The NADH oxidase of Streptococcus pneumoniae : its involvement in competence and virulence

Authors

  • Isabelle Auzat,

    1. Laboratoire d'Enzymologie et de Biochimie structurale du Centre National de la Recherche Scientifique, 91198 Gif-sur Yvette, France.,
    Search for more papers by this author
  • Sabine Chapuy-Regaud,

    1. Laboratoire de Génétique et Physiologie Bactérienne, Centre Hospitalo Universitaire de Rangueil, Université Paul Sabatier, 31403 Toulouse cedex France.,
    Search for more papers by this author
  • Gisèle Le Bras,

    1. Laboratoire d'Enzymologie et de Biochimie structurale du Centre National de la Recherche Scientifique, 91198 Gif-sur Yvette, France.,
    Search for more papers by this author
  • Delphine Dos Santos,

    1. Laboratoire de Génétique et Physiologie Bactérienne, Centre Hospitalo Universitaire de Rangueil, Université Paul Sabatier, 31403 Toulouse cedex France.,
    Search for more papers by this author
  • A. David Ogunniyi,

    1. Molecular Microbiology Unit, Women's and Children's Hospital, North Adelaide, South Australia 5006, Australia.
    Search for more papers by this author
  • Isabelle Le Thomas,

    1. Laboratoire de Génétique et Physiologie Bactérienne, Centre Hospitalo Universitaire de Rangueil, Université Paul Sabatier, 31403 Toulouse cedex France.,
    Search for more papers by this author
  • Jean-Renaud Garel,

    1. Laboratoire d'Enzymologie et de Biochimie structurale du Centre National de la Recherche Scientifique, 91198 Gif-sur Yvette, France.,
    Search for more papers by this author
  • James C. Paton,

    1. Molecular Microbiology Unit, Women's and Children's Hospital, North Adelaide, South Australia 5006, Australia.
    Search for more papers by this author
  • Marie-Claude Trombe

    1. Laboratoire de Génétique et Physiologie Bactérienne, Centre Hospitalo Universitaire de Rangueil, Université Paul Sabatier, 31403 Toulouse cedex France.,
    Search for more papers by this author

Marie-Claude Trombe. E-mail trombe@CICT.fr; Tel. (+33) 561 322 974; Fax: (+33) 561 322 620.

Abstract

A soluble flavoprotein that reoxidizes NADH and reduces molecular oxygen to water was purified from the facultative anaerobic human pathogen Streptococcus pneumoniae. The nucleotide sequence of nox, the gene which encodes it, has been determined and was characterized at the functional and physiological level. Several nox mutants were obtained by insertion, nonsense or missense mutation. In extracts from these strains, no NADH oxidase activity could be measured, suggesting that a single enzyme encoded by nox, having a C44 in its active site, was utilizing O2 to oxidize NADH in S. pneumoniae. The growth rate and yield of the NADH oxidase-deficient strains were not changed under aerobic or anaerobic conditions, but the efficiency of development of competence for genetic transformation during growth was markedly altered. Conditions that triggered competence induction did not affect the amount of Nox, as measured using Western blotting, indicating that nox does not belong to the competence-regulated genetic network. The decrease in competence efficiency due to the nox mutations was similar to that due to the absence of oxygen in the nox+ strain, suggesting that input of oxygen into the metabolism via NADH oxidase was important for controlling competence development throughout growth. This was not related to regulation of nox expression by O2. Interestingly, the virulence and persistence in mice of a blood isolate was attenuated by a nox insertion mutation. Global cellular responses of S. pneumoniae, such as competence for genetic exchange or virulence in a mammalian host, could thus be modulated by oxygen via the NADH oxidase activity of the bacteria, although the bacterial energetic metabolism is essentially anaerobic. The enzymatic activity of the NADH oxidase coded by nox was probably involved in transducing the external signal, corresponding to O2 availability, to the cell metabolism and physiology; thus, this enzyme may function as an oxygen sensor. This work establishes, for the first time, the role of O2 in the regulation of pneumococcal transformability and virulence.

Introduction

Streptococcus pneumoniae is an anaerobic aerotolerant pathogen. This bacterium is mainly extracellular and infects the human body in quite different environments. In the upper respiratory tract, where S. pneumoniae resides as a commensal, it grows under a partial pressure of oxygen (O2) close to that of the atmosphere. When S. pneumoniae develops as a pathogen to cause pneumonia, otitis media and meningitis, it often colonizes regions that are much less oxygenated and, in some cases, almost O2 free. Oxygen is thus a variable feature in the environment of S. pneumoniae, a situation the bacterium may have to adapt to in the absence of haem proteins and cytochromes. A likely adaptation is the defence against oxidative stress, and it is probable that S. pneumoniae possesses enzymes similar to those found in other anaerobic organisms that are induced in environments with a high level of O2. The NADH oxidases are one class of enzymes that are thought to detoxify molecular O2 by catalysing its reduction by NADH into either H2O, or H2O2 (Higuchi, 1992). Indeed, NADH oxidase activity in S. pneumoniae extracts has been reported since the early work of Smith et al. (1959), and has been purified recently as a soluble flavoprotein that reoxidizes NADH and reduces molecular O2 to water (I. Auzat et al., manuscript in preparation).

The presence of one or several NADH oxidases in S. pneumoniae may also be relevant to the metabolic regulation of competence for genetic exchange. In S. pneumoniae, competence under the control of a quorum-sensing oligopeptide, the competence-stimulating peptide (CSP) and its dedicated two-component system ComD–ComE corresponds to a physiological state, distinct from vegetative growth, which develops in exponential cultures (for a recent review see Havarstein, 1998). Competence is induced in bacteria growing in rich medium and is influenced by parameters such as pH (Chen and Morrison, 1987; Trombe et al., 1992) and cation concentration (Fox and Hotchkiss, 1957; Tomasz and Hotchkiss, 1964; Trombe et al., 1992; 1994; Trombe, 1993; Dintilhac et al., 1997). The appearance of competence during the exponential growth phase is associated with a marked metabolic stimulation (Lopez et al., 1989). This metabolic acceleration associated with competence could result from either a higher rate of glucose degradation and/or a more efficient energetic metabolism with a higher yield of ATP per glucose molecule. S. pneumoniae is thought to derive most of its energy from the glycolytic breakdown of glucose into lactate. In the fermentation reaction, at most two ATP molecules are produced per glucose consumed, and lactate is excreted as the by-product of reoxidation of NADH by pyruvate. The presence of NADH oxidase would allow S. pneumoniae to reoxidize a fraction of the glycolytic NADH, using O2 instead of pyruvate, thus improving the efficiency of glucose catabolism (Condon, 1987) and allowing the high ATP level required for the uptake of transforming DNA to be reached (Clave and Trombe, 1989). Although S. pneumoniae has an anaerobic energetic metabolism, spxB, encoding a product homologous to pyruvate oxidase (Spellerberg et al., 1996), producing the energy-rich intermediate acetyl-PO4, provides evidence that it can use molecular O2 to metabolize pyruvate.

The present work describes the molecular and genetic characterization of Nox, an NADH oxidase that uses O2 as a second substrate and produces H2O. The level of Nox was not changed by conditions that triggered competence. However, mutations that abolish NADH oxidase activity have a striking influence on the competence and the virulence of S. pneumoniae, with no significant effect on its growth in the presence of O2. These results suggest that NADH oxidase activity is important for the regulation of genetic exchange and the virulence of this pathogen, possibly by sensing and adapting to O2 concentration in the environment.

Results

Genetic characterization of NADH oxidase

NADH oxidase was purified from S. pneumoniae as described in Experimental procedures and subjected to N-terminal amino acid sequence analysis. Despite ambiguity for the first two cycles, this N-terminal sequence (X-X-I-V-V-V-G-A-N-H-A-G-T- - - -) is homologous to those of the NADH oxidases from Enterococcus faecalis, Serpulina hyodysenteriae and Streptococcus mutans (Fig. 1). These enzymes catalyse the same reaction, utilizing O2 as a second substrate, producing H2O (Ahmed and Claiborne, 1989; Higuchi et al., 1993; Stanton and Jensen, 1993), as does Nox from S. pneumoniae (J.-R. Garel et al., manuscript in preparation). The nucleotide sequence of the gene encoding this NADH oxidase in S. pneumoniae was obtained as described in Experimental procedures. Following our determination of the nucleotide sequence of nox (GenBank accession number AF014458), 98% of the genome sequence of S. pneumoniae was released by The Institute for Genomic Research (TIGR). The 5′ and 3′ regions of nox were located on fragments 4146 and 4294 respectively, with a gap of 48 bases; this corresponds to fragment sp42 of the newly organized contigs.

Figure 1.

. Nox conservation in different pathogens. Alignment of the amino acid sequences deduced from the gene sequences of H2O-producing NADH oxidases from S. pneumoniae (this work), S. pyogenes (released at http://dna1.chem.uoknor.edu/strep.html), S. mutans (GenBank accession number D49951), S. hyodysenteriae (GenBank accession number U19610), E. feacalis (GenBank Accession Number X68847 S45681) and M. genitalium (GenBank, Accession Number U39706). The star at position 44 shows the essential cysteine residue (Ahmed and Claiborne, 1989; Mallett and Claiborne, 1998; this work). The segments translated into primers PDG1 and PDG2 for PCR amplification of the nox probe are shown by white arrows. Boxes in the S. pneumoniae sequence show the positions corresponding to the reverse oligonucleotides P5RV1 and P3RV2 used for direct PCR sequencing, and black arrows show the beginning of the nucleotide sequence thus determined.

Localized mutagenesis of nox and construction of nox mutant strains

In order to obtain strains with non-functional Nox proteins, the nox gene was mutated in vitro in three different ways: (i) insertion of the aphA-3 cassette conferring kanamycin resistance into the unique internal EcoRV site of a 1317 bp internal fragment cloned into pNox2, leading to pTSS1 (Table 1); (ii) mutation of codon 71 to generate the stop codon TAA and a restriction site for AflII; and (iii) mutation of the cysteine codon TGT at position 44 to generate an isosteric serine codon TCC and a restriction site for BamHI. The mutated fragments cloned into the pBluescript vector, generating plasmids pN1 and pN6 respectively (see Experimental procedures and Table 1), were used to transform Cp1015, resulting in allelic exchange with the wild-type gene.

Table 1. . Strains and plasmids (Morrison et al., 1984; Bethesda Research Laboratories, 1986; Short et al., 1988; Peeters et al., 1988).Thumbnail image of

Western blot analysis of crude extracts from these different strains, using polyclonal antibodies directed against Nox, showed that strain Cp8054, carrying the C44 → S mutation, produced the same immunoreactive material as the wild-type strain Cp1015, whereas this immunoreactive material was absent from strain Cp8056, carrying the K71 → stop codon mutation (Fig. 2A and B). No NADH oxidase activity could be measured in crude extracts from any of the four strains, Cp8054 with the C44 → S mutation, Cp8056 with the K71 → stop codon mutation (Fig. 2C) or Cp 8050 and 23450 with disrupted nox (not shown). Therefore, the loss of NADH oxidase activity owing to the single C44 → S mutation in strain Cp8054 is similar to that obtained owing to the absence of Nox protein in strain Cp8056. The absence of detectable NADH oxidase activity in Cp8054 indicates that the C44 residue is part of the active site, as was shown using the corresponding C42 → S replacement in Nox from E. faecalis (Mallett and Claiborne, 1998). The absence of activity in strain Cp8056 and in any nox mutants demonstrates that no other enzyme capable of using NADH to reduce molecular O2 to either H2O or H2O2 was present. This suggests that nox is probably the only gene in S. pneumoniae encoding NADH oxidase activity, or at least the only gene expressed. It should be noticed that two open reading frames (ORFs) showing some homology with Nox were deduced from the sequenc of the S. pneumoniae genome. Sequence alignment revealed that homologies exclude the N part of the protein containing the essential residue C 44.

Figure 2.

. Influence of nox mutations on the amount and activity of NADH oxidase and on the efficiency of competence development throughout growth. A. Mutated nox alleles in strains Cp8054 (C44 → S) and Cp8056 (K71 nonsense). B. Nox level in crude extracts of Cp1015 (1), Cp8054 (2) and Cp8056 (3). Two hundred and fifty nanograms of protein, corresponding to 5 × 106 cfu, was subjected to Western blot analysis; purified Nox (a, 10 ng; b, 100 ng). C. NADH oxidase activity in crude extracts of Cp1015 (1), Cp8054 (2) and Cp8056 (3). D. Competence yield throughout growth of Cp1015 (1), Cp8054 (2) and Cp8056 (3). The results are geometrical means of n independent experiments ± standard deviations.

Growth of nox mutant strains

The growth of all the nox mutant strains under aerobic conditions was the same as that of their corresponding isogenic wild-type strains, Cp1015 or S6, as measured by biomass accumulation, growth rate and viable counts (Fig. 3A illustrates a representative experiment). Cultures contained mainly diplococci, and long chains were absent from the cultures of the nox mutants. Similar growth rates and yields of Cp1015 were observed under aerobic and anaerobic conditions in a medium in which O2 had been depleted by a mixture of carbon dioxide, hydrogen and nitrogen (Fig. 4A and data not shown). This suggests that NADH oxidase is not an essential enzyme and that the reduction of its activity, either by mutation or by limitation of its substrate (O2), did not result in any marked selective disadvantage (or advantage) for vegetative growth.

Figure 3.

. Influence of the C44 → S mutation on the kinetics and efficiency of competence development during exponential growth. A. Biomass accumulation as measured by changes in OD400. Arrows indicate the window in which competence developed. Lower arrows, Cp8054; upper arrows, Cp1015. B. Time profile of transformant recovery. Cp1015 (grey bars) Cp8054 (dark bars). In this representative experiment, the total percentage of transformants in cultures was 12.9 and 3.9 for Cp1015 and Cp8054 respectively.

Figure 4.

. Influence of O2 on bacterial growth, competence yield and cellular Nox level. Biomass accumulation (A) and transformation (B) of strain Cp1015 were measured after incubation under aerobic or anaerobic conditions. Samples of the corresponding cultures (250 ng proteins, 2 × 106 cfu) were then analysed using Western blotting (C). Quantification of the bands gave 30.7 and 31.7 arbitrary units for aerobiosis and anaerobiosis respectively.

Competence development in nox mutant strains

Competence allowing DNA utilization is autoregulated and develops in exponential growing cultures as several (two or three) successive waves. It is possible to accelerate competence induction, so that it develops within 10 min at 37°C, by the addition of extra competence-stimulating peptide (CSP). The influence of the various nox mutations on competence was investigated in bacteria subjected to both growth regimes.

Strains Cp8054 and Cp8056 showed the same threshold activation level (3 ng ml−1) in response to CSP and gave the same transformant yields as Cp1015. This indicates that the absence of NADH oxidase activity did not cause any defect in the response to CSP or in DNA uptake or recombination. However, the transformability level in these mutants showed a 60–80% reduction if extra CSP was not added (Figs 2D and 3B). The time profile of competence development in nox strains was also affected. Transformability appeared as a single sharp peak at the onset of the exponential phase (OD400 = 0.05) and could be observed over a short period of time. This pattern was quite different from that observed in strain Cp1015, the transformability of which appeared much later and lasted much longer, spanning almost the entire exponential phase (between OD400 0.1 and 0.6), with usually two or three competence waves (Fig. 3, A and B as representative). We have verified that the defects in transformability resulted from the defect of NADH oxidase, owing to the mutation in nox. The nox::aphA-3 allele in strain Cp8050 was replaced by the nox+ allele by transforming Cp8050 with pNox2 and selecting KanS transformants. This resulted in a nox+ strain, Cp8052, with a similar competence time profile to Cp1015. Moreover, in a comparative experiment, the yield of transformants for both Cp8052 and Cp1015 reached 6%, in contrast to Cp8050, which yielded only 1% transformants. Therefore, the altered competence development in Cp8050 was indeed a result of nox disruption. We checked, using Western blot analysis, whether variations in the cellular level of Nox might be related to some features characteristic of competence regulation in the wild-type strain (data not shown). This analysis did not reveal significant changes in the cellular Nox level in cultures, whether or not they were activated by the CSP, showing that nox does not belong to the genetic network regulated by CSP.

Influence of oxygen on competence and the cellular level of Nox

NADH oxidase uses molecular oxygen as a substrate. We checked whether oxygen deprivation might confer ‘nox-like’ properties to nox+ bacteria.

The growth of Cp1015 was not significantly altered by the absence of O2, as shown by the similar yield (OD400 measurements; Fig. 4A), viable counts and growth rate, with a doubling time of 32 (± 4) min under aerobic and anaerobic conditions. However, competence was markedly affected by O2. Under anaerobic conditions, the transformation yield of Cp1015 cultures dropped from 9% to 2% (Fig. 4B), a result reminiscent of data obtained with nox mutant strains, suggesting that the reaction catalysed by NADH oxidase is involved in the regulation of competence by O2. It has been shown that this effect is not shared by another oxidase utilizing O2 as a substrate, namely the pyruvate oxidase encoded by spxB (Spellerberg et al., 1996).

A stop codon mutation was introduced into spxB, and strain Cp8400, carrying the mutated allele (see Experimental procedures and Table 1) was evaluated for competence development. Both the kinetics and yield of transformant recovery in Cp8400 were identical to the wild-type strain analysed in parallel (data not shown). This suggested a specific role for NADH oxidase in competence regulation and raised the question of the effect of O2 on the regulation of the level of Nox. Indeed, no significant difference in the amount of Nox was detected between bacteria from aerobic and anaerobic cultures as measured by Western blotting (Fig. 4C). Therefore, the 80% decrease in transformability under anaerobiosis (Fig. 4B) was a result of the absence of O2 and not of the absence of the protein Nox. O2 probably regulates the activity of Nox, but does not change its cellular level, suggesting that O2 availability, via its control of NADH oxidase activity, modulates the development of competence throughout growth.

Attenuation of growth, persistence and virulence of S. pneumoniae in animals

It has previously been suggested that competence development and experimental virulence of S. pneumoniae might be related (Azoulay-Dupuis et al., 1998). The possible involvement of Nox in virulence was investigated in an intraperitoneal model of infection in BALB/c mice by assessing and comparing the survival patterns of mice infected with strain S6 and its isogenic nox insertion mutant, strain 23450, as detailed under Experimental procedures. As shown in Fig. 5A, the median survival time of mice infected with 2.5 × 106 cfu of strain S6 was ≈9 days, whereas no deaths were recorded in the group of mice infected with a similar inoculum of strain 23450 over the 3 weeks of study. This difference is highly significant (P << 0.001), suggesting that the mutation in nox resulted in a severe in vivo attenuation of strain 23450. The difference between the overall rate of survival of mice infected by the different strains is also highly statistically significant (P << 0.005).

Figure 5.

. Growth pattern and virulence of strain S6 and its isogenic nox mutant derivative 23450 carrying the aphA-3 inserted allele. A. Groups of 12 BALB/c mice were challenged intraperitoneally with 2.5 × 106 cfu of the indicated strains. Survival time was recorded and the broken lines indicate the median survival time for each group. B. Groups of 25 BALB/c mice were infected with ≈1 × 103 cfu of either S6 or 23450, and 5 mice, selected at random from each group, were sacrificed at the indicated intervals. The numbers of organisms recovered from the blood and lungs are expressed as the mean (± SEM) log10 cfu values for each bacterial strain.

In a similar experiment, in which the challenge dose of the two strains was increased to 1 × 108 cfu, the median survival time of mice infected with strain S6 was less than 1 day compared with a median survival time of ≈4 days for mice infected with strain 23450 (data not shown). The difference remained statistically significant (P < 0.01), indicating that the loss of NADH oxidase activity in 23450 did indeed attenuate virulence.

In order to obtain further insight into the putative regulation of virulence by Nox, the abilities of both strains to establish a bacteraemic infection and to persist in organs were compared after intraperitoneal infection of mice with ≈1 × 103 cfu of the two strains. Within 24 h, appreciable numbers of both strains were found in the blood and lungs (Fig. 5B) and also in the livers and spleens (data not shown). Organisms were also recovered in the brains of the mice at this time, although counts were lower (data not shown). This indicates that the nox mutation does not have a major impact on establishment of a systemic infection at day 1. However, over the next 4 days, strain 23450 was cleared from the various tissues, the most pronounced effects being observed in the blood and lungs (Fig. 5B and data not shown). Some residual numbers of strain 23450 were recovered on day 3 and day 4 in blood, and on day 3 for lungs, a likely consequence of mouse-to-mouse variation. In contrast, significant numbers of strain S6 were recovered from these sites over the entire period of observation (Fig. 5B). The differences in the numbers of strains S6 and 23450 recovered from blood, lungs, livers and spleens over the 5 days were statistically significant (P < 0.05, Student's t-test). A similar trend was also seen in brains, but total numbers of pneumococci were appreciably lower and the differences between S6 and 23450 did not reach statistical significance (results not shown).

The results show that the absence of NADH oxidase activity did not abolish virulence but that it is required for its optimal expression, and that nox is probably involved in modulating, rather than triggering, virulence, as demonstrated by the increased survival rate of the mice infected with a strain carrying an insertion in nox. Our findings also attest to the requirement for a functional copy of the gene for persistence in tissues, at least within the context of the challenge model used in these studies. These effects of nox disruption were not associated with specific regulation of the level of other putative virulence proteins, including pneumolysin, autolysin (LytA), neuraminidase, hyaluronidase, PspA, PsaA (Paton et al., 1993) and SpsA (Hammerschmidt et al., 1997), as determined by a comparative Western blot analysis of strains 23450 and S6 (data not shown). It is noteworthy that growth and persistence of nox mutants in lungs was very low relative to S6, suggesting an important role for NADH oxidase in this particular niche, which is the primary route for S. pneumoniae infection.

Discussion

Despite an energy metabolism that is predominantly anaerobic in S. pneumoniae, growth/persistence in animals, probably culminating in virulence, and also transformability were significantly influenced by the loss of activity of the NADH oxidase encoded by nox, suggesting that O2 availability is regulating these complex responses.

The ability to reoxidize a fraction of its NADH, using oxygen instead of pyruvate, provided a metabolic advantage for the cell (Condon, 1987). Also, it has recently been found that NADH oxidase overexpression caused a shift from homolactic to mixed-acid fermentation in Lactococcus lactis (Lopez de Felipe et al., 1998). In S. pneumoniae, the metabolic impairment due to nox mutations was not severe enough to slow vegetative growth.

NADH oxidase has been proposed to function as a defence against oxidative stress. In S. pneumoniae, which probably lacks catalase activity (Sneath et al., 1986), the presence of genes encoding superoxide dismutase (SOD) (Poyart et al., 1995; P. Andrew, personal communication), is consistent with the hypothesis that oxygen could be damaging to the cell. Indeed, protein oxidation in Escherichia coli has been implicated in complex physiological responses leading to stasis (Dukan and Nystrom, 1998). There are two ways in which Nox could be involved in defence against oxidative stress. First, reduction of potentially dangerous oxygen to inoffensive water is a detoxification mechanism in itself (Higuchi, 1992). Second, its regulation of competence allows the capture of DNA as a source of nucleotides and DNA fragments for the repair of O2-induced damage to the chromosome.

Culture transformability is significantly modulated by oxygen rather than abolished, although culture-to-culture variations giving inhibition levels ranging between 50% and 100% are frequently measured. First, the absence of oxygen during anaerobic growth caused an average 80% decrease in competence. Second, a similar average decrease resulted from the interruption of oxygen flow into metabolism when the nox gene was mutated. It is tempting to correlate these two observations and conclude that the nox mutations have affected competence because they have induced a ‘metabolic anaerobiosis’ similar to anaerobic growth. This does not exclude other putative mechanisms mediating the bacterial response to O2 in addition to Nox. In any case, changes in culture transformability were not correlated with variations in the cellular level of Nox, but rather with an NADH oxidase activity that was either decreased (by O2 limitation) or absent (after a nox mutation). This is the first report showing that S. pneumoniae transformability is under the control of oxygen and also that O2 control involves a metabolic enzyme. The mechanism of this regulation remains to be elucidated.

The virulence of S. pneumoniae in an intraperitoneal model of sepsis in BALB/c mice was significantly attenuated by a nox insertion mutation. Two lines of evidence have suggested that the virulence of S. pneumoniae might be related directly or indirectly to the regulation of its competence. The first concerns the regulation and genetic organization of lytA, which encodes the major pneumococcal autolysin. LytA is a putative virulence protein (Paton et al., 1993). Competence and LytA-dependent autolysis are triggered by the same environmental factors (Trombe et al., 1992), in agreement with the finding that lytA belongs to a late competence operon (Martin et al., 1995; Pearce et al., 1995). The second relies on the finding that the dmb mutations that alter the kinetics of calcium transport also affect, at the same time, competence regulation, LytA-dependent autolysis (Trombe, 1993; Trombe et al., 1994) and virulence in mice (Azoulay-Dupuis et al., 1998). The marked attenuation of virulence of the nox insertion-mutant strain as compared with wild type could be due either to a metabolic change (via the NAD+/NADH ratio or the ATP yield) or to a weaker defence against oxidative stress in vivo. Interestingly, pyruvate oxidase encoded by spxB has been identified as a determinant of adherence and virulence of pneumococci (Spellerberg et al., 1996). Like NADH oxidase, pyruvate oxidase could be involved in ATP production as it produces acetyl-PO4, a precursor of ATP. In addition, the by-product of pyruvate oxidation (hydrogen peroxide) might trigger more general adaptive responses, allowing growth and persistence in organs and probably culminating in virulence. In any case, the reduced virulence of mutants impaired in their pyruvate oxidase or NADH oxidase activities provides credence for the role of O2 in the virulence of S. pneumoniae mediated by O2-utilizing enzymes, and for the relationship between virulence and oxidative stress and/or energetic metabolism.

Nox proteins that catalyse the reoxidation of NADH by molecular O2 to produce H2O have been found in other organisms, such as E. faecalis (Ahmed and Claiborne, 1989), S. mutans (Higuchi et al., 1993), Streptococcus pyogenes (Mallett and Claiborne, 1998), Giardia duodenalis (Brown et al., 1996) or Serpulina hyodysenteriae (Stanton and Jensen, 1993). These proteins are highly conserved, so that the presence of Nox could also be inferred in Mycoplasma genitalium on the basis of DNA sequence analysis (Fraser et al., 1995). Enzymes with NADH oxidase activities have also been described in Entamoeba histolytica (Lo and Reeves, 1980) and Trichomonas vaginalis (Linstead and Bradley, 1988). All these organisms are anaerobic pathogens that infect parts of the mammalian body in contact with air, such as the gastrointestinal, pulmonary or urogenital tracts, and are thus potentially subject to oxidative stress. Nox might contribute to a more general O2 detoxification/adaptation mechanism in these bacteria, including virulence expression. If the disappearance of their NADH oxidase activity also results in a decrease in their virulence, as observed in S. pneumoniae, then inhibitors targeted against Nox warrant investigation as potential therapeutic agents.

In conclusion, the NADH oxidase coded by nox is the first enzyme described in S. pneumoniae that utilizes molecular O2 as a substrate. The complex phenotype of nox mutants suggests that this NADH oxidase could have a crucial role in the sensing of O2 and in transducing this environmental signal into metabolic changes that affect the global physiology of the bacteria, both in laboratory cultures and in animal hosts.

Experimental procedures

Bacterial strains and culture conditions

All the strains used in this study are listed in Table 1. E. coli was kept and grown under standard conditions (Sambrook et al., 1989). For the pneumococcus, growth media and conditions have been described elsewhere (Clavé and Trombe, 1989). Anaerobic growth was carried out in media placed overnight in an atmosphere comprising 10% O2, 10% H2 and 80% N2 in a Forma Scientific station before inoculation in order to chase dissolved O2, and growth was followed under the same anaerobic conditions. For virulence studies, strain S6 and its nox insertion-mutant derivative (strain 23450) were grown at 37°C overnight on blood agar (supplemented with 90 μg ml−1 kanamycin for strain 23450). Bacteria were then inoculated into meat extract broth with 10% horse serum (serum broth), and grown statically for 4 h at 37°C to give ≈5 × 108 cfu ml−1. Production of type 6 capsule was confirmed by the quellung reaction using antisera obtained from Statens Seruminstitut.

Mutant strains of S. pneumoniae were constructed by allelic exchange obtained by CSP (500 ng ml−1)-induced transformation of the wild-type strains Cp1015 and S6, with non-replicative plasmids carrying the minimal mutated allele or chromosomal DNA from S. pneumoniae, prepared as described by Saito and Masura (1963). Selection of recombinants was based either on the antibiotic resistance trait (rifampicin 2 μg ml−1 and kanamycin 40 μg ml−1) or by restriction analysis of the relevant amplicon.

Purification of NADH oxidase

Purification of Nox, from an exponential batch culture (OD400nm = 0.4) of S. pneumoniae strain Cp1015, was performed at 4°C. The protein with the NADH-oxidizing activity was purified to homogeneity, using the following four steps: (i) ammonium sulphate precipitation, with the Nox activity precipitating between 55% and 80% saturation; (ii) two successive ‘negative’ dye-affinity chromatographies on Matrex Blue A and Red A columns that did not bind Nox activity but retained a large fraction of the other proteins; (iii) an ion-exchange chromatography on QA-Trisacryl, eluting with a salt gradient; (iv) a gel filtration on Ultrogel ACA34. This purification yielded about 0.3 mg of pure Nox per 10 g of cells.

Protein concentration was determined according to Bradford (1976). The N-terminal amino acid sequence was determined using an Applied Biosystems 470A protein sequencer.

NADH oxidase activity measurements

NADH oxidase activity was measured spectrophotometrically by monitoring the initial rate of decrease in NADH absorbance, A340 nm, at 28°C. The assay mixture contained 0.2 mM NADH and the second substrate O2 was provided by air-saturated buffer, composed of 0.1 M Mes [2-(N-morpholino)ethane sulphonic acid], 0.05 M N-ethylmorpholine and 0.05 M diethanolamine, pH 8.2. One unit of NADH oxidase activity corresponds to the oxidation of 1 μmol of NADH per min, using a molar extinction coefficient of 6220 M−1 cm−1 for NADH at 340 nm.

Cloning and mutagenesis of nox and spxB

DNA manipulations were performed as described in Sambrook et al. (1989). Plasmids and oligonucleotides are described in Table 1. The strategy used to obtain the nucleotide sequence of nox was based on the N-terminal amino acid sequence of NADH oxidase and on homologies, and is illustrated in Fig. 1. Briefly, the nucleotide sequence of nox was obtained in two steps. First, an internal 1.3 kb polymerase chain reaction (PCR) fragment was obtained with degenerate oligonucleotides PDG1, 5′-GT(TA)GT(TA)GT(TA)GG(TA)GC(TA)AA (TC)CA(TC)GC(TA)GG, designed from the N-terminal sequence of NADH oxidase of S. pneumoniae and PDG2, 5′-TT (AG)AA(AG)TG(TA)GG(TAC)A(AG)(AG)AA(AG)AA(TAC)A (CAG)A(G)TC, corresponding to the conserved C-terminal residue (Fig. 1) and cloned into pMOSBlue to give pNox2 (Table 1). Second, using direct PCR sequencing employing S. pneumoniae chromosomal DNA as template and reverse primers P5RV1, 5′-AAAATTATCCAACATGGT and P3RV2, 5′-ATTGCAATTAGCATGGGAA, corresponding to a site within the already known sequence and located near the 5′ or the 3′ end respectively (Fig. 1). We could thus determine separately two short sequences that diverged from each of these PCR primers and partially overlapped the 3′ and 5′ terminal segments of the known 1317 bp fragment. These sequences were sufficient to reach the ATG and TTA punctuation codons and cover the entire coding sequence of nox. This coding DNA segment (without its promoter and terminator) was amplified using PCR with chromosomal DNA as template and PNOXATG, 5′-AAACATATGAGTAAAATCGTTGTAGTC and PNOXTAA, 5′-CTTAAGGCCAGATAGCTC as primers (Table 1). The amplified DNA fragment was cloned into pMOSBlue vector to generate plasmid pIANOX1 (Table 1). The nucleotide sequence of the fragment has been verified and can be retrieved from the database (GenBank Accession Number AF014458).

For insertion mutagenesis, the kanamycin-resistance gene, aphA-3, was isolated as a 1.3 kb HincII fragment from pPJ1 (Peeters et al., 1988) and cloned into the EcoRV site of pNox2 (Table 1) to generate the recombinant plasmid, pTSS1 (Table 1). Strains Cp1015 and S6 were transformed with pTSS1 to yield kanamycin-resistant mutants Cp8050 and 23450 respectively, each carrying aphA-3 (Table 1). The nox+ allele was restored in the nox::aphA-3 mutant strain, Cp8050, by transformation with pNox2 in the presence of CSP (500 ng ml−1) (Table 1), and selecting for loss of kanamycin resistance. Fragments with a TAA codon at position 71 and a TCC codon at position 44 (Fig. 2) were generated using PCR using primers PN1, 5′-TTGGAAGCTTAAGGTGCTAAAGTTTACATG and PN6, 5′-CCTAGGATCCGGAATGGCTCTTTGGATTGG respectively, and the complementary primer PNC1, 5′-ATTTGAAGCAAGAGCGATATAGCTTGTATC. The resultant recombinant plasmids pN1 and pN6, containing fragments cloned into the EcoRV or SmaI sites of pBluescriptSK, were used to transform Cp1015 to give strains Cp8056 and Cp8054 respectively (Table 1). The AflII and BamHI restriction sites generated by these mutations allowed the screening of transformed colonies after amplification of the PN4–PN5 fragment, 5′-CCACGCTGGTACAGTATG and 5′-TGCTACTTCATGGTTGTC respectively. Usually 2%–4% recombinant clones were obtained using this process. All the derivatives of nox were analysed by nucleotide sequencing using PNS1, 5′-CCAACAACGGCGATACGGTC and PNS2, 5′-GAACCTGTAGCGAAAATC primers, either in the transforming plasmids or after their transfer into the chromosome.

For spxB (GenBank accession number L39074) mutagenesis a POX1–POX2, 5′-GAGTTTTGAGCAGATTTTTA–5′-GAAAATCAAAGAATAAAGTC, fragment cloned into pTSS4 (Table 1) was obtained with the mutagenic primer POX3, 5′-CACTCAGCTGATTGATGGACGCTTTG, introducing the S34 → stop codon mutation and the PvuII restriction site. The sequence of the mutated allele was verified on the POX3–POX4 5′-GCATCCATGTTCAATTCG amplicon.

Production of antibodies and Western blot analysis

Rabbit antibodies against NADH oxidase were obtained by three successive subcutaneous injections of purified NADH oxidase in complete Freund's adjuvant at 3-week intervals. Three weeks after the last injection, blood was recovered, centrifuged and the serum stored at −80°C. Antibodies against pneumolysin, autolysin, hyaluronidase and neuraminidase were obtained using the same procedure, and antibodies against SpsA were provided by Drs Chhatwal and Hammerschmidt.

For Western immunoblots, fresh cells were lysed with 0.4% sarkosyl. Cell lysates containing 250 ng of protein (corresponding to 5 × 106 cfu in Tris 0.02 M, pH 6.9, 0.2% SDS and 2% glycerol), were subjected to SDS–PAGE, electrotransferred to nitrocellulose, reacted with specific antibodies and stained with the ECL Western Blotting kit (Boerhinger Mannheim). The transfer efficiency was checked by staining the membranes with a Ponceau solution and the reactive bands were quantified with a digital densitometer system.

Competence

Competence was detected using transformation tests performed between 0 and 6 h in cultures growing at 37°C in alkaline cat transformation medium (CTM) (Clavé and Trombe, 1989). At 30 min intervals, 100 μl aliquots were withdrawn and mixed with transforming rif-23 DNA (1 μg ml−1) from Cp1016 (Table 1). After 20 min at 20°C, cells were mixed with blood agar medium and incubated for 2 h at 37°C before adding an agar layer with or without rifampicin. Colonies were counted after overnight incubation at 37°C, and transformation efficiencies were calculated as the ratio of the viable counts on plates with and without rifampicin.

For CSP-induced competence, early exponential growing cultures (OD400 = 0.05–0.1) in CTM at pH 7.0 were activated by CSP in the range 3–500 ng ml−1 for 10 min at 37°C. Competence was measured using transformation tests as described above.

Virulence studies and growth patterns in mice

For these studies, 6- to 8-week-old inbred male BALB/c mice were obtained from the Central Animal House of the University of Adelaide and used at the Animal House of the Women's and Children's Hospital, North Adelaide, South Australia.

In the virulence studies, two sets of experiments were performed in parallel. In the first experiment, groups of 12 BALB/c mice were infected intraperitoneally with ≈2.5 × 106 cfu of either strain S6 or strain 23450. In the second set of experiments, mice were infected with ≈1 × 108 cfu of these strains. The survival of the infected mice was closely monitored for 21 days and death was taken as the end point. Differences in median survival times of the mice were analysed by the Mann–Whitney U-test (one-tailed). The overall survival rates of the mice were compared using the Fisher exact probability test.

In order to assess the abilities of the two strains to persist in vivo and establish a systemic infection, groups of 25 mice were infected intraperitoneally with ≈1 × 103 cfu of either strain S6 or strain 23450. On a daily basis, five mice from each group were exsanguinated from the retro-orbital plexus, then sacrificed by cervical dislocation and the spleen, liver, lungs and brain of all five mice from each group were homogenized separately in sterile phosphate-buffered saline (PBS, pH 7.2) and diluted as necessary. Viable counts were determined by culturing blood, liver, spleen, lung and brain homogenates on blood agar plates overnight at 37°C. Stability of the nox mutation was confirmed by quantitative culture on blood agar containing kanamycin; in all cases, the viable counts obtained were similar to those obtained using blood agar without kanamycin. Differences in the relative number of bacteria recovered from the tissues were analysed, using the Student's t-test.

Footnotes

  1. Part of this work has been presented at the Euroconferences on Bacterial Neural Networks Intercellular Communication and Signal transduction, Dourdan 5–9 May 1999.

Acknowledgements

The authors are grateful for fruitful discussions with Drs A. Claiborne and F. Lederer. This work was supported by grants from CNRS (UPR9063) Université Paris 6 (927–0300), Université Paul Sabatier, Toulouse (J.E.1968), La Fondation pour la Recherche Medicale (FRM), l'Association pour la Recherche contre le Cancer (ARC) and the National Health and Medical Research Council of Australia.

CSP was prepared by Rhône-Poulenc Rorer France, SpsA-antibodies were provided by Drs Chhatwal and Hammersmidt. Experiments requiring anaerobic conditions were performed at Ecole Dentaire, Université Paul Sabatier, Toulouse.

Ancillary