of xseB, we engineered a strain carrying an inducible xseB allele designated xseBind. We first quantified the transcription of xseB using various IPTG concentrations and compared it with that of cell-associated colony-forming units (cfu) and bacteria grown in control medium (Fig. 4A). The ratio xseB/xseA of cell-associated bacteria was identical to that obtained with the inducible allele at an IPTG concentration of 0.1 mM. The consequences of the induction of xseB were then tested. In the absence of IPTG, the growth on plates containing nalidixic acid of the strain carrying the inducible allele was similar to that of an xseB mutant (Fig. 4B). At higher IPTG concentrations, there was an increased resistance to nalidixic acid. When the IPTG concentration reached 0.01 mM, the strain expressing the inducible xseB allele had a level of resistance to nalidixic acid identical to that of the wild-type strain. A similar correlation was observed when UV survival was tested (Fig. 4C). Moreover, we analysed the response to EMS of the xseB-inducible mutant in a recJ- background (as single mutations did not have a clear effect), at different concentrations of IPTG (Fig. 4D) and observed a correlation between the concentration of IPTG and survival to EMS. Altogether these data show that the upregulation of meningococcal xseB is responsible for an increased ability of the bacteria to repair its DNA.
Role of meningococcal Exo VII in the occurrence of frameshifts
In E. coli, the Exo VII is implicated in frameshift avoidance (Viswanathan and Lovett, 1998). We therefore carried out experiments in order to determine whether Nm Exo VII fulfilled a similar role. Briefly, hmbR and hpuA/B are the two haemoglobin receptors present in the strain used in this study. Both genes are subject to phase variation; therefore, if one gene is mutated, the rate of off-to-on switching of the other gene can be determined by assessing the ability of the bacteria to grow on a plate where haemoglobin is the only iron source. We determined the rate of hpuAB phase variation in a derivative of Nm Z5463 carrying a mutation in hmbR (hmbR::Spr)) and which has a hpuAB gene in the off phase, as previously described (Richardson and Stojiljkovic, 1999; 2001). No difference in the occurrence of frameshifts was observed between the wild-type strain and the xseB mutant when bacteria were grown on solid or in liquid medium. But, surprisingly, when bacteria were grown on cells, the rate of frameshifting of the wild-type strain was increased by 1.9 when compared with bacteria grown in the control medium (P < 0.05). It should be pointed out that the xseB mutant showed the same rate of frameshifting in both conditions (Table 2). These data were consistently reproduced six times. Supporting the role of the upregulation of xseB in regulating the rate of frameshifting is the fact that the phase variation of the hpuAB gene of the strain containing the xseBind was increased in the presence of 0.1 mM IPTG. To determine whether natural transformation was responsible for that increase of frameshifting, the rate of phase variation was determined in a pilT derivative of the Z5463 hmbR-hpuA/Boff strain, which is therefore not competent for uptake of exogenous DNA (Wolfgang et al., 1998). We observed that this strain behaved as the wild-type strain, i.e. a twofold induction of phase variation after cell contact was observed (Table 2). These experiments reinforced the above data indicating that the rate of frameshifting is increased upon interaction with host cells and in addition suggest that this is an intracellular process that does not need exogenous DNA.
Table 2. Induction of frameshift frequency.
Increase in frameshift frequency between different conditions
Bacteria grown in medium/ cell-associated bacteria
Like other bacteria, Nm has evolved a large panel of DNA repair systems which function to maintain the integrity of its genome (Kline et al., 2003). But in contrast to E. coli, the meningococcus does not possess an SOS system to respond to DNA stresses. Thus, the discovery that a gene implicated in DNA repair (xseB, the small subunit of Exo VII) was induced by contact of Nm on human cells prompted us to study its function in the meningococcus. In this work, we demonstrate that Nm has a DNA repair system that is induced upon interaction with host cells. This induction requires the upregulation of xseB, and is also responsible for an increase of the frequency of phase variation.
In E. coli, the single-strand DNA-specific Exo VII shares redundant functions with three other exonucleases: Exo I, Exo X and RecJ, among which only RecJ is present in Nm. We therefore studied the consequence of a mutation in both RecJ and the Exo VII subunits. Our data revealed that the Nm RecJ and Nm Exo VII may play a role in: (i) the RecBCD pathway which is implicated in repair of lesions caused by nalidixic acid, (ii) the RecF-like pathway that is involved in UV repair and (iii) a pathway that removes lesions induced by EMS. In addition, our results suggest that, as in E. coli, these exonucleases present redundant functions.
Our analysis of the regulation of the transcription of xseA, xseB and recJ during bacteria–cell interaction confirmed that only xseB is upregulated when bacteria interact with host cells. We subsequently demonstrate that the induction of xseB after cell contact increases the ability of the bacteria to repair their chromosome. This hypothesis has been confirmed by the construction of a strain in which the gene xseB is under control of the lac promoter and in which we observed a correlation between the ratio of transcription of xseA to xseB and survival in the presence of nalidixic acid, EMS and UV irradiation treatment (Fig. 4). Biochemical studies have shown that in E. coli, Exo VII is composed of one large subunit (XseA) and four small subunits (XseB) (Vales et al., 1982). One possible explanation for the requirement of the upregulation of xseB to optimize the activity of meningococcal Exo VII is that XseB needs to be overexpressed to obtain the optimal ratio of the two subunits. However, it can not be excluded that this protein has another role.
Despite their role in mismatch repair in E. coli, the single-strand exonucleases mutants did not show a mutator phenotype (Burdett et al., 2001). This is also observed in Nm (data not shown). The hypotheses proposed are either that an unknown exonuclease is able to substitute for the RecJ and Exo VII functions or that the activation of mismatch repair in the absence of the exonuclease gives rise to lethal chromosome damage (Burdett et al., 2001).
Neisseria meningitidis is a bacterium whose genome shows a high plasticity and contains a large number of genes that undergo phase variation (Snyder et al., 2001; Martin et al., 2003), this latter characteristic promoting genetic diversity and facilitating survival in the face of selective pressure within the host. The mechanism used to modify the number of polymorphic repeats is not known but is thought to be linked to slippage of DNA polymerase during replication (Viguera et al., 2001). Mismatch correction has been shown to reduce the occurrence of frameshifting (Richardson and Stojiljkovic, 2001). The mechanism by which this repair pathway can reduce chromosomal sequence changes due to slippage is not understood. However, the fact that the pilT mutant, which is unable to acquire external DNA, behaves like the wild-type strain suggests that frameshifting is an intracellular event that does not need any exogenous DNA. In E. coli Exo VII has been described as important to avoid frameshifting (Viswanathan and Lovett, 1998). In Nm, the basal rate of frameshift mutations in bacteria grown on solid medium was not modified by a xseB mutation. On the other hand, contact with host cells increased the frameshift rate of the wild-type strain by twofold, but had no effect on that of the xseB mutant (Table 2). This result was reinforced by the fact that the rate of frameshifting increased with the upregulation of XseB in the strain carrying an xseBind allele. The role of xseB in phase variation has only been tested using a gene whose phase variation relies on a modification of a stretch of Gs. The role of xseB induction on repeats of tetramers or pentamers remains to be shown.
When Neisseria come in contact with cells, several genes are co-ordinately upregulated, presumably to aid the adaptation of the bacteria to this new environment. Among these is a gene implicated in DNA repair: the small subunit of Exo VII. The necessity for the meningococcus to raise the expression of such a gene may reflect the increase rate of DNA lesions in the bacteria caused by the adverse environment such as oxygen or modification of the pH. Moreover, the increase of phase variation after cell contact could be a benefit for the bacteria as it increases the variation in a population and hence the probability of producing a subpopulation of bacteria well adapted to a given condition.
Because Nm Z2491, whose genome has been completed by the Sanger centre, is not transformable, we opted for Z5463, formerly designated C396, another strain that was isolated from a patient with meningitis in The Gambia in 1983 (Achtman et al., 1988). Z5463 is a naturally transformable serogroup A strain that belongs to the same sequence type as strain Z2491 (Sarkari et al., 1994). Neisseria were grown at 37°C in 5% CO2 on GC Medium Base (GCB) (Difco) containing Kellogg's supplement, or in GC-liquid medium [1.5% Proteose Peptone (Difco), 0.4% K2HPO4, 0.1% KH2PO4, 0.1% NaCl, with 12 µM FeSO4 and Kellog's supplements]. E. coli strain TOP10 (Invitrogen) was used for DNA cloning and plasmid propagation. E. coli were grown on Luria–Bertani (LB) agar or in liquid medium. For antibiotic selection of E. coli strains, kanamycin (Km) and spectinomycin (Sp) were used at 60 µg ml−1, erythromycin (Em) was used at 150 µg ml−1 and chloramphenicol (Cm) at 15 µg ml−1. To select strains derived from meningococcus strain Z5463, the Km concentration was 200 µg ml−1, Sp was used at 75 µg ml−1, Em at 3 µg ml−1 and Cm at 5 µg ml−1.
For rifampicin resistance assays, 108 bacteria were serially diluted and plated onto selective media (GC media with 4 µg ml−1 rifampicin). The ratio of rifampicin-resistant cfu to total viable cfu was calculated.
Mutants were engineered in strain Z5463 either by in vitro transposition mutagenesis as previously described (Pelicic et al., 2000) or by insertion of an antibiotic resistance cassette into an appropriate restriction site. Mutants and wild-type strains were having identical growth curves.
For the construction of an inducible xseB gene, a NotI restriction site was inserted immediately upstream of the Shine-Dalgarno sequence of the xseB gene by PCR. The inducible promoter was then introduced into this site. The oligonucleotides used for the amplification of the two chromosomal fragments of strain NmZ2491 necessary for the introduction of the NotI site were A1XB-rep-Not (5′-GCGGC CGCGGTTGAAACCCCGCCACTTGGA-3′), A2XB-2224 (5′-TCCACAACTGCACCCACCGCGCCGAACCCG-3′), B1XB-SD-Not (5′-GCGGCCGCCCGAAAAGGAAACACGATGAAG -3′) and B2XB-isp-up (5′-TGTGCGGGATTTCGTTTTCA GACGGCAAAA-3′). The pHSX-ermC-lacIOP plasmid (gift of H.S. Seifert) (Seifert, 1997) was cleaved with NotI, to release a 3.1 kb fragment containing ermC, lacIq, and the tandem lac operator promoter sequences, tacOP and UV5OP. This fragment containing the inducible promoter was cloned into the NotI site engineered upstream of xseB. The resulting IPTG-inducible xseB allele was designated xseBind and introduced by transformation into Nm and selected using Em.
Transformation and recombination assays
Homologous recombination was measured by the ability to introduce by transformation the spectinomycin resistance gene of strain NMA0900::Sp into the tested strains (Morelle et al., 2003). The transformation–recombination ratio was calculated as the number of antibiotic-resistant cfu per total viable cfu.
Cell culture and infection
HUVECs (promoCell) were cultured and infected as already described (Morelle et al., 2003). Briefly, for infection, 106 bacteria in exponential growth phase were left in contact with the cells for 30 min, the supernatant was then removed and cells were gently washed with infection medium (IM) [RPMI 1640 medium with glutamax (Life Technologies) supplemented with 10% heat-inactivated FCS]. Monolayers were then washed every hour and fresh medium was added. The monolayers were harvested at different time points.
Bacteria were harvested either as HUVEC-associated cfu or from control medium (IM). RNA was obtained as previously described (Morelle et al., 2003) using Trizol reagent (Life Technologies). The complete removal of DNA was checked by the absence of signal in a PCR using the RNA preparation as a template.
Real-time PCR assays
These assays were performed as previously described (Morelle et al., 2003). Reverse transcription was performed using oligonucleotides complementary of the 3′ end of the genes and the SuperscriptTM II reverse-transcriptase. Reverse-transcription products were diluted 1:10 and 1:100, and real-time PCR was run on an ABI PRISM 7700 apparatus (Applied Biosystems) using SYBR-Green PCR Master Mix, according to the manufacturer's instructions. Transcription was quantified using the comparative threshold cycle (Ct) method (Schmittgen et al., 2000). Experiments were carried out three times and freshly extracted RNA was used each time. The statistical method used to test whether differences observed were significant was the Student's t-test. Differences were considered as significant when P < 0.05.
Primer extension was performed as described (Sambrook et al., 1989). Primer (2.5 µM) (XseB-ext-A: 5′-TTCAAG GCGCGACAAGGCTTCTTCAAA-3′) and 10 µg of total RNA were precipitated with ethanol. Pellets were resuspended in hybridization buffer (40 mM PIPES, 1 mM EDTA, 0.4 M NaCl, 80%), incubated at 80°C for 10 min and left for 1 h at room temperature. The mixture is then precipitated with ethanol and redissolved in 25 µl of reverse transcriptase buffer: 2.5 µl of AMV reverse transcriptase buffer (Finnzymes), 1 mM each dNTP (except dATP), 2 µM [α-35S]-ATP, 40 units of Rnasin (Promega) and 40 units of AMV reverse transcriptase. The reaction was incubated at 42°C for 10 min and then 1 mM dATP was added. The Mix was incubated at 42°C for 30 min, then purified with phenol-chloroform, ethanol precipitated and redissolved in TE. Samples were heated 5 min at 95°C in loading buffer (80% formamide, 10 mM EDTA, 1 mg ml−1 xylene cyanol FF, 1 mg ml−1 bromophenol blue) and analysed by electrophoresis through a 6% polyacrylamide/urea gel.
The size marker was a dideoxy sequencing reaction ladder, which was obtained by sequencing a PCR product amplified with primers XseB-extA (5′-TTCAAGGCGCGACAAGGCT TCTTCAAA-3′) and Upstream-XseB (5′-GCAAGAATTTG GTTTACACCACCTCCA-3′).
Nalidixic acid sensitivity
Meningococci grown for 18 h on plates were resuspended to 107 cfu ml−1 in 0.9% NaCl and 100 µl of dilutions were plated onto GCB plates containing various concentrations of nalidixic acid. Results are the average of duplicate measurements from three independent experiments and are expressed as the ratio: (cfu on plates containing nalidixic acid/cfu on control plates) × 100.
UV radiation survival
Sensitivity to radiation was assayed by plating different dilutions of bacteria onto GCB agar plates and exposing the plates to UV (254 nm) at the indicated fluences (mJ cm−2). Fluence is the radiation dose and correspond to the energy received per unit area. The data points represent three independent experiments, each conducted in triplicate.
Treatment of meningococci with a DNA-alkylating agent
Meningococci were resuspended at a cell density of 107 bacteria per millilitre in 0.9% NaCl. Methanesulphonic acid ethyl ester (EMS) was added to a final concentration of 1%. Meningococci were then incubated at 37°C, at various time points, aliquots were removed, and bacteria were washed once in 0.9% NaCl and plated on GCB plates to assess for survivors (Hill, 2000). Experiments were performed in duplicate and all variants were tested concurrently.
To test the resistance to EMS of cell-associated cfu, the incubation time was 40 min.
hmbR and hpuA/B are the two haemoglobin receptors present in the strain used in this study. Both genes are subject to phase variation; therefore, if one gene is mutated the other gene off-to-on switching rate is measured by the ability of the bacteria to grow on a plate where haemoglobin is the only iron source. We determined hpuAB phase variation frequencies in a derivative of Nm Z5463 carrying a mutation in hmbR (hmbR::Spr)) and which is hpuAB-off. Experiments were performed as previously described (Richardson and Stojiljkovic, 1999; 2001). Bacteria were resuspended at a cell density of 104−105 bacteria per millilitre and different dilutions were plated on GCB plates containing or not 50 µM desferioxamine mesylate and 100 mg ml−1 haemoglobin. In such a media bacteria that do not express HmbR or HpuA/B are unable to grow. The hpuAB off-to-on switching rates are reported as the ratio between the number of colonies on selective plates and the number of colonies on non-selective plates. Six independent experiments were performed, and statistical significance was determined by a Student's t-test.
We thank C.R. Tinsley for stimulating discussions and careful reading of the manuscript. We are grateful to M.A. Lety for help with primer extension. This work was supported by grants from the Institut National de la Santé et de la Recherche Médicale, Université Paris V-René Descartes, and S.M. was a recipient of a fellowship from the Fondation pour la Recherche Médicale.